The URL can be used to link to this page

ST9302-SY 920301 - - HEALTH EFFECTS OF EXPOSURE TO POWERLINE-FREQUENCY ELECTRIC AND MAGNETIC FIELDS ELECTRO-MAGNETIC HEALTH EFFECTS COMMITIEE '<J 1l/ I\" Jv \ ~ 'v:-O. ;/\y~\~ ~ _,J:> '/' \P ~ '" ~ \ c..Sc . ,;). . \; ~"'~~' . rJ\. '\ \) ...r- '\('Cl' I".rf' , ).. \V- \ ( \0- {\ (0 .,L \ ,.-~ ~, J Q ~ ~v ~ "I' \pC PUBUC UTll..lTY COMMISSION OF TEXAS AUSTIN, TEXAS MARCH, 1992 MEMBERS OF THE ELECTRO-MAGNETIC HEALTH EFFECTS COMMFr'FEE Patricia A. Buffier, Ph.D., M.P.H. Richard A. Beauchamp, M.D., B.S.E.E. Committ~ Chairperson Environmental Public Health Physician (Formerly) Ashbel Smith Professor and Bur~u of Disease Control and Epidemiology Professor of Epidemiology Texas DeparUnent of Health University of Texas Health Science Center at Houston Austin, Texas School of Public Health Houston, Texas H. Alan Higgins, M.E., P.E. Manager, Strategic Analysis B. G. Burgess, P.E. Southwestern Pubtic Service Company General Manager - System Engineering Amarillo, Texas Houston Lighting & Power Company Houston, Texas Stephen H. Linder, Ph.D. Associate Professor, Management and Policy Sciences Sttzie B. Kent, M.S.H.P. * University of Texas Health Science Center at Houston Health Physicist School of Public Health (Formerly) Chief, Standards Houston, Texas Development Program Bureau of Radiation Control Milton E. McLain, Ph.D., C.H.P. Texas Department of Health Director, Office of Radiological Safety Austin, Texas Texas A&M Univenity College Station, Texas Gary L. Smith, Ph.D. Chief, Radiological Assessment Program Paul L. Zweiacker, Ph.D. Bureau of Radiation Control Manager, Environmental Planning Texas Department of Health Texas Utilities Services Austin, Texas Dallas, Texas- * Conlribuling, Former Cornmillet Member The Committee gratefully acknowledges the editorial consultation and assistance of: Mary Thorpe Parker, Ph.D. Chief, Ecological Evaluation Program Bureau of Radiation Control Texas Department of Health Austin, Texas and the computer and document handling assistance of: Joe Castleberry Engineering Technician Electric Division Public Utility Commission of Texas Austin, Texas ICopies of this publication have been distributed in compliance with the State Depo~iwr~ Law, and am available for public use through the Texas State Publications Deposiwr~ Prog, ram at the Texas State Library and other state depository libraries. PREFACE The Electro-Magnetic Health Effects Committee has completed its initial examination of the literature and research involving electric and magnetic fields (EMF) and public health. This report is the result of three years of work by the Committee and represents a thorough study and analysis of the EMF issue. This report contains the Committee's review of EMF engineering and exposure assessment, epidemiologic studies, experimental studies, judicial issues, regulatory issues, and policy issues, and includes the Committee's recommendations to the Public Utility Commission of Texas. The conclusions and recommendations in this report represent the consensus of the Committee, and do not necessarily reflect the opinions of the Commission or the Commission Staff. The Committee was originally proposed by a Commission task force that was organized to review the rules, practices, applications, and forms concerning transmission line certification in Texas. The task force identified numerous on-going studies concerning EMF and public health and believed that this issue required additional monitoring by qualified individuals. In February 1988, the task force recommended that the Commission appoint a Committee to study the EMF issue and report its findings annually to the Commission. The Committee met for the first time in January, 1989. The Public Utility Commission of Texas recognized the increase in concerns regarding exposure to EMF and its potential effects on human health. The Commission agreed with the task force recommendations and on April 18, 1988, resolved that a Committee be appointed to study the literature and monitor the research concerning the possible health effects of exposure to electric and magnetic fields. The Commission originally selected seven members and added an eighth member in September 1989. The members of the Committee represent the research community, the public health community, and electric utilities. They hold credentials in medicine, epidemiology, biology, engineering, health physics, bio-statistics, and public policy. The Committee members have served as volunteers and have not been reimbursed by the Commission for travel expenses or for the significant amount of time each member has devoted to this project. The Public Utility Commission of Texas owes the Committee members its sincere thanks and appreciation for the exceptional effort and commitment to this project. Engineering Section Engineering Section Electric Division Electric Division Public Utility Commission Public Utility Commission iii TABLE OF CONTENTS OVERVI EW ..................................................................................................xiii 1, Introduction and Background ...................................................................... xiii 2. Engineering and Exposure Assessment ........................................................ xiii 3. Epidemiologic Studies of EMF Exposure ....................................................... xiv 4. Experimental Studies of EMF Exposures ...................................................... xvii 5. Judicial Issues ..........................................................................................xviii 6. Regulatory Issues ...................................................................................... xviii 7. Policy Issues and Options ........................................................................... xix CONCLUSIONS AND RECOMMENDATIONS ............................................xxi 1. Standards .................................................................................................xxi 2. Siting Criteria ............................................................................................xxi 3. EMF Research ........................................................................................... xxii 4. Public Forum .............................................................................................xxii 5. Education Of The Public ............................................................................. xxii 1.0 INTRODUCTION AND BACKGROUND ............................................1-1 1.1 Introduction ..............................................................................................1-1 1.2 Background .............................................................................................. 1 ol 2.0 ENGINEERING AND EXPOSURE ASSESSMENT ...........................2-1 2.1 Introduction ..............................................................................................2-1 2.2 Summary ..................................................................................................2-1 2.3 Electric and Magnetic Field Fundamentals .................................................... 2-2 2.4 Exposure Assessment Fundamentals ........................................................... 2-19 2.5 Measurements .......................................................................................... 2-23 2.6 EMF Exposure Estimates ............................................................................ 2-30 2.7 Preliminary Field Measurements .................................................................. 2-35 References ............................................................................................................2-39 Table of Contents 3.0 Epidemiology of Health Effects and Exposure to EMF .....................3-1 3.1 Introduction ...............................................................................................3-1 3.2 U.S. Cancer Mortality Rates and Trends ....................................................... 3-8 3.3 Epidemiologic Studies Involving EMF Exposures ............................................ 3-10 3.4 Discussion ................................................................................................3-18 3.5 Conclusions ...............................................................................................3-22 3.6 Recommendations ..................................................................................... 3-22 References ............................................................................................................3-57 4.0 EXPERIMENTAL STUDIES .............................................................. 4-1 4.1 Introduction ...............................................................................................4-1 4.2 Summary .................................................................................................. 4-2 4.3 Effects on Animal and Human Behavior ....................................................... 44 4.4 Cancer ......................................................................................................4-7 4.5 Development and Growth ........................................................................... 4-9 4.6 Endocrine System and Immunity ................................................................. 4-10 4.7 Biological Mechanisms ............................................................................... 4-14 References ............................................................................................................4-23 5.0 JUDICIAL ISSUES ............................................................................5-1 5.1 Purpose ....................................................................................................5-1 5.2 Introduction ...............................................................................................5-1 5.3 EMF Proceedings ....................................................................................... 5-1 5.4 Conclusions ...............................................................................................5-5 6.0 REGULATORY ISSUES ...................................................................6-1 6.1 Introduction and Background ....................................................................... 6-1 6.2 Standards and Limits .................................................................................. 6ol 6.3 General Rationale for Health-Based Exposure Standards ................................ 6-1 6.4 Scientific Basis for EMF Standards .............................................................. 6-2 vi Table of Contents 6.5 Existing Standards ..................................................................................... 6-3 6.6 The Situation in Texas ............................................................................... 6-8 6.7 Conclusions ..: ...........................................................................................6-8 References ............................................................................................................6-9 7.0 POLICY ISSUES AND OPTIONS ..................................................... 7ol 7.1 EMF Policy and Political Institutions ............................................................ 7-1 7.2. Contrasting Interpretations by Science and the Courts .................................. 7-8 7.3. Rhetoric and Pul~lic Interpretation ................................................................ 7-10 7.4 Contending Definitions of the Public Policy Problem ...................................... 7-12 7.5 Conclusion: Multiple Interpretations and Institutional Design .......................... 7-16 References ............................................................................................................7ol 8 GLOSSARY OF TERMS ..............................................................................G-1 APPENDIX A COMPUTER CALCULATION OF ELECTRIC AND MAGNETIC FIELDS ..........................................................................A-1 A. 1 345-KV Transmission Line Configuration ........................................................... A-1 A.2 Corona Electric Field Report .............................................................................. A-2 A.3 Corona Magnetic Field Report ........................................................................... A.4 Transpac Electric Field Report ........................................................................... A-7 A.5 Transpac Magnetic Field Report ........................................................................ A-9 A.6 Expocalc Electric Field Report ........................................................................... A-11 A.7 Expocalc Magnetic Field Report ........................................................................ A-14 A.8 Comparison of Programs' Calculated Results ...................................................... A-17 APPENDIX B - FUNDAMENTALS OF EPIDEMIOLOGY ............................. B-1 B. 1 Epidemiologic Methods .............................................................................. B-1 B.2 Sources and Validity of Data ...................................................................... Bo5 B.3 Comparability and Bias .............................................................................. B-9 B.4 Association or Causation? .......................................................................... B-11 B.5 Statistics: Risk Estimates ........................................................................... B-13 vii Table of Contents B.6 Statistics: p-Values, Confidence Intervals and Significance ........................... B-16 B.7 Statistics: Type II Errors and Power ............................................................ B-19 References ....... '. ....................................................................................................B-2 1 APPENDIX C - RESULTS OF EMF SURVEY .............................................C-1 C.1 Siting ........................................................................................................C-1 C.2 Zoning ......................................................................................................C-6 C. 3 Condemnation ........................................................................................... C-7 C.4 Tort ........~ .................................................................................................C-12 C,5 Other ........................................................................................................C-13 viii TABLE OF FIGURES Figure 1-1. Schematic illustration of the stages in an electrical system used to transfer power from the generator via transmission and distribution lines to an end user .................................................... 1-2 Figure 1-2 The electromagnetic spectrum ............................................................................................ 1-3 Figure 2-1. Alternating sinusoidal wave shape for current or voltage ......................................................... 2-2 Figure 2-2. The electromagnetic spectrum shown by frequency and wavelength ........................................... 2-3 Figure 2-3. Field strength varies with distance from the source according to inverse, inverse.squared or inverse.cubed relationships ..................................................................................................2-4 Figure 2-4. Average diurnal variation of the atmospheric potential gradient ................................................ 2-5 Figure 2-5. A typical three-phase single-circuit AC transmission line ........................................................ 2-5 Figure 2-6. The maximum electric field intend profile for 500-kV, 345-kV, 230-kV, and 138-kV transmission lines ...............................................................................................................................2-6 Figure 2-7. The electric field ellipse at a point in space .......................................................................... 2-7 Figure 2-8. The maximum magnetic field hteral profile for 500-kV, 345-kV, 230-kV, and 138-kV transnilsion lines ..............................................................................................................2-8 Figure 2-9. Electric field profiles at lm above ground for single-circuit 345-kV transmission lines with conductors 63, 53, 43, and 33 feet above the ground. 7. .............................................................. 2-8 Figure 2-10. Magnetic field prof~es at lm above ground for single.circuit 345-kV transmission lines with conductors 63, 53, 43, and 33 feet above the ground ................................................................. 2-9 Figure 2-11. Critical distance (Led) for electric field from a 345-kV transmission line ................................... Figure 2-12. Electric field profiles for phase conductor bundle spacings of g, 18, and 36 inches for a single- ci~uit 345-kV transmission line ............................................................................................ 2-10 Figure 2-13. Magnetic field proftles for phase conductor bundle spacings of g, 18, and 36 inches for a single- circuit 345-kV transmission line ............................................................................................ 2-10 Figure 2-14. Electric field profiles for phase conductor spacings of 17.5, 27.5, and 37.5 feet for a single- circuit 345-kV transmission line ............................................................................................ 2-11 Figure 2-15. Magnetic field profiles for phase conductor spacings of 17.5, 27.5, and 37.5 fcct for a single- circuit 345-kV transmission line ............................................................................................ 2-11 Figure 2-16. Electric field profiles for single-circuit 345-kV transmission lines with flat (horizontal), delta (equilateral) and vertical phase gecmetries ............................................................................... 2-12 Figure 2-17. Magnetic field profiles for single-cirouit 345-kV transmission lines with flat (horizontal), delta (equilateral) and vertical phase geometries ............................................................................... 2-12 Figure 2-18. Residential magnetic field sources include appliances, grounding systems and overhead distribution lines (primary, secondary, and net current) .................................................... · ..........2-14 Figure 2-19. Illustrations of three types of electric field meters: ................................................................ 2-17 Figure 2-21. Diagram of a magnetic field meter .................................................................................... 2-18 Figure 2-22. Lateral profile and plan view of IEEE standardized procedure for conducting survey measurements of the electric and magnetic fields from powerlines ................................................. 2-25 Figuro 2-23. Sample output of magnetic field exposure history from Electric and Magnetic Field Exposure (EMDEX) Meter ............................................................................................................... 2-29 Figure 2-24. Comparison of electric field profile calculated by BPA's CORONA, EPRI's EXPOCALC and APPA's TRANSPAC ......................................................................................................... 2-33 jx Table of Figures Figure 2-25. Comparison of magnetic field prof~e calculated by BPA's CORONA, EPRI's EXPOCALC ar/d APPA's TRANSPAC .......................................................................................................... 2-34 Figure 2-26 ...................................................................................................................................2-36 Figure 2-27 ...................................................................................................................................2-38 Figure 3-1 - Crude Cancer Mortality Rates, 1930-87 MaidFemale/Total, Adults and Children ......................... 3-46 Figure 3-2 - Total Cancer Mortality, Rate~, 1930-87 MaldFemalefYoud, Adults and Children ......................... 3-46 Figure 3-3 - Total Cancer Mortality, 1930-87, Male and Female Adults ...................................................... 3-47 Figure 3-4 - Total Cancer Mortality, Male and Female Children ................................................................ 3-47 Figure 3-5 - Lung Cancer Mortality, 1930-87, Male/FcmaledTotal Adults and Children .................................. 3-48 Figure 3-6 - Cigarette Consumption v. Lung Cancer ............................................................................... 3-48 Figure 3-7 - Total Cancer Mortality (Minus Lung), 1930-87, Male/FemaledTotal Adults and Children ................ 3-49 Figure 3-8 - L~ukemia Mortality, 1930-87, Males/Females/Total Adults and Children .................................... 3-49 Figure 3-9 - Leukemia Mortality, 1930-87, Male and Female Adults (20-+85 yrs) ........................................ 3-50 Figure 3-10 - Leukemia Mortality Ram, 1930-87, Male and Female Children (0-19 yrs) ................................ 3-50 Figure 3-11 - Brain and CNS Cancer Mortality, 1930-87, Males/Females/Total Adults and Children ................. 3-51 Figure 3-12 - Brain and CN5 Cancer Mortality, 1930-87, Male and Female Adults (20-85+ yrs) ...................... 3-51 Figure 3-13 - Brain and CNS Cancer Moratlity, 1930-87, Male and Female Children ..................................... 3-52 Figure 3-14 - Breast Cancer Mortality, 1930-87, Male and Female Adults (20-85+ yrs) ................................. 3-52 Figure 3-15 - Power Consumption v. Cancer Mortality, 1930-87, Male Adults (20-85 + yrs) ........................... 3-53 Figure 3-16 - Power Conaumption v. Cancer Mortality, 1930-87, Female Adults (20-85+ yrs) ......................... 3-53 Figure 3-17 - Power Cort~umption v. Cancer Mortality, 1930-87, Male Children (0-19 yn) ............................. 3-54 Figure 3-18 - Power Consumption v. Cancer Mortality, 1930-87, Female Children (0-19 yrs) .......................... 3-54 Figure 3-19 - Leading Causes of Death in the U.5. During Each Decade from 1900 to 1987 ............................ 3-55 Figure 3-20 - U .5. Life Expectancy at Birth, Main and Females, 1900-1990' .............................................. 3-55 Figure 3-21 - U.5. Popuhtion Distributions, 1900 - 1987 ......................................................................... 3-56 TABLE OF TABLES Table 2.1 - Equivalence Between Magnetic Field Units .................................................................... 2-4 Table 2.4 - 60-Hz magnetic flux densities near various appliances ................................................... 2-13 Table 2.2 - 60-Hz electric field levels at the center of various rooms in a typical U.S. home ............... 2-13 Table 2.3 -Typical 60-Hz electric field levels at 30 cm from 115-V home appliances ........................ 2-13 Table 2.5 - Residential magnetic field source characteristics ............................................................2-16 Table 2.6 - Gaussmeters and Dosimeters: .....................................................................................2-27 Table 2.7 - EMDEX Sampling Intervals ..........................................................................................2-28 Table 3.1 - Childhood Cancers and Residential EMF Exposures ........................................................ 3-24 Table 3.2 - Adult Cancers and Residential EMF Exposures ............................................................... 3-26 Table 3.3 - Total Cancer and Occupational EMF Exposure ............................................................... 3-27 Table 3.4 - Leukemia and Occupational EMF Exposure .................................................................... 3-29 Table 3.5 - Brain/CIS Cancer and Occupational EMF Exposure ......................................................... 3-35 Table 3.6 - Other Sites and Occupational EMF Exposure ................................................................. 3-39 Table 3.7 - Childhood Adverse Effects and Paternal/Maternal EMF Exposure .............................. .- ......3-44 Table 4.1 - Summary of Observations/Conclusions of Experiments to Determine Behavioral Effects of EMF Exposure, as Detailed in Section 4.3 .......................................................... 4-15 Table 4.2 - Summary of Observations/Conclusions of Experiments to Determine Effects of EMF Exposure on Cancer Initiation and Promotion, as Detailed in Section 4.4 .............................. 4-17 Table 4.3 - Summary of Observations/Conclusions of Experiments to Determine Effects of EMF Exposure on Development and Growth, as Detailed in Section 4.5 ...................................... 4-18 Table 4.4 - Summary of Observations/Conclusions of Experiments to Determine Effects of EMF Exposure on Endocrine Ststem Function and Immunity, as Detailed in Section 4.6 ................ 4-19 Table 4.5 - Summary of Observations/Conclusions of Experiments to Determine Effects of EMF Exposure on Biological Mechanisms, as Detailed in Section 4.7 .......................................... 4-22 Table S. 1 - Breakdown of EMF related Proccedings ........................................................................5-2 Table 6.1 - Recent International Standards for 60-Hz Fields ............................................................. 6-3 Table 6.2 - State EMF Standards for Transmission Lines ............................................................... 6-7 Table 7.1 - The Tradeoffs in Responses to Uncertainty .................................................................. 7-3 Table 7.2 - Problem Definitions & Policy Options ............................................................................7-13 xi OVERVIEW Electric and magnetic fields have been the subject of 1. introduction and Background scientific study since the 19th century. Energy On April 18, 1988, the Public Utility Commission of content from EMF is much lower than that from Texas (PUC) established thc Electro-Magnetic ionizing radiation (such as x rays) and is too low to Health Effects CommiUcc for the purpose of causc hcating cffccts. Even so, observations of somc addressing the possible health effects of powerline- biological cffects combined with findings from frequency electric and magnetic fields. Thc cpidemiologic studies havc increased thc public's Committec was charged with thc responsibility for concern about possiblc human health effects. It is researching the literature, monitoring on-going not clear which properties of the EMF environment research, and reporting their findings annually to the among many should bc measured, e.g., field Commission. This CommiUee was established as an intensity, duration of exposure, etc. In addition, it is clear that home appliances can produce magnetic independent review body which has served without fields as strong or stronger than those from compensation. Commitlee members were drawn transmission lines. Nonetheless, the public generally from the research community, the Texas Department views involuntary exposure to be more of a healffi of Health, and suppliers of electric services. They sr~ fiunilisr wiffi the scientific literature on electric hu, u~ than voluntary exposure. Research into this and magnetic fields (EMF) and ~e meffiodology imperrant issue is continuing largely through the employed in this area. This report is the result of efforts of the U.S. Department of Energy and the Electric Power Research Institute. Many their efforts. uncertainties remain. This report concentrates on The report is divided into this Overview and major scientific, regulatory, and judicial aspects of EMF. sections on the following topics: (1) Introduction and Background, (2) Engineering and Exposure 2. Engineering and Exposure Assessment, (3) Epidemiology of Health Effects and Exposure to EMF, (4) Experimental Studies, (S) Assessment /udicisl Issues, (6) Regulatory Issues, and (7) Policy Demonstration of a cause-and-effect relationship Issues and Options. In addition, appendices to the between observed health effects and exposure to text Lr~included. EMF is basically dependent on the accurate Before the 19'70's, healffi issues associated wiffi assessment of exposure to EMF and the resulting electricity were limital to safety issues related to absorbed dose to cells, organs, and body. The electrical shock. Then writers of a few reports from electric and magnetic fields resulting from the Eastern Europe suggested certain health cffeets in everyday use of alternating current arc complex, individuals exposed to electric and magnetic fields. varying in properties such as wave shape, frequency, In the mid-19?0's, the State of New York began a S- harmonic content, and transients (spilces). In vivo year, $S-million EMF research program. By the late laboratory studies on animals and in vitro studies on 1980's, the scientific literature contained many EMF cells, as well as ¢pidemiologic studies, have failed to reports. The uncertainty inhereat in such work has clearly identify any single field exposure parametcr caused public concern because of the suggestion of as a major agent in the induction of adverse health cancer and other health effects. effects. The usual problems associatexl with applying data obtained from laboratory animals to humans are Although public concern over EMF health effects particularly important in the evaluation of EMF has focused principally on transmission fines, such health effects. Furthermore, field measurements of fields are produced by all electrical devices in exposure to electric and magnetic fields are by everyday use. necessity limited to a determination of basic Electric and magnetic fields ire produced by voltage environmental properties. Most exposure differences and current flow changes in electric assessments to date have been based on long-term transmission lines. Electricity generalion and average exposure rates. In this process, important transmission is accomplished in three stages: (1) data may not be recorded, and effects of exposure generation and passage through a step-up (dose) rate may be missed. transformer, (2) transmission through high-voltage The Institute of Electrical and Electronic Engineers lines, and (3) passage through a step-down (IEEE) has established standards for methods used in transformer and transfer to lower-voltage distribution measurement of EMF from power transmission lines. lines. Alternating current (60 Hcflz, or cycles per Development of standard methods for measuring second) is standard in North America. EMF in other environments, such as residences, is xiv Health Effects of Exposure to Powerline Frequency Electric and Magnetic Fields needed. Various instrumcnu to measure EMF are mG nearby. The normal combination of dksuncc avaklable commcrcia~y. These are capable of from an appliance and infrequent use reduces the reliable measurement of individual EMF parameters, possible significance of this source of EMF but no one mcasurcment system exists for completely exposure. characterizing EMF in the environment. Rcgazdless of the imperfcct (and perhaps inaccurate) 3. Epidemiologic Studies of uaturc of current cx~osure as3cssmcnt methods, data EMF Exposure so obtained Lrc esscntial to the scientific evaluation of possible health cffccu, In the laborazory, Epidcminlogy is the study of the incidence and conditions of EMF exposure can be carefully distribution of human discase and injury, controlled. Assessing the exposure of the public to Epidcmiologista organize the study of the complex EMF is, however, besct by a multitude of process of discuc causation in terms of the discase complicating factors that determine the effect of thc agcnt, the cnvironmcnt, ~nd the host. Epidcmiologic ~clds as wcU is actual exposure. This situation studies arc organized into two types: descriptive and causes confusion when an effort is trade to apply analytic. Descriptive cpidcmiologic studies explore causal relationships established in controlled pau. crns of disease in whole populations laboratory studies to human populations, (correhtional studies) or specific subgroups in a population (cross-sectional studies), Analytic studies In situations where it is difficult or impossible to chancterizc subjects th~ do or do not have a m4kc actual EMF measurements, exposure rate specific disease (cue-control studies) or subjects estima~ can be generated by appropriate computer who share a common risk factor for a disrasc (cohort calculations. Rcliablc programs exist for calculation studies). Of major concern in all types of of EMF in the vicinity of power transmission Lines, cpidcmiologic studies is the potentlad for bias and and more capable programs designed to calculate confounding factors. Bias is avoided by stringcn~y magnetic fields in the more complex residential defining subject selection criteria and maintaining indoor cnvLronmcnt arc under dcvclopmenL quality control over mcuurcmcnt procedures. When potential health effects of EMF from Confounding factors arc ~ccounted for by transmission lines u~ evaluated, background EMF understanding the complex interrelationships ne~l___~ to be cormidcrgd, The average naturll between exposure and disease. magnetic field of the earth at Texas latitudes, which In cpidcmiologic studies of EMF and cancer, is static in contrut to such fields in most scientists have alXcmlXcd to dcfmc the incidence and "technologically cahamccd" environments, is around distribution of health effects in populations exposed 500 mi]liGauss (raG), to electric and/or magnetic fields, However, the effectiveness of these studies has boca limited by the The natural electric racial in the atmosphere is 130 volts pot macr near thc carth's surface, As in the use of indirect, imprecise, and/or inaccurate case of the natund magncfic field, the natural electric mcasurgs of exposure, Uncertainty in exposure mcasurcmcnts is magnified by ~c abscnco of a field is essentially static, while electric fields due to plausible biological cffcct mcchazdzm in any EMF- use of ciccrxicity in the home or proximity to power transmission lines Lrc alternating at any, c of 60 cancot association and by the difficulty of formulating a dose-response relationship. No proper Hertz. mcuurc of EMF cxposurc has bccn defined, The magnetic field (flux dcnsity) dirccdy beneath a The exposure assessmc:~t mcthodologies curren~y in 345-kV transmission rmc carrying an average load is use 8xc surrogate or indirect mcuures of exposure, about 130 mG, Design of tnnsmission lines can cxposur~ models, and ~cki mcuur~mcats, |ndircct strongly affect the magnitude of thc EMF gcncrated by the lines, Generally slx~cing, raising ~c height EMF exposure mcasuru which have been used arc of a line above the ground r~duces the strength of wire configuration codes, job titlea, and census codes (indicators of occulmfion)- Exposur~ models based EMF outside the rights-of-w~y, Burying on hiztoricaJ a_,t, have bccn used to project exposure trimsmission cables, however, does not assure a vahes. Field measurements provide screening significant reduction in the exposure to magnctic information for short-term exposures but may not fields, give good indications of average long-term It has reccn~y bccn found that the avcragc magnctic cxposures. field intensity within a U.S. home ranges from 0.5 to Various categories of wiring configurations have 1.0 mG and that arcrage residential clcctric ~clds bccn dcvised by rescarchcrs to substitute as range from 5 to 20 volts per metcr, Operating mcasurcs of exposure in homes. These include, for clcctric applianccs, for example, an clcctric can opcncr, may gcncra~c a magnctic field up to 20,000 example, very high current configurations (VHCC) Overview with Conclusions and Recommendations xv and ordinary low current configurations (OLCC), an epidemiologic study. Risk ratios less than 2.0 are beth of which are dependent on the proximity of a likely to be affected by bias or confounding; risk dwelling to specific types of powerline wiring factors greater than 5.0 are more likely to reflect a true increase in risk. A causal hypothesis may bc configurations. further strengthened when experimental evidence is Job titles have also been used as surrogates for available. Laberatory or experimental studies exposure to EMF. Occupational cpidcminlogic completed under controlled conditions provide studies have focused on telecommunications valtmblc data regarding the generality of a hypothesis workers, electrical engineers, and other occupations that is being considered. considered to be exposed to EMF. However, actual exposures in these groups were hrgely unknown and An essential component of cpidcmiologic evidence in were assessed on the basis of exposure categories. the study of human cancer and its causes is time One study attempted to determine actual occupational trends for various cancers. In the United States, exposures by using portable dosimeten for such data have been compiled by the American measuring individual exposures to EMF. Even Cancer Society back to 1930. Data for the years within a single job category, considerable variability 1930-1987 for various cancer sites including lung, in field exposures was found. leukemia, brain, breast, and total cancers for males and females and adults and children were complied Field measurements have shown some promise when for this report. One of the important findings in used in comparison with wh'ing configurations and these data is the effect of the shifting age distribution for linking spot measurements to 24-hour average of U.S. population. Total cancer moxulity rates magnetic fields. However, a single 24-hour appear to have doubled over the last 50 years, but measurement may yield imprecise results. A model when adjusted for changing age distributions, the based on measurement data seems to provide a better rate increases by only about 209~ over the same time index than the measurements alone. interval. An increase in the size of older age groups Exposure assessment studies are also subject to necessarily leads to an increase in the number of confounders. Subjects may be exposed to people dying of diseases associated with old age, carcinogens in the environment as wcH as to EMF. which include cancer. In addition, the total age- A truc confounder will be related to beth EMF and adjusted cancer mortality rates change dramatically cancer. In one BMF study, traffic density was when lung cancer deaths are subtracted. Mortality studied as an indirect measure of exposure to vehicle rates for cancer minus lung cancer have remained emissions and benzene (both related to cancer), and nearly constant for males and have actually a statistically significant association between canccr decreased for females over the period of study. and traffic density was found. In another study the Of particular impoxuncc to this study are the 'wire code effect' was most pronounced among findings for male adult lcukcmia and brain/central females, older children, those living in multi-family nervous system (CNS) cancers, and male and female housing, disadvantaged porsons, and those whose childhood lcukcmia, brain/CNS, and total cancers. mothers smoked during pregnancy. These outcomes Mortality rates for all these cancer sites were indicate the importance of other factors in correiation undergoing substantial increases prior to the with cancer risk. exponential growth period (beginning in 1945) in In order to assess the results of cpidcmiologic studies U.S. electric power consumption. In general, of EMF, one must consider both internal and mortality rates for thesc cancers began to level off or cxtornal validity. Internal validity is concerned with decline alkcr the period of rapid increase in electric the criteria, procedures, attention to confoundors, power consumption. and chance that go into designing and potforming a In previously published reviews of EMF study. Extornal validity is concerned with how the cpidemiologic studies which are cited in this report, results of a study can be generalized and whether the results from both residential and occupational study addresses the causal nature of the association settings wen: analyzed. Two initial studies of EMF between EMF and disease. and disease were done by Wcrtheimer and Leeper in After confuming that a study is internally valid, 1979 and 1992. In the former study, the authors epidcrniologists follow several guidelines to aid in found a excess of high-current wiring configurations the determination of external validity. These include near former homea of children who had died of strength, consistency, spcci~city, temporality, dose- cancer, and in the htter study, the authors found an response gradient, biological plausibility, coherence increase in adult cancer mortality associated with of evidence, and effect of intervention. The high-current wiring configurations. Other EMF magnitude of risk ratios or strength, for example, reviews included a study of residential childhood can be used to partially assess the external validity of lcukemia and exposure to EMF with a summary odds ratio of 1.33, a residential exposure study of xx Heaah Effects of Exposure to Powerline Frequency Electric and Magnetic Field~ Different rcaponscs and their rationalca itrc tied to prcblcm is the most appropriate, one is lcet to decide different views of what constitutes the key problem largely on the basis of prc-cxisting beliefs and vitlues in the EMF debittc. There hitve bccn itt least four that each of us brings to the EMF issue. different ways to dcfmc the EMF "problem", each In this instancc, thc valuca of experts alone may with distinctivc views of thc scientific cvidcncc, of thc proper role for scicncc to play, and of the proper provide too narrow a basis for lcgitimating one dcfmition of the problem over others. Recognizing perception of risk. More importantly, each this limitation, the Committcc rccomxncnds that, until dcfutition carrica it policy prcscription along with it. scicncc can provide it clcarcr path, statc officials In thc abscncc of a conclusivc body of scientific should engage the public in open discussions of both findIngs that would provide it firm grounding for thc cvidcncc to datc and the public valuca that dcciding which of the four ways of constructing thc in~ucnco its intorprctation. xvi Health Effects of F. xposure to Powerline Frequency Electric and Magnetic Fields childhood cancer of the CNS with an odds ratio of beset with the same problems as thc lcukcmia 2.44, an occupational exposure study for leukemia studies, and half of the studies have produced with a risk estimate of 1.18, an occupational inconclusivc results. Significant results were exposure study for mycloid lcukemia with a risk reported for associations of job catcgories with estimate of 1.46, and scvcral other occupational malignant mclanoma and cyc cancer, but not for exposure studies. Many of thc studies mentinned in testicular cancer. Finally, threc of six studies of the reviews cited in this report iackai precise childhood canccr/pattrnal occupation which were assessments of exposure. One study for male breast eval~sd in this report showed significant results. cancer among tclcphonc workcn reported a Five studies were compl~_-~d_ on adverse reproductive standardizal incidence ratio of 6.$, which has bccn cffccts, and in threc of thcsc studies statistically intcrpracd to lend support to thc proposal that I:MF significant associations wcrc found with spontaneous increases cancer risks by intcrfcring with mclatonin abortion, fraluency of abnormal prcgnancy, and production. Howcvcr, mclatonin production may bc congenital realformation. independently affcctat by the shiftwork of thc Adding to the pubtic's concern over canccr and EMF subjects. has beam thc misusc of cancer cpidcmiologic data. Childhood cancers associatat with residcntial EMF Because cancer incidence dam arc generally exposures wcrc explored in five studies. Two of thc unavailable, limc-trcnd studies are usually based on five studies exploring associations of EMF with total mortality rates. Mortality rates can bc cxpresscd in childhood cancers rcportcd significant associations scvcral ways, but in order to present atruc picture, with odds ratios of 2.22 and 2.10. With regard to two facton affccting such presentations must 5c childhood leukcmia, none of thc studies showat a taken into account: data must bc agc-adjustcd to consistent association with EMF, but two produccd account for shifting age distributions in thc odds ratios of 2.35 and 2.10. Thc findings for population ofthc U.S., and improvements in medical childhood tumor of the C:NS wcre also inconsistcut care which have dramatically dccrcued the with only one study producing a mndcratcly cicvatcd proportion of thc population dying from infcctious odds ratio of 2.86. dise. Ascs must hc healed. Consideration of these two factors produccs a much different vicw of the present Adult canccrs associatcd with residential exposures importance of cancer as a cause of death. In wac evaluatai in five studies. Only two out of the addition, no positive correlation is sccn between age- five studies producai significant results. Onc adjustcd cancer mortality trca~ds and increases in repox*,cd an association of !~MF with totst canccr U.S. electric power consumption, which onc would (odds ratio of 1.28) and associations for lymphomas, expect to sen if an I~MF relation exists. (Lung cancer of the CNS, utcrus, and breast. Thc othcr cancer, of coursc, remains at the top of the list for reported a significant association with lung cancer. cancer mortality.) tn addition, four of the ~vc studies rcportcd weak associations of P-MF with adult lcukcmia. Utilizing cancer statistics, risk managcn in Federal regulatory agancies scck to achicvc protcction of This report evaluates occupational EMF studies in public heslth and the environment while responding association with all cancer sites, leukcmia, tumors of to thc rcquircrnents ofthc Officc of Management and thc CNS, mclanoma, and othcr canccr sites. In Budget, dcfcodin~ thc tcchnological and cconomic addition, this report cvahmtes several studies which feasibility of a proposcd action, and following examined the associations bctwccn paternal legislative mandates. Ultimatcly, risk managers occupations having potential for EMF cxposurcs and respond to spcci~c problems based on assessments childhood cancers and advcnc cffccts on which arc formal using aocclXcd sciunti~c critcria. reproduction. This report also cvaluatcs 15 studies Howcvcr, scvcral mndcls for risk asscssment havc of thc association of occupational exposures with all cvoived. Neat approaches to risk assessment are canccr sites. Bccausc of diffcrcnccs in definitions, being formulatal which recognizc the importancc of mcthndology, and othcr inconsistcncies, it was a scientific approach to risk dccisions. impossiblc to determine any causal relationships. Lcukcmia incidenco among occupationally exposed Historically, thc attcntion surrounding I~MF grew out individuals has bccn given thc most attention, and of public concern in thc 1960's for thc aesthetic and results for this sitc arc suggestivc of a causal nuisance problems relatat to high voltage association. Among 15 studies of lcukcrnia and transmission lines. Reports in thc late 1960's and !~MF, several havc yielded weak, but statistically early 1970's by Soviet scicntists concerning possible significant, results. Howcvcr, problems with health effects of EMF changed the focus of public confounding factors and inaccurate cxposurc concern. Western scicntists failed to conftrm the assessments limit the nscfuinesS of thc lcukcmia Soviet findings, cxccpt that a study in Dcnvcr in the results. Cancer of thc CNS has also received iatc 1970's scorned to confirm the earlicr studies. increased attcntion. Studies of this site have bccn Ncgativc findings did not ease public concerns. Overview with Conclusions and Recommendations xix International organizations such as the World Health evidence upon which to develop a health-based Organization, the International Radiation Protection standard, the Commission may make use of Section Association, and some countries (the United 23.44 of the Public Utility Reguiatory Act, which Kingdom and Australia) have addressed the EMF addresses new construction. Section 23.44 is based issue. These groups have found that the scientific on American National Standards Institute (ANSI) and data suggesting health effects due to long-term National Electric Safety Code (NESC) standards. If cnvlronmental EMF exposure arc not persuasive. EMF standards are issued by ANSI and/or NESC, the Commission could readily adopt them as guides. In the United Statcs, the Congress, scvcral fedcral A question remains, howcvcr, about regulatory agencies, a few institutes, and some national jurisdiction over city-owned utilities in siting associations have performed some preliminax7 work questions. Mother option for the PUC is to defcr to on the EMF issue. Although the federal government the Texas Dcpartmcnt of Hcalth which has the has no clear mandate or authority to take rcguhtory ultimate responsibility . for dcvcloping statcwidc action concerning 60-Hz EMF and the existing hcalth standards. evidence doca not compol immediate action, some federal action has occurred. Congress has hosted hearings to collect testimony on the issue; the U.S. 7. Policy Issues and Options Environmental Protection Agency has conducte4 a The current status of scientific evidence regarding review of EMF scientific literature; thc U.S. EMF health effects is unclear. There is no definitive Department of Energy has maintained a strong indication that EMF exposure can affect health, and rescareh program in thc area of basic EMF science; there are no data that establish convincingly that it the U.S. Department of Transportation is evaluating does not. In fact, as is often the case in situations *maglev' trains; and several other agencies have involving very low probability cause/effect maintained a more limited involvement in the area. relationships, it may not cvcr be possible to prove an Organizations like the National Council on Radiation effect or the lack of an effect. Protection and Measurements and thc American National Standards Institute have not pursued the With rospect to thc EMF health cffects issue, state issue at a rigorous level but may do so when the legislatures fred themsclves in a quandary. scientific results become less speculative. Acceptance of false positivc conclusions may result Associations like the National Association of in a significant expenditure of taxpayers' money and Regulatory Utility Commissioncrs and the divert attention from efforts to seek the truc source Conference of Radiation Control Program Directors of any increased risk. By contrast, not acting on have urged greater federal involvement. faisc negative conclusions is likely to bc interpreted by the public as irrcspons~lc disregard for citizens' At present, the only gencraily applied national safety. Therefore, it seems reasonable to expect standard for EMF- is the National Electric Safety legisiatures to actively support efforts to resolve the Code, which deals with reducing shock hazards from conflict. transmission lines. This code is not intended to provide protection from possible long-term health Rcguhtory agencies normally address scientific uncertainty, such as the EMF health effects question, cffects due to chronic exposure. through procedural mechanisms similar to those used Bccausc of the lack of fedcral leadership on the EMF in the courts and legislatures. The details of the issue, the states have responded individually. The mechanisms vary considerably depending on the result is varied and lacks consistency. The states' nature of the regulator/agency and its legishtivc responses have fallen into four categories: (1) take charter. Political pressures to 'do something' about no action, (2) study and report on the issuc, {3) fund the EMF issue may result directly or indirectly in the rcacaroh, and/or (4) use reguhtory authority to search for reguhwry relief, especially if no action is establish standards. At least one common thread achieved at the judicial or leghiativc levels. runs through these efforts: In the absence of a f~rm dose-response relationship or intended results, no In at least 17 states, legislative or administrativc method for evaluating the benefit of EMF standards agencies have formally considered the possibility of is avaLlabic. In the body of this report, the health effects as a result of EMF exposure. Committee details the actions taken by seven states. Responses rangc from dismissal of the question duc to hck of cvidencc (Wyoming) to codification of Texas powerline siting problems are similar to those formal EMF limits in transmission lines (Florida). in othcr states. In some cases siting permit Courts and legislatures are activcly considering applications have been contcstexl, and the applicants actions in several states. have been taken to court. A health-based standard would have simplified thc siting process by providing design criteria to achieve compliance. Without clear Overview with Conclusions and Recommendations xvii In conclusion, much disagreement exists over the Nonethclcas, the Committee believes that, based on relationship, if any, between EMF and disease. its evaluation of the hboratory and cpidemiologic Available epidemiologic evidence has produced literature, there is at this time no conclusive evidence limited conclusions. Findings related to lcukemia to suggest that EMF due to electric power remain suggestive, and associations with cancer of transmission lines poses a human health hazard. The the CNS and other cancer sites arc inconclusive. CommiRec believes that this conclusion is basically corroborated in other EMF literature summaries and In order to improve the quality of future EMF background rel~rts prt~pared by expert scientific and epiderniologic studies, the Committee often several research panels. r~,ommcndations. The exposed population must be well defined. There should be more than one The following observations can be summarized on reference cohort. More work needs to be done to the basis of the studies evaluated by the Committee: accurately assess the complex nature of EMF The interaction of variables which control actual exposure. New EMF measurement technologies exposure to EMF is poorly understood. need to be explored. The relationship, if any, Undoubtedly, the inconsistencies and contradictions between residenthi wiring configurations and EMF found in the scientific literature arc due, at lust exposure nee. As to be studied. The biological basis partly, to this fact. of any health effects in humans needs further study. Epidcmiologic results should provide guidance for Under certain circumstances, animals and humans new experimental studio. Special care must be can detect and avoid electric fields. However, no taken in future studies to control for confounders and research to date has presented any conclusive to avoid internal inconsistencies. evidence that these fields, detected or not, produce any deleterious and/or long hating impacts on animal 4. Experimental Studies of EMF or human behavior. EXpOSUreS One of the current 'm6dels for carcinogenisis involves two steps, initiation and promotio.n. The The Committee examined the mulls of numerous initiation step involves direct or indirect permanent laboratory experiments, comprising in rive (alive) damage to the cell's genetic material (DNA). studies of EMF effects on animals (e.g., rats, Ionizing radiation and certain chemicals have been baboons) and in vitro (test tube) studies at the identified as cancer initiators. Prom6tinn is cellular level. These studies focused on animal characterized by uncontrolled cell growth (tumor behavior, cancer initiation and promotion, formation) after exposure to an initiator, which developmental and growth effects, endocrine system causes or allows the expression of genetic damage. and immunity and cell-cell (membrane)interactions. Neither electric nor magnetic fields are energetic While the quantity and quality of EMF research have enough to cause damage to DNA, and it is generally improved dramatically in recent years, the EMF accepted that power frequency fields are not cancer effects d~_¢_, base is still in a state of infancy when initiators. However, scientists have suggested that compared to the research literature on other potential EMF may be a cancer promoter. No firm environmental exposure risks. Although hboratory conclusions can be drawn on the promotion theory at studies generally provide a greater opportunity to this time. Hypotheses are only now being advanced. control extraneous variables than do epidemiologic Additional information is clearly nee. xied. and field studies, many opportunities still exist for Most of the EMF studies reviewed by the Committee sources of error to enter into even the best designed found no teratogenic effects during embryonic study. It is possible that the EMF literature, like development or during postnatal growth. A few most scientific literature, contains false positives and studies do show effects. Some show effects only false negatives. The Committee has found that the under 'pulsed' fields, which are not normally scientific literature on EMF contains results of associated with 60-Hz alternating current laboratory studies that were performed under a transmission. Certain studies show effects using one variety of exposure mettics (e.g., frequencies, field animal strain, but no effects with another. A high intensities, exposure duration, earth's static magnetic incidence of effects is observed in the controls of field). Thus, the inconsistencies and contradictions various studies, making interpretation of the data of study findings may be due to unknown errors difficult. Overall, these hboratory studies tend to and/or the numerous aforementioned laboratory lead to the conclusion that there is no proven conditions. This circumstance makes it difficult to detrimental effect on prenatal development or sort through the literature, interpra the evidence, postnatal growth from exposure to EMF. and draw definite conclusions with respect to EMF effects. It has been suggested that exposure to EMF can affect animal immune systems. Whole-animal studies xviii Health Effects of E;cposure to Powerline Frequency Electric and Magnetic Fields have not shown such an effect, but certain cellular Review of the information available on EMF-related studies indicate possible effects. Hypotheses neexl to judicial proceedings shows that, to date, little weight be developed and tested before any definitive has been given to EMF health effects claims by conclusion can be drawn. objectors, intervenors, and plaintiffs. Due to public peroeptions of potential hazards and scientific Several studies suggest that EMF exposure causes interest, however, the I~-MF issue is assured changes in the function of animal endocrine systems. continued involvement in judicial proceedings. To For example, reduction in fight-time mclatonin respond to these continuing concerns, the utilities are production and alteration of biological rhythm have developing strategies including keeping up with bcem recorded in animals cxposod to 60-Hz fields. F. MF-rolated research, complying with regulations Numerous physiological effects duc to melatonin regarding the planning and siting of facilities, reduction have been hypothesized, but the potential surveying pubtic swaghess about I~MF health health effects duc to such reduction needs further effeeti, and developing public education and investigation. information programs. Many in vitro studies have shown no effect on eelIs exposed to BMF, while others have shown positive 6. FtegulatorV Issues effects. Although the results of these studies art complex and inconclusive, a growing number of As powerlinc-frtqueney transmission grids have positive findings imply that, under specific expanded, so have the health concerns of those ennditions, BMF can produce cellular changes. For citizens living, working, or going to school close to example, levels of calcium which is involved in the power lines. Some citizens believe that regulations regulation of numerous physiological processes have are necessary to protect pubtic health. Such been shown to be affected in several test systems. regulations are being contemplated and enacted in The significance of these results is unknown. some states. Although effects have been observed at the cellular Several approaches to regulations can bo considered level, with most being attributed to changes fur controlling power line plaeem.ent. Specific occurring at the cell membrane, the actual circumstances my dictate which approach is used. binphysical and/or bioiogiual mechanism is When adopting rtgulations, a gnvcmment agency unknown. Various mechanism have been may use a standard (an acknowledged criterion for postulated, but all art speculative. More rtseareh is comparison) or a limit (a specified level which is needed to evaluate these mechanism. If a restrictive). Other options art to use a guideline (an mechanism is established at the cellular level, this optional standard or limit) or a criterion statement will support the positive laboratory and (usually a document for making informed decisions cpidemiologic studies. about regulations). Ordinarily, regulations which arc protective of health art based upon health risk 5. ,Judicial !sstles assessments, an approach which takes into account all the evidence and weighs benefits versus risk to Although the BMF health effects issue is still actively assign an aceept~lc level of safety. debated in scientific circles and the pubtic press, it If health-based rtguhtions designed to protect the has been a factor in several types of judicial pubtic or exposed workers art contemplated for proceedings for some time. An increase in judicial transmission fine siting, exptieit health data are proceedings on this issue is expected. As used here, required. At present, bowever, no such data exist, 'judicial' includes siting, zoning, condemnation, and nor is there any other rational approach for setting tort proceedings. The PUC is concerned 0nly with exposure roguhtions tu protea pubtie health. Before transmission line siting considerations. occupational regulations can be adopted, a consistent An early concern about BMF health effects was health effect must be found which is rtlated to a cxpreased during Public Service Commission measure of EMF exposuro (such as frequency, hearings in New York in the mid-1970's on a intensity, or time). The necessary basic g-MF data proposed 765-kV transmission line. Since then, would then be combined with the so-called 'healthy many proceedings have involved presentation of worker' criteria whioh define possible exposure time evidence relative to the !~MF health effects question. on the job and basic human physiological quantities. Over 200 proceedings involving BMF eases related Similarly, biological evidence, qtumti~cation of dose, to power transmission and substations have been and risk assessment information must bc available to rel~rted. Of more local interest, nine Texas electric set regulations for populations. Any regulations utilities have reported one or more proceedings written in the absence of the mentioned data would whero EMF or other health effects issues were offer no protection and could possibly hinder further raised. investigations into real health effects. Overview with Conclusion. rand Recommendation~ xxi CONCLUSIONS AND RECOMMENDATIONS The foGowing arc ~c Commitxec's overall false sense of security, diverting resources from conclusions and recommendations regarding evaluating a genuine risk as.sociar~cd with some other standards, siting criteria, research, and public environmental factor. education. The Committee concludes that r~gulatory activities should be divorced from the EMF issue, at this time, 1. Standards and that the Public Utility Commission of Texas (PUC) take action regarding the EMF health effects 1.1 CO!'ICJUSiOI"tS issue only when, or if, action can be justified on a public health basis. If such action is required, the The Committee has examined much of the current Committee concludes that the issue bc referred to the EMF scientific literature. Many cpidcmiologic Texas Department of Health (TDH), since the PUC studies have investigated the possibility of an does not have authority over all EMF sources (e.g., association between disease and ruidencc near appliances, home wiring). The TDH is the state installations transmiU~ng electricity. Epidcmiologic agency with authority in health matters. studies have most frequently investigated the poulbnlty of an as,oclatiun b,~en variou, types of 1.2 Recommendations cancer and exposure to EMF. To date, the results from these epidcmiologic studies have been The CommitRe recommends that neither the PUC inconsistent and inconclusive. nor other statg authorities atzcrnpt to set EMF standards through guidelines, regulations, or The results of the !aboratoi'y studies evaluated by the legislation. Committee arc also inconsistent and in some cases inconclusive. However, it is apparent that under Should new evidence cmcrgc establishing a clear specific exposure conditions, biological changes do association of human health effects from EMF occur. It appears that many variables (e.g., exposure, justifying promulgation of standards, the frequency, intormity, exposure duration, field Comrnittcc rcconunends that the EMF issue bc orientations) can affect the ruults of these studies, referred to the Texas Dcparunent of Health. which undoubtedly play an important part in the br. ons~tcncies reported in the litoramrc. 2. Siting Criteria Thc Commitlee believes that, hascd on its cvaluation of thc cxisting EMF rescaroh, thc evidence at this 2,1 Conclusiorls timc is insufficient to concludc that cxposurc to EMF The CommiRcc concludes that at prcsenr, thc from clcctric power transmission lines poses an immincnt or signi~cant public health risk* In general, cx~sting cfitcria uscd by thc PUC for siting thc Committcc's evaluation is corroborated by othcr transmission lines appear to bc adcquatc. Thc EMF litcraturc sumrnarics and background reports. Committcc concludes that a plan for enginccrini; interventions is not warranted at this fimc. The Thc CommRtcc concludes that at prcscnt thcrc is Commilzgc noted that "prudent avoidancc" in siting insufficient cvidcncc rcgard ing human health effects of transmission lines has bccn thc dc facto of EMF to providc thc basis for a health-bascd philosophy in thc PUC criteria sincc 19?6, by standard. Thc CommilXcc can find no reason to avoiding population ccntcrs, historical sites and creato arbitrary numbers to usc u a desired icvcl of existing facilities. Mallgrs of safcty and rii;hts-of- cxposurc, bccausc thc usc of such numbcrs cannot way criteria havc influenced thc sclection of routcs. be argued or ticfended on thc basis of scientific Based on cu~r,~ cviduncc, thc Commiucc finds this cvidencc. The primary objectivc of the Committee is approach adcquato and aocclXablc. the protection of public health, and the Commiu. cc can find no scientific argumcnt to support standards, 2.2 Recommendations cithcr through guidancc or through rcgulatory criteria. The CommRtcc rccommcnds that thc PUC continue its policy of dc fact0 "prudcnt avoidancc" in the The CommiUt, c has rcvicwed various statc EMF siting of transmission Emcs. Wc further recommend standards. However, the usc of numbers for an that, at this timc, thc PUC not expand cxjstini; arbitrary standard in the abscncc of scicntific routing critcda to includc conccrns rcgarding health justification sets adc facto risk lcvcl which is not cffects of BMF cxposurcs. suppoxtcd by avai/ablc cvidcncc. Usc of such numbers, which is stricdy political, can gcncratc a xxii Health Effects orsExposure to Powerline Frequency Electric and Magnetic Fields 3. EMF Research 4. Public Forum 3.1 Conclusions 4.1 Conclusions While the quantity and quality of EMF research have Since the mid-1970's, EMF has been an issue in incrcas{xi, the EMF effects data buc, however, is over 200 legal proceedings involving the electric still in a "sate-of-infancy" when compaxcd to the utility industry in the U.S. Approximately 75% of research litorstur~ on risks to other environmental these occurred during siting of electrical facilities, exposures. Continued research will help to reduce primszily transmission lines, and during the cuii~,,~l level of ambiguity inherent in EMF condemnation proceedings. findings, while increasing the certainty of research The Committe~ recoguizcs that the EMF health issue results and confidence in research conclusions. The is oaten introduced in legal proceedings (e.g., Committee concludes that the research agenda hearings). When EMF health effects concerns ar~ developed and funded by the Electric Power addressed in this setting, the expense and delays in Research Institute (EPRD should continue together siting decisions ~ un~ntsrranted. The quasi-judicial with enhanced federal funding. A considerable rule of a state rcgulatory agency may facilitate the number of laxlie and well-designed studies sr~ establishment of a better srcns thin the judicial arena currently under w~y. These studies offer the for the expression of differing views and conflicting potential of more conclusive information thin exists evidence. The Conunite= recognizes that, regarding at present. Increasing the number of studies without transmission lines, the PUC has jurisdiction. a coherent research plan is not likely to contribute to However, transmission lines are not the only source resolving the current inconsistencies in available of EMF exposurc. It is anticipated that s public rcsearch results. A carcfuBy coordinated and forum sponsored by an apprupriate state agency, comprehensive national research agencla with addressing all exposures, would pruvide it adequate fuming from a mix of governmental and nonadversaxial setting for the review of concerns non-governmental sources is needed. regarding the EMF health effects issue. Th~ Committe~ also notes the need for a more systematic rn, icw of avaiiabl~ research results to 1[.2 Recommendations resolve the inconsistencies in the published studies and to identify sr~ts of needed resetrch. Most The Committee recommends that the Texas reviews of available evidence conducted to a_._t,. hive Department of Health assume the leadership rule in not employed rigoruus t,~view criteria and sponsoring any public informational meetings for the quantitative methods such as those used in meta- exchange of EMF informslion. Such meetings, if analysis. Such reviews would entail extensive deemed necessary, can be scheduled in conjunction rcanalysis of data frum studies specifically selected with the release, by the PUC, of the Committec's because they satisfy study design criteria. A review annul repotu. of this type was beyond the resources available to th~ Commiuee. 5. Education Of The Public The application of limited state resources to the EMF issue cannot be justified at this time, when more 5.'[ CorlCJUSiOflS direct public health benefits can be derived frum The Committee hu not addressed the important need other uses. for public education regarding mk assessment, as this was beyond the scol~ of its work. However, the 3.2 Recommendations committee recogniz~ the need for follow-up The Committee recommends that the PUC continue research and education in public risk perception and to review research findings fi~m on-going studies of risk communicalion in conjunction with the association of EMF with human health effects as dissemination of risk information to the public. these data become available. The PUC should The Committee also did not investigate the question continue this review thruugh the Committee on EMF of personal options such as prudent avoidance. Health Effects. However, with the present uncertainty regarding the The Committee recommends that Texas not develop association of EMF exposure with health effects, there is no clear indication regarding what specific s specific EMF research program at this time. aspects of EMF exposure should be avoided. The public needs to be informed about the EMF health effects issue, and involved in the discussion of Overview with Conclusions and Recommendations concerns arising from this issue. At present, the 5.2 Recommendations TDH and utility companies do respond to public concerns regarding potential health effects of The Committee recommends that electric utility environmental exposures by speaking to public companies and the TE)H continue to be responsive to groups and by distributing writXcn information. the public's nccxi to have general information Some electric utility companies also make EMF regarding potential EMF health effects, and continue mca.suremcnts for their customers. to provide ncc~cd information through brochures, audioviaual pre,sentationa, and field measurements. 1-1 1.0 INTRODUCTION AND BACKGROUND Effects and Exposur~ to EMF, Experimental 1.1 Introduction Studies, Judicial Issues, Rcguiatory Issues, and The Public Utility Commission of Texas (PUC) Policy Issues and Options. recognized the increased concerns r~garding exposure to powerline-frequency (i.e., 50 and 60- 1.2 Background hertz) clcctric and magnetic fields (EMF) and their Until recently, serious inquiry about biological possible cffccts on human health. On April 18, effects sssociated with electricivy was limited to 1988, the Commission resolved that a Committee (i.e., Electro-Magnetic Health Effects Committee) safety issues, primarily those identified with electrical shock. Several events in the 1960s and be appointed to research the literature and monitor 1970s prompted inquiry about the biological the on-going research on the health effects of effects of exposure to electric and magnetic fields electric and magnetic fields from electrical (EMF). RcporU from the Soviet Union in the transmission and distribution lines, and report annually their findings to the Commission. early 1960s suggested neurological and cardiovascuiar effects in workers exposed to CommiU;ee members were originally appointed in electric and magncrjc fields. The controversy December, 1988 and met for the first time in increased in 1974 when the New York Public January, 1989. In May, 1989 the Commiuec Scrvicc Commission began hearings on a proposed issued an Interim Report. Additional members 765 kilovolt (kV) transmission line, resulting in the wore added in 1989 and 1990. Thc following New York Electrical Utilities funding a 5-year, $5- report represents the fu'st complete review, and million EMF research program. Many articles will be up-dated with new research findings, conclusions and recommendations on an annual concerning biological cffects of exposure to basis. clecU"ic and magnetic ~ckls have appeared in the technical literature during the past two decades. The mereben of this Commiucc represent the Many more have been published in the popular research col~munivy, the state public health agency press. These events have increased the public's and the suppliers of clcaric services. The concern over possible health effects when exposed members hold credentials in medicine, to EMF. cpidcrniology, biology, biochemistry, engineering, Electric fields are produced by the voltage applied health physics, political science and biostatistics. to a wire and are measured in volts pot meter Commiucc members are familiar with the current scientific EMF literature, and are actively engaged W/m). Magnetic fields are produced by the current flowing through a wire and are measured in reviewing hundreds of documents and published EMF research findings. in terms ofgauss (G). The amount of power that a line transmits is thc product of its voltage and This CommiV~c is similar to others that have current. Power systems are designed to hold potformed literature reviews and made voltages relatively consUmt, while currents recommendations regarding possible health effects increase and decrease depending on the power from these fields. However, it differs in one demand. Therefore, for a given voltage ~c major respect: the Committee was esublished electric field will remain relatively constant over because of the foresight of the Public Utility time, but the magnetic field will increase or Commission of Texas and not by lcgisiativc or decrease depending on power demand. judicial mandates. There are basically three stages in generating The CommiUgc has neither been placed under a electricity and moving the clectricivy from thc deadline to produce a recommendation nor has it electric stations to the end user (Figure 1-1). been funded to conduct its review or conduct new First, electricity is generated at an electrical research. It serves as an independent review body generating station g about 20,000 volts (20 to evaluate scientific evidence and to provide kilovolts). The power is then passed through a recommendations and advice to the Commission transformer which increases the voltage so that the on the possiblc health effects associated with EMF power can be transported with minimal losses. In exposures. the second stage, electriciVy is transported over high voltage transmission Ones (i.c., 69 to 765 This rcpor~ is divided into seven major sections: kV). Introduction and Background, Engineering and Exposure Assessment, Epidcmiology of HeaRh 1-2 Health Effects of Exposure to Powerant Frequency Electric and Magnetic Fields GENERATION TRANSMISSION DISTRIBUTION END USER Primmy Secondary 20kV 60W to 50OkV ~ 2kV to 59kV ~ 120/240V Figut~ 1-1. Schematic illustration of the rages in an electrical system used to transfer power from the gcnuator via t~ion and distribution liege to an end user. (l~odified from Office ofT~chnology Aseeament Report-Biological Effects of Power Frequency Ele~tri~ and Magnetic Fields). Currently in the United States there arc more than lines, EMF are produced. But these fields are also 300,000 miles of alternating current (AC) pnxluccd in homes, offices and other buildings, powerlines ranging from 115 to 765 kilovolts due not only to the proximity of the transmission OcV). However, 500 kV is the highest operating and distribution lines, but due to electrical wiring voltage currently used in Texas. in the facility and the use of electrical appliances (e.g., can openca, hair dryers, video display Transmission lines connect to substations where terminals, m-ten, electric blankets). Thus, the the voltage is reduced and power is transferred to sources of exposure to fields arc numerous, lower-voltage distribution lines. In the third stage, exposure to fields is ubiquitous and if a true distribution lines deliver power locally to human hmtlth hazard exists, the response will individual users. The distribution system is need to be comprehensive, involving society as a composed of two voltage levels. One is a whole. 'primary' circuit (2 to 59 kV) that delivers power from a substation to a distribution transformer. Electric and magnetic fields arc not something From there the power flows through a new. Scientists have had a good understanding of 'secondary' circuit to an end user. The them since the nin~e_-nth century. For example, 'secondary' circuit voltage is low enough processes in the earth's core give rise to the (120/240 volts) to operate household electrical esrth's magnetic field. Unlike the alternating fields associated with transmission lines and appliances, lights, etc. appliances, the earth's magnetic field does not The electricity we use in our homes, offices, etc. alternate, but is static. is alternating current (AC) in contrast to direct cuii~t~ (De) which is like that produced by The EMF from powerlines and appliances are of batteries. Alternating current does not flow in one extremely low energy and frcclucncy. They are direction, but instead alternates back and forth. markedly different in frequency (i.e., Hz) from The current used in North America alternates back ionizing radiation (e.g., gamma rays, xxays, and forth 60 times per second, which is called 60 ultraviolet rays) in the electromagnetic energy hertz (60 Hz), compared with 50 times per second spectrum (Figure 1-2). Not only is the energy in (50 Hz)in Europe and other countries. the 60-Hz frequency not great enough to cause ionization, there is not even enough energy to heat Although the major public concern has been tissue as is the case for microwaves. The usociatcd with EMF exposure from transmission non-ionizing and athcrmal (i.e., non-heating) lines, EMF arc also present whenever electricity is characteristics of EMF produced from 60-Hz used. As electricity is generated at electrical frequencies arc two of the reasons why some generating stations and transferred to homes via scientists believe that these fields could not induce transmission tines, substations and distribution biological changes. However, biological changes Introduction and Background 1-3 30 10 1015 1013 I0I · 10 ENERGY TYPE Figur~ 1-:2 The clcctromag, ntljc spectrum havc bccn obscrvcd, trader cxpcrimcntaI cxposurc l~xposurc duration (i.e., acute vs. chronic) must conditions, and thcsc observations havc incrcascd also bc considered. Since most appliances arc thc concern for possiblc human health cffccts. used infrcqucntly and for short duration, their cxposurc may bc of less importancc. Also, it is Elcctric ~cld-~ can bc easily blocked by mcca, realized that the public pcrccivcs involuntary buildings, earth and othcr objccts. Howcvcr, cxposurc (c.g., transmission line) to bc morc of a magnctic fields arc not ca~ily blocked and can pass hcalth hazard than voluntary cxposurc (e.g., through buildings, earth, and humans. appliances). Somc individuals havc prcscntcd thc contcntion During the past dccadc, extensive rcscarch that EMF cxposurc to clcctrical applianccs is just programs evaluating the possiblc health cffccts of as great a p~tcntial health hazard as cxposurc to cxposurc to EMF havc bccn pcrformcd in the U.S. EMF produced from transmission lincs. Bccausc The U.S. Department of Energy (DOE) and thc the intensity of EMF decreases rapidly as onc Electric Power Research [nstitutc (EPRI) havc moves away from a source, thc fields at thc edgc sponsorcd much of this rcscarch. Thcsc studies of thc rights-of-way for a transmission linc (sincc havc helped answer many unknowns, but many the source is :25' to dO' above the ground) may not questions rcmain. Notwithstanding thcsc bc much grcatcr, and in ccrtain cascs may bc lcss unknowns in the scientific data base, these than, thc ficids next to an clcctrical appliance uncertainties have triggcrcd public, rcgulatory, (c.g., oven, hair dryer, clcctricai shaver, can and judicial involvements. This report cvaluatcs opcncr), cspcciaily for magnetic ticIda. Such thc EMF scientific litcraturc and addresses the comparisons arc of value, but ~cld intensity is rcgulatory and judicial involvement in the issue. only onc of thc important variablc3 to bc considcrcd in cvaluating potential hcalth cffccts. 2-1 2.0 ENGINEERING AND EXPOSURE ASSESSMENT and cellular studies have not clearly identified any 2.1 Introduction s~gle metric of exposure. Exposure assessments have been relegated to measuring only the convenient and Studies of the possible effects of power,line frequency simple proponies of an environment due to the electric and magnetic fields on health are hampored by obscurity of the metric and the scarcity of sophisticated problems in measuring exposure. Exposure should not instrumentation necessary to measure aspects more be confused with dose. Exposure is the simultaneous complex than time average field strengths. occurrence of some agent (e.g., electric or magnetic fields) in the presence of a subject (e.g., human), Many exposure assessments to date have relied on whereas dose is the amount of agent actually interacting average exposure during a sampling poriod. Inherent with the subject. Dose inherently involves a thorough in the averaging process is a loss of information; the understanding of the cause and effect relationship more subtle aspects of EMF, such as windows, between the agent and biological effect. Currently, transient exposure, etc., are obscured when science is only beginning to understand this relationship instantaneous field values are averaged. Also between electric and magnetic fields and their averaging fails to portray any temporal variations. interaction with biological systems. In this section the acronym EMF will be used for electric and magnetic Exposure assessment is important to cpidcmiologists, fields. biologists, and regulators. To correlate a disease with a suspected agent, cpidcmiologists must bc able to The essence of exposure assessment is determining, measure ~e difference in exposure among the subjects through direct measurement or estimation, the amount for the metric of interest and for a host of possible of a catml agent occurring in the subject's confounders present in a real environment. To environment. The assay of this agent is often called the investigate a suspected interaction, biologists must bc "metric of exposure"--the quantity that explici~y is able to design cxporiments that accurately mimic related to dose and the one we want to measure. exposure in the real environment. If their Ideally, science must first identify the mechanism by extrapolations of laboratory experiments to the real which the agent affects the subject before wc know environment arc to bc believable, they must be able to what to look for. For example, if a subject drinks a simulate accurately and control the exposure of their glass of chocolate milk and then breaks out in 'a rash, subjects to the suspcctod agent. And finally if science what caused the rash? Was it the milk, the chocolate, does identify a public health hazard, regulators must bc some by-product of the reaction between the milk and able to identify explicitly which aspects of EMF are chocohte, the color of the mixture, or the material threatening public health and at what level those aspects from which the glass was made? Most exposure should be limited. assessments have assumed that the average magnetic or electric field strength found in the subjecCs 2.2 Sum maW environment is the "metric of exposure." Yet, several studies, both in vitro (cclluhr) and in viv0 {animal), The Committee has reviewed the major elements of have suggcstext other aspects of the electric and magnetic fields, besides average field strength, may be exposure assessment through reviewing the literature, examining computer models and communication with the measure sought. manufacturers and users of EMF measuring equipment. Wc fred that for making survey measurements of I=MF Like with the host of properties associated with the associated with powerlines, commercial instrumentation glass of chocohte milk, any one of the propot-ties is readily available and acceptable standards, specifying manifested in the EMF environment could be the metric how these measurements should bc made, have been of exposure. For crampie, associated with every field published. However, for measuring exposure, only a arc aspects of wave shape, frequency, harmonic few choices of commercially available instruments content, and transients (spikes), and, if transients are exist, and the instrumentation to make thorough and present, their host of properties. Furthermore, the intensive engineering measurements must be custom exposure metric could bc the variability of the field or assembled. Also, there are no standards to specify how perhaps the number of times the subject passes in and exposure and engineering measurements should be out of the field. The exposure metric could be the made. occurrence of fields in certain windows of frequency and/or amplitude, or even more complex, some type of When it is impossible or not 'feasible to actually interaction between the strength and orientation of an measure cxposurc, EMF exposure can bc estimated by cxterual field and the earth's magnetic field. Animal using computer models, spot measurements, and 2-2 Health Effects of Exposure to Powerant Frequency Electric and Magnetic F~eld~ surrogates. Several computer programs exist to Current is the movement of charge through a conductor accurately estimate field levch for the simplistic and is measured in Amperes (A). A circuit is created geometries usually found around transmission lines. when a continuous path for the current is formed. With The Electric Power Research Institute (EPRD is direct current (DC), like thatproduced by a battery, the currently developing a program to calculate magnetic current flows in one direction at a constant level; fields found in the more complex residential geometries whereas with alternating current (AC) both the level composed of distribution circuits, house wiring, and and direction of the flow change periodically with time. ground raum paths. EPRI has also developed a Frequency is the number of these complete cycles that program to estimato exposure basal on time,-weighted the alternating current undergoes in one second and is averages of field strength. Under controlled exprcssai in cycles-per-second or Hertz (Hz). conditions, spot measurements may be combined with Electrical power systoms in North America operate at the subjects' activity patterns to estimate exposure. 60 Hz, while 50 Hz is prcdominato elsewhere, Surrogates must be used with great care sinc~ they including all of Europe. For 60-Hz altoroating current, oRen suggest other factors besides powerline EMF, 60 cornplea cycles occur every second with the which my be associated with cancer. currcnt's direction reversing during each cycle. Pr~limitmry studies show that electric fields in the home Conceptually similar to water being pumped through a arc not greatly affected by outside powerlines, but these pipe, electrical current is *pushed* through s conductor line my be important contributors to interior rtmgnctic by a diffcrcnco in electric *preasure' or potential fields. EMF in the work place is s'unilar to that in the betweal the ends of the conductor. This difference in home. High current devices appear to be more petcntlal is measured in volts and is called voltage. prevalent in the work place than high voltage devices, With alternating curt,rot, both the voltage and current so higher magnetic fields arc moro likely than electric vary sinusoidally, as Figure 2-1 shows. fields. Little ~_~¢-_ exists on EMF in other area. 2.3 Electric and Magnetic Field Fundamentals a.a. Introduction This section lays · foundation of the basic concepts about electric and magnetic fields fundamental to understanding cxposuro assessment. The reader already possessing · basic understanding of powerline fields may wish to skip this section and refer to it or the -- glossary as needed. 2.3.2 Basic Elactrical Concepts The source of both electric and magnetic fields is electric charge. Charge can be either positive or negative. Like charges repel and opposito charges attract. This electric force acting Ixnv~n charges is about a billion-billinn-billion-bi!ilon times (1036) stronger than the force of gravity between the two Figure2-1. Alltrnating sinusoidal wave shape for curr~nt or charges. voltage. Th~ quantity flows ons dirsction during the first half of the cyck: and r~v=rs~ di~-ction during the s~cond. 0EPRI, 1989) A conductor is any material that allows electrons to move freely and to railstribute charge. At some level 2.3.3 Field Concepts of voltage, most rna~rlals t~come conductors. Metals are the b~st conductors. When electrons in a material arc not frcc to move, the material is called an insulator. A set of values of a physical quantity at different points This property of opposing the movement of electrons is in space can be represented as a field. An example of called resistivity. In a wire, resistivity is expressed as a simple field is the ternpcraturo across the State of resistance and is measured in Ohms. An ideal Texas at noon on January 1, 1990. Each geographic conductor has zero resistance and an ideal insulator has point in the state has associated with it a measurable infinite resistance. value of temperaturo. By associating a temperature Engineering and Exposure Assessment 2-3 reading with every reporting point in the state, we middle of the spectrum, in a small frequency band, is could construct a temperature field. visible light; different frequencies of light produce different colon. Below visible light are frequencies The above example is a scalar field, where the property that produce infrared, microwave, and radio waves, being mcasurrxl is a value easily read on a single scale. while above visible light are ultraviolet, x, gamma, and More pertinent to the electric or magnetic fields is the cosmic rays. The product of frequency and wavelength idea of a vector field, where each point not only has a of electromagnetic radiation is always a constant-the value associated with it, but the value is oriented in a speed of light. Thcrefore, the highcr the frequency the specific d~on. One example of a vector field is the shorter the wave length. A 60-Hz powcr frequency has trajectory of each fragment of a hand grenade during a corresponding wavclcngth of 5,000 kilometers (about an explosion. A snapshot would show that each piece 3,000 miles). In comparison, the wavelength of a of shrapnel is travelling at a specific speed in a certain television transmission is about 3 meters. direction. Wc can describe the explosion in terms of this vector field. Another example of a vector field is The way the electric and magnetic fields from a source the pattern formed by water spraying from the end of a of electromagnetic energy appear to an observer fire hose. A vector field describing the flow will consist depends on the distance to the source in comparison to of the speed and direction of each molecule of water at the wavelength of that source. V~icn the distance from every point in the flow. the source is large compared to its wavelength, the electric and magnetic fields arc linked together as 2.3.4 The Electromagnetic Spectrum electromagnetic radiation. The area where this linking occurs is called the 'far" or 'radiation' zone. At As shown in Figure 2-2. the electromagnetic spectrum anything greater than atomic distances, visible light will encompasses the' frequency range of all electromagnetic always appear as a radiation. energy; Near the bottom of the spectrum are extremely low frequency (ELF) waves like powerline fields and When the distance from the souroe is small with respect near the top are very energetic cosmic rays. In the to wavelength, the electric and magnetic fields appear as separate quantities. Earth based observers are _ always in the so called 'near" or "static' zone of power Uses Frequency $pec~aJ ri~lon~ Wav~l~rl~lq frequency fields because of their long wavelength. sxmon,~ ,~r"'m Therefore, power frequency fields behave as separate, ~0-'~,. independent, non-radiating electric and magnetic fields. So when studying power frequency fields, we consider s*'osm~ '°'~m the electric and magnetic fields as separate quantities and not as electromagnetic radiation. ,a*.. 2.3.5 Electric and Magnetic Fields Sun Lamps ~°'Tm Electric and magnetic fields ar~ vector fields. Within ,e4m the field. the electrical forec produced by the field on a v~t~et mx~e'4~ ~? ~c'sm unit charge can have a different magnitude and Imadia~t ~ ,z ~ ~°~"~ direction at each point in space and time. These fields ,o-.m are dcfmcd by the forces exerted on electrical charges. wadar Sumt.~,fmquancV ,C-'m Electrical charges cause electric fields. which can be 3x~o*~ U~rahkJ.~eey . ~m described in terms of electric field strength (E) with Talemien Very ri0hfm~k~y ,cm units of volts per meter (V/m). The electric field is Radio 3 · IO *H~ High fme~m~c'y ,oom dcfmed by the force it exerts on a static unit of charge. Mectiumfrequeetcy LOOOm The electric field is a function of the voltage of the m~o.= t~vhqu.m~ ~0.ceom source-the higher the voltage the stronger the field. v~t ~v;,,quo.cy ,c~zo m Transmission line electric fields are typically measured --~ .. ~.eeo.oco m in thousands of volts (kilovolts) per meter (kV/m). Power Tmnsmismi~ ~ 0.3)0,000 m 3 ~k ,~o.o~.~x~ m Moving clectricai charges cause magnetic fields. Just as the electric field is dcf'med by the foroc exerted on a stationary unit of charge, the magnetic field is defmcd F~mu'e 2-2. The electromagnctlc spectrum shown by frequency by thc forec exerted on a moving unit of charge. The and wavelength. At a frequency of 60 Hz and a wavelength of magnetic field is usually measured in terms of its 5,000 ks, powcrlin~m are st the bouom of the figure. magnetic flux density (B), although some instrumcnts Frequencies less thsn 300 Hz are designated ms extremely low frequency ('ELF), (EPRI, 198~) 24 Health Effects of F.;cposure to Powerlint Frequency Electric and Magnetic F~elds may be calibrated in magnetic field strength (H). The The magnetic field is independent of the voltage, but magnetic field strength and flux density are rehted to depends on current in the conductor and the distance to each other by a permeability constant ~) i.e., B=/~H. the conductor. The magnetic field increases with more The most common units of magnetic flux density are current and increases the closer you get to the source. the gauss (G) and testa (T) and for magnetic field Unlike the electric field, the magnetic field from a strength it is the ampere/meter (A/m). Table 2. i shows powerline exhibits a great temporal variability since it the equivalence between units. Powerline magnetic is a function of the circuit loading, which varies by fields are usually described in terms of thousandths of a time of,lay and season of the year. gauss (milligauss or raG) or millionths of a testa (microtesta or/~T). The magnetic field drops off with increasing distance from the source. The electrical and physical Table2.1-Equivalence l~twe~n Mapstic Fidd Units characteristics of the source dictate how rapidly this decrease occurs. Generally, magnetic field levels will unis o mO T t,T A/m doerease according to one of three relationships with ~ I 100o 0.0ool 10o so distance: invertely with distance, inversely with the ma o.ool t 1o-7 0.1 0.os square of distance, or inversely with the cube of distance. Figure 2-3 illustrates these three relationships T 104 :07 i 106 soo.000 for a source of the same strength. Doubling the t,T 0.01 10 10.6 I 0. S distance will decr~tse the field to one-half under the A/m 0.0125 12.5 1.25x10'6 1.25 I inverse relationship, to one-quarter under the invene squared, and to one-eighth under the inverse cubed. The conditions under which these relationships occur The electric field at a point is a function of the voltage will be discussed later. of the source and the distance to the source. The electric field strength increases as the voltage is raised Earth's Electric Field. The earth possesses an or the distance to the source is reduced. Because essentially static electric field, which is vertically- utilities design their systems to maintain powerline directed with a strength about 130 V/m near the voltage levels within st fairly usn~w range over time, surface. It is caused by the separation of charge the electric field at specific point from a particular between the earth and the ionosphere. Together they powerline will vary little with time and can almost be form a capacitor with the earth being the negatively consider~ constant. charged plate and the atmosphere being the positively Fngutt 2-3. Field strength varies with distan~c from thc source according to inverse, inverse-squared or invers~-~ubed relationships. Engineering and Exposure Assessment 2-5 chargexi plate. Lightning maintains the potential varies between the equator and the ]~oles. The vertical difference by transferring the excess charges. On component of the geomagnctic field is greatest at average, about 2000 thunderstorms arc occurring at magnetic poles, reaching about 670 mG and falls to any time, and there arc about 100 lightning flashes per zero at the magnetic equator. Conversely, the second worldwide. The field follows a diurnal cycle as horizontal componcnt's maximum of about 330 mG shown in Fignrc 24. Fields of 10 kV/m or higher can occurs at the equator and is zero at the magnetic poles. occur during thunderstorms. {EPRI, 1989) (EPRI, 1989) Earth's Magnetic Field. The earth also possesses an Man-made Power Frequency Fields essentially static magnetic field. Current flowing Overhead Transmission Lines. The most common through the earth's molten interior is believed to be the means of transporting electric power is by overhead source of the gcomagnctic field. Its magnetic flux alternating-current transmission lines. Transmission density averages about 500 mG at middle latitudes, but lines arc oitcn groupal by their design operating voltages. Two groups arc high voltage (less than 345 kV) and extra high voltage (345 kV and above). A typical transmission line has three phase conductors per E(Vtm) circuit. Multiple, or "bundled," conductors for each phase arc used at higher voltages to control corona- 120 related effects (such as audible noise) or to increase 110 ~/~ power handling capability on heavily loaded lines. Figure 2-5 shows a typical transmission [inc. In each phase conductor, the sinusoidal voltage or current wave is otlt of phase with the othcr two phasc conductors by one-third of the wavelength. The transmission line is said to be balanced if the '[ I I I I vector .sum of the phase voltages and phase currents 0 6 12 18 24 add up to zero. Ideally no currents will bc flowing in 14~tn Gt4T the shield wires or the ground. Becausc of differences in the electrical characteristics of each phase and the diffcring amounts of single phasc load attached to each phase, transmission lines arc seldom precisely r~re 2-4. Average diurral variation of the atmospheric balanced. Even undcr balanced conditions, some pot¢ntial gradient. Th, peak occurs ,gar 7 p .m. Greenwich current will bc induced into the shield wires, unless M~sn Tim (GMT) and is saociat,d with peak thunderatonn they arc isolated into short segments. activity around the globe (EPRI, 1989). Transmission lines arc identified by thcir nominal F'~mre 2-S. A typical thr~c-phasc single-circuit AC transmission line. CEPRI, 1989) 2-6 Health Effects of E~posure to Powerline Frequency Electric and Magneac Fields phase-to-phase voltage. Typical nominal voltages in minor axis. Figurc 2-7 shows the field ellipse. Texas arc 69, 115, 138, 230, 345, and 500 kV. In Electric current in the transmission Line . phase practice, actual operating voltage w;tll be within a few conductors produces a magnetic field. Figure 2-8 percentage points oft he nominal voltage. shows calculatc<i magnetic flux densities proides at 1 Electric fields near transmission lines arc usually meter above the ground for diffcrcnt transmission tines. calculated or measured at a height of 1 m above the The peak magnetic field beneath a 500-kV line carrying surface. The strength of the electric field from a 2,150 megawatts (MW) is about 450 mG. For a single- transmission line is dependent on three factors: the circuit 345-kV transmission l~nc carrying 1,050 MW is operating voltage of the line, the height of the about 330 raG. The peak magnetic ficld for a 2.30-kV conductors above the ground and thc distance from the line carrying 350 MW is about 170 mG and the peak l;mc to thc point of messuroment. Figure 2-6 shows the magnetic field for a 138-kV line carrying 112 MW is clectric field profilcs for a 500-kV, 345-kV, 230-kV, about 95 raG. and 138-kV single-circuit transmission lines. Because Unlike the clectric ficld, thc presence of most objects the ficld profile is symmetrical about the center of the does not perturb the magnetic ficld thus making transmission fine for symmetrical arrangements of thc shielding very difficult. phase conductors, often only onc-half of thc profdc is shown. Thc calculations assume an open, flat surface For balanced conditions, thc transmission linc's in the area about thc transmission linc. Conducting magnetic field will decrcasc with thc square of the objects such as vegetation, buildings or fences will distance from the line. Generally transmission line 'perturb,' or distort the electric field. phase currents wiJl be better balanced than distribution linc phasc currents. If an unbalanced condition occurs, The clc~lric field that permeates space surrounding a thc resulting magnetic field would be proportional to transmission line can be described as a rotating vector the dcgrec of unbalance in the phase currents (i.e., the field. The electric field at each point in space may net current) and would decrease as distance between have a different magnitude and direction and varies the line and point of interest increases. cyclically st the powerline frequency. The loci of the field vector describes an ellipse, with the maximum The design of the transmission line will influence the electric field occurring along the semi-major axis and strength of the power frequency fields under and next the minimum electric field occurring along the semi- to the transmission line. The parameter of transmission line design having the most effect on the strength of Comparison of Electric Field Profiles 8000 500 kV 7000 '~'6000 345 kV ~---5000 230 kV 4000 3000 138 kV 2000 1000 0 ~ 0 50 100 150 200 250 300 350 400 Feet Fqure 2-6. The maximum electric field lateral profile for 500-kV, 345-kV, 23D-kV, and 138-kV tammission lines. The profiles arc symmetrical about center of the line (located at 200 t~.). Conductors arc at minimum clearance conditions and the field ngasurcd st lm above the ground. Engineering and Exposure Assessment 2-7 at some point on the ground, conductor height must bc y increascd for points within the critical distance (i.e., usuaUy locations within thc right-of-way) but decreased for points greater than the critical distance (i.e., usually locations outside the right-of-way). Since the magnetic field is unaffected by the presence of the ground, this critical distance phenomena does not affect the magnetic field profiles. Magnetic field intensities, for nil points on the ground, will always bc reduced with increased conductor height. Another parsmeter of transmission line design affecting only the electric field strength is the size of the phase conductors. Figures 2-12 and 2-13 shows the effect on the magnetic and electric fields respectively of doubling X and halving the 18 inch bundle spacing. Decreasing the diameter of the phase conductors or, for bundled phases, decreasing the bundle spacing decreases the electric field strength, but has no effect on the magnetic field strength. Drastic changes are rcquixcd to affect significantly the electric field at ground level, and the opportunities for manipulating phase diameter are ~ limited by mechanical, electrical, and cost constraints. The distance between phases can affect the electric and magnetic field levels. Figures 2-14 and 2-15 show the effect on the electric and magnetic field pref'dcs, " Figure 2-7. The electric field ellipse ·t · point in sp·ce. The respectively, for the nmc transmission line with different phuc spacings of 37.5', 27.5', and 17.5'. The nmximum field occurs ·long the semi-major axis and the canccling effect of one phase upon the others suggests mininmm ·long the semi-minor axis. The field vector rotates at that more compact lines will have lower fields at the per frequency. (EPRI, 1982) ground level. The amount of compaction is limited by the corona performance, tower construction and fields is the height of the conductors above ground. spacing, and National Electric Safety Codc's clearance Figures 2-9 and 2-10 show the electric and magnetic considerations. The orientation or configuration of the field profiles for a single circuit 345-kV transmission phases also can have a significant impact on the ground ]inc with conductor heights of 33', 43', 53', and 63'. level fields. Figures 2-16 and 2-17 show the electric As the figures show increasing the height of the and magnetic field pro Cdes for three possible phase conductors lowers the field strength under the imc. configurations: vertical, horizontal (fiat), and an equilateral (delta). All three configurations have the But, the distortion of the electric field by the earth same phase spacing (27.5') and minimum conductor preduccs a curious effect: raising the conductors height (33'). The vertical and fiat configurations have sHghdy increases the electric ficld's strength at points the highest maximum ground-level electric field under greater than a ccrcain distance from the ]inc. This the line while the delta has the lowest. The electric field distance is called the critical distance. Figure 2-11, outside the right-of-way is lowest for the vertical which shows the resultant electric Reid prordcs for the configuration and highest for the fiat. The fiat two extremes (33' and 63'), more clearly shows this configuration has the highest maximum ground-level effect. With increasing conductor height, the electric magnetic field and the broadest protalc, whereas the field strength decreases at those points locatcci less than delta and vertical present lower maxima and more thc critical distance, the electric field stays the same at compact protales. But, vertical con figured lines require the critical distance, and it increases at distances the taUcst towers and therefore are the most expensive. greater than the critical distance. For transmission lines composed of more than one The critical distance depends on the distance between circuit, the phase sequencing can have a dramatic effect phase conductors and the dismatt of the phase on the ground lcvc] fields. Also using a lower voltage bundles. The significance of the critical distance is that distribution circuit beneath a transmission c~rcuit can in cases where it is dcsirablc to reduce the electric field sometimes raise or lower ground level fields. Health Effects of Exposure to Powerline Frequency Electric and Magnetic Fields Comparison of Magnetic Field Profiles 500 kV @ 2150 MW 450 400 W 350 345 kV @ 1050!W ~-~-E300 ,250 230 kV @ 350 MW .~2 200 ~ceo~150 138kV@ 112 :E 100 0 50 100 150 200 250 300 350 400 Feet F'qure 2-8. The maximum magnetic field lateral profile for 500-kV, 345-kV, 230-kV, and 138-kV transmission lines. Lines arc at maximum operating load of 2150, 1050, 350, and 112 MegawatLs, rcspe. ctively. Conductors ar~ at minimum clcarame conditions and field measured at lm above ground. Electric Field Profile for Different Conductor Heights 6000 h = 33' 5000 h = 43' ~ 4000 h = 53' i.~ 3000 1000 0 50 100 150 200 250 Distance from Center of Line (ft.) Figure 2-9. Electric field profiles at lm above ground for single-circuit 345-kV transmission lines with conductors 63, 53, 43, and 33 feet above the ground. The profiles arc symmetrical about center of line and only one side is shown. Engineering and Exposure Assessment 2-9 Magnetic Field Profile for Different Conductor Heights h=33' 200 180 ~ h=43' --- 160 h = 53' ~ 140 120 '~ 100 "~ 40 .-,'-~.~:~:~:.~, 20 0 50 100 150 200 250 Distance from Center of Line F'~qu'e 2-10, Magnedc field profiles at lm above ground for single-circuit 345-kV transmission lines with conductors 63, 53, 43, and 33 feet above the ground. The profiles are synunetrical about center of line and only one side is shown. Electric Field Profile for Different Conductor Heights 6000 5000 h = 33' ~., 4000 h=63' ~3000 2000 " ' nce (Lcd) 1000 , ," ' ' ' ' ' . . ............. 0 50 100 150 200 250 Distance from Center of Line (ft.) F~,utt 2-11. Critical dinnee (Led) for electric field from a 345-kV transmission line. The intersection of the field profiles occurs at Led. Increasing conductor height lowers E-field inside Led but raises E-field outside of Led. 2-10 Health Effects of F. xposure to Poweraunt Frequency Electric and Magnetic Fields Effect of Bundle Spacing on Electric Field Profile d=36" 6000 d= 18" 5000 =" ~_, 4000 tr, 3000 · _ '~':~ ~.:,~:~ 2000 1000 0 50 100 150 200 250 Distance from Center of Line Fipre 2-12. Electric field profiles for phase conductor bundle spacings of 9, 18, and 36 inchca for a single-circuit 345-kV transmis~on line. Lines with smaller conductors (i.e., closer bundle spacing) have lower electric ticIda. Effect of Bundle Spacing on Magnetic Field Profile 200 - 180 ', d -- 9", 18", 36" 160 ~'~140 "",,~.. .~ 120 ',,. "100 ", ~ 80 "",. '- 60 "'.,-,. :~: 40 "'--,,_ 20 0 50 100 150 200 250 Distance from Center of Line Figure 2-13. Magnetic field profiles for plug conductor bundle spacings of 9, 18, and 36 inches for a single-circuit 345-kV transmission lirg. Conductors size (i.e., bundle spacing) has no effect on magr~tic field str~ngffi. Engineering and Exposure Assessment 2-11 Effect of Phase Spacing on Electric Field Profile Dp = 37,5' 6000 Dp = 27.5' 5000 i .~"' ~\ ""' '., = ' ' 4000 ~3000 2000 1000 " "~':'::~'=';~="~-'='= ...... -r .... 4 0 50 100 150 200 250 Dis~nce from Center of Line ~ 2-14. Electric field profiles for phase conductor spacings of 17.5, 27.5, and 37.5 feet for a single-~ircuit 345-kV transmiseion lin~. Compact transmission lines (narrower phase spacings) have lower electric fields. Effect of Phase Spacing on Magnetic Field Profile 250 Dp = 37.5' 200 Dp = 27.5' 7.5 150 100 50 0 0 50 100 150 200 250 Distance from Center of Line Figure 2-15. Magnetic field profiles for phase conductor spacings of 17.5, 27.5, and 37.5 feet for a single-circuit 345-kV transmiseion line. Compact transmission lines (narrower-phase spacings) have lower magnetic fields. 2-12 Health Effects of Exposure to Powerlint Frequency Electric and Magnetic Fields Comparison of Electric Field Profiles 6000 VERTICAL 5000 _ '~' ,," "', DELTA ~, 4000 ,,? ~x~x ~'~,,~,,~ ~3000 2000 '-,,, 1000 0 50 100 150 200 250 Dis~nce ~om Center of Line (~.) Fainre 2-16. Electric field pro~iea for single-circuit 345-kV transmission lines with fiat (horizontal), delta (equilateral) and vertical phase geometries. The phase spacing and minimum conductor height is the same for each configuration. Comparison of Magnetic Field Profiles FLAT 200 180 / 160 ' ~EE140 DELTA .~--~120 "100 .o 80 VERTICAL 60 40 20 0 50 100 150 200 250 Distance from Center of Line (ft.) Fngure 2-17. Magnetic field profiles for single-circuit 345-kV transmission lines with fiat (horizontal), delta (equilateral) and vertical phase geometries. The phase spacing and minimum conductor height is the same for each configuration. Engineering and Exposure Assessment 2-13 Fields Near Underground Conductors. Placing Table 2.3 - Typical 60.Hz electric field levels at 30 cm from conductors underground rather than suspending them 115-V home appliances. (Source: WHO, 1984) overhead from towers or poles changes the characteristics of the fields. Underground conductors Appliance V/m rely upon rubber and plastic materials for insulation Electric Blanket 250 instead of air; therefore, the phase conductors can be Broiler 130 located much closer together than is possible with Stereo 90 overhead lines. Sometimes all three phase conductors are combined into a single cable. With the conductors Refrigerator 60 closer together, the fields tend to be lower due to the Electric tmn 60 cancellag effect from the phase differences among the Hand Mixer 50 conductors. The electric field is further reduced or Taster 40 eliminated by the earth, or in the case of a cable, a Hair Dr/er 40 Funded shield which sometimes encircles the phase Color TV 30 conductors. Although the fields can be lower, people can be in much closer proximity to the underground Coffee Pot 30 conductors than overhead, so exposuro levels may be Vacuum Cleaner 16 similar to those of overhead transmission lines. Thus lr~._-,tndescent bulb 2 putting conductors underground may not necessarily The magnetic field produced by most appliances is guarantee that exposures would be reduced. from a lcop of wire or many-turn coil. Necessarily, Transmitting power through underground cables is because of the compact size of most appliances, the tremendously more expensive than using overhead diameter of this loop or coil is small and achieves high transmission lines. Not only do high voltage magnetic fields with either high cu~,ts or multiple underground cables themselves cost more per foot and turns. At distances larger than the diameter of the coil arc more expensive to install than overhead the field approximates a three dimensional dipo le field transmission lines, but due to the higher capacitive (Monitor Industries, undated) and decreases with the reactance found in cables, more eirouits will be cube of the distance (see Figure 2-3), therefore falling required to carry the same amount of power. off very quickly. Fields in the Home Environment. Any use of Table 2.4 shows the magnetic flux densities at distances electricity will produce electric and magnetic fields. of 3 era, 30 cm and 1 m from several appliances. At Electric field levels measur~cl at the center of different 30 era, levels range from 0.03 tzT to 30 ttT (0.3-300 rooms, typical of housing in United States, arc shown raG). Notice how rapidly the magnetic field decreases in Table 2.2, while Table 2.3 lists levels measured 30 with increasing distance. Unlike electric field levels in cm (about 1 foot) away from various home appliances. the home, magnetic flux densities close to some The data in the tables are from limited measurements household appliances are higher than encountered and should be considered anecdotal. They do suggest under transmission tines (EPRI, 1989). However; the general range of levels that may be encountered in when comparing the time-duration exposure to the home, although wide variability should be expected. appliance fields, two points must be considered: Preliminary measurements, (Bracken, 1988), indicate that the electric fields found in residences result from Table 2.4 - 60.Hz magnetic flux densities near various internal sources (house wiring, appliance, etc.) rather appliances. (Source: WHO, 1987) than external sources such as transmission and distribution tines. App!~snee Magnetic Flux Densityr Table 2.2 - 60=Hz electric field levels at the center of various 3cm 30cm lm rooms in a typical U.S. home. (Source: WHO, 1984) Can openers 1000-2000 3.5-30 0.07-1 Hair dr/era 6-2000 0.01-7 < 0.01 -0.3 Location V/m Electric shavers 15-1500 0.08-9 <0.01-0.3 bundr/Room 0.8 Drills 400-800 2-3.5 0.08-2 Dining Room 0.9 Mixers 60-700 0.6-10 0.02-0.25 Bathroom 1.2-1.5 Portable heaters 10-180 0.15-5 0.01-0.25 Kitchen 2.6 Ble__~_--rs 25-130 0.6-2 0.03-0.12 Bedroom 2.4=7.8 Television 2.5-50 0.04-2 0.01 -. 15 Living Room 3.3 Irons 8-30 0.08-0.15 0.01-0.025 Hallwa}, 13.0 Coffee makers 1.8-25 0.08-0.15 <0.01 Refrigerators 0.5-1.7 0.01-0.25 <0.01 2-14 Health Effect.v of Erposure to Powerlint Frequency Electric and Magnetic F~tlds · Appliance fields, generally, exist only a small consist of both primary and secondary conductors fraction of the time-most appliances are off (wires). Three separate sources of magnetic ticIda can mort than they are on. be identitlai for distribution lines: balanced currents in primary wires, balanced currents in secondary wires, · Because the fields of most appliances fall off and net current that is the vector sum of all individual wire currents. rapidly with distance, the artoz in which the fields are elevated due to operation of the The primary camca power from a step-down appliances are small compared to the total transformer at the substation to the pole-top living area. transformers on the distribution line. The primary may include a neutral wiro, which may or may not bc Residential background magnetic fields, away from connectai to the secondary neutral wire. The appliances, range from 0.05 to 1 ~T (0.5-10 raG) secondary carries power from the pole top transformer (EPRI,1989). Figaro 2-18 shows the major sources of to the customers' service drops. The secondary usually magnetic fieids, which arc distribution lines, residential consists of two energized wires at the nominal grounding systems, unusual wiring configurations residential voltage of 120 V (240 V between the two within the residence, and nearby transmission lines. wires) and the neutral, which is at ground potential The following will discuss each of these sources. (zero volts). The secondary serves several residences, while service to each home is supplied via a service The power distribution lines that gird the alleys and drop, which is generally a threc-wiro line connected to sirotis of residential neighborhoods are a source of the secondary conductors. magnetic fields in the home environment. These lines // -- Phase wires metallic t --- Neutral wsres water main Fagme 2-18. Residential magnetic field so,axes include appliances, grounding systems and overhead distribution lines (primary, secondary, and net current). Other possible sources arc unusual wiring, underground cables and nearby transmission lines.(EPRl, 1989) Engineering and E;cposur e Assessment 2-15 Safety precautions and engineering practice dictate the Ground currents produce a very non-uniform magnetic number, location, and type of grounds in the field within the living space of a residence because of distribution system, but these multiple ground the widely differing distances and convoluted routings connections provide multiple return paths for the load of the grounding system's current paths. The field also current. The primary's neutral is grounded at each varies greatly in time since the ground current changes distribution transformer and at regular intervals along every time a 120-V appliance operates. the line. If a portion of the load current returns through any of these ground paths instead of through House wiring is gonehilly not a significant souroe of the neutntl conductor, an unbalanced condition ~ magnetic fields. When the supply and return currents occur and a net current will exist, producing a magnetic are equal, the opposite fields cancel and produce very field in the residerico. little field at distances of more than a few inches from the wiring. In other situations, which may not violate As mentioned earlier, the fields from balanced currents any electric codes, more than one supply or return path decreue proportional to the square of distance, is present. This might happen when controlling one whereas the fields from unbalanced currents (i.e., net socket of a duplex outlet from a remote light switch, current is not zero), decrease proportional to distance. wiring adjacent outlets from different sources but using At locations far from the line, the net current becomes only one neutral return for both, and controlling one a major contributor to the magnetic field. Therefore light fixture from two switches. predicting the magnetic field level at such locations bued only on balanced conditions oRen under- Two methods of wiring may be commonly found in estimates the actual level ff a net current is present. houses: Romex cable and 'knob and tube." Modern The magnetic field caused by a net current in the practice is to use Romex cable, which combines distribution system will be more uniform throughout a supply, return and ground wires closely together in one residence than that caused by balanced currents. cable. Older houses may contain "knob and tube" wiring, which uses discrete supply and return wins The grounding system at the residence can be a very often separated by several inches and supported in the significant ~ouree of magnetic fields in the home attics and wails by a series of porcelain standoffs environment. It consists of the conneetious between (knobs) and insulating tubes. Experience has shown the service dr~p's neutral wire and' the electrical that houses with knob and tube wiring generally have ground at the residence. A common safety practice is higher fields. to ground the neutral wire at the service entrance by cotmeeting it to a metallic water pipe. This allows However, by far the most common situation found in neutral current from household appliances to return house wiring that results in unbalanced currents is through both the service drop neutral and the ground cotmeeting the neutral to ground at a location in the cormeetion to the water line (Figure 2-18). The current residence besides the service entrance. In houses with returning on the water pipe (the ground current) flows meallie plumbing, the neutral is usually connected to to the water main, and then to neighboring water pipes the plumbing by a ground strap. Some current may and service drop's neutrais. While done for safety flow through the ground strap back to the transformer reasons, grounding the neutral wire at the service serving the house, resulting in an unbalanced condition. enmmee eta cause large current loops that are a source of magnetic fields. Electrical appliances (garbage disposals, ice makers, whirlpool baths, refrigerators, etc.) connected to the The amount of the house current flowing back to the plumbing also may provide additional ground return distribution system through water pipes or other ground paths for the neutral current. These can produce return paths may be small. However, if some of the multiple current loops and be a source of magnetic current does return by a way other than the neutral fields within the home. wire, it most likely will be through the water pipe. Ontario Hydro Resetrob (Mader, 1989) has shown that they can predict residential magnetic fields by combining a measurement of the outside ambient magnetic field with the magnetic field calculated to be produced by ground paths currents. They calculate the ground path magnetic fields by measuring the current in the water pipe. Other possible ground return paths are through cable television lines, telephone lines, ground rods, connections to steel reinforcing rods in coneme floors and foundations, and equipment connected to ground. 2-16 Health Effects of E;cposure to Powerline Frequency Electric and Magnetic Fields Table 2.5 - Residential magnetic field source characteristics. (Source: EPRI, 1989) 5ource Spatial Distribution in living space Temporal distribution Harmonic Content Tramminion Lines Practically uniform Relatively uniform Practically zero Distribution Prlm~ Non-uniform Diurnal cycle Low 3rd harmonic 0-5 %) Distribution Secondary Non=uniform Very non-uniform High hemtonic content Net Current 51igh~y non-uniform Non-uniform High 3rd harmonic C20-150%~ Grounding Systent Very non-uniform Very non-uniform High up to I 1-17th hamtonics Unusual Wiring Very non-uniform Very non-uniform May be high Appliances Extrcntcly non-uniform Extrcntely non-uniform Depends on aFpliance EPR/(EPRI, 1989) lugs summarized (Table 2.5) the Magnetic field metca operate by detecting the voltage residential magnetic field characteristics for different induced into a probe by the magnetic field. The probe souroes. The table includes appliances as field sources clement is a shielded wit= coil. When the coll and observations on the harmonic content of the fields. encounters a time-varying magnetic field, a voltage will be induced in the coil bued on Faraday's Law of 2.3.6 Field Measurement Fundamentals Induction (see Figure 2-21). The device can be made moro sensitive by increasing the number of turns in the Although the type and purpose of a measurement coil The induccd voltage is proportional to magnetic dietarm the type of equipment used, all instruments rely flux density perpendicular to the plane of the coil. The on a few fundamental principles to detect and measure probe may be composed of a single element or ~ree EMF. Electric fields can be measured by three types orthogonad elements that simultaneously measure the of field meters as Figuro 2-19 shows: the free body magnetic fields in all three geometric planes. meter, the ground rofcrencc meter and ~c electro-optic meter. The free body meter metsum electfie field Hall=effect Gnuas meters, which measuro magnetic flux stzcngth by metering the current or time varying charge density from DC to several hundred hertz, arc also induced between the halves of a dipolc probe (see available, but because ~ey suffer from low sensitivity Figure 2-20). For many instruments the case itsclf is and from saturation effects due to the earth's magnetic the probe. The free body meter is self contained, field, authorities do not recommend their use allows measurements above at ground plane, and does (Mienklan, 1988). not require a known ground reference. The ground refcrence type meter measures the electric field 2.3.7 Cyclotron Resonance strength by metering the current flowing between a probe placed in the ~cld and a known ground To cxplain some inconsistent and ambiguous results of reference. Its use is restricted to measurements on flat bioeffects exporiments, some scientists have postulated grounded surfacet The electro-opticAl meter uses the a complex interaction between the earth's DC magnetic Pockets effect to measure electric field strength. The field and an extcrmtl AC magnetic field (i.e., from a Pockale effect is the chamgc in refractive properties of powerline). At specific combinations of intensity and certain crystals in the presence of an applied electric orientation of the two magnetic fields, it is suggested field and is propordomd to the f=st power of the that certain charged molecules (ions) of biological electric field strength. Although compact, the clectro- significance exhibit a resonance phcnomcna at optical meter lacks sensitivity in fields of less than 5 frequencies near the powerline frequency. This kV/m and is exportsiva. The free body and ground phenomena is known as cyclotron resonance. rofcrence meters have been available commercially for about ~e last fieteen years, whereas ~e electro-optical meter only recendy. Engineering and Exposure Assessment 2-17 E S~ probe (a) Box-type probe (b) E Insul~ljon Electrodes Parallel-pINe probe (c) Ground-reference probe (d) '~' E light in Polarizer Quamist wave plate Pockdim dled clystal light out Analyzer INTENSITY Eledro-optic probe (e) TIME Figure 2-19. Illustrations of three types of electric field meters: 1. free body: (a),Co),(c) 2. ground reference: (d) 3. electro-opticah (e) CIEC, 1987) 2-18 Health Effects of Exposure to Powerant Frequency Electric and Magnetic Fields conditions to mimic human exposure levels. For example, investigators might use a 65-V/m electric field on rats to simulate the same exposure of a lO-V/m field on humans, whereas they might use a 35-V/m field if the study was being done with pigs. The electric field level used in a study ~ depend on the size, shapo and posture of thc exposed subject, the pal~ular clcctric paramctcr (electric field, induced current, or inducai current density) being investigatai, and thc degree to which thc cxposai subjcct is groundcd. The clcctric ficld is perturbcd by conducting objects in thc field. Conducting objccts can act to increasc locally ! the elect, tic field intonsity. The amount of this concentration dcpcncts on the shapo and size of the objects and the orientation of that object to the electric field. The field intensity above the head of a standing v I grounded porson, may be as much as 15 to 20 times the ' unpon'urbed field (EPRI, 1988), whereas the field Georrtetrjc J above the bach of pigs or rats is increased seven or axis four times, respectively. Concentration effects vary based on the specific part of the body and the subject's posture, but in general, the concentration of the electric field will be greatest about those surfaces with smaller Him 2-20. Dingram of · f~e body electric field meter. The radii of curvature. device measures the current induced between its two isolated conducting halvet The induced current is proportional to the The amount of current induced in an exposed subject strength of the electric field. (EPRI, 1989) depends on the strength of the surface electric field on the subject's body, the conductivity of the subjcct's tissue in which the current flows and how well the ~' subject is grounded. To accurately simulate biological ~ effects caused by cu~f~.,~t, allowances for the differences between the conductive cress-section to the induced current of the human and subject's bodies must be made. The level of the electric field can be adjusted to get the same current density, which is the amount of current flowing threugh a given unit of area. Internal current density varies among different animals """t'~'v in a way similar to the variation of surface electric field strength. Thus the strength of the applied electric field will depend on the type animal and also the desired current density. FagurmZ-21, Diagram ofa magnstic field mt~r. The voltage The magnetic field is unporturbcd by pinenee of (V) induced in the coll will be proportional to the flux density biological materials; still in exposure studies, the (B) of the external magnetic field. (EPRI, 1991) strength of the magnetic field must be adjusted to compen_tq3_e for the differences in the perimeters of the human's and exposed subjects' bodies. Faraday's law, 2.3.8 Biological Scaring of EMF which relates the induced electric field to the external magnetic field, can be used to calculate the apprepriate The establishment of precise exposure of EMF in magnetic field levels for various body radii. laboratory studies on animals and tissue samples is complicated. Since EMF interaction in experimental animals can be very different frem humans, it is often necessary to scale the field to duplicate the suspected Engineering and Fjcposure Assessment 2-19 when the effects of an agent kre known. The 2.4 Exposure Assessment techniquca of the assessment are oftcn prcscribcd. Fundamentals Asscasmcnts in the novcl catcgor/occur whcn little is known about cither the agent or its effects, or both. 2.4.1 Introduction They can provide new information to further our understanding of the agent. For example, an exposure This section develops a working defmitinn of exposure assessment and was adapted from materials pr~-cd assessment might seek to define where the agent by Robert M. Pattenon, Temple University for: occun, in what mounts, and under what conditions, or Scminar on Ncw EMF Epidcmiologic Results And it might try to dcfmc who is exposed and the detail of Their lm lications, October 16-19, 1990. Its objcctivc their cxposure. In the latter case, it could bc an is to create an undcrstanding of the basic goals of integral part of a hcalth effects study that is looking for a causc and effect relationship. There may bc littic exposure assessment rather than to list strict criteria or pcrformancc guidclincs. From these goals, tictailed guidancc as to how or even what to measure. Most guidclines and criteria could bc dcrivcd as nccdcd. It extant assessments of EMF cxposure are in this also discussca thc general tools and strategica of categoxT. cxposure assessment, and describes the pat~icuhr instruments used with EMF. Data from EMF cxposure An asscssmcnt for regulatory compliance is relatively studies arc usai to explain what is known about thc straightforward bccausc the agent and thc techniques sourccsandnaturcofcxposurctothescagcnts. for measuring it arc given. Whcn trying to link environmental cxposurcs to an effect, however, there is 2.4.2 EXposure v. Dose normaUy a well-defined cffeet but littic notion of the cause. Thc cffeet might bc an increase in a certain discase in a population. Thc causativc agent, however, Exposure (of a person to an agcnt) is the joint can bc very difficult to determine because of the variety occurrence in spacc and time of the person and thc of cxposurcs of those with the disease. Evcn with a agcnt. It is different from dosc, which is the amount of strongly suspect agent, adcquate exposure assessmcnt thc agent interacting with thc peTson. As expressed by bccomca critical in identifying a cause and effect thc Council on Environmental Quality (CEQ, 1989), relationship. The better thc cxposurc data, the more "dose is the concentration or quantity or risk agent surely atruc link can bc uncovered. reaching tissues, organs or coils within the exposed organisms whcro damage may occur." Applying this The causativc agent may bc identified by using distinction, one could bc exposed momcntarily to a cpidcmiologic studies. Causality cannot bc established vapor that is toxic when inhaled and, by not breathing, until eertain supportivc characteristics of association have no dose. Onc could bc cxposcd to a systemic, between an agent and a discasc is demonstrated. The poisonous chcmical spihed on the skin, but unless it is characteristics include strength, consistency, absorbed through the skin thcrc is no dose and no spccificity, temporal relationship, biological offoct. plausibility, and coherence of evidcncc. However, bccausc of thnc, cost, inherent difficulty, or othcr A human cxposure assessmcnt gives a qualitative and constraints, cxposurc assessment is often the study quantitative picture of the cxposure of pooplc to agents. clcmcnt doric with least accuracy. In fact, as discussexl In general, exposure asscssmcnts try to dcfmc "(1) later, thc parametcr that should bc measured is substances that target organisms, species, or sometimes not even known. environments; (2) thc intensity of cxposure; (3) in what way; (4) for how long; and (5) under what conditions" According to a report by the Office of Scicncc and (CEQ 1989). Technology (1985), "cxposure assessment is often the most resource-demanding portion of thc (risk) 2.4.3 Reasons for EXposure evaluations...* Thc Task Force on Environmental Assessments Canccr and Hart and Lung Disease (1981) has stated, *Many authorities agree that the weakest link in our Thc reason for making exposure asscasmcnts can bc understanding of thc environmental health studies is our placed in two broad categories, dcacribed as routinc knowlcdgc ofhumancxposure." and novel. 2.4.4 Elements of an EXposure Routine assessments are common in cnvironmcntal Assessment regulatory compliance activities and in the day-to-day practicc of industrial hygicnc. They are conducted With a dcfmition of exposure and an appreciation of thc importance of cxposure asscssmcnt, the clcmcnts of an 2-20 Health Effects of Exposure to Powerline Frequency Electric and Magnetic Fields exposur~ assessment can be simply stated: Measures avenged value, the peak value, the time that the value ofaposure should mimic the receptor's experience of exce~s some stated level, or something else. The interest. From this statement, volumes could bc fdled time-averaged concentration, the peak concentration, with cookbook instructions covering nearly every and the amount of time that the concentration exceeds a conceivable situation. However, that is not necessary given level are all different exposure roetries. or even desirable. An exposure assessment can bc Depending on the agent, the values in each case may designed with two keys: the first is knowing what to be expressed in units, such as parts per million, mimic; the second is knowing how to mimic it. This is milligrams per cubic rueS:r, or milligauss. Consider explored using a common example. the experience of into-eat in terms of another familiar example, cxposuro to carbon monoxide. Suppose that an exposure assessment is to be made of indoor, ambient tt~nperature (like EMF, a physical An exposure assessment for cat~n monoxide might bc agent). Guidelines are found describing rules for made with a personal monitor, a device that one can setting out a monitor to measure air temperature in a wear and that measures ambient concentrations. One dweffmg, and they state that the monitor shaH: choice could be that it measures the average of all concentrations over some long time period, such as · operate continuously days. However, the body responds to carbon monoxide on a shorter time scale, so perhaps it would · be located at a height of 1.5 meters above the be appropriate to measur~ over a shorter period. CArbon monoxide exposure can be measured floor, essentially continuously, but the body may not respond rapidly enough to warrant continuous measurement. · be placed on an inside wall, away from drafts Besides, a peak level that occurs during exhalation and direct sunlight and provide an electrical would not really matter. Some intermediate time scale signal reflecting its measurement. would then be best, and it need be no shorter than the time it takes to complet~ a breath. Thus, a rational Based on these guidelines, a monitor is positioned and basis for an assessment of human exposure to carbon operated to assess (and r~tct to) air temperslure. monoxide might be: the experience of interests is the average concentration over the time it takes to complete Of course, this monitor is a thermostat for operating a one breath. (A monitor that sampled only the air home heating and cooling system. Consider how it inhaled would be even better but difficult and expensive might be deployed given the objective of mimicking the to produce.) Because the action of carbon monoxide in experience of inZrest. That experience is our comfort, the body is known, and because an indicator of which relates to the temperature in the room. The exposure persists, this example can be carried further. mimicking is done by measuring temperature with a th~rrnostat, which then drives the furnace or air The ambient level relates to the experience 0f interest condifioner accordingly. The thermostat is not placed for exposure, but the real experience of interest for on the ceiling because no one occupies that space. It is assessing the effects of carbon monoxide exposure is not placed on an outside wall because it is relatively the level at which it has combined to form cold (or hot) there. Similarly, no one ordinarily sits in carboxyhemogiobin in the blood. From the earlier drafts. The thermostat should be located to mimic definitions, that would really be an assessment of dose, one 's experience of air temperature and thermal not exposure. Hence, a better assessment would be comfort in the home. Sometimes, one thermostat is not based on dose, because that is where the biological adequato because of variability in the home effects are produced. Notice, however, that the environment, and two thermostats are used. That is just mechanism must be understood first. Knowing nothing a response to spatial variability of temperalure; about carboxyhemogiobin, one would assess exposure additional locations ar~ added to be sure that the to carbon monoxide. experience of interest is continnaHy mimicked. The point is that the same advice expressed by the The best assessment is conducted by the bodies of guidelines can be obtained by thinking about what is to those exposed. Blood carboxyhemoglobin level mimics be mimicked by measurements. the experience of interest better than monitoring could ever do because it reflects not only the level of With that broad idea of the objective of an exposure exposure but also the bedy's response. It acts as an assessment, consider the latter part of the above "internal monitor." The emergence of biological statement, i.e., the "experience of interest." This is monitoring in the practice of industrial hygiene is based sometimes expressed (or obfuscated) as the "exposure on uncovering similar responses to other agents. An metric.' The exposure metric tells what characteristic internal monitor for EMF would greatly increase of exposure is being measured. It might be the time- Engineering and Exposure Assessment 2-21 confidence in cpidemiologic studies of hypothesized directly related to current and proximity to the source. health effects from EMF. Because thicker wires can carry more current than do thinncr wires, and because field intensity decreases Carbon monoxide is an easy example for exposure or away from a source, it is presumed that higher dosc, bccausc the mechanism for its effects is known. magnetic fields occur closer to thicker lines. EMF is more difficult because of the "illusory nature of Worthtimer and Leeper (1979) divided their study reported effects'. As with most environmental agents, population into groups having presumably different EMF exposure can be recorded as time-averaged magnetic field exposures according to factors that measurements. With confli~ng laboratory results, with included wire thickness, voltage rating, and proximity reports of effects in windows of power, frequency and of nearby lines. temperature, and with conflicting results from cpidcmiologic studies, an exact idea of what should bc Others have employed source strength and distance in measured does not cxist. Alter considering ambiguous different ways. Some studies have used distance from results, some researchers have speculated that the sources such as transformers and substations as well as experience of interest is not the field frequency or from overbead lines ($trumza, 1970; McDowall, 1986; intensity per sc, but rather may be movement in and Coleman et al., 1985). Some studies of workers have out of a field of some undorefrained parameters. Given used the strength of the source, for example the voltage the range of possibilities that this notion uncovers, the of equipment at thc work place (Knave et al., 1979; 'parameters probably are indcterminablc as well as Nordstrom etal., 1983). undctermined. This seems to arise from a faith that a cause-and-effect rehtionship certainly exists, and we For studies of workers, however, the most commonly can uncover it if we just measure the right way. used cxposuro surrogate has been occupational classification, such as electrician, electronics worker, 2.4.5 ~xposurs Assessment Tools electrical engineer or welder. Virtually all published occupational studies to date have used this surrogate Exposure assessments can use one or a combination of (e.g., Milham, 1982, 1985a, 198Yo; Wright ctal., three basic tools: surrogates, models, and monitoring 1982; McDowall, 1983; Coleman ctal., 1983; Callc (CEQ, 1989). Surrogates are "stand-ins' for the actual and Savitz, 1985; Pearoe ctal., 1985; Tornqvist ct al., agent, and they ar~ assumed to have attributes that 1986; Spitz and Johnson, 1985; Olin et al., 1985; Stern rehtz to the agent under study. For example, wire cfal., 1986). It is not known whether workers in their codes arc surrogates for actual exposure to background so-ca!led 'electrical occupations" arc in fact exposed to residential magnetic fields. Models are used to elevated EMF levels. They are, of course, exposed to estimate or predict exposure from basic, available other agents, such as welding or soldering fumes and information and, for EMF, physical principles. solvent vapors, which might be confounding exposures. Monitoring would be expected to come closet to An exposure assessment based merely on surrogates is replicating exposures. at best speculative. Exposure Surrogates. Instead of monitoring exposure Exposure Models, The distinction between surrogates difec~y, a substitute or surrogate is oPen used- An and models can be difficult to draw. Somewhat as exposure surrogate is defined as a factor that is used to exposure assessment blends into dose assessment, represent exposure to an agent when a direct measure surrogates blend into models. The wiring configuration of exposure is not available. Number of cigarettes code might be thought of as a crude model for EMF smoked per day is a surrogate for exposure to cigarette exposure because it has some physical foundation: smoke. A stationary monitor collects surrogate data for thicker wires carry more current, which produces thc exposure of a mobile population. (In the extreme, higher field levels. Computer-generated models for one might argue that virtually all exposure assessment estimating electric and magnetic field exposure levels uses surrogates.) $urrogates fred wide and often have been developed by the Elcctxic Power Research creative application in health effects studies. Their use Institute (EPRI) and other groups, including the in EMF studies is illustrative and is outlined here. Bonneville Power Administration. Probably the most familiar surrogate for EMF exposure EPRI has developed the "EXPOsure CALCulation' of the general population is the wiring configuration of model, EXPOCALC, for transmission line electric and the electric power distribution network. It was used by magnetic fields. It is working on another, called Worthtimer and Lccpcr (1979) in their original study of RESICALC, for distribution lines and residential childhood cancer and has since been used repeatedly by sources. EXPOCALC calculates field levels near them and others (e.g., Fulton ctal., 1980; Worthelmer transmission lines and incorporates a time-activity and Leeper, 1982; Savitz ct al. 1988; Severson et al., model (discussed below) for estimating exposure. 1988). The basic idea is that magnetic field intensity is Model inputs include the environmental setting and the 2-22 Health Effects of Esposure to Powerline Frequency Electric and Magnetic Fields physical and electrical parameters of the line; outputs Other devices can give continuous readings of the ~cld include contour maps of field intensity and exposure intensity. When connected to a recording device, such histograms. RESICALC is similsr in concept, but it as a chart or tape recorder, they can collect data over will focus on the home and include the effects of time or space for later analysis. The most sophisticated distribution lines, household wiring, and unbalanced instruments now include a built-in computer to control retu~-n currents through water pipes. data recording and recovery. An example of the state- of-the-art is the EMDEX, also developed by EPRI. Kaune (1987) has built a model of magnefc field levels The EMDEX uses three orthogonal coils to measure based on a sample of Seattle residences, using the magnetic field, and it is capable of sampling at ram statistical relationships between wiring configurations from about once each second to about once every five and measured field levels. The most impoxunt input to and one-half minutes, giving a range of collection this "crnpirical regression model" is the number of frequency that might bc matched to environmental service drops near the residence. Distance from conditions. distribution lines correlates only weakly with measured ticIda. It is important to note that a statistical model 2.4.6 Exposure Assessment Strategies based 'on data collected in one locale may not be valid elsewhere. Using the objective of mimicking the experience of interest, some broad exposure assessment strategies can E_ix>sure Monitors. EMF monitoring instruments have be developed. First, though, it is recognized that not advanccd very rapidly in recent years, due lsxgcly to all EMF exposure is continuous. So, hcking an work by EPRI. Their technical bases of operation was exposure metric, the ideal exposure assessment would described in the section 2.3.6. yield continuous data on the exposure of the population being studied. The simplest way to do this would be by Monitors may bc classified as survey or personal, equipping everyone with a personal, continuously depending on their design and intended use. Survey recording monitor. But what happens when someone monitors require an operator and arc suitable for engages in an activity during which the monitor cannot stationary mcasurctncnts. Personal monitors do not be worn? What if the size of the study popuhtion is require an operator and can bc worn on the body, greater than the available number of monitors or the continuously muting ticIda to which one is exposed. resources to use them? Now something less than the continuous, popuhtion-widc ideal must suffico. How Many companies manufacture monitors and section this is done is what exposure assessment strategies arc 2.5.3 dcscribea in detail several instruments that are about: coming as near the ideal as possible under the commcroially available. They differ in how they technical constraints of available monitoring methods record mcasuren~cnts. Some, useful mainly for and the resouroc constraints of time, money, and surveys, indicate only the instantaneous field intensity, subjects. perhaps on a dial, with no way of recording or averaging. Some average aH field levels over time. Exposure was defined earlier as requiring a person and They yield an exposure metric of timceavcragcd levels, an agent to be at the same place and time. A expressed with units of kV/m or raG, which could be continuous porsonal exposure monitor essentially used to distinguish among average exposures of accompanies a person evcry,,vhcre and continuously workers. If the length of time of exposure is noted, the measures the amount of agent. The same measured total, integrated exposure could bc expressed by data could be generated by knowing the concentrations multiplying the average by that length of time, with of the agent at all times and places, and then units of OcV/m) hr or mG hr. The more sophisticated suporimposing these on the persoWs movements. instruments of this type collect d_s3n separately for Instead of using a monitor that moves with the person, different field level ranges; for example, <2 kV/m, 24 stationary monitors that record everywhere kV/m, 4-7 kV/m, and so on. Rather than averaging all continuously would be used (at least everywhere that exposures, they average in each of the ranges. the person goes). Variations and blends of these two approaches-personal continuous monitoring (or The Average Magnetic Field Exposure Meter, or possibly modeling or use of surrogates), or continuous AMEX, is illustrative of a simple, time-integrating, monitoring of locations combined with knowledge or personal monitor for magnetic field exposure. assumptions about a pcrson's activity (again, with Developed by EPRI, the AMEX measures exposure by possible use of modeling or surrogates)--cncompass storing the charge rehtcd to the cu, t~fi~ inducctl by the exposure assessment strategies for EMF. Historical magnetic field in each of three, mutually perpendicular assessments must rely on the latter approach. The coils. The total charge is proportional to the product of exact details are the result of the study constraints and the field intensity and the exposure time. i.e. integrated the invcstigator's imagination. exposure. Engineering and Exposure Assessment 2-23 Instead of continuous data, average or instantancous 2.4.7 Confounders data arc usually gathcrcd. Instantaneous data, also callcd grab samples, reflect the conditions at onc A confounder is an agent that is associated with both location and time, for example, spot mcasutcmcnts in the presumed, causativc agent and the cffcct being the center of a room. Average data are rcferrcd to as studied. Confounders arc important considcrations in integrated samples and can be avcragcd in space or exposure assessments because of thcir possible role as time, or both. Many grab samples can be compiled the true causative factor. Many potential confounding into an integrated sample, if conditions arc assumed to facton have been identified that are rclcvant to be constant or smoothly varying between samples. epidcmiologic studies of EMF. When occupational When data ~rc collected in the center of a room and classification is the exposure surrogate, the association used to represent the conditions throughout the room or may be confounded by exposures to other agents such residence, spatial averaging has been usumed. as solvents. In community studies that rcly on wiring Implicit time averaging is assumed when measurements configuration cedes for exposure assessment, rcsults taken only poriodicaUy arc used to represent continuous may bo confounded by other urban environmental cxposurc. (True spatial averaging of some chemical factors that have the same routing and usage density pollutants in air can be done with sophisticated sensing features u electric distribution lines. These factors techniques; an equivalent capability does not exist for could include air pollution from local strcc~ traffic, gas EMF.) lines, potable water and gwcr lines, or tolcphonc lines. Not all of these might be plausible confounders. On Another approach employs survey techniques, similar the other hand, benzene is a known carcinogcnic to an opinion poll: Individuals arc singled out as being component of automobile exhaust and may be a rcprescntazivc of the larger population, and their confounding factor. Well-designed studies include exposure is monitored. The rcsuRs arc cxtrapolatcd to exposure assessments of plausible confounders for an the larger population. study subjects. Still another approach uses time-activity patterns 2.4.8 Conclusion (patterns of how and where people spend their time) with mcuur~nent a_,~,~ fzom different Exposure assessments arc performed to document 'microenvironments,* or separately identifiable Icvcla of agents having known effects or to elucidate loc~tions. Microenvironments ~rc defined so that suspected ones. While vital to health effects studies, exposure chuactcristics can bc assumed similar within good exposure assessments arc ottcn lacking, in large spccitic microcavironmcnts, such as homes, but part duc to their inherent difficulty. Wc can design an dissimilar between microenvironments, such as ideal exposure assessment by following the simple between homes, substations, and offices. Monitoring notion that our measurements should mimic the studies characterize levels of the agent for each rcccl~or's experience of interest. For EMF studies the microenvironment; activity patterns can be developed important experience of interest is unclear, although the by sociological research. The time-activity patterns teU practice has been to measure time-averaged levels, as is where pooplc arc and how long they spond there, while done with other environmental agents. the microenvironment data supply the exposures. When combined, they yield an assessment of time- Tools for marking exposure include surrogates, avcraged exposure. models, and monitors. Surrogates, which arc substitutes for a measure of exposure to the real agent, Each new exposure assessment offers the opportunity have often been used in exposure assessments. EMF for a unique strategy. The major danger in this, exposure moniton have become rapidly more however, is that the ability to compare the results of sophisticated in' rccent years and arc finding wide use different studies can be compromiscd by differences in in a number of cpidcmiologic studies. how exposure is assessed. That is why it is important for new studies to include the techniques of prcvious Starting with the ideal assessment, strategies can be ones and for protocols to evolve to a common point. designed that accommodate technical and resource However, since wc do not know which, if any aspect of limitations. However, exposure data arc sufficiently EMF is bioactivc, or could bc the agent, using standard variable that careful exposure assessments will be measurement techniques might prccludc identification needed to identify any health effects on EMF. of the truc agent, if such exists. 2.5 Measurements This section will examine the types of measurements porformed in assessing fields and exposure, review any 2-24 Health Effects of Exposure to Powerline Frequency Electric and Magnetic Fields recognized standards for measuring ticIda from alternative method is to takc survey measurements of powerlines, and survey commcrciaUy available all areas that the subject is Likcly to frequent during a instruments used to makc measurements. normal daily routine and to record the time spent in cach arcs. Although limitcd in applicability for human Thc objcctivc of a study will Largely prescribe thc EMF subjects, this method is suitablc for caged animal mcasurcmcnt. Thc extent, accuracy, and duration of studics. The equipment designed to make exposure the data coUcction wi[l dictatc thc type of mcasurcmcnts is spe~ializcd basccl on thc application. instrumentation nscd. Thc mcasurcmcnts may bc Some dcvicca that arc to bc worn emphasize light covcrcct by an cxisting standard that further specifics weight and compact sizc whilc retaining mcasurcnlcnt the typc cquipmcnt and proccdurc uscd. capability. 2.5.1 Measurement Categories Eneinccrin~, Measurements. Enginccring mcasurcmcnts go bcyond the single point-in-space and Three specific categorics of mcasurcrncnts (Dict~ch, instant-in-time approach uscd by survey mcasurcmcnts 1988) arc offcn madc: survey measurements to find thc and thc periodic sampling or timc weighted averaging ~ckt strength at specific locations at a spcci~c time, usal in cxposurc mcasurcmcnts. If, as discusscd in cxposurc mcasurcrncnts to cstinlatc the cxposurc of a the introduction of the cxposQro asscssmcnt (section subject over time, and engineering measurements to 2.1), thc truc metric of cxposuro turns out to bc more morc completely characterize thc naturc of the fields in complicatcd than avcragc field level, thcsc morc an cnviroruncnt. cxtcnsivc mcasurcrncnts w~l bc ncccasary to f'md thc prcscncc and amount of thc particular property. Survey Measurements. Survcy measurements t'md the strength of EMF at a spccific point-in-space and Enginccring mcasurcmcnts attcmpt to morc completely instant-in4imc. Often mcasurcancnts will bc doric characterizc the EMF environment at a location ovcr a scrlally at scvcral points on a path perpendicular to thc long duration. Engineering mcasurcmcnts may include transmission line to fad thc fino's latcral field profic. ~cld strength, field polarization, field oricntation, Maximum and minimum ~cld strength may bc temporal variations, spatial variations, harmonic dctermincd by rotating the mcasuring probe about the content, wavc slutpc, transient content, and sourcc field cillpsc. The instant-in-timc naturo of thcsc identification. Engineering measurements would bc rncasurcrnents prccludcs detecting long term variations rcquircd to asscss the subjcct's exposure to fields of of the told quantities. The ~ llaturc of thcsc spcci~c w~indo~vs of frcqucncy and amplitude, or the measurements allows temporal variations occurring occurrcncc of complex interactions between ~cld during the measurements to confound determination of strcngth and oricntation of thc powcrlinc's fields and an accurate field proftic. the gcomagnctic field. Thc equipment most often nscd for survcy The cquipmcnt captoyed for making cnginecring mcasurcmcnts is the serf-contained frcc body meter. mcuurcrncnts is specifically dcsigncd and customizcd Thc meter may havc probes to mcasuro clectric ~clcis to thc problem undcr study. A typical system would or magnctic tclds. The mcter providcs either an consist of multiple sensors (each spcciaUy fabricated) analog meter movement or a digital display to show the dcsignccl and calibratccl to interface with a multichannel root-mcan-squarc (RI~S) valuc of the mcasurcd data acquisitionsystem- quantity. Survey meters havc no rccording capability, although some may offcr outputs for stand alone 2.5.2 Measurement Standards rccordcrs. They arc point-in-spacc and instant-in-timc devices. Selective electronic filtering permits only the The Institute of Electrical and Electronic Enginecrs mcasurcmcnt of 60-Hz powcrlinc frcquency ticIda. But, (IEEE) trst published a standard in 19'79 for making some matt3 allow ovcrriding the sclectivc fitcring, measurements of powcrtinc electric and magnetic folds crabling mcasurcrnent of harmonics or other entitled *IEEE Rccommended Practiccs for frcquencics bcyond the fundamental frcqucncy. Measurement of Elcctrical and Magnetic Fields from AC Powcrlincs.* Moro rcccntly, IEEE rcvised this Exposuro Mcasurcrncnts. Exposure mcasurcmcnts standard (ANSI/IEEE, 198'7) and the Anlcrican introduce thc time paramcter into thc ficld mcasurcmcnt National Standard Institute (ANSI) has approvcd it. rcgimc to fad the cxtent of exposure to power The standard cstablishcs a uniform procedure for frequency tclds that a subject cxpcricnccs. Such conducting survey mcasurcmcnts of powcrlincs and mcasurcrncnts arc of primar~ importancc to human and calibrating instruments, or morc spccitcally as stated in animal cxposurc studies. Thc most scourate way to the standard: mcasurc cxposurc for human subjects is to outfit the person with a recording cxposurc monitor. An Engineering and Exposure Assessment 2-25 "The purpose of this standard is to establish along a line parallel to the powerline (longitudinal uniform preccdures for the measurement of profdc). First, measurements of the lateral profile power freclucncy electric and magnetic fields should bo made at mid-span where the conductors are from alternating current (ac) overhead closest to the ground. Measurements should bc made powerlines and for the calibration of the meters from the center of the line to at least 30 m (100 R.) used in these measurements. A uniform beyond the outside conductor. At least five equally procedure is a prerequisite to comparisons of spaced measurements should be made under the phase electric and magnetic fields of various overhead conductor. Complete profde measurements should powerlines. These procedures apply to cormmenec in the region of inter~t beyond the outside measurement of electric and magnetic field conductor and progress successively to the opposite levels close to the ground. They also can be side of the fight-of-way. tentatively applied to electric field measurements near an energized conductor or Next, the longitudinal proftlc of the electric field should structure with the limitations outlined in ... be made, beginning at mid-span at the point (as this standard. * determined from the lateral profde) of the greatest field strength. Measurements at five nearly equal For measuring electric fields, the standard specifies the consecutive increments along a path parallel to the line ug of a free body meter, and for measuring magnetic should be made in both directions from the mid-span fields, a shielded coil probe connected via a shielded point for a total distance equal to one span. cable to a shielded detector. Acceptable calibration equipment and procedures for both instruments are Magnetic field measurements are made using the same preseribed bythcstandard. procedure as for the electric field*s lateral and longitudinal profdes. The standard specifies that all As Figuro 2-22 shows, the electric field measurement measurements (both electric and magnetic fields) be ptoeeduro is to make several measurements along a line made with the sensor at the height of 1 m above the perpendicular to the powerline (lateral profde) and ground. The standard specifies that the combined errors due to all sources (ealibration, temperature effects, I~ interference, operator proximity, etc.) should not exceed 10% for both electric and magnetic field i measurelnents. Recently the International Electrotechnical Committee i , ~ has also published a standard (IEC, 1987) for ,~'~'~'~"~' : ~.. .~ _. measuring power frequency electric fields. The IEC ',i~ ~, ,~ i ~ ,i, ,.x, standard is similar to the IEEE standard, with the exception that IEC recognizes the electro-optical ,,mm ~,. __,, ,.. ~.. ,, .,__ ,,.. ,,, ., _;:~ ._ electfie field meter, which is not widely used in the =,_,u ,~ ~ ~ ,t vmu~ ~r,,u ~ -. ~ United States. While these two standards address survey type · .~ measurements, no published standards exist that specify procedures for making exposure and engineering type measurements. B , /]~ 2.5.3 Commercial instrumentation ,~.~ ...,~r=,·. A review of manufacturers' literature shows that ,.- several magnetic field metors and a few electric field .e~,~,_..~. meters are commercially available. Although the ~.u.a mr. v~. ~ ~ ..~ ~ ~ o~m Committee's review of commercially available instrumentation is not exhaustive, it does indicate what Figur~ 2-22. Latenl profile and plan view of lEEE standardized is on the market and the capability of the equipment. Microwave News (Microwave News, 1989) published procedure for conducting survey measurements of the electric an extensive summary list of commercial instruments and magnetic fields from powedines. and that summary is shown in Table 2.6. Most of these (ANSI/IEEE, 1987) instruments are for making survey-measurements, 2-26 Heallh Effects of ~xposure to Powerlint Frequency Electric and Magnetic F~elds while only = couple are available for making exposure- EFMC has replaced the "De·o" meter with a low cost measurements, and none are available for making electric and magnetic field meter (Electric Field engineering-measurements. Please refer to the Measurement Co., 1989) consisting of a 3.5 inch fundamental section for a description of the general purpose digital multimeter and special electric measurement principles used in these instruments. and magnetic field probes. The probes convert the field values to a voltage signal allowing the actual field Survey Instruments. Survey instruments measure quantifies to be read directly on the rnultimeter's liquid power frequency fields at a single point-in-space and crystal display. The probes produce 1 mV of output for instant-in-time. Several instruments are commercially 1 mG of magnetic flux density or I V/m of elecu-ic available. field strength. The Model MiI6PLUS can measure electric fields as small as O. 1 V/m and magnetic fields Although no longer being manufactured, one of the ts weak as 0.1 raG. The unit's upper range is over 30 most widely used and versatile survey instruments is kV/m and a few hundred gauas for the electric and the "De·o" Power Frequency Field Meter (Electric magnetic fields respectively. The Mll6PLUS costs Field Measurements Co., 1985). It is a multi-function about $250. instrurnent-a free.body electric field meter with accessory probes and inputs for measuring other field Monitor Industries of Boulder, CO. sells a single- quantities. Besides electric and magnetic fields, the function magnetic field meter (Model 42A) (Monitor instrument can measure space potential, and the open- Industries, undsted) that uses a 6-inch diameter pick-up circuit voltage and short-circuit currants induced in coil to sense AC magnetic fields. Offering stable low- objecU by the fields. The unit's measurement range is level measurements, the Model 42A provides true RMS from 1 to 100,000 V/m for electric fields and 0.125 to reading with a fiat response over a frequency of 40 to 12,500 mG for the magnetic flux density. The 1000 Hz. The design avoids the problem of movement instrument reads magnetic field strength in A/m, which of the- probe in the geomagnetic field saturating the must be multiplied by the conversion factor of 12.5 to instrument. Another interesting feamro is an internal obtain magnetic flux density in milligauss. speaker allowing · qualitative indication of the frequency content of the magnetic field being The meter features · large analog dial, probes for measured. This is handy when attempting to identify magnetic field and space potential, and an eight foot the various magnetic field sources that might insulated handle used to isolate the meter fn>m the simultaneously be present. The 42A has an optional effects of the operator's body during electric field linear mode in which the meter's sensitivity is measurements. Output jacks on the meter allow the use proportional to the frequency of the magnetic field- of an oseilloscopc for further analysis of these fields. fields of higher frequency give larger readings. The price of the instrument with the optional linear mode is When using the meter on its more sensitive magnetic about $325. field ranges, the operator must take care to hold the probe motionless while taking readings. Any Holadsy Industries of Eden Prairie, MN produces an movement of the probe will cause a saturation of the extremely low frequency (ELF) field strength meter instrument by the earth's magnetic field. Electric Field (Holaday Industries, 1989) that measures both electric Measurement Company (EFMC)of West Stockbridge, and magnetic fields. The sensitivity range of their MA. formerly sold the meter for about $1200. model HI-3600-02 is I V/m to 200 kV/m for electric fields and 1 mA/m to 200 A/m for magnetic fields. It features true RMS detection, automatic ranging, maximum reading hold, and waveform signal output. An interesting option is a fiber optic coupled remote sensor for measuring electric fields without perturbing the field. Engineering and Exposure Assessment 2-27 Table 2.6 - Gaummmetetm and Dosirnetera: (Source: Microwave News, 1990) c . y,/.~.~ c.,.-'-~- .... ;""-'-; "-- w~Jme ~-'"'1 Ne. of StoW, Aeeunq (In./llm,.) C'~-- F.w. tlLhg. MolstdiMS $~(} 0-12kHz 0.1G-2Ok-G/3 2~t ix7xl.iq 6121}HamgingM~msRL Modd9'~0 $1,.~0 lOItz-t01d'tz 10nsC~-20kG/4 2.~% 8.8x4.~t11/~ df6m~. O~msto FL 32:007 Modal ~ $7.,0GO :21)F,z-IOklt mG-3GO tG/6 .0% tatT.,~14/19 Irmids. A~I _,~-,-~, 4048 (407) ~ Mod~ 99a3 M,800 ]01~-5OId~ raG-3 MG/7 1.0% lt17.~tt6/'36 eulpa m an oaillmoo~ ' ~A~,o/oF.21mmnlm, ltg. MFIdlO f~,~l} 51t~-IOki~ 0,1mCf.|0G/4 2.0% 15.2~4.(at10/d.6 13um=anlas_-:=_--'-mi ummfunudmomspa~r. p.O. tst961 ,iamhhmnlmm~ PA (21~) 3S'/-517,~ FAX C21~ 3667~12 Elmati: Fsdd m Moll 116 S75 ~-H= 0.1 mG-3IOG/4 3.0% 1.~tt..~7.,~.4 - -- k326 Moddl16+ FL~0 ~O-Fa 0.tnsG-2DOG/4 3.0% 4.75x2.,wotg/2 dlgiml~. it. ~,,,~,,,;~1o, MA012Ri EMI:~X-C $7.,01~ 40~0o I~z O.t mC.,2~ C. d4 3.0% 1,114,116_511.3 indmls~'. EMDEX (413) 637-1929;, FAX (413) 637-2926 POlnt 10412 (415) 1~-251|1 FAX (415) C=-='-~=Dr-~''~t Amo2--:'-_- LCD Eianlmltm~ Dmign, hm. ACGM-I $d~O 1-150 Hz 0.1 raG.-9 6/7. 1.0% 2x/aTII 9100 ~ Fmmu,/ ACGM-2 $990 20-150 i~ 0.1 mCf-9 ~ .0% 2~417/1 .~O, (612} 818-74"/3 Hdsday ~ Is~. H1-3~O4}2 11,195 50/(i}-Hz 0.1 raG- 20 615 1.0% l.M..~t!7/2.g ' is (wlhrdimmam' ou~s~,fetdB/dt 141~ Mmlt Dr. m) mmmasmsms. VDTNLF Eis~Pmiti,,IdN~d4 (612} 934-e2~ FAX (il2} 934-:MIM -_;_:: -_.._:.._~ ~ ~ Illlid i k. i!~-109 $59~ ~ HI tIG-2~14 2.0% 3014x7,O.9 LCD ,:~;' ~, E-fllld mst:homm(si~)lm~ :~lBmdmmM$oSL avuiidi.~il.-llgavildlfot Madmyra Elm:trtmim Dmiln Ame,~iam, In=. ddAG $495 O-100 HI 10 p6-2 G/3 0_5% 4x7..~.2~.9 '--'--~' with m~h fmld nmatullmima,~ablfor P. dmm, VA22DgO Cl1~) 471-1445 C .... "~ Vaylt Audio ...:..,l~.. u..~.s Merejim' hailmaim Mod~l ~ $350 40 HI-I kH~ 0.01 raG-2.5 G/12 7-10% 2.1x3.1xT.g/1.0 - - - 6112F~atmiinCany~n 003) 442-37/3 C--:-~-: Ea I ~ Pmiu~imhsmim, lag. Dooimmm' $1,650 ~O-HI S}~G-4G/ 5.0% 613x1~O_5 Ouq~no, ~ . Stm~ .~101 tatsram ~t., ~L Mortreal, 311,101 (~-20 MHI) ~adatm6e da~ of dam. Med=l 379102 Qodg, H4P-2R9, Casmda (514} 34~-2~1]1~ FAX (,~14) 731-112 .SabCempmingCampm~ ~-~Mmmr $1~ 20HI-3Okitx t~C,-230m6/7 5.0% 6t~4,~}.7 Sa~....~t -' Nmlam, MA {}2IN Mmm' Sl75 5HI-tld-b. 0.1m6-2fifim611 3.0% 5.Sx3.3xl.5/0-1 ~dm. Bahmmm, mfor (611) dUM-7771, (~0) 222-]0fi3 (I-40 kltg} $29.95 Sdmm~m. Appllml To,:tmoi~y Mall EMI $90 tO HI-t kHz 0.610+ mG/I 5.0% 5_sx3.1xl_5/l)-8 · 2fiOMiluan~., thitlR CRams for mnlm Im, dmmO.~ mt Nm,lmm. MA. ~ MO.0OAvk) Mail EMIO Inmu0, imp (617) 320-9900. (~0) 3~6.5500 di,play. C-:-'- Joim~ ' · 2-28 Health Effects of E~cposure to Powerlent Frequency Electric and Magnetic Fields Sydkraft of Sweden rnarkeu separate electric and recording of the field readings. The soilware, which magnetic field meters (Sydin'att, 1988). The magnetic runs on an IBM Personal Computer (PC) or flux density meter operates over a range of 0.001 to compatible, retrieves the field data stored in the 2,000 ~T and is accurate within 5~ over the entire instrument's microcomputer and then analyzes and range. The manufacturer has provided both an analog displays the readings. and a digital display for displaying values in/~T of magnetic flux density. Their electric meter is a free- The EMDEX instrument is compact and portable-it body meter and covers the range of 0 to 40 kV/m. measures about 6.0 x 4.5 x 2.0 inches and weighs 16 Electric field strength is displayed directly on a digital ounces. It can simultaneously measure and record the display. avenge electric field on the user, the three orthogonal components of magnetic flux density, and the rotational Integrity Electr~}nics and Research of Buffalo, NY sells motion of the meter in the geomagnetic field. The a Model IER-109 magnetic field meter (Integrity electric field mcter's measurement range is from 0 to Electronics and Research, 1989) that measures 50 kV/m and the magnetic flux density range is 0 to magnetic flux density over a range of 2 to 2000 raG. 25,000 raG. The accuracy over all but the highest tier The narrow band response of this unit prevcots it from of the range is 5%. EMDEX has sampling intervals measuring any contribution from harmonics are any from one per second to one every 5.45 minutes. frequency removed from 60 Hz. The unit has a Standard sample intervals and their resulting data sensitivity of 0.001 mG and is accurate within 2%. collection durations are shown in Table 2.7. Other features include audio and visual alarms, a 200- mV chart teeorder output for dosimctry, and an Table 2.7 - EMDEX Sampling Intervals (Source: EPRI, 1998) optional 3=axis magnetic field probe. The IER-109 sells for about $600 and many accessories arc Sample lnt-rval ApproximateData Collection available. Crime Between Samples) Time Limit I ~cond 10.6 Hrs. <0.4 days) Exposure Instruments. For making cxposuro 2.5 seconds 26.5 Hrs. <1.1 days) measurements three instrument are commercially 5 seconds 53.0 Hrs. (2.2 days) available: EPRI's EMDEX, Positmn's Electr~magnetic Dosimeter, and Combinova's MFM10. Although the 10 seconds 106.0Hrs. <4.4days) prinmry purpose of these instruments is to measure 15 seconds · 159.0 Hrs. (6.5 days) field levels over time, they can also be used for making 30 seconds 318.1 Hrs. <13.2 days) survey measurements. 60 seconds 636. I Hrs. ~26.5 da~s~ EMDEX (Electric and Magnetic Dose Exposure) is a The settware allows the user to download the data from hardware and soRware exposure assessment package the EMDEX instromcnt to an IBM PC . Once resident developed by EPRI (EPRI, 1988). The hardware is a in the PC, other programs process and analyze the compact self-contained electric and magnetic field data, allowing it to bc displayed in graphical or tabular meter coupled to a microcomputer for periodic Engineering and Exposure Assessment 2-29 form. An example of the graphlea1 display is shown in exposure with a single axis probe worn on the wrist. Figure 2-23. Measurements from the single axis probe did not capture the subjcct's truc exposure. Conscqucntially, Although the EMDEX was designed for exposure EPRI is developing (EPRI, 1989) a new version of the measurements it is very useful for making survey AMEX that will have a 3-axis magnetic probe and measurements. Electric Field Measurement Company should solve this problem. A new more compact, offers an enhancement package for the EMDEX that improved model of the EMDEX meter called includes a bicycle wheel distance measuring device. EMDEXII is also under development. The EMDEX can be programmed to read field values at an increment of distance as measured by the bicycle The Elcetromagnetic Dosimetcr is a 5-channel wheel. Profiles and contours of powertines, recording dosimetcr (Positron Industries, 1990) substations, and even residences can be rapidly made manufactured by Position Industries under a license using this feature. The enhancement package will from Hydro Qucbec/]REQ. The unit simultaneously include soRwarc to generate field profiles of powerlines measures and records electric field, magnetic field in and buildings. three axes, and electromagnetic radio frequency (RF) disturbances. The Positron Dosimac. r is considerably The EMDEX is available from and available from lighter and smaller than the EMDEX instrument, it is Electric Field Measurements (under license from EPRI) about the size of a Sony Walkman. for $2000. The enhancement package is $800. The instrument measures and records electric and Although not as capable as the EMDEX nor magnetic fields at power frequencies. The values of commercially available, EPRI has also developed the the field measurement is stored within 16 separate bin Avenge Magnetic Field Exposure (AMEX) meter. The registers. Each bin represents a specific frequency AMEX is similar in idea to a radiation film-badge range; therefore, the intensity of the field is indicated dosimeter. It measures integrated magnetic field by the number of bins filled. The 16 bins are divided Flux Dwn'm. itv vet'mum. Tiww SeWS · i -- lwmg ,.. . . :: 13 .Se t8.88 18,58 11.88 tt .58 t2 Sg 12.58 13.88 t3.58 Tine or ]3a~ (hrs). Seq. Start: Thu Feb 22 89:c:3:tS 19ge Figran 2-23. Sample output of magnetic field exposure history from Electric and Magnetic Field Exposure (EIViDEX) Meter. Data from the meter is download to a personal computer for analysis and graphicai display. 2-30 Health Effects of Exposure to Powerline Frequency Electric and Magnetic Field~ m a non-linear manner over the full sensitivity range of 2.5.4 Conclusions the unit. Instantaneous readings of thc electric, magnc~c, and RF impulse bin levels are displayed on a liquid crystal display (LCD). The Committee concludes the following regarding measurement of electric and magnetic fields: The instrument can measure electric fields ranging from 0.61 to 20,000 V/m and magnetic fields ranging · Field measurements fall into 1 of 3 from 0.12 to 2000 raG. The magnetic field is categories: survcy. cxposure, orengine~ring. measured, recorded, and stored separately for each axis (X,Y,Z). Electric and magnetic fields can be · The instrumentation, procedures, and sampled at 1, 5, or 60 second intervals providing standards (if any) used to make field continuous recording periods of 8, 24, and 168 hours measurements will depend on the purpose and C7 days), respectively. The fifth channel effectively type of study being performed. counts th~ number of high intensity RF pulses (greater than 200 V/m) occurring in a frequency range of 5 to · Standards for making survey measurements 20 MHz. of the fields from AC powerlines have been Data from the instrument is download over a standard published by ANSIflEEE and IEC. No RS-232 port to an IBM compatible desktop personal standards exist for making exposure or computer. Sottware is provided to disphy graphs of engine-ring measurements. chronological records of the fields at two different time resolutions, total delivered dosages, average values, as * Survey instruments are readily available from well as hiamgrams of the d~ta. several manufacturen for measuring both electric and magnetic fields. Exposure The Electromagnetic Dosimeter is available for either instruments ax~ available, but only from a few 50 or 60-Hz frequencies from Positron Industrn, INC. manufacturers. Packaged engineering of Montreal, Canada. It cost about $1650. systems are not commercially availablo. Combinova AB. a Swedish company, manufactures a :2.6 EMF Exposure Estimates rex:ording magneti~ field meter (Ergonomies, INC., 1989). Their instrument, denoted as MFM 10, is larger (15 x 10 x 4.5 inches) and heavier (6.5 !bs.) than Olten for a variety of reasons, assessing exposure the EMDEX and it only measures magnetic fields. The requires estimating instead of making actual field measurements. This section will desert'be three measurement range is 0 to 10 G with a 2% accuracy. The MFM 10 can store more than 4,000 readings and methods of estimating exposure: computer models that like the EMDEX download the readings to a PC for calcuhte exposure, the extension of spot measurements analysis. Available from Ergonomics, INC. of to characterize exposure, and substitute properties, .Southampton, PA, the MFM 10 costs about $6,800. caUed surrogates, which are assumed to be indicative of EMF exposures. Eneineering Instruments. The Committee's search did not identify any commerciaUy available systems for Several situations occur when estimating may be performing the vepJ detailed and exhaustive preferable to measurement: when attempting to assess measurements needed to characterize completely the exposure for conditions when measurements are not EMF environment. Although the components of possible, which is during extraordinary operating necessarily sophisticated systems are available (i.e., conditions not easily created; when attempting to field probes, multichannel data loggers, and powerful quantify historical exposures, which is necessary in personal computers or engineering workstations), the occupational and residential epidemiologic studies; or Committee made no effort to study the integration of when the time and expense of making measurements is these elements into useful systems. Although most of prohibitive and large amounts of data are required. the components do exist in 'off-the-shall' form, the Three of the most common means of estimating Commitno notes one key ingredient is missing: the exposures are computer modeling, extrapolations of solbA, are to analyze and make sense out of the vast spot measurements, and the use of surrogates. mount of daut required to be coUected. 2.6.1 Computer Models Several computer programs are available that calculate the electric and magnetic fields produced by powerlines. An informal survey of Texas utility Engineering and Exposure Assessment 2-3 1 companies, shows three programs arc used (APPA), an extensive set of computer programs predominantly: BPA's CORONA, EPRI's (Brown, 1989) for the design and analysis of electric EXPOCALC, and APPA's TRANSPAC. power transmission lines. The programs arc divided into three categories: environmental effects, mechanical BPA's Corona. The Bonneville Power porformance, and electrical performance. The Administration's Corona and Field Effects Program programs run on the IBM PC or compatible with at (CORONA) was first dcvclopat as a main-frame least 512 K of memory and two floppy disk drives. computer FORTRAN code for predicting audible noise, radio and telcvlsion interfertmcc and ozone production The environmental effects group contains four from AC transmission lines. The program was later programs to calculate electric fields, magnetic fields, exlcnded to calculate corona loss, electric field, space audible noise, and radio noise. The mechanical potential, magnetic field, and conductor surface performance group features two programs to calculate gradkmts for either AC or DC powerlines. It calculate~ conductor temperature rise and current carrying nominal field values for DC lines, but does not. capacity, and conductor sag and blowout (i.e., physical calculate ion-cahanccd field effects. The program can displacement of the line duc to wind loading). The calcuisle the effects from a combination of up to 50 electrical performance group contains six programs that phase or pole conductors and overhead ground wires, calculate line electrical impaiance and admittance but it is unable to analyze hybrid systems consisting of parameters, line capability, line shield and insulator both AC and DC conductors. losses, switching surge performance, lightning performance, and tower footing ruistance. These With the arrival of more powerful PC's, users ported programs arc independent of each other and can be run CORONA from mainframes to IBM PC or compatible in any order, but they all use the same input data sets. systems. Data input to the program can be from either the keyboard or disk flies. Input da~ arc a physical Input data to TRANSPAC are very similar to the input description of the Free-the number of phase and ground to the BPA's CORONA, although ' more conductors, the number and size of the subconductors straightforward. TRANSPAC accepts phase-to-phase making up bundled conductors, and the physical voltages and allows the phase relationships between arrangement with respect to some reference datum of phases to be specified explicitly by a phase angle in the phase and ground conductors. Other data inputs arc degrees inste~,_d of using complex values for each the line-to-neutral voltage of each' phase conductor, the phase's phase-to-neutral voltage. TRANSPAC is phase currents, and a host of environmental slightly limited because it allows only half the total parameters, and sensor height and location ,!_~. number of phase and shield conductors as CORONA- 25 versus 50. In most applications, problems seldom The user has the option of generating one or all of the involve configurations of more than four or five following rel~rts: audible noise, radio interference, separate circuits on a transmission right-of-way. television interference, ozone concentrations, corona loss, electric field, and magnetic field. Each report is a The electric field program's output report lists the tabular listing of the output data. The program does not average maximum subconductor surface gradient for feature any graphical output capability; however it does each phase and shield group. It generates a tabular offer the option of creating a flic of calculated results in listing and graph of the magnitude of the maximum data interchange format (DIP"). The DIF flie can easily electric field strength that occurs at each increment be loaded into any one of several popular spreadsheet along the lateral profile. Whereas CORONA lists the packages, such as Lotus 123 or MicroSoit's Excel, for magnitude and angle for the maximum and the vertical generating graphical plots of the powcrline's lateral and horizontal components of electric field strength, prof~c. TRANSPAC lists only the magnitude of the maximum with no detail about its orientation. Both CORONA and Appendix A contains CORONA's electric and magnetic TRANSPAC are limited to a maximum of 100 data field output reports for a typical 345-kV transmission points. llne. TRANSPAC's magnetic field calculation accounts for BPA has placed CORONA in the "public domain" and the effect of currents in the shield wires and ground permits it to be freely copied and distributed. The only re. turn currents. Neither CORONA or EXPOCALC stipulation is thai BPA be credited as the source of the offer this capability. The user must supply earth program. It is free to anyone for the asking and in resistivity and permeability to use this feature. wide use across the country. Inclusion of the shield wires and ground return paths affect the contribution to the magnetic field results APPA's TRANSPAC. David R. Brown, INC. has slightly as shown by comparison of the output in developed, for American Public Power Association 2-32 Health Effects of F. xposure to Powerline Frequency Electric and Magnetic Fields Appendix A. Because of this ability TRANSPAC may The approach EXPOCALC uses is fu-st to characterize be the program of choice when: the physical geography of the study area. Second, the subject's activity patam is mappal to the physical · Calculating magnetic fields at locations a geography to calculate time spent m each location. great distance from the tint, wherc Third, electric and magnetic contours for the electric contributions from ground return path can bc sources arc calculated for the area of the physical significant; or geography. Fourth, the model calculates the ccluivalcnt field for the specific activity by applying an activity factor to the calculated unpermeoed field. The activity * Proving in rcgulatory and judicial prc-:__-:~ing factor describes a known physical reference condition that all effects (including shield wire and (e.g., current induced into a body standing erect, with ground return currents) have been accountod arms at side) to the condition occurring in the activity for in the calculation of magnetic field levels. (e.g., percent of reference current for a body sitting on a tractor with arms raised). FiP. h, the model calculates The magnetic field roport is similar in format to the the index of exposure by summing the product of the electric field report. It lists the calculated currents in time spent at each location and the value of the field at each shield wire and then features a tabular listing and that location for all activities. graph of the magnitude of the maximum magnetic flux density in milligauss. Again CORONA provides more EXPOCALC features function menus, built in data information by giving both the magnitude and angle of editor, and graphicai output dircc~y to the screen or the maximum magnetic flux density and its vcrtieal and printer. The user has the options of creating a new horizontal components. TRANSPAC offers the choice case by entering data from the keyboard, or loading of including or ignoring the effects of the currents in and modifying an existing case, or running an existing the shield wires and earth returu currents. case by loading an existing data f~c. TRANSPAC allows the reports to be sent to the printer EXPOCALC's input has the physical description of a or to a disk file. Unlike CORONA's DIF file option, study arcs surrounding the transmission tint, the TRANSPAC does not offer an easy method of electrical description and physical orientation of the transferring output data to other programs. It is phase bundles and shield wires, and thcphasc-to-phasc possible to use a text editor to extract the data from the voltage and current for each phase. EXPOCALC output file for "importing* intq. a spreadsheetprogram. assumes a three phase transmission line with each phase separated by 120 electrical degrees. TRANSPAC is available to members through APPA EXPOCALC often the unique capability to account for for a nominal distn'bution fcc of $150 and available to the shielding of the electric field by buildings, fences, anyone cisc through David R. Brown, INC. for $2,995. vegetation and other objects. EXPOCALC also AppondLx A contains a sample run of a typical 345-kV accounts for the sag in the transmission conductor transmission line. between supports in calculating the electric and magnetic field contours. EPRI's EXPOCALC. EPRI has sponsored the development of a computor model that assess personal Users can select to run an electric field analysis, exposure to the electric and magnetic fields of magnetic field analysis, or both. They have the option transmission lines. EXPOCALC (Encrtcch, 1988) of viewing the results graphically, through contour combines field calculations similar to those performed maps or histograms, or generating the results in a by CORONA and TRANSPAC with activity modeling tabular format. They can direct the tabular results to to csthnatc the fiffac a person engaged in a particular the screen, printer, or a data fde. Although activity spends in various levels of fields from neatJoy EXPOCALC generates horizontal contour plots, it is transmission lines. The program is designed for use on unable to generate lateral prof~es. a microcomputer by persons with limited computer knowledge. Engineering and Exposure Assessment 2-33 Like TRANSPAC, EXPOCALC calculates the different electric phasing and teinporal magnitudes. magnitude of the maximum electric strength and Sources of magnetic fields considered ar~ distribution magnetic flux density. It omits giving the angle of the primary and secondary conductors, water pipes, magnitude. The program also does not support any household wiring, and other sources. cl_a_ta exchange formats, so transferring data to other programs requires editing the output files into a form 2.6.2. Spot Msasursmsnts other programs can accept. A copy of the input and output reports for the typical 345-kV transmission line Another method (5ilva, 1988a) of assessing exposure is arc included in Appendix A. to use a single point-in-space, instant-in-time value obtained from a survey measurement to represent the EXPOCALC is available to EPRI mcmben for a field at a location. The chief shortcoming of this nominal distribution fe~ and available to anyone else proceduro is that it ignores temporal variation of the under license from EPRI through Enertech Consultants fields. This may be acceptable for electric field for$1200. exposure estimauu, since the electric field from tnnsmlssion and distribution lines are almost constant. Comparison of Results. The electric and magnetic However, magnetic fields vary constantly and such an fields for the same 345-kV transmission line were approach prevents characterizing the magnetic field's calculatai by CORONA, TRANSPAC, and spatial and temporal variability. The spot measurement EXPOCALC and the results comparaJ. The lateral approach can bc improved by sampling the magnetic profile was plottai by importing the results into field over a period. But, that period (usually 24 hours) MicroSoWs EXCEL spreadsheet. The proides start at may not be long enough to capture weekly and seasonal the center of the tnnsmlssion line and extend out 500 variations. Also the method portrays the cntir~ area's feet. As the plots for beth the electric and magnetic fields based on the field in only one location. fields show (see Figures 2-24 and 2-25), all three programs give almost identical results-the three curves 2.6.3. Surrogates overlay each. A common method to assess exposure to EMF is to Other Programs. EPRI is developing another choose a substitute physical quantity that may reliably program called RESICALC (Sussman, 1988) that will indicate exposures. Use of a surrogate is o~cn calculate residential magnetic fields. RESICALC desirable because of economic and time constraints. allows development of complex residential models that Surmgate. s that have been used to estimate electric and contain multiple line and point sources each having magnetic field exposure are various wiring codes, Elednc Fiddl ~,d by Progrlm~ ~4~KV~n~eC;i~laifT~Line 0 o 50 100 15o 2oo 250 300 35o 4o0 450 500 t)ia~n~ from c, ntef Line BPA -- ~ -- F'~gure 2-24. Compadaon of electric field profile calculated by BPA's CORONA, EPRI's EXPOCALC and APPA's TRANSPAC. The three program yield almost identical values and the three curves overlay one another. 2-34 Health Effects of Exposure to Powerline Frequency Electric and Magnetic Fields Magn~(i~ Fields Calculated by Program~ 160,0 140.0 120.0 ~ 100.0 ~ SO.0 o- 60.0 40.0 20.0 .0 0 50 100 150 200 250 300 350 400 450 500 Distance from Center Line (Ft.) Fal~re ],-~,,{, Compari=on of mala~¢ 5¢1d profile calculated by BPA's CORONA, EPRJ's ~]~ICALC and APPA'= ~SPAC. The ~ee pm~q'am ~ield almost idemicai values and the ~ curves overlay one sao*.ber. distance to powcdjnes, number of service drops, (confoundors) li~ely to be =ssociated with causin8 strcngth of the source, and occupstional classificJtion. cancer. W=chtel m=intains Oat wiring codes m=y be surrogites for the ,~ and tTp¢ of housing construction, One ~arLr~ code system dcvelolx~l by Wertheimcr and the density of housing, and the mount of vehicuiar Leepcr (Wefthelmet, 1979) attempts to characterize the traffic in an mrca. historical exposure of a location to rrmgnetic fields based on the mount of eurrcnt the external wiring can Confounden associated with the age and type of carry and the proximity of the wiring to the location. housing censtruetion my be: presence of toxic Locations arc divided into onc of several categories building material (e.g., fommidehyde, asbestos}, such as: seepage of radon gas, and proximity to toxic waste sites or contaminated water sources. Thc confoundcrs VI~CC- Very HigA Current Configuration associated with traffic and housing density may be air pollution and noise lcvcls, and lack of recreational O]:ICC- Ordinary High Current Configuration space. In fact, Savitz has rocently analyzed the Denvcr locations (Savitz, 1989) for correlations between childhood cancer and traffic density and found risk OLCC- Ordinary Low Current Configuration ratios very similar to those associated with magnetic ficld exposure. I/LCC - Very Low Current Configuration 2.6.4 Conclusions The important factor in application of a surrogatc is proper understanding of the accuracy of the indicator · Scvcral computer programs are available to and whether the surrogate represents only the suspected quantity, For instance, Bracken (Bracken, 1988)noted accuratcly calculate clcctric and magnetic that Wertheimer-Leeper code only accounted for about fields. 20% of the variation found in one magnetic field measurcment study. Keam (ICeam, 1988) concluded, · CORONA, TRANSPAC, and EXPOCALC aP, er rcanalyzing the data from the various Denver arc three programs in use by Texas utility epidemiological studies, that the Wertheimer-Leeper companies. All three programs yielded almost cede is not a good surrogate for magnetic fields. identical results for the sample problem. Also, critics C~/achtel, 1988) claim that wiring codes * Under controlled conditions, spot can confound thc rclationship between magnetic ficlds mcasurcments may be combined with the and canccr sincc they may be indicative of other factors Engineering and Exposure Assessment 2-35 subject's activity patterns to estimate As part of an cpidemiologic study, magnexic ficld measurements were conducted in 483 Denver exposure. residences under low and high power conditions · Surrogates must bc used with great care since (Savitz, 1987). The mean low power value was 0.76 mG with a standard deviation of 0.79 mG, while the they can oRen be indicative of other factors mean'and standard deviation for high power were 1.05 besides power frequency EMF, which also and 1.3 raG. Wachtcl (1986) pointed out that the may be associated with cancer. diffcrcncc is not statistically significant. 2.7 Preliminary Field An c-pidcmiologic study in Se, aulc incorporated electric Measurements and magnetic field mcuuremcnts to assess residcntia! exposures of cancer cues and controls (~unc ctal., Dais hsvc bccn conccted thst charactcr~c field levels 1987). Dais wcrc coLlecteel for 24 hours in 43 homes. in residences and work places. Source The clcc~c fields mcuured near a wall in the family characterizations have included mcuuremcnts near rooms averaged 33 V/m, a value ~cater than the sho~- applia~cu, u well u sources specific to ccrisin term mcuuremcnts t~cn in the center of the room. occupations. The dais collection procedures have The difference was attributed to differences in robed from spot, ira mcuuremcnts at one location, proximit7 to electrical wb'inB. The researchers found to continuous arcs monitoring of a location, to personal no relationship between electric and magnetic fields, monitoring. The findings of maay of these studies, between electric fields and wirinB configurations, and were dcisiled in mstcr~ls prcl~srcd by Robert M. between 24-hour and spot Hcctric field mcuuremcnts. Patlcrson, Temple University, for; Seminar on New They concluded that localized mcasuren~cnts are not EMF Epidcmio]ollic Results and Their Implications, useful for characterizing electric field exposures. October 16-19, 1990, are summar~7,cd below. A somewhat different picture emerged from the 2.7.1 Residential Exposure ms~ncdc field dais in the Ksunc study. The mean value in ~c family room was 1.0 raG, with a standard deviation of 1.2 mG and a median of 0.6 raG. Mcuuremcnts of electric fields in 10 rural residences Simulisncous measurements in ~c bedroom gave 1.0, in W'aconsin and Michigan showed electric field 1.4, and 0.:5 mG for the mean, standard deviation, and intensities robing from 2 to 65 V/m, with a mean median, respectively. There was a diurnal variation valuo of 16 VIm and a median of 14 VIm. MaBnaic that coincided wiffi electric utili~ loads, pcaYjng in the field mcuuremcnts st the same locations yielded a morning and evening, with low levels in the very early range of 0.07 to 1.6 raG, with a mean of 0.8 mG and a toomanB. For an individual residence, there was no median intcnsi~ of 0.4 mG (IITRI, 1984). correlation with power consumption. There was a significant correlation between twenty-four-hour and The EPA measured 25 homes in Las Vcgas and found short-term levels. The researchers concluded that the an average of 5.8 VIm for the electric fic]d and 2.:5 mG predominant sources are external to the home and that for the magnetic field. The ranges were 2 to 12.7 VIm Found return currents could bc an important source. and 0.6 and 7.8 mG (Tell, 1983). Both the Denver and Seattle studies tested the In New Jersey, Cao]a ctal. (1983) mcuured the Wcrthcimcr-Lccpcr wL,-c code apinst their measured electric fields at homes that were 1610 m, 95 m, and dais. Savitz (1987) found that the w~c codes gave the 1682 m from a 500-kV transmission line and found a correct average ranYjng of magnetic field levels among mBc of I to 20 VIm. Mcuuremcnts in a Pittsburgh homes. Kaunc ctal. (1987) found some correlation apartment revealed a non-uniform distribution of bct~vccn measured levels and wb-c codes, hut the wLrc electric field levels throughout (Florig, 1986). codes explained only 20% of the variance. As discussed carlact, hc constructed a reFcssion model of Barnes (1985) reported electric field levels measured in field levels. The most important predictor was the thirty-six homes in Denver under two conditions. number of service drops near (within 43 m)of the Under "low power" conditions, with lights and residence. The presence and proximity of transmission appliances off, the mean value wa~ 7.5 VIm. "High lines were also important, as was the number of power" conditions, with llghts and appliances on, primar7 phase conductors. There was no significant produced a mean of l0.4V/m. correlation with dislance to proarT and secondary distribution lines. 2-36 Health Effects of Erposure to Powerline Frequency Electric and Magnetic Fields Silva ctal. (1988) measured magnetic fields in 91 Figure 2-26 summarizes many of these data on a buildings in six states and found the vertical component common plot (Bracken, 1988). Geometric means and of the field to be dominant. Values mcasured at head standard deviations are shown when available. When level had a mean of 1.2 mG and a standard deviation of these were missing, the data were assumed to be log- 2.5 raG. Thc maximum was 63 raG; the data appeared normally distributed and a transformation was made. to follow a log normal distribution. In houses, thcy Missing standard dcviations were assumed to be 1.37 found that field levels depended on the method used for raG. The figure shows consistency among data in the grounding. Of 31 homes with a local ground rod, the United States and Sweden, with lower values recorded mean vertical magnetic field was 0.65 raG. It was 1.9 in the United Kingdom. mG for homes grounded to the water system. Electric fields near appliances average 30 to 60 V/m at In other studies, a median level of 0.15 was found for one foot, but they fall off rapidly at greater distances. 44 homes in the United Kingdom (Mycrs ctal., 1985). Lcvcls can reach 960 VIm at the chest under an clectric Tomcnius (1986) recorded an average of 0.7 mG at the blanket (Florig, 1986). Preston-Martin et al. (1988) front door of 2098 dwellings in Sweden. rcl~oaed magnetic fields of about 24 mG under clectric Measurements at lgl locations in lg residences in Los blankets. They estimated that electric blanket use Angeles had a geometric mean of 0.6 mG and a increases 'overall exposure'to electric fields by less geometric standard deviation of 2.g (Bowman et al., than 50%, and to magnetic fields by less than 100~. 1988). The ninety-filth percentlie was 3.4 raG. Mcasurements of magnetic field exposure due to According to Stuchly (1986), levels of 80 to 120 mG appliance usc averaged 21 mG with a standard have been reported in Germany in rooms whcre deviation of 76 mG for measurements taken at the belt thermal storage electric heating is in use. (Silva et al., 1988). The high magnetic fields near RESIDENTIAL MAGNETIC FIELD MEASUREMENTS With Geometric Standard Deviations 10.000 t tit E 1.000 - T T --- , * c ~ 0.100 .[ 0.010 relure 2-26. Engineering and E;cposure Assessment 2-37 appliances indicate that their contribution must be Bracken (1985) reported the use of personal elcctric explicitly considered in any careful assessment of total field exposure monitors to measure cumulative exposure. exposures of utility employees. Highest exposures, 1.7 kV/m hr, occurred for linemen. Exposures generally 2.7.2 Occupational Exposure roac with the voltage of the ¢q. uipmcnt, and dally maxima ranged from 5.1 to 7.6 kV/m hr. Electric field measurements in fourteen commercial and retail locations in rural Wisconsin yielded a mean of Personal exposure data have bccn collected for work, 4.8 VIm and a standard deviation of 4.3 V/m, about non-work, and sleep periods in a study of 36 one-third the values found in residences. The magnetic Canadians-20 utility workers and 16 officc workers fields averaged 1.1 raG; the standard deviation was 2 (Deadman, 1988). Thc timc-wcightcd arcrage of one mG (IITRI, 1984). wcck's data yielded a gcomctric mean-electric-field exposure of 3.1 raG. It was 1.9 mG for the office Bowman ctal. (1988) sampled 114 work sites. workers. Both groups had a lcvcl of 1.5 mG whilc "Electrical worker" environments had a geometric sleeping. Whilc at work, the utility workers' cxposures mean fields of 4.6 V/m and 5 raG. Sccretarics had averaged 48.3 V/m and 16.6 raG. Office workcrs wcre values of 2 to 5 V/m, 3.1 mG if thcy used a Video cxposed to a gcomctric mean level of 4.9 V/m and 1.6 Display Terminal (VDT), and 1. imG if they did not. raG. For powerline workers, the overhead !inc cnvifonmcnt yielded geometric means of 160 V/m and 42 raG. A 2.7.3 Exposure SUITIITlary value of 57 mG was determined for underground lines. Other findings included 298 V/m and 39 mG at a Bntckcn (1988) has summarized the characteristics of distribution substation. Radio and tclcvision repair EMF exposure as follows: shops yielded 45 V/m, while AC welding produced 41 raG. . Internal sources of electric field seem to predominate in residences. Electric feld3 in Swedish workers in a 400-kV substation spont most of residences are highly variable, source dependent, their time in fields below 5 kV/m, with brief exposures and not easily predicated. above 15 kV/m (Knave ctal., 1979). A study of Canadian linemen and substation workers used . Residential electric fitld~ are in the range of I to measuremenU and task activity patterns to estimate I00 V/m, with area fields typically in the range of daily exposures (Stupps and Janischcwskyj, 1979). 5 to 20 V/nt Estimates ranged from $0 to 60 0cV/m)-hr for 500-kV linemen to 13 0cV/m)-hr for 115 to 2.30-kV lineman, · Electric fields in public areas and occupational and 12 0cV/m)-hr for substation workers. Using a settings are comparable with residential exposures personal exposure meter, which was worn on the ann, except when well-defined high-voltage sources art Malc ct al. (1984) measured exposures of electrical present. workers in the United Kingdom. The dcvicc had a threshold of 6.6 0cV/m)-hr. Adnong 166 transmission · Transmission lines and other power trattrmission workers (cquipmcnt rated 132, 275, or 400 kV) and facilities represent sources of high electric field 121 distribution workers (equipment rated at 132 kV or levels in oladoor areas, t~la levels are .$frongly bolow), only 26 transmission workcrs and two dependent on shielding by objects. distribution workers had cumulativc, lO-day exposures above the threshold. Among the 26 transmission workcrs with measurable exposures, the median daily * Residential magntticfield~ are strongly influenced exposure value was 1.5 (kV/m)-hr por day, and thc by external sources such as ground currents, maximum valuc was 24.3 0cV/m)-hr per day. transmission lines and nearby distribution lines. Appliances represent a source o/highly variable Fanncrs whose land is crossed by high-voltage fields, but predictive modeling of field levels may transmission lines represent another cxposed population be possible. with higher-than-normal peak exposures. Using a combination of measured and modeled concentration . Domestic magnefic field levels are typically in the data, it was dctcrmined that the annual exposure of this range 0/0.5 to 1.0 mG but average levels can be group might range from 10 to 120 kV/m hr, with much higher. diffcrenccs being attributed to the voltage of the lincs (EPRI, 1985). Peak cxposures ranged above 8 kV/m. Engineering and Exposure Assessment 2-43 Werthcimer N, Leeper E. Adult cancer related to electrical wires near the home. International Journal of Epidtmiology 1982;11:345-55. WHO. Environmental Health Criteria 35: Extremely Low FrequenCy (ELF) Fields. Geneva; World Health Organization, 1984. WHO. Environmental Health Criteria 69: Magnetic Fields. Geneva; Worl Health Organization, 1987. Microwave News. ELF Gaussmettrs and Dosimeters, Microwave News, January/February 1990, p. 8-9. Engineering and Exposure Assessment 2-39 References ANSI/IEEE Std 6,~-1987 (Revision of ANSI/IEEE Std 64~1979). IEEE Standard Procextures for Measurement of Power Frequency Electric and Magnetic Fields from AC Transmission Lines, Institute of Electrical and Electronic Engineers, New York, New York, 1987. Barnes F. Exposure data needs and measurement procedures for the University of Colorado epidemiologie study. Workshop on measurement of non-uniform and fluctuating 60-Hz electrical and magnetic fields; U.S. Environmental Protection Agency, Las Vegas, Nevada; 1985 February 12-13. Bowman JD, Garabrant DH, Sobel E, Petera, IM. Exposure to extremely low frequency (ELF) electromagnetic fields in occupations with elevated leukemia rates. Applied Industrial Hygiene 1988; 3:189-94. Bracken TD. Analysis of BPA occupational electrical field exposure data. Final report for Bonneville Power Administration, Vancouver, Washington; 1985. Bracken TD. The 60-Hz electric and magnetic fields in the environment. EPRI Seminar on power-Frequency Electfie and Magnetic Field Exposure Assessment; Colorado Springs, Colorado; 1988 October 12-14. Brown, DR. TRANSPAC: A solhare package for the design and analysis of electfie power transmission lines, Program User's Manual, David R. Brown, INC., Austin, Texas, 1989. Caola RJ, Deno DW, Dymek VSW. Measurement of electric and magnetic fields in and around homes near a 500 kV trammission line. IEEE Transactions on Power Apparatus and Systems 1983. PA$-102:333847. Calle EE, Savitz DA. Leukemia in occupational groups with presumed exposure to electrical and magnetic fields. Hew England lournal of Medicine 1985; 313: 1476-77. CEQ. Risk Analysis: A Guide to Principles and Methods for Analyzing Health and Environmental Risks. Council on Environmental Quality, Executive Office of the President, Washington, D.C.; 1989. NTIS Order Number PB 89-137772. Coleman M, Bell J, Skeet R. Leukemia incidence in electrical workers. Lancet 1983;i:982. Coleman Met al. Leukemia and electromagnetic fields: a case-control study. IEE International Conference on Electfie and Magnetic Field in Medicine and Biology; London, 1985 December 4-5. IEE Publication No. 257. Deadman JE, Camus M, Armstrong BG, Herous P, Cyr D, Planto M, Theriault G. Occupational and residential 60- Hz electromagnetic fields and high frequency electfie transients: exposure assessment using a new dosimeter. American Industrial Hygiene Association Iournal 1988;3: 189-94. Dietrich FM. AC electfie and magnetic field measurements. EPRI Seminar on power-Frcqueney Electric and Magnetic Field Exposure Assessment; Colorado Springs, Colorado; 1988 October 12-14. Electric Field Measurement Co. Model 113 Power Frcqueney Field Meter Instruction Manual, Electric Field Measurement Co., W. Stockbridge, Massachusetts, March 1985. Electfie Field Measurcment Co. Model 116PLUS 50/60 Digital Electric and Magnetic Field, Product Brochure, Electric Field Measurement Co., W. Stockbridge, Massachusetts, September 8, 1989. Engineering and Exposure Assessment 2-41 Male JC, Norris WT, Watts MW. Exposure of people to power-frequency citeeric and magnetic fields. Twenty- third Hartford Life S~icnces Symposium; Richland, Washington; 1984 October 2-4. McDowall ME. Lcukcmla in electrical workers in N~v Zealand. Lancet 1983;i:811-12. McDowall ME. Mortality of persons resident in the vicinity of electricity transmission facilities. British Journal of Cancer 1986;53:271-79. Mllham S Jr. Moz~tlity from lcuk~mla in workers exposed to electrical and magnet. ic fields. New England Journal of Medicine 1982;307:249. Milham 8 Jr. Mortality in workers cxposexl to cleetremagnetic fields. Environmental Health Perspectives 1985a;62:297-300. Millram S Jr. Silent keys: lcukemia mortality in amateur radio operaton. Lancet 1985b;i:812. Misaivlan, M. AC Electric and Magnetic Field Me~r Fundamentals, EPRI Seminar on Pov~r-Frequcncy Electric and Magnetic Field Exposure Assessment; Colorado Springs, Colorado; 1988 October 12-14. Monitor Industries. Model 42A Exploratory AC Milligaussme~r, Instruction Manual, Monitor Industries, Boulder, Colorado, COndated). Myers A, Cartwright RA, Bonnell JA, Male JC, Cartwright SC. Overhead power lines and childhood cancer. International Conference on Electric and Magnetic Fields in Mc~licine and Biology; London; 1985 Dec. c~b~r. IEE Conference Publication No. 257: 126-30. Nordstrom S, Birkc E, Gustavsson L. Reproductive hazards among workers at high voRagc substations. Bioeltctromagnttics 1983; 4:91-101. Offic~ of S4:ienc~ and Technology Policy. Chemical careinogcns: s review of the science and its associated principles; 1985 February. Federal Register 1985 March 14; pp. 10371-442. Olin R et al. Mortality experience of electrical engineers. British Journal of Industrial Medicine 1985; 42:211-12. Pe~rec NE et al. lxukcrnia in clc~etrical workers in Nwv Zealand. Lancet 1985;i:811. Positron Industries, Inc. Electsmagnetic Dosimet~r Specification Sheet, Positron Industries, Inc., Montreal, Canada, COndate, d). Preston-Martin S, Peten 3M, Yu MC, et al. Myelogenous leukemia and electric blanket use. B~oelectromagnetics 1988;9:207-13. Savitz DA. Case-control study of childhood cancer and residential expesuro to electTic and magnetic fields. Final report to the New York State Department of Health Power Lines Project; 1987. Savitz DA et al. Case-control study of childhood cancer and exposure to 60-Hz magnetic fields. American Journal of F. pidemiology 1988;128:21-38. Savitz, DA. Association of Childhood Cancer with Residential Traffic Density, Scandinavian Journal of Work, Environment, and Health,1989: 15:360-363. Severson RK et al. Acute nonlymphocytic leukcmia and residential exposure to power frequency magnetic fields. American Journal of Epidemiology 1988;128: 10-20. Engineering and Exposure Assessment 2--41 Male JC, Norris WT, Watts MW. Exposure of people to power-frequency electric and magnetic fields. Twenty- third Hitford Life Sciences Symposium; Richland, Washington; 1984 October 2-4. McDowall ME. Leukemia in electrical workers in New Zealand. Lancet 1983;i:811-12. McDowall ME. Mortality of persons midcat in the vicinity of electricity transmission facilities. British Journal of Cancer 1986;53:271-79. Milham S Jr. Mortality from lcukcmia in workers exposed to electrical and magnetic fields. New England Journal of Medicine 1982;307:249. Milham S Jr. Mortality in workers exposed to elcctromagnctic fields. Environmental Health Perspectives 1985a;62:297-300. Milham S Jr. Silent keys: leukemia mortality in amateur radio operators. Lancet 1985b;i:812. Mi~an, M. AC Electric and Magnetic Field Meter Fundamentals, EPRI Seminar on Power-Frequency Electric and Magnetic Field Exposure Assessment; Colorado Springs, Colorado; 1988 October 12-14. Monitor Industries. Model 42A Exploratory AC Milligaussmctcr, Instruction Manual, Monitor Industries, Boulder, Colorado, (Undatcd). Myers A, Cartwright RA, Bonneli IA, Male JC, Can'wright SC. 0verhe~d power lin~s and childhood cancer. International Conference on Electric and Magnetic Fields in Medicine and Biology; London; 1985 December. IEE Conference Publication No. 257: 126-30. Nordstrom S, Birkc E, Gustavsson L. Reproductive hazards among workers at high voltage substations. Bioeltctromagnetic~ 1983; 4:91-101. Office of Science and Technology Policy. Chemical carcinogens: a review of the science and its associated principles; 1985 February. FederalRegister 1985 March 14; pp. 10371-442. Olin R et al. Mortality experience of electrical engineers. British Journal of Industrial Medicine 1985; 42:211-12. Pcarcc NE et al. Leukcmia in electrical workers in New Zealand. Lancet 1985;i:811. Positron Industries, Inc. Electromagnetic Dosimctcr Specification Sheet, Positron Industries, Inc., Montreal, Canada, (LIndate, d). Preston-Martin S, Peters JM, Yu MC, et al. Myclogenous lcukcmia and electric blanket use. Bioelectromagnetics 1988;9:207-13. Savitz DA. Case-control study of childhood cancer and residential exposure to electric and magnetic fields. Final report to the New York State Department of Health Puwcr Lines Project; 1987. Savitz DA et al. Case-control study of childhood cancer and exposure to 60-Hz magnetic fields. American Journal of Epidtmiology 1988;128:21-38. Savitz, DA. Association of Childhood Cancer with Residential Traffic Density, Scandinavian Journal of Work, Environment, and Health,1989: 15:360-363. Severson RK ctal. Acute nonlymphocytic Icukemia and residential exposure to power frequency magnetic fields. American Journal of Epidemiology 1988;128: 10-20. Engineering and E~posure Assessment 2-39 References ANSI/IEEE 5td 644-1987 (Revision of ANSI/IEEE Std 644-1979). IEEE Standard Procedures for Measurement of Power Frequency Electric and Magnetic Fields from AC Transmission Lines, Institute of Electrical and Electronic Engineers, New York, New York, 1987. Barnes F. Exposure data needs and measurement procedures for the University of Colorado epidemiologic study. Workshop on measurement of non-uniform and fluctuating 60-Hz electrical and magnetic fields; U.S. Environmental Protection Agency, l.,aa Vegas, Nevada; 1985 February 12-13. Bowutah JD, Garabrant DH, Sobel E, Peters, JM. Exposure to exlxcmcly low frequency (ELF) electromagnetic fields in occupations with elevated leukcmia rates. Applied Industrial Hygiene 1988; 3:189-94. Bracken TD. Analysis of BPA occupational electrical field exposure data. Final report for Bonneville Power Administration, Vancouver, Washingion; 1985. Bracken TD. The 60-Hz electric and magnetic fields in the environment. EPRI Seminar on Power-Frequency Electric and Magnetic Field Exposure Assessment; Colorado Springs, Colorado; 1988 October 12-14. Brown, DR. TRANSPAC: A software package for the design and analysis of electric power transmission lines, Program Uscr's Manual, David R. Brown, INC., Austin, Texas, 1989. CaoLa RJ, Deno DW, Dymek VSW. Measurement of electric and magnetic fields in and around homes near a 500 kV transmission line. IEEE Transactions on Power Apparatus and Systems 1983. PA5-102:3338-47. Calle EE, Savitz DA. Leukemia in occupational groups with px~esumed exposure to electrical and magnetic fields. New England Journal of Medicine 1985; 313: 1476-77. CEQ. Risk Analysis: A Guide to Principles and Methods for Analyzing Health and Environmental Risks. Council on Environmental QuaLity, Executive Office of the President, Washington, D.C.; 1989. NTIS Order Number PB 89-137772. Coleman M, Bell J, Skcet R. Lcukemia incidence in electrical workers. Lancet 1983;i:982. Coleman Met al. Leukcmia and electromagnetic fields: a case-control study. FEE International Conference on Electric and Magnetic Field in Medicine and Biology; London. 1985 December 4-5. FEE PubLication No. 257. Deadman JE, Camus M, Armstrong BG, Herous P, Cyr D, Planto M, Theriault G. Occupational and residential 643- Hz electromagnetic fields and high frequency electric transients: exposure assessment using a new dosimeter. American Industrial Hygiene Association Journal 1988;3:189-94. Dietrich FM. AC electric and magnetic field measurements. EPRI Seminar on Power-Frequency Electric and Magnetic Field Exposurc Assessment; Colorado Springs, Colorado; 1988 October 12-14. Electric Field Measurement Co. Model 113 Power Frequency Field Me. mr Instruction Manual, Electric Field Measurement Co., W. Stockbridge, Massachusetts, Maroh 1985. Electric Field Measurement Co. Model I16PLUS 50/60 Digital Electric and Magnetic Field, Product Brochure, Electric Field Measurement Co., W. Stockbridge, Massachusetts, September 8, 1989. Engineering and Exposure Assessment 2-43 Werehelmet N, Leeper E. Adult cancer related to electrical wires near the home. International Journal of Epidemiology 1982; 11:345-55. WHO. Environmental Health Criteria 35: Extremely Low FrequenCy (ELF) Fields. Geneva; World Health Organization, 1984. WHO. Environmental Health Criteria 69: Magnetic Fields. Geneva; Worl Health Organization, 1987. Microwave News. ELF Gaussmetcrs and Dosimetcrs, Microwave News, January/February 1990, p. 8-9. Engineering and Exposure Assessment 2-37 appliances indicate that theix contribution must be Bracken (1985) reported the use of personal electric explicitly considered in any careful assessment of total field exposure monitors to measure cumulative exposures of utility employees. Highest exposures, 1.7 exposure. kV/m hr, occurred for linemen. Exposures generaBy 2.7.2 Occupational Exposure rose with the voltage of the equipment, and daily maxirna ranged from 5.1 to 7.6 kV/m hr. Electric field meuuremcnts in fourteen commercial and retail locations in rural Wisconsin yielded a mean of Personal exposure data have bccn collected for work, 4.8 VIm and a standard deviation of 4.3 V/m, about non-work, and slccp poriods in a study of 36 one-third the values found in residences. The magnetic Canadians-20 utility workcrs and 16 office workers fields averaged 1.1 raG; the standard deviation was 2 (Deadman, 1988). The time-weighted average of one mG (IITRI, 1984). week's data yielded a geometric mean-electric-field exposure of 3.1 mG. It was 1.9 mG for the office Bowman et al. (1988) sampled 114 work sites. workers. Both groups had a level of 1.5 mG while 'Electrical worker* environments had a geometric sleeping. While at work, the utility workers' exposures mean fields of 4.6 V/m and 5 raG. Secretaries had averaged 48.3 V/m and 16.6 raG. Office worker~ were values of 2 to 5 V/m, 3.1 mG if they used a Video exposed to a geometric mean level of 4.9 V/m and 1.6 Display Terminal (VDT), and 1.1 mG if they did not. raG. For powerline workcrs, the overhead line environment yielded geometric means of 160 V/m and 42 raG. A 2.7.3 Exposure Summan/ value of 57 mG was determined for underground lines. Other findings included 298 V/m and 39 mG at a Bracken (1988) has summarized the characteristics of distribution substation. Radio and television repair EMF exposure as follows: shops yielded 45 V/m, while AC welding produced 41 raG. · Internal sources of elecgr~c field seelfi to predominate in residences. Electric fields in Swedish workers in a 400-kV substation spent most of residences are highly variable, source dependent, their time in fields below 5 kV/m, with brief exposures and not easily predicated. above 15 kV/m (Knave et al., 1979). A study of Canadian linemen and substation workers used · Residential electric felda are in the range of I to measurements and task activity patterns to estimate IO0 V/m, with area fields typically in the range of daily exposures (Stopps and Janischcwskyj, 1979). 5 to 20 V/rn. Estimate~ ranged from 50 to 60 OcV/m)-hr for 500-kV linemen to 13 OcV/m)-hr for 115 to 230-kV lineman, · Electric fields in public areas and occupational and 12 (kV/m)-hr for substation workers. Using a settings are comparable with residential exposures personal exposure meter, which was worn on the arm, except when well-defined high-voltage sources are Male et al. (1984) measured exposures of electrical present. workers in the United Kingdom. The device had a threshold of 6.6 0cV/m)-hr. Among 166 transmission · Transmission lines and other power transmission workers (equipment rated 132, 275, or 400 kV) and facilities represent sources of high electric feld 121 distribution workers (equipment rated at 132 kV or levels in outdoor areas, but levels are strongly below), only 26 transmission workers and two dependent on shielding by objects. distribution workers had cumulative, lO-day exposures above the threshold. Among the 26 transmission worken with measurable exposures, the median daily * Residential magneticfields art strongly influenced exposure value was 1.5 0cV/m)-hr per day, and the by external sources such as ground currents, maximum value was 24.3 OcV/m)-hr per day. transmission lines and nearby distribution lines. Appliances represent a source of highly variable Farmers whose land is crossed by high-voltage fields, but prtdictive modeling of reid levels may transmission lines represent another exposed population be possible. with higher-than-normal peak exposures. Using a combination of measured and modeled concentration · Domestic magnetic field levels are typically in the data, it was determined that the annual exposure of this range of O. 5 to 1.0 mG but average levels can bt group might range from 10 to 120 kV/m hr, with much higher. differences being allributccl to the voltage of the lines (EPRI, 1985). Pcak exposures ranged above 8 kV/m. 3-1 3.0 Epidemiology of Health Effects and Exposure to EMF 3.1 Introduction ~,,~gory of descriptive ¢pidcmiologic studies. In correlational studies, measures that represent F. pidemiology is the study of the distribution and characteristics of entire populations are used to determinants of diseases and injuries in human describe a particular disease in rehtion to some factor populations. That is, epidcmiology is concerned with of interest, such as age, time, utilization of health the frequencies and types of illnesses and injuries in services, or consumption of a food, medicalion, or groups of people and with the factors that influence other product. Instead of considering whole their distribution. This implies that disease is not populations, as in correhtional studies, case reports or randomly distributed throughout a population, but case series describe thc experience of a singlc paticnt rather that subgroups differ in the frequency of or group of patients with a similar diagnosis. Thc third dlffcrent diseases. Knowledge of disease distributions major type of descriptive study is the cross-sectional or can be used to investigate causal factors and thus to lay prevalence survey, in which exposure and disease the groundwork for programs of prevention and control status are assessed simultaneously among individuals in (Mausnerand Kramer, 1985). a well defined population (Hennekens and Buring, 1987). Epidemiolog~ts have organized the complex processes that lead to disease in various ways. One useful way to Once a disease has been identified and categorized with view thc causes of disease is in terms of the agent, the respect to person, phcc, and time, ana/yac or etiologic environment, and the host (Friedman, 1980). When a tpidtmiologic studies arc often employed to test factor must be present for a disease to occur, it is specific hypofficses, estimate chronic health effects, and called a necessary agent of that disease. An agent may to suggest potential means of disease prevention be a necessary but not sufficient cause of disease (K. leinbaum et al., 1982). The major types of analytic because suitable conditions within the host and the or etinlogic cpidcmiologic studies include case-control environmental must be present for disease to develop. and cohort or follow-up studies. In a case-control Host factors are usually intrinsic, whereas factors in the study, subjects are selected on the basis of whether they environment are extrinsic. Host factors affect do {cases) or do not (controls) have a particular susceptibility to dlscaso; environmental factors disoasc. The proportions of subjects wiffiin each group influence exposure and may indirectly influence having histories of various exposures or other susceptibility as well. The interactions of these two characteristics of interest arc then compared. Case- sets of factors determine whether or not disease control studies provide a solution to the problems develops (Mausner and Kramer, 1985). inherent in the study of diseases with extended latehey periods. Also, case-control studies permit the 3.1.1 Types of Epidemiologic Studies evaluation of a wide range of potential cliologic Epidemiologic study designs are more fully discussed exposures that may bc related to a specific disease and in Appendix B (Fundamentals of Epidemiology, Section of the interrehtionships among these factors. Case- I), and are only briefly described below. control studha are particularly useful in the examination of rare diseases (Hcnnekcns and Buring, In general, epidcmiologic studies (which undertake no 1987). manipuhtion of the study factor) may be categorized as either descriptive or analytic (i.e., etiologic). The second major type of analytic cpidcmiologic study Descriptive tpidtmiologic studies arc usually conducted is the cohort or follow. up study. In ~is type of study, a when little is known regarding the occurrence, the group {or several groups) of individuals is defined on natoral history, or the risk factors for a particuhr the basis of the prcscncc or absence of exposure to a disease (Klcinbaum et al., 1982). The objectives of suspected risk factor for a disease. There are two types such studies include identification of the paRcans of of cohort or follow-up studies: prospective {or disease occurrence in relation to variables such as concurrent) and retrospective (or historical). These person, place, or time and the generation of more two types of cohort studies differ in terms of when specific hypotheses regarding ctiologies. Descriptive exposure and disease occur in relation to the onsct of studies provldc essential data to public health the study. At the time exposure status is defined, all administrators who use this information to plan potential subjects must be free from the disease being programs for prevention and control of disease. studied, and all discase-free eligible participants are Epidemiologists also rely on descriptive studies to lay followed for a period of time to assess the occurrence the groundwork for hypothesis-testing cliologic studies of that disease. As a result of this design, cohort (Hcnnckcns and Buring, .1987). studies can provide information on the full range of health effects of a single exposure. When feasible, Correlational studies, case reports or case series, and they are the prefcrrcd method of study since the results cross-sectional studies are included under the gcncral 3-2 Health Effects of E~cposure to Powerline Frequency Electric and Magnetic Fields are less subject to bias than other study types. (or 100, if that multiplying factor has been used), and However, since cohort studies are generally very time- no excess (or decreased) risk is present. Values less cormuming and expensive, they are oPten conducted than 1.0 (or 100) represent 'decreased risk," and only after a hypothesized rehtionship has been values greater than 1.0 (or 100)indicate "increased explored and evaluated in a case-control study risk" for the study group compared to the standard (l-lennekens and Buring, 1987). Therefore, these two population (see Appendix B). types of studies are complementary. A similar ratio can be calcuhted using morbidity or 3.1.2 I=stimatss of Risk in I~pidsmiologic incidence data, thereby producing a standardized StUcliss incidence ratio (SIR) which is interpreted in the same manner. An SIR that is equal to 1.0 (or 100) means Rates of disetse and ratios of rates derived from that the observed and expected numbers of cases are comparisons of exposed and unexposed grouln in equal, indicating no increased ri~k for the development epideminlogic studies are used to provide estimates of of the particular disease. An SIR of 1.5 (or 150) means risk. In a follow-up (cohort) study, an actual disease that there are 50% more eases in the study group than rate can be tabulated separately for both the exposed expected. and nonexposed groups. These rates can then be compared in several ways to develop a quantified An entirely different ratio which appears frequently is expression of risk in relative or absolute terms. the proportional mortality ratio (PMR). Although PMR and SMR studies appear superficially similar, they are The most common measure of risk in a cohort follow- quite different and are derived from different types of up study is the relative risk (RR). It is defined as the data. Proportional mot*,ality expresses the proportion ratio of the incidence of disease in the exposed group of all deaths that are due to a partieuhr cause. For divided by the incidence of disease in the nonexposed example, for deceased individuah who were employed group. The relative risk is an indication of the degree in a particular industry, 20 percent of the deaths may of risk for disease (either increased or decreased) have been due to cancer, whereas heart disease may among the exposed relative to the nonexposed. This have accounted for 35 percent of the deaths. Most value, therefore, provides an estimate of the importance PMR studies are done when the investigator only has of the exposure under study. A relative risk of one information regarding the people who have died but (1.0) indieate~ no association between exposure and does not have data on the total number of persons (or disease. person-years) at risk. Under these circumstances, the The standardized mortality ratio (SMR) is frequently only items that can be compared are the proportions of used to estimate risk in epidemiologie studies when the all deaths that were due to cancer in one occupation only information available is the number of deaths that with the proportion due to cancer in another group. have been observed among the study population. It Since data for the total popuhfion at risk are not provides a means of comparing the mortality available, a mortality rate cannot be determined. experiences of populations that have different In contrast to cohort studies, which determine disease distributions of important variables such as age, sex, or rates for exposed and nonexposed groups, case-control race. To assess whether the 'observed' number of studies start with diseased and non-diseased groups and deaths is excessive, rates from a standard popuhtion then determine exposure histories for individuals in the are used to calculate the number of 'expected' deaths two groups. Since this approach does not permit the for the study group had they succumbed to the disease determination of actual disease rates, a relative risk at the same rate as the standard population. cannot be determined. One can, however, compare The study group is fwst divided into a number of 'exposure ratios' between diseased and non-disused "strata,' which usually include 5- or 10-year age groups and, under certain conditions, these 'exposure ratios' can be used to estimate the rehtive risk by groups within each sex and race or ethnieity category. The expected number of deaths for each stratum is calculating the relative odds of exposure, or odds ratio calculated by multiplying the number of persons (or {see Appendix B). person-years of observation) in that stratum by the The odds ratio (OR) provides a reasonable estimate of cerreaponding stratum-specific mortality rate in the the relative risk if the disease in question is relatively standard population. The total 'expected' number of rare (e.g., specific cance~), if the exposure is deaths is then obtained by summing the expected relatively common, and if there are no serious biases in humben calculated for each stratum. Once this value the design or conduct of the study. The odds ratio is has been obtained, the SMR is calculated by dividing interpreted exactly the same as the relative risk. If the observed number of deaths by the expected number equal to one (1.0), the odds ratio suggests no (occasionally, the resulting ratio is multiplied by 100 to association between the exposure and the disease. eliminate the decimal places). When the observed and expected numbers are equal, the SMR is equal to 1.0 Epidemiology of Health Effects and Exposure to EMF 3-3 3.1.3 Assessment of Validity in studies, bias in the selection of study subjects can occur Epidemiologic Studies when the procedure used for the selection of cases is different from that used for selection of controls, as In any scientific study, careful attention must be given when the procedure used to identify disease status to the validity and reliability of the data. All data varies with exposure status. In cohort studies, selection collection methods involve some degree of inaccuracy bias occur~ when the exposed group is selected from a and variability. Concern for these threats to data popuhtion with a different overall probability of quality is more pronounced in observational studies of disease than that from which the uncxposcd group is human populations, as is the case with most selected. Systematic differences in the way cases and cpidemiologic studies. controls choose to participate in either a case-control or In asscasing the results obtained from cpidemiologic a cohort study also represent important sources of bias studies of human health effects thought to bc associatecl in cpidcrniologic studies. with cxposuro to powerline-frequency electric and/or Information bias refers to a distortion in the estimated magnetic fields (BMF) (or any other suggested ctiologic effect that results from a random or systematic flaw in factor), two aspects of study validity must bc the measurement or classification of either the exposure considered: internal validity, as determined by status or the disease outcome. Measuremenl error can accuracy and reliability of measurements, and external be either random (duc to chance variations) or validity or the gcncralizability of the results. nonrandom (duc to systematic bias). Random errors INTBRNAL VALIDITY may be introduced into a study as a result of the variability inherent in most physical measurements. The concelt of internal validity addresses the question, This form of error may bc reduced to a certain extent 'Is this study capable of providing an unbiased and by employing multiple measurements and using the quantitatively accurate cstixnate of risk?' Threats to the resultant mean value. Nonrandom errors may occur as internal validity of a study include: biases in the a result of the improper or inconsistent calibration of selection of study groups, erron and biases in field measuring instruments or laboratory equipment. tncasuremcnts and classification of disease or exposure, Other sources of nonrandom measurement errors confounding of the disease-exposure association by include a defective questionnaire or an interview other risk factors, and the possibility that apparent schedule that fails to elicit the intended responses or an associations may bc duc to chance. Each of these inaccurate diagnostic procedure that either facton must be addressed in evaluating the internal overestimates or underestimates disease status validity of an cpidcmiologic study and weighing the (Klcinbaum ctal., 1982). credibility of any conclusions based thereon. Misclass~fwafion b/as is a form of information bias Bias in Epidcmiologic Studies. B/as may bc defined as which typically affects the analysis of data in which a systematic error introduced into an cpidcmiologic both the disease and the cxposuro variables arc study that results in an inaccurate estimate of the dichotomous. As with measurement errors, association between exposure and risk of disease misclassi~cation bias may also bc either random or (Hcnnekcns and Buring, 1987). Since cpidemiologic nonrandom. studies involve free-living human beings, even the most rigorously designed investigation will have the potential Random or nond/fferen//a/misclassidfication introduces for one or more types of bias and/or confounding. A iraprecision but may also bias the association towards number of sources of bias may distort the association the null in a study seeking to evaluate an exposure- between exposure and disease observed in a particular disease association. Thus, random errors in the study (Sackerr, 1979). The major sources of bias estimation of exposure will generally lead to an multirag from the employed methods of study design undctestimation of risk, and can introduce a spurious and analysis may be conveniently grouped under thc curvilincarity into the estimated dose-response function. headings of selection bias, information bias, and For example, in the Radiation Effects Research confounding bias CKlcinbaum ctal., 1982). Foundation (RERF) studies, random errors in the dose estimates resulted in a 10-15 percent underestimation of Selection bias refers to an error in the estimate of an the relative risk (SchuU, 1991). This type of error may effect that is duc to systematic differences in the be partially overcome by sufficiently increasing the characteristics of those who are selected for study and sample size of the study. those who are not (Hcnnckcns and Burrag, 1987). For instance, a telephone interview survey will exclude By contrast, nonrandom errors in the estimation of households that have no telephone. This exclusion may exposure can distort or bias the relative risk either underestimate the proportions in the population with upward or downward, and, conscqucnUy, represcnt a certain characteristics, such as age, cthnicity, and serious threat to the validity of a study. Differential socioeconomic status; thus, the truc proportions remain misclassification occurs when the errors in classifying unknown (Corey and Freeman, 1990). In case-control individuals occur differentially among study groups, 34 Health Effects of Exposure to Powerline Frequency Electric and Magnetic Fields thus distorting the comparison of ram and the resulting and not duc to chance, is accomplished through the use rate ratios. Differential misclassi~cation of the of appropriate statistical techniques such as significance exposure or disme status may occur in epidcmiologic testing with p-values and confidence intervats (CI). studies, especially when surrogate measures or These techniques are discussed in the tutorial on subjective estimates are used to determine exposure or epidemiologic methods in Appendix B. disease status of the study subjects. This form of bias EXTERNAL VALIDITY can occur when knowledge of the disease status of the eases and controls (in a ease-control study) influences Assessing the external validity of an epidemiologic the exposure classification, or when knowledge of the study involves an evaluation of the genernliT~bility of exposure status of the study subjects (in a cohort study) the association. It addresses two questions: (1) "Is the influences the disease classification of the subjects. association of disme with exposure consistent with To avoid or substantially reduce this type of bias, well causality?" (2) "Is the disease-exposure association observed in a specific study likely to hold for other designed studies generally employ some form of similarly exposed populations?" When bias, blinding to insure that the investigator is unaware o[ the confounding, and chance have all been determined to disease status of the individual when the exposure be unlikely explanations of a particuhr f'mding in an status is being classified and vice vcrsa. Also, when cpidcmiologic study, it is then necessary to decide persons who become lost to follow-up differ, with whether the observed association of disease with respect to both exposure and outcome, from those who exposure can be considcrai to be causal (Klcinbaum ct remain in the study, the observed association can be al., 1982; Hcnnckens and Buring, 1987). A number of differentially biased. For instance, a statistically guidelines to assist with the judgment of the causal significant increase in lung cancer mortality was nature of an enviro nmattal association were initially observed in a cohort of workers producing urea- and proposed by Hill (1965). These guidelines include the mclaminc-formaldchydc ruins, but the increase could strength of the association, consistency of the data, not bo specifically attributed to formaldehyde exposure spcci~city of the association, temporality of exposure because of incomplete work histories for the workers and disease, dosc-ruponsc gradient, biological lost to follow-up (Bcrtng~i ctal. 1986). plausibility, coherence of the evidence, and the effect Confomsding bias in cpidemiologic studies is a of intervention. distortion in the apparent association between a disme StrcnRth of the Association.. For cpidemiologic and an exposure by a third factor that is causally cvidcnco, the strength of the association, as measured related to the disease under study and is also associated by the maguitudc of the risk ratio, is useful in with the exposure under study but is not a consequence determining whether the exposure affects the risk of of the exposure. Therefore, the dlseasc may be either developing the disme. Specifically, the stronger the partially or totally attributable to the third factor and association (or the greater the risk ratio), the more not to the exposure under investigation. The pruence likely it is that the association is causal. In general, of confounding in a disease-exposure association can be weaker associations do not lend as much support to a assessed by controlling for the effect of any of the causal interpretation (Klcinbaum ctal., 1982). Monson cxu'ancous factors which may be associated with both (1990) has proposed some guidelines for assessing the disease and exposure. Unlike selection bias and magnitude o[ relative risks. These guidelines suggest information bias, which arc primarily introduced by the that relative risks of less than 2.0 are more likely to be investigator or study participants, confounding is a due to bias or confounding, whereas relative risks function of complex interrelationships hawcon various greater than 5.0 (depending of course on sample size) cxposuru and disease (Hcnnckens and Buring, 1987). are more Ilkely to be reflective of a true increase in Chance in Epidemioloaic Studies. Before attempting to risk. assess the causal nature of an association, it is Consistency of the Data. Since cpideminiogy is by its necessary also to determine whether the difference in !tature observational, it is never possible to achieve the rates is likely to have bccn duc to chance. Because of degree of control possible in experimental studies. the random variations that occur in population samples, Evidence to support a judgment of a cause-and-effect the case and control groups (in a case-control study) relationship cannot be persuasive unless the association are unlikely to have exactly the same proportion of is consisUmtly observed in a number of studies. cxposai and non-exposed persons, even when no Consistency among multiple studies, conducted by association exists between exposure and disease. different investigators, at various times, using Likewise, the exposed and non-exposed groups in a alternative methods, in a variety of geographic or cohort study are unlikely to have exactly the same cultural settings, and among different populations, incidence of (or mortality from) a particular disme provides strong evidence for the gcncralizability of a even though there is no cause-and-effect (or protective) cause-and-effect rolationship (Hcnnckcns and Buring, relationship. Determining how large a difference must 1987). Conversely, when multiple studies sccm to bc be to establish convincingly that it is likely to be real Epidemiology of Health Effects and Exposure to EMF 3-5 producing inconsistent or contradictory results, chance confident one can be in rejecting a causal interpretation associations, .confounding factors, and/or study bias are on the basis of this guideline (Kleinbaum ct al., 1982). more likely to be present, and the argument for a Coherence of the Evidence. Coherence of the cause-and-effect relationship is weakened. cpidcmiologic, biologic, and other evidence implies that Spcci~city of the Association. Spcci~city of thc a causal interpretation is not scriously in conflict with association is anothcr aspect to bc considcrcd in current knowlcdgc of thc natural history and biology of wcighing the external validity of a study. If thc thc discasc (Greenland, 1987). For example, morality association is limited to a specific cxposure and/or to a from lung canccr was initially higher in males than in particular discasc, there is a strong argument in favor fcmalcs bccausc smoking was more common in males. of causation. However, in chronic disease As more females adoptai the smoking habit, morality cpidcmiology, multiple-causation is gcncrally more rates for lung cancer among females havc tended to likely than singlc causation. Therefore, if specificivy of approach those secn among males (Hammond, 1966). thc association has been demonstrated, a causal The Effect of Intervcntinn. The strongest support for interpretation can bc made with greater confidcnce; if the causation hypothesis is provided by cxpcrimcntal specificiVy does not cxist, a causal interpretation must studies or intervention studies in human populations be made more cautiously {Greenland, 1987). (Hill, 1971). If cxposure to an agent is truly causal, TemporaliVy of Bxposure and Di;ease. The temporal then removal of thc agcnt should result in a decrease in relationship of an association is particularly relcvant the disease rate. Since the association of smoking with with diseases of slow development (Hill, 1971) or lung cancer was recognized, many physicians havc pwlonged latency periods, such u cancer. Does a stopped smoking. This change was followed somc 20 particular occupational or environmental exposure lead years later by a ticcrease in lung canccr mortality to an increase in some form of cancer? If the among physicians (Doll andPeto, 1976). association is to be considerai causal, thcn thc relevant An expanded discussion of epidemiologic study exposure must prccale the occurrcncc of canccr by a methods is found in Appendix B. sufficient interval of tLmc to account for thc discasc latency period. This also holds for diseases with short 3.1.4 EMF ~xposure Assessment in incubation or htency periods {e.g., symptoms of acute Epidemiologic Studies toxicity). If the interval from exposure to onset of diseasc is too long, the argument for causation is Thc assessment of EMF cxposure secrns to bc the factor of greatest unccrtainty in the epidemiologic weakened. studies of EMF and cancer. Many studies published to Dose-Response Gradient. If thc association date have been somcwhat flawed in methodology, in demonstrates a dose-response gradient, the causc-and- part due to the use of indirect, imprecise, and effect hypothesis is supported {Grecnland, 1987). For unverified surrogate mcasures of cxposure. Thc cxamplc, the causal association of smoking and lung possibility for misclassification ofthc exposure status of cancer was strongly supported by the cvidencc that thc an individual can bc random or systematic (i.c., mortaliVy rate from lung canccr increased linearly with nonrandom), thus resulting in cithcr dccreased the numbcr of cigarettes smoked dally {Doll and Hill, precision or overt distortion in the assessment of an 1950). association. Biological Plausibility. Thc causal nature of an In gencral, some form of exposure gradicnt or association is more readlly accepted if there is a diffcrential among study subjocts is nccessary for an biologically plausible hypothesis to support causality. effective environmcntal or occupational cpidemiologic However, biological plausibility is determined, to a study. If exposures were homogeneous, it is unlikcly certain extent, by the current state of knowledgc that an exposure-disease association could bc (Grecnland, 1987). Conscqucntly, such phusibility investigatcd. Thc choice of grouping subjects into cannot always be demanded of · hypothesis, since the dlchotomous or multi-stage exposure classifications current state of knowledge may simply be inadequate to often infiucnces the results of the statistical analyses explain the observations. An impressive number of (Flegal et al., 1986). associations of environmental exposures with cancer !~-XPOSURB METRIC was identified from epidcmiologic studies prior to knowledgc regarding the biological mechanisms. The appropriate direct incasure of exposure in Notablc examples include the association of cigarette cpidemiologic studies of EMF is still bcing ticbated, smoking and lung canccr, and association of specific and this is an important arca of ongoing research. The agcnts such as asbestos, bcnzenc, and ultraviolet abscnceofamechanism for a biologic cffectto explain radiation with human canccr. Convcrsely, the less that thc EMF-canccr association crcates unccrainty about is known regarding thc etiology of a diseasc, the less what, cxactly, should bc thc appropriatc "cxposure metric.' Candidates for the cxposure metric might 3-6 Health Effects of Exposure to Powerline Frequene/ Electric and Magnetic Fields include the magnelic field component, the clccuic field measurements arc used to estimate the unmcasurablc component, a combination of the two, or possibly the historical exposure. orientation of the EMF with respect to the earth's static Wrxrim~ Confi2uration Codes. The most commonly magnetic ~cld. Certain frequcncics may bc of greater employed surrogate measure for residential BMF or lesser significancc. Transicnts (sudden changes in exposure is thc clcctric power transmission or power an electric or magnetic field) may be important, and distribution line wiring configuration code (WCC). In certain field intensifies (i.e., the window effect) have the initial study of childhood cancer in Denver, been suggcstai as possibilities to help explain some Werthcimcr and Leeper (1979) catcgorizai high- limited and aplarcntly contradictory experimental data. ct~ii~t configuration (HCC) homes as those close to a To date, cpidcmiologic studies have not reported an number of specific types of wiring which had the association between directly-measured electric or potential to carry high currents. All other wiring magnetic fields (or other exposure metric) and he4tlth configurations wcrc considered low-current effects, but this may bc because exposure assessment configurations (LCC). In their later study of cancer in has been inadalu~_tc. adults, Worthelmer and Leeper (1982) expanded their Magnetic fields have recently been regarded as the wiring coding system to four categories: very high exposure component likely to be of greatest concern, current configurations (VHCC), ordinary high current but this emphasis 'is not based upon clear-cut configurations (OHCC), ordinary low current cpidcmiologic cvidenco of an association between configurations (OLCC), and end pole configurations. cancer and measured magnetic fields. Electric fields This latter category referred to houses situated beyond have received less allcntion in recent studies of cancor the pole at the end of · secondary line with no and residential exposure because few or no biologic distribution wires running past. effects have been demonstrated in experimental animal In the case-control study of childhood cancer in Denver studies. In addition, while electric fields are relatively by Savitz ctal. (1988) which attempted to replicate the easily shielded by structures and other barriers, Worthelmer and Leeper (1979)study, five categories magnetic fields arc not. In fact, wiring configuration were used for the wiring configuration codes: VHCC, codes wcrc originally dcvisal to predict magnetic fields OHCC, and OLCC. plus a very low current rather than electric fields. Regardless, it is somewhat configuration (VLCC) and a buried category. Thus, premature to dismiss electric ftclds as a possible agent, these two studies used similar, but not completely especially in the studies socking to evaluate comparable, coding systems. occupational EMF exposures. Field Measurements and Their Correlation with Wire EXPOSURE MODELS ~9nfi~uration Codes. An estimate of the reliability of Exposure models generally consist of some form of wiring configuration codes can be made by comparing algorithm derived to estimate EMF exposure based on the magnetic field measurements in 432 homes made by a set of known physical parameters such as wiring Savitz et ·1. (1988) at times of high and low power configuration, distance from wirea, average voltage and usage with the field measurements near 41'/homes current, and behavior patterns. The use of exposure made by Worthelmer and Leeper (1982). The field models may be suporior to the direct use of exposure measurements made by Savitz el al. (1988) were measurements for current exposures because the obtained in a number of rooms inside each house, and influenco of such factors as appliance use over time can an average value for each house was used, while those be factoral into a model. Exposure models using made by Worthelmer and Leeper (1982) were obtained available historical information that can be validated at a point close to the pan of the house nearest the using current measurement data are particularly useful distribution wires. The porcent of homes with for studies dealing with past unmcasurcd exposures. measured fields gre~_~ter than 3 milligauss (raG) and the median snd maximum values for wire code categories F. XPOSURE SURROGATES suggest fairly good agreement between the two coding In retrospective case-control studies, it is ot~cn systems, but poor discrimination because of necessary to establish or define certain surrogate considerablc ovcrlap bctwccn categorics. measures of exposure which will help to quantitatc · Worthelmer and Lccpcr (1982) suggested that wire ilkely historical exposure which cannot be directly codes remain stable over long poriods of time and, measured. The use of such surrogate measures to infer therefore, might provide better measures of historical potential individual EMF exposure is not uncommon, field levels, but this suggestion is not based on but the process is fraught with uncertainty. On the quantitative evidence. The measurements made by other hand, surrogate measures, though potentially Savitz et al. (1988) in the case-control study in Denver inaccurate, may result in less exposure misclassi~cation homes mentioned above appear to offer some support than when inherently variable, short-term, spot for this suggestion, as does the study in the Seattle area by Kaunc et al. (1987). In their study, Kaunc el al. Epidemiolog'y of Health Effects and Exposure to EM~F ~-7 (198~) found a slighdy stronger correlation between been considered in most of the studies using these spot measurements and 24-hour averages for measured methods. In addition, potential confounding factors magnetic fields (r:0.$) than between wiring codes and such as socioeconomic status have been described, but 24.-hour averages (r--0.41). Although this correlation not carefully estimated or controUed in most studies. has been interpreted as suggesting fairly good Dc Guir~ ctal. (1988) and V&gcr6 ct al. (1985) agTecmcnt, a very large percent of the actual variation selected subjeers who had cvcr worked in the remains unexplained (75-83% when 1-r2 is used to telecommunications industry. OIL. et al. (1985)studied eatimatc the unexplained variation). W'xre code eleetricai engincen, Mllham (1985b) and Peaace ctal. surrogate data have recently been correlated with (1989) chose several occupations which they magnetic field levels in a relatively unsophisticated considered to bc exposed to EMF. However, MLlham approach to modeling (Flynn, 1990). If magnetic fields excluded cleetrical engineers because hc thought the arc like other environmental agents, a single 24-hour EMF exposure of electrical engineers was *infrequent' measurement will bc a very imprecise indicator of long- and their potential social class might bias the mortality term or historical exposure levels at the measurement ratio. location. An exposure model based on specific measurement data combined with historical local power In cpidcmi61ogic studies where information was consumption trends might provide a better index of past abstracted from death certificates, census codes for exposure than isohted short-term measurements occupation were used (Thomas et al., 1986, 198'7) to themselves. eatimatc EMF cxposuro. In the htter study, the job entry in a study subjeet's work history was assigned a Intuitively, one would expect that, if there is an three-digit standard industrial classification (SIC)code association between magnetic fields and cancer, the for industry (Office of Management and Budget, 19'72) association would bc stronger for the moro directly and a 1980 Census code for occupation (Bureau of the measured fields than for wiring codes. The fact that Census, 1982). In these studies, EMF exposure was Savitz et al. (1988) found the reverse to bc truc in their estimated indirectly by using a surrogate measure study may suggeat that the association of wiring codes without knowledge of actual EMF exposures. The with cancer may not bc causal, or that spot occupational titles considered as surrogates were measurement clata do not provide accurate bases for assumed to indicate a higher levels of EMF exposure exposure classification. In fact, none of the cancer than other job titics, but the intensities of individual odds ratios calcuhted for magnetic fields measured exposures were unknown. In the study by Linct al. under high and low power usage conditions was (1985), occupations wcrc grouped according to !cvcl of statistically significant in the Savitz ct al. (1988) study. likely exposure to EMF. The exposure category for This lack of association may bc duc to the smaller each occupation was determined in consultation with an number of study subjects for whom data on measured industrial hygienist, an occupational physician, and a magnetic fields were available, as well as to the radiation physicist. However, data on duration of variability in the measured data. occupational exposure were not available. It was only when the wire code at the subjcct's Preliminary studies are underway to describe the residence two yeats before diagnosis was used in the correlation of job title surrogate data with measured analysis that the odds ratio for one category achieved occupational EMF exposures (Peters, 1990). In order statistical significance. This ratio is somewhat to measure actual occupational EMF exposures, Peters imprecise duc to the small number of observations, ctal. (1990) quantified exposure to EMF among the which included only 8 cases and 2 controls in the cleetrical occupations and among a representative VHCC e-e_e_egory (Savitz et al., 1988). Since a lower sample of non--cleetrical occupations. An EMDEX participation rate was observed for controls in the dosimeter was used to measure cleetric and magnetic VHCC category than other WCC categories, it has field exposures over an entire work shiR at 2.:5-second bccn suggested that the elevated risk ratio could have intervals. The study indicated considerable variability resulted from differential participation of controis by in magnetic field cxposures w~thin a single job category exposure status. The potenthi for this type of selection depending on the different tasks being pcrformexl. This bias is now under study (Poole and Trlchopoulos, finding demonstrated the importance of cstifnating the 1991), as was strongly recommended by participants at average exposure over a typical shiik, as well as the EPRI Workshop on EMF Epidcmiology (EPRI recording peak and transient exposure patterns. For Proceedings from CarmclWorkshop, 1991). each job category of non-electrical workers, the 3ob Titles and Industry Codes. For occupational average magnetic field exposure over a shire was shown cpidemiologic studies, job titles and industry codes are to bc both lower and less variable than that for commonly used as surrogate measures for estimating electrical workers. potential occupational EMF exposure. However, the classification scheme has not yet been standardized (Lewis, 1990). Also, the duration of exposure has not 3-8 Health Effects of Exposure to Powerfine Frequency Electric and Magnetic Fields CONFOUNDING IN EMF STUDIES from which the cases are drawn. The method of selecting controls by random digit dialing, as was done Study populations may be exposed to a variety of in the two major studies of childhood cancer and EMF known or suspcctod carcinogcns as wcll as to EMF. exposure (Savitz ctal., 1988; London ctal., 1991), For cxamplc, in studio associating residential EMF may result in under asccrtainmcnt of controls from the exposure to the risk of cancer, some a=cntion should be lower sociceconomic groups and controls with a given to othcr residential exposures tha~ are known to greater stability of residence than cues. There is some be associazcd with cancer, such as radon gas and evidence suggoting that such biaso may have benzcnc cmissions, which may also be correlated with occurral in beth studies; this possibility is being the otimaxc of EMP exposure usai in the study, studied by several investigators (Poolc and Study of the same Denver subjccu used in the Savitz ct Trichoupoulos, 1~91; EPRI Workshop Proceedings, al, (1988) study of magnetic fields, for example, 1991). showed a weak, but statistically significant, association 3,2 U.S. Cancer Mortality Rates between cancer and traffic density, a surrogate for motor vchiclc emissions and benzene exposure (Savitz and Trends and Pcingold, 1989). Some aucrnpt was made in the It is instructive at this point to cxaminc the time trends Savilz ct al. (1988) study to cval,_~t_~_ the eftnets of other for the various cancers of concern, In thc subsalucnt cxposores on canca risk, opccially thoso exposures review of the cpidcrniologic evidence regarding EMF which may 'confound' or distort the possiblc exposure and adult or childhood cancers (scc section association between magnetic fields and cancer, A truc 3.3), the sito of concern include: total cancer, confounding factor would be one tha~ is reiar~cd to beth lcukcmia, central nervous system (CNS) cancer, and, wiring codes and canca. One suggorgd factor is to a lesser extent, breast cancot, Since lung cancer has traffic density, which Savitz ct aL (1988) examined, come to rcprocnt such a large porecntagc of total concluding that 'although traffic density did seem to bc cancers and tends to dominate the overall trend, this associa~l with both canca incidcnco and wire codes, site is also considcral. those associations wcr~ not strong enough to confound the association ber, vccn wire codes and cancer.' Much of the concern surrounding the EMF issue has Although no data were provided to support this arisen from reported increases in various childhood assertion, it seems likely tl~ the elcvaXcd risk ratios and/or adult cancers, Consequently, cancon have bccn may have ruultod from some unrccognizal bias, rather examined dichotomously by age. For this analysis, the than confounding, which was bclXcr controUed in the word 'childhood' means porsons of ago 0 through 19 study by Savitz ctal. (1988) than in the study by years and the word 'adult' means persons of age 20 Worthtimer and Lccpcr (1979), and above, Age- and sex-specific cancer mortality rates for the United States were obtained from the In general, analysis for potential confoundcn would be American Cancer Society for each of the lcading limitai by the small sampic*sizo available for specific canccr sites and for cach ycar back to 1930 cancers and would generally not bc very informativc, (Silverberg, 1990), From these data, annul, agc- However, the 'wire code effect' was reported to bc adjustcd mortality rates were calculated for (male and most pronounced among fcmalo, oldcr children, those female) children and adults. A variety of additional who lived in multi-family housing, low social class, and terminology, analytical techniques, and ds~_n_ sources those whoso mothers smoked during pregnancy. These are typically utilizal in thc evaluation and quanti~cation obsorvadons could imply cffcct modification or that of morbidity and mortality data. Some of the more othcr corrclatos of wire coda are responsible for the commonly used data sources and their limitations are reported association. reviewed in Appendix B. Also, a number of additional Additional and expanded discussions of exposure commonly used terms and tcclmiqucs for analyzing assosmcnt arc found in section 2,0, Appendix A, and morbidity and morality ~ arc briefly describcd in in the tutorial in Appendix B on exposure assessmcnt in Appendix B or defined in the Glossary, ~idaniologic studio, 3,2,1 Total Cancer Mortality SELECTION OF CONTROLS IN EMP STUDIES Prom Figure 3-1, it can be sccn that the crude cancer The published studio of EMF exposure and cancer mortality ~t_~ in the U.S. for main and fcmales cmphuizc thc importance of sciccting a control or combined have ncarly doubled since 1930. However, reference population that is truly representativc of the whcn the shilting age distribution of the population at study population. The potcntiaJ for bias duc to risk is udcen into consideration through agc-adj ustcd differential paucrns of response in the selection of nun, it is apparent that the increase has bocn a much controls can occur in case-control studies when thc more modest 19 percent (Figure 3-2)- Thc age- procedure for sclccting controls does not ensure a adjustcd rates for fcmales have actually ticcreased by reprcscntativc sample from thc underlying population about 10 pcrccnt over the pcrind from 1930 to 1987, F. pidemiology of Health Effects and Exposure to EMF 3-9 but the rates for males have gone up 64 percent. When began a sharp and steady decline from 1960 to 1987. adult cancers arc examined separately from childhood Pan of this pattern was duc to greatly improved cancers, the time-trend patterns arc still very similar to chemothcrapcutic measures for the childhood lcukcmias the totals for all ages combined, but quantitatively, the which have increased the expected five-year survival adult-only, age-adjusted ratca are approximately 60 for this disease. percent higher (Figure 3-3). The age-adjusted rates for 3.2.5 Brain Cancer Mortality total cancers in children (Figure 3-4) increased 54 percent from 1930 to 1945, but leveled off and then The time trend for brain and central nervous system began decreasing so that they art now about 30 to 40 (CNS) cancer is similar to that for lcukcmla, with a percent lower than in 1930. slightly mort gradual rise from about 0.7 in 1930 to 3.2 in 1955 (Figurt 3-11). At that point, the increase 3.2.2 Lung Cancer Mortality becamc even mort gradual, with ratca reaching 3.9 Lung cancer in males has bccn the leading form of deaths per 100,000 in 1969. Since then, the ratca have cause- and sox-spcci~c cancer mo~allty in the Uniteel remained rtlativcly stable in the range of 3.8 to 4.3 States since it surpassed breast cancer in females in the through 1987. There is a aught irregularity in the trend early 1960's. hcntly, lung cancers account for between the years 1978 and 1979. This occurred with about 28 percent of all cancer deaths nationwide. From the switch from the 8th Revision to the 9th Revision of about 1950 to 1970, lung cancer among males the International Classification of Diseases (i.e., underwent its porind of most rapid growth, while the switching from ICD-8 to ICD-9). Prior to 1979, rates among females did not show appreciable increases 'Malignant Ncoplasms of the Brain' included some until around 1965 (Figure 3-5). Over the 57-year secondary ncoplasms; since then, only primary tumors porind from 1930 to 1987, the age-adjusted lung cancer have been included. If this is taken into account, thert mortality rates among males and females have is still a gradual increase occurring (3.8 in 1979 to 4. i increased by 1510 and 1282 percent, respectively. in 1987). However, since the early 1980's, the ratca among When adult brain and CNS cancers (Figurt 3-12) are males have shown a tendency to stabilLzc in the range examined independently of childhood cancers, the trend of 72 to 74 deaths per 100,000 population. Lung is similar to the combined rates, but they art about 43 cancer mortality among childrtn (not shown) is to 51 percent higher. Childhood brain and CNS cancer generally very low (in the range of 0.05 to 0.15 deaths mortality per 100,000 popuhtion increased from 0.25 per 100,000). for males and 0.21 for females in 1930 to about 1.73 3.2.3 Total Cancer Mortality {Excludin0 for males and 1.28 for females in 1954 (Figure 3-13). Lung Cancer) The rates then began a gradually decreasing trend which has accelerated since about 1970. By 1987, When lung cancer deaths art subtracted from total childhood brain and CNS cancer mortality was back cancers, the effect on the time trends is impressive down to 0.75 and 0.60 for males and females, (Figure 3-7). Total cancer (minus lung) among males respectively. Whereas brain/CNS cancer and lcukcmia has remained virtually unchanged since about 1945, accounted for less than 10% of all cancer deaths in while among females, the rates have dccressed adults in the 1980's, these sites accounted for mort significantly by about 23 percent. Part of this declining than 50% of aH cancer deaths among childrtn. trend is duc to the steadily decreasing stomach cancer mortality rates, which, in 1930, accounted for about 28 3.2.6 Breast Cancer Mortality to 38 deaths per 100,000 population for females and Breast cancer mortality has remained rtmarkably stable males, respectively, while by 1987 they had decreased for females, with only minor fluctuations in the range to the range of 3 to 7 (data not shown). of 39.8 to 45.0 over the entire period from 1930 3.2.4 Lsukamia Mortality through 1987 (Figurt 3-14). Male breast cancer mortality has always been below 0.75 deaths per Mortality from lcukcmia underwent a rtlatively steady 100,000, and the time-trend has shown a gradual but increase from 1930 to 1960 (Figuaz 3-8). Since around steady decline to the range of 0.35 to 0.42 in the mid- 1968, the rates have gradually declined and, in 1987, to-late 1980's. they accounted for about 8.0 and 4.9 deaths per 100,000 in males and females, respectively. The time 3.2.7 Selected Cancer Mortality versus trends for adult-only lcukemlas art very similar to total Electric Power Consumption lcukcmias except, quantitatively, they art about 30 to Electric power consumption in the United States is 40 percent higher (Figurt 3-9). When childhood plottcd along with cancer mortality for the selected sites lcukcmias art examined separately, it can bc seen of concern in Figures 3.15 through 3.18 in order to (Figurt 3-10) that mortality rates increased sharply compart them with the temporal rtlationship of tobacco (100 to 200 percent) over the period from 1930 to consumption versus lung cancer mortality shown in 1950, plateaued between 1950 and 1960, and then 3-10 Health Effects of Exposure to Powerline Frequency Electric and Magnetic Fields Figure 3-6. This comparison provides a superficial, Subsequently, over 80 epidcmiologic studies have been but useful, summary of the secular trends for electric published which investigate the potential adverse health power consumption and cancer mortality, even though effects of residential and occupational EMF exposures. it is not known whether electric power consumption is a Studies of EMF and cancer reported through early good predictor of individual magnetic field exposurts. 1991 (and subsequently published) were selected for In fact, in the United States it is unclear whether or not the health effects literature review detailed in this there has been an increase in exposure to EMF since section. Also reviewed were several unpublished 1950 concomitant with the increase in electric power studies for which full t~ports or manuscripts were availabe. In addition, several previous review articles colBBIBptiOIL on EMF were summarized. From Figure 3-15, we see that male adult leukemia and bnin/CNS cancer mortality rat= were increasing 3.3.1 $ummarY oflarsviousRsvisws substantially before the exponential growth in U.S. Since the prolifcration of epidemiclogic studies of EMF electric power consumption was significantly and cancer began, an increasing number of review undoBray. Then, as electric power consumption began studies have also been published. Eight of these its rapid increase, leukemla and brain/CNS cancer studies were included in our review of the EMF mortality began to level off. Male adult breast cancer mortality has continued a gradual but attady decline in literature (,$avitz and Calle, 1987; Aldrich and Easterly, 1987; Ahlbem, 1988; Coleman and Beral, spite of the rapid increases in electric power consumption (Figure 3-15). The mortality time-trends 1988; Nair etal., 1989; Theriault, 1990; Hutchinson, 1991; Jauchem and Merritt, 1991). for female adult leukemia and brain/CNS cancer (Figure 3-16) show similar patterns. Female breast ,%ritz and Calle (1987) reviewed 11 studies of cancer mortality has remained relatively stable over the leukemia and occupational exposure to EMF, reporting entire period of rapid growth in electric power that there was a 'modest' excess risk for total leukemia consumption (Figure 3-16). among men in exposed occupations and an 'enhanced" risk elevation for acute leukemla, especially acute Leukemia, brain/CNS cancer, and total cancer mortality, for both male children (Figure 3-17) and myelogenous leukemla. They concluded that, the studies were inherently limited because of the absence female children (Figure 3-18), all show an increasing of exposure characterization, but that telegraph, radio, lend before electric power consumption had begun its and radar operators; power and telephone linemen; and major growth. Then, as power consumption rates electrical and electronic engineers showed the most inoreagd significantly in the 1945 to 1950 time period, consistent results and warranted further study. childhood leukemia, brain/CNS cancer, and total cancer all leveled off and began to decrease. Aldrich and Easterly (1987), in addition to reviewing a number of experimental (i.e., animal, plant, and cell It should be emphasized that examination of mortality data is not the preferred method for examining secuhr tissue) studies of EMF exposure, also reviewed and summarized 14 epidemiclogic studies of cancer and changes in population risk for childhood cancers birth defects associated with occupational EMF because of the impact of improved Uv. atment for childhood cancers on the mortality rates. Examination exposures. On the bash of the generally low risk levels observed in these studies, they concluded that if of cancer incidence would be preferable, but reliable a human cancer risk does exist, it is likely to be very incidence data for childhood cancers are not available small, perhaps on the order of 2.0 or less, and then for the extended time period of interest (i.e., 1930 to only for highly specific groups in the population. These 1987). authors suggested that future epidemiclogic studies of 3.3 Epidemiolofk: Studies the possible carcinogenio effects of EMF should take Involvin9 EMF Exposures into account other potentiaBy confounding exposures. Ahlbom (1981i) reviewed nine studies which focused In 1979, Worthelmer and Leep,:r published a study primarily on residential EMF exposure, and he repor, ing a greater number of electrical wiring concluded that, although the childhood studies seem to configurations that presumably carried high current indicate an increased risk for cancer, 'so many near the former homes of children in Denver who had methodological and theoretical concerns have been died of cancer when compared with the former homes raised against these studies that the findings must be of controls. In a second study, published in 1982, the considered highly uncertain. Ahlbom also concluded same authon r~ported an incnmaed cancer mortality that the studies on adult cancer and residential among adults who resided at several locations in exposures, provided little evidence for an association Colorado which appeared to be associated with high with all cancers together or with leukemia. While it current wiring configurations. These two studies were was not possible to daermine whether exposure to the ftrnt to suggest possible human cancer risks maguctic fields increases the risk of cancer, the associated with exposure to EMF. Epidemiology of Health Effects and Exposure to EMF 3-11 information in this review did suggest that research in for confoundca, and inaccurate exposure this area should be pursued. assessments (Theriault, 1990). Coleman and Beral (1988) reviewed seven studies of 2) The five case"control studies of lcukcmia and cancer and residential EMF exposure resulting from occupational EMF exposures conducted insts, lhtions transmitting electricity and an additional 11 subsquently were generally considered to be studies of cancer and occupational EMF exposure. more informative because of improved study They concluded that thero was no clear association desigr~s and hrgcr numbers of leukemia cases between cancot risk and residence near installations studied. These studies provided support for transmitting clectricity. Combined data from the the possible association of lcukcmia with occupational studies indicated a significant cxcess of occupational ElyiF cxposurc, but all were total leukemia and of acute myeloid leukcmia, with risk plagued with one major weakness: cxposure estimates of 1.18 (95% CI, 1.09-1.27) and 1.46 (95% assessment. EMF exposure was gcnerally CI, 1.27-1.65), respectively. However, they reported inferred from job titles and occupational that it was not clear whether the increase was specific histories secured through postal to cexmin types of work within the electrical industry. questionnaires or transcribed from They also concluded that, from the available data, it registration forms rather than from actual was not possible to deltrmine whether the increases in measurement CTheriault, 1990). leukemia were due to EMF or to other factors to which 3) Seven case"control studies on brain cancer the electrical workers wore exposed. and or.~zupational exposures were reviewed; In an extensive review of the biologic effects of power- most showed elevated odds ratios for frequency electric and magnetic fields for the Office of electrlcity-rchted occupations. In these Technology Assessment (OTA), Nair el al. (1989) studies, the numbers of cases were relatively reviewed five studies of childhood cancer and large, and, in three of the studies, an apparent residential EMF exposure, three studies of adult cancer dose-response relationship was observed. and residential exposure, and about 20 studies of However, as with the leukemia studies, leukemla, brain cancer, and total cancer in connection exposures wero estimated on the basis of the with occupational EMF exposures. They concluded reported occupations or securtxl through that there was an indieslion that occupational exposuro postal questionnaires CTherlault, 1990). in "electrical oceul~..tions" was associated with 4) The risk ratios in the cohort studies of enhanced leukemla risk, but they pointed out that electrical workers generally wore not as "associated" means 'occurs together with" and does not highly elevated as in the me"control studies, imply a cansatire link. With brain cancer and total and few excesses were statistically significant- cancer, the evidence was somewhat less substantial, Only one study OVlilhanm, 1988) reported a and their overall conclusion was that the available significant excess of acute myeloid leukemia; evidence was too weak to allow any firm conclusions. only one (Matanowski el aL, 1989) observed In a paperprepared for a plenary session of the NIOSH an excess of total leukemia; none relaorted scientific workshop, Theriault (1990) reviewed the significant excesses for brain cancer. epidemiological evidence for the risks of cancer and However, skin melanoma appeared to be other adverse health effects associated with fairly consistently elevated in five of the occupational exposure to 60/50 Hz EMF. In his cohort studies (Therlault, 1990). review, he grouped the studies into seven categories: 1) 5) Two me-control studies of eye melanoma cancer hypothesis generating studies, 2) leukernla case- were reviewed; Swerdiow (1983) reported an control studies, 3) brain cancer me-control studies, 4) elevated odds ratio (reported as the cohort studies of electrical workers, 5) eye melanoma proportional registration ratio)for electrical me"control studies, 6) welding and exposure to EMF, and electronic workers while Gallagher et al. and 7) studies of male breast cancer. On the basis of (1985) did not Fad an excess in st similar these study groupings, Theriault concluded that: group of workers (Theriault, 1990). 1) Pooled analysis of 12 of the early 6) Occupational studies conducted among proportional mortality ratio (PMR) studies has welders are important with respect to india minimal but significantly elevated potential health effects of EMF exposure risk estimates for total leukemia and acute because of the presumably high electric and myeloid leukemia. However, these magnetic field exposures. Review of 15 exploratory studies wero limited by study cancer studies in welders indicated an excess design, small numbers of observed deaths, of lung cancer but slightly decreased risk for weak statistical analyses, lack of controlling leukemla (Theriault, 1990). 3-12 Health Effects of Exposure to Powerant Frequency Electric and Magnetic Fields 7) A cohort study of telephone workcn by combinccl), tumors of the brain/CNS, total and various Matanoski (1989) rcportcd a standardized spcci~c lcukcmias, and other sclcctcd cancer types or incidence ratio of 6.5 (95% CI, 0.79-23.5) for sites such as mclanorna, cye cancer, and breast cancer malc breast cancer bascd on 2 cases obscrvcd arc rcpot~d. In terms of advcrsc effects on and 0.3 expected. Also reviewed was a case- reproduction, a number of outcomes associatcct with control study by Dcmcrs ctal. (1990) who residential and occupational cxposurc to EMF wcrc noted clcvatcd odds ratios for male breast reported in thc studies rcvicwai. Thc following is an canccr in w~3rkcrs potentially exposed to outlinc of this plan of review: EMF (OR=l.8; 95% CI, 1.0-3.2). These RESIDENTIAL EMF EXPOSURES findings wcrc intcrprctai as lending support to the hypothesis that EMF may incrcasc Childhood Cancc_.____ canca risk by interfering with thc mclatonin (6 studies, scc Table 3.1) hormonal system Crhcriauit, 1990). Adult Cancers. In a rcvicw of canccr studies with residential EMF cxposurc, Hutchinson (1991) rcpot~ai a statistically (5 studies, sccTablc 3.2) significant summary odds ratio of 1.33 (95% CI, 1.06-1.67) for childhood lcukcmia and a history of OCCUPATIONAL EMF EXPOSURES residential cxposurc to EMF. His analysis was basal Total Canccr. on five prcvious studies CWcrthcimcr and Lccpcr, (10 studies, sccTablc3.3) 1979; Fulton ct sl., 1980; Tomchins, 1986; Savitz ct aL, 1988; Coleman et al., 1989). Among the five Lcukcmia. studies, only onc (Wcrthcimcr and Leeper, 1979) reported significant increases in lcukcmia. Thc (28 studies, sccTablc3.4) statistical signiflcancc of the summary odds ratio Brain/CNS Cancer. (Mantcl-Hacrmzcl) may bc duc to significant hctcrogcncity of thc componcnt odds ratios (18 studies, scc Tablc 3.5) (Hutchinson, 1991). Mclanoma and Other Canccr Sites. The summary odds ratio for childhood CNS cancer and (:22 studies, scc Tabic 3.6) a history of residential cxposurc to EMF was statistically significant, with an odds ratio of 2.44 (95% pATERNAL/MATERNAL OCCUPATIONAL EMF CI, 1.70-3.$3) in Hutchinson*s (1991) analysis. This EXPOSURF. S analysis was based on thrcc studies that cxamincd brain ~anccrs. canca (~Vcrthcimcr and Lccpcr, 1979; Tomchins, 1986; Savitz ct al., 1988). (6 studies, scc Table 3.7) Jauchcrn and Mcmtt (1991) rcvicwcd a widc variety of Congenital Malformations, Spontaneous cpidcmiologic studica and rcvicw axticlcs addrcssing Abortions, and/or lntrautcrinc Growth cancer and other effects rcportcdly associated with Retardation. cxposurca to EMF. Bccausc of thc numcrcus inconsistcncics and dc~cicncics of thc studies (6 studies, scc Tablc 3.7) rcvicvnxi, they concludai that thcrc is currently no RESIDENTIAL EMF EXPOSURES definitive cvidcncc of an association bctwccn cxposurc Childhood Cancers. Six major cpidcmiologic studies to EMF and the alleged cffccts. cxamining various childhood canccrs and rcsidcntial 3.3.2 Fieview of Specific Studies EMF cxposurc wcrc rcvicwal (Wcrthcimcr and Thc studies rcvicwcd for this report arc dividcd into Lccpcr, 1979; Fulton ctal., 1980; Mycrs ctal., 1985; three sections according to thc circumstanccs of thc Tomchins, 1986; Savitz ctal., 1988; London ctal., presumed EMF cxposurc: residential cxposurcs and 1991) (Table 3.1). occupational cxposurcs of the cases, and prenatal (or Four of thcsc case-control studies cxamincd at total preconception) occupational exposures of the parcnt(s) cancers (V~'crthcimcr and Lccpcr, 1979; Myers ctal., ofthc childhood cases. Rcportcd health consequences 1985; Tomchins, 1986; Savitz ctal., 1988). Three of associatat with cxposurc to EMF includc various thc four studies rcportcd statistically significant rcsults cancers in adults and children and adverse cffccts on for total cancer. Wcrthcimcr and Lccpcr (1979) the fctus or reproduction. Under residential EMF reported an excess in total cancers (OR=2.22.; 95% CI, cxposurca, studies conductcd among both children and 1.58-3.12) for homes near electrical wiring adults wcrc rcvicwcci. Among occupational groups configurations suggestive of high-current flow. The potcntially cxposcd to EMF, total cancers (all sites study further rcportcd that thc association *appearcci to Epidtmiology of Hearth Effects and Exposure to EMF 3-13 be dose-related.' Tomenius (1986) reported significant Four of the five studies yielded inconsistent and weak mulls (OR=2.10; p<0.05) when magnetic fields at associations of leukemia in adults with presumed the dwelling were higher than 0.3 microtcsla (~T) (i.e., residential EMF exposure (McDowall, 1986; Preston- 3.0 mG). Savitz et al., 1988) rcportexi a slight excess Martin et al., 1988; Severson et al., 1988; Coleman for total cancer for HCC wiring codes (OR= 1.53; 95% and Bell, 1989). Them was no consistency among CI, 1.04-2.26) but non-significant mulls for measured studies for the examination of any specific type of magnetic fields 2.0 raG. Meyers ctal. (1985) found leukemia. Two studies focused only on total leukemia no increased ris~ for total cancers. (McDowall, 1986; Coleman and BeU, 1989), and the study conductod by Severson et al. (1988) was limitccl All six of the studies examined childhood lcukcmia and to non-lymphocytic lcukcmia. Preston-Martin et al. residential EMF exposure, but the rcsulls appeared to (1988) examined acute and chronic mycloid lcukcmias be inconsistent. Two of the studies ('vVcrthcimcr and and observed non-significant odds ratios which wcrc Lccpcr, 1979; London et al., 1991) reported statistical less than one. significance on the basis of Wire code configurations with odds ratios of 2.35 (95% CI, 1.55-3.56) and 2.15 OCCUPATIONAL EMF EXPOSURES (95% CI, 1.08-4.26), respectively. However, when Fifty-one cpidcmiologic studies which wcrc designed to London et al. (1991) analyzed the data on the basis of detect possible associations between various cancer the measured 24-hour average magnetic field strength sites and occupational EMF exposure wcrc reviewed in the child's bairoom, the resulls wcrc not significant (Tables 3.3, 3.4, 3.5, and 3.6). These studies focused (p for trend = 0.74). The rcffmlning studies did not report any significant associations of lcukcmia with on a variety of different occupational group generally classified as electrical workers, including electricians, EMF. electrical engineers, electric power station operators, Among the six residential EMF studies reviewed, three linemen, and others. They wcrc carded out in different had no ~a_t_a on childhood CNS or brain cancer, and countries and employed a variety of different study three rcportai elevated odds ratios: Worthtimer and designs, including cohort, case-control, and PMR Lccpcr (1979) with an OR of 2.86 (95% CI, studies. The reported rcsulls wcro found to be 1.64-4.98), Tomchins (1986) with an OR of 3.7 generally inconsistent. (p<0.05), and Savitz et al. (1988) with an OR of 2.04 Total Cancer. Ten of the 51 cpidcmiologic studies (95% CI, 1.11-3.76) for high current configuration examined total cancers associated with presumed (HCC) wiring codes. When Savitz et al. (1988) occupational EMFcxposurc (Howe and Lindsay, 1983; analyzat on the basis of magnetic field strength _ 2.0 VAgcr~ and Olin, 1983; BarregArd et al., 1985; raG, the CNS/brain cancer results wcrc not significant Millram, 198Yo; Olin et al., 1985; VAgcr~ et al., 1985; (OR, 1.04; 95% CI, 0.22-4.82) TSrnqvist et al., 1986; Lin, 1987; Milham, 1988; Adult Cancers. Five cpidcmiologic studies of adult Gub~ran et al., 1989) (Table 3.3). One study was canccn and residential EMF exposure wcrc based on proportional data (PMR), and the remaining rcvicwed(%Vcrthcimcr and Lccpcr, 1982; McDowall, nine studies wcrc of a cohort design. 1986; Preston-Martin et al., 1988; Scvcrson et at., Of the ten studies examining total cancer, three cohort 1988; Coleman and Bc~, 1989) (Table 3.2). studies (Howe and Lindsay, 1983; VAgcr~ and Olin, Significant rcsuils for total cancer wcrc reported in only 1983; Lin, 1987) and one PMR study (Milham, 1985b) one of the studies. The case-control study by showed a weak association between presumed Wcrthcimcr and Lccpcr (1982) reported data which occupational EMF cxposuro and total cancer. Although produces an odds ratio of 1.28 (95% CI, 1.08-1.52; the risks wcrc only s~ghtly elevated, they wcrc all p(0.005) for total cancer. They also reported statistically significant. One additional study (T~rnqvist "significantly high C-ratios" (i.e., the ratio of the ctal., 1986) demonstrated a weakly positive but non- number of case-control pain with the case exposure significant SMR. Four studies (BarregArd ctal., 1985; hlghcr to the number of pairs with the control exposure Olln et al., 1985; Miiham, 1988: Gub~ran ctal., 1989) higher) for lymphomas and cancer of the nervous reported risk cstimatcs for total cancer and EMF system, uterus, and breast, but the individual numbcn exposure that wcrc less than 1.00. wcrc not presented. In comparing the rcsulls of the above ten studies, it is The cohort study by IvicDowall (1986) reported a in~portant to rcnncmbcr that the definitions for "total significant elevation for lung cancer (SMR, 2.15; 95% cancers" wcrc not identical. In addition, industr), codes CX, 1.18-3.61) when the cohort was grouped by and job titles for the "exposed" workers differed from distance from electrical installations, but icukcmias study to study. These disparities in definition and (SMR, 1.43; 95% CI, 0.04-7.96) and total cancers methodology, combined with the inconsistencies of the (SMR, 1.03; 95% CI, 0.68-1.50) wcrc not significantly data, make it impossible to determine whether there is a clcvated. 3-14 Health Effects of F. xposure to Powerfine Frequency Electric and Magnetic Fields likely eauul usoeiation between total cancers and 1987; LinetetaI., 1988; MLIham, 1988; Garland etal., occupational EMF exposure. 1990). The remaining six cohort studies reported non- significant risk estimates close to (or less than) 1.00. Leukemia. The possible uaoeiation of leukemla with The sample sizes in the cohort studies were generally occupational EMF exposure has received considerable large enough to achieve adequate statistical power, but attention and stimulated numerous occupational were marginal or insufficient when total lcukcmias cpidcmiologic studies. Twenty-eight studies which were separated into the various lcukcmla subtypes and examined the association of lcukcmla with EMF when exposed workers were segregated into different exposure wctc reviewed CTablc 3.4). Eight of these job categories. studies were bucd on propo~onal ~t, (PMR or PIR studiu), 11 were cohort studies, and 10 were cuc- In seven of the 10 cuc-cont~l studies reviewed, control studies, one of which a/so reported PMR data sntd~tically significant associations wcre obscrved in (McDowcll, 1983). Overall, the cue-control studies one or more of the specific occupational groups tendcd to produco the greater number of significant presumed to bc exposed to EMF (McDowmil, 1983; findings (seven out of 10 studies) followed by thc Gilmam ctal,, 1985; Pcarcc ctal,, 1985; Flodcn ct al,, PMR/PIR studies (five out of cight) and the cohort 1986; Stern c~ al., 1986; Preston-Martin and Peters, studies (five out of 11). 1988; Pcarcc ctal., 1989), Excesses for total lcukcmla were sccn in four cuc-conlxol studies (Oilman ctal., Five of the cight PMR or PIR studies (Milham, 1982; 1985; Pcarcc ctal., 1985; Stern, 1986; Pesrec ct Wright ct 81,, 1982; Callc and Savitz, 1985; Mllham, 1989) and cloysteal risks for acute mycloid lcukcmla 1985a and 1985b) dcmonstnt~ weak, but statistically were reported in four studies (McDowall, 1983; significant tuociations, Milham (1982) studied Mi]ham, 1985a; Flodin, 1986; Pcarcc ctal,, 1989), lcukcmia mofiality among workcn exposed to EMF and reported a PMR of 1.37 (95% CI, 1.15-1.62), McDowall (1983) studied acute mycloid lcukcmla in Wright ct m/. (1982), in a similar study looking at selected cloctrical workcn and found a relative risk of incidence data, found a PIR of 1.73 (95% CI, 1.10o 2,3 (95% CI, 1.4-3,7) for all electrical occupations. 2.59) for acute lcukcmia and a PIR of 2.07 (95% CI, Gllnmn ctal. (1985) observed elevated risks for total 1,30-3.14) for acute myclogenous lcukcrnla, Callc and lcukcmla (OR=2,53; p <0.05), myclogcnous lcukcmla Savitz (1985) observed a PMR of 1,86 (p<0.05) for (OR=4,74; p<0,05), and chronic lymphocytic total lcukcmla and a PMR of 2,57 (p<0,05) for acute lcukcmia (OR=6,33, p<0.05) among underground lcukcmla aunong clccl~cal cnginccn. They Mso coal miners, ~ ctal, (1985) examined MI adult observed a PMR of 2,35 (p < 0.05) for total lcukcmla male cancer cuca in the New Zealand Cancer among radio and telegraph operaton. In a study of Registry. Their findings sugguted an increase in the artmteur radio ol~cratora, Mi]ham (1985a) reported risk of lcukcmla among electronic equipment significant elevations in the PMR's for mycloid asscmblcn (OR, 8.17; 95% CI, 1.49-44.7) and lcukcmia and total lcukcmla (2,81 and 1.91, radio/television repairmen (OR, 4.75; 95% CI, 1'59- respectively). In a death certificate study in 14,2). Coggon ctal. (1986b) noted that five out of 29 Washington state, Mi]ham (1985b) observed a PMR of patients with acute mycloid lcukcmia had worked in 1.36 (p<0,01) for total lcukcmla and a PMR of 1.62 electrical trades, but the statistical aignificancc of this (p<0.01) for acute lcukcrnla among all electrical finding was not reported. Flodin ctti. (1986)studied occupations, cases of acute myclogcnous lcukcrnia and reported an cloysteal odds ratio of 3,8 (95% CI, 1.5-9.5) for In general, the validity of the findings from a PMR cloctrlcal workers (a category which included electrical study depends on whether the deaths included in the tcclmicians, electrical wcldcn, and computer-tclcphonc PMR arc represcrttativc of all deaths in the total mochahies). Stern ct M, (1986) studied lcukcmia cxpo~:d population. A PMR study is a retiablc among naval shipyard workers and reported indicator of risk only when the healthy worker effect is significantly increucd odds ratios for total lcukcmla of equal strength for the disease of interest (c,g,, among electricians (OR, 3,0; 95% CI, 1,29-6,98) and lcukcmia) and for MI causes of duth in the exposed for mycloid lcukcmia among welders (OR, 3.83; 95% population (Chcckowxy ct al., 1989), In addition, the CI, 1,28-11.5). In this study, a detailed history of potential for misclassi~cation exists, in that the occupational radiation exposure was obtained and the information used in most PMR studies is ascertained analysis controlled for these and other occupational from duth certificates or canccr registries. Accuracy exposures. Preston-Martin and Peters (1988) reported of the information of exposure and case dot'tuition a highly elevated odds ratio (OR, 25.4; 95% CI, 2.78- (diagnosis) are not guaran~__-~__. PMR studies, 232,5) for chronic mycloid lcukcmla associated with thcrefore, arc less reliablc as a basis for estimating prior employment as a wcldcr. Pcarec ctal. (1989) risks than arc other typ~ of cpidcmiologic studies. obscrved an clcvated risk among radio/tclcvision Among the 11 cohort studies, five reported onc or repairmen (OR, 7.86; 95% CI, 2.20028.1) and power more significant finding (Lin, 1987; TSrnqvist ctal., Epidemiology of Health Effects and Exposure to EMF 3-15 station operators (OR, 3.89; 95% Cl, 1.00-15.2) but uncertainty of the exposure assessments. In most of the not among electrical linemen. studies revicwexi, exposure to EMF was inferred on the bash of job title and/or employment in a specific In general, the case-control studies demonstrated a industry. Furthermore, most studies failed to carefully stronger and more consistent association between consider and control for potential confounding factors, lcukcmia and occupational EMF exposure than did the such as specific relevant occupational exposures, which cohort studies, although, in some cases, the high odds may have account~l for some or all of the apparent ratios were associated with wide confidence intervals increases in morbidity or mortality. Most importantly, (Pearee et al., 1985; Stem el: al., 1986; Preston-Martin the majority of the studies with the strongest study and Peters, 1998). design (i.e., the cohort studies) did not report a Preston-Martin and Peters (1988) reported a highly significant association of EMF exposure with leukemia elevated odds ratio (OR=25.4) in a study of 137 risk. Few of the studies of electrical workers chronic myeloid leukemia cases, 19 of whom reported accounted for potential confounding factors such as prior employment as welders. However, based on an exposure to polyehlorinated biphenyls (PCB's), earlier review of 15 cancer studies in welders (with a solvents (e.g., benzene), or ionizing radiation from pooled total of 146 leukemia c. ues) Stem (1987) radon or radon daughtors, from diagnostic radiography, described an excess for lung cancer but not for and from x-rays produced by high-voltage cathode ray leukemia. The job title groupings used in these studies, tubes (CRT's) commonly used prior to the 1970's. however, were not comparable. In general, studies of Brain/CNS Canee.r. Following leukemia, brain or CNS welders suffer from a major confounding factor which cancer has been the second most frequent site to be arises from the concurrent exposure to high levels of investigated with respect to occupational EMF metal fumes, a number of which ag known human exposure. Eighteen occupational studies of brain/CNS eareinogens. Furthermore, since some welders employ cancer were reviewed and are listed in Table 3.5. gas welding methods (e.g., acetylene, hydrogen, town Although different morphologie types were examined in gas, and propane) instead of are welding, exposure to some studies, most did not focus on a specific type of EMF in this occupational category may be highly brain/CNS cancer. One of the studies reviewed used variable. proportional (PMR) data, seven were cohort studies, Increased risks for several different types of leukemia and 10 were of a ease-control design. were detected in a study of underground coal miners by Milham (198Yo), in a death eenifieate study in Gilman et al. (1985). T~mqvist el al. (1987) also Washington state, found a PMR of 1.23 (p(0.05) for reported an elevated $MR of 2.1 for acute myeloid deeedents with occupation coded as any one of nine leukemia among miners and rockblasters. Miners are eloctrieal occupations. presumed to be exposed to EMF through overhead lines used for distribution of power to lights and mining Two of the seven cohort studies of bnin/CNS cancer equipment and through electrically operated trolleys reporW. d significant SMR's or SIR's CLin, 1987; used for transportation of men and materials. However, McLaughlin el al., 1987). In a study of Taiwan magnetic field exposure in underground coal mines is Electric Power Company employees, Lin (1987) low enmpared to aboveground measurements in observed an SMR of 4.10 (95% CI, 1.77-8.08) for residential areas, and electfie fields have not been brain/CNS cancer. MeLaughlin el al. (1987)found an measured in coal mines (Giltnan el al., 1985). Also, a SIR of 1.4 (95% CI, 1.02-1.87) for intracranial gliomas number of significant confounders for this occupational among welders and metal cutters and an SIR of 1.1 group, such as exposure to radioactive mineral dusts, (95% CI, 0.98-1.23) among workers in the machinery radon and radon daughters, and diesel exhaust and electronics industry. missions, were not taken into account. Combined, the Six out of ten ease-control studies repor~.d statistically undocumented exposure levels and presence of possible significant odds ratios (Linet al., 1985; Thomas et al., confounders weaken the argument that the excess 1987; Speers et al., 1988; Loomis and Savitz, 1989; leukemias resulted from EMF exposure. Other studies peatee et al., 1989; Preston-Martin et ai., 1999). An which looked at leukemia mortality among a wide exposure-response relationship between CNS cancer range of occupational groups (Howe and Lindsay, and presumed EMF exposure was observed in four of 1983; Blair el al., 1985; Linel el al., 1988) do not the six studies (I. in et al., 1985; Thomas el al., 1987; confirm the excesses in leukemia among miners. Speers el al., 1988; Preston-Martin el al., 1989). The findings of an association between ieukemia and In a study of glioblastoma multifurme and astroeytoma potential occupational EMF exposures in the studies deaths, Lin el al. (1985) grouped occupations according reviewed are modestly suggestive of a possible causal to probability of exposure to EMF. An OR of 2.15 association. However, the conflicting data and the (95% CI, 1.104.06) was reported for the "definite" design limitations of the studies cannot be ignored. EMF exposure category, and lower OR's were found at Most important among these limitations is the 3-16 Health Effects of ~xposure to Powerline Frtquency Electric and Magnetic Fields the lower Icyeli of probable EMF exposure. Electric Mclanoma and Other Cancer Sites. With increasing and telephone linemen/serviccmen seemed to have the attention being given to the possibility of an association highest mortality (SMR, 3.73; 95% CI, 2.24-5.82), but between occupational EMF exposure and lcukcmia or electricians and electronic engincen/technicians were brainYCNS cancer, more studies have begun looking also elevated (SMR's 2.28 and 2.50, respectively). In for associations with other sites. Twenty-two Thomas et al. (1987), microwave and radio frequency occupational studies wer~ reviewed which examined (MW/RF) EMF exposure was estimated on the basis of presumed EMF exposure and specific cancer sites length of employment in various categories of other than leukemia and brain/CNS cancer (Table 3.6). electronics and electrical jobs. A significantly elevated Two studies used proportional data including PMR's odds ratio of 2.3 (95% CI, 1.3-4.2) was found among and proportional registration ratios (PRR's), 12 were men pr~umed to be exposed to MW/RF EMF for 5-19 cohort studies, and eight employed case-control study years. designs. In a study of east Texas residents by 5peen et al. Swerdlow (1983)studied total eye cancer as a surrogate (1988), occupations wcr~ grouped according to for eye melanoma in England and Wales and reported probability of exposure to EMF. Electric and significant PRR's for electrical and electronic workers ~lephone company employees, electricians, electronic in 3 out of g years studied. However, engineers, and railroad and telecommunication professional/technical workers (with no specific EMF engineers were among the group categnrized as having exposures) were found to have significantly elevated "definite exposure* to EMF. The OR for the 'probable PRR's for 5 out of the 8 years. In a case-control study exposure" group was 2.86 (p<O.04), and there wore in westem Canada, Gallagher et al. (1985) examined six cases and no controls in the "definite exposure" ocular melanomas among electrical and electronic group, producing a significant Linear trend with workers and found no increased risk. Significantly p<0.01. In a death certificate study, Loomis and elevated 5MR's for malignant melanoma (of the skin) 5avit-z (1989) reported an OR of 1.5 (95% CI, 1.0-2.1) wore demonstrated in thr~c out of four cohort studies for brain cancer among all electrical workers (V~tger6 and Olin, 1983; V~tger6 et al., 1985; De Guire combined. Pearee et al. (1989) found an OR of 4.74 et al., 1988). Viger6 and Olin (1983) studied various (95% CI, 1.65-13.6) for brain cancer among electrical cancon among workers in the electronics industry and engineers, but the OR for electricians was not reported a relative risk of 1.35 (95% CI, 1.05-1.76) for significantly elevated. Preston-Martin et al. (1989) skin mehmoma. Viget6 et al. (1985) examined various examined the occupational and other risk factors for cancers among workers involved in the manufacture of primary brain tumors (glionm, including astrocytemas, telecommunications equipment and found an SMR of and meningiomas) among rnales in Los Angeles 2.6 (95% CI, 1.3-4.5) for melanoma among all County, 1980-1984. They found thai the odds ratio for workers. De Guire et al. (1988) found 10 cases of skin prior employment in jobs likely to involve high melanoma among telecommunication workers, exposure to electric and magnetic fields was significant producing an SIR of 2.7 (95% CI, 1.35-5.02) for males only for gliomas and that the risk was greatest for (no female cases were seen). None of the three case- astrocytomas. control studies that examined malignant melanoma among electrical and electronic workers reported Half of the studies reviewed (Olin et al., 1985; VAger~ significantly increased risks. et al., 1985; Coggon et al., 1986b; T~rnqvist et al., 1986; Magnani et al., 1987; Milham, 1988; Gub~ran et Only two out of 11 studies looking at lung cancer al., 1989; Reif et al., 1989; Lewis, 1990) failed to among persons presumed exposed to higher-than- demonstrate a statistically significant association normal levels of EMF reported significantly elevated between EMF exposure and cancer of the CNS. results (V/tger~ and Olin, 1983; Milham, 1985b). Furthermore, two me-control studies (Reif et al. 1989; Viger~ and Olin (1983) studied various cancers among Lewis, 1990)actually showed reduced odds ratios for workers in the electronics industry and reported a brain/CN5 cancer in the groups pr~umably exposed to relative risk of 1.52 (95~ CI, 1.35-1.72) for lung EMF. As with the occupational studies of leukemia, cancer in males among all electronics workers. the estimates of potential EMF exposure wore based on Milham (19gSb) examined death certificates in job titles and/or industry codes. Consequently, the Washington state and observed a PMR of 1.14 assessment of EMF exposure in the studies of CNS (p<0.01) for cancer of the lung, trachea, and bronchus cancer was generally considered to be inadequate. among all electrical occupations. Three of the studies While an apparent dose-response relationship between rc~rtecl significantly lower risk for lung cancer among surrogate measures of EMF exposure and CNS cancer various electrical occupations (Blair et al., 1985; was observed in a few of the studies, the significance of T~rnqvist et al. 1986; Milham, 1988). these ~'mdings remains questionable, given the Two studies examining female breast cancer and EMF uncertainty of the exposure assessments. exposure, one occupational (V~ger~ et ai., 1985) and one of home electric blanket use (Vena et al., 1991), Epidemiology of Health Effects and Exposure to EMF 3-17 reported no increased risk. Matanoski (1989,1991) Childhood Cancers. Six studies exploring the possible reported a standardized incidence ratio of 6.5 (95% CI, association of childhood ncurobiastoma or cancer of the 0.79-23.5) for male breast cancer among telephone CNS with presumed paternal EMF exposures were workers. This finding was based on 2 eases observed reviewed (Spitz and Johnson, 1985; Nasca et al., 1988; and 0.3 cases expected. Demcrs et al. (1990,1991) Wilkins and Koutras, 1988; Johnson and Spitz, 1989; noted elevated odds ratios for male breast cancer in W'tlkins and Hunalley, 1989; Bunin et al., 1990) (Table workers potentially exposed to EMF (OR, 1.8; 95% 3.7). Three of the six studies (Spitz and Johnson, el, 1.0-3.2). Most important among the studies of 1985; W'ilkins and Koutras, 1988; Johnson and Spitz, male breast cancer is the study by Tynes and Andersen 1989) showed a statistically significant association (1990) who observed 12 c~ues with 5.8 expected between ncuroblastoma or childhood cancer of the CNS among Norwegian workers with potential EMF and paternal occupations with potential EMF exposure. exposures (SIR, 2.07; 95% CI, 1.07-3.61)- These The magnitude of the association reported in these findings have been interpreted as lending support to the three studies was similar. W'tlkins and Hundley (1989) hypothesis that EMF may increase cancer risk by calcuhtetl the odds ratios for childhood neurobhstomas interfering with melatonin production (Theriault, 1990). using a variety of methods to estimate potential paternal They may also be interpreted as suggesting the EMF exposure based on occupation and industry importance of shiR-work and its effects on mchtonin c~xles. While the resulting odds ratios were generally production as an independent risk factor for male breast comparable to the previously reported positive studies, cancer. they were not statistically significant. The studies associating paternal occupational EMF exposure with Only one study (Vtgcr6 and Olin, 1983) out of five cancer in offspring are difficult to interpret since an studies examining bhddcr cancer and occupational effect, if real, would imply that EMF acts as a cancer EMF exposure reported a minimally increased risk initiating agent. To date, none has suggested a cancer a~R, 1.22; 95% CI, 1.02-1.47). Of six studies looking initiating mechanism for EMF. Most investigators at stomach cancer, only one (Howe and Lindsay, 1983) believe, instead, that if EMF affects cancer risk at all, it observed an increased risk (SMR, 2.33; p<0,05) for does so by cancer promotion or growth enhancement. telephone, telegraph, and power linemen and servicemen. Two out of eight studies (Howe and The magnitude of the association of childhood CNS Lindsay, 1983; Vigor6 and Olin, 1983) examining tumors with paternal EMF exposure was not large cancer of the colon or intestine (excluding rectum) enough to support a causal inference. In addition, the repox't.d significant elevations. One study(Milham, findings were inoonslstent in the studies reviewed 1985"o) out of five considering cancer of the pancreas (Nasca et al., 1988; W'~kins and Koutras, 1988; observed a very minimal but significant increase in risk Johnson and Spitz, 1989; W'ilkins and Hunalley, 1989; (PMR, 1.17; p<0.05). None of the five studies with Bunin et al., 1990) (Table 3.7). Two studies of CNS data on prostate cancer_ showed significant risks. Of tumors and one of neuroblastomas reported statistically three studies reviewing liver cancer data, only one significant results (Spitz and Johnson, 1985; W'dkins (Lin, 1987) reported statistical significance (SMR, and Koutras, 1988; Johnson and Spitz, 1989). 1.54; p<0.01). Although not an occupational study, However, the observed paternal occupations thought to Verreauit et al. (1990), in a case-control study of white be associated with childhood cancer of the CNS in males with testicuiar cancer in western Washington, these studies were not specific electrical occupations, found that the odds ratio for disease among electric but instead represented various job titles. There were blanket users was not elevated with respect to controls. no exposure-response relationships observed in these studies. In general, the weight of evidence from studies examining *other cancer sites* is even weaker than that Congenital Malformations, Spontaneous Abortions, for lcukemia and brain/CNS cancer. Many of the and/or Intrauterinc Growth Retardation. Five studies positive findings came out of hypothesis-generating (Hcmminki et al., 1980; NordstrSm et al., 1983; studies which looked at dozens of different occupation Wertheimer and Leeper, 1986; Nordstr6m ctal., 1987; or industry codes and possibly 10 to 20 different cancer Worthelmer and Leeper, 1989) (Table 3.7) have been sites. Under these conditions, statistically significant conducted to evaluate the potential adverse effects of results are to be expected on the basis of chance alone, occupational exposure to EMF on reproduction. Three and are, at best, weak arguments for possible of these studies demonstrated a statistically significant causation. As mentioned previously, lack of exposure association between presumed occupational exposure to assessment and failure to account for possible EMF and adverse effects of reproduction, including confounding factors also weaken the evidence for an spontaneous abortion, frequency of abnormal association between EMF and "other cancer sites.' pregnancy, and congenital realformation (Hcmminld et al., 1980; Nordstr6m et al., 1985; Nordstr6m et al., pATERNAL/MATERNAL EMF EXPOSURES 1987). One study in Italy reported an odds ratio of 5.9 for oligospormia and azoospormia among radio electric 3-18 Health Effects of Exposure to Powerline Frequency Electric and Magnetic Fields workers, but the results were not statistically significant technologically advanccd countries. However, some (95% CI, 0.86-40.2). tend to cite cancer statistics in a manner which appears to support the hypothesis that thc U.S. and thc rest of 3.4 Discussion the technological world are in the midst of a cancer 3.4.1 Studies of EMF and Adverse Health 'epidcmic.' Ralph Nader (1981) reports that cancer in the U.S. has increased dramatically in this centuzT. Effects Such a conclusion is based on the changing pcrccntage Since the late 1960's, increasing public av. ention has of all deaths tha~ are alIributablo to cancer since 1900. barn directed tovard the electric power companies and The supporting figures are quoted as 3 percent in 1900, the extra-high-voltage (I~HV) transmission lines and the 9 percent in 1930, and 20 perecnt in 1975. Likcwlsc, lower voltage distribution lines nccessary to deliver that Larry Agran {19T7) reports that only one in every power from the generation facillties to the substations Americans died of cancor in 1900, while nearly one in and on to the consumers. Initially, the conccrns wcre ever7 five died of cancer in 1975. focused pr4.marily on the adverse aesthctic impact of Changes in Cancer Mortality Ratio. In 1900, about 47 having large, unsightly towers and wires cluttering the percent of all deaths were duc to infcctious diseases; by skylines. However, the clcctric and magnetic ficlds the 1970's, the proportion had dropped to only 6 assocbiecl with EHV transmission lines also produced a percent (Figure 3-19). Over the tbnc period in which number of nubance effects such as audible noises, TV the proportion of deaths duc to cancer was showing a and radio signal interference, and unpleasant shocks dramatic rise, influenza, pneumonia, tuberculosis, and when touching ungrounded metal objects (e.g., can, other infectious disease Muses of mortality were trucks, or farm vehicles) whi/e standing under the El-IV undergoing dramatic reductions. These decreues transmission lines (Nair et al., 1989). resulted from a number of factors including improved With the growth of the environmental movement in the sanitation procedures, widespread purification of late 1960's and early 1970's, anention began shifting drinking water, and the development of antiblotics and more toward the possibility that power transmission or vaccines for many of the potentially lethal childhood distribution lines might have an adverse effect on and adult infectious diseases. Thus, people who once human and animal health. These fun were reinforced would have died in childhood or in the prime of their when Sovict and eastern European investigawrs lives as a result of some infectious disease had begun rq}ortcd an increasai number of nourologic, living to a much older a~e. The average life cardiovascular, hematologic, and other nonspecific expectancy st birth increased from about 47 years in ailments among workers in EHV switchyards 1900 to over 75 years by 1987 (Figure 3-20). These (Annova, 1966; Sazonova, 1967; Filippov, 1972; factors resulted in a shift in thc age distribution of the Korobkova, 1972). S~wc nogative studies are entire U.S. population, characterized by a frequently ignored or minimizcd, it did little to case the disproportionate increase in the number of people in the concerns of the public when Western scientists older age group (Figure 3-21). An incvitablc conducted similar studies and failed to co nfLrm the consequence of such a shift is an increase in the earlier findings (Strumza, 1970; Robtree, 1976; lsscl, percentage of people dying of diseases which are 19'7'7; Knave, 1978). typieaUy associated with older age such as heart diseasc and cancer, although since 1970, age-adjusted mortality When Werthcimcr and Leeper (1979) reported an duc to coronary heart discue and stroke has decreased clevstlon in lcukcmia, CNS tumors, and total eanccr markedly. mortality among Denver children who lived in homes that wore cvaluated as near 'high current Sclectinl~ a Morbidity/Mortality Data Set for Analysis. configuration" powor transmission or distribution lines, Since survival times for various cancers may differ widespread alt,-ntion became focused on the issue of considerably from site to site and (with improvements cancer as a possible adverse health effect of cxposurc in treatment methods) over tiznc as weU, it is usually to ELF electric and/or magnetic ficlds. As beforo, the preferable to look st trends in cancer incidence rates as negative study of Fulton a al. (1980) {which used opposed to cancer mortality rates. However, historical similar methods to study childhood lcukcmia cascs in trends in cancer incidence rates cannot be examined Rhodc Island) did little to case public concern over the unless there has been some sort of long-term EMF/canccr issue. registration of all newly diagnosed cases of cancer from the geographical area in question. This long-term data 3.4.2 Historical Yrends in the Cancer collection process generally requires the extensive Mortality Ratio case-finding cfforLs of an active, population-based, Most authors who are knowledgeable about the cancer registry. historical trends of canccr recognize that thc discasc The National Canccr Institute's (NCI's) Surveillance, did not suddenly begin in the 20th century and that it is Epidcmiology, and End Result (SEER) Rcgistries {see not limir~l to the United States and other Appendix B) have been collecting such canccr Epidtrniology of Health Effects and Exposure to EMF 3-19 incidence data from selected geographical locations in one population as it changes over long poriods of time the country since the early 1970's; a very few states (Figure 3-,,). have canccr registries that go further back than the Age-Specific Mortality Rates. Onc method of SEER registries. Consequently, there aro no avadablc controlling for changes in age distribution is to sources of cancer incidence data that would permit a compare the age-specific mortality rates for a particular meaningful time-trend analysis of a large segment of disease in the two populations. For this analysis, the the population back to the early 1930's. The only annmtl deaths arc generally tallied for each five-year or feasible alternative, then, is to cxarninc cancer ten-year age group, and thc number of dcaths for each mortality ,4,r,, which are derived ultimately from death age group is dividcci by the popuhtlon in that age certi~catu, which in turn have been recorded with group. The ruuit for each age group is then multiplied relative consistency across the U.S. since the early by 100,000 and expressed as the number of deaths per 1930's and, in some states and/or cities, as far back as 100,000 population por year. Thus, the age-specific the early 1900's. Despite the impact of improved mortality rates for any age group of onc popuhtion can treatment on the ratio of incidence to mortality rates, be compared appropriately with the same age group of mortality data can be relied on for a useful analysis in any other population. However, when the mortality this regard. Poolc and Trichopoulos (1991) for rates are likely to be different depending on race and/or instance dcmonstratcd that data from thc Connecticut sex, it may also be necessary to look at age-, race, and CAncot Registr7 (considcrai to be thc be. st long-tcrm, sex-specific mortality rates for the two popuhtions population-bascd d_ata on cancer incidence in the U.S.) before a valid comparison can be nutde. do not show a secular increase for childhood lcukemla or CNS tumon. Aite-Adiusted MortaliW Rates. Anothcr tacthod of controlling for age distribution differences is to The Cause-Specific Mortality Ratio. In gencral, thc calculate the age-adjusted mortality rate for cach examination of time trends for the various cause- population. This method has some advantage over the specific mortality ratios (as employed by Agran, 1977 age-specific mortality rates because the age-adjusted and Nader, 1981) is of limited nscfulness. Since the mortality rate is expressed as a single rate for the entirc sum of percentages for all causes of dcath must always population. However, this characteristic may equal 100 percent, when the porcentagc for one cause sometimes conceal a difference in the two popuhtions of death goes down, the porccntagc for one or more that the age-specific mortality rates will reveal. For othercauscs of death must, ofncceasity, go up. When example, if a particular disease were occurring at a such a shift in the cause-specific mortality ratios higher-than-expected rate for a particular age group in occurs, it is not always clear which change is the cause one of the populations, this effect could be canccled by and which is the effect. Consequently, it is nccessary a lower-than-cxpectai rate in some other age group. In to utilize a more stable and mcaningful incasure for this situation, thc age-adjusted mortality rates could be eval,_m_f_ng time trends for the various causes of death. the same for the two popuhtions, and thc difference Annul or Crude Mortality Rate. Onc such measure is would become apparent only if the age-specific the annual mortality rato or cause-specific mortality rate mortality rates were cemparcd for each age group. which is expressed as the number of people dying from Interpretation of Observed Trends. Whcn canccr a particular cause per 100,000 population per year. mortality rates are properly adjusted for changes in Rather than being normalized to the total number of population age distributions over time, the resulting people who died in a particuhr year, the annual trends become significantly more meaningful in an mortality rate for each cause of death is normalizexi to cpidemiologic context. While the individual mortality the total population from which the deaths occurred. rates for some of the cancer sites are in a state of flux, This measure is often refcrrai to as the crude mortality most of the changes arc relatively small and have little rate to distinguish between this figure and an age- impact on the wtal cancer mortality, as some are adjusted mortality rate discussed below. However, increasing while others are decreasing. One of the annual morality rates also have somc weaknessca notable exceptinns is the time-trend for lung cancer which can lead to misinterpretations of the mo~ality which has undergone a major increase over the past 50 dam. years, placing this site on the top of thc list for cancer The U .S. popuhtion has bccn shiiting toward ,the oldcr mortality (Figure 3-5). age groups (Figures 3.20 and 3.21). As this happens, While the exact relationship betwecn individual EMF the crude mortality ratca for diseases which are exposure and the crude surrogate measure (total U.S. characteristic of the oldcr age groups will increase power consumption trends) has not bcen clearly (Figure 3-1), and it may appear, on supcrficial analysis, established, it is not unreasonable to assume that thcre that there is an emerging problem with thosc diseases should bc somc degree of positive correhtion betwccn (Figure 3-19). Thus, it is neccssary to control for *the two. If such exposures arc indcexi causing a differences or changes in age distribution whcn significant increase in a number of different types of comparing two different populations or when analyzing 3-20 Health Effects of Exposure to Powerline Frequency Electric and Magnetic Field~ cancer, then there also should bc some sign of 5. The principle of a zero tolerance for correlation between power consumption and cancer carcinogenic exposures should bc established mortality trends. In the preceding analysis, none of the for all areas of legislation, and only in cases long-~erm cancer mortality trends for the sites of where carcinogcnic contamination is concern show any patterns which would suggest an unavoidable should exceptions bc made to the association with EMF exposure when plotted together zero tolerance rule. with the surrogate measure, unlike the plot of cigarcUe These recommendations formed the basis for the consumption in pounds per capita together with lung FDA's "Delancy Clause' (which, in principle, bans cancer morality (Figure 3-6). While this form of analysis does not disprove such a possible association, every carcinogenic additive from the food supply) and aho gave me to the generally accepted philosophy that the inconsistency between the hypothesis and the chemicah should be presumed *guilty until proven obgr/Rinns suggest, that the epidemiologic evidence innocent.' In the face of contradictory data (a~ when should be somewhat more substantial before a causal one study reporu that a substance causes cancer while association can be reasonably assumed. another report* that it does not), regulatory scientists a.4.a CarctnoClsn POliCy arlel Rsgumtion maintain that *prudence* requires that the positive study should be accepted over the negative. This ba early as 1964, the *WHO Expert Committee on the policy has been extended so that all that is required to Prevention of Cancer,' chaired by Sir Richard Doll, establish the careinogenicity of a chemical is a concluded *that the rn~jo~ty of hutnan cancer is technic~lly inadequate study, and that such a study may potentially preventable' (WHO, 1964). In 1969, even outweigh a well done study which doe, not show Higgin~on estimated that 90 percent of cancen were a cancer effect. influenced by *environmental facton* and, thus, were preventable (1969). Subsequently, writers have quoted While such a system for interpreting carcinogenic Higginzon's 90-poreent figure and equ~_t_ed the term bioa~aays is not warranted on a strictly scientific b~is, *environmental factors* with *man-made chemicals* or u it is clearly bhued toward making Type I errors 'environmental pollution.' It is essential to note that (i.e., taking regulatory action when none is needed), it Higgin~on treed the word *environmental* in a much is aho cle,rly preferable to a system that fails to take broeder serae to indic, te all external faetora including any action until there i~ clear proof that a substance is tohaeco and alcohol usage; dietary, social, and cultural having a detrimental effect on human health. However, habit,; reproductive and infectinu~ dim:ag history; full implementation of the 'guilty until proven radiation exposures; occupational exposures; and innocent* philosophy would be tantamount to ignoring exposure to either man-made or natural caroinogens in the scientific bt~i~ for regulation, and it would render the air, water, or food. meaningless what i~ now, and hu been, a prudent and effective system for protecting public health. In 1970, the Ad Hoc Committee on the Evaltation of Furthermore, it would seriously damage the credibility Low Level-, of Environmental Carcinogens made a of regulatory scientist, if every chemical that had not number of recommendations to the Surgeon General been tested for carcinogenicity were suddenly declared that have since become aimoat axiomatic among most to be a suspected carcinogen and, therefore, were to be scientizt, and regulators when addressing the issue of *banned* from food and drinking water supplies. In environmental carcinogens: recent yean, there has been growing recognition of the 1. Any substance that has been shown to cause limitations of the rodent lifetime careinogenisis cancer in animals should be considered biotssay (Lave et al., 1988). Even if all the testing carcinogenic and, therefore, a probable or could be done and alI the positive substances identified, possible cancer risk for humans; there would still be a tretnendous technological gap which would make it virtually impossible to insure that 2. No level of exposure to a known chemical all of the atapect substance~ were permanently careinogen should be con-,idered eliminated from the food and water supplies. Unless toxicologically insignificant for humans (i.e., alcohol, tohaceo, air pollution, and foods containing no no-effect threshold can be **aumed); natural carcinogena were also banned or eliminated, it 3. No chemical substance should be assumed iz highly unlikely that the removal from the food and water supply of all trace carcinogens that have resulted rife for human consumption without proper negative lifetime carcinogenic bioassays; from humans and their technology would have any appreciable effect on cancer incidence or mortality 4. In caroinogenic bioasuys, negative results rate,. should be auporaeded by positive findings, and positive result, should remain definitive unless new evidence conclusively proves prior result, were in error; and Epidemiology of Healzh Effects and Exposure to EMF 3-21 3.4.4 Carcinogenic Risk Assessment and exposure limits designed to protect worker health under Risk Management the Occupational Safety and Health Act). Risk managers at feeloral regulatory agencies are Although regulatory statutes differ markedly, they seeking to achieve multiple objectives which include: rarely specify how agencies are to perform risk assessments of human exposures to potentially harmful · protection of public health fwm widespread agents. Risk assessment practices have therefore exposure to toxic agents; evolved at federal agencies and become routine through · protection of highly exposed and/or sensitive formal and informal agency guidelines. While statutory groups from toxic agents on the basis of mandates and risk assessment practices differ among equity or fairness, even when exposure is not agencies, the various regulatory cultures share a common policy viewpoint: namely, that risk managers wide~pr~td; should err onthe side of safety when making regulatory · protection of the natural environment and decisions in the face of scientific uncertainty. ecosystems from the adverse effects of toxic agents on behaff of both current and future Risk assessment is an analytical tool rather than an end in itself. Since the tool is used for a wide variety of generations; risk management decisions, it is critical that risk · responsiveness to public concerns about assessors and managers forge a constructive human health and environmental risks, even collaboration. Risk assessments should address the when risk assessors are skeptical about the needs of risk managers in a rigorous, objective, and magnitude of the risks posed by toxic agents; timely fashion. Some assessments need to be more and rofmexi than others, depending on the importance of the decision. · economic efficiency achieved by adopting protective regulations when the marginal For chemicals that are known or suspected to cause social benefits of risk management exceed the cancer, federal agencies have adopted standard risk marginal social costs of risk management (at assessment procedures. When data are insufficient to least where an agency's legislative mandate complete a full risk assessment, specific "default" does not protu~it consideration of assumptions are consider~l appropriate . The default economics). assumptions are designed to err on the side of safety in the absence of scientific knowledge. In centrut to Over the years, an adversarial relationship between the Executive Office of Management and Budget (OMB) more defined risk assessments, these standard assessments do not require extensive case-specific data. and the federal regulatory agencies has developed in part because it is usually impossible to simultaneously Despite their scientific limitations, standard risk achieve all of these objectives. The parties in this assessments of caroinogens are useful screening tools adversarlal relationship appear to assign differing (i.e., they help identify potential problems and indicate degre~ of importance to the achievement of the exposures that are not worthy of further concern). various rogulatory objectives. Given the different They may also provide a basis for rogulatory decisions objectives, it should be expected that OMB and the in cases where risks are potentially significant and the agencies might have different ideas about what is a estimated costs of risk reduction are too small to justify good risk assessment and what is a good risk a more refined risk assessment. management decision. Congress and federal agencies are now considering Federal roles are adopted on the basis of legislative new approaches to risk assessment. The Committee on mandates which provide risk managers with varying Risk Assessment Methodology of the National degrees of discretion. When discretion is restricted, it Academy of Sciences (NAS) is looking into ways to may reflect congressional determination to achieve improve the risk assessment process. Another NAS specific reguhttov/objectives. Some statutes (e.g., the C0mmitte¢ on risk assessment is being formed in Toxic Substances Control Act) authorize risk managers resix>nso to the malldates of the Clean Air Act to weigh the risks, costs, and benefits of alternative Amendments of 1990. The President' s Science courses of action even though such factors are not Advisor, Dr. D. Allan Bromley, has launched an always considered. Other statutes (e.g., the Delaney interagency committee to explore improvements in risk Clause covering careinogenic food additives and the assessment practice, while several federal agencies National Ambient Air Quality Standards under the have internal groups working toward the same goal. Clean Air Act) compel the federal government to base The EPA is considering revision of its risk assessment decisions exclusively on health considerations. Still guidelines, and the EPA Science Advisory Board other statutes order risk managers to reduce human recently expressed its intent to offer comments aimed at health risk to the maximum extent that is technically accelerating EPA's review of the guidelines. Hearings and economically feasible (e.g., the permissible 3-22 Health Effects of F~posure to Powerline Frequency Electric and Magnetic Fields on risk assessment were recently held by a exposures (e.g., ionizing radiation, solvents, PCB's), subcommittee of the U.S. House of Representatives, which may confound the association between EMF and which is another indication of the growing interest in cancer, have either been neglected or not considered reform of the risk assessment practices. fully. In addition, the healthy worker effect seen in studies o[ occupational groups when compared to the 3.5 Conclusions general population is much more pervasivc than Several studies in the United States and abroad have differences in exposures at work. It may also imply differences in sociocconomic status and access to examined the relationship between cancer and exposure medical care, thus resulting in earlier and more to EMF. There is still considerable disagreement accurate diagnosis of disease. In residential studies, regarding which outcomes may be related to exposure the incorporation of EMF exposures from other of EMF. Lcukemh, brain cancer, and skin mclanoma are reported in several studies to be associated with sources, both inside and outside the home, must be exposure to EMF. An unidentified common considered, as well as other potential confounders such denominator among these three sites of cancer is as exposure to indoor air pollutants. suspected, but none has been identified. However, 2- In order to be able to identify a specific disuse conclusions reached from the available epidemiologic agent or disease mechanism, there is a need to improve evidcnco arc limited. However, findings of the studies the assessment of EMF exposure for electrical workers of lcukcmla remain suggestive because, in soyoral of and for commui3ity psidents as well. The assessment the cpidcmiologic studies, an increased occurrence of of exposure to EMF is the most important issue in leukcmia is observed among workers and reskients future research of this association. Because of the presumed to be exposed to EMF. diversity of the cancon and other adverse health effects Cancer of the CNS is the second site given reported to date, if EMF is truly causing any of these considerable attention as possibly associated with EMF effects, it is reasonable to expect that several cxposuru. Findings from studies of CNS cancer also mechanisms may be involved. This may bc especially remain inconclusive. Only half of the epiderniologic true for the childhood cancers which would involve studies reviewed report an excessive mortality of CNS rehtively short exposure times and much shorter cancer among workers potentially exposed to EMF. latehey periods compared to adult cancers resulting However, an exposure-response relationship is from lengthy occupational exposures. observed most consistently only for the studies of CNS 3. Assessment of EMF exposure should not be limited cancer. to the use of industrial codes and iob titles ~n Studies have suggested other adverse effects of EMF occupatiorud studies end wirin~ confilzurations in pside~t~l studies. The exposure to EMF includes including other cancer sites, other mechanisms (e.g., transmission and distribution lines, video display paternal exposure), reproductive effects, and neurologicnl effects. To date, these findings also terminals, various electrical appliances, and remain suggestive, but not conclusive of a role for background sources. Estimation of overall exposure is EMF exposures in the development or progression of critical when a number of different sources are considered for the same person. Since considerable disease. variability is observed in magnetic field exposure There is anccd for further studies to resolve current within a single job category depending on different inconsistencies in available research results. To tasks performed, it will be important to estimate the contribute to the resolution of these research questions, average as well as other characteristics of exposure future studies will need to use more refined methods of over an 8-hour shift, and to give some consideration to assessing EMF exposures, to addreu the role of nonoccupational exposures. potentially confounding exposure~ as well as sources of In residential studies, it is important to determine bias, and to fastidionsly select refcrenco populations. whether wiring configurations, field measurements or 3.6 Recommendations soma other methods provide the better basis for estimating historical exposures. In occupational The following statements represent some suggestions stodiu, thejob-exposure matrix approach may prove to for future studies. In many of the epidcmiologic be a useful approach. However, it remains to be studies reviewed, adverse health effects have been determined whether the job-exposure matrix approach reported for electrical workers and children with or personal measurements will provide the better paternal or residential exposure to EMF. method for estimating historical EMF exposures. If the 1. While it is clear that something is responsible for a job-exposure matrix approach is preferred, it will be reported difference in cancer risk in the available necessary to develop high-quality data for use in such studies, it is premature to conclude that EMF is the an approach. factor involved. In occupational studies, other F, pidemiology of Health Effects and Exposure to EMF 3-2~ 4. Improved methods to measure exposure to electric openton in Europe may be an excellent choice of a and magnetic fields arc nce, ded~ some of which arc study population since they are exposed to constant under development. Future epidemiologic studies levels of EMF over the work shift. should adoiX quantitative measures of exposure. The 9. Outcomes due to pre- and post-natal exposures questions to be addressed in accurately assessing EMF should be examined separately. Children arc cxposurc arcthcsources of cxposureand the aspects of considcrcd to bc more vulnerable than adults to exposure (frequency, intensity, inXcrmim:ncy, diurnal potential carclnogcnic effects of EMF, and in uttro timing, resonance factors, etc.). Detailed information exposures are thought to bc more serious than post- on the history of BMF cxposure n-,~d--s to be obtained. natal exposures. Idcntificalion of appropriate iatcncics should bc punuai. The importance of induction time and thc ]0. The assoclatlnn of cancer with cxposure to EMF importance of duration of exposure also should bc must bc examined in li~.ht of current knowled2c from usessod. previous epidemiologic studies conducted in human populations and from exporirnental studies involvin~ 5. It is important to consider unrecognized laborato animals. Cancer increases postulated from confounders and sources of bias in existing and future EMF exposure are presumed to be dose-dependent and epidernioloeic studies, even though few confounders cumulative. There is a need to delineate the biological are presently known for the cancer sites reported to be basis for hypothesized health effects in humans, and to associated with EMF exposure. Results from many establish whether a dose-response relationship exists. studies have been severely criticized because of the lack of control for confounding and bias. 11. The results from epidemiologic studies should provide auidanc¢ for the design of future laboratory 6. Psrticu_ls_r efforts should be made to avoid possible investi2ations. To fully assess the question of syst___~_tic errors or misclassification in future studies. causality, mcchanlstic investigations in the laboratory For instance, while the use of waterbed heaters or electric pads my appear to be independent of electric are necessary. blanks use (the one tending to preclude the other), in 12. It will be 3-5 years before the results of the studies reality they would bo negatively correlated. Asking currently underway are available. In the meantime, only about use of electric blankets might result in a except for studies to pursue specific questions that have systcmslic error. In addition, abstractlag histories of been proposed by previous studies, a large undireeted EMF exposure from death certificates is a possible expansion of epidemiologic research of EMF is not source of differential misclassification if the person indicated. Several large cpidemiologic studies of the coding the occupational information is also aware of the association of EMF with cancer and adverse cause of death and the research question. Similarly, reproductive outcomes are curnntiy underway in the inquiries regarding a history of EMF exposure in United States, Canada, and in other paris of the world. interview studies my result in a recall bias. Such Studies in progress include investigations of childhood systematic errors could result in overesthnating or cancer, specifically ieukcm la and CNS tumors, and underestimating the association of exposure to EMF studies of occupationally exposed populations. with disease. However, overestimation is much more Although these most recent studies are still hampered likely. by the lack of knowledge regarding the relevant EMF exposure components that may account for the EMF- 1. The use of more than one reference cohort of at- cancer association, they have been designed to address !e-_s~-equai size to that of the exposed cohort is the major limitations of the earlier studies. The results suezested. Some previous studies have provided of these studies will, no doubt, provide valuable conclusions which arc of questionable validity because information regarding the nature of any associations insufficient attention was paid to comparable unexposed between EMF and cancer or EMF and adverse populations. The use of multiple reference cohorts would facilitate an assessment of the homogeneity of reproductive outcomes. risk and the tenability of a pr/or/judgments (lViieuinen et sl., 1990). All the rcfcnmcc cohorts should be comparable to the exposed cohort in terms of the a prior/judgments regarding the relevant issues, except EMF exposure. Comparison of exposed workers with several reference cohorts of unexposed workers may also be helpful to overcome the bias which results from the healthy worker effect. g. For future studies it will bc important to define a clearly exposed population and to quantitatc the EMF exposure in this population. Electrical transportation 3-24 Health Effects of E;cposure to Powerline Frequency Electric and Magnetic Fields EpidemioloAO/ of HeaRh Effects and Exposure to EMF 3-25 346 Health Effects of E. xposure to Powerline Frequency Electric and Magneac Fields Epidemiology of Health Effects and Exposure to EMF 3-27 3-28 Health Effects of Exposure to Powerline Frequency Electric and Magnedc Fields Epidemiology of Heallh Effects and Exposure to EMF 3-29 3-30 Heallh Effects of Exposure to PowerFine Frequency Electric and Magnetic Fields TABLE 3.4. LEUKEMIA AND OCCUPATIONAL EMF EXPOSURE (Continued) Investigator Location Population Method Used for Type of Number of Cases Results (Including SMR's, SIR's, PMR's, PIR's, Studied Exposure Estimate Study or Study Subjects OR's, or RR's, and 95% crs and/or p-values) Olin el al., Sweden Mortality among Graduated with a Cohort 24 cancer deaths, Graduates with M.S. in Electrical Engineering: 1985 1,254 electrical master of science 108 total death. s Leukemia: SMR=0.9 (N=2, O. i-3.2) engineering graduates degree in electrical from the Royal engineering Institute of Technology, 1930-79 Peatee et New Zealand 546 male leukemia Occupation coded CC 546 leukemia All electrical occupations: al., 1985 cases, age>20 yrs, as any one of 8 cases, 2184 OR,= 1.70 (N=i8, 0.97-2.97, p<O.06) registered with the electrical occupation conlrols with Electronic equipment assemblers: New Zealand Cancer groups other types of OR=8.17 (N=4, 1.49-44.7, p<O.02) Registry, ! 979-83 cancer Radio/television repairmen: OR=4.75 (N=7, 1.59-14.2, p<O.OI) Remaining jobs not significantly associated with leukemia Coggon et Cleveland. 2.942 Male residents Occupation and CC 29 acute myeloid Electrical and electronic workers: al.. 1986b Humberside. of study area. ages industry obtained leukemia cases, AML: N=5 (RR not reported) and Cheshire 18-54, diagnosed from postal 2,942 total cancer (including the with cancer, 1975-80 questionnaire cases used as Wirral), controls Flodin et Sweden AM L cases among Occupation coded CC 59 AML cases All electrical workers: al., 1986 residents, age 20-70 as any one of 3 354 controls AML: OR=3.8 (N=8, 1.5-9.5) yrs, dangnosed in five electrical occupation medical clinics in groups, controlling Sweden, 1977-82 for gamma radiation exposure from all sources Stem et al., New Leukemia mortality Detailed shop and CC 53 leukemia Electricians: 1986 Hampshire among 24,545 white job history deaths Total !eukemia: OR=3.00 (N= I I, !.29-6.98) male on-shore controlling for 212 controls Myeloid leukemia: OR=2.33 (N=6, 0.77-7.06) workers at the radiation and Lymphatic leukemia: OR=6.00 (N=5, !.47-24.5) portsmouth Naval solvent exposure Working in electrical shop: Shipyard, 1952-77 Total leukemia: OR=2.57 (N = 10, I. 1 !-5.96) Myeloid leukemia: OR=2.12 (N=5, 0.64-7.10) Lymphatic leukemia: OR = 3.80 (N = 5, !. 13 - 12.8) Welders or working in welding shop: Total leukemia: OR=2.25 (N=7, 0.92-5.53) Myeloid leukemia: OR=3.g3 (N=6, 1.28-11.5) TABLE 3.4. LEUKEMIA AND OCCUPATIONAL EMF EXPOSURE (Continued) and Date of Study Studied Exposure Estimate Study or Study Subjects OR's, or RR's, and 95~, Cl's and/or p-values) ' Tgrnqvist et Sweden Cancer incidence Job title codes 'Cohort 236 cases among Power linesmen: al., 1986 among 3,358 power ~ower linesmen, !.~ukemia: SMR= 1.25 (N= 10, 0.60-2.30) linesmen and 6,703 463 cases among Power station operators: - ~ower station station operators Leukemia: SMR=0.96 (N= 16, 0.55-1.57) operators, 1961-79 Lin, 1987 Taiwan Cancer mortality Employment with Cohort 733 total deaths Electric power company workers: (Abxtr) from 733 death Taiwan Electric including 180 total Leukemia/lymphoma: ce~ificates of male Power Company eancef deaths SMR=2.00 (!.i7-3.21, p<0.OI) employees, ages 64, of Taiwan Electric Power Company, 1971-85 TSrnqvist et Sweden Cancer incidence in Job titles Cohort 325 leukemia & Electrical engineers &technicians: al., 1987 147,134 persons in 401 CNS cancer Total leukemia: SIR= 1.3 (I.0-1.8) (Absrr) 14 occupational CLL: SIR= Miners and rockblasters: groups AML: SIR=2.1 (!.0-1.8) [sic] Juutilainen et Finland Finnish Cancer Occupation codes Cohort 17 cases of All electrical occupations combined: al., 1988 Registry and Central leukemia Total leukemia: SIR= ! .23 (N= 17, 0.72-1.97) Statistical Office of Linemen and cable jointers: Finland Total leukemia: SIR=:}. 13 (N=4, 0.85-8.00) Linet et al., Sweden 5,351 leukemia cases Occupation codes Cohort Swedish Cancer- ANL: SiR- I. ! (N =42. NS} Environment CML: SIR= I. 1 (N =27, NS) Registry, i 96 ! -79 Electric line workers: CLL: SIR=I.9 (N=I3, p<O.05) Mining and quaflying industry (not specific for EMF exposure): CLL: SIR= 1.3 (N =27, NS) ANL: SIR= 1.2 (N = 19, NS) CML: SIR=I.I (N=9, NS) Miners, quarr),men, and related workers: (not specific for EMF exposure): ALL: SIR = 1.0 (N = I, NS) CLL: SIR= 1.4 (N=21, NS) ANL: SIR=I.5 (N~-17, NS) CML: SIR= 1.4 (N =9, NS) TABLE 3.4. LEUKEMIA AND OCCUPATIONAL EMF EXPOSURE (Continued) ~~::=::~l~: Investigator Population Method Used for Type of Number of Cases Results (Including SMR's, SIR's, PMR's, PIR's, and Date of Study Studied Exposure Estimate Study or Study Subjects OR's, or RR's, and 95 · C!'s and/or p-values) Milham, Washington 67,829 licensed License class as Coho~ 2,485 total deaths Amateur radio operators: 1988 and amateur radio an amateur radio Total leukemia: SMR=I.24 (0.87-1.72, NS) Myeloid Leukemia: SMR= 1.40 (0.g3-2.20, NS) California operators, 1979-84 operator AML: SMR=i.76 (1.03-2.85, p<0.OS) CML: SMR=0.86 (0.17-2.50, NS) Obrams, U.S. 222 leukemia deaths lob titles Nested i26 male Leukemia deaths: 1988 among employees of CC leukemia deaths Cable splicer: OR=0.95 (0.47-2.04, NS) AT&T, employed 2 96 female Outside line tech: OR=0.67 (0.36-1.27, NS) yrs or more, 1975-80 leukemia deaths Installation: OR = i .04 (0.66-1.60, NS) Central office: OR= i.20 (0.71-1.85, NS) preston- Los Angeles 137 patients, age 20- Prior employment CC i 37 chronic Prior employment as a welder: Martin and County, CA 69, with chronic as a welder myeloid leukemia CML: Adjusted* OR=25.4 (N= 19, 2.78-232.5) Peters, 1988 myeloid leukemis cases and diagnosed 1979-85 137 age-, race-, * Adjusted for other variables in conditional Iogistic and sex-matched regression analysis (Other variables in model include: controls tad dose to bone marrow from diagnostic radiography, ever lived on a farm, and self or first degree relative with Down's syndrome or thalassemia minor). Gub~ran et Switzerland ! ,916 painters and Job titles Cohort 78 cancer cases, Cancer incidence d_n_~: al., 1989 1,948 electricians 52 cancer deaths Electricians: Leukemia: SIR.= 1.25 (N=2, 0. 154.52) with residence in Cancer mortality data: Canton of Geneva, Electricians: 1971-84 Leukemia: SMR=I.43 (N=2, 0.17-5.16) Loomis and 16 States, Deaths from Usual occupation CC 1,694 leukemia All electrical workers: Savitz, 1989 U.S. leukemias among and industry deaths (including Total leukemia: OR=0.9 (0.6-1.3, NS) (.4bstr) males with known recorded on death 474 AML deaths) AML: OR=0.9 (0.5-1.8, NS) occupation, 1985 certificate controls from all Electrical and electronic technicians: other causes of Total leukemia: OR = 1.9 death Electricians: Total leukemia: OR= 1.8 TABLE 3.4. LEUKEMIA AND OCCUPATIONAL EMF EXPOSURE (Continued) Investigator Location I Population Method Used for Type of Number of Cases I Results (Including SMR's, SIR's, PMR's, PIR's. and Date of Study Studied Exposure Estimate Study or Study Subjects OR's, or RR's, ned 95 % Crs ned/or p-values) Pearce et al.. New Zealned All male cancer Job title CC 534 leukemia Electrical engineers: 1989 patients, age>20 cases Total leukemia: OR= 1.54 (0.38-6.27, NS) years. registered in 19,904 total Radio/television repair: New Zealand Cancer cancer cases as Total leukemia: OR=7.86 (2.20-28.1) Registry, 1980-84' controls Power station operators: Total leukemia: OR=3.89 (I.00-15.2) Linemen: Total leukemia: OR=2.35 (0.97-5.70, NS) Total electric workers: Total leukemia: OR= 1.62 (I.04-2.52) Chronic leukemia: OR =2.12 (I. 19-3.76) Acute leukemia: OR= 1.25 (0.62-2.54, NS) Gallagher et British Leukemia deaths Occupation PMR 65 deaths from Age 20 yr and older: al., 1990 Columbia, among male residents leukemia Electrical engineers: (Letter) Canada of British Columbia, Total leukemia: PMR = 1.70 (N = 8, 0.73-3.35) 1950-84 Electrical & electronic assemblers & repairmen: Total leukemia: PMR=2.08 (N=5, 0.67-4.86) Metal mill workers (not specific for EMF): Total leukemia: PMR= 1.60 (N= 12, 0.82-2.80) Garland et Sne Diego, Leukemia cases in Occupation HP 102 cases of Total leukemia (compared with SEER population): al., 1990 CA white male US Navy Cohort leukemia Electricine's mate: SIR=2.4 (N=7, 1.0-5.0) personnel, recorded Sonar technician: SIR= 1.6 (N=3, 0.3-4.8) at the Naval Health Electronics teclmicine: SIR--I. ! (N=5, 0.4-2.5) Research Center, San Radioman: SIR= I.I (N=4, 0.3-2.7) ABBREVIATIONS USED: ACO - Astrocytoma CNS - Central Nervous System PIR - Propoflionate Incidence Ratio ALL - Acute Lymphocytic Leukemia EMF - Electric ned/or Magnetic Fields PMR - Proportionate Mortality Ratio AML - Acute Myelogenous Leukemia GBO - Glioblastoma PRR - Proportional Registration Ratio ANL - Acute Nonlymphocytic Leukemia HP - Historical Prospevtive Study RF - Radio Frequency CC - Case-Control Study MW - Microwave RR - Risk Ratio or Relative Risk CI - Confidence Interval NS - Not Significant SIR - Standardized Incidence Ratio CLL - Chronic Lymphoid Leukemia OR - Odds Ratio SMR - Standardized Mortality Ratio CML - Chronic Myelogenous Leukemia Rev. 4/9/92 TABLE 3.5. MAJOR EPIDEM1OLOG1C STUDIES OF CANCER INCIDENCE/MORTALITY BRAIN/CNS CANCER AND OCCUPATIONAL EMF EXPOSURE Investigator Location Population Method Used for Type of Number of Cases Results (Including SMR's, SIR's, PMR's, PIR's, and Date ] of Study I Studied I Exposure Estimate Study or Study Subjects OR's, or RR's, and 95% Cl's and/or p-values) Linet al., Maryland 951 brain tumor Usual occupation SMR/ 951 deaths from Electrician: 1985 deaths in adult white and industry CC total brain tumors GBM & AST: SMR=2.28 (N= 13, !.21-3.90) male Maryland recorded on death 519 glioblastoma Other brain: SMR=I.70 (N=8, 0.73-3.35) residents, age:> 20 cerlificate multiforme and ~.lectric or electronic engineer &technician: years, 1969-82 astrocyloma GBM & AST: SMR=2.50 (N= 18, !.48-3.95) (GBM & AST) Other brain: SMR=2.07 (N= 12, 1.07-3.61) deaths Utility (electric & telephone) lineman/serviceman: GBM & AST: SMR=3.73 (N= 19, 2.24-5.82) Other brain: SMR= 1.90 (N=8, 0.82-3.75) All electrical occupations: GBM & AST: SMR=2,78 (N=50, 2.06-3.66) Other brain: SMR= 1.90 (N=28, 1.27-2.75) Glioblastoma and astrocytoma risk with occupations grouped by categories of estimated EMF exposure: Definite: OR=2.15 (N=27, 1.10-4.06) probable: OR= 1.95 (N=21, 0.94-3.91) Possible: OR=1.44 (N=128, 1.06-1.95) Milham, Washington 12,714 deaths among Death certificate PMR 2,649 total cancer All electrical occupations: 1985b State male residents, occupation coded deaths (out of Brain cancer: PMR= !.23 (N= 101, p <0.05) age>20, who to any one of 9 12,714 total worked in 9 electrical electrical deaths) occupations, 1950-82 occupations Olin ctal., Sweden Mortality among Graduated with a Cohort 24 cancer deaths, Graduates with M.S. in Electrical Engineering: 1985 i ,254 electrical master of science 108 total deaths Brain cancer: SMR= 1.0 (N =2, O. !-3.7) engineering graduates degree in from the Royal electrical Institute of engineering Technology, 1930-79 V/tger6 et al., Sweden Cancer morbidity Employment at Cohort 102 male cancer Telecommunications workers: 1985 among 2,9 i 8 workers work sites cases and Nervous system: at three work sites involved in 37 female cases Males: SMR= !.0 (N=5, 0.3-2.3) manufacturing research, telecommunications development, and equipment, 1958-79 manufacturing of telecommunication equipment TABLE 3,5. BRAIN/CNS CANCER AND OCCUPATIONAL EMF EXPOSURE (Continued) and Date of Study Studied Exposure Estimate Study or Study.Subjects OR's, or RR's, and 9S% Cl's and/or p-values) Coggon et Cleveland, 2,942 Male residents Occupation and CC 97 Brain/CNS Electrical engineering industry: al., 1986b Humbcrside, of study area, ages industry obtained cancer cases, Brain/CNS: RR= !.9 (N=3, NS) and Cheshire 18-54, diagnosed from postal 2,942 total cancer Electrical and electronic workers: (including the with cancer, 1975-80 questionnaire cases used as Brain/CNS: RR=2.0 (N=7, NS) Wirral), controls U.K. T6rnqvist et Sweden Cancer incidence Job title codes Cohort 236 cases among Power linesmen: al., 1986 among 3,358 power power linesmen, Nervous system: SMR= 1.49 (N= 13, 0.80-2.56) linesmen and 6,703 463 cases among Power station operators: power station station operators Nervous system: SMR=0.97 (N= 17, 0.57-1.56) operators, 1961-79 Lin, 1987 Taiwan Cancer mortality Employment with Cohort 733 total deaths Electric power company workers: (.4bstr) from 733 death Taiwan Electric including 180 total Brain/CNS: SMR=4.10 (1.77-8.08, p<O.OI) certificates of male Power Company cancer deaths employees, ages 25- 64, of Taiwan Electric Power Company, 1971-85 Magnani et Cleveland, 1,265 male residents Occupation and CC 1,265 total deaths Electrical engineering: al., 1987 Humberside, of study areas, ages Industry codes due to Brain/CNS Brain/CNS: RR=O.9 (0.2-4.3) Cheshire, .e, 18-54, who died of plus job-exposure cancer & 4 other Electrical &electronic workers: the Wirral, any I of 5 cancers matrix sites; 4,470 age, Brain/CNS: RR= 1.3 (0.7-2.5) U.K. under study, 1959-63 sex, and residence and 1965-79 matched controls : McLaughlin Sweden 3,394 incident cases Occupation and Cohort 3,394 incident Electrical occupations or industries: et al., 1987 of intracranial glioma industry codes cases of Electricians and electronic workers: registered with the intracraniai SIR--0.9 (N:75, 0.71-1.13, NS) Swedish Cancer- gliomas Welders and metal cutters: Environment SIR= 1.4 (N=46, 1.02-1.87, p<O.05) Registry, 1961-79 Machinery and electronics industry: SIR-- I.I (N=313, 0.98-1.23, p<O.05) Other electronics: SIR= I.2 (N=86, 0.96-1.48, NS) Non-electrical occupations with significant elevations: Biologists: SIR:2.3 (N= I3, 1.23-3.93, p<O.05) Medical professionals: SIR=I.7 (N=26, !.11-2.49, p<O.05) Potters, kilnmen, and glass workers: SIR= 1.6 (N=25, 1.04-2.37, p<:O.05) TABLE 3.5. BRAIN/CNS CANCER AND OCCUPATIONAL EMF EXPOSURE (Continued) Investigator Location Population Method Used for Type of Number of Cases Results (Including SMR's, SIR's, PMR's, PIR's, and Date I of Study I Studied I Exposure Estimate Study or Study Subjects OR's, or RR's, and 95% Cl's and/or p-values) Thomas el Northern 435 death certificates Occupation and CC 435 deaths from All male electrical & electronic workers al., 1987 New Jersey; listing brain or other industry codes. brain tumors. ever exposed to MW/RF radiation: Philadelphia, CNS tumors as cause grouped by likely 386 controls who RR= !.6 (N--69, 1.0-2.4). PA; and of death in white exposure to died from other Design, manufacture, repair, or installation southern males, age > 30 MW/RF radiation causes of electrical or electronic equipment: RR=2.3 (N=44, 1.3-.4.2) Louisiana years, 1978-81 Engineers, teachers, technicians, repairers, and assemblers with no MW/RF exposure: RR=3.9 (N--28, 1.6-9.9) Milham, Washington 67,829 licensed. Amateur radio Cohort 2,485 total deaths Amateur radio operators: 1988 and amateur radio operators Brsin/CNS: SMR= 1.39 (0.93-2.00, NS) California operators, 1979-84 Speers et al., East Texas 202 deaths from Usual occupation CC 202 CNS tumor Electricians and electrorues workers: 1988 glioma among male and industry of deaths (gliomas) All combined: OR=2. I i (0.77-5.81, NS) residents of 40 employment as 238 controls Grouped by estimated EMF exposure category: counties in east recorded on death None: OR= 1.00 (N =92) Texas, 1969-78 certificate Possible: OR= !. 15 (N=68, 0.73-1.81) Probable: OR=2.86 (N= I !, 0.8-10.3) Definite: OR= oo (N=6, control=O) Occupation/industry not specific for EM F exposure: Utility workers: OR= 13. I (!.33-129.0) Trucking industry: OR=6.65 (!.05-42.2) Gub,~ran et Switzerland 1,916 painters and Job titles Cohoa 78 cancer cases, Cancer incidence data: al., 1989 1,948 electricians 52 cancer deaths Electricians: with residence in Brain: SIR=I.18 (N=2, 0.14-4.25) Canton of Geneva, Painters (not specific for EMF exposure): 1971-84 Brain: SIR= 1.43 (N =3, 0.29-4.17) Cancer mortality data: Electricians: Brain: SMR= ! .54 (N=2, O. 19-5.56) Painters (not specific for EMF exposure): Brain: SMR=O.53 (N= i, 0.01-2.93) Loomis and 16 States, Deaths from Usual occupation CC 1,095 brain All electrical workers: Savitz, 1989 U.S. malignant brain and industry cancer deaths Brain: OR= 1.5 ( ! .0-2. I) (Abstr) tumors or leukemias recorded on death Electrical and electronic technicians: among males with certi ticate Brain: OR = 3. I known occupation, Electric power repairers and installers: 1985 Brain: OR=2.4 TABLE 3,5, BRAIN/CNS CANCER AND OCCUPATIONAL EMF EXPOSURE (Continued) Investigator ] Location ] Population Method Used for Type of ] Number of Cases Results (Including SMR's, SIR's, PMR's, PIR's, and Date of Study Studied Exposure Estimate Study or Study Subjects OR's, or RR's, and 95% Cl's and/or p-values) ~carce el al., New Zealand All male cancer Job title CC 431 brain cancer Electrical engineers: OR=4.74 (I.65-13.6) 1989 patients, age2:20 cases, Electricians: OR--- 1.91 (0.844.33, NS) years, registered in 19,904 total Total electric workers: OR,- 1.01 (0.56-1.82, NS) New Zealand Cancer cancer cases as Registry, 1980-84 controls Preston- Los Angeles 272 Brain/CNS Job titles and CC with 202 glioma cases, All electrical occupations (employed) 5 ye.,rs): Martin et al., County, CA cancer cases among occupations matched 70 meningioma Gliomas: OR= 1.8 (0.74.8) 1989 black &white males, involving high pain cases Meningiomas: OR--0.7 (0. I-5.8) ages 25-69, 1980-84 likelyhood of Astrocytomas: OR=4.3 (I.2-15.6) exposure to EMF Risk for gliomas increased with increase in number of years worked: (p for trend =- 0.008). Reif et al., New Zealand 452 Brain/CNS Current or most CC 452 brain/CNS Electrical workers: 1989 cancer cases among recent occupation cancer cases &: Brain: OR=0.78 (N,~8, 0.39-1.59) males, age>20, in st the time of 19,452 cases of Occupational groups not specific for EMF exposure: New Zealand Cancer registration other cancers used Plumbers and welders: Registry, 1980-84 as controls Brain: OR=2.02 (N--8, 0.99-4.12) Sheet and structural metal workers Brain: OR-~2.10 (N--II, 1.15-3.82) Livestock farmers: Brain: OR=2.59 (N~-II, 1.41-4.75) Lewis, 1990 Coastal Incident cases of Occupational CC 375 cases of No increased risk by occupation: Texas and neuroglial tumors, history obtained neuroglial tumors Ever employed in an electrical occupation: Southem white and non-white by interview of the CNS OR~0.65 (0.40-1.09) Louisiana males, ages 20-79, (confirmed Usual occupation listed as electrical: disgnosed 1980-84 pathologically), OR=0.76 (0.33-1.73) : 450 controls No increased risk by estimated exposure categories: High: OR=O.80 (N~-37, 0.40-1.64) Medium: OR:0.88 (N=3, 0.06-21.8) Low: OR=0.45 (N=20, 0.19-1.03) ABBREVIATIONS USED: AST - Astrocytoma GBM - Glioblastoma Multiforme PMR - Proportionate Mortality Ratio CC - Case-Control Study MW - Microwave RF - Radio Frequency CI - Confidence Interval NS - Not Significant RR - Risk Ratio or Relative Risk CNS - Central Nervous System OR - Odds Ratio SIR - Standardiz~ Incidence Ratio EMF - Electric and/or Magnetic Fields PIR - Proportionate Incidence Ratio SMR - Standardiz~l Mortality Ratio R~v. 4/9/92 TABLE 3.6. MAJOR EPIDEMIOLOG1C STUDIES OF CANCER INCIDENCE/MORTALITY OTHER SITES (Eye, Breast, Testis, Skin Melanoma, etc.) AND OCCUPATIONAL EMF EXPOSURE Investigator Location Population Method Used for Type of Number of Cases Results (Including SMR's, SIR's, PMR's, PIR's, Studied Exposure Estimate Study or Study Subjects OR's, or RR's, and 95 % crs and/or p-values) ana t,a,~ I ~'' ~'"'~Y I ...... I ""' I a ' Howe & Canada Cancer mortality Occupation and Cohort 4,203 tolal cancer Linemen and servicemen (telephone, Lindsay, among 415,201 males industry codes (by deaths telegraph. and power): 1983 of the Canadian both 2-digit and 3- Intestine (except rectum): SMR=3.53 (N=8, i.52-6.96, p<O.OI) Labor Force, 1965- digit codes) Stomach cancer: 73 SMR=2.33 (N--6, 0.86-5.07, p<O.05) Miners, quarrymen, and related workers: Lung: SMR=-I.O0 (N=31, NS) Stomach: SMR= 1.07 (N= I0, NS) intestine: ' SMR=O.75 (N=6. NS) Pancreas: SMR= 1.04 (N =6, NS) Prostate: SMR=O.28 (N=I, NS) Bladder: SMR=0.39 (N=I, NS) Brain: SMR--I.I5 (N=4, NS) Swerdlow, England and All cases of eye Occupational PRR 2,159 cases of eye Electrical and electronic workers: 1983 Wales cancer in persons order listed as cancer in males, Eye cancer in males: age> 15, registered electrical and 2,125 cases in 1971:PRR=I.67 (N=I, NS) in any of the 14 electronics worker females 1972:PRR=7.14 (N=5, p<0.Ol) population-based 1973:PRR=2.63 (N=5, NS) 1974:PRR=4.44 (N=4, p<0.05) regional cancer registries, 1962-77 1975:PRR=5.41 (N=4, p<O.05) Professional. technical workers, and artists (not specific for EMF exposure): Eye cancer in males: 1971:PRR=2.39 (N=i6, p<0.OI) 1972:PRR=4.00 (N=6, p<0.OI) 1973:PRR=2.29 (N=II, p<O.05) 1974:PRR=2.18 (N=!2, p<O.05) 1975:PRR=2.70 (N=!O, p<O.OI) VAger6 and Sweden Cancer incidence Employment in HP/ !,855 male cancer All electronics workers: Olin, 1983 among 54,624 male the electronics Cohort cases and Males: and 18,478 female industry 1,009 female Lung: RR= 1.52 (N=273. !.35-1.72) workers. ages 15-64, cancer cam Larynx: RR= 1.46 (N=36, 1.05-2.03) Bladder: RR= 1.22 (N= 122, 1.02-1.47) working in the electronics industry Colon: RR= 1.20 (N= 138, 1.02-1.43) in Sweden, at time of Skin melanoma: RR= 1.35 (N =59, 1.05-1.76) 1960 census Females: Cervix uteri: RR=i.14 (N=417, 1.04-1.26) TABLE 3.6. OTHER SITES (Eye, Breast, Testis, Skin Melanoma, etc.) AND OCCUPATIONAL EMF EXPOSURE (Continued) Investigator Location Population Method Used for Type of Number of Cases Results (Including SMR's, SIR's, PMR's, PIR's, and Date I of Study I Studied I Exposure Estimate Study or Study Subjects OR's, or RR's, and 95 % C!'s and/or p-values) I I I BJair et al., U.S. Cancer mortality among Usual occupation as Cohort 107,563 total Electricians: 1985 293,958 U.S. Veterans, reported by subject deaths from all Stomach: SMR=I.22 (N=9, 0.56-2.31) age 31-84, who held in questionnaires causes Intestines: SMR=0.74 (N= I I, 0.37-1.33) U.S. government life- administered in Rectum: SMR=O.75 (N=4, 0.20-1.90) insurance policies in 1954 and 1957 Lung & Bron: SMR= 1.18 (N=36, 0.83-1.64) December 1953 and of Prostatei SMR--0.87 (N= I I, 0.44-1.56) whom 107,563 were Bladder: SMR= 1.40 (N=7, 0.56-2.88) deceased as of Jan I, Lymphoma: SMR=0.87 (N=7, 0.35-1.80) 1970 Electrical engineers: Stomach: SMR=O.63 (N=7, 0.25-1.29) Intestines: SMR=I.06 (N=26, 0.99-1.55) Rectum: SMR=0.80 (N=7, 0.32-1.66) Lung & Bron: SMR=O.68 (N=27, 0.45-0.99) Prostate: SMR=0.96 (N=20, 0.58-1.48) Bladder: SMR= 1.53 (N= 12, 0.78-2.65) Lymphoma: SMR= 1.29 (N = 16, 0.74-2. I0) Telegraph, telephone, & powerline servicemen: Intestines: SMR=I.39 (N=I2, 0.72-2.44) Lung & Bron: SMR=I.29 (N=23, 0.82-1.94) Prostate: SMR=I.07 (N= 8, 0.46-2.10) Gallagher et Westem All cases of occular Job and industry CC 65 cases of Electrical and electronic workers: al., 1985 Canada melanoma in persons titles from detailed occular melanorna Occular melanoma: No increased risk age 20-79 registered in occupational history & 65 age- and Government workers (primarily indoor managerial) cancer registries of sex-matched (Not specific for EMF exposure): British Columbia, controls Occular melanoma: OR=3.5 (N=43, p=O.OO6) Alberta, Saskatchewan, and Manitoba, 1979-81 Milham, Washington 12,714 deaths among Death certificate PMR 2,649 total cancer All electrical occupations: 1985b State male residents, occupation coded to deaths (out of Pancreas: PMR= i. 17 (N= 174, p < 0.05) age>20, who worked any one of 9 12,714 total Lung, trachea, in 9 electrical electrical deaths) and bronchus: PMR= !. 14 (N=789, p<O.OI) occupations, 1950-82 occupations Lymphoma: PMR= 1.64 (N=51, p<O.OI) Olin et al., Sweden Mortality among 1,254 Graduated with a Cohort 24 cancer deaths, Graduates with M.S. in Electrical Engineenng: 1985 electrical engineers from master of science 108 total deaths Intestines: SMR=0.8 (N=3, 0.2-2.5) the Royal Institute of degree in electrical Pancreas: SMR=0.9 (N=3, 0.2-2.5) Technology, 1930-79 engineering Lung: SMR=0.4 (N=4, O. I-I. I) Melanoma: SMR=3.2 (N=3, 0.7-9.4) TABLE 3.6. OTHER SITES (Eye, Breast, Testis, Skin Melanoma, etc.) AND OCCUPATIONAL EMF EXPOSURE (Continued) Investigator Location Population Method Used for Type of Number of Cases Results (Including SMR's, SIR's, PMR's, PIR's, and Date [ of Sludy ] Studied I Exposure Estimate Study or Study Subjects OR's, or RR's, and 95% C!'s and/or p-values) VAger6 et Sweden Cancer morbidity Employment at Cohort 102 male cancer Telecommunications workers: al., 1985 among 2,918 workers work sites cases and Lung (males): SMR=0.9 (N= I0, 0.4-1.6) at three work sites involved in 37 female cases Colon (males): SMR= 1.5 (N= !0, 0.7-2.?) nmlufacturing research, Kidney (males): SMR= 1.9 (N= I0, 0.9-3.5) telecommunications development, and Prostate (males): SMR= i .2 (N= 15, 0.7-1.9) equipment, 1958-79 nmnufacturing of Melanoma (males): SMR=2.S (N=8, !.!-4.9) telecommunication Breast (females): SMR=0.6 (N=?, 0.34.3) equipment Melanoma (females): SMR=2.g (N=4, 0.8-7.2) Melanoma (total): SMR=2.6 (N= 12, 1.3-4.5} Coggon et Cleveland, 2,942 Male residents Occupation and CC 198 stomach, Electrical engineering industry: al., 1986a Humberside, of study area, ages industry obtained 176 colon, Stomach: RR=0.9 (N=4, NS) and 1986b and Cheshire 18-54, diagnosed from postal 167 rectum, Colon: RR= !.2 (N=6, NS) (including the with cancer, 1975-80 questionnaire 738 bronchus, Rectum: RR=0.7 (N=4, NS) Wirral), 64 reelsnores, Lung &Bronchus: RR=0.6 (N=IO, NS) U.K. 97 lymphores Melanoma: RR= 1.3 (N=2, NS) cases with 2,942 Lymphores: RR= 1.4 (N=4, NS) total cancer cases Electrical and electronic workers: used as controls Colon: RR= 1.2 (N=8, NS) Rectum: RR=0.9 (N=6, NS) Lung & Bronchus: RR=0.8 (N= 16, NS) Mclanoma: RR=0.7 (N=2, NS) Lymphores: RR= 1.7 (N =7, NS) T6mqvist et Sweden Cancer incidence Job title codes Cohort 236 cases among Power linesmen: al., 1986 among 3,358 power power linesmen, Stomach: SMR= 1.20 (N=25, 0.77-1.77) linesmen and 6,703 463 cases among Lung/trachea: SMR=0.66 (N= 17, 0.38-1.05) power station station operators Kidney: SMR= 1.33 (N = operators, 1961-79 Power station operators: Stomach: SMR=O.95 (N=42, 0.69-1.29) Lung/trachea: SMR=0.71 (N=40, 0.50-0.96) Kidney: SMR= 1.26 (N =29, 0.84-1.81) Lin, 1987 Tatwan Cancer mortality Employment with Cohort 733 total deaths Electric power company workers: (Abstr) from 733 death Tatwan Electric including 180 total Liver: SMR= 1.54 (I.204.95, p<O.O!) certificates of male Power Company cancer deaths employees, ages 64, of Tatwan Electric Power Company, 1971-85 TABLE 3.6. OTHER SITES (Eye, Breast, Testis, Skin Melanoma, etc.) AND OCCUPATIONAL EMF EXPOSURE (Continued) and Date of Study Studied Exposure Estimate Study or Study Subjects OR's, or RR's, and 95% Cl's and/or p-values) Magnani et Cleveland. 1,265 male Occupation and CC ! ,265 total deaths Electrical engineering: al., 1987 Humberside, residents of study Industry codes due to kidney, Pancreas: RR=0.8 (0.2-3.9) Cheshire. & areas, ages 18-54, plusjob-exposure esophageal, Kidney: RR=I5.2 (N=4, 1.7-136.0) the Wirral, who died of any I matrix pancreas, Electrical & electronic workers: U.K. of 5 cancers under melanoma, & Esophagus: RR= 1.4 (0.6-3.4) study, 1959-63 and CNS cancer; Pancreas: RR=0.6 (0.2-1.6) 1965-79 4,470 age-, sex-, Melanoma: RR= 1.2 (0.2-5.9) and residence- Kidney: RR= 1.7 (0.7-4.6) matched controls De Gui re et Montreal, 9,590 workers at Job titles and Coboff IO male cases Telecommunication workers: al., 1988 Canada telecommunications person-years of of malignant Malignant melanoma of the skin: company, 1976-83 observation melanoma of skin SIR=2.70 (N = tO, ! .3 i-5.02) Milham, Washington 67,829 licensed License class as Cohofl 2,485 total deaths Amateur radio operators: 1988 and amateur radio an amateur radio Stomach: SMR= 1.02 (0.68-1.45, NS) California operators, 1979-84 operator Large intestine: SMR= I. ! I (0.89-1.37, NS) Liver: SMR=O.65 (0.33-1.17, NS) Respiratory: SMR~-0.66 (0.58-0.76, p<O.05) Vrosta,e: SMR=I.14 (0.90-1.42, Kidney: SMR=O.94 (0.57-1.48, NS) Gub,~ran et Switzerland !,916 Painters and Job titles Cohofl 78 cancer cases, Cancer incidence data: al., 1989 i ,948 electricians 52 cancer deaths Electricians: wi,h residence in Lung: SIR=0.96 (N-- 16.0.55-1.57 Canton of Geneva, Bladder: SIR=0.67 (N--3, 0.14-1.95) 1971-84 Painters (not specific for EMF exposure): Lung: SIR= 1.47 (N=40, i.05-2.00) Bladder: SIR= l.71 (N= 13, 0.91-2.93) Pearce et al., New Zealand All male cancer Job title CC 19,904 total Total electrical workers: 1989 patients with known cancer cases Stomach: OR= 1.06 (N=25, 0.70-1.59) occupation, age>20 Colon: OR=- !.06 (N=52, 0.79-1.41) years, registered in Rectum: OR=l.16 (N=38, 0.83-1.62) New Zealand Liver: OR= 1.52 (N=7, 0.71-3.25) Cancer Registry, Pancreas: OR-=.O.91 (N= 12, 0.51-1.62) 1980-g4 Lung: OR=0.88 (N--89, 0.69-1. I I) Melanoma: OR=O.70 (N=24, 0.46-1.08) Prostate: OR=0.96 (N--52, 0.71-1.29) Testis: OR~--0.78 (N= 12, 0.41-1.47) Bladder: OR= I.O0 (H~-35, 0.67-1.50) TABLE 3.6. OTHER SITES (Eye, Breast, Testis, Skin Melanoma, etc.) AND OCCUPATIONAL EMF EXPOSURE (Continued) Investigator [ Location Population Method Used for Type of Number of Cases Resulls (Including SMR's, SIR's, PMR's, PIR's, or Study Subjects OR's, or RR's, and 95% Cl's and/or p-values) ana ua~e ua ~u,~y I Studied I Exposure Estimate I Study I Damera et U.S. 227 Males with breast Occupation CC 227 incident cases Any expesed job: al., 1990 cancer from I0 of breast cancer in Breast cancer (males): OR= 1.8 (N=33, i .0-3.7) (Abstr) and population-based males Electricians, telephone linemen, electric power 1991 SEER registries, 1983- 300 controls from workers, power plant operators: 87 medicare list Breast cancer (males): OR=6.0 (N= 13, i .7-21.5) Radio broadcast and communication workers: Breast cancer (males): OR=2.9 (N=7, 0.8-10.2) Tynes and Norway 37,953 Male workers, Job title Cohort 12 cases of male All electrical occupations: Andersen, age>20 years, with breast cancer from Breast cancer (males): SIR=2.07 (N= 12, 1.07-3.61) 1990 824,321 person-years Cancer Registry possible EM F of Norway exposure, 1961-85 Verreault et 13 Counties White males, ages 20- History of electric CC 182 cases of Electric blanket ever used in prior I0 years: al., 1990 of Westem 69 yrs, with testicular blanket use testicular cancer, Tcsticular cancer: RR= !.0 Washington cancer diagnosed 658 controls from Serethomas: RR=0.7 (N=27, 0.5-1.2) 1981-84 same geographic Nonseminomas: RR= 1.4 (N=30, 0.9-2.3) areas Risk by cumulative electric blanket us~ (months): 0 months: RR= 1.0 (N = 125) 1-24 months: RR=O.9 (N=30, 0.5-1.3) 25-120 months: RR= ! .2 (N =27, 0.7-1.9) Matanoski et New York 50.582 Telephone Job titles _ Cohort 2 Cases of breast Central office technicians (Average daily magnetic field al., 1991 state workers in one state- cancer in males exposure = 2.5 m(3): SIR=6.5 (N=2, 0.79-23.5) wide company, 1976- from 206,067 Cable splicen (Average daily magnetic field person-years of exposure = 4.3 raG): SIR=O (N=0) 80 observation Vena et al., Western Postmcnopausal Electric blanket CC 382 cases of Electric blanket used to warm bed only: 1991 New York women, ages 41-85, use female breast OR=0.64 (N=28, 0.39-1.05, NS) State 1987-89 cancer and Used to warm bed and sometimes throughout night: 439 controls OR=0.64 (N=30, 0.40-1.05, NS) Used continuously throughout night, not every night: OR=I.31 (N=68, 0.88-1.95, NS) Used daily in season through the night for I0 years: OR=1.25 (N=32, 0.73-2.16, NS) ABBREVIATIO 4S USED: CC - Case-Control Study NS - Not Significant RF - Radio Frequency CI - Confidence Interval OR - Odds Ratio RR - Risk Ratio or Relative Risk EMF - Electric and/or Magnetic Fields PIR - Propoflionate Incidence Ratio S1R - Standardized Incidence Ratio HP - Historical Prospevtive Study PMR - propoflionate Mortality Ratio SMR - Standardized Mortality Ratio MW - Microwave PRR - Propoflional Registration Ratio Roy. 4/9/92 TABLE 3.7. MAJOR EPIDEMIOLOGIC STUDIES OF CANCER OR ADVERSE REPRODUCTIVE OUTCOMES CHILDHOOD ADVERSE EFFECTS AND PATERNAL/MATERNAL EMF EXPOSURE Investigator ] Location Population Method Used for Type of Number of Cases Results (Including SMR's, SIR's, PMR's, PIR's, and Date of Study Studied Exposure Estimate Study or Study Subjects OR's, or RR's, and 95% crs and/or p-values) Hemminki et Finland Women in metal Maternal occupation Cohort 195 spontaneous Electrical workers of Union of Metal Workers: al., 1980 industry abortions Spontaneous Abortions (ate per 100 births): compared with All Finnish women: 10.3 Electrical workers: 16.5, p< O.O01 all Finnish Before joining Union: I0. I women After joining Union: 14.6, p<O.0OI Nordstrom et Sweden 542 Employe~s Paternal occupation Cohort 26 complicated Paternal exposure to high voltage: al., 1983 at Swedish realformations Increased frequency of abnormal pregnancy power plants outcome (p<O.O01). Spitz & Texas Childhood Palereal occupation CC 157 deaths from Electrical workers (father): Johnson, deaths due to neuroblastoma in OR~2.13 (N~ 17, i.05-4.35) 1985 neuroblastoma children Electronics workers (father): OR= I !.8 (N~6, 1.40-98.6) Wertheimer Colorado 1,256 families Usage of eleclric Cohort 130 retarded Proportion of births with above-median gestation & Leeper, giving birth in blankets and heated growth, 5 significantly greater in exposed group. Five congenital 1986 1982 waterbeds congenital defect, defects reported in 193 users vs one in 335 non-users. and 24 abortions Abortion rates were 6.3-7.8% in users and 4.2% in non-users. Nordstrom et Sweden Employees at Paternal occupation Cohort 482 employees An increased frequency of congenital real formations al., 1987 Swedish power responded to was found (p<O.OOi) among offspring of fathers who stations questionnaire worked at high voltage substations. providing data for 26 mallfunctions among 866 pregnancies. Nasca et al., New York Incident cases of Paternal occupation CC 338 cases with Definition for paternal EMF Exposure: 1988 State primary CNS primary CNS tumor Narrow: OR= 1.70 (N= IS, 0.80-3.59, NS) tumors in Broad: OR--I.61 (N=I9, 0.83-3.11, NS) children Wilkins & Childhood Palernal occupation CC 491 dealhs from Electrical assembling, installing, and repairing: Koutras, dealhs due to at time of childs birth brain cancer OR=2.7 (N= 19, 1.2-6.12) 1988 brain cancer TABLE 3.7. CHILDHOOD ADVERSE EFFECTS AND PATERNAL/MATERNAL EMF EXPOSURE (Continued) Investigator Location Population Method Used for Type of of Cases Results (Including SMR's, SIR's, PMR's, PIR's, Studied ExpOsure Estimate Study or Study Subjects OR's, or RR's, and 95 % Ci's and/or p-values) Buiatti et al. Florence, Italy Residents in Paternal occupation CC 112 cases Radioelectric workers: , Oligospermia and azoospermia: 1989 attending clinic OR=5.89 (N=5, 0.86-40.2, NS) Johnson, Texas Childhood Paternal industry and CC 499 deaths from Paternal industry with potential EMF ~xposure: 1985; deaths due to occupation CNS turnors in OR= 1.64 (N-=-25, 0.96-2.82, NS) Johnson & CNS tumors children Electronics manufacturing: OR=3.56 (N=7, 1.04-12.2) Spitz, 1989 Paternal occupation with potential EMF exposure: OR=i.44 (N=28, 0.88-2.38, NS) Electricians: OR=3.52 (N=7, !.02-12.1) Wertheimer Oregon Oregon slate Ceiling cable eleclric Cohofl 380 ilboflions of Increased felal loss rate during October-January season & Leeper, birth records heating in homes which 143 wero when EMF exposure is presumed to increase. 1989 for 1983 & Irated (37.4%), 142 cases used in 1985 analysis Walkins & Ohio Childhood Paternal industry and CC 101 cases of Occupational EMF exposure as deftned by: Hundley, neuroblastoma occupation neuroblastoma Hsieh el al.: OR= I. I (N =24, 0.5-2.3, NS) Deal~n & 1990 incident cilses Henderson: OR= 1.6 (N=4, 0.3-9. I. NS) Lin el al.: Narrow: OR= 1.9 (N=6, 0.4-9.7, Broad: OR=0.7 (N=19.0.3-1.5, NS) Bunin et ill., Pennsylvania Childhood Paternal occupalional CC 104 cases of Paternal EMF exposure prior to conception: 1990 neuroblastoma hislories obtained by neuroblasloma Workers: OR= 1.6 (0.5-6.2, NS) incident cases telephone interview Assemblers: OR--4.0 (0.4-19.2, NS) Paternal EMF exposur~ during pregnanc~t: Workers: OR=0.4 (0.1-1.6, NS) ABBREVIATIONS USED: CC - Case-Conlrol Study OR - Odds Ratio RR - Risk Ratio or Relative Risk CNS - Cenlral Nervous System PIR - Propoflionate Incidence Ratio SIR - Standardized Incidence Ratio EMF - Electric and/or Msgnetic Fields PMR - Propo~ionate Mortality Ratio SMR - Standardized Mortality Riltio NS - Hol Significant Rev. 4/9/92 3-46 Health Effects of FJcposure to Powerline Frequency Electric and Magnetic Fields Crude Cancer Mortality Crude Mortlily Rlt~' ~r 100.000 220 180 ' 160 140 120 ' 100 30 35 40 45 50 55 60 65 70 75 80 85 Year ~ilum 3-1 -Cmdc Caaccr Mortalit), Rates, 1930-87 Maldl%malcrrotal, Adults and Children. Total Cancer Mortality 200 14~ 1:~ 100 ' Maim 'Femalee mawTotals 30 35 40 45 50 55 60 65 70 75 80 85 Year 3-~ - Total Ca~cc~ Mo~=llW, l~lcs, 19)0-87 MddFcmalcri'ml, Adults =nd Children. Epidemiology of Health Effects and Exposure to EMF 3-47 Total Cancer Mortality Male and Female Adults 320 ' Femm~ ~ .... 280 240 ~ 200 160 ........................................................ 30 35 40 45 50 55 60 65 70 75 80 85 Year Figure 3-3 - Total Cancer Morudity, 1930-87, Maic and Female Adults. Total Cancer Mortality Male and Female Children 6 4 30 35 40 45 50 55 60 65 70 75 80 85 Year · Adjureled to US 1970 IDOf~b'tm Figure 3-4. - Toh, I Cancer Mortalhy, M,,lc and Fcm~lc Children. 3-48 Health Effects of E~posure to Powerline Frequency Electric and Magnetic Fields Lung Cancer Mortality Age-Adjueed Reee' pet ~00. CO0 Po~ 120 ' Melee ~ Females eeeTcMls 100 30 35 40 45 50 55 80 65 70 75 80 85 Year · Adiuead m U.$. t070 pe~b~. Figure 3-5 - Lung Cancer Mortality, 1930-87, Male/Female/Total Adults and Children. Cigarette ConsumpUon v, Lung Cancer PeetdNC~ Deee, d100K 1~ 10 50 8 ...................... 40 6 ................... 30 4 ..................................... 20 2 .............................. 10 25 30 35 40 45 50 55 60 65 70 75 80 85 90 Year Figtire 3-6 - Cigarette Consumption v. Lung Cancer. Epidem~ology of Health Effects and E~posure to EMF 3-4.9 Total Cancer Mortality ' MelEe ' ~0 ........................................................ 30 35 ~ 45 ~ ~ ~ 65 70 75 ~ 85 Y~ 3-7 - T~I C~¢r Mo~li~ ~inu~ ~n~), 193~87, M~Ic~IeR~I Adulu ~ ~ild~n. Age-Adjueted Rele~' per 100,000 Pop. 10 8 6 4 2 ' Maims ' Females roTorals 0 ........................... 30 35 40 45 50 5~ 60 65 70 75 80 85 Year · k:ljum,d to US 1970 pe0ula~mrt Figurc 3-8 - Lcukcmia Mortality, 1930-87, Malcs/Fcmalc~Total Adults and Children. 3-50 Health Effects of Exposure to Powerline Frequency Electric and Magnetic Fields Leukemia Mortality Male and Female Adults AO,-nlmeNRm'~,~100.000~. 14 10 ............. 6 4 : 2 - L' Main Yeu Figure 3-9 - Leukemia Mortality, 1930-87, Male ud Female AdulU C20-+85 yes). Leukemia Mortality Male and Female Children Age~Reeee'lert00,0001~ 4 i 3 2 1 30 35 40 45 50 55 60 65 70 75 80 85 Year · Adjueeed to US. ~970 ix~d4.~e. ~i~ut~ 3-10 - L~uk~m~e Mortality Rates, 1930-87, Male and Fcma|¢ Children (0-19 yrs), Epidtmiology of Health Effects and Exposure to EMF 3-51 Brain & CNS Cancer Mortality 4 ' 3 2 I ' Is ' Females mTotel~ 0 ........... 30 35 40 45 50 55 60 65 70 75 80 85 Year A~ju. Ed to LI$1910 fe~aeUet. 3-! 1 - Brain and CNS Caacer Mortslity, 1930-8'7, M,,ledFemalesrl'otsl Adults aad Children. Brain & CNS Cancer Mortality Male and Female Adults Aee,a~)uelel Retee' W loo,ooo ~ 6 -- 4 ' Year i. Adjueled to US. 1970 populeae~. F~urc 3-].2 - Brain and CNS Cancer Morta|hy, ~930-~7, lVle|¢ and Fcma|¢ Adults (20-8~+ ~)- 3-52 Health Effects of F, vposure to Powerline Frequency Electric and Magnetic Fields Brain & CNS Cancer Mortality Male and Female Children ~ge-Muma R,m- ~r loo,olxl 2.0 ' 1.0 0~ i,,,,,,,,,,,,,, .............. , .... ,,,~,,,~,, ..... 30 35 40 45 50 55 60 65 70 75 80 85 Year Filu~ 3-13 - Br~in and CNS Cancer Mortality, 193G-87, M~le and Female Children. Breast Cancer Mortality /~.-A61ueedAIteVOelOO.OOOPa~ 100 100 10 10 1 '1 0.1 ,,,, .... ,, ....... ,,,,,,,,,.,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 0.1 · A~ju~d~oUS. lgT0pop,Mao. ~ure 3-14 - Br~st Cancer MorTality, t930-87, Male and Fcmsie Adults (20-85+ Epidemiology of Health Effects and Exposure to EMF 3-53 Power Consumption v. Cancer Mortality Male Adults 15 2500 12 2 -t- Lm - 2000 6 ' lOOO 3 --- ' 500 0 , ........ 0 30 35 40 45 50 55 60 65 70 75 80 85 5-Year Period Figure 3-15 - Power Consumption v. Cancer Mortality, 1930-87, Male Adults (20-85+ yrs). Power Consumption v. Cancer Mortality Adult Female 50 2500 40 - - 2(00 30 - '1500 20 - 1000 10 - 500 0 0 30 35 40 45 50 55 60 65 70 75 80 85 5-Year Pedod Figure 3-16 - Power Consumption v. Cancer Mortality, 1930-87, Female Adults (20-85 + yrs). 3-54 Health Effects of F. xposurt to Powerant Frequency Electric and Magnetic Fields Power Consumption v. Cancer Mortality Male Children Oem~e/lOOK Pop 8ilo. KW-Hm 10 2500 8 - 2000 8 - ~.vC~ - 154:X} =l=Leeemle 4 ' ' 1000 30 35 40 45 50 55 60 65 70 75 80 85 5-Year Figure 3-17 - Power Consumption v. Cancer Motlality, 1930-87, Male Children (0-19 yrs). Power Consumption v. Cancer Mortality Female Children 7 2500 6 · 2000 5 1500 4 3 ,_ 1000 2 - - 500 1 - 0 30 35 40 45 50 55 60 65 70 75 80 85 5-Year Pmiod Figure 3-18 - Power Consumption v. Cancer Mortality, 1930-87, Fcmalc Children (0-19 yrs). Epidemiology of Health Effects and Exposure to EMF 3-55 Leading Causes of Death in the U.S. Pe-cetl d Ill 100% 15% ' 25% 0 10 20 30 40 50 60 70 50 81 Year of Death · Ai F~)a/tmd oa de~h'~erti~lM (delth regi/'ttm elltel Ody lg00-30}. Figu~ 3-19 ' Leading Cau~s of l~lth in ~hc U.S. During Each Ducadc from 1900 to 1987. U.S. Life Expectancy at Birth ymafU~mr-wl~m~ed 80 70 40 30 0 10 20 30 40 50 60 70 80 90 Year of Birth 3-20 - U.S. Life E~c~y at Bi~, Males aM Fe~lea, 19~1990'. 3-56 Health Effects of Exposure to Powerant Frtqutncy Electric and Magnttic F~tlds U.S. Population Porcontago Distributions lawcrania, o~ Total Popumuon m 10 - 19 ' .-'.. ' ~ 70 Rm ~ 10 ~ ~ ~ ~ ~ 70 ~ ~ Y~r 3-21 - U.S. ~ula~oa ~bu~, 1~ - 1987 F_.pideraiology of Health Effects and Exposure to EMF 3-57 References Agran L. 1977. The Cancer Connection: And What We Can Do About It. New York: St. Martin's. Ahlbom A. 1988. A review of the cpidemiologic literature on magnetic fields and cancer. Scan J Work Environ Health. 14:337-43. Aldrich TE and Easterly CE. Electromagnetic fields and public health. Environ Health Perspect. 75:159-71. Annegets JF, Schocnberg BS, Okazaki H, and Kurland LT. 1981. Epidemiologic study of primary intracranial ncoplasms. Arch Neurol. 38:217-9. Assnova TP and Rakov AN. 1966. Health conditions of workers exposed to an electrical field of 400-500 kV in open distributing installations. Giglena truda iprofessional'nye zabolevan~ia. 5:50-2. (Cited in Hauf, 1982). BarregArd L, Jarvholm B, and Ungeth~m E. 1985. Cancer among workers exposed to strong static magnetic fields. Lancet. October 19, 2:892. Bcrta~,-zl PA, Pesatori AC, Badice L, and Zocchetti C. 1986. Exposure to formaldehyde and cancer mortality in a cohort of workers producing resins. Scan J Work Environ Health. 12:461-8. Blair A, Walrath J, and Rogot E. 1985. Mortality patterns among U.S. veterans by occupation. 1. Cancer. J Nat Cancer lnst. 75:1039--47. Brenner H, Savitz DA, JSckel K-H, and Grecnhnd S. 1992. Effects of nondifferential exposure misclassification in ecologic studies. Am J F. pidemiot 135:85-95. Brownson RC, Rcif JS, Chang JC, and Davis JR. 1990. An analysis of occupational risks for brain cancer. Am J Public Health. 80:169-172. Buiatti E, Barchielll A, Geddes M, Nastasi L, Kricbel D, Franchini M, and Scarstill G. 1984. Risk factors in male infertih'ty: a case-control study. Arch Environ Health. 39:266-70. Bunin GR, Ward E, Kntmcr S, Rhec CA, and Meadows AT. 1990. Neuroblastoma and parental occupation. Am J F.p/dem/ol. 131:776-80. Bureau of the Census. 1982. 1980. Census of popuhtion: classified index of industries and occupations. Washington, DC: U.S. Govt Print Off. Calie EE and Savitz DA. 1985. Leukemia in occupational groups with presumed exposure to electrical and magnetic fields. (Letter). New Engl J Med. 313:1476-7. Checkoway H, Pearce N, and Crawford-Brown D. 1989. Research Methods in Occupational Epidemiology. New York: Oxford University Press. Codd MB and Kurland LT. 1985. Descriptive cpidcminlogy of primary intracranial ncoplasms. Prog Exp Tumor Res. 29:1-11. Coggon D, Pannett B, Osmond C, and Acheson ED. 1986a. A survey of cancer and occupation in young and middle aged men. I. Cancers of the respiratory tract. BrilJ IndMed. 43:332-8. Coggon D, Panneli B, Osmond C, and Acheson ED. 1986b. A survey of cancer and occupation in young and middle aged men. II. Non-respiratory cancers. BritJlndMed. 43:381-6. Coleman M, Bell J, and Skeet R. 1983. Lcukemia incidence in electrical workers. Lancet. April 30, 1:982-3. Coleman M and Bcral V. 1988. A review of cpidcmiological studies of the health effects of living near or working with electricity generation and transmission equipment. Int J EpidemioL 17:1-13. Coleman MP, Bell CMJ, Taylor H-L, and Primic-Zakclj M. 1989. Leukcmia and residence near electricity transmission equipment: a case-control study. Bri~ J Cancer. 60:793-798. Corey CR and Freeman HE. 1990. Use of telephone interviewing in health care research. Health Scrv Res. 25:129-44. 3-58 Health Effects of Eacposure to Powerfine Frequency Electric and Magnetic Fields De Guire L, Therlault G, Imrra H, Provencher S, Cyr D, and Cue BW. 1988. Increased malignant melanoma of the skin in workers in a telecommunications industry. Brit J Ind Med. 45:824-8. Demers PA, Thomas DB, Rosenblare KA, Jimenez LM, McTierrmn A, Sealsberg H, Stemhagen A, Thompson WE), Curnen M, Satariano A, Austin DF, Isacson P, Greenberg R$, Key C, Kolonel L, and West D. 1990. Oeeupardonal exposure to electromagnetic radiation and breast cancer in males. Abstracts of Papers Presented .t the Twenty-Third Anntuff Meeting of the Society for Epidemiologic Research. Am J F4/dtm/oL 132:7754. Demers PA, Thonms DB, Rosenblan KA, limenez LM, McTiernan A, Sealsberg H, Stemhagen A, Thompson WE), Cumen MGM, Satariano W, Austin DF, Isaeson P, Greenberg RS, Key C, Kolonel LN, and West DW. 1991. Oeeupslionnl exposure to electromagnetic fields and breast cancer in men. Am I Epiderniol. 134:340-7. Doll R and Hill AB. 1950. Smoking and careinoma of the lung: preliminary report. BritMtd.l. 2:739-758. Doll R and POto R. 1976. Moaality in relation to smoking. Twenty yean' observations on male British doctors. Br/t Mtd,l. 2:1525-1527. Doll R. 1991. Progress against cancer: an epidemiologie assessment. Am J Ep/dem/o/. 134:675-88. Electric Power Research Institute. 1991. Procealings from Workshop on EMF Epidemiology, Camel, CA. February 2- 5. Filippov V. 1972. Der Ein~uss yon elektrisehen Wechselfeldern anf den Mensehen rrhe effect of alternating electfie fields on man]. In: 2. Int. Colloquium fiir die Verh~tung yon Aioeitsunillen und Berufskrankheiten dutch Elelarizitit, K61n, Berufsgenossenschait der Feinmechanik und Elektroteehnik, imp. 170-177. (Cited in Haul, 1982). Flegal ruM, Brownie C, and Hm JD. 1986. The effects of exposure sisclassification on estimates of relative risk. Am I EpidrmioL 123:'736-51. Flodin U, FrcdrHcason M, Axclson O, Porsson B, and Hardell L. 1986. Background radiation, electrical work, and ~omc othcr exposures associated with acutc mycloid lcukcmia in a caso-refcrcnt study. Arch Environ Health. 41:77-84. Flynn MR, West S, Kaunc VerY, SavRz DA, Chen C-C, and Loomis DP. 1990. Validation of cxpcrt judgment in asscmsing occupational cxlx)surg to magnctic ~clds in the utility industry. Appl Occup Environ Hyg. 6:141- 145. Friedman GD. 1980. Primer ofEpidtmiology. New York: McGraw-Hill Book Company. Fulton JP, Cobb S, Preblc L, Lennc L, and Forman E. 1980. Electrical wiring configurations and childhood Icukemia in Rhodc Island. AmJEpidera~oL 111:2926. Gallaghcr RP, Elwood JM, Rootman ~, SpinclH Jl, Hill GB, ThrelfaU WJ, and Birdsell JM. Risk factors for ocular mclanoma: western Canada mclanoma study. ~ Nat Cancer lnst. 74:775-8. Gallaghcr RP, McBridc ML, Band PR. SpineIll lJ, Threlfall W], and Yang P. 1990. Occupational electromagnctic ~cld exposure, solvent cxposure, and lcukcmia. (Lcllcr). ~ Occup Med. 32:64-5. Garland FC, Shaw E, Gorham ED, Garland CF, White MR, and Sinshcimer P]. 1990. Incidence of lcukcmia in occupations with potential electromagnetic ~cld exposure in United States Navy personnel. Am J Epidemiol. 132:293-303. Gilman PA, Ames RG, and McCawley MA. 1985. Lcukcmla risk among U.S. White male coal miners../Occup Med. 27:669-71. Grecnland S. 1987. Evolution of Epidemiologic Ideas. Epidcmiology Resources Inc., Boston. Gub~ran E, Uscl M, Raymond L, Tissot R, and Swectnam PM. 1989. Disability, mortality, and incidence of cancer among Geneva painters and electricians. Brit ~lnd Med. 445: 1623. Epidemiology of Health Effects and E;cposure to EMF 3-59 Hammond EC. 1966. Smoking in relation to the death rate of one million men and women. In Haenszel W. (cd.) Epidcmiological approaches to the study of cancer and other chronic diseases. Nat Cancer lnst Monogr. 19:127-35. Haul R. 1982. Electric and magnctic fields at power frequencies, with particular reference to 50 and 60 Hz. In: Sucss MJ, ed. NonionL:ing radiation protection. WHO Regional Publications, European Series No. 10: 175-97. Helseth A, Langnmrk F, and Merk SJ. 1988. Neoplasms of the central nervous system in Norway. I. Quality control of the registration in the Norwegian CAncer Registry. APMI5. 96: 1(X32-8. Hclscth A, Mark SJ, Johanscn A, and Tretli S. 1989. Ncoplasms of the central ncrvous system in Norway. IV. A popuiation-basccl cpidcmiological study of mcningiomu. APMIS. 97: 646-54. HemminkS K, Nicmi M-L, Koskincn K, and Vaiaio H. 1980. Spontaneous abortions among women employed in the metal industry in Finland. Ira Arch Occup Environ Health. 47:53-60. Hcnnckcns CH and Buring JE. 1987. Ep/dem/o/ogy/n Medicine. Boston: Little, Brown and Company. Higginson J. 1969. 'Present Trends in CAnoct Epidcmiology," in Proceedings of the Eighth Canadian Cancer Conf~g~ncc, Honey Harbour, Ontario, p. 40. Higginson J and Muir CS. 1979. Environmcntal carcinogcnisis: misconccptions and limitations to cancer control. J Nail Cancer lnst. 63:1291-8. Hill A. 1965. The Environmcnt and disease: Association on causation? Proc R $oc Med. 58:295. Hill A. 1971. Pr/nc/p/e$ of Medical $tatistics. New York: Oxford. University Press. Howe GR and Lindsay JP. 1983. A follow-up study of a tcn-porccnt samplc of thc Canadian Labor Force. 1. Cancer mortality in males. 1965-73. J Nail Cancer lnst. 70:3744 Hutchinson GB. 1991. CArcinogcnic Effect of Exposure to Electric Fields and Magnetic Ficids [prepared for workshop on Future Epidcmiologic Studies of Health Effects of Electric and Magnetic Fickis, February 6-8, 1991 Carmcl, California. Isscl I, et al. 1977. Tauglichkcits- und Eignungsuntcrsuchungcn an Elcktromontcuren -Erlautcrung an cincr neucn Untersuchungsanwcisung [The uscfuiness and suitability of cleotromcchanics - explanation of a new approach to examination], Deutsche GesundheiLvwesen. 32:1526-31. (Cited in Haul, 1982). Jauchcm JR and Mcrritt JH. 1991. Thc cpidcmiology of cxposurc to clcctromagnctic ~clds: an overview of thc recent literature. J Clin Epidemiol. 44:895-906. Johnson CC. 1985. An cpidcmiologic case-control analysis of childhood intracranial and spinal cord tumors in relation to paternal occupation at birth. Ph.D. Dissertation. University of Texas, School of Public Health. Johnson CC and Spitz MR. 1989. Childhood nervous system tumors: an assessment of risk associated with paternal occupations involving usc, repair or manufacture of electrical and electronic equipment. Ira J Epidemiol. 18:901-7. Juutilaincn J, Pukkala E, and Laara E. 1988. Rcsults of an cpidcmiological canccr study among electrical workers in Finland. l Bioelectricity. 7:119-21. Kaunc WT, Stcvens RG, Callshah N/, Severson RK, and Thomas DB. 1987. Residential magnetic and clcctric fields. Bioelectromagnetic$. 8:315-35. Kcam DW. 1988. W'trc coding configurations are poor surrogettcs for magnetic field exposures. Radiation Protection in Australia. 6:82-6. Kclscy JL, Thompson WE), and Evans AS. 1986. Methods in Observational Ep/dem/o/ogy. New York: Oxford University Press. pp. 47-53. Klcinbaum DG, Kuppcr LL, and Morgenstcrn H. 1982. Epidemiologic Research, Principle and Quantitative Methods. London: Lifetime Learning Publications. 3-60 Health Effects of F. xposure to Powerline Frequency Electric and Magnetic Fields Knave B, et aL 1978. Exposure to electric fields: An epidemiological health examination on long-term exposed switchyard worken. Arbete och halsa. (Cited in Haul, 1982). Kor~bkova VP, Morozov UA, Stolyarov MD, and Yakub UA. 1972. Influence of the Electric Field in 500 kV and 750 kV Switchyards on Maintermnce Staff and Means for Its Protection. Conference International des Grands R~eaux Electriques (CIGRE), Paris, (Report 2.3-06). (Cit~l in Haul 1982). Last MI. 1988. A Dictionary ofF_,pidtmiology. (2nd edition). New York: Oxford University Press. Lave LB, Ennever FK, Rosenknnz HS, and Omenn G-~. 1988. Information value of the rodent biossssy. Nature. 336:631-3. Lewis I. 1990. Employment in Electrical Occupations and the Risk of Neurological Tumors. Doctoral Dissertation. Houston, TX: UTSPH. Lin RS, Dischinger PC, Conde J, and Farnil KP. 1985. Occupational exposure to electromagnaic fields and the occurnnce of brain tumors: an analysis of possible associations. J Occup Med. 27:413-9. Lin R$. 1987. Mortality patterns gong employees of electric power company in Taiwan (Me~ting abstract). Biological effects from Electric and Magnetic Fields, Air Ions and Ion Currents Associated with High Voltage Tansmission Lines: Contnctors Review. Kansas City: Department of Energy and Electric Power Research Institute. Liner MS, Mslker HSR, McLaughlin JK, Weiner JA, Stone BJ, Blot WI, Ericsson JLE, and Fraumeni JF Jr. 1988. Leukemias and occupation in Sweden: a registry-based s. rmlysis. Amer J lnd Med. 14:319-30. London, SJ, Thorruts DC, Bowman JD, Sobel E, Cheng T-C, and Peten JM. 1991. Exposure to residential electric and magnetic fields and risk of childhood leukemia. Amer J Epidiol. 134:923-37. Loomis DP and Savitz DA. 1989. Brain cancer and leukemia mortality among electrical workers. (Abstract). Am J F_.io/dtm/oL 130:814. Magnsni C, Coggon D, Osmond C, and Acheson ED. 1987. Occupation and five cane, ca: a case-control study using death certifie~_t_es. Br J lnd Med. 44:769-76. MaUmoski GM, Breysse P, Elliott EA. 1989. Cancer incidence in N~v York telephone worken. Poster at DOFdEPRI contracton review meeting. (Portland, Oregon, November 15). Matanoski G, Breysse PN, and Elljoel EA. 1991. Electromagnetic field exposure and male breast cancer. Lancet. 337:737. Matanoski G, Ellion EA, Breysse PN, and Lynberg MC. 1991. Leukemia in telephone linemen. Amer J Epidiol. (In Press). Msusner JS and Kramer S. 1985. F. pidtmiology, An Introductory TexL Philadelphia: W.B. Saunders Company. McDowall ME. 1983. I.,~uk=mia mortality in electrical worken in England and Wales. Lancet. 1:246. McDowall ME. 1986. Mortality of persons nsident in the Vicinity of electricity transmission facilities. Brit J Cancer. 53:271-9. McLaughlin IK, Maiker HSR, Blot WI, Malker BK, Stone RI, Weiner JA, Ericsson ILE, and Fraumeni JF Jr. 1987. Occupational risks for intracnmial gliomas in Sweden. J Nat Cancer lust. 78:253-7. Miettinen OS and Rossleer CE. 1990. Man-made mineral fiben and lung cancer - Epidemiologic evidence regarding the causal hypothesis. ~can J Work Environ Health. 16:221-31. Milhsm Slr. 1982. Mortality from leukemia in workers exposed to electrical and magnetic fields. (Letter). New Engl J Med. 307:249. Milham S Jr. 1985a. Silent Keys: Leukemia mortality in amateur radio operators. Lancet. 1:812. Milham S Jr. 1985b. Mortality in workers exposed to electromagnetic fields. Environ Health Ptrsptct. 62:297-300. Epidemiology of Heaah Effects and Exposure to EMF 3-61 Mllham S Jr. 1988. Increased mortality in amateur radio openton due to lymphatic and hemaWl~ietic malignancies. Am J Epidemiol. 127:50-4. Mortson RR. 1990. Editorial commentary: Epidemiology and exposure to electwmagnetic fields. Am i Epidemiol. 13:774-5. Myers A, Cartwright RA, Bonnell JA, Mole CK, and Can'wright 5C. 1985. Overhead power lines and childhood cancer. Paper presented at the International Conference on l~lectric and Magnetic Fields in Medicine and Biology, London, 1985. (Abstract). Myen A, Clayden teD, Canwright RA, and Cartwright SC. 1990. Childhood cancer and overhead powerllnes: a case- control study. BrilJ Cancer. 62:1008-14. Nair I, Morgan MG, and Florig HK. 1989. Biological Effects of Power Frequency Electric and Magnetic Fields. Office of Technology Assessment (OTA), Congress of the United States. Hasca PC, Baptiste MS, MacCubbin PA, Metzger BB, CarlWn K, Gnenwald P, Annbrustmacher VW, Earle KM, and Waldman J. 1988. An epidemiologic case-control study of central nervous system tumors in children and paternal occupational exposures. Am ! Epidemiol. 128:1256-65. Nordstr6m S, Birke E, and Gustavsson L. 1983. Reproductive hazards among workers at high voltage substations. Bioelectromagnetics. 4:91-101. Nordstr6m S, Nordenson I, and Hansson-Mild K. 1987. Genetic and reproductive hazards in high-voltage substations. In: U .S. Department of Energy. Interaction of Biological Systems with Static and ELF Electric and Magnetic - fields: Twenty-third Hartford Life Sciences Symposium Richland. WA: Pacific Northwest Laboratory. pp. 487-93. NTIS-841041. Obrams G. 1988. Leukernia in Telephone Lineman. University of Michigan Dissertation Service. Office of Management and Budget. 1972. Standard Industrial Classification Manual, 1972. Washington. DC: U.S. Govt Print and Off. Olin R, Vtger6 D, and Ahlbom A. 1985. Mortality experience of electrical engineers. BritJlndMed. 42:211-2. Pearoe NE, $heppard RA, Howard JK, Fraser J, and Lilley BM. 1985. Leukemia in electrical workers in New Zeahnd. Lancet. 1:811-2. Pearce N, Reif J, and Fraser J. 1989. Case-control studies of cancer in New Zealand electrical workers. lnt J Epidemlol. 18:55-9. Paers IM. 1990. EMF exposures by job category, results and interpretations. 1990 EPRI Utility Seminar. Peters JM. 1990. EMF Exposures by Job Category Loreseated at 1990 EPRI Utility Seminar, New EMF Epidemiologic Results and their Implications, Austin, Tens, no publication available]. Peters JM. 1990. EMF Exposures by Job Category, Results and Interpretations. (Presented at 1990 EPRI Utility Seminar. October 16-19. Austin, Tens) Poole C and Trichopoulos D. 1991. Extremely low frequency electric and magnetic fields and cancer. Cancer Causes and Control (In Press). Preston-Martin S and Peters JM. 1988. Prior employment as a welder associated with the development of chronic myeloid leukemia. BritJ Cancer. 58:105-8. Preston-Martin S, Peters JM, Yu MC, Garabrant DH, and Bowman JD. 1988. Myelogenous leukemia and electric blanket use. Bioeltctromagnetics. 9:207-13. Preston-Martin S, Mack W. and Henderson BE. 1989. Risk factors for gliomas and meningiomas in males in Los Angeles County. Cancer Res. 49:613743. Reif JS, Peatee N, and Fraser J. 1989. Occupational risks for brain cancer: a New Zealand cancer registry-based study. J Occup Med. 31:863-7. 3-62 Health Effects of Exposure to Powerline Frequency Electric and Magnetic Fields Roberge PF. 1976. Elude de l'aat de sant~ des ~lectriciens d'entredAen px'e~s~ ~ l'entretien des pos~es 735 kV de l'Hydn}-Qua~c, Quebec, Hydro-Qu~bec. (Cited in Haul, 1982). Rothman KJ. 1986. Modern Epidemiology. Boston: Little, Brown and Company. pp- 23-34. Sack~tt DL. 1979. Bias in analytic research. J Chronic Dis. 32:51-63. Savitz DA and Callc EE. 1987. Lcukcmia and occupational exposure to electromagnetic fields: review of cpidcmiologic surveys. J Occup Med. 29:47-51. $avitz DA, Wacht~l H, Barnes FA, Iohn EM, and Tvrdik JG. 1988. Case-control study of childhood cancer and exposure to 60.Hz magnetic fields. Am I F. pidem/oL 128:21-38. Savitz DA and Fcingold L. 1989. Association of childhood cancer with residential traffic density. Scan l Work Environ Health. 15:360-3. Savitz DA, Pcaxcc NE, and Poolc C. 1989. Mcthodological issues in the cpidcmiology of clcctromagnctic fields and cancer. Epidemiol Rev. 11:59-78. Savitz DA, John EM, and Klcckncr RC. 1990. Magnetic field exposures from clcctric appliances and childhood cancer. Am ./Epidem/oL 131:763-73. Sazonova TE. 1967. Physiological and Hygienic Assessment of the Working Conditions at the Outdoor Switchyards 400 and 500 kV. In: i>rocccdings of the Institute of Labour Protection under the All-Union Central Council of Tragic Unions, No. 2. (Cited in Hauf, 1982). Schoenberg BS, Schoenbcrg DG, Christinc BW, and Gomcz MR. 1976. The epidemiology of primary intracranial ncoplasms of childhood, a population study. Mayo Clin Proc. 51,51-6. Schull V, rJ. 1991. Personal Communication, November 27. Severson RK, Stevcns RG, Kaunc WT, Thomas DB, Hcuscr L, Davis S, and Sever LE. 1988. Acute nonlymphocytic lcukcmia and ruidcntial cxposuro to power frequency magnetic fields. Am l Ep/dem/oL 128:1020. Silverberg E. 1990. Personal Communication. Amer/can Cancer ~ociely. Speers MA, Dobbins JG, and Miller VS. 1988. Occupational exposures and brain cancer mortality: a preliminary study of East Texas residents. Am J lnd Med. 13:629-38. Spitz MR and Johnson CC. 1985. Ncurxsblastoma and paternal occupation: a case-control analysis. Am J Epidtmiol. 121:9249. Stem FB, Waxwcilcr RA, Beaumont J J, Lee ST, Rinsky RA, Zumwa/dc RD, Halpcrin WE, Bicfoaum PJ, Landrigan PJ, and Murray WE. 1986. A case-control study of lcukemia at a naval nuclear shipyard. Am J Epidemiol. 123:980-992. Stem RM. 1987. Cancer incidence among welders: possible effects of exposure to cxtr~rncly low frequency electromagnetic radiation (ELF) and to welding fumes. Environ Health Perspree. 76:221-9. Strumza MW. 1970. The effects of proximity to high voltage electrical wires on human health. Archives des maladies profe~sionaellts, de mldecine du travail et de sicuritl sociale, 31: 269-76. (Cited in Hauf, 1982). Swcrdlow AJ. 1983. Epidcmiology of cyc cancer in adults in England and Wales, 1962-1977. Am I Ep/dem/ol. 118:294-300. Thcriault G. 1990. Health effect of electromagnetic radiation on workers: cpidemiologic studies. NIOSH. Proceedings of Workshop, January 1991. Thomas TL, Fontham ETH, Norman SA, Stemhagen A, and Hoover RN. 1986. Occupational risk factors for brain tumors, a easc-rcfcrcnt death-certificate analysis. Scan J Work Environ Health. 12:121-7. Thomas TL, Stollcy PD, Stcmhagen A, Fontham ETH, Blceckcr ML, Stewart PA, and Hoover RN. 1987. Brain tumor mortality risk among men with electrical and clectronics jobs: a case-control study. J Nat Cancer Irkst. 79:233-8. Epidemiology of Health Effects and Exposure to EMF 3-63 Tomcnius L. 1986. 50-Hz electromagnetic environment and the incidence of childhood tumors in Stockholm County. Bioeltctromagnetics. 7:191-207. TSmqvist S, Norell S, Ahlbom A, and Knave B. 1986. Cancer in the electric power industry. Brit J Ind Med. 43:212- 3. TSmqvist S, Persson T, and Knave B. 1987. Cancer incidence in occupational groups exposed to low frequency electromagnetic fields. (Abstract). Stockholm Sweden: Andera Ahlbom, National Institute of Environmental Medicine. Tyncs T and Andersen A. 1990. Electromagnetic fields and male breast cancer. Lancet. 336:1596. Vlgexi D and Olin R. 1983. Incidence of cancer in the electronics industry: using the new Swedish Cancer l~nvironmcnt Registry as a screening instrument- Brit J lnd Med. 40:188-92. Viger6 D, Ahlbom A, Olin R, and Sahlstcn S. 1985. Cancer moeoidity among workers in the telecommunications industry. Brit J lnd Med. 42:191-5. Vcna JE, Graham S, Hellmann R, Swanson M, and Brasurc I. 1991. Use of electric blankets and risk of postmcnopausal breast cancer. Am J F. pidemioL 134:180-5. Vcrrcault R, Weiss lqS, Hollcnbach KA, Strader CH, and Dating JR. 1990. Use of electric blankets and risk of tcaticular cancer. Am J F_.p/demioL 131:759-62. Werthcimer N and Lccpcr E. 1979. Electrical wiring configurations and childhood cancer. Am J EpidemioL 109:273- 84. Worthelmer N and Lccpcr E. 1982. Adult cancer related to electrical wires near the home. lnt J Epidemiol. 11:345- Worthelmer N and Lccpcr IL 1986. Possible effects of electric blankets and heated water beds on fctal development- Bioelectromagnetics. 7: 13-22. Werehelmet N and Lccpcr !a. 1989. Fctal loss asseciatcd with two seasorrel sources of electromagnetic field exposure. Am J F.p/demiol. 129:2204. Whotan E. 1985. Toxic Terror. Ottawa, IL: lameson Books. Wildund K, l~irthom J, and Eklund G. 1981. An application of the Swedish Cancer-Environment Registry- Lcukcmia among telephone operators at the Tclecommunications Administration in Sweden. Inter J Epidemiol. 10:373- 376. W'~kins JR III and Koutras RA. 1988. Paternal occupation and brain cancer in offspring: a mortality-based case-control study. Am J lnd Med. 14:299-318. Walkins JR III and Hundlcy VD. 1990. Paternal occupational exposure to electromagnetic fields and ncuroblastoma in offspring- Am J EpidemioL 131:995-1007. WHO. 1964. Prevention of Cancer: Report of a WHO Export Committee, WHO Technical Report Series. 276:1-53. Wright WE, Peters JM, and Mack TM. 1982. Lcukcmia in workers exposed to electric and magnetic fields. Lancet. November 20:1160-1. 4-1 4.0 EXPERIMENTAL STUDIES environmental variables, the Committoe conccntratcd 4.1 Introduction on results of in vivo and in vitro controned laboratory experimentation. Consequently, this rcport does not Rcscarch on the biological effccts of electric and addrcas the rcsults of field studies involving crops, magnctic ~clds (EMF) can bc dividcd into two basic livestock, and naturally-occurring vcgcUtion and catcgories: (I) cpidcminlogic me. arch, previously wildlifc. discu~scd in section 3.0, and (2) laboratory cxpcrimcntation comprising in vivo (alive)sttKlic~' of ~ review and cvaluation ~ three typeJ of EMF cffccts on humans, lower animab (c.g., rat~, lltcraturc~ourcc~: baboons, ctc.), and in vitro (test tubc) studh at thc collularlcvcl. 1. Litcntturc summarica and background papers prcparcd by rcgulatory agcncies, scicnti~c or Although, laboratory stndic~ gcncrally providc a mcdical socicdc~. Duc to their complctcness, greater opportunity to control cxtrancous variables as thcsc documcnt~ wcrc drawn upon heavily to compared with cpidcmiologic and field studies, thcrc support data cvaluations and conclusions regarding still exist many opportunities for sources of crror to specific biological cffccU. The following cntcr into thc best designed laboratory study. It is summaries and background papcrs wcrc among possible that EMF scientific litcraturc, like aU scientific those rcvicwcd by thc Committcc: lltcraturc, contains somc false positives (i.c., showing cffcct~, whca they truly do not exist) and some faLsc . Congress of the United States Office of negatives (i.c., showing no effects, whcn they truly do Technology Assessment (OTA), 1989. exist). Although thc results of cxisting EMF research raise a number of intcresting questions, ~c complcxity · World Hcalffi Ofgamzation (WHO), 1984. of study findings and thc possibility of thesc crron makc it difficult to sort ~rough thc literature, interpret the evidence, and draw definite conclusions with · American Institute of Biological Sciences respect to EMF cffccta. (AIBS), 1985. Whilc the quantity and quality of EMF research havc · Florida Electric and Magnetic Fields Science improvcd dramatically in recent years, thc EMF cffccts Advisory Commission, 1985. data buc ia still in a relative "state of infancy" when comparcd to the research literature on othcr * New York State Power Lines Project environmental cxposurc ri~ks (e.g., ionizing radiation). Scientific Advisory Panel, 1987. Continued cxporimentation will hclp to rcduco thc currcnt level of complcxity and inconsistency inherent · Crcascy and Goldberg, 1989. in EMF findings, while incrcasing ~c "robustness" (strcagth) of tc~t resuh and con~dcncc in researchers' · Report to the California State Legislative by conclu~inn~. Continued study should includc: standardized and broad-bucd rcplication of studies, the Califorma Public Utilities Commission m expansion of the number of experimental species cooperation with the Califorma Dcpamnent exposed to EMF, a high lcvcl of quality control ofHcalthScrviccs. 1989. maintained across disciplinary lines (i.e., biology, chemistry, enghtccring, etc.) and, a balanccd and · Report by T. Sykes and Ping Li for the ~cxiblc cross-section of cpidcmiologic, in vitro and in Washington State Institute for Public Policy. vivo studies. 1990. Thc Commitice examined a numbcr of ~t~ sources on · Oregon EnL'rl~ Facility Siting Council. 1990. thc binlogical cffccts of clcctric and magnctic Topic~ included EMF cffccu on animal and human 2. Sccondary sources wriU~n or edited by mcmbcr~ behavior, cancer, growth and development, cndocrinc of ~c scicnti~c community, and public testimony and immune system functions, and biological prcscntcd by medical/research cxports rccognizcd mcchanismL In an effort to diaccm and cvaluatc thc by ~cir pocrs via professional honors and awards ottcn subtic cffcct~ of EMF, and to mitigatc ~c as wcll as professional appointments. Thc~c potentially confounding cffcct~ of naturaliy-occurring sources wcrc also uscxt to corroborate statcmcnts 4-2 Health Effects of Exposure to Powerline Frequency Electric and Magnetic Fields and conclusions presented in the aforementioned Reviewers of this document arc advised that the results iiu:raturc summaries and background papen. from these laboratory studies should be consider~t together with epidemiologic study results. Each 3. Prinmry research reports published in scientific provides information that is needed to evaluate the journals. These sources were consulted primarily overall EMF health issue. Epidemiologic studies to validate statements and conclusions regarding provide scientific observations on hunurns that could biological effects presented in literature summaries not be obtained from human laboratory tests, and and beckground papers referenced in the conversely, laboratory studies provide information from tests performed on lower animals, that cannot be Committee's report. performed on human subjocts. During the course of this review and evaluation, the Committee found that the EMF scientific literature 4.2 Stlmf~slFY contains results of laboratory studies that were performed under a variety of conditions, which The Committee believes that based on its evaluation of increases the complexity of reaching conclusions existing research literature, there is at this time no regarding potential EMF effects. Some of these conclusive evidence to suggest that EMF due to electric conditions arc presented as follows; studies were power transmission lines pose a human health hazard. performed: Results of the studies reviewed arc summarized in Tables 4.1 to 4.5 A point-by-point synopsis of the · using a variety of animals, human subjects, scientific fmdiogs which led the Committee to this and tissues; evaluative judgemeat is presented hercin, followed by detailed accounts of experimental results and · using different age groups and sex conclusions. The Committee feels that this evaluation basically is corroborat~l in other EMF literature differences; summaries and background reports prepared by expert scientific and rescaroh panels (e.g., World Health · using EMF frequencies other than the 60 and Organization, American Institute of Biological 50-Hz frequencies used to transmit and Sciences, State of Florida Electric and Magnetic Fields distribute electricity in the United States and Science Advisory Commission, Office of Technology Europe; Assessment, The California Public Utilities Commission in cooperation with the California · some considered the natural DC magnetic Department of Health Services, Washington State field and others did not; Institute for Public Policy, and New York State Power Lines Project Scientific Advisory Panel.) · exposing animals and tissues to higher fields than would normally be present in nature, A numbor of the in vitro studies reviewed, reported even al~er scaling the exposures to account some degree of effects on cells exposed to EMF. while for the differences in the lower animal's body many othen have reported no effects. Although cellular changes may occur in vitro, these changes may size as compared to humans; and not express themselves in the whole animal. For example, when a cell is pertueoed by an external agent, · using different exposure durations (e.g., acute other cellular processes may compensate for the and chronic) which may have over or under- change, so that there arc no overall adverse effects on exposed the animal or tissue as compared to the organism. Although the results of these in vitro what happens in nature. studies arc complex and inconclusive, the growing number of positive findings has suggested that under In an attempt to structure the broad spectrum of specific conditions even weak EMF can produce existing literature, and create a better understanding of changes at the cellular level. the EMF data base, the Committee categorized the report's experimental laboratory studies evaluation into As is the case with in vitro studies, some in vivo five major sections: (1) Behavior, (2) Cancer, (3) studies have shown positive effects of exposing the Growth and Development, (4) Endocrine System and whole animal to EMF, while other studies have Immunity, and (5) Biological Mechanisms. These reported no effects. Although it is not certain that the major seaion headings are subdivided further to effects observed in lower animals exposed to EMF will address spe0ific research topics and discuss separately also be expressed in humans, the assumption made the results of in vivo and in vitro studies. when dealing with human health is that similar effects may also be manifested in humans. This relationship is Experimental Studies 4-3 not always assumcd when extrapolating obscrvcd dccarboxylasc-ODC), and thc other is the ccllular effects to wholc animals. study of cell membrane modulators which control or inhibit calcium cf~ux or cause an The results of thc laboratory studies cvaluatai by the internal rcdistribution of calcium, that may Committec arc inconsistent and in some cases stimulate promotion of cancer. No firm inconclusivc, but enough information has been rcportccl conclusions an bc drawn on the promotion so the following observations can be nutde: thcory at this time. Hypotheses arc only now being advanccd. Indirect results arc · It appears that many variables can possibly compatible with the hypothesis that EMF may affer, t the results of laboratory studies, be a cancer promoter. Additional information including frequency, field intensity, exposure is clearly needed. duration, earth's static magnetic field, and frequent periods of exposure (that is chronic · Most of the EMF studies reviewed found no exposure may be less perturbing than a effects during embryonic development (i.e., pattern that involves frequent periods of tcratogcnic effects) or during post-natal exposure and non-exposure). Undoubtedly, growth. A few studies do show effects, some these variables play an important part in the of which occun'cd only under 'pulsed' fields, inconsistences rcpon~ai in the literature. which arc not normally associated with 50 or 60-Hz AC transmission lines. Overall these · It has beon shown that animals, including laboratory studies tend to lead to the hunurns, can detect fields under certain conclusion that there is no effect on conditions and this stimulus results in certain development or growth from EMF. physiological changes. Also, under given However, several recent studies performed conditions animals may avoid electric fields, on chicken eggs exposed to 'pulsed' ficlds but such actions have bccn shown to be have shown possible tcratogenic effects when mostly transient, sometimes lasting only a few exposure occurs during early embryonic minutes. Electric fields have not caused development. Reports on chicken continued aversiv· (avoidance) behavior, tcratogenesis, however, have shown except at fields high enough to produce an contradictory results, and studies using electric shock. it is felt that these behavioral chicken embryos arc of limited use in effects recorded in the literature do not predicting tcratogenic hazards in humans. constitute a health risk. · Based on several studies, EMF exposure · It is generally accepted that pewer-fralucncy apparently causes changes in the function of fields do not cause damage to genetic material the endocrine system of animals. For (e.g., DNA, chromosomes), as in the case example, reduction in nighttime mclatonin with higher(ionizing)frcqucncies. However, production and alteration of an animal's changes have been observed in the rate of biological rhythms have been recorded in DNA synthesis and in the production of animals exposed to 60-Hz electric fields. altered protoins, by interfering with the These observed changes arc within the range transcription by the RNA. The specific of normal changes observed in function of the mechanism responsible for these changes or endocrine system when stimulated from other the significance of these changes on the whole external stimuli (e,g., temperature, noise, organism is unknown. fight/dark). Numerous physiological effects have been hypothesized, due to melatonin · It is generally accepted that powcr-fralucncy reduction. However, the potential effects of fields arc not cancer initiators, since they do EMF on endocrine functions needs further not cause genetic damage. However, investigation. scientists have suggested that EMF can be a cancer promoter. Two general classes of in · Exposure to EMF has been theorized to affect vitro studies arc being performed to test the the animai's immune system. Whole animal promotion theory. One is the use of a studies (i.e., in vivo) have not shown such an biochanical marker (i.e., ornthnithinc effect. However, certain ceUular studies 4-4 Health Effects of Exposure to Powerline Frequency Electric and Magnetic Fields (i.e., in vitro) have shown effects, while such fields can be sensed by certain species under others have not. This inconsistency may be certain conditions. Stem et al. (1983) and Stern and due to the *window effect.' That is, effects Laties (1985) reported that rats detected electric fields may be observed at certain frequencies and as low as 3 kV/m, and as high as 10 kV/m. Kato et al. intexmities, but not at those below or above (1989) reported that the body hair and whiskers of rodents vibrated when the animals wen exposed to the 'window'. No definitive conclusions can electric fields; however, such sensitivity has been be drawn from the existing dats base. shown to be extremely weak as opposed to sensitivities However, since the immune system is a major to visual, audio, touch, smell, and taste stimuli surveillance mechanism that protects the host (Terrace, 1988). 5tell and Adcy (1988) found that animal from a variety of diseases, it is Spraguc-Dawlcy rats exposed to 60-Hz electric fields important to further explore effects of EMF and monitored for detection of fields, as a function of on the functions of the immune system. circadish rhythm activity exhibited no differences in Hypotheses need to be developed and tested detection pcrfonnanco during high or low activity before any definitive conclusion can be periods. Their data did not indicate · difference in the drawn. rat's ability to detect an electric field up to 25 kV/m, during different phases of their circadin rhythm cycle. · Since power-frequency fields (i.e., 50 and 60- Hz) at environmental levels arc not Perception of EMF in humans typically is accompanied by movement of hairs on the back of the neck, arms, sufficiently intense to break chemical bonds and/or hands, in response to thc electric and magnetic or warm the tissue, many scientists have felt force of the fields. Kato et al. (1989) reported that that insufficent energy could bc transferred to when hair follicle rcceptors arc excited by an electric biological systems to induce any changes. field, the hairs actually vibrate. However, recent studies have suggested that under certain conditions these low intensity In relation to the ability of humans to detect an electric fields can induce changes at the cellular level. field. Terrace (1988) noted that various studies have The actual biological mechanism is unknown, reported threshold levels ranging from 9 to 27 kV/m. but various ones have been postulated. All of Dcno and Zaffanella (1975, as cited in Stern ctal, the propom-~d mechanisms arc speculative at 1983) reported humans being able to detect electric this time. More mechanistic research is fields between 5 and 15 kV/m. Graham and Cohen needed to test the plausibility of human health (1985) found that 90% of seated humans could detect an electric field of 9 kV/m. Also, 20 Hz is the effects. If a mechanism is established at the frequency reported at which the greatest sensitivity to cellular level, this will support the results EMF has been observed in humans. Thi~ frequency is from the positive cpidcmiologic and in vivo much lower than the frequency used for AC power line laboratory studies. transmission (50 and 60-Hz). 4.:3 Effects on Animal and Human Tucker and 5chmitt (1978) found that, out of 200 Behavior human~, none could perceive amagnetic field of 7.5 - 15 gauss. Graham and Cohen (1985) found that humans could not detect magnetic fields up to 0.4 A review of extant research literature evaluating EMF effects produced by high-voltage AC powerlines on gauss. animal behavior can be divided into four main topics: detectability, ·voidanco, activity, and performance. 4.3.2 Avoidanc~ Observable and/or undctoctable and subtle effects of EMF on the biochcmical/physiological integrity of The fact that certain animals and humans can detect the whole organisms may cause changes in these presence of EMF does not provide the information behavioral areas, some of which may be disruptive to necessary to prcdict whether the organism will fred that the animal. As stated by Salzingcr (1989), 'the stimulus pleasant, annoying, or innocuous. Given that exquisite sensitivity of behavior makes it an effective an animal could in some way detect the presence of early warning system for illness.' EMF, and that detection resulted in physical discomfort and/or physiological distress, an avoidance behavior 4.3.1 Detoctabilit~, response might be expected. However, experimental results to date are inconclusive. With respect to Results of experiments porformed on the ability of avoidance, Hjcresen et al. (1980) showed that as various test animals to detect EMF have shown that electric field strengths increased, rats avoided electric fields of 75 kV/m. However, a subsequent re-analysis F~pertmeraal $1udies 4-5 of the data showed that the rats avoided these fields effects (Smith et al., 1979; Graves et al., 1985) while only during their sleep/rest periods (i.e., light patti of others report a variety of effects (Hjeresen et al., 1980 the 12 hour light/12 hour dark cycle). Hjeresen et al. and 1982; Rose·berg et al., 1981 and 1983). In (1982) observed simihr results in female miniature pigs Salzingcr's summary report (California Public Utilities exposed to 30 kV/m; that is, the pigs avoided the fields Commission Report, 1989), hc states that there is some during their sleep/rest period. However, Cream et al. increase in activity in exposed adult animals and some (1982, cited in California Public Utilities Commission evidence of reduction in activity when the exposure 1989) using the same apparatus u in Hjcrescn et al. takes place pre-and/or per···taLly. (1982) study, observed no effects on rats exposed to electric fields up to 100kV/m. Smith and Justesen (1977, cited in California Public Utilities Commission Report, 1989) reported a slight Stern and Laties (1987) found that, under many increase in the motor behavior of mice in the presence conditions, even in a 100-kV/m electric field (60 Hz), of 60-Hz magnetic fields; however, this increased these fields were not a very aver·ire stimuli for rats. activity was not sustained and was observed only at the In a sacchar·n-flavored water experiment. Cream eta]. onlet of field charge. Davis et al. (1984) found no (1984) concluded that exposure of moderate duration to changes in mice activity levels in the presence of DC or 60 -Hz electric fields (133 kV/m or less) did not AC magnetic fields. Groh et aL (1988)found that a produce taste·version learning in rats. In · number of response to an electric field (especiaLly observed phase instances, reactions to electric fields have been shown shifts in light/dark-induced carcad·an rhythms) to be transient, sometimes lasting only · few minutes depended on field strengths (> 25 to 35 kV/m), a (Hackman and Graves. 1981; Graves et al., 1977 cited seasonal light/dark effect sensitivity, and exposure in Florida Electric and Magnetic Fields Science during susceptible phases oft he carcad·an cycle. They Advisory Commission, 1985). Researchers typically observed a threshold of 25 to 35 kV/m below which the attribute this initial response to the animals' recognition majority of mice showed no consistent, m~asurable of an environmental change, (i.e., "what-is-it?") to response, and · maxamazed effect above 100 kV/m, but which the animal rapidly adjusts (Hackman and no simple dose-dependent response for activity, or Graves, 1981; Rose·berg et al., 1981). respiration. They concluded that, based on results of their experiments, ali known potential long-ten health Based on a review of recent research summaries and risk effects of high-intensity electric fields could be background papers, the Committee finds that exposure attributed to their actions as carcad·an reguiatora. of laboratory animals (primarily rodents) to a variety of electric fields has failed to produce any conclusive Working with social groups of baboons, Rogers et al. evidence of physical discomfort or internal distress. (1988) found that exposure to 60-Hz electric fields Further, electsic fields have not been shown to be produced changes in posture and positions, usually at aver·ire stimuli, unless the field is intense enough to the onset of field charge. They suggested that the produce electric shocks. Results of conditioning animals reacted (huddled together) perhaps to reduce studies have failed to yield any evidence that EMF can field strength, and to increase shielding, all protective motivate an animal to avoid such fields. or that such responses to perceived stimuli. However, these fields can cause conditional feelings of internal reactions were not consistent across all experiments unp!e~ntness or stress. FinaLly, studies using and were temporary in nature. The huddling effect was biological indicators as direct measures of stress have found to be dependent on the strength of the fields. failed to yield any ~vidence that EMF are aversave After three days of exposure, proximity values (i.e., stimuli. These conclusions basically coincide with baboon locations relative to one another) returned to findings reached by. the Office of Technology normal, and re-exposure of previously exposed animals Assessment (OTA. 1989), by testL, nony given by Dr. produced no effects. Rogers et al. (1988) concluded H. 5. Terrace CTerraee, 1988) and the excellent that exposed animals reacted to the field as a threat by summary prepared by 5alzinger in the California Public huddling at the beginning of exposure, "learning" after Utilities Commission Report (1989). some time that there was no danger, and returning to normal patterns. 4.3.3 ACtiVity Based on a review of behavioral research performed to Another area of research drawing the attention of date, the Committee concludes that positive resuRs behavioral scientists is the possible effect(s) of EMF on from research on EMF effects on activity show that general animal activity as controlled by circaalia· observed responses appear to be tams·eat or·eating rhythms and physiochemieal regulators. To date, there r~sponses (i.e., responses to stimuli not previously is a general hck of consistency in these studies encountered). Under normal environmental conditions, regarding EMF effects. 5ome studies have reported no there appears to be no indication Umt exposure to EMF 4-6 Health Effects of Exposure to Powerline Frequency Electric and Magnetic Fields alters circadian rhythms. Where effects on circadian wward a reduction in response rate was evident in the rhythms have been reported, findings have come from offspring. studies concluded under unnatural laboratory conditions that attempt to elirninate any cue to time of day. The effects of 60-Hz EMF on performance of human Additionally, exposure-~ reported to be capable of subjeca have been studied by a number of ngArohen. altering cizcadian rhythms have been much higher than Gibson and Moroney (1974) attempted to evaluate the those that would be encountered in the immediate effect of EMF on hutnan perfonnanco as measured by vicinity of a high-voltage transmission line. a battery of standard intellectual tests in the presence and absence of such fields. In each instance, EMF The Committee's findings parallel the conclusions failed to produce any decrement in performance. drawn by the American Institute of Biological Sciences Gamberale et al. (1987) examined 26 utility linemen in (1985) in their post-1977 literature review on biological the laboratory over two days, measuring behavioral and human health effects of extremely low frequency performance, EEG, mood scales, subjcx, tive symptoms ele~'tromagnetic fields, by Terrace (1988) in his and various blood chemistry parameters. No statistical summary of research regarding EMF effects on difference was observed between exposed and control activity, and by the Office of Technology Assessment groups which could be attributed to exposure to EMF. (1989) in their literature review of the biological effects In two double-blind experiments, exposing male of power frequency electric and magnetic fields. These humans to an electric field of 9 kV/m and a magnetic publications concluded that none of the subtle effects of field of 200 mG, Graham et al. (1988) reported a electric and magnetic fields wport~ in the scientific slowing of the heart rate and changes in the central literature constitute a definite health risk, and that nervous system. These changes tended to occur soon observed behavioral effects appear to represent an after field onset or offset, suggesting that changes in initial awareness response to an external stimulus, to exposure may be more important than exposure which the organism quickly adjusts. Although 60-Hz duration. However, exposure to higher levels (12 fields appear to have an effect on the periodicity of kV/m, 300 mG), resulted in no consistent significant physiological functioning, it has not been proven that differences between the exposed and control groups. these effects arc harmful or even prolonged. They postuhted that exposure may interact with biological systems only in certain limited "windows" of 4.3.4 Psrformsiftcs stimuhtion, i.e., for a particuhr frequency, some field intensities may produce an effect but intensities below A large number of research studies have evaluated the or above the "window" do not. effects of EMF on the performance of learned behavior in anhnals, including humans. Again, study results A number of researchers have reported similar results, have b~n inconclusive. Coolho et al. (1987)found that i.e., that EMF effects observed at a particular field performance rates of tension, foraging, and grouping strength appear to be "tuned out' by changing the field stereotype behavior in baboons exposed to 30 kV/m frequency or intensity. and 60 kV/m were significantly elevated during the exposure period as compared to pre and post-exposure This apparently unusual relationship between EMF and periods. However, some of these same researchers biological system interaction tends to complicate the found that baboons exposed for six weeks to 30 kV/m usually clear correlation between a measure of or 60 kV/m electric fields exhibited only transitory exposure to a physicaYchemicai agent "dog' and the behavioral changes (Rogers et al., 1988). Behavior consequent effect. patterns returned to normal in one to three weeks alter exposure. The threshold for electric field detection was The apparent "window" nature of these effects may 13 kV/m, and no effects were detected for acquisition, imply that effects seen at particular frequencies may not porfurmanceor responso to food offerings. Orr et al. be observed at other, higher or lower values. (1987) reported that responses of baboons to food Similarly, the olden-applied 'more is worse" rwvard were affected by 30 kV/m and 60 kV/m (60 rehtionship between dose and effect used so often Hz) electric fields upon initial field exposure, but when applied to chemical exposure does not appear to responses returned to normal within a few days. hold, since for a number of experiments hrger values of an applied field have not caused a hrger, or in some Salzinger et al., 1987 (as cited in Salzinger, 1989) cases any effect, compared to a reduced field. found no effects on the performance of a memory task by adult male rats exposed for 72 hours to 60-Hz Alternatively, changes in conditions of exposure, or electric and magnetic fields of I gauss and 30 kV/m. changes associated with differences in field onset or However, when these same experiments were offset, may be relatively more important than either performed on adult female rats exposed while pregnant duration or intensity of exposure to a steady-state field. (22 days) and for the fwst eight days of life, a trend For example, OTA (1989) reported that background Experimental Studies 4-7 static field conditions (i.e., how a field is applied can be drawn. Various cancer types and locations in relative to earth's natural static magnetic field) may the body (i.e., organs) have been attributed to EMF influence the observed effects of an applied EMF. exposure. However, it appears, based on the aforementioned cpidcmiologic studies, that possible The Committee feels that based on the results of studies cancer types attributed to EMF are being narrowed reviewed, it appears that EMF effects observed under somewhat to lcukcmias in children, and tumors of the laboratory conditions have no long-lasting effects on central ncPv'ous system in adults. lower animals or humans in short-tom memory or cognition. However, possible effects on learning needs 4.4.2 Carcinogenesis Models further investigation. Effects such as slightly slower reaction times and transient slowing of heart rate can be An integral part of dcl'ming any possible relationship attributed to oricnting responses which can bc evoked between cancer and EMF is the establishment of a by a variety of natural stimuli as well as EMF. plausible biological mechanism. It has been proposed by certain scientists that EMF may promote cancer 4.3.5 Conclusions aRcr another agent has initiated the cancer forming process. The carcinogcncsis process has bccn The Committee feels that, al~cr a review of existing described using both a two-stage model, and a three- research literature, wc can conclude that EMF stage model. behavioral research to date has shown no effects of any significance to human health risks related to Stage 1 is the "initiation," which results in a permanent psychological functions/behavioral response. (i.e., non-reversible) change or mutation of the ccil's genetic material (i.e., DNA). Initiation is caused by When detected, EMF appear to elicit a "something's carcinogcnic agents, like ionizing radiation and certain there' recognition response, similar to normal reactions chemicals. Stage 2 is called "promotion," in which the to previously uncncountcrcd stimuli. No research to initiated cells expand into a visible tumor after repeated date has presented any conclusive evidence that these exposure to the promoting agent (e.g., phorbol esters). fields, detected or not, produce any deleterious and/or These changes may not be permanent, and can be long-lasting hnpacts on lower animal or human reversed in some cases. In the two stage model, the behavior. These views are shared by other scientific tumor may bc benign or malignant. In the three stage and research agencies such as the Office of Technology model, the benign tumor becomes malignant by passing Assessment (OTA, 1989), and The American Institute through a third stage called "progression." of Biological Sciences (AIRS, 1985). 4.4.3 Cancer Initiation 4.4 Cancer Generally, it is relatively accepted by most scientists 4.4.1 Cause - Effect Relationships that EMF are not cancer initiators, i.e., cells exposed to EMF have not shown any change or mutation in the DNA material, a requirement of an initiator. Human Determining the cause of the various human cancer types and their specific location (i.e. organ) has been lymphocytcs (i.e., white blood cells) showed no effects ' when exposed to 60-Hz fields (Cohen, 1986; Cohen et and is continuing to bca difficult task. The discussion al., 1986a; 1986b as cited in Crcascy and Goldberg; in section 3.0 (National Cancer Rates) provides information on various cancer rates since 1950. Our Livingston ctal. 1986). No effects were observed in mouse bone marrow cells (Carstcnscn, 1987) and understanding of the cause-effect relationships of Chinese hamstar ovary cells (Rccsc ctal. 1988) various cancers has greatly increased over the past 30 years. Wc know that certain risk factors are associated exposed to EMF. Also, it was the conclusion of the New York State Powerlines Project Scientific Advisory wiffi certain cancer types. For example, smoking is Panel (1987), that it was unlikely that electric and related to lung and oral cancers, the sun's rays (i.e., magnetic fields damage human chromosomes. ultraviolet) to skin cancer and a high-fat diet to cancer of the breast and colon tract. However, research to determine an association between EMF exposure and However, other studies have attributed DNA damage to cancer is still in the infancy stage, when compared to EMF exposure. Nordcnson and Hansson (1987) the volumes of research literature on other reported chromosome damage in human amniotic cells environmental exposure risks (e.g., at least 50 years of exposed to 50-Hz sinusoidal and 20-kHz sawtooth data on ionizing radiation). The literature does not magnetic fields. But no chromosome damage was contain years of scientific studies covering the 50 and observed in switchyard workers during a study by 60-Hz frequencies, from which definitive conclusions Bauchingcr et al. (1981). d'Ambrosio ctal. (1985) found an increased percentage of chromosome 4-8 Healzh Effecls of F. xposure Io Powerline Frequency Electric and Magnetic Fields aberrations in bovine poriphcral blood lymphocytes genetic factors arc considered to constitute the initiation (i.e., white blood cells) exposed to 50-Hz electric fields step. For example, mice with a high frequency of for 72 hours. E1 Nahas and Oraby (1989, cited in mammary cancer (e.g., the C3H strain), and rats prone Creascy and Goldberg, 1989) exposed Swiss male mice to lcukemia (e.g., the ACT strain) might be suitable to 100, 170, 220, and 290 kV/m 50-Hz electric fields models. for 24 hours. The high exposuro was to scale up to humans. They found no increase in micronuclear 10 Vitro Studies. The growth enzyme ornithinc polychromatic erythrocytes (i.e., red blood cells)in the dccarboxylase (ODC) has been used as an indirect bone marrow exposed to 100 kV/m, but significant marker of the promotion step in carcinogcnesis. It has increases occurnx:l in those mice exposed to the higher been shown that ODC production increased in those fields. Since the exposure system was not described in cells exposed in vitro to EMF, depending on amplitude, great detail, the observed effects may have been due to frequency and exposure duration (Byus and Adey, "microshocks'at the higher exposure levels. 1988). This increase, however, does not necessarily mean that since ODC increased, these cells are on their Study results are inconsistent and a fee researchen do way to becoming a tumor and EMF are cancer report effects on genetic material. it is possible that promoten. A variety of stimuli, which do not play a effects may be observed at a field of 50 Hz, but not at role in careinogencsis, will induce ODC activity (e.g., 60 Hz, a possibility stated by Crcascy and Goldberg drugs, hormones). In addition, the enhancement of (1989) in their summary report; Extremely Low ODC activity by EMF was much less than for known Frequency - Electric and Magnetic Fields and Cancer: promoters, such as phorbol esters. However, it is A Literaturo Reviee. However, given the prr. sent fn'mly established that all cells that are becoming status of the EMF data base, it is the position of the tumors have increased ODC activity, and have lost Committee that EMF are not cancer initiators. their ability to control the activity of the enzyme. Permanent changcs in DNA material or abnormalities caused by damage to cellular DNA production or repair Frazier et al. (1989) tested the hypothesis that EMF can mechanisms would provide evidence that EMF acts as promote transformation of initiated cells. They a cancer initiator. However, to date researchers have exposed cells (i.e., C3HIOT1/2) to various exposure been unable to demonstrate any conclusive detrimental combinations; and initiator (i.e., 60Co irradiation), a differences be~ecn the DNA structure and function of known promoter [i.e., 12-O-tetradecanoyl-phorbol-13- EMF-exposed and unexposed cells. This conclusion acetate (TPA)] and to O. 1, 0.75 or 6.0 G of 60-Hz AC concurs with Rosen's (1988). position in his royjew of magnetic fields. The radiation exposures significantly electric and magnetic field health concerns, wherein he increased the transformation frequencies and, as state~: *Studies of genetic damage are relevant to the predicted, TPA increased transformation frequencies of initiation stage. Changes in DNA production or repair, irradiat~i cells by approximately 10 times. However, or the occurrence of abnormalities, would be manifest exposure to the magnetic fieIda did not significantly as portoanent changes .in the DNA material. To date, alter transformation frequencies of either initiated or data demonstrate no difference between cells exposed promoted cells as compared to sham-exposed cells, an to electric and/or magnetic fields and sham conditions." indication that magnetic fields are not carcinogenic promoters. 4.4.4 Cancer Promotion Other in vitro studies have shown that cells exposed to EMF apparently are not cancer initiators; however, EMF undergo changes in enzyme activity (Cain et al. certain scientists believe that EMF are possible cancer 1987) and ceil-to-cell communication (Fletcher et al. promoters, i.e., they begin the second stage of the 1937), changes that are similar to those caused by cancer-forming process (tumor formation). Certain known cancer promoters (e.g., phorhol esters). studies have not shown EMF to be a cancer promoter. Phillips et al. (1986 cited in the California Report, However, the curr~nt dala base is insufficient to either 1989) observed enhanced growlh rates in malignant accept or reject the promotion hypothesis. Most in human colon cancer cells exposed to EMF, but Cohen vivo studies have trimsplanted tumors into normal (1987) could not duplicate the same results. Aho, animals, or exposed these animals to known cancer Adolphe et al. (1987 as cited in Creasey and Goldberg, initiators and then exposed them to EMF. As stated by 1989) observed no effect on growth of malignant Creasey and Goldberg (1989) "no studies have been human uterine cervical cancer cells in culture exposed done that expose normal animals to extremely low to 50-H z EMF. Goodman et al. (1989) observed frequency (ELF)EMF and follow the spontaneous induced quantitative changes in measenger RNA developmeat of tumors." They recommend using (MRNA) and proteins in human cells (HLGO) exposed rodent strains having a high natural incidence of to sinusoidal signals with repetition rates at 60 Hz and malignancies to test the promotion hypothesis. These 72 Hz. animals are genetically predisposed toward cancer and Experimental Studies 4-9 Although the ruults of in vitro studies arc far from change in growth, development or overall health of the proving or disproving that EMF arc cancer promoters, exposed chicken embryos. Sikov ctal. (1987) reported such findings cannot be ignored and additional malformations in Hanford miniature swine offspring of information is needed. EMF-cxposed sows. Durfcc ctal. (1975, cited in the Florida Report, 1985) exposed chicken embryos to In Vivo Studies. There have been several in vivo fields of 0.001 to 3.6 kV/m at a frequency of 45 Hz to studics using verious anin.all for testing the EMF- 75 Hz during and after incubation and observed no cancer promotion hypothesis. The majority of studies effects on fertility, hatchability, survivabillty, weight have shown no cffccu, with a few cxperimcnu even gain, or behavior. Sandstrnm et al. (1987) exposed showing beneficial effects. Leung ct al. (1988b) fertilized hen eggs, during the farst two days of reported no significant differences between electric development, to magnetic fields with an asymmetrical field exposed and uncxposed rats in the number of rats saw-tooth waveform, with no observed significant that developed mammary tumors, but did find an increase of abnormalities. However, Delgado et al. increag in the number of mammary tumors per tumor- (1982) reported abnormalities in chicken embryos bearing rats exposed to 7, 12-dimethylbenzenc (a) exposed to low frequency pulsed magnetic fields (fields anthracene (DMBA) and 60-Hz (40 kV/m) electfie which are turned on quickly for only a brief period - fields as compared with those only exposed to the not normally found associated with transmission lines) tumor inducing chemical. Chandra and Stefano (1978) during early development. Ubeda et al. (1983) also found no effect on the growth characteristics of mouse observed effects in chick embryos exposed during the mammary tumors exposed to magnetic fields before first 48 hours of development to pulled electric and (i.e., in vitro)or after being transplanted (i.e., in vivo) magnetic fields. Subsequent independent studies were into healthy mice. Thomson et al. (1988) observed no unable to replicate these teratogenie effects (Silken a effect on the incidence or progression of !>388 !eukemia al. 1986; Maffeo et al. 1984). cells implanted in mice exposed to 60-Hz magnetic fields. Batkin and Tabrah (1977 cited in the California In an attempt to resolve the contradictory results from Report, 1989) aetnaHy reported a decrease in mouse these studies, the U.S. Office of Naval Research and tumor growth in. response to 60-Hz magnetic field the U.S. Environmental Protection Agency sponsored exposure. the "henhouse project". This project consisted of replicaling the same experiment at six independent 4.4.5 Conclusions laboratories around the world. Each laboratory exposed chicken eggs to extremely weak unipolar pulsed The Committee is of the opinion that the present magnetic fields with the same set of characteristics, and literature indicates that EMF are not cancer initiators, evaluated the same set of parameters: egg ferlility, and but indirect results are compatible with the hypothesis embryo abnormalities in development and growth. that EMF may be can'cer promoters. The results When the data were combined for all six laboratories, neither prove nor disprove the cancer promotion an overall inerase in the proportion of abnormal chick hypothesis. Additional information is clearly needed. embryos was found for the exposed embryos. However, the exact proportion of abnormalities in the 4.5 Developmentand Growth exposed population varied from lab to lab. Two of the six laboratories reported a significant increase in the proportion of abnormal embryos, while four labs 4.5.1 TeratoOenlc Effects observed no effects (Bennan et al. 1988a, 1988b). Since these results were reported for pulsed magnetic Numerous laboratory studies have been performed to fields, not usually associated with powerline evaluate the effects of EMF exposure on the transmissions, implications of these findings for lower development and growth rates of lower animall. animall and humans exposed to 60-Hz fields are highly Teratogcnic effects are those that occur during uncertain. This conclusion was drawn in the summary embryonic development, and if effects occur, they are report prepar?.d by the Office of Technology manifested in the malformation of the off-spring. Assessment (1989). In addition, the chicken embryo Various in vivo studies using different test animall have has been rejected as a standard model to be used in been performed to determine if teratogcnic effects evaluating the teratogenie potential of chemicals and, occur due to EMF exposure. therefore, may not be a good model for evaluating the teratogenic potential from EMF exposures. Graves et al. (1985) exposed over 20,000 chicken embryos to 60-Hz electric fields from 0.1 to 100 kV/m, An important consideration in the use of chick embryos and looked for effects during incubation, at hatching to monitor for possible teratogenie effects of EMF may and after hatching. They observed no significant be the e~ticality of time of exposure relative to 4-10 Health Effects of F. xposure to Powerfine Frequency Eltctric and Magnetic Fields embryonic development. For example, Mardn (1988) development re.arch to chtto: 'There is no observed effecU when exposure to pulsed EMF reproducible ~cienti~c basis for implicating power occurred during the fiat 24 hour, of incubation, but frequency electric and/or magnetic fields as being found no effect, if the embryos were exposed lator in cauaative of genetic chtmage that is reflected in growth their development. Allo, as sated by Creaaey and and development." Goldberg (1989) genetics may play an important pan in the variability of study results, since it is known that 4.5.3 Conclusions chickens of the ~une attain are not ncceaurily genetically similar. Supporting this possibility is In aumnury, moat of the EMF atudic~ reviewed by the Martin 'a (1989) finding that effects were observed in CommilJ. ee .how no effect during embryonic one strain (i.e., White Leghorn) of chick embryo, development or during poat-ntttal growth. A few exposed to pul~ed fields, but not in another (i.e., Arbor .tudie. do report effects, with some showing effects Acre). Another point that makc~ intorpretation of the under "pulm:xt" magnetic fields, which ~r~ not normally data more difficult is the observation made by Chernoff found u~ciated with 50 and 60-Hz AC tmumisminn (1989) that there is a high incidence of abnormal liner Certain atudic~ show effects using one strain, but embryos found in the control eggs in m~ny no effects with another. Also, high incidence of effect. invc~tigationL Thu~, the results observed may be due are observed in the controls of various studie~, making to unrecognized factor,, unrelated to EMF. interpretation of the data, more difficult. Overall, thc~e laboratory studies tend to lead to the conclusion that 4.5.2 Rsproductive Effects there is no proven detrimental effect on development or growth from EMF. Scto et al. (1983) exposed rm to an electric field of g0 kV/m for 21 hr/day until approximately 120 days of 4.6 Endocrine System and age. There was no statistically significant effect on hTIrrlunity food and water uptakc. However, there were significant growth patxcrn effects from four to eight wccka of age, which were not observed beyond eight weeks. Stuchly Propor functioning of an anin~!'a endocrine (hormone) et al. (1987) found no statistical differences between system depends on the precisely coordinm:d oporation rats exposed to magnetic fields of an unsymmctrical of all endocrine organs. An upset to any part of the s~wtooth waveform, and those not exposed. Sikov et system could trigger visible and measurable deleterious al. (1978, 1984) exposed rats to electric fields prior to effects. From a health pcr,pcctivc, potential effects of mating, and continued exposure of the pregnant EMF on endocrine function is an importlnt question. fcmlcs and observed no effects on fetal length or weight or on internal or cxtorrmi realformations. 4.6.1 Hormonel Effects However, they did obsorvc more atiHbirths in the cxpoRd group in ono of their cxporimcnts. Rommcrein Cannaciu et al. (19T/, u cited in C~mcnscn, 1987) et al. (1984) exposed rats to an electric ficid and found rcl~ortcd that initial increases in the adrerm] bormom no diffcrenco between the exposed group and the cpincphrinc and nor~-~incphrinc (which control uncxposed group (i.e., sluim group) in incidence of vesoconstruction and mediate transmission of nerve realformations in the off-spring of the rffst generation. impulses) in rats, apparently causnd by exposure to 200 However, when the fu~t gencratinn was rebred, a kV/m electric field for 6 to 72 hour,, stabilized significant increase in ms]formed fatuses were quickly. No further increases wcrc noted during a observed in the exposed group, but no such subsequent 12 day chronic exposure perind. malforrrmtiona were observed in a second replication. Additionally, there were no detectable ncurologicai or Lotz and Saxon (1984) reported reduced weight in trait ncurocndocrinc changes enrrelated with these increucd Rhceua monkeys (1 to 54 months old), chronicsHy hormorml increases. Since no deleterious effects were cxpom:d to EMF, but no weight reductions in fcmalct observed, the ruearcher, concluded tl~t the elevated ncurocndocrinc secretions induced by. clcctrlc field Smith et ,,l. (1981) exposed mice to very low electric exposure wcrc within the norms] response experience fieida and monitored the growth and development of for the test animalL Frcc et el. (1981) showed thlt over 1400 mice in 128 litter,. Neither fertility, number prolonged exposure of rats to a strong 60-Hz electric of mice born, litter size, nor sex ratio was affected. field (i.e., 68 kV/m, 80 kV/m) slightly lowered the Bcnz eta]. (1987, as cited in OTA, 1989) reported no plasma corticoatcronc levels together with that of effect on 3,000 mice exposed to EMF over three tcstostoronc and prolactin. They concluded thax 60-Hz generations. electric fieida may bring about subtic changes in the endocrine system of rots, and the changes may be Aaronson (1988) in his review of the biological effects related to alterations in cpisodic rhythms. Jollcy c~ al. of EMF concludes the following, regarding growth and (1983) found a reduction in insulin and calcium re]cue Experimental Studies 4-11 by bolated rabbit islcU of Langcrhans tissue. Udintscv 4.6.2 Circadian Rhythms et at. (1986, as cited in Creucy and Goldberg, 1989) showed that cxposur~ to an alternating magnetic field (50 Hz; 200 gauss) activated the hypothalnmo- Recent research has suggested that EMF can affect hypophyseal-adrenal system (affecting behavior, endocrine balance and function by altering an metabolism, maintenance of body temperature, etc.) in organisms' circadish (24 hrs.) and ultradian (less than rats. Levels of steriods and adreno eo~cotrophic 24 hra.) biologic rhythms. 'These rhythms are biologic hormone (ACTH) were elevated as well as plasma and processes controlled by both external stimuli (i.e., tissue free fatty acids and phospholipids, but prolonged light/ark periodicity) and hormones. The internal repeated exposur~ inhibited the activity of the clocks or pacemakers for these rhythms are believed to endocrine system. Evidence of increased lipid be Iotatill in the hypothalamus gland, and can be peroxidation after chronic exposure to EMF is influenced by specific external stimuli including postulated to represent a type of stress reaction te~'nperature, noise, light/dark, etc. Results to date (Creasey and Goldberg, 1989). appear to show that EMF can influence these rhythms. Groh et ai. (1988) reported that the circadian rhythms !n vitro studies performed by Lymangrover et al. (i.e., phase shill, dyschronion, torpor) of rats and mice (1987), using adrenal rat tissue, showed that 60-Hz were effected when exposed to electric fields of 100 fields greatly stimulated cellular response to ACTH. kV/m. Animals raised under an 8:16-hour light=dark Knrtashev and Ivanova (1988) reported activation of the cycle were less sensitive to electric field exposures than adrenal and thyroid systems in mice; however, Quinlan animals raised under a 16:8-hour light-dark cycle, In a et tl. (1985)failed to show any generalized activation summary report by Groh (1989), he concludes that EMF are circadian zeit gerbers (environmental cues). of the hypothalamo-hypophyseal-adrenal system in exposed rats. Although, they did observe a statistically However, the importance of EMF as an environmental significant increase in growth hormone production. cue relative to other more widely recognized external Michaelson and Lu (1988) showed that rats exposed to stimuli, such as light, is unknown. an electric field of 80 kV/m for four hours at 71 hour intervals exhibited no persistent changes in adrenal Much work is presently being done to determine the gland function (as measured by corticosterone possible effect of EMF exposure on the functions of the production), and no indication of physiologic or pineal gland. The gland is located at the base of the neuroendocrine stress. A study by Leung et al. brain in man and functions as a neuroendocrine (1988a) repot~xI that rats exposed to electric fields transducer, converting neurological input to hormonal exhibited a statiaticaHy significant increase in the output, primarily melatonin. An excellent review on incidence and severity of chromodacryorrhes (a gland the known functions of the gland and the possible secretion indicative of stress). The authors suggested EMF exposure effects are given by Reiter (1990) and that rats exposed to electric fields are subjected to a W'dson and Anderson (1990). chronic low-level stress. Michaelson (1987) monitored hormone balance in acclimatized (unstressed) and non- Melatonin production has a circadian variation, high at acclimatized (stressed) rats subjected to 50, 80, or 100 night and low during the day. Exposure to light of kV/m sustained or interrupted electric fields (60 Hz) sufficient intensity during the dark cycle leads to an for various times up to five hours, as well as daily immediate drop in mealtonin in several species, exposure t~-petition for five successive days, or at 48- including man. However, every slzcies seems to have 72 hour intervals. Utilizing a highly consistent and ~t different light inhibition threshold. In man, the pincai sensitive hormone assay system, he rcpot~d that gland receives its indirect sensory input from the retina. sustained and intermittent gO kV/m exposure may elicit The gland functions in the endocrine system as an subtle reguiatory adjustments of endocrine levels, but inhibitor on most other endocrine glands. Reiter (1990) that these levels were within the threshold for a in his review, concludes that the functions of the pineal recognized physiological stressor. He concluded that gland is to keep the animal in appropriate synchrony within the constraints of the expe:"nental design, there with its external environment. The gland contintutliy appeared to be no portu~afion of the endocrine system apprises the animal of the environmental state and in the rats subjected to the fields applied, and expressed adjusts its physiology accordingly. doubt that prolonged exposure would cause further detrimental effects, since adaptation and biological Wilson et al. (1981 and 1986) rcport~ that exposure of varigbility would help to moderate such influences on rodents to 60-Hz electric fields (i.e., 39 kV/m effective endocrine balance. field) can upset the pineal gland's circadian rhythm. They reported that prolonged exposure (30 days) of rats to an electric field significantly reduced the nighttime rise in melatonin and serotonin-N-acetyl 4-12 Health Effects of Exposure to Powerline Frequency Electric and Magnetic Field~ tansferag. The onsce of the effect occurred within Along with other obgrv·ble and mmurable three wccks ·Rer exposure initiation, and recovery w, parameters, aninml reproducfon and dcvciopmcnt can observed in less than three days ·~cr exposure ressation. Anderson et ·1. (1987) and Reiter ct ·1. accura~ly indicate the relative health of · synchronized endocrine system, because the ability of mammalian (1988) found that rats exposed to electric fields of 10, 65, or 130 kV/m from conception to 23 days of age animals to conceive and carry a pregnancy to term requires s highly coordinated sequence of endocrine- showed · reduction and time-shift of peak nighttime pineel melatonin. However, no dose-response mediated steps. rehtionship was observed. Young ·dult rats exposed for three weeks to 60-Hz electric fields resulted in The majority of studies to date have dcmonstrat~ no desynchronization of pinca] gland rhythms, and deleterious effects of exposure to EMF on reproduction statistically significant decreases in nighttime levels of and development. Ccrretclli et aL (1979) found that mclatonin and serotonin N-·cetyl trsnsferase (a blood long-term (i.e., two months) exposure to an electric vessel constriction enzyme) when compared to enntrol field of 100 kV/m had no effect on rats' fertility. Sikov animals. Changes were ·leo noted in pineel serotonin et el. (1984) and Bcnz ct el., 1987 (as cited in OTA, and N-·ceytl serotonin levels. Rats exposed to 10, 6S, 1989) using identical electric fields prior to and during or 130 kV/m 60-Hz electric fields from conception to gestation, found no changes in rat or swine mating 22 days showed stable but significantly suppressed performance or futility. Albert et ·1., 1984 (as cited in melatonin secretions compared to controls. OTA, 1989) was unable to demonstrate any signif'he·at Additionally, peak production of pineel mclatonin in aH differences between exposed and uncxposed developing exposed groups showed a phase shift delay of rats. Faro (1980)observed reduced growth in female mice cxposed to a 240 kV/m electric field, but not in approximately one to two hours when compsrcd to controls, and the effect appeared to be an "All or males. But there were no affects on the number of born None" response. and surviving progenies. Several other studies, Sikov et el., 1984, and Konerman and Monig, 1986, (as cited in New York State Powerlines Project Scientific In conjunction with the possible effects of EMF exposure on the eircsdian rhythm of animals, due to the Advisory Panel, 1987) observed no differences in animal weights after prolonged exposure to electric many known functions of the pineel gland, numerous fields. Hilton and Phillips (1981) also observed no other physiological effects have been hypothesized. effects on growth of rats and mice exposed to an Gland dysfunction has been implicated as a possible factor effectlag illness. This is owning partially to the electric field of 100 kV/m. They ·ttributed this to · Iterations in normal biological rhythms. The gland is eliminating or minimizing secondary factors (e.g., hypothesized to be involved in the etiology of corona ozone, harmonic distortion, spark discharge) which are associated with certain hboratory exposures. depressive illness, although its exact role remains unclear. The reduction and phase-shifting of the circadian peaks in meiatonin concentrations have been 4.6.4 Immune System Effects suggested as factors effecting mood changes. Stevens (1987, as cited in W'dson et el. 1990) suggested that Another theory that has been postuhtcd is that EMF reduction in melatonin concentration due to EMF may impact the body's immune system. The immune exposure in rats may result in incrcased circulating system is a network of cells and tissues which ace as · estrogen levels. This m·y stimulate mammary tissuc m·jor surveiihnce mechanism that protects the host prolifcration and thus increase breast cancer risk. organism from a variety of diseases, including cancer. Wilson (1988) suggests that pineel dysfunctions may In vivo research which attompts to assess the integrity contribute to the onset of depression or may exacerbate of the immune system, examines the reactions of the existing depressive disorders in humans. It has also whole organism to various challenges. Though been hypothesized that melatonin may have · typically more difficult ahd time consuming, in vivo stimulatory effect on the immune system. testing is a more ·ccurate reflection of the immune systems' cap·city for response, than extrapolation from It appears very likely that EMF cxposure can alter in vitro studies. pineel gland functions by decreasing mehtonin synthesis and release. However, given the many The literature on the interaction of EMF with the possible functions of the gland, it remains to be immune system exhibits a wide variety of conflicting determined whether EMF effects on the gland's results. Only data from exposure of whole animals to functions represent · health risk to humans. 50 or 60-Hz electric or electric and magnetic fields can be said to show some consistency, and those effects 4.6.3 Other Biological Functions consist of relatively small or no effects. No conf'trrned repeatable research to date, (as reviewed by Bockman, 1989) has shown that exposure to electric and magnetic Esperimental Studies 4-13 fields adversely affects the organism's immune system. Hackman and Graves (1981) showed that Lyle et al. (1988) reported inhibition of the allogenic transient/minimal rises in blood corticostemnc levels in cytotoxicity (ability to kill cells) of a normal white blood coil line by 60-Hz sinusoiclal electric fields. animals initially exposed to an electric ~cld return to V+fmters (1986, as cited in OTA, 1989) in studies on normal levels within 15 minutes; subsequent field human and dog white blood coils (leukocytes) exposures produc,~cl no further effe,~s. Cerrctelli et al. concluded that extremely low-frequency EMF had no (,1979) found no differences in mortality for electric effects on molecular binding sites, immunoglobulins, or field exposed and unexposed groups of mico injected with the bacterium Staphlococcus nvoEenes after the synthesis of DNA, RNA or protein. However, he exposure to 25 kV/m electric fields for up to 42 days. reported that cultured human colon cancer coUs showed Krueger and Reed (1975) noted no differences in a mitogenic response (induced coil division), and resistsrico to natural killer coils after 24 hours of mortality for mice exposed to an electric field of 100 kV/m (75 Hz) for 21 days and injected with influenza exposure to a 60-Hz magnetic field. Field-exposed coils also showed an increased ability to multiply, virus. Morris and Ragan (1979) and Morris and Philtips (1982, 1983) found no significant differences in compared to unexposed eancor colIs. However, Cohen serum immunoglobulin (an antibody protein) and coil- (1987), in an attempt to duplicate W'mter's (1986, as cited in OTA, 1989) results, found no significant mediated response of mico chronically exposed to low- effects of the fields on the proliferative ability of the level 60-Hz electric fields. Stopps and Janischcwsky same two cell lines. (1979, as cited in Word Health Organization, 1984) discerned no differencos in the general health or in Some of the variability reported for in vitro and in vivo levels of lymphocytes (white blood coils) and other blood coils of workers maintaining high-voltage experimentation is attributed to possible "window effects" for frequencies and intensities of EMF. equipmerit and transmission lines in Ontario, Canada. Finally, Morris ctal. (1979, 1982, 1983, and 1988) Cadossi et al. (1986) reported that inhibition of cellular found no observable changes in either mouse T or B responses occurred upon exposure to low frequency lymphocyte coILs in r~ponse to mitogens (foreign pulsing electric and magnefc fields, resulting in an induced voltage of 10 mV, in contrast to stimulation substance, s) following exposure to 60-Hz EMF for two hr.lday for 30 to 60 days or at 100 kV/m for 30 to 150 seen at other intensities above and below this level. days. These studies show that exposure of intact animals to EMF greater than those experienced under Francoschi et al. (1986) obtained a bimodol response at high-voltage AC powerlines produced no identifiable low phytohemagglutinin (protein that causes clumping deficit in immune coil function. of red blood cells) concentrations in which the effect of EMF was inhibitory rather than stimulatory at specific Data from exposure of isolated immune system frequencies and intensifies. Additionally, mixed components (i.e., various animal cell types) to electric/magnetic fields were found to enhance extremely low frequency EMF in vitro have been responses not seen under electric or magnetic fields con fficting. A series of reports have claimed that alone. exposure of human iymphocytes to pulsed EMF stimulates their response to mitogens (substances that Based on research evidenco presented to date, the induco mitosis-nuclear repfieation) such as Committee finds no clear or consistent evidence that phytohemagglutinin, as measured by DNA synthesis extremely low frequency (ELF) electric and magnetic fields (including 60 Hz) has any physiologically (Hellman et al., 1985; Erailia et al., 1985; C. antini et al., 1986; Franceschi et al., 1986; CAdossi et al., detrimental effects on the blond, immune system, or on 1986). In contrast, Conti et al., (1983, as cited in their cellular elements. These views are shared by Creasey and Goldberg, 1989) using square-wave pulses OTA (1989) in it's comprehensive review of immune (not found ~ssociated with AC transmission lines) at 1 system studies. tO 40 micr0s~conds and 200 Hz reported that lcctin- stimulat~l mitogcnesis (call ~ivision) of human 4.6.5 Gonclusiona lymphocytes (white blood cells) was inhibited. The Committee's summarized conclusions regarding Studies with sinnsoidal ELF/EMF have generally the present state of research of the effects of power shown modest depression of various components of the frequency EMF on endocrine and immune functions immune process. Philllps (1986) found that 60-Hz areas follows: magnetic fields alone, or combined electric and · Researchers have demonstrated no evidenco magnetic fields, inhibit~l natural killer coil-induced cytolysis (coil destruction) of irradiated coliform- of metabolic disorders in lower animals or bacterial coils (Colo 205) in vitro. 4-14 Health Effects of F. xposure to Powerline Frequency Electric and Magnetic Field. v humans exposed to EMF at levels generated and waste products to exit the cell as required to by high-voltage AC transmission linu. support life functions. · Studlea have shown that mtcgratcd The cell mcmbranc also uses . unequal ion functioning of the endocrine systems' conccntratlorm to transmit external signals to the ccll's intcrinr by regulating selective entry of molcculer and enmponcnts govcrning complex biologic torn. Th~ most important of thcac ions arc Calcium functions such as conception, growth and (Ca++), Sodium (Na+), Chloridc (Cl-), Hydrogen dcvclopment, appear to porform normally in (H+), and Potassium OC+). Entrance end exit of there the prcscncc of electric and magnetic ficlds inns occurs through ion channels in the ccU mcmbranc, such as those cxpcricnccd under high-voltage which open and close in response to ionic AC power liner. concentrations and the binding of molcculer such as hormoner. Additionally, corrals membrane-bound · Rcacarch evidence prercntocl to clato appcan enzymer (attached to the cell membrane) take part in to reject the hypothesis that acute or the syntheris of molecules and control initial actions of prolongeel exposure to EMF equivalent to or ~omc drugs. up to several timer stronger than fields experienced under high-voltage AC Most postulatccl modcls point to thc cell membrane as the specific site of interaction bctwccn EMF and the transmission liner results in biologic cell. Adcy (1986) suggerts that EMF interacts with thc disruption of endocrinc or immunologic glycoprotcin rcccptor sites on thc ccU mcmbranc, systems. which arc involvcd in transductlon of signals to the ccH's intorior. However, the effect studicd most 4.7 BioloiJical Mechanisms frcclucntly has bccn the apparent nonlinear pattcro of Calcium (Ca+ +) ion rclcasccl from cells, which results 4.7.t EMF - Cellular interaction following ELF exposure. Normally, Ca+ + flow govcrns a numbcr of bodily pmccasca such as muscle Although certain cpidcmiologic studies show · possible contraction, egg fertilization, and coil division. Ca+ + weak association between EMF and certain discuer, ions flow across ccU mcmbraner in response to signals and certain experimental laboratory studier report from outside thc cell, acting u a rn~scngcr in response biological cffccu, the scientific community generally to specific clcctrochcmical signals of specific intcnsity believed until recently that power-frequency fields (i.e., and frcclucncy. Howcvcr, in vitro EMF exposure 50 and 60 Hz) could not transfer enough energy to studies have shown unusual rater of Ca++ cfflux biological systems to induco any changer. Unlike other (outward flow). Liboff (1983, 1985, and 1987 as cited energy sourcu (c.g, x-rays, microwaver) in Crcascy and Goldberg) postulatcd a mcchanistic powcr-frcclucncy fields do n6t have enough energy to theory based on cyclotron reronanco to explain this break chemical bonds or warm tissuer. Ionization and cffcct, i.c., EMF of the proper frcqucncy and intcusity, warming of tissue arc two wcH catablishcd physical in combination with the carth's D.C. magnetic facld, mechanisms known to cause biological damagc. can affect thc movement of Ca++ ions through Howcvcr, rcccnt studier have dcmonstratcd that undcr channels in the coil membrane. Batwin and Adcy certain conditions power-frequency ficlds can induco (1976, as cited in OTA, 1989) observed a decrease in changes at the cellular level. Thus, certain mechanisms cfflux of calcium from chick brain tissue exposed at (i.e., biological models) have bccn postulated to explain frcclucncy and amplitude windows around 6 Hz and 16 these ohacrvcd changes. An cxccUcnt summary of Hz and at 20 V/m, while Blackman ctal. (1982, 1985a, thcac possible mechanisms is prcaentccl by Crcascy and 1985b) reported an increase in calcium cfflux in chick Goldberg (1989). brain tissuc exposed to a complex sc~er of frcqucocy/amplitudc windows. Bc!lossi (1986) obscrvccl no difference in Ca++ cffiux in neonatal 4.7.2 Call Membrane Models chick brains cxposcd in vitro to uniform or non- uniform static magnetic fields ranging in intensity from The cell mcmbranc is thc boundary of the cell, 2,0001o 9,000 gauss. Blackwell and Rccd (1985)failed maintaining the ccll's structural integrity and to fred any signs of change in cxploratory activity and cobtrolling flows of materials and energy signals into of barbiturate-induced sleeping time in male mice and out of the coU. Under normal conditions, this membrane tammils information from the cell surface exposed to 50 to 400 VIm at 15, 30, and 50 Hz. Both of these parameters can bo affecteel by changes in the to the cell interior, mediating ccUular ructions, and central nervous system (CNS) associatccl with calcium also act~ as a highly selective filter, maintaining an changes. !t is possible the strength and relative unequal concentration of ions on either side of the ccU orientation of the carth's natural static magnctic field membrane. This mechanism allows nutrienU to enter Experimental Studies 4-15 (i.e., 0 Hz) in the laboratory superimposed on the is worse" approach may not be an effective induc,©d AC field contributes to the inconsistency of the approach. data. · Observ~l effects apparently induced by 4.7.3. Conclusions application of EMF may bo influenced by the positionleg of the field relative to the earth's In regard to the mechanism theories evaluated by natural static magnetic field. Also, the researchers to date, the Committee has re, ached a apparent "window" nature of observed effects number of conclusions. These conclusions are requires the evaluation of a wide range of supported in general by the Office of Technology Assessment (1989) in their background paper on the frequency/amplitude scojngs during biological effcas of power-frequency electric and experimentation. magnctk fields: · All of these mechanisml are speculative at · in the cas~ of EMP expolure, it is possibk best. More research is ne~led before these that the 'window effect" replac,~s the conc4:p~ theories can be proven or disproven. of larger (more intense) fields resulting in a Although very si~culative at this time, this grcaU:[ effect than that caused by a less type of research is very important. If a intense field. If such a "window effect" mechanism is established at the cellular level, relationship between EMF then this may support'the results from the frequency/amplkudc and biological reactions positive cpidemiologic and in vivo laboratory should be substantiated, regulations of EMF studies, and aid in the design of new studies. exposure by w,O. ing standards based on "more Table - 4.1 - Summary of Obs-rvations/Conchslons of Experiments to D~termine Behavioral Effects of EMF Exposura, as Detailed in Section 4.3 It~r.~t~. Exl~rimeat Obss~atlmss/Cmsclusions Stern et al (1983) & D~tection of electric fields by rats. Rats detected electric fields as low as 3.0 kVIm and as high as 10 Stem and l_-I!~_- (1985) kV/m. Rato ct al (1989) Exposure of rodents to electric fields to Body hair and whiskers of rodenu vibrated when rodents were determite animals ability to detect field. exposed to electric fields. Sicll and Adcy (1988) Detection of 60-HZ electric fields by Rats exhibited no differences in detection performance during roderoe ·s a function of cireadian rhythm high/low ectivity periods, up to 25 kV/m exposure. activity. Smith et al (1979) Expomra to power-frequency electric No observed effect of power frequency electric fields on general fields to determine effects on general activity of animals tested. acdvity of animals. Graves ct al (1985) Exposure to power frequency electric General activity not affected by power frequency electric fields. fields to determine effects on general activity of chicks. Smith and June:sen(1977) Exposure of mice to 60-HZ magnetic Repo~cd a slight increase in mouse motor behavior (movement) fields to dctermlnc effects on motor in the presence of60-HZ ralgnetic ~eldl; however, increased behavior. activity was not sualaincd, and was obgrvcd only at the onset of field charge. Hjcresen et al (1982) Dclgrmlnation of avoidance of electric Female pigs avoided electric fields of 30 kV/m during their fields by pig,. sleep/rest period. cretin ct al (1982) Avoidance ofelcctric fields by roderos. No avoidance in rat expo0cd to I00 kV/m. Cretin et al (1984) Avoidance of 60-HZ electric fields by Exposure of moderate duration to 60-HZ electric fields (133 rodents. kv/m or less) did not produce tam aversion in rats. Stem and Latics (1987') Exposure to 60-HZ electric fields to Undcr many conditions, fields as high as 100 kV/m were not a determine avoidance behavior by vcry aversire stimuli for rats. rodents. Hackman and Graves Exposure of rodents to electric fields to Reactions to electric fields, sometimes lasting only · few minutes; (1981) Rognbcrg et al determine avoidance behavior. response aUfibuted to recognition of environmental change, to (1981) which the animal quickly adjusts. 4--16 Health Effects of F~posure to Powerline Frequency Eltctric and Magnetic Fields Table 4.1 (Continued) itm,,.lu.,. F, xperimmt Oba~rva6ons/Concluaiom Davis st al (1984) Effect on general ·ctivity level of No changes in mice ·ctivity levels in the presence of DC or AC _r~i_e.,..L.. magnetic fields. ~ st al (1988) Effect oo geeend activity level of Obgrv~d phase shil~ in light/dark-induced cit~·dian rhythm, rodents as controll~l by ciradian dep~nd~l on field mrengths (greater than 25 kV/m) gasonal rhy0~ms and phyaiochemical regulaton. light/dark effect sens:-tivity, and exposure during susceptible phags of the cireadlan cycle. i~low 25 to 3SkV/m, majority of rake showed no cornlatent measurable response, and maximized effect at greater than 100 kV/m. No evideze of a simple dose- dependera reapong for ectivity or respiration. All known potential iong-~rm health risks of high intensity electric fields aUributed to actions as cir~edian regulators. Rogera st al (1988) Activity/behavior patterns of baboons Exposure caused haboons to huddle together, at the onset of field usually expogd to 60=HZ f~ld,. charge. Animals may have huddled to reduce field strength and izreag shielding, aH prol~tive rnspongs to perceived stinmli. Reactions nm eonsis~nt, and temporary, huddling dependent on field mr~ngth. Baboons reacted to field as a threat by huddling at b~ginnlng of exposure, 'learning" after soma time that there was no danger, and returning to normal pal-ms. Re-exposure produced no eftcote. Orr st al (1987) Perfornmnee pau~ms ofbaboom Responses to food rewerd affected by el~:ttic fields upon initial expoMd to 30 kV/m end 60 kV/m expomre; responMs returned to norhal wilhin a few days. Sal~n_o¢rstal(1987) P¢rforman~eofmilexposedto¢leclric l',loeffecilonpcrforman~eofememoryteshw·so~servedin and magnetic fields. adult male rail exposed for 72 hours to · magnetic field of i .0 lau. and an electric field of 30 kv/m. Ho~¢ver, when pregnant female rail were expov. d, e trend toward · reduction in rate was evident in the offaprinR. HUbfANS Stem st el (1983) Dste~tion of electric fields. Humans were able to dete~t electric fields bstwcen 5 kV/m and 15 kV/m. Graham and Coh.t Detection of magnetic and electric fields. Determined that 90% of sealed humans conld dete~t en (1985) field of 9 kV/m, but none could dst~:t magn,~tic field up to 0.4 Is... Tucker and S~iuniU I:)¢t~tion ofmalnetk fields. Out of 200 humans, none cooid deW~t · mainstic field of 7.5 to (1978) 15 C, amb~ral~ st el (1987) Performing· of 26 utility lin¢m¢n Measuremenu of behavioral parroman·e, EEG, ~ scale·, ¢xi~ to 60-HZ EIV!F in the laboratory subjective symptoms and vetions blood chemimry par·mater· over two dais. revealed no stetistical differaneea in exposed vs. control groups. Grabera st al (1988) Double-blind experimanu exposing male Results showed slowing of heart ate, end chengea in central humans to an electric field of 9 kV/m nervous system neeurring soon alter field ongt or offset, tnd · magnetic field of 200 raG. suggesting exposure changes mor~ imporient than exposure Subsequent ¢xpoaure to 12 kV/m, 300 duration. Exposure to higher levels showed no raG. differen~ea between controls and exposed. Exposure may interact with biologicel ey~ma only in limited 'wirelows' of stimuletion, i.e., for · pankuler frequeney, some field intensities produce effect., end th~ intensities above/below the 'window" don't. FJperimentai Studies 4-17 Table 4.2 - Sununary of Obzrvations/Conclusions of Experingnts to Determine Effects of I~4F Exposure on Canecr Initiation end Promotion, as Detailed in Section 4.4 CANCER INITIATION Cohen (1986); Cohen et Exposure o~human lymphocytes (white Expo~d cells did not show any chsnge or mutation in the DNA el, (1986 s&.b);Livingmon blood cells) to 60-HZ i~{F fields. material, i.e., no evidence of cancer initiation. · tal (1986) C. ars~ngn (1987 Exposure of mouse bone-marrow cells to Exposed cells did not exhibit any change or mutstion in the DNA 60-HZ fields. material. Rees~ et al (1988) Exposure of Chinese hamster ovary cells No changes or mutations in DNA material were ob,,-rved. to 60-HZ fields. Nordemon & Hansson Examinstion of human amnic~ic cells, Researchers reported chromosome damage due to magnetic (1987) exposed to sinusoidal and sawtooth exposure. msSnetic fields. Bmghinger et al (1981) Examination of swit~;hyard workers Researchers reported no chromosome damage due to EMF exposed occupationall}, to E,MF. exposure. d'Ambrosio ¢t al (1985) Exposure of bovine white blood cells to Increased percentage of chromosome abcrrstions in bovine 50-HZ fields. periphcrsl blood I},mphoc}'tes exposed for 72 hours. El Nahas & Oraby (1989) Exposure of Swiss male mice to 100,170, No increase in micro nuclear polychromatic er~throcytes (red 220, and 290 kV/m S0-HZ electric fields blood cells) in the bone marrow of mice exposed to i 00 kV/m, for 24 hours (high exposure used to scale but significant inereases in fitice exposed to higher fields. up to humans). CANCER PROMOTION IN VITRO Byus and Adey (1988) Exposure of cells in-vitro to F_.MF, to EIdF exposure increased production of the growth hormone examine changes in hormone production. ornithine decarboxylase (ODe), depending on amplitodc, frequency and exposure duration (increase in ODe production serves as indirect marker to support cancer promotion theory) Note: incrcasc in ODe production doesn't nccesaarily mean that exposed cells will become tumors and F_Jv!F arc cancer pronmters, since a variety of stimuli; which do not play · role in carcinogenesis, can induce ODe activity. In addition, the enhancenest of ODe activity by EMF was much leo than for known promoters~ such as pho~ol esters. Frszicr et al (1989) Tested the hypothesis that magnetic fields Expomre to the magnetic fields did not significantly alter can promote traneformation of initiated tranefornation frequeocies of either initiated or promoted cells as ccHs. compared to sham-cxpo~ cells, an indication that magnetic fields arc not carcinoicnic promoters. Cain ¢t 81 (1987) Exposure of cells in-vitro to E,%4F to Cells exposed to F.J~F were observed to undergo changes in determine changes in enzyme activity. enzyme activity (changes ~imiiar to those caused by known cancer promoters, (e.l~.p phorbol esters}). Fletcher ¢t al (1987) Exposure of cells in-vitro to E/dF to Cells exposed to EIv!F were observed to undergo changes in cell- detect changes in cell-to-cell to-cell communication (changes that arc similar to thog c-used conununication. b}, known cancer promoterst (e.l., phod>oi esters:)). Phillips et al (1986) Exposure of mallSnest human colon Observed enhanced growth rates in malignant human colon cancer cells to EJ~F. cancer cells. Cohen (1987) Exposure of malignant human colon Cohen (1987) could not duplicate !~illips (1986) results. cancer cells to EMF. Adolphe et al (1987) Exposure of malignant human uterine Observed no eftoct on growlh of malignant human uterine cervical cancer cells in culture to S0-HZ cervical cancer cells in-vitro. E!v!F to determine effects on Irowth. C.,oodtnan et al (1989) Exp~ hutnan cells to slnusoidal with Observed induce~l qnantltstive changes in messenger RNA end repetition rates of 60 HZ and 72 HZ. proteins. CANCER PROMOTION IN VIVO Lmst~ et al (1988b) Expomre of rats to 7, 12-dimethyl Observed no significant diffcrenee between exposed and benzene (a) anthrsccne and 60 HZ, 40 unexposcd rate in the number of rats that developed mammary kV/m electric field compared to rate tumors, but did find an jeerease in the number of tumors per exposed Io the chemical onl},. tomor-bcarlnf rats. 4-18 Health Efftc~ of Exposure to Powerline Frequency Eltctric and Magnetic Fields Table 4.2 (Continued) Chandra and St~fano Exposure of mouse mammary ramors to No effect on the growth characterisles of mouse mammary (1978) magnetic fields, and transplanted into turnon cxpos~i to magnetic fields before or after being healthy mice to cuisine changes in transplanted into healthy mice. Thomson et al (1988) iraplantation of !'388 Icukcmia cells in Observ~cl no effect on the incidence or progression of F3gg mice exposed to 60-HZ magnetic fields leukemia cells implanted in mice expo0ed to 60-HZ magnetic to detect effects on the incidence or fields. pmareseion of ,_k_~_ cancer cells. BatirdnandTabrah(1977) ExpoaureofmouMmmors~vivo) to Areport~ddm:reaMinnmuMmmorgrowthinresponMto60- 60-IIZ magnetic fields to dst~t changes !iZ angelic field cximsure. Table 4.3 - Summary of Observations/Conclusions of Experiments to Delermine Effects of i~dF Exposure on Development and Grow~, as Dinlied in Section 4.5 ' p----~_.~-- Expe~_.----,~_.e' Observa6om/Conchisiom TERATOGENIC EFFECTS Graves ct al (1985) E~xpomre of more than 20,000 chick embryos to No obMrved significant change in gromah, devcloprncut, 60-HZ electric fields from. 1 to 100 kV/m to or ovcrsll health of the expoMd chick embryos. detect effects on incubation and development at end after hatchlnm~- Sikov et al (1987) Expend Haasford miniature swine to EMF. Observed realformations in offspring of 0ows exposed to EMF. Durfne et al (1975) Expomre of chick embryos to .001-3.6 kV/m No observed effects on chicken fertility, hatchability, electric fields at 45 HZ to 75 HZ during and survivability, weight gain, and behavior. after incubation to detect effects on fertility, hatchability, mrvlvability, weight gain, and Sandsstem et al (1987) Exposed fe~ilized hen eggs to magnetic fields No significant increases in abnormalities were observed with an asymmetrical uw-aooth waveform. in fertilized h~n eggs, which were expogd during their ~m two days of devdopnwnt. Dmigado ~ al (1982) Expomar~ of chick embryos to low-frequency Abnormalities attested in chicken embryos exposd to pulsed magnetic fields during early development. low-frequency pulsd magnetic fields. (No~: pulsed magnetic fields are not nornmlly auoclatsd with trmnamlmsion lines). lYocda et al (1983) Expogd fertilized chick embryne to pulsed Observed effects in chick cabryes expogd during the ,*l~,~ic and ma..netic ~_,e!d--". first 48 hours ofdevelopmam. Sialnsn et al (1986) Maff~o Exposure of chicken embryos to low-freqnency Neither resurcher was able to replicate the terato~nic et al (1984) pulgd magnetic fields during early develoVment. effects reposed by l:Migedo et a!(1982) and ~ et ol (1983). Berntan et al (198h 19881>) "Henhouse Project" replicating same ext~riment Bas~l on dsla combined for all six labs, remits showed at six indet~ndent laboratories, exposing chick overall increag in proportion of abnormal chick eggs to extremely weak pulsed magnetic fields embryos in groups exposed to puismi magnetic fields. under same set of charscterlaticm and evaluated Exact proportion of abnormalities varied among labs: for same set ofparen~t~rs: egg fertility, cabrye ' two labs repotmt increagd nhnormaliti~s, four labs abnommliti~s fe.ility, in d~velopmena and ~ r~poaed no differences. (No~: pulsmt magnetic fields growth. are not normally associated with transmission lines; exlrapolatlng remits to 60 -HZ effects on bunruns ia highly uncertain. Martin (1988) Exposure of chick embryos to pulsed EMF to Effects observed when expomre occurred during the detect effects during early development. first 24 hours of incubation, but no effects if embryo~ exposed later in their development. Ma~in (1989) Exposure of two different chicken atrsitm to Effects were observed in the Whits Leghorn main, but pulme_d field__.. not in the Ad~or Acre Main. Experimental Studies 4-19 Table 4.3 (Continued) ~ Experimmt Observations/Conclusions REPRODUCTIVI~ EFFECTS Site it al (1983) Exposure of rats to electric fields for three No effect on food and water uptake, but effects on generaflora to ditect realformations and changes growth paltern~ were observed from four to eight weeks in Fowl/development. of aget but not beyond eight weeks. Stuchly it al (1987) Exposure of rats to unsynunctricai sawtenth No statistical diffcrenccs between rats exposed and nminctic fields to detect reproductive effects. those not exposed. Sikovital(1978; 1984) Exposure of rats prior to mating; exposure of No cffects on fetsl leniqh and weight or on pregnant female rats to detect effects on fetal imernal/exlernal realformations; more stillbirths in one devclopmcnt, birth ratu? itc. of the exposed groups. Ronunerein it al (1984) Expo0ure of rats to an electric field to detect No differences observed between the expogd group and changes in incidence of realformed offspring. the unexpoaed (sham) group in the first generation. When fiat generation febred, a significant incrnsc in realformed fcmacs observed in exposed group. No such realformations found in subsequent replication. Lotz and Saxon (1984) Exposure of Rhcsus monkeys to EMF. Observed reduced weight louin males chronically exposed to EMFt but not in females. Smith it al (1981) Exposure of mice to very low electric fields to Neither fertility, litter size, stillborn/live born ratios, nor detect changes in growth and development of sex redo were affected. over 1~400 mice in ]28 litters. lknz et al (1987) Expomre of 3,000 mice to EMF over 3 No effccte observed. generations to detect reproductive/devciopmenUd effects. Table 4.4 - Suetmary of ObservationMConclusions of Experiments to Determine Effects of EMF Exposure on Endocrine Ststcm Function and Immunity, as Detailed in Section 4.6 ~ Exptwimml Observa~om/Conchtsiom HORMONE EFFECTS Cannaciu it al (1977) Exposure of rats to 200-kV/m electric field for Initial increases in ad;cnal hormoOes epinephrinc and six to 72 hours to detect changes in hormone norepincphrine (apparendy caused by exposure) kvcla. stabilized quickly. No further incruses were noted during a subsequent 12-day chronic expire period. No delectable neurolngical or neuro endocrine changes correlated with increased honm>t~l secretions. i:levated neurocndocrinc secretions induced by exposure were within normal response experience for test animals. Frr~ it al (1981) Prolonged exposure of rats to strong 60-HZ Observed siightJy lowered plasma corticosterone levels, electric fields to determine effect(s) on hormone together with slightly lower teslcmcrone and prolactin levcla. levcla. Changes which may be related to alterations in episodic rhylms. Udintsev et al (1986) Exposure of rats to an alternating magnetic field Exposure activated hypothalamo-hypophyseal adrenal (50 HZ; 200 gauu) to determine effects on system (affecting behavior, metabolism, body bonnonal balance. tempereamre, ctc.). Steroid and adreno coaicotrophic hormone (ACTH) were elevated as well as plasma and tiseue-fre~ fitty acids and phoapholipids, but prolonged repeated exposures inhibited activity of endocrine system. Lymangrover it al (1987) Exposure of rat edrenal cells to 60-1tZ fields to E/vlF exposure apparendy stimulated cellular response to detect changes in cellular respormc. ACTH. Kartaahrv & Ivanova (1988) F. xpomare of mice to 60-HZ fields to determine Activation of the adreml and Ihyroid system observed. effecu on adrenal and thyroid systems. Quinlan it al (1985) Exposure of rats to 60-t!Z fields to detect Failed to show any generalized activation of the changes in edrenal system function. hypoOmlamo=hypephyseal-ndrenel system, but an increase in growl hormone production was observed. 4-20 Health Effects of F, xposure to Powerline Frequency Eltctric and Magnetic Fields Table 4.4 (Continued) R~_~_~k~- Exlsslsnmt Observa~ions/Conchssions Michelson & Lu (1988) Exposure of rets to an electric field of 80 kV/m Rats exhibited no pcreiScnc changes in sdrgnal gland for four hour it 71-hour intervals to dstec| function, ind no indlcttion of physiologic or ~hsnsoes in sdrenal ~land function. ncurocndocrinc sireu. Lcung et al (198h) Exposure ofrets to clectrlc fields to detect Rats exhibited · satisdcally significant incruse in the indicstions of ass. ~cidencc end severity of chromodscryorrhca, suggeslng rals exposed to electric field· are subjected to a chronic low-level siren. Michaelson (19877 Exposure of ecclimalizgd and noa4cclln~dzed Sumalncd and in2rndl2n~ 80 kV/m cxpomre may elicit rats to 50, 80, or 100 kV/m sumis~l or subtle relulatory sdjusuncnts of endocrine !cycle, but is~mspted electric fields (60 HZ) for up to five level· were within the threshold for a recognized hours sad dsily exposare repetition for five physiological sUessot, W*ilhin design coasirainU, successive dsys, or st 48-72 hour in2rvmls, to penurhstioa of rat endocrine sys2m. Doubtful thst de2ct changes in hormo~ bahace. proloaged expoaure would cause further dsuimcaud effects,since adaptation stud variability would help mealcrete shy infiucncee on cndncrinc halancc. CIRCADIA~ RHYTHMS Rciter et al (1988) Exposure of rats to electric fields of 10, 65, or Reduced peak nighttime plncal melatonin level· and 130 kVIm from conception to 23 days of age to shifts in clr~adlan n'~ylhma. No dosc-depcndenl ~;~cT chsn,~es in cir~-a~-,, rhythma. relstionships wss observed. Anticgee et al (19877 Exposure of rats to 0, 10, 65 or 130 kV/m G0- Slablc but significantly suppressed mclatonin secretions HZ elcclrlc fields from conception to 22 days to in exposed animals. Peak production of pineel dc2ct changes in circsdlan rhythms su indicated mslatonin in all exposed groups showed · phase shitt by changes in hormone level·. delay of approximately one to two hours, compared to controls. Effect appeared be an 'All or None' response. Changes also noted in pineel sercXonin and N-scetel sero/onin. Groh et al (1987) Exposure of rodcalm to 60-HZ electric field· to Csn upset the pineel gland's clr~sdlan rhythms for the detect changes in circndlan rhythms as evldsnced synthesis end secretion of mHstonin, with recovery by changes in hormonal production. observed in less than three dsys sfier cessation or field cxlKmure in Wilson's ct slt (1986;) study. W'risen et al (1981; 1986) Prolonged exposure of rets to electric fields to Significantly reduced the nighttime rise in mslatoin with dc2ct field-induced changes in circadish recovery observed in Icu than three days after exposure rhythms. cemtion. OTHER BIOLOGICAL FUNCTIONS Cerre2Ui ct al (1979) Loag4crm exposure (i.c., two months) to sn No effects on rat fertility. electric field of 100 kV/m to detect changes in fcfiility. Sikov et al (1984); !knz et al Expomre of rels and swlnc to identical elcclric No changes in rat or swine reeling performsnee or (1987) fields prior to and during ges/stlon to dslcct fertility. ~cld-~,-a,,~ed ~,-,,ps in masln~ and ferlility. Albe~ et al (1984) Exposure ofdeveloplng rats to EMF to Unsblc to demonsirate any differeucee between exposed determine effects, if shy, On growth and and unexposcd dsvclo!~ing rats. Fsm (1980) Exposure of mice to an electric field of 240 Observed reduced growth in female mice, but kVIm to determine field-induced effects on males, no effect on number of offspring or surviving dcvHopmcnt. l~ro~cnies. Sikov ct sl (1984); Prolonged exposure of vsrious Isborstory No diiTcrenccs in animal wclghts between exposed and Konermsn snd Monlg animals to elecuic fields to detect field-induced control groups. Hilton and Philllps (1981) stuibu~c no (1986); Hilton and Phillips effects on development, effects observed duc to climinsting or minimizing (1981) secondary laboratory effects (c.g,, harmonic dlstofilon, spark dischsrge), IMMUNE SYSTEM Hscksmm and Graves (1981) Expomare of Igsl animals to clccUic fields to Transiem/minimal rises ia blood corticeScrone levels in clc4cct chanlcs in blood hormone levels. animals initially expo~P,,d to an electric field returned to nonnsl level· within 15 minutes; subscquenl field exposure produced nu furlher F, vper~'nental Studies 4-2 1 Table 4.4 (Continued) Rammrdter Experimmt Observsfoas/Coaclusions Cctretelli et sl (1979) Exposure of mice to 25 kV/m elecuic fields for No differences in morality for exposed and unexposed up to 42 days, followed by injection with the groups of mice. bacl-rium Staphlo=coccus pyogengs to determine field-induced effects on immune s},stem. Kaleget & R~,d (1975) Exposure of mice to 100 kV/m electric field for No differences in mortality between comrol and exposed 21 days, and injec*,'d with influenza vina to mice. detect effects on mouse immune defenses. Morris and Ragan (1979); Chronic exposure of mice to low-level 60-HZ No changes in serum immunoglobulin and cell-mediated Morris snd Phiilips (1982. electric fields to detect field-induced changes in responses. 1983) antibody production and ceil-mediat~l response to infection. Stepps and Janisehlwsky Examination of general health and blood No differmaces in the i,~neral health or in tcvels of (1979) ch~mimry of workca maintaining high-voltage lymphocy~s (whit~ blood cells) and other blood cells. equipmant and transmission lines to detect ~ld- induced effects. Morris et al (1979. 1982, Exposure of mice to 60-HZ EMF for two No observable changes in either mouse T or B 1983, and 1988) hour/day for 30-60 days or at 100 kV/m for 30 lymphocyte cells. to 150 days, followed by exposure to mitogerm, to detect changes in inunune responses. Hellman et al (1985); Emilha Exposure of human lymphocytes to pulsed EMF Exposure to pulsed EMF stimulated lymphocyxc et al (1985); Camini et al to determine any field-induced changes in response to mitogens such as ph3mlemagglutanin,as (1986); Franc~schi ct al cellular response to mitogens. measured by DNA synthesis. (1986~i Cadossi et al ~1986) Conti et al (1983) Exposure of human lymphocyles to square-wave Leetin-stimulated mitogenesis (cell division) of human pulsed electric fields (not found associated with lymphocytes was inhibited. transmission lines) to evaluate field-induced effects. Faillips (1986) Exposure ofirradhated bacterial cells to 60-HZ 60-Hz magnetic fields alon~. or combined EMF magnetic or combined electric/magnetic fields to inhibited natural killer cell-induced cytolysis of determine cellular immune response. irradiated coilform bacteria in vitro. Lyl= ~t al (1988) Expire of white blood cells to 60-HZ Inhibition of the silogenie cytotoxicity of a normal whit~ slnusnidal electric fields to detect inhibition of blood cell line was observed in the exposed cells. the silogenie c]rtotoxiciiy of the cells. W'tre (1986) Exposure of human and dog blood cells to Extremely low EMF had no effects on molecular extremely low-frequency EMF to determine binding sites, immunoglobulias, or the synthesis of effects on molecular binding, immune responses DNA, RNA, or proloin. Cultured human colon cancer or protein synthesis. cells showed · mltogenic response, and reel·lance to natural killer cells ·tkr exposure. DNA synthesis in human skin fibroblasts was evaluaP. d after exposure to · 60-Hz magnetic field. Cohen (1987) Duplication of Winters (1986) experiment. Observed no significant effects. Cadoui ct al (1986) Exposure of cells to iow-fre~cluency pulsing EMF Inhibition of cellular responses nccurred st an induced to examine for inhibition/stimulation of cellular voltage of 10 mV, in contrast to stimulation observed at responses. other intemities above and below this level; possible 'window effect" for specific intensities of EMF. Franceschi ~t al (1986) Expire of red blood cells to pure and A bimodal respons,c at low ph3aohemagglutinin combined electric and magnetic fields to examine concentrations in which EMF effect was inhibitory stimulative/inhibitory cellular responses. rather th·n stirnulative at specific frequencies and intensifies. Mixed EMF fields enhanced responses not seen under electric or meltnetic fields alone. 4-22 Health Effects of F~posure to Powerline Frequency Electric and Magnetic Fields Table 4.5 - 5timmary of 0b~rvaliona/Concluaions of Ex~rlm~nts to Del,'rmine Effects of EMF Exposure on Biological Mechanisms, as De~ai~l in Section 4.7 EMF-CI~-I_I_ULAR INTEIL~C~FION !lawin and Adey (1976) Exposure of chick brain liamg to I~4F at varying A de~raaa,- in efllux of calcium from chick brain lime at frequencies and amplitudes to det~t chan~es in frequency/amplitodc windows around 6 HZ and 16 HZ, cellular calcium metabolism. and at 20 V/re. Blackanan et al (1982, Exposure of chick brain time to E~IF at varyin~ An increag in calcium efflux at a complex grits of 1985a, 1985b) frequencies and amplitudes to dent changes in fr~'=quency/amplitude ~vindows. cellular calcium metabolism. BeHoui (1986) Expomar~ of ~hick brain time m matic magnetic No o~s~rv~d effect on calc'ium efilux in chick brains kids in an a~mpt.to replicat~ efHux exposed in vitro to uniform and nonuniform static experiments. magnetic fglds ranging from 2,000 to 9,000 gauss. Experimental Studies 4-2,3 References Aaronson, S.A. 1988. Molecular and cellular biological cffects of power frequency electric and/or magnetic ticida. Testimony given in New York State Court of Claims; D. Zappavigrm vs. State of New York and The Power Authority of the State of New York; 345-kV Marcy-South Transmission Line. Adcy, W.R. 1986. The sequence and energetics of ceB membrane transductive coupling to intraceilelar enzyme systems. Bioelectrochem. Biocnerg. 15:447456. Adolphe, M.; Friese, V.; Porter R.; Cabanes, J. 1987. Effect of a low-intensity magnetic field on the proliferation of several lines ofeultured ceLLs. C.R. Soc. Biol. (Paris) 181(3): 282-286. Albert, E.; Cohen, G; Avelllno, L.; Komhousor, G; Yoshioka, A. 1984. Electronmicroscopic observations on rat cerebellum and hippocampus at~r exposure to 60.Hz electric fields. Annual Meeting Bioeleettomagnefca Soc. 04-- 6:52. American Institute of Biological Sciences (AIBS) 1985. Committee on Biological and Human Health Effects of Extremely Low Frequency Electromagnetic Fields: Post-1977 literature review. Technical Report AD/A152731, prepared for the Naval Electronic Systems Command by The American Institute of Biological Sciences. Anderson, L.E.; W"dson, B.W.; Rciter, R.J., 1987. Influence of 60.Hz electric fields on the pincad gland in the developing rat. [Meeting Abstract] Pacific Northwest Laboratory, Richland, WA., US DOF./EPRI Coatactor Review Projcet Resume, Novcmbor, 1987. Batkin, S. and Tabtab, F.L. 1977. Effects of alternating magnetic field (12 gauss) on transplanted neuroblastoma. Res. Cornre. Chem Pathol and Phamutcol 16:351-362. Bauchinger, M.; Haul, R.; Schmid, E.; Desp, J. 1981. Analysis of structural chromosome changes and SCE after occupational long-term exposure to electric and magnetic fields from 380.kV systems. Radiat. Environ. Biophys. 19(4):235-238. Bawin, S.M. and Adey, W.R, 1976. Sensitivity of calcium binding in cerebral tissue to weak environmental electric fields oscillating at low frequency. Proceedings of the National Academy of Sciences 73(6):1999- 2003, June. Bellossi, A. 1986. Lack of an effect of static magnetic field on calcium cf~ux from isolated chick brains. Bioelectromagnetics 7:381-386. Benz, R.D.; Carsten, A.L.; Baum, J.W.; Kuchner, A.V. 1987. Mutagcnicity and toxicity of 60-Hz magnetic and electric fields. Technical Report, Final Report to the New York State Power Lines Project, Wadsworth Labs, E-297, Empire State Plaza, Albany, New York, Bcrman, F_.; Hause, D.E.; Koch, W.E.; Lcal, J.D.; Martin, A.H.; Martucci, G.; Mild, K.H.; Monahart, J.C. 1988a. The Henhouse Project: Effects of pulsed magnetic fields on early chick embryos (Me~ing Abstract). Abstracts of the Bioclectromagnetic Society, 10th Annul Meeting, 19-24 June, Stanford, CT. P.14. Bennan, E; Hotsac, D.!~.; Koch, W.E.; Leal, J.D.; Martin, A.E.; Martucci, G.; Mild, K.H.; Monahart, J.C. 1988b. Effect of pulsed magnetic fields on early chick embryos (Meeting Abstract). Abstracts of the Bioeleetxomagnc~c Society, 10th Annual Meeting, 19-24 June, Stanford, CT., P.72. Blackman, C.F.; Benane, S.G.; Kinney, L.S.; Joines, W.T., and House, D.E. 1992. Effects of ELF fields on calcium- ion efflux from brain tissue in vitro. Radiation Research 92:510-520. 4--24 Health Effects of Exposure to Powerline Frequency Electric and Magnetic Fields Blackms., C.F.; Benane, S.G.; House, D.E.; Joinca, W.T. 1985a. Effects of ELF(l-120 Hz) and modulated (50 Hz) RF fields on the cffiux of calcium ions from brain tissue in vitro. Bioclcctromagnetics 6(1): 1-11. Blackman, C.F; Benant, S.G.; Rabinowltz,/.R.; HouR, D.E.; Joinca, W.T. 1985b. A role for the magnetic field in the radiation-induced cffiux of calcium ions from brain tissue in vitro. Bioclectronutgne~cs 6(4):327-337. Blackwell, R.P. and Reed, A.L. 1985. Effects of electric field exposure on some bodica of CN$ trouul in the mouR. Bioclcctrotrmgnetics 6: 105-107. Bocknmn, R.S. 1989. Effects of power frequency electric and magnetic fields on endocrine sad immune function. Testimony given in New York State Court of Claims; D. Zappavigna vs. The State of New York ~nd The Power Authority of the State of New York; 345-kV Msxcy-3outh Tamsmission Line. Byus, C.V. and Adcy, W.R. 1988. Increaucd omithint dccarboxylasc activity in cuRured coils exposed to athcrmal modulated microwmvcs sad to 60-Hz c!cctric fields. Division of Biomcdic. al Seiencca, University of C_.Mifornia, Riverside, CA. and Veterans Administration Moclical Center, Loma Linda, CA.; US DO E/EPRI Contacton Review Project Resume, October 30-November 3, 1988. C. adossi, R.; Ceccherelli, G.; EraHis, G.; ToreHi, G.M. 1986. Effect of ELF electromagnetic fields on leetin-induced lymphocyte prolifcration [M~ting Abstract]. Abstracts of The Bioclectromagnetics Society, 8th Annual Meeting, 1-5 June, 1986, Madison, WI., P.30. Cain, C.; Mailsrider, M.; Strum, S. and Adey, W.R. 1987. The role of calcium on the effect of pulsexl electromagnetic fields on parsthyroid hormone- relat~cl CAMP accumulation of primary bone coils. [Me~tlng Abstract] Loma Linda University and VA Medical Center, Loma Linda, CA US DOE/EPRI Contractors Review Project Resume, November 1987. California Public Utilities Commission. 1989. Potential health effects of electric and magneSAc fields from electric power facilities. California Public Utilities Commission, 5050 Van Ncas Avenue, San Francisco, CA 94102. Cantini, M.; Cossarizza, A.; !~raani, F.; Cadossi, R.; C~cchereHi, G.; Tcnconi,R.; Gatti, C.; Franccschi, C. 1986. Enhancing effect of low frequency pulsed electromagnetic fields on leetin-induced human lymphocyte proliferation. J. Bio~!ectricity 5(1):91-104. Carsteasen, E.L., 1987. Biological effects of transmission line fields. Elsevier Publishing Company, New York, 1987. Carmacin, R.; Groza, P.; Daneliuc, E. 1977. Effects of a high-tension electric field on the secretion of antidiur~tic hormone in rats. Rev. Romaa. Morpho!., Embryol., Physioiogie 14:79-83. Cerre~Hi, P.; Veicsteinas, V.; Margnnato, V.; Cantone, A.; Malaguti, C.; and Viola, D. 1979. 1000-kV project. Research on the biological effects of S0-Hz electric fields in Italy. In: Proc. 18th Annual Hsnford Life Seienc~s Symp. Richland, WA. National Technical Information Servico, Springfield, VA. Chandra, S. and Sttfano, S. 1978. Effect' of constant and alternating magnetic fields on tumor coils in vitro and in vivo. Proc~:dings of the Eightx~..-nth Annual Hanford Life Seienco Symposium, Richhnd, WA, PP. 436- 446. Chernoff, N. 1989. Addressing the possible human health effects of electric and magnetic fields from electric power lines. A critical evaluation of studies of reproduction and developmental effects in humans and animals. Report: Potential Health Effects of Electric sad Magnetic Fields from Electric Pov~r Facilities, pre~mred for California St-to L~gislature by California Public Utilities Commission and California Dcparm~nt of Health Sciences. September 15, 1989. Experimental Studies 4-25 Coclho, A.M., Jr.; Easlcy, S.P.; Taylor, L.L.; Rogers, W.R. 1987. Effects of electric fields on social stress behavior. Behavioral Medicine Laboratory, Southwest Foundation for Biomedical Research and Department of Biocnginccring, Southwest Research Institute, San Antonio, Texa~; US DOHEPRI Contractor Review Project Resume, November 1987. Cohen, M.M. 1986. in vitro genetic effects of clectrornagncdc fields. New York State Power Lines Project. Wadsworth Centor for Laboratories and Research, Albany, NY 100 pp. Cohen, M.M. 1987. The effects of low-level electromagnetic fields on cloning of two human cancer cell lines (Colo 205 and Colo 320). New York State Power Lines Project, Wadsworth Center for Laboratories and Research, Albany, NY, 69pp. Cohen, M.M.; Kunska, A.; Astcmborski, J.A.; McCulloch, D.; PaskewiLz, D.A. 1986a. Effect of low-level, 60-Hz electromagnetic fields on huma~ lymphoid cells: I. mitotic rate and chromosome breikage in human peripheral lymphocytes. Bioelectromagne~cs 7(4): 415-523. Cohen, M.M.; Kunska, A.; Astemborski, J.A.; McCulloch, D.; Paskcwitz, D.A. 198613. The effect of low-level 60-Hz elecLromagnedc fields on human lymphoid cells. !I. Sister-chromatid exclumges in peripheral iymphocytes and lymphoblastoid cell lines. Mutat. Res. 172(2): 177-184. Conti, P.; Gigante, G.E.; Cifonc, M.G.; Alesac, E.; !anni, G.; Reale, M;Angcletti, P.U. 1983. Reduced mitogenic stimulation of human lymphocytes by extremely-low-frequency electromagnetic-fields. FEBS Lett. 162(1): 156-160. Crcasey, W'~linm A.; Goldberg, Robert B. 1989. Extremely low frequency electric and magnetic fields and cancer: A literature review. Information Ventures, Inc., 1500 Locust Street, Suite 3216, Philadelphia, PA. 19102 - Report prepared for Electric Power Research Institute. R!Y2965-12. Cr~im, J.A.; Lovely, R.H.; Kaunc, W.T. and Phiilips, R.D. 1984. Attempts to produce taste-aversion learning in rats exposed to 60-Hz electric fields. Bioelectromagnetics 5: 271-282. Crcim, J.A.; Lovely, R.H. and Phillips, R.D. 1982. Role of previous exposure history on preference to minimize exposure to 60-Hz electric fields in rats. Fourth Annual meeting, Bioelectorrnagnetic Society, Los Angeles, CA. d'Ambrosio, G.; $c, aglionc, A.; De Bcrnardino, D.; Lioi, M,B.; iannuzzi, L.;Mostacciuolo, E.; Scarfi, M.R. 1985. Chromosomal aberrations induced by ELF electric fields. In: Symposium on Bioelectrical Research in Italy, R. C, adossi, Ed. J. Bioclectricity 4(1): 279-284. Davis, H.P.; Mizumori, S.J.; Allen, H.; Rose. nzweig, M.R.; Bennett, E.L.; and Tenfordi, T.S. 1994. Bchaviontl studies with mice exposed to DC and 60- Hz magnetic fields. Bioelectromagnetics. 5:147-164. Deigndo, J.M.R.; Leal, J.; Monteagudo, J.L.; and Garcia-Garcia, M. 1982. Embryological changes induced by weak extremely low frequency electromagnetic fields. J. Anat. 134:533-551. De. no, D.W.; and Zaffanclla, L.E. 1975. Electrostatic effects of overhead transmission lines and stations. In Transmission Line Reference Book 345-kV and Above. Palo Alto, CA. Electric Power Research Institute, pp 248-280. Durfee, W.; Chang, P.W.; Polk, C.; Smith, L.T.; Yates, V.J.; Phnte, P.R.;Muthukrishnan, S.; and Chen, H. 1975. Extremely-low-frequency electric and magnetic fields in domestic birds. Univ. of Rhode Island Technical Report. Phase I (continuous wave) Kingston, RI. 4-26 Health Efftcts of Esposurt to Powerlint Frtqutncy Eltctric and Magnttic Fields El Nuhu, S.M.; Oraby, H.A. 1989. Micronuclei formation in tomuric cells of mic~ exposed to 50.Hz electric f~ids. Env. & Mol~. Mutag~n. 13(2): 107-111. Emili, 121.; Torolll, 6.; Ceo~her~lli, G.; Doneill, A.; Ferrari, S.; Zucehini, P.;Cadossi, R. 1985. Effect of low- frequency, low-energy, puising electromagnetic fields on the mponR to loain ttimulation of human normal and chronic lymphoe)tie leukemia iymphoeyt~. J. Bioelectrieity 4(1):145-161. Ftun, W.Z. 1980. Long-term biological efteels of very intense 60 Hz electric fieldt on mi~. IEEE Tranuctions on Biomedical Eng. BME. 27:376-381. Fletcher, W.H.; Shiu, W.H.; lshida, T.A.; and Adey, W.R. 1987. A modulated microwave field and'tumor promoters may inhibit cell-to-cell communication and cause an increased insensitivity to cytotoxic lymphokines and tumor necrosis factor. [Meeting Abstract] Loma Lincla University and VA Medical Center, Loma Linda, CA. US DOE/EPRI Contractors Review Project Resume, November, 1987. Florida Electric and Magnetic Fields Science Advisory Commission. 1985. Biological effects of 60-Hz power transmission lines. Technical Report, prepared for the Florida Department of Environmental Regulation, Tallahasscc, Fla., Msrch. Franceschi, C.; lkrsani, F.; Marin, C. 1986. Extremely-low-frequency electromagnetic fields (EMF) affect phytohemagglutinin (PHA)-induced blastogenesis of human lymphocytes [Meeting Abstract]. Abstracts of The Bio4:lectromagnetics Society, gth annual meeting, 1-5 June, 1986, Madison, WI., p. 29. Frazier, M.E.; Reese, J.A.; and Morris, J.E. 1989. Analysis of transforming and/or promoting potentials of 60.Hz magnetic fields in C3H10TY2 cells. [Meeting Abstract]. Pacific Northwest Laboratory, Richland, WA. US DOE/EPRI Contractors Review Project Resume, November 1989. Free, M.J.~ Kaune, W.T.; PhiHips, R.D.; Cheng, H.C. 1981. Endocrinological effects of strong 60-Hz electric fields on rats. Bioclectromagnetics 2:105-121. Gamberale, F.; Olson, B.A.; Lindh, T.L.; Wennberg, A.; Hagman, M.; and Tomquist,S. 1987. Acute effects of ELF electromagnetic fields. A field study on hnemen working at 400.kV. [Meeting Abstract] National Institute of Occupational Health, S17184 Soina, Sweden. US DOE/EPRI Contractors Review Project Resume, November, 1987, Gibson, R,So; and Moroney, W,F, 1974, The effect of cxtr~rnely-low-froquency magnetic fields on human performance: A preliminary study. Technical Report, NAMRL Aerospace Medical Research Laboratory, Pensacola, Florida. Goodman, R; Wei, L.X.; Xu, J.C.; and Henderson, A.S. 1989. The effect of varying signal amplitude r~sults in quantitative changes in transcripts and polypeptides in human cells. [Meeting Abstract]. Columbia University Health Science, New York, NY US DOE/EPR! Contractors Review Project Resume, November, 1989. Graham, C.; Cohen, H.D.; Cook, M.R.; Gerkovlch, M.M.; and Riffle, D.R. 19gg. Effects of intermittent exposur~ to 60.Hz fields on human physiology. [Meeting Abstract] Midwest Research Institute, Kansas City, MO., US DOE/EPRI Contractors Review Project Resume, October 30.November 3, 1908. Graham, C.; and Cohen, H.D. 1995. Influence of 60.Hz fields on human behavior,physiology, and biochemistry. New York State Power Lines Project. Graves, H.B.; Certer, J.H.; Bankoske, J.W.; Cooper, L.; and Poznaniak, D.T. 1977. Perceptibility and Electrophysiological ruponse of small birds to intense 60.Hz electric fields. IEEE Transactions on Power Apparatus and Systems, PAS-97(4): 1070-1073. Experimental Studies 4-27 Graves, H.B.; McCort, W.D.; Austin, M.C.; Reexl, T.J.; Barnick, J.S.; and Jackfert,R.J. 1985. Effects of 60-Hz electric fields on embryo and chick development, growth, and behavior. EPRI Research Project 1064-1 Final Re. port, Vol. 1. Elec~xic Power Research Institute, Palo Alto, CA. Groh, K.R.; 1989. Addressing the possible human health effects of electric and magnetic fields from electric power lines: A critical evaluation of studies of circadish rhythm effects in humans and animals. Report: Potential Health Effects of Electric and Magnetic Fields from Electric Power Facilities, prepared for California State Legishture by California Public Utilities Commission and California Department of Health Services. September 15, 1989. Groh, K.R.; Ehret, C.F., and Readey, M.A., 1988. The actions of high-str~ngth60-Hz electric fields on circadish rhythms in small rodents. [Meeting Abstract], Div. of Biological and Medical Research, Argonne National Laboratory, US DOE/EPRI Contractors Review Project Resume, November, 1987. Hacknan, R.M.; and Graves, H.B. 1981. Corticostorone levels in mice exposed to high-intensity electric fields. Behav. Neural Biol. 32:201-213. Hellman, K.B.; Brewer, P.P.; Fowler, A.K.; Hellman, B.; and Swicord, M.L. 1985. The effect of electromagnetic fields on lymphoeyte function: enhancement of mitogenic stimulation. Technical. Report, Abstract, Seventh Annual Meeting of the Bioelectromagnetics Society, San Francisco, CA., June 16- 20. Hilton, D.I. and Phillips, R.D. 1981. Growth and metabolism of rodents exposed to 60-Hz electric fields. Bicelectromagnetics 2:381-389. Hjeresen, D.L.; Kaune, W.T.; Decker, J.R.; and Phillips, R.D. 1980. Effects of 60-Hz electric fields on avoidance behavior and activity of rats. Bioelectromagnetics 1:299-312. Hjeresen, D.L.; MiHer, M.C.; Kaune, W.T.; and PhiHips, R.D. 1982. A behavioral response of swine to a 60-Hz electric field. Bioelectromagnetics 3:443- 452. Jolley, W.B.; Hinshaw, D.B.; Knierim, K.; Hinshaw, D.B. 1983. Magnetic field effects on Calcium ef~ux and insulin secx~tion in isolated rabbit islets of [.angerhans. Bioelectromagnetics 4(1): 103-106. Kartashev, A.G.; Ivanova, L.A. 1988. Chronic effect of an alternating electric field on the endocrine system of albino mice. Gig Sanit (5):9-12. Kato, M.; Ohta, S.; Shimizu, K.; Tsuchida, Y.; and Matsumoto, G. 1989. Detection -Thishold of 50-Hz Electric Fields by Human Subjects. Bioelectrnmagnetics. 10:319-327. Konermann, G.; and Monig, H. 1986. Untersuchungen uber den Ein~ess statischerdragnet felder auf die pranatele Entionsidung tier maus. Radiologe 26:490- 497. Krueger, A.P.; and Reed, E.J. 1975. A study of the biological effects of cemtin ELF electromagnetic fields. J. Biomettor. 19'.194-201. Leung, F.C.; Rommerem, D.N.; Miller, R.A.; and Anderson, L.E. 1988a. Exporixnental observations in rats exposed to 60-Hz electric fields. Tenth Annual Meeting of the Bioelectromagnetics Society, Stamford, CT., June 19-23, Page 68, Abstract. Leung, F.C.; Rommer¢im, D,N.; Stevens, R.G.; W"dson, B.W.; Buschborn, R.L.; Anderson, L.E. 1988b. Effects of electric fields on rat mammary tumor development induced by 7, 12-dimethylbenzene (A) anthracene 4-28 Health Effects of F. xposure to Powerlint Frequency E!tctric and Magnetic Fields [MeetingAbstract] .. AbstracU of the Bioelectromagne~cs Society, 10th Annual Meeting, 19-24 June; Stamford, CT. P.2-3. Liboff, A.R. 1983. On the uature of the electromagnetic interaction with coih(Mecting Abstract). Tranuctions of the third umual mealrig of Bioelectric Repai~ and Growth Society 3:24. Liboff, A.R. 1985. Cyclotron rr~orane~ mechaniam for electromagnetic energy Ixasmfer to colh (Mocting Abstnuzt). Abgrm:u of Bio¢l~ztromagnctic Society, 7th Annual Me~ng, 1-3. Liboff, A.R. 1987. Cyclotron r~ormnco eff~.U at 60-Hz. [Me~ting Abstract].Detutrtment of Physics, O~,hnd University, Rochester, M!, US DOF_SEPRI Contractor'a Review Proj~t Resume, Novembor, 1987. Livingston, G.K.; Gandi, O.P.; Chattorjco, L.; Witt, K.; and Roti Roti,l.L. 1986. Reproductive integrity of nmmmmlim~ colh expogd to 60-Hz electromagnetic fields. ConUgtor's Final Report, N~v York $ta~ Power Lines Projoct No. 218209, 45pp. Lotz, W.G.; $axton, J.L. 1984. Aaaessment of ELF growth effects in young Rheaus monkeya (M~ting Abstntct). in: Abatract, of Bioele~tromagne~ic. Society, 6th Annul Meeting, p.9. Lyl~, D.B.; Ayctg, R.D.; $heppard, A.R.; Adey, W.R. 1988. $uppr~aion of T-lympho~yte cytotoxicity following exposure to 60-Hz sinusoidal electric fields. Bioelectromagnetics 9(3):303-313. Lymangrover, J.R.; Keku, E.; Hsich, S.T.; Dunlap, W.P.; and Seto, Y.J- 1987. Direct power-frequency electric field effects on mamalian endocrine tissue. Environmental Research 43:157-167. Maffeo, S.; Miller, M.W.; Carstenaen E.L., 1984. Lack of effect of weak-low-frequency electromagnetic fields on chick development. J. Anat., 139:613. Martin, A.H. 1988. Magnetic fields and time-dependent effect~ on development. Bioclectromagnetica 9(4):393-396. Ma~jn, A.H. 1989. Electromagnetic fields and the developing embryo. [Meeting Abstract]. The University of Western Ontario, London, Ontario. US DOE/EPRI Contractor'a Review Project Resume, November 13-19, 1959. Michaelson, S.M. 1987. Neuroendocrine aspects of ELF exposure. Univenity of Rochester Medical Center, Rochester, NY; US DOE/EPRi Contractors Review Project Resume, November, 1987. Michaelson, S.M.; and Lu, S.T. 1988. Electric field exposure and physiologic"stress." [Meeting Abstract]. Department of Biophysica, University of Rochester, School of Medicine and Dentistry, Rochester, NY; US DOE/EPRI Contractor's Review Project Resume, October 30-November 3, 19gg. Morris, J.E.; Frnzier, M.E.; McChnahan, B.J.; Buschborn, R.L.; Anderson, L.E.1988. Effects of 60-Hz electric ~elds on immune response in r~ts [Meeting Abstract]. i~ttelle, Pacific Northwest Laboratories, Richltnd, WA.; US DOE/EPRI Contractor'a Review Project Resume. October 30-November 3, 1958. Morris, J.E.; Ragan, H.A. 1979. lmmunological studies with 60-Hz electric fieIda in: biologic4tl effects of extsemely- low-frequency electromagnetic fields. Proceedings of the 18th Annual Hartford Life Sciences Symposium, Edited by PhiRips, R.D.; Gillis, M.F.; Kuane, W.T.; Mahlum, D.D. pp.326-334. U.5. Department of Energy. M~rris, J.E.; and Phi!llps, R.D. 1992. Effects of 60-Hz electric fields on specific humoral and coHular components of the immune system. Bioelectromagnctics 3:341-347. Morris, J.E.; and Philiipa, R.D. 1983. Erratum. Bioelectromagnctica 4:294. E~cperimental $tudies 4-29 New York Sate Powerlines Project Scientific Advisory Panel, 1987. Biological Effecu of Power line Fields. Technical Rcpofi, prepared for the New York State Power lines Project, Wadsworth Labs, E-297, Empire State Plaza, Albany, New York. Nordenson, I.; Hansson, K. 1987. Clastogenic effects of low-intensity magnetic fields [Meeting Abstract]. Abstracts of Bioelectromagnaics Society, 9th Annual Meeting, p.62. Office of Technology Assessment (OTA) 1989. Biological effecU of power frequency electric and magnetic fields. Background Paper prepared for the Congress of the United States Office of Technology Assessment by Department of Engineering and Public policy, Carnegie Mellon Univenity. 103 pages. Oregon Energy Facility Siting Council. 1990. Repofi on human health effects from exposure to 60-Hz electric and magnetic fields from high voltage power lines. Oregon Department of Energy, 625 Marion Street, N.E, Salem, OR 97310. Off, J.L.; Smith, H.D.; Lucas, J.H.; Moore, G.T.; and Rogcrs, W.R. 1987. Effects of 60-Hz electric fields on operant behavior. [Meaing Abstract]. Southwest Research Institute, San Antonio, Tens; US DOE/EPRI Contractors Review Project Rcsumc, November, 1987. Phiflips, J.L.; Rutledge L.; and Winters, W. 1986. In vitro exposure to electromagnetic fields: Changes in tumor cell properties. int. J. Badiat. Biol. 49(3):463-469. Quinlan, W.J.; Parondas, D.; Lcbda, N.; Pettit, S.; and Michaelson, S.M. 1985. Ncurocndocrinc parameters in the rats cxpo~d to 60-Hz electric field. Bioelectromagnetics 6:381-389. Rcitcr, R.J.; et al. 1988. Reduction of the nocturnal rise in pineal mclatonin levels in rats exposed to 60-Hz electric fields in utero and for 23 days after birth. Science 42:2203-2206. Reiter, R.J. 1990. Effects of light and stress on pineal function. Book:Extremely Low Frequency Electromagnetic Fields: The Question of Cancer. Battelle Press, 505 King Avenue, Columbus, OH 3838 Rceae, J.A.; Jostcs, R.F.; and Frazier, M.E. 1988. Exposure of mammalian cells to 60-Hz magnetic or electric fields: Analysis for DNA single-strand breaks. Bioeicctrnmagnetics 9(3):237-247. Rogers, W.; Coelho, A.; Cory, W.; Easley, S.; Lucas, J.; Orr, J.; and Moore, G.1988. Investigation of effects of 60- Hz electric and magnetic field exposure on operant and social behavior and on the ncuroendocrine system of nonhuman primates. [Meeting Abstract]. Southwest Research InstitUte and Southwest Foundation for Biomedical Research, San Antonio, US DOE/EPRI, Contractors Review Project Resume, October 30- November 3, 1988. Rommercin, D.N.; Kaune, W.T.; Buschbum, R.L.; Phillips, R.D.; and Sikov, M.R.1984. Reproduction and development in rats exposed to 60-Hz electric fields. In: Proc. 23rd Hartford Life Science Symposium, Richland, WA. National Technical Information Service. Springfield, VA. Rosen, L.A. 1988. A review of electromagnetic field health concerns. W/L Associates, Ltd. 120 West Church Street, Suite 4, Frederick, Maryland 21701: pp.23. Rosenberg, R.S.; Duffy, P.H.; and Sacher, G.A. 1981. Effects of intcrmittent60-Hz high voltage electric fields on metabolism, activity, and tomperature in mice. Bioelectromagnetics 2:291-304. Rosenberg, R.S.; Duffy, P.H.; Sacher, G.A.; and Ehefi, C.F. 1983. Relationship between field strength and arousal response in mice exposed to 60-Hz electric fields. Bioclectromagnetic 4: 181-191. 4.-30 Health Effects of Exposure to Powerline Frequency Electric and Magnetic Fields $ai~inger, K. 1989. AddresSing the possible hutnan health effects of 50-60 Hz electric and magnetic fields from electric power lines: A critical evaluation of studies rogarding possible !uaroing and behavioral effects in humans and animah. Repo~: Potential Health Effects of Electric and Magnetic Field From Electric Power Facilities, prepared for CAlifornia Sate Legislature by California Public Utilities Commission and California Department of Health Services, September 15, 1989. SaiT. inger, F.; Froinutrk, S.; McCullough, M.; Phillips, D.; Birenbaum, L.; Coil, G.;and Paduano, J, 1987. Behavioral effects of ELF. Contractors Final Report, New York State Power Lines Project. Contract 1218204. Sandstrom, M.; Hanson Mild, K.; Loulxup, 5. 1987. Effects of weak pulsed magnetic fields on chick embryogenesis. Work with display units g6, International Scientific Conferonce on Work with Display Units, Stockholm, Sweden, May 12-15, 1996, B. Famve:P-G Wideback, Eds. Amsterdam; Elsevier Science Publishers B.V., pp- 135-140. Seto, Y.J.; Hsieh, S.T.; Majean-Cimrqois, D.; Dunlap, W.P,; and Lyrrmngrover, J.R.1983. Food censumption, water intake and growth data on rats chronically exposed to high-intensity 60-Hz fields. J. Bicelectricity 2:197-205. Sikov, M.R.; Montgomery, L.D.; and Smith, L.G. 1978. Developmental toxicology studies with 60-Hz electric fields. In: Proc. 18th Annual Hanford Life Sciences Symposium, Richland, WA. National Technical Information Service. Springfield, VA. Sikov, M.R.; Montgomery, L.D.; Smith, L.G.; and Phillips, R.O. 1984. Studies on prenatal and postnatal development in rats exposed to 60.Hz electric fields. Bicelectromagnetics 5:101-112. Sikov, M.R.; Rommereim, D.N.; Bcamer, J.L.; Buschbum, R.L.; Kaune, W.T.; Phillips,R.D. 1997. Developmental studies of Hanford miniature swine exposed to 60-Hz electric fields. Bioelectromagnetics: 8:229. Sisken, B.F.; Fowler, 14 Maynard, G.; Ryaby, J.P.; Ryaby, J.; and Pills, A.A.19g6. Pulsed electromagnetic fields and nonnll chick development. J. Bicelectricity: 5:250-34. Smith, L.G.; Buschbum, R.L.; and Sikov, M.R. 1981. Effects of continuous electric field exposure on four generations of mice. EPRI Interim Report. Electric Power Re, eareh Institute. Palo Alto, CA. Smith, M.T.; D'AndRa, J.A.; and Gandhi, O.P. 1979. Behavioral effects of strong60-Hz electric fields in rats. J. Microwave Power 14:223-228. Smith, R.F.; and JusteR~, D.R. 1977. Effects of a 60-Hz magnetic field on activity levels of mice. Radio Science, 12:279-285. St:H, M.E.; and Adey, W.R. 1988. Behavioral studies of the mechanism of electric field detection. [Meeting Abstract]. Jerry L. Pottis Memorial Veterans Hospital, US DOE/EPRI Contractors Review Project Resume, October 30- November 3, 1988, Stem, S.; and Laties, V.G. 1985. 60-Hz electric fields: detection by female rats. Bioelectromagnetics 6: 99-103. Stern, 54 and Latles, V. 1997. Failure to train an avoidance response to electric fields, [Meeting Abstract]. University of Rochester, NY, US DOE/EPRI, Contractors Review Project Resume, November, 1987. 5tern, 54 Laties, V.G.; 5tancampiano, C.V.; Cox, C.; and deLorge, J.O. 1993. Behavioral c~____n:'tion of 60-Hz electric fields by rats. Bicelectxomagnefics 4:215-247, Steveas, R.G. 1987, Electric Power and Breast Cancer: A Hypothesis. Am. J. Epiderrnlal. 125:556 - 561. F~cperimental Studies 4-31 Steppe, G.J.; and Janiachewaky, J.W. 1979. Epidemiologieal study of workers maintaining HV equipment and transmission lines in Ontario. Vancouver, B.C., Canadian Electrical Association. Stuchly, M.A.; Ruddick, J.; Lccuycr, D.W.; Robinson, K.; Tan, K.; and Wong, J.;1987. Tcratological study of rats exposed to magnetic fields. [Meeting Abstract]. Health and Welfare Canada, Ottawa, Ontario, Bin-Reaearch Laboratories Ltd., Scnncvillc, Qucbcc; US DOE/EPRI Contractors Rcvicw Project Resume, November, 1987. Sykes, T. and Ping, L. 1990. Possible health effects of electric and magnetic fields from electric power tines: A summary of s~ienti~c studies. Washington State Institute for Public Policy, The Evergreen State College, Olympia, WA 98505. Terrace, H .S. 1988. The effects of power frequcncy clcaric and magnetic fields on behavior. Testimony given in New York State Court of Claims; D. Zappavigna vs. State of New York and Power Authority of the State of New York; 345-kV Marcy-South Tntnsmiasion Line. Thomson, R.A.E.; Michalcson, S.M.; Nguyen, Q.A. 1988. Influence of 60-Hertz magnetic fields on leukemia. Bio~lectromagnetics 9(2): 149-158. Tucker, R.D.; and Schmitt, O.H. 1978. Tests for human perception of 60-Hz moderate strength magnetic fields. IEEE Trans. Bio. Med. Eng. BME- 25:509-518. Ubeds, A.; [zal, l.; Trillo, M.A.; Jimenez, M.A.; and Delgado, J.M.R. 1983. Pulse shape of magnetic fields influences chick embryogenesis. J. Anat. 137:513-536. Udints~v, N.A.; Iranov, V.V.; Moroz, V.V. 1986. Effect of low-frequency magnetic fields on metabolism, and its regulation. In: Biological effects of electromagnetic fields. Their application and standardization. Collection of Scientific Papers. Pushchine 1986. USSR Acad. of Sciences, Res. Center for Biological Studies, Inst. of Biological Physics, pp. 94-108. W'dson, B.W.; Anderson, L.E.; Hilton, D.I.; Phillips, R.D. 1981. Chronic exposure to 60-Hz electric fields: Effects on pineal function in the rat. Bioelectromagnetics 2:371-380. W'dson, B.W. 1988. Chronic exposure to ELF fields may induce depression: Bioelectromagnetics 9: 195-205. W'dson, B.W. and Anderson, L.E. 1990. ELF electromagnetic-field effects on the pineal gland. Book: Extremely low frequency electromagnetic fields: The Question of Cancer. Battelle Press, 505 King Avenue, Columbus, OH 3838. W*dson, B,W,; Chess, E.K4 and Andcraon, L,E, 1986. 60-Hz clectric field effects on pineal mclatonin rhythms: Time courac for onset and recovcry, Bioclcctromagnctics, 7:239-242, W*mton, W,D, 1986, Biological functions of immunological]y reactlye human and canine cells influcnccd by in vitro exposure to 60-Hz clcctric and magnetic ficlds, Technical report, prepared for Thc Ncw York State Power Lines Project, Wadswor*~ Labs, E-297, Empire State Plaza, Albany, New York. World Health Organization. 1984. UnitaJ Nations Environment Programs, and The International Radiation Prot~tion A~ao~iation. Environmental Health Criteria 35 - extremely low frequency (ELF) fields. Technical Report, World Health Organization, Geneva, Switzerland. 5-1 5.0 JUDICIAL ISSUES 5.1 Purpose a~c only those proceedings reported in response to a survey of EEl member companies, American Public Power Association (APPA) members, and the National The sole purpose of this section of the Report is to Rural Electric Cooperative Association (NRECA) provide the Public Utility Commission with information rehtivc to the growing level of activity in regulatory members. and judicial arenas dealing with the clccLric and magnetic fields (EMF) health effects issue as it pertains Since the February 1989 EEl survey, an additional 115 to siting and constructing high voltage transmission proceedings have bccn identified, bringing the total lines. Conclusions, if any, as to whether or not EMF number of reported proceedings involving EMF to 201. poses a health risk should be drawn from the scientific This report does not suppose that aH proceedings have bccn reported or identified but does include data from: evidence presented clscwhero in this report. 5.2 Introduction A. Utilities that proviously have identified an EMF health effects proceeding in response to The cleclxic and magnaic fields (EMF) health effects industry surveys; issue is complicated and surrounded by controversy. Further, it is an issue that beth the scientific B. All investor-owned utilities in the State of community and society as a whole must addms. Texas as wcH as aH but one of the municipal Answers to some questions concerning EMF and health utility systems in the State of Texas having may, in time, come from research, but until the more than 10,000 rcsidcntial customcrs; questions have bccn resolved, the EMF issue will bc addressed in courtrooms, regulatory arenas, and before C. A random sampling of other utilities including legishtivc bodies. As a result, electric utilities can investor owned, municipal, and rural utilities. expect to face the EMF health effects issue each time they seek the necessary approvals for siting and The reported proceedings are tabulated in Appendix C. constructing high-voltage power lines. 5.3.2 Types of Proceedings One of the earliest reported public concerns over the possible health cffccu of EMF from high voltage transmission lines was expressed in the mid 1970s Each reported proceeding has been categorizcd by type during hearings before the Ncw York Stato Public u either (a) siting, (b) zoning, (c) cendcmnation, (d) Service Commission over a 76~-kV transmission line tort, or (c) other. A brief description of each proposed by New York utilities to import hydro-clectric proceeding type follows: power from Canada. Earlier environmental concerns raised about high-voltage transmission lines were a. Siting ProceedinEs. Sitingproceedings are normally related to visual impact or aesthetic issues, corona held before State Public Utility Commissions, Public effect, and audible noise. Also, there were earlier Service Commissions and other state agencies where concerns of safety such as electric shock and utilities are seeking Certificates of Public Convenience and Necessity (CCN) or other similar typos of conductors falling. Some of theso earlier issues along with the health effects issue were brought up during the approvals for proposed transmission line projects. The New York hearings. agency having jurisdiction generally must assess (a) the need for the project, Co) alternatives to the proposal, and (c) the environmental and/or other considerations 5.3 EMF Proceedings of the proposed project to minimize adverse impacts. 5.3.1 General Considerations In the State of Texas, projects requiring CCN certification must meet the requirements of Section The Edison Electric Institute (EEl) has been collecting 54(c) of thc Public Utility Rcgulatory Act (PURA). and disscminating information on procccdings (Ccrti~cates of Convcnicncc and Necessity shah bo involving EMF health cffects since 1984. In thc most granted on a non=discriminatory basis aficr recent EEl survey "Electric and Magnetic Field Cases" considcration by the Commission of thc nccd. , . and Summary Rclx>rt, Fcbruary 1989, there were 86 on such factors as community values, recreational and "reportcd EMF proceedings" involving electric utilities park areas, historical and acsthctic values, throughout thc United States. 'Rcpo~cd proceedings" cnvironmcntal integrity and.. .). Health Effects of Exposure to Powerline Frequency Electric and Magnetic Fields Utilities, in siting proceedings, normally bear the Of the five categories of proceedings outlined in Table burden ofprnving the need for the project and that the 5.1, the Public Utility Commission of Texas will be project will not have unaccoptable environmental involved primarily with siting issues and may expect the EMF hcalth effects issue to be raised in impacts. proceedings involving the CCN proccss for siting of b. Zoning Precccdings. Zoning proceedings arc transmission lines. shnilar in many respects to siting proceedings. Zoning proceedings are normally held before local zoning 5.3.3 Rssults Of tho Rsportsd EMF boards and other local governmental bodies, and, like Procssdln0s siting proce~ings, the utility generaBy bean the burden of proof. a. Siting Proceedings. Of the 72 reported siting proceedings identified in which EMF health effec. ls A major difference between siting and zoning w~re raised, 64 are final with eight pending. proceedings is the type of hearing procedure that is employed. In siting proceedings, st-to agencies In 40 of the rt?orted final proceedings, the authorizing normally convene full evidentiaryhearings, with expert bodies either did not address the EMF issues, witnesses, and generally follow established roles of determined the field levels to be acceptable, or found evidenez. Zoning proceedings frequently involve that the risk of EMF health effects was unproven or of public hearings, hy witnesses, Limited cross insufficient magnitude to prohibit authorization of the examination of witnesses, and loosely applied rnles of facility. evidence. Also, local politics may phy a significant role in the proceedings. In seven of the proceedings the facilities were approved with various stipuhtions by the authorizing bodies that c. Condemnation Proceedings. The major issue in a field measurements were to be taken before and condemnation proe__--,~-__ing is the amount of construction, EMF research would be monitored and compensation owed to a landowner as a result of a rcpofa~l beck on an annual or semi-sumual basis, or utility exercising its right of eminent domain and taking magnetic fields were to be minimized. (Appendix C; the landowner's property for a project. In Siting #s 48, 53, 54, 66, 67, 71, & 72) condemnation proceedings the landowner is entitled to 'just compensation' which usually means full In one proceeding the utility made application to compensation for the property taken, plus damages, if construct and operate a 230-kV transmission line end any, to the remainder caused by the taking. rehted substation. The landowners raised the EMF health effects issue; the utility submitted reports on the d. Tort Proceedings. Tort proceedings are those EMF health effects issue; the landowners dropped all whereby the phintiff alleges that the defendant hu EMF chims and the 230-kV line was approved when a committed a wrongful act which resulted in either new route was agreed upon. (Appendix C; Siting #6). personal injury or other damage suffered by the phintiff. !n one proe__--*~-_-ing involving a 230-kV line, the EMF health effects issue was raised but there were no e. Other Proceedings This category encompuses specific findings related to EMF. An alternate route various proceedings which cannot be easily categorized was proposed. (Appendix C; Siting as any one of the above. Of the 201 EMF proceedings reported in the most In one proceeding involving a 138-kV line, siting was recent updaXu to the survey, 75~$ were either siting or approved and intervenors' request that the utility condemnation relatul. The breakdown by type is provide indenmifs:ation was denied. (Appendix C; shown in Table 5.1. Siting #10). Table 5.1 -Breakdown of EMF related Proceedings in one proceeding, the utility filed a written report on the EMF issue and the line was approved and % of Total subsequently built. (Appendix C; Siting #11) Type N.~k~__ Fmal Pendin~ Psxgeedinns In one proceeding the opponents to · transmission line Siting 72 64 S ~6 presented EMF testimony by an engineer. The utility ZMi,,=a 17 11 6 8 produced a company engineer. The line was approved Ce-~_----'~, 7~ 67 11 39 and upheld by the State court. (Appendix C; Siting To~ 12 7 5 6 O0ser 22 19 3 !1 Total 201 168 33 100 Judicial l~suts 5-3 In one proceexling EMF testimony was presented. The In a 1984 siting procee, ding, Salt River Project, authorizing body made no finding regarding EMF but Phoenix, AZ was denied a Certificate of Environmental denied certificate on other grounds. An amended Compatibility by the Arizona Corporation Committee. application was filed by the utility and the Certificate The denial was based, in part, on EMF health effects was granted. (Appendix C; Siting//16) raised during the proceedings. The main health effect raised was cancer, particularly leukemia and brain In one proceeding an application for a transmission line tumors. (Appendix C; Siting was withdrawn by the utility aRer a rcassessment of the need for the line. (Appendix C; Siting #27) b. Zoning Proceedings. Of the 17 zoning proceedings reported, 11 have been fmalized or settled, six are In two proceedings, the applications were approved but pending. the utility decided not to build the lines. (Appendix C; Siting #s 35 & 36) In two proceedings the zoning councils imposed, either directly or indirectly, magnetic field limits. (Appendix In one proceeding, the authorizing body found that C; Zoning #s 6 & 9) health hazards were conjecture and had not been established. The Supreme Judicial Court affirmed this In two proceedings the zoning councils approved finding ami remanded the proceeding on other issues transmission lines aRer the utilities voluntarily modifie~d not related to EMF. (Appendix C; Siting//38) the routes. (Appendix C; Zoning #s 7 & 14) In one proceeding involving a high-voltage direct In four proceedings EMF health effects testimony was current (DC) transmission line, the hearing examiner presented and the requested facilities were approved. recommended denial of a certificate based on purported (Appendix C; Zoning #s 5, 11, 12, & 16) health effects from air-ions. EMF was not an issue. The utility withdrew the application. (Appendix C; In one proceeding an application to construct a Siting#12) substation was withdrawn without prejudice after hearings were held on the EMF health effects issue. In one proceeding, a State Corporation Committee (Appendix C; Zoning//10) granted a Certificate of Environmental Compatibility for a 500-kV transmission line, but did not make a In one proceeding it was determined that county beards direct ruling on the EMF issue. (Appendix C; Siting are prcempted from setting EMF standards once the //4) State Public Service Commission has authorized the facility. (Appendix C; Zoning #13) In one proceeding involving a 500-kV line the application was approval. Prior to construction, an in one proceeding, the utility, Tri-State Generation and intervenor group (COPE) raised the EMF issue and Transmission Association, Denver, Colorado, was successful in modifying the proposed route. The withdrew its application for a high voltage transmission utility applied for site certificate amendments for four line aRer the proceeding was remanded to the Grand route changes and these amendments have been County Colorado Planning and Zoning Commission to approved. In addition, the utility modified the circuit reconsider cancer promotion and other EMF health configuration from a flat arrangement to a delta effects raised during the application proceeding. arrangement to reduce the magnetic field levels. Mountain Parks Electric, Tri-State's local distribution (Appendix C; Siting #45) customer, built distribution lines to serve the load. (Appendix C; Zoning//17) In one proceeding involving a 500-kV line the siting board denied certification on the basis of the c. Condemnation Proceedings. Of the 78 unavailab ility of forthal EMF standards. EMF condemnation proceedings reported, 67 are final with regulations were adopted and the line was npproved on 11 pending, In 19 0f the proceedings it was not rexhand. (Appendix C; Siting t21) possible to determine from the reports whether the EMF issue affected the awards. However, in 13 of the In one transmission line siting proceeding, the EMF proceedings the c, ouR either refused to decided the issue was raised, witnesses testified, and approval for EMF issue or excluded all testimony on EMF health the line was granted. The intervenors appealed to the effects. In addition, there were nine reported County Planning Commission and the approval for condemnation actions in which the judge or jury made certification was rescinded. The utility rerouted the specific findings that the landowner was not entitled to line. (Appendix C; Siting #37) additional damages based on the EMF issue. In 24 5-4 Health Effccu of Exposure to Powcrlinc Frcqucncy Elcctric and Magnctic Fields procccdings, scttlcmcnts wcrc reached which wcrc 5.3.4 Texas Proceedings apparcndy agrccablc to thc parries involved. Nine utilities operating whoBy or part~.Uy w~th~n the In onc pro,:_~'d__ing, Houston L. ilthtinl~ & Power Stale of Tcx~ rcpoflt. donc or more pro,:_,_-~_:l__ings where Company (I-IL&Pl v. Klcin lndcpcndcm ,School District EMF or other health cffccu issues wcrc raiscd. Thesc (Klein1 (No. 395-7S5, Civil Court at Law No. 1, Hgrm proceedings arc briclly summarizcd below. County, TX), Klein wu avardcd $100,275 in damages. HL&P rclocatcd thc !inc off school propcrty. In Dockc~ 5023 (1984) bcfurc thc Tcxu Public Utility (Appcndix C, Condcmnation t24) Commission, the heating cxamincr rccommcndcd denial of a CCN rcquest by Ccntral Puwcr & Light In anothcr procccding, Louisimm Powcr& Light Company (CP&L), a -ubsidis~y of Central & Company (LP&L) v. MobIcy (1986), the Court of Southwest $crvicc~ Corporation, for a 400-kV dircct AppcaLs of Louisiana hcld it was pcnnissiblc for thc current (DC) transmission linc in southeast Tcxas. Thc trial judgc to forcc LP&L to pay for ehreages to hind if cxamincr detcnnincd the applicant~ had not met the damages wcrc cau/t,d by fear of advcrsc EMF health burdcn of proof to show timt the lmc would not cffccu. (Appendix C; Condemnation #32). adversely affcct thc health of individtmh who livc and work adjaccnt to thc linc. It should bc notcd tl~t th__c d. Tort Procccdines. Of the 12 roportod Tort issuc in thi._s procce. dine wai air-ions, no.J EMF. CP&L proccodings, scven arc final and fivc arc portcling. withdrcw the application. In four of thc fmalizcd procccdings, thc cases wcrc Ccntral Powcr& Light Company (CP&L) filca an dismisscd. (Appcndix C, Tort #s 1, 4, 5, & 10) application for a 345-kV transmission linc bctwccn thcir Lon Hill and Colcto Crcck stations (Docket In onc proceeding, the court found that thc utility could #9305) EMF testimony was f~cd by CP&L. Rcgional not bchcld for the tort of trespus sincc it (the utility) hearings wcrc hcld and environmcatal issues, othcr wu in rightful possession of thc property under the than EMF, surfaccd during thesc rcgion~l hearings. statc condcmn~tion I-w. (Appcndix C; Tort #6) Thc procccding is ponding whilc CP&L invcstigalc~ rcrouto~ to avoid thc environmcntally scnsitivc arcas. in one proceeding, thc landowner claimcd trespass and nuisa~cc as a resuR of EMF. A trial court dismissed thc Houston Lighting & Puwcr Company (NL&P) wu complaint stating that thc Public Utility Commission sucd by Klein Indclgndent School District in was thc propcr forum to resoivc such issues. On condemnation action. Thc School District was awardcxi appeal, the I-ndowncr succ, ccdcd in having the nui~ncc $104,275 in actual damages. During thc pcndency of · nd trc~pus claims rcinstatcd. The cuc sclllcd with no thc appeal of this cuc, thc transmission linc could not firml ruling on thc EMF issuc. (Appendix C; Tort #9). bc uscd duc to an injunction issucd by the trial court. For its opcrational nccds, thc Company rclocatcd this In onc proc, ccding thc utility found ancw sito for a linc. An award of $25 million in punitive damages was proporal substation aitcr local resident. cxprc.scd ovcrturncd on appcal by thc Tcxas Court of Appeals on concorn ovcr thc proximity of thc proposcd substation grounds tl~t HL&P wu rightfully in possc~sion of the to a local high school. (Appendix C; Tort # 11) property under Tcxu Condcnumtion Law. Thc Tcnts Suprcmc Court in essence uphcld the Appeals Court c. Other Procccdinl~s, Of thc 22 reported "other" dccision by rofusing to hear thc procccding. A motion procccdings, 19 arc rmal and thrcc arc pending. for rcconsidcration wu rilcd by the School District and this motion was dcnicd by thc Tcxas Suprcmc Court. Thcrc appears to bc littic commonality among the proceedings classified u "Othcr". Howcvcr, in at least Scott, ct alv. Houston Lilzhtin~ & Power Company cight of thc proceedings, hearings of various types (No. 87-058967, District Court for Harris County, wcrc held, thc EMF issues wcrc prcscntcd, and a Tcxaa) is a personal injury lawsuit fdc, d against HL&P variety of f~cilidc~ wcrc Ipprovcd. in thrcc of thc on Dcccmbcr 14, 1987. Mr. Scott's home ls locatcd procccclings, rc-olutions, ordin,,ncc~, and/or initiatives adjaccat to thc right-of-way of a double circuit 345 kV wcrc puscd to prohibit construction of transmission transmission line (Singlcton-TombaB-King). Mr. Scott lines abovc certain voitagc lcvcls. Bricf dcUi~ of thcsc sufTcrcd from a brain tumor diagno-cd in 19g7, and it and thc rcauaindcr of thc procccdings .so found in i~ allcgcd that the tumor was cithcr causcd or aggravatcd by thc clcctric and magnetic fields Appendix C-Othcr. cmmuting from the trimsmission linc. Mr. Scott, et furthcr alleges that HI.,&P is liabic for his tcnnin~l health condition becausc the Company (HL&P) knew Judicial Issues 5-5 of potential health hazards resulting from EMF since mcthods of insuring system reliability ~o the Austin area 1975. This case has been dismissed. while minimizing human exposure to EMF. Rainwater v. Houston Lighting & Power Company In 1979, LCRA applied for and obtained a CCN for a (No. 87-058968, District Court for Harris County, 345-kV line and associated substation. The Public Texas) was filed on December 14, 1987 by Mr. Scott's Utility Commission adopted a hearing examiner's sister, Bcvcdy Scott Rainwater. Ms. Rainwater sold ruling that then: wen: no proven hcalth effects and HL&P an easement for the Singlcton-Tomball-KAng added a provision that the Commission could amend or Transmission Line Corridor, and charged HL&P with n:vokc the Ccrti~cate if future research shows that real estate fraud. Ms. Rainwater alleged that HL&P cxposurc to cicctric fields causes adverse health cffccts. knew of potential health effects resulting from transmission line electric and magnctic fields and that In addition, LCRA has had a total of 11 fight-of-way had HL&P revealed that information to her, or had she (ROW) condemnation proceedings involving EMF. known about it, she would not havc sold thc eascmcnt Eight of thc cases settled out of court and three went to rights. This casc has been dismissed. trial. LCRA prevailed in all threc court actions. During the proceedings much EMF testimony was Texas Electric Service Company1 v. Robert Carl presented by both sides, however, then: was no Berger, ct al (Cause No. 85A-216, 97th Texas Judicial mention of EMF in any of the final rulings. District), May 1987, involved a condemnation procccding for a right-of-way/easement to construct a In the early 1980s, City Public Service (CPS) of San 345-kV high-voltage transmission line across Bergcr's Antonio was involved in condemnation proceedings property. The primary issues were land values relative involving a 345-kV line in the Stone Oak area. Alter to the actual land taken and land values before and alter the landowners raised the EMF issue and presented the taking of the land and compensation due Bcrgcr, et expert testimony, the landowners were awarded an al. amount slightly lcss than the initial valuation of the condemned property. CPS has had seven other ROW Berger's attorney failed in his attempt to introduce condemnation cases involving EMF setfled out of EMF health effects testimony into evidence. The jury court. ruled against Berger and awarded an amount less than the value of the land set in the original condemnation Brazos Electric Power Cooperative, Inc. (BEPC) hearing. reported three condemnation proceedings involving 138-kV transmission lines. In one proceeding (Brazos Matador Cattle Company v. Southwestern Public Electric v. TheIres Ray), the utility offered $42,000 for Service Company (No. 1981 110th Judicial District a 4.2 acre parcel out of a 100 acre tract of land. The Cou~, Mo~ey County, Texas) May 1, 1987. The landowner claimed that the total 100 acres was EMF issue in this condemnation proceeding involving damaged and asked for $1,000,000. Testimony on the TUCO-Oldaunion 345-kV line was whether or not EMF issues was pn:sented. The jury set the value of the electric and/or magnetic fields extended bcyond the the 4.2 acres at $77,000. In the final order, then: was limits of an casement onto adjacent lands. The jury no mention of EMF by either the jury or the judge. In found that the fields do extend beyond the easement the other two proceedings (BEPC v. Maddie and BEPC Limits. The Court found the line to be constructed in v. McAHum) trial juries awarded the iandownen accordance with aH applicable ststc and federal slightly less than originally awarded by the special rcgulations and declined to award damages for EMF commissionen courts. No mention of EMF was made trespass or grant an injunction against operation of the in either of the final orders. line. Bluebonnct Electric Coopcrative (BBEC) reported one The City of Austin and the Lower Colorado River condemnation proceeding involving right-of-way for a Authority (LCRA) engagcd in a study to dctcrmine new 138-kV transmission line (BBEC v. Colhoun). configurations for a proposcd 345-kV loop around the Discovcry is in progrcss with no trial date set. City of Austin. The original plan for a 345-kV loop was vetoed by the Austin City Council. A second plan 5.4 COFICIUSiOFIS utilizing 138-kV lines was incompatiblc with LCRA's system rcquixx,-ments. Thc involved utilities havc been The EMF issue is a complex issue that is being studied directed by ~e Austin City Council to determine by electric utilities, scientists, rcgulatory agencies, and public health officials. To date, thc rcsearch has not 1 Now a division of Texas Utilities Electric dcmonstrated any adverse health cffects associated with Company. cxposure to EMF. Because of its complexity, thc EMF Health Effects of Exposure to Powerline Frequency Electric and Magnetic Fields issue will continue to be raised in future legal , Stay abreast of current and emerging proceedings. This is p~rti~uhriy true in siting and zoning proe__,x'd__ings where reguhtory and/or local research; governmental bodies are required to consider impacts on public health and welfare. · Continue to plan and sis= facilities in accordance with the rules set forth by the In future condemnation proceedings it is expected that Public Utility Commission of Texas; opponents of high voltage powerlines will continue to emphasize the public fear issue. This issue evolved * Develop an awareness of the public's based on the contention that the electric and magnetic curiosity about EMF health effects; fields extend beyond the edge of the right-of-way and pose health risks or that the public's fear of health risks · Develop public education and information associated with EMF affects the market value of the pregl~ms; and property. · Continue funding EMF health effects research Since few tort actions with EMF implications have been projects either through organizations like the flied to d_s_t_,,, it is difficuit to predict whether tort litigation will increase in the future. Tort litigation has Electric Power Research institut~ (EPifl) or the potential to expose utilities to the greatest liability through other nationally recognized research sinco tort proceedings need not be restricted to organizations. propused or newly constructed high-vohage powerline projects, but may involve putential EMF health effects Regulatory and governmental agencies also need to related to distribution facilities. stay abtr~st of the emerging research in order to make fair and prudent decisions in siting and zoning Because of the EMF health effecU issue, utilities face proceedings where the EMF health effects issue is the future with uneeminty, and can only attempt to raised. Likewise, public health officials must develop minimize EMF litigation. Utilities, by necessity, need an awareness of the issue such that fact can be to: separatui from fiction in deciding if EMF poses a risk to public health. 6-1 6.0 REGULATORY ISSUES National Council on Radiation Protection and 6.1Introduction and Background The term "limit" is a narrower term and refers to the The increase in demand for electricity and the need to establish new corridors for transmission lines was faced level which is restrictive, e.g. speed limit. A Limit is by opposition in the 1970's who alleged that high- precise and can be measured, such as an electric field voltage lines might have an effect on the health of near- limit in kilovolts per meter. If a standard or limit is by residents. A notable example was the New York recommended or optional, this is usually called a Power Authority experience, which resulted in guideline or guidance level. Guidelines are not heightened interest and expanded research into possible dcvclopexi in a formal rulemaking procedure. There is health effects due to electric and magnetic fields (EMF) usually no authority for restriction or penalty if the caused by transmission lines. These activities and guideline is not used and other approaches are taken. cases involving controversies, contested lines, and litigation led to consideration of establishing Criteria or criteria documents may refer to coUcctio ns environmental cxpesure standards or limits. Electric of information, data, or evidence which can bc used to field standards were also set as early as 1976 in make decisions or informed judgcmcnts about Minnesota. standards, guidance, or limits. Criteria documents have been compiled by groups such as World Health In this chapter, the Committee examines actions that Organization (Vv'HO) and National Institute for have been taken in response to the call by some for Occupational Safety and Health (NIOSH). standards to protect the public health. Thesc efforts at regulating the electric and magnetic fields will bc In the EMF context, a standard can be established reviewed to determine if they are related to health which specifies restrictidns on transmission lines. effects. These may include additional siting criteria, specifications for corridors, height and other design 6.2 Standards and Limits features, and lcvcis of calculated or measured electric and magnetic field strengths. Those field strength lcvcls are oRcn referred to as the limits. It is important here to deCree terms that are used in regulatory language. Thc use of terms may vary in 6.3 General Rationale for Health- different jurisdictions; statutes and regulations may contain specific definitions for certain terms and should Based Exposure Standards be referenced. A standard is an acknowledged measure for comparison; it may be qualitative or A health-basexl chronic exposure standard restricts quantitative. Examples of qualitative standards are exposure to cnvironmcotal agcots which may cause design specifications (such as steel construction), adverse health effects in some or all populations. requirements for training, or credentials of personnel. Efforts at sating this type of EMF standard have bccn Quantitative standards include distances, made in reccut years to alleviate concern about l. hc concontrations, or time periods. Standards may be possible health effects from chronic exposure to EMF. included as part of a quality assurance or performance The basis for these standards is of particular program and may be adolXcd voluntarily. When impofiancc when considering regulatory implications. standards are adopted by a governmental body, they The Commitxcc examined the existing and developing become mandatory and restrictive and may result in standards to sec if such an effort is applicable. The penalties as a consequence of govcrnmcot cnforecmcnt objective was to determine if a health-based exposure authority. standard has been established, or can be justified, as a possiblc consideration by the PUC. Standards are derived in a variciy of ways. A pionccring group or industry may chcosc a design or Health-based exposure standards can only be protocol that others follow. Standards may be alter health risk assessments occur. (Soc previous gcocratcd by consensus; i.e., represcotatives of sections regarding hadth effects, risk and intereatcd groups (such as scicmists or coginccrs) cpidcmiology.) A risk assessmcot compiles and conycoo to analyze the facts, discuss, and debate until analyzes available knowledge, scicoti~c froclings and agrccmcnt is reached. Such groups include the clinical ruuRs, and other evidence to make a American National Standards Institute (ANSI) and the conclusion about the health outcomes or deleterious effects from an (>-2 Health Effects of F, xposure to Powerfine Frequency Electric and Magnetic Fields then given to possible b~nc~U which may outweigh the Another timitation in using current data to generate a ncgativc outcome or risk, and the realization that "zero health-based standard is that no dosc has bce. n derived. risk* cannot bc the goal for cnvironmcntal agents. Thcrc is no consistr, nt rcaponsc rchiting a metric of the Attention is also dircctcd at acthods and technology ticIda with a health outcomc, and no dccision can bc avaihiblc to achieve a dcsircd level or condition, madc about safe or unsafe levels or ticIda. practical restrictions, and fmancial constraints on thc managcmcnt of thc risk. This process is inherent in Onc cannot employ thc methodology uscd for other public hcalth policy and rcguhitions; such as cnvironmcotal pollutants nor justify modcls similar to chlorination of public drinking water supplies to rcducc thosc used for ionizing radiation, microwavc and radio the risk from watceoornc discues, or thc cstablishmcnt frcqucocy radiation, or chcmieal cxposurcs. For of programs to immunize chiidrcn and adults. cxamplc, it would bo vct7 convcnicnt if thc same approach used in sctting standards for lead could bc Thc decision for how to manage a risk may involvc used. Lead is a heavy rectal, commonly found in the setting a standard for cxposurc. This standard can bo environmcnt of an industrialized socicty. Various derived by applying a margin of safety or utiliTing standards for cxposurc to lcad have boca act which thrcshokls to gcocratc limits with accclaa51c risks. If have boon dcfived from clinical tests and observations thc standard is ugd in thc rcguhtory manner, it is of patients exposed to had. Thcrc arc occupational formally proposed, offcrcd for scrutiny and public cxposurc limits for many diffcrcnt lead compounds. commcnt, and adoptcd. This formal rulcmaking Thcrc arc limits on the Icad contont in drinking watcr. proccas is specified in administrative codes for thc Thcrc arc blood concontration lcvcls of lead which arc political jurisdiction at fedcral, state, or local Icvcis. used to trigger decisions in the work phico, or to signal investigation of childrcn exposed to agents such as 6.4 Scientific Basis for EMF lead-based paint. It is not clear whcthcr additional Standards EMF re, search will complctc thc missing information, fad another cause, or if now mechanisms ~ gcncratc a diffcrcnt framework of standard setting. Tlg 60-Hz Thcrc arc somc complicating factors in the prcse, nt statc data cannot at this time bo adaptcd to this framcwork of knowledge about 60 Hz EMF which prccludc string because thc risk is nicthcr known nor quantifmblc, and health-basod standards at this timc. The usual no dose ruponse has been found. assumptions uscd to act litnits (both occupational and population standards) arc not valid in this cuc. Thcrc 6.4.1 Occupational Standards arc insufficient and inconsistent data on which to establish a health risk. Thcsc arc dctailed in prcccding Thc rationale and justification for occupational sections of this rcport and in sevcral litcraturc rcvicw standards can bc based on assumptions and the use of papors' biological models that have bccn dorived from the scientific cxporimcnts. These modcls may bc used to The tmsic usumption in protcction from exposure to calculatca dose and dose rcsponsc, i.c., a biological most environmcntal hazards, such as chemicals result, usually in avcragc-sizcd adults with normal (bcnzcnc, pcaticides), or physical agents, (ionizing ranges of physiological functions, such as brcathing radiation or ultraviolct radiation) is that incrcascd rates or cxcrction rates. This se-callcd 'healthy cxposurc, either to higher levels or for longcr duration, workcr' is usually assum4xl to bo working an 8-hour increases the risk and severity of deleterious cffccts. day and away from thc work phco for thc rcanaining 16 Morc exposure is worse; less exposure is better. This hourl in a day. '['his httcr portion of the day aRowa the may not bc the casc in EMF cffects based on present body to rccuporatc and repair biological damagc or scicnti~c cvidencc. It has bccn hypothesized that thcrc insult that may havc bccn causcd by thc chemical or could bo 'window effects* for different frcquencies physical agent. and diffcrcnt amplitudes. Thesc have becn dcmonstratcd in human exposures as well as in vitro In thc casc of chcmicah in thc work phicc for which cxpcrimcnts in which a rcsult was obscrved at onc lcvcl dclctcrious or lethal concentrations arc wcll- and disappcare, d at others, oRcn highcr or strongcr than documentcd, thc cxposurc to a worker is limited to a thc rcactivc one. !t thcrcforc b~comes possible to select a standard which could bo lower, but is in the spcci~cd concentration which may includa a time response 'window' and could result in diffcrcnt duration for thc cxposurc to thc chcmical or cxposurc groups. A possible implication for the carcinogen. in industrial hygienc, this is called a ~ rcguhtory issuc is that prohibition of highcr-kV lines wcightcd-avcragc. Good examples in the litcraturc arc may forcc construction of additional lower voltagc lincs found for lead, benzcnc, and many othcrs. for thc load demand reliability. Those additional lines may 'expose* a greater numbcr ofpcoplc. Indication of the complcxitics in 60-Hz standard sctting should bo noted when onc comparcs thc rnctrics and Regulalory issuts field characteristics detailed in the literature'~ Some 6.4.2. Environmental Standards for the re. searchers use a single measure of field strength or General Public flux density; others attempt an avenging; others Itpply more elaborate engineering approaches. No data as yet indicates what "number" best det'mes the field for Population-based standards for the general public have purposes of establishing health effects. The advances been used for various kinds of exposures. Examples of in exposure assessment and characterization of the these kinds of standards include limits set on chemical constituents in drinking water, air pollution indices fields has been noted in recent years and may contribute to a greater understanding of the science. including emission levels from industries and automobiles, asbestos, radon gas in homes, and Another technique in setting occupational standards is contamination levels in food and produce. These to establish a ceiling or upper limit. Of impor, anco in standards may take into account the individuals in the this context is the standard rg~commended by the population who are more sensitive to the environmental American Conference of Governmental Industrial agents. The most sensitive are usually fatuses, young Hygienists (ACGIH) in "Threshold Limit Values (TLV) children, the aged, ill, or handicapped. Such standards and Biological Exposure Indices for 1989-1990." may assume that the exposure is continuous (24 hours ACGIH added a section on "Extremely Low Frequency per day, 7 days per week) and chronic (will occur over (ELF) Electric Fields" up to 30 kHz, and recommend the next S0 years or a lifetime). The standards must 25 kV/m for a routine occupational limit. They also rely on biological evidence, quantification of dose, have an existing 'Static Magnetic Fields' limit of 600 extrapolation to the popuhtions at risk, hboratory gauss, whole body exposure. These field levels are ruults, and measurement of what is in the environment. Evidence of this type is incomplete and levels to 'which it is believed that most workers may be exposed repeatedly without permanent biological insufficient at this time to support EMF standards. effects..." They are not to be used as a demarcation Setting a standard based on concorn about a rare between safe and dangerous levels. condition, such as childhood cancer, in the abscnco of a biological mechanism and dosimetric response, does The ACGIH stated in 1989 that they had "not found not provide any real protection to the public. sufficient information to propose a TLV," but will continue to study "Low Frequency Magnetic Fields' 6.5 Existing Standards and "Static Electric Fields." Subsequently, the ACGIH Physical Agents Gommittlz prepared a Notice of 6.5.1. International Standarde Intended Change for 1990-91, concerning 'Sub-Radio frequency (30 kHz and below) Magnetic Fields, and (See Table 6.1 for a summary.) Sub-Radio frequency (30 kHz and below) and Static Electric Fields." That proposed exposure vadue was This discussion of international work will focus on the adopted in the 1991-92 edition. It limits worker selected countries because of the long-standing parallels exposure to a permissable magnetic flux density of I with the United States in the areas of environmental mT at 60 Hz. policies. Although there has been some interesting effort in the Soviet countries, it is difficult to rehte to western circumstances or extrapolate how the standards are implemented and enforced. It is possible with the dramatic political changes that in a few years much Table 6.1 - Recent ln~rnational Standards for 60-Hz Fields Sta~sdard Electric Field MaiMtic Fscld IRPA (apublic) 5 kV/m 1000 n~3 IRPA (occupational) 10 kV/m 5000 mG United Kingdom (all~ 10.233 kV/m 1630 A/m ~20 G *~ Australia ( public--all day) 5 kV/m 0.1 mT Australia ( publlc-limi~d day) 10 kV/m I mT Australia (occupational--all dsy} 10 kV/m 0.5 mT Australia (occupational limited da~,) 30 kV/m 5 mr** * Flux density in tissue ** Maximum 2 hours per day 6-4 Health Effects of FJposurt to Powerline Frequency Electric and Magnetic Fields more open dialogue will occur about their work. restrictive considerations appear necessary. In the abwaac,~ of ad~quat~ data and in lieu of ~onclusions a. The International Radiation Proteaion Association about possible long-term cffccts, precautionary (IRPA) issued 'Interim Guidelines on Limits of measures arc suggested, i.e., keeping exposurc~ Exposure to 50/60 Hz Electric and Magnetic Fields'in ALARA (As Low As Rcasonably Achievable). A 1990 which arc bucd on protection from induced subscqucnt Environmcntal Health Critoria69: Manetic currents and 'eatabllthed effects of exposure.' IRPA Fields (1987) expandcd the discussion, but was unable stated that an association with cancer wu not to make conclusions about chronic exposure; it was considered because present data does not provide 'any completed prior to the more recent studies of magnetic basis for health risk assessment useful for the fields. These criteria detail the scientific data base and development of exposure limits." (IRPA, 1990) rationale for re. commendations, and may contribute to evidence for possible rccommcodations or standards for b. The United Kingdom approvcd a guidance standard the IRPA, NRPB, or NCRP. in 1988. The National Radiological Protection Board (NRPB) chose not to differentiate between occupational d. Australia cstabllthcd *Interim Guidelines on Limits versus the public exposure at low frequencies because of Exposure to 50/60 Hz Etcctric and Magnetic Fields of the absence of · 'scientific justification.' The (1989)' expanding groups but following the rationale NRPB guidelines cover the frequency range up to 300 stated in the other examples nutcd. (Commonwealth of GHz and arc bucd on the specific energy absorption Australia, 1989) rate and thermal cffcets for the higher frequencies. At frcquencico below 30 MHz, induced current is 6.5.2. tJnjtodStato. considered; the elcctrlc and magnetic fields arc derived separately in root-mcan-sqnarc values. The electric No national standards exist for the regulation of long- field strength is specified in V/m and magnetic field term health effects from 60-Hz electric and magnetic strength in A/re. For 60 Hz, the levels arc 10233 V/m fields, nor does a federal agency have a clear mandate (10.2 kV/m) and for 50 Hz, 12280 V/m (12.3 kV/m). or spcci~c authority to rcguiate or take the initiative. These arc similar to the 9 to 10 kV limits in Table 6.2 In the absence of federal direction, the states have for states in the United States. reacted in various ways to the issue. The magnetic field strength is limited to 1630 A/re. Interest at the Congressional level has bccn evidenced The magnetic flux density in tissue is also given as an by hearings in Washington to receive t~stimony from alternative to the equivalent magnetic field strength; the federal agencies, research scientists, industry that limit is 2rot or 20 G. representatives, and concerned citizens. Those hearings include the House Subcommittee on Natural The NRPB stated that they agree with the International Resources, Agriculture Research and Environment Nonionizing Radiation Commltt~c (INIRC)of the IRPA Hearing on EMF Research Bill HR 4801 (July in its conclusions, that: 1990); House Interior General Oversight and 'there is at present insufficient biological and Investigation Subcommittee (March 8, 1990); House Subcommitte~ on Water and Power Resources (October epidcmiologieal data to make a health risk assessment or ~ven to determine whether there 6, 1987). The hearings have involved extensive is a potential hazard to health with regard to argument on the research needs, the limited fedettl athcrmal effects of clcctromagnetie fields.' support and the role of the federal agencies. They further explain that 'the experimental evidence Environmental Protection Agency (EPA) is the federal for biological studies is oftml statistically weak and agency with authority to establish national proves difficult to reproduce." Regarding the environmental standards for water and air pollution, population-based and occupational cpidcmiologic toxics, and hazardous wastes. The Agency does not studies, they conclude that if the risks arc real, they arc currently have a legislative or judicial mandate within the range regarded as 'tolerable' and 'should regarding cxposurcs to EMF as it has for r~gulating other environmental areas, such as air standards under not unduly concern individuals.' (NRPB, 1989) the aegis of the Clean Air Act, or to set other standards for hazardous waste or water contamination. Nor has c. 'The World Health Organization (WHO) published the Agency been placed under court order, as in the Environmental Health Criteria 35: Extremely Low chemical carcinogen cases, to set standards based on Frequency (ELF) Fields in 1984. The document health risks, and scientific health-risk assessment. On reviewed the existing worldwide research and reached Scptembcr 29, 1988, EPA announced the phasing out conclusions on the existing studies. For intermittent of its activities in nonionizing radiation because of exposures to electric fields below 10 kV/m, no agency priorities and limited resource. However, the ReguiaWry Issues staff more reamfly undertook a review of the scientific Some national associations have also cxprcasai concern literature and consideration of policy implications that about the EMF issue and urged a greater federal role. may lead to more federal involvement, and funding Others havc commissioned study groups or allocations. commL-~c_n_, have proposed policy statements, or passat rc~olutions on the issue. These groups include The Dcpa~ment of Energy (DOE) has been the only the National Association of Rcg uhtory Utility federal agency to phy a consistent role in the EMF Commissioncn, the Association of thc State and issue for the many years since intercat heightened. Tcrriterial Health Officers, National School Boards DOE has a compcRing role because EMF concerns Association, the Conference of Radiation Control involve the energy distribution and use in this counb'T. Program Dirccton, Inc. and othcrs. Much of thc Thc DOE has made significant contributions by concern arisc, bccausc of the divcnity of rcsponsc~ in maintaining funding of basic research and monitoring thc states and potential problcms with confficting the scicncc. ovcrlap, or incompatible recommendations or standards. Othcr federal agencies have potential roles if certain health cffccts should bo dcmonstratcd for EMF. Thc The National Elcetric Safety Codc (NESC) has bccn Dcpartmcnt of Transpo~ation is evaluating "maglcv" cmploycd for many years by thc industry and many trains which travel via high-spccd magnetic lcvitation. state rcguhtory agencics as a standard for reducing Thc Food and Drug Administration has bccn a lead potential shock hazards from transmission lincs. The agency in maUcrs involving radiation and in rcguhting NESC is used in dcsign criteria for lincs by placing consumcr products, both ionizing and nonionizing limits on induced current levels of 5.0 milliampcrcs radiation with potential cxposure to the public, such as RMS (root mean square), in objocts in proximity of thc tclcvlsinns, microwave ovcns, and lascrs undcr Public lincs. Assumlxions must bc madc to estimate thc Law 90-602, and subscqucnt Public Law 101-629, Safe hrgcst vchiclc or cquipmcnt anticipatcd to bc prcscnt Medical Dcviccs Act of 1990. Should additional undcr the lincs. Thc code in practicc is used in thc studies with occupational exposures indicate anccd for determination of right*of-way and clearance dlstanccs action, the Occupational Safety and Health for spcci~cations of line construction and placemcnt, Administration may become more involved. Othcr and to limit the maximum lcvcl of the clcctric ~cld on agcncic~ with potentlaUy expanded rolu are those with the right*of-way. (BPA, 1989) interrXts and ruponaibHitics for worker and public health: National lnatituto of Health, National Institute of The NESC is catablishcd for thc prevention of clcctric Occupational Health and Safety, National Cancer shock and is based on safety concerns from a singlc Institute, and the Centers for Disease Control. acute exposure. It is not intended to protect from possiblc long term health cffccts duc to chronic Outside thc federal agencies, there has bccn exposure. preliminary work by the National Council on Radiation ProtwAion and Measurements (NCRP), by the 6.5.3 Other States American National Standards Institute (ANSI), the American Conference of Govcrnmcntal Industrial In responsc to thc EMF health effects lssuc, and this Hygienists (ACGIH) and the Institute of Electrical and lack of national-lcvcl initiative, the individual states Electronics Engineers, Inc. (IEEE). Thcsc groups are have cxcrclscd a variety of options; some leading to a noted for developing scientific guidance documents regulatory response; however most staW. s have found describing measurements, cxposurr~, and proW. ction no basis for regulation. The few regulatory rcsponscs practices. Dcvcloping conscnsus standards is done by are a hodgc-podgc of restrictions with differing committees of scicntists, cogincers, and othcr exports rationalc, and incentivcs. This continucs since the in the spccializcd area. These recommended standards 1970's whco the concern was for electric rich o~cn lcad to an industry or regulatory standard. The exposures, until recent in~crcst turned to magnetic r~guhtory community consitch this cxpcrtisc in fields. Thc states* responsc~ can bo dcscribcd in four adopting its own s*tandards, even to the ado!~ion groups: (1) take no special action, (2) study and rgport vcrbathn of 8omc portions. With the cxcc!~on of on thc issue, (3) undct~akc and fund research, and (4) ACGIH as noted in thc abovc scction on occupational use rcgulatory authority and cstablish standards. standards, thcsc groups havc not pursued the issuc to thc rigorous lcvcl they employ to makc formal In the fwst group, many statcs have chosen not to takc rccommcndationa. Until the science becomes lus additional action, but continue to review each siting speculative and scientific con,cnsu can bc approached, lssuc on a cuc-by-ca,c basis. they may no~ accelerate the cffo~. Statcs in thc sccond group havc dircetcd a formal review of the scientific literature by a state agcncy or 6-6 Htahh Efftcts of Exposure to Powerfine Frequency Electric and Magnetic Fields an ad hoc committee of experts. The states ustutHy field limits and childhood cancer. At what levcl is the request a recommendation or report for the legislature outcome of cancer expected and at what lcvcl will it be or regulatory authority. Virginia has an on-going reduced? Since no dose response has been review with periodic xq~orting; annual updates have domonstratcd, no protoction or prevention can be followed the initial report. Other states requested a measured. Table 6.2 shows the standards for single written report such as Washington and Oregon. transmission lines that have been in phce in the Montana used this approach for the electric field specified stau. question and subsequently adopted electric field limits based on the recommendations. These various state 1. Minnesota's standard for electric fields dates reviews are usually undertaken for political concerns, back to 1976 u a result of the heating for the have lithe funding or support and have made limited W'mnipeg-Twin Cities Transmission Project. contribution to understanding of the issue. There The Environmental Quality Board set the 8- appears to be a need instead for collaborative effort on kV/m limit in the right-of-way based on a national level, rather than 50 separate studies, since suggestive evidence at the time that 10 kV/m the scope of the problem is national and international is a level which protects the public health and (notably Canada and the northern U.S.). welfare. They reduced that level based on the fact that lower fields could be achieved The third type of state response is represented by New using available t~hnology. This limit is used York and California. Both states implemented research for new tamsmission lines of 200 kV or programs funded by the utilities and administored by higher. stato agencies. Maine's legishture considered this option in 1989, but decided to collect and evaluate 2. North Dakota Public Service Commission existing data and research fmdings. Maryland has also employs a 9-kV/m electric field limit which undertaken a research program. was selected alter the issue of biological effects arose in some of their siting eases. In The fourth response option that a few states have taken lieu of a scientific rationale, the level was is the adoption of reguhtory standards imposing limits chosen to respond to potential concern for on the electric and/or magnetic fields. A tabulation of health effects. the existing limits for transmission lines for the seven 3. Oregon like Montana has a codified rule for states is shown in Table 6.2. The electric field limit electric field limits for tansmission lines. was set based on the consideration in the ea~y 1970's that if a biological effect exists, it is due to the electric The 1980 rule requires the Oregon Energy field component interacting with, or influencing, the Facility Siting Council to use the 9-kV/m limit in the right-of-way, i.e., in areas chemical and electrical nature of hutnan physiology. Later research implicating magnetic fields is evidenced accessible to the public. They adopt the NESC, but encourage an effort to keep the in the Florida rule in 1989 and the preliminary induced current level to be as low as rolemaking activities in New York and New Jersey. The rationale and the development of these limits is reasonably achievable. The intent of the standard is to protect the public health. In impotumt for us to consider.1 1990, a study was commissioned to review the literature. (Oregon, 1991) The difficulty in examining the standards is fmding any method to evaluate their benefit. The electric field 4. Montana set electric field limits as a prudent limits have been in use for several years: for the approach toward a public health-based NESC-based limits for induced current, the protection standard. The standard was recommended in against shock is a benefit. For the limits set on the an extensive review of the available evidence fields for presumed protection from chronic exposure, on electric fields. They ennmmissioned a no quantifmble outcome is estimattd, therefore no scientist in biceleetromagnetics research to benefit has been offered or substantiated. Even more evaluate the scientific work and suggest difficult to evaluate is the presumed link to magnetic action if warranted. (Sheppard, 1983) Concluding no serious hazards had been proven, the report did offer a limit as a precautionary level to provide protection from I Banks (1989a, 1989b, 1990) hu researched the any potential health effects due to chronic regulatory process and evidentiary record for the exposure. In adopting the rules, the states in great detail and can be consulted for further Montana Board of Natural Resources and reference. Sicsin (Microrave News) aiso reports Conservation included the l-kV/m limit at political and scientific events regarding reguhtory edge-of-right of way, but specified it for issues. ruidential areas, and provided the landowner Regulatory Issues (>-7 an option to waive that limit and thereby the field lcvcla generated by existing 345-kV reduce the width of the right-of-way. They lines. The limit is used in determining width also added a 7-kV/rn peak limit at mad of fight-of-way. An added limit for electric crossings. The provisions include a welfare- ficlds in the right-of-way on public roads is 7 based ttandard as well to reduce the kV/m; 11 for private roads and 11.8 for other annoyance characteristic from audible noise, terrain. This limit is similar to the National further extending the width of the right-of- Electric Safety Code, but restricts induced current to 4.5 mA instead of 5.0 mA. In way. recent years following the extensive research 5. New Jersey has a nominal 3-kV/m limit at the program, considerable activity occurred edge of right-of-way which has been used toward letting a magnetic field limit. An since 1981 as a basis for response to evaluation of the existing 345-kV lines in the complaints concerning eating lines. The state has been performed to suggest a limit basis was the status quo condition at the time that would be equity-based, i.e., they would and the lack of any evidence on adverse not permit additional lines which would cause human health effects at that level. fields higher than have already existed. (That Consideration was given to annoyance from level is approximately 200 raG.) transient spark discharge at 6 kV/m, ~o a margin of two was used to set the 3 kV/m. 7. Florida's legislature empowered the They are currontly involved in rulemaking to Department of Environmental Regulation to set a magnetic field limit following a directlye adopt roles for transmission lines in 1983 from the Commission on Radiation which would restrict the electric and magnetic fields. Upon the advice of a scientific Prote, etion. advisory group of experts, the Department 6. New York set an interim standard of 1.6 did not exercise that prerogative. Following kV/m edge of right-of-way limit (350-foot legal proceedings and a aecond scientific width) in 1978 at the commencement of the advisory group, limits were recommended. Power Lines Project. The intent was to set Rulemaking and adoption of magnetic and the width of the right-of-way to status quo, electric field limits followed, setting limits for · · Table 6.2 -$lalt EMF Slantlards for Tammission Lines Slat~ Electric Field ~kV/m) Magnelic Field {tnG) Location Application M~n;.~;ala 8 in ROW > 200 kV qonh DakoU 9 in ROW Informal Oregon 9 in ROW 230 kV, > 10 miles (ca:lifted) Montana I edge of ROW > 69 kV /codi~ed) 7 in ROW road erossinEs New l~i'-~I 3 ** edEe of ROW Guideline for complaints New York 1.6 200raG*** edge of ROW > 125 kV, > 1 mile 7 in ROW public roads 11 in ROW privatt roads 11.8 in ROW other t~rvain 500 kV lines Flodda (codi~ed) 10 in ROW Single Circuit 2 200 edge of ROW Single Circuit 2 250 edge of ROW Double circuit 230 kV or leu= 8 in ROW 2 150 ed}e of ROW · Compiled from Banks(1989), OTA(1989), MWN(8/89), FL Statute · * Under consideration · ** Final 9111190 PUC 'Interim Standards' 6-8 Health Efftct~ of F_:cposure to Powerline Frequency Electric and Magnetic Fields lines of 69 kV and ·bove, and includes other coda and standards thai art generally substations. The electric field limit ·t the accepted by the industry, except us modified by edge of the fight-of-way is 2 kV/m; in the this commission or by munic~oal regulations right-of-way, it is 10 kV/m for 500-kV lines within their jurisdiction. Each utility shall and 8 kV/m for 230-kV lines or lower. The constract, install, operate, and maintain its magnetic field limit is specified only for the plant, structures, equipment, and lines in edge of the right-of-way: 150 mG for 230-kV accordance with these standards, and in such lines or lower, 200 mG for single circuit 500- manner to best accommodate the public, and to kV lines and 250 mG for double circuit 500- prevent interfereacts with service furnished by kV lines. other public utilaits insofar us practical The standards were set in response to It appears that if electric and magnetic field exposure concerns and political motivations regarding standards arc adopted by the ANSI, NESC or another the possibility, not evidence, of adverse group that the industry acccpu, the PUC can readily health effects. The standard hu therefore adopt them as guides. If they are to be implemented as bccn catcgorizcd as a welfare-based, not limits for aH siting cases, this may require rulemaking health-based standard. The standard is and inclusion in the substantive rules. subject to review during 1991. In examining the rule for the standards of construction, 6.6 The Situation in Texas a question arises if one considers the possibility that municipal r~gulations may bc adopted. Can any or all The EMF issue in Texas is similar to the experience in cities adopt exposure limits for EMF and thereby other states. When attcrnpting upgrades or selecting restrict a PUC permitted line? How serious would be new corridors, the utilities are faced with a potentially the effects of several different limits for one service lengthy and expensive legal situation. Historically, area7 The ramifications may be important in future there have bccn challenges to their siting propouls; cases, such as the case in which an Austin'city however, they were oltcn based on monetary and ordinance restricted lines which were necessary for contractual arguments and were settled by negotiating ]and davciopment ostensibly referencing possible public the price and arrangements for land acquisition. The health concerns. The ordinance was rucinded in 1989. decision making did not grca~y involve ton claims and legal battles alleging health claims, The PUC also has the option of deferring arguments and potential hcalffi effects issues to the Texas Some more recent siting cases have been in the courts Department of Health (TI)H). In effect, the PUC could and appellate courts for years. Should standards be rely on 'TDH's expertise and authority to identify the established, ~c siting process might be simplified, presence of health risks to ~c public, advise other Once a standard is set, a utility can use ticsign critorla apprupriatc authorities, and participate in the and pntcticc to achieve compliance. The standard development of health-bued standards if they become would ostensibly rernovc the argument for health justified. Should future evidence suggest personal life- effects or greatly tcduco the chances for appeal. style changes, for pregnant mofficrs, children with Barring the complication, for a moment, that there is no genetic or officr traits which altor their risk for clear or conclusive evidence on which to develop a childhood lcukcmia, or occupational health health-based standard, what options exist in Texas? considcnttions, health agencies have the needed In Article HI, Scc, 16, (a), the PURA gives to the PUC expertise and experience. "**the general powor to regulato and supervise the 6.7 Conclusions business of every public utility within its jurisdiction.." Additionally the PUC "*** shall make and enforce rules In the area of regulatory issues, a wide variety of reasonably required in the exercise of its powers and actions has been taken by legislative and administrative jurisdiction .... " bodies in reapease to public and media concerns about potential health effects. The range of ruponsc is gnat The substantive rules promulgated under ~is authority and the nttionalc and reuons vary. There has bern currently contain · provision which is usociatcd wiffi some formal and some informal use of standards, the EMF issue. It is addressed in the reference to the although standard setting has been controversial and of construction of new service, Section 25.44(a), New unknown benefit. The majority of states have Construction: determined attcr their own research, or based on Standards of constrswtion, In determining x,q~om from experts, that the most apprnpriatc standard practice, the commLfsion will be response at this time is to defer any standard-setting g,,ided by the provi~ior~f of the American activity until scientific evidence is more convincing and National Standards Instit~e, Incorporated, the consistent so ~at effective and ·ppropriatc actions can National Electric Safety Code, and any $sch be devised. Issues References American Conference of Governmental Industrial Hygienists (ACGIH), Threshold [.imit Values and ]~ioloeical Exposure ladices for 1989-1990i ACGIH. Cincinnati; 1989. American Conference of Governmental Industrial Hygienists (ACGIH), 1990-1991 Threshold Limit Values for Chemical SubsUmce~ sod Physical AEents and Bioiolical Exposure lndieesi ACGIB, Cincinnati; 1990. American Conference of Governmental Industrial Hygienists (ACGIH), [991-1992 Threshold Limit Values for Chemical Substances and physical Agents and Bioiolical Exposure Indices: ACGIB, Cincinnati; 1991. Banks, R.S., Reiulation of Overhead Power Transmission Line l~lectrie and Ma2netie Fields; Robert S. Banks Assoeiatea, Inc. / Edison EltwArie Institute; Minneapolis; 1984. Banks, R.S., The EMF Health Effects Issue: An Overview of Developments, 1988-1989.; Robert S. Banks Assoeiatea, Inc.; Minneapolis, 1989a. Banks, R.S., power Lines: Kids. Schools, and Caneelf; Robert S. Banks Associates, Inc.; Minneapolis, 1989b. Banks, R.S., EMF Res~ulation: A Look at the Present Status; Robert S. Banks Associates, Inc.:, Minneapolis, 1990. Bonneville Power Administration (BPA), Electrical and Biological Effects of Transmissiop Lines,; DOE/BP-945; Portland OR: 1989. California Public Utilities Commission; California Department of Health Services, Dral~ Report: Potential Health ]~ffecta of Electric Power Facilities; San Francisco, July 1989. Commonwealth of Australia, Interim Guidelines on Limits of Exposure to 50/60 Hz Electric and Magnetic Fields (1989); National Bealth and Medical Research Council; Canben'a, 1989. Florida Statutes, Chapter 17-274 F.A.C., Electfie and Magnetic Fields, January 18, 1989. International Radiation Protection Association (IRPA), International Non-lonizing Radiation Commitlee, 'Guidelines on Limits of Exposure to Radio frequency Electromagnetic Fields in the Frequency Range from 100 kHz to 300 GHz," Bealth Physics 54: 115-123, January 1988. International Radiation Protection Association (IRPA), 'Interim Guidelines on Limits of ExposUre to 50/60 Bz Electric and Magnetic Fields," Health Physics 58:113 -122, January 1990. Microave News ELF News: M/A 1989, M/J 1989, J/A 1989; New York, 1989. National Radiological Protection Board (NRPB), "Guidance as to Restrictions on Exposures to Time Varying Electromagnetic Fields and the 1988 Recommendations of the International Non-Ionizing Radiation Committee," NRPB-GS 11, Oxfordshire, UK, 1989. Of~c~ of Technobgy Assessment (OTA) Biological Effects of Power Frequency Electric and Magnetic Fields: ~; Congress of the United States, June 18, 1989. Sheppanl, A.R., Biological Effects of Hizh Volta2e AC Transmission Lines; NTIS PB 83-207241; 1983. World Health Organization (WHO/IRPA), Environme_n_tal Hea_!th Criteria 35: Extremely Low Frequency (ELF} FieIda; WHO, Geneva, 1984. World Health Organization (WHO/IRPA), Environmental Health Criteria 69: Magnetic Fields; WHO, Geneva, 1987. 7.0 POLICY ISSUES AND OPTIONS The proposition that electric and magnetic fields (EMF) management options but include different ways of from such sources Is high-voltage transmission line. s, understanding =nd frsming the key policy problem. distribution facilities, home wiring, and electric appliinces, may pose a risk to health has become Political institutions can b¢ expected to take different widely recognized is a legitimate and worrisom¢ policy stands on the EMF issue, depending upon how hypothesis in hood of thorough investigation (S¢¢ OTA, they collectivcfy weigh evidaxcc and judge 1989). lust in the past year or so, several acceptability (Se¢ Clark, 1990). Pirt of the account Congressional Hearings have been held, the Office of offered hero will trace those differences to an Technology Assessment, Dclmrtmcnt of Energy, and institution's structural and procedural sltributes. With the Environmental Protection Agency have issued the problem posed by EMF exposur~ open to report-% and a number of states have convened export interpretation, however, disagreements over policy will panels to investigate the possibility of adverse health span the language used to characterize EMF as a public effects from EMF. Media envcrage has moved in concern as well as its empirical character. Accounts of parallel, providing a window for the public on the contending claims for action and inaction must reach emerging debate. Coverage in ncwspapars and beyond institutional features to the contest over soc,.'~l newsmagazines emphasizes state initiatives and the and political meaning. This section will examine public positions of various st=kcholdcrs, while more policy on EMF on beth of these levels. First, we will spccializai poriedicais, such as Science and IEEE suggest several ways that judgmental biases abeut risk Spectrum, sum~ scientific findings to dato and from EMF appear to paraU¢l the structural and speculate on their import. Exposes in the New Yorkcr, procedural features of different political institutions. and Family Circle offer accounts of porsonal tragedy Given the pattern of institutional experience with EMF and exposure to EMF. Despite all of this aRcntion, the across the states, the discussion will focus more intently scicoti~c cvidenco suggesting a link between magnetic on regulatory agencies and the courts than on other fields and disease is considcrai inconclusive by many institutions. Atteution will then turn to questions of exports and appears inadequate to support an how EMF has bccn depicted in public discourse, how authoritative claim one way or the other (See Morgan its status as a problem is being defined, and to the and Nair, 1990). Further, there is a notable lack of assumptions and policy prescriptions that foBow. This consensus on whether EMF, as currently understood, analysis will include the role of rhetorical devices, constitutes a public policy problem. Even among those principally analogy, and consider four distinct problem assured of its problemstic status, there is no unifying dcrmitions and their import for the design and choice of presumption abeut the type of problem it represents or competing policy options. the collective action it warnrots. 7.1 EMF Policy and Political In the absence of conclusive scicoce on the matter, the Institutions interpretations of other social and political institutions have effectively come into play, each bringing a somewhat different perspective to the EMF issue and a Unlike assessments for most hazardous substances, distinctive way of accommodating the perceived ground rules for mmuring exposure to EMF, and thus uncortainty over adverse health effects. The result is a characterizing the health risk, arc themselves open to somcwhat confusing mixture of warnings and debate. Estimates of risk published in the scientific reassurances, of calls for more study, or for immediate literature vary widely with dct'mitions of exposure and action. Typically, such wide variation in responses by the exposed population. sRcntion to confounding both people and institutions to environmental risks can influences. and to the precision and accuracy of be attrib,_,t_~ to well documented biases that influence measurement. In the view of the scientific community, porcclxions of mk and judgmatts about its the primary means to overcome these inconsistencies is accepUbRity (Fischhoff, ctal.. 198~). In the cue of to accumulate better evidence hhmugh more research, EMF, however, there is no benchmark for such assuming that subsequent studies will introduce judgment grounded in a scientific consensus on the successive corrections and refinements, eventually empirical nature of the risk at issue. Differences in allowing investigators to converge on a stable set of empiresis and interpretation, then, extcacl beyond the valid ruults. Arguments to quicken the paco of workings of judgmental biases to encompass the basic research sometimes appeal implicitly to prejudico meaning assigned to the EMF issue. Policy questions against social and political methods of "resolving" abeut EMF exposure, accordingly, are not limited to scientific uncertainties and 1he deairabillty of clear-cut choices among control strategies or preanpting them. Nevertheless, in the interim, there will be efforu to anticipate possibly conclusive 7-2 Health Effects of F~cposure to Powerline Frequency Electric and Magnetic Fields findings, and to take action to mitigate thcir liltely similar ways across st·to lines. A few basic concepts social and economic impact. Until the scientific adapted from the literature on decision theory will be community has fashioned a consensus on EMF health used to characteriz~ the judgmcntal biases of selected effect, claims about how best to meet assumed institutions and to simplify their link to structural and mitigation needs or to manage suspected risks, must procedural differences. Since the relation between roly principally on a carefully limited selection of science and political institutions sets the context for scientific sources and largely non-scientific backing. much of the policy debate over EMF, science itself will be treated as a social institution whose procedural and Political institutions in our society can provide such structural features will serve as a basis for comparing backing, offering alternative means for re. solving scientific with non-scientific ways of framing EMF uncertainty, and standards of proof and persuasion that questions. are very different from those of organized science. Accordingly, there is likely to be conffict among these 7.1.1 Judl]msnta[ Blasss institutions over how much risk there is, and how quickly they must decide. Several factions within the In the face of unceflainty over a given hazard, scientific community compound this co·fflet by airlag institutions can choose either to act or not act, their disagreements over the significance of EMF risk depending upon their bias, and may be proved wrong in non-scientific forums. Risk-accepting and risk- in either case. Although mistakes are to be expected, aversiva factions have formed to disput~ the magnitude not all mistakes carry the same consequences. of harm suggested by the available evidence. Conflict Borrowing an analytical distinction from de~ision of this sort can easily test public patience with the pace theory, the core issue in the choice of action versus of scientific progress on EMF issues and confuse calls inaction appears as a tradeoff between two kinds of for more careful study with veiled attempts at strategic errors: the false positive, acting when no action was delay. Critics of postponing action to await further warranted; and the false negative, failing to act when research assign self-serving motives not only to the action was warranted. Since, at the thne of the choice, stakeholders, such as the electric power industry, but the uncetlainty over the validity of the acJ. ion has yet to also to the scientists who will conduct the studies. be resolved, the recognition of error is necessarily retrospective. Nonetheless, attention to the tradeoff Across the state~, · number of institutions, including beforehand reveals · great deal about an ins'tatution's courts, legislatures and administrative agencies, have weighing of the possible political and economic explicitly addressed the unc.~.~ainty over the health consequenc. t~ connected with each error and its relative effects of EMF. While the scientific community tolerance for one kind over the other (Linder and continues to play · prominent role in the EMF issue, McBride, 1984; Page, 1978). Each particular tradeoff other institutions whose judgment does not re~luire an affords · somewhat different role to science and invites accumulation of research results are inclined to decide a different measure of discretion and pr~lictive more quickly on action or inaction despite the judgment on the pa~ of the policy maker (Gaines, remaining uncertainties. Their decisions wffi be based 1990). on interpretations of incomplete information, structural and procedural factors, and on their expectations of As an illustration, crim'.mal courts, porlmps out of social science and the findings it may eventually produce. consensus rather than judicial temperament, are The resulting diversity of action and understanding predisposed toward false negatives since they assign an makes it difficult to find any reliable cue. on the single, intolerable political cost to false positives; in this case, most appropriate, public policy response. conviction and punishment of the innocent. Accordingly, their threshold for action is set quite high, The principal task in this section of the report is to demanding · close scrutiny of plausible evidence of examine the judgmental biases, ·long with their guilt, their warrant for action. In contrast, civil courts structural and procedural correlates, that lie at the roots manifest · less pronounced bias, shitking more of the of the diversity among policy responses to EMF. costs of error to the claimants in deference to the status Social and political institutions rather than whole states quo (See Huber, 1958). wiil serve as the focus. Notably, most of the conventional wisdom about statr..Ao-state variation in Greater complexity can be introduced into the tradeoff policy making based on differences in political, by recognizing the parallel between the costs of errors demographic, and economic cbar,,cteristics appear not and the values assigned to earfeet responses. Not all to apply to the EMF issue. Knowing which states are correct responses share the same annsequences nor are considering EMF questions will be not nearly as valued in quite the same way. An insfitution's inaction, informative as knowing which institutions within these even if that response is proved correct by subsequent states have assumed jurisdiction, since we expect developments, may have very different consequences institutions sharing judgmental biases to respond in than a correct response that involved action (V(ddavsky, 1990). Further, the relative value of each politically. Further, the gray areas 4~vccn harmful type of correct response, from this perspective, can be and harmless, substantial and undermeal, have been linked to the relative costs assigned to each kind of excluded to highlight the workings of the "bin" notion error. Institutions that place a high value on correct and to clarify its connection to certain institutions. rejections of ambiguous evidence and on the avoidance of false positives, for example, will tend to have a To the cxtont that such a bias exists, similar institutions higher threshold for action than other institutions icing across the start should respond in predictably similar the same level of uncertainty. In contrast, those ways to uncertainty, employing similar conventions of placing a high value on anticipating the need for action interpretation and sharing judgments that distinguish and who fad false negatives to be the more them from other kinds of institutions. Logically, we objectionable of the two kinds of erron will tend to would expect the utility commitions of any two states, have a lower threshold for action, again independently for example, to be more similar to each other in how of the balance of the evidence. they deal with scientific uncertainty, than either would be to a Health or Environmental Agency within theh- The simplest way to represent the bias in how respective state. Under this intorpretation, it clearly institutions structure the uncertainty surrounding matters which kind of institution has assumed adverse health effects is in a two-way table. On the jurisdiction over the EMF issue, since different left-hand side of Table 7.1 appears the two altornative institutions witl structure uncertainty differently and positions that actors can take based on inconclusive will employ different policy tools to assist in its evidence. The tint is to conclude that substantial resolution. adverse effects are indeed likely and, accordingly, that widespread precautionary measures are in order. In The tradeoff between false-positives and false-negatives other words, the risk posed to the public may be or judgmental bias unique to each institution is unacceptable and something should be done. This effectively supported by both structural and procedural stance need not involve a denial of the inconclusivehess features. Typically, however, one set of features will of the evidence but might be based instead on an appear more prominent in a given institution, aversion to the possible consequences of postponing permitting a more direct link to its imputed bias. action in error. Avoiding a false negative outweighs Structural features, including norms, incentives and other concerns. The second position holds that adverse formal organization will be emphasized in discussing health effects are largely unsubstantiatcd and, thus, that the legislature, media, and scientific establishment, control measures are premature. From this viewpoint, while procedural ones will be the focus in a later until the risk to the public is clarified, or at least shown treatment of regulatory agencies and the courts. to be potentially large, no action is warranted. Here, it is the consequences of acting without a ftrm warrant The Science Community. The scientific community as that ms most troublesome. an institution has a high stake in avoiding false positives, public claims of an association lator found to The two possible resolutions based on prospective have been unjusti~ed, for example, since its authority evidence that will ultimately prove one or the other of in public policy matters is linked beth to its public these positions wrong appear at the top of the table: at image of skepticism and to the porccived conservatism some time in the future, either health effects prove to of its procedural mechanism for validating evidence bc significantly adverse, or not. The four discrete (Large and Michie, 1981). When this image is shaken outcomes appear in the body of the table. It should be because of premature claims that mislead the public or noted that this construction assumes that the evidence their officials, the credibility of science as a whole on the health effects of EMF will eventually become suffers. For instance, the scientific assurances given to conclusive, if not scientifically, then socially or Table 7.1 - The Trsdcoffs in Responses to Uncertainty RESPONSE EVENTUAL RESOLUTION EMF Proves Harmful EI~4F Proves Hsrmleu Yes, Correct Rcsponsa "Falsg Positive" Error Risk is Substantial [precautions prove wsrrantsd [sacrifice trust & investment in precaution] No, "Fals~ Negative" Error Corrsct Reapong Risk is Not Yst Defmed [assung r~sponsibility for delayed response] [caution proves warmme, d] 7-4 Health Effects of Exposure to Powerline Frequency Electric and Magnetic Fields the public throughout the dccadc of thc 1950's that inconclusive, thcre are no easy footholds for supporting nuclear testing and radiation exposure posed no gnvc onc's position. Not knowing thc chances of ultimately danger, in thc vicw of some analysts, created a being proven wrong, onc is left to decide, at lcast in credibility problem with which the proponents of pan, on the basis of loyalties and how these affect nuclear power must still contend (Bupp and Denin, one's views oft he possible erron. 1978). Inferences about the relative benefits to the scientific community of either kind of correct response Science has traditionally served the public health are more ambiguous. profession not only in the identification of environmental hams, but also in their mitigation. A Timely claims of problems in need of public action are number of successful collaborations has resulted in a prime sourc~ of social support for the scientific pubtic policy for controlling exposures to both natural enterprise; however, a large part of this support and artificial substances, ranging from benzene to involves a reliance on the capacity of science to offer chloroform. For many within the profession, thes~ protection from the ill-founded claims of othen on successes reinforce a bias toward precautionary action society's resourcts. Assuming that neither of these and confidence in predictive judgment. A more subtle values can compensate for the costs of false positives, effect has been to create a powerful analogy that shapes we can assume a relatively high threshold within subsequent considerslions. Ambiguous risks can be organized science for supporting action in the face of given clear definition by assuming parallels to earlier uncertain evidence (Jasaaoff, 1990). Ironically, for the success stories, perhaps most poignantly to stories individual scieatist, the false negative may r~pnsent the involving the identification and control of hazardous more serious risk; a pattern of claims establishes one's exposures to widely used substano~s, such as asbestos reputation and status, even if some proportion of the and lead. Proven models of exposure, causal relations, claims are eventually refuted by the larger process of appropriate control measures, patterns of institutional testing and replication. support and opposition, and recipes for suc. t/.ssful policy campaigns fred their way into current The Media. Several othe~ institutions can effectively discussions about EMF. The policy import of these exploit this irony by invoking the authority of science analogies will be discussed later in this section. while inviting investigators to preempt the scientific process by making their intramural claims public. The 7.'l .2 $tats Logislaturos print and electronic media have procedural devices for validating evidence, to avoid false positives, although Although state legislatures vary widely in their by design they are less time and resource intensive and structures, pafiisaaship, and professionalism, at the therefore less stringent than those of science. The false level of their individual mnmben, it is safe to assume negative, in contrast, represents a threat to their that they share a common desh~ to be reelected. Part economic prosperity. The media enverage given to of this desir~ is channeled into eftotis to mitigate EMF has b~en prompted in part by the general climate conffict either by building compromises, responding to of public concern about concealed threats to health. it public sentiment, or by handling divisive issues can also be seen, however, as an effort by news precedurally. Another part leads to efforts to transmit dir~ors and managing editon to minimize the risk of material or symbolic benefits to their constituents while missing a potentially significant phenomenon, given spreading or deflecting costs that might accrue to them. their p~rception of a shift on the part of some scientists To a large extent, the value associated with either type now intent on doing the same. Being a part of a of correct response resisting action that was possible "$coop' may have sufficient appeal, for some unwarranted or taking action that was/will reflect in science as well as the media, to outweigh the risks of partisan affiliations, one political party taking pride in sounding a false alarm. the former and the other pa~y in the latter. The Public H~lth Profession. The values of the In the case of crmn, however, partisanship is likely to public health profession, grounded in an ethic of be eclipsed by more elemental civic values linked to disease prevention, militate in favor of precautionary our common political tradition. The false positive measures in the face of uncertain harms. Here again, transiates into a waste of taxpayen' money, and the the individual scientist may be torn between conflicting false negative into irr~sponsibk disregard for pubtic imperatives: on the public health side, advocate action safety. Neither of these consequences is likely to be just in case the harm is real, venus, on the scientific palatable to kgislaton of either party. As · result, we side, maintain an open-minded yet neutral stance until can expect legislatures in general to seek · hedge sufficient evidence is in. In this instanco, choosing is against erron of either type by committing r~sourc~s to not so much · w_~__-_,r-r of professional judgment as it is · allcmpt to resolve the existing uncertainty. For some personal de~ision about where one's primary loyalties legislatures, sensitivity to these pot4mtial ensts will be naid~. When uncertainty is pervasive and evidence is coupled with external pressure for · political Policy laues and Options 7-5 resolution, perhaps in response to local conflicts over treatment of uncertainty set down in their authorizing who should bear the costs of the uncertainty. In this legislation (Gelpe and Tarlock, 1974). Agencies instance, legislatures can either develop an additional charged with regulatory responsibilities over functional hedge by freezing current conditions (and the policy areas, environmental quality, energy, or health, distribution of costs) until more becomes known about for instance, tend to follow quasi-legislative procedures the potential ri~ks, or they can try to exit altogether by involving a notice and comment process, and informal reassignin g jurisdiction over the controversy to an hearings or consultations on any rules, orders, or administrative agency. decisions to be adopted as part of the state's Administrative Code. The setting of mandatory The choice between these three legislative strategies - standards at both the federal and state level tends to commit to further study, frt~zt conditions, or reassign conform to this pattern of informal rulemaking, in part, jurisdiction - will probably be dictated more by the as a way of capitalizing on agency expertise. structural and historical conditions within a given state's government, than by any particuhr substantive In Texas, for example, agency decisions based on feature of the EMF issue. Some legislatures are more informal rolemaking are considered under the same inclined to devote resources to the further study of presumptions of legitinmcy that apply to legislative technial issues, perhaps by virtue of being better ones. The agency need only present a reasoned equipped to act on technical information, while other justification backed by more than a '...scintilla of legislatures, deferring to the expertise of administrative evidence...' to meet legal challenges to its decisions. agencies, will assign jurisdiction to one of them should Such deference to agency discretion permits a hrge the issue not yet be under formal consideration (Feller, measure of predictive judgment to color the agency's et al., 1979). Still other legishtures will choose to act treatment of uncertainty, especially in sitnations where based on the initi""""ative of their 'peer" or neighboring scientific evidence appears inconclusive (AFFRA, legislatures, or will await federal guidance (Sabatier 1989). and Whiteman, 1985). In contrast, agencies charged with assigning property In any event, political pressure to do something rights, duties, or privileges, or granting permission of legishtively about EMF concerns is likely to arise from some sort tend to adhere to quasi-judicial procedures unresolved conflict surrounding transmission line siting involving evidcntlary hearings based in testimony under disputes, with parties seeking legislative relief on the oath, depositions and writs of subpoena, the hceis of disappointing judicial or administrative compilation of a formal record, and ease-by-case settlements. Each attempt to site a new transmission deliberation. Regulatory decisions on ratemaking, line or upgrade an existing one raises the possibility of licensing and certification of various types tend to public opposition and, in turn, creates an opportunity follow this more forrealized procedure across the states for disputing EMF chims. Where the conflict is and at the federal level (Breyer, 1982). By analogy, especially contentious or where opposition has been regulatory 'standard setting" in this context is more widespread, the legislature is more likely to consider likely to involve the promulgation of supplemental action. Legislative activity in states such as Califorola, duties or guidelines for obtaining, or complying with Florida, New York, and Washington, can be traced to the conditions of, one's authorization. basic conflicts over the location of transmission lines and a perceived need to mitigate them politically (See The quasi-judicial procedures that frame regulatory Banks, 1989). In Texas, by way of contnst, the Public responsibilities for the taking or giving away of Utility Commission took the initiative to gather and something of value have far more extensive "due interpret technical information on EMF health effects process" protections for the affected patties than does before any forrnal consideration of the issue by the informal rulemaking. Here, the resolution of stat~ legislature. uncertainty rests on the weight of the formal rt~cord, with questions of fact and of law settled in much the 7.1.3 Stato Rogulato~ Agoneion same way as in non-jury civil cases. The agcncy's discretion, however, is more narrowly defined and its Administrative agencies charged with regulatory predictire judgment given less opportunity to influence responsibilities generally address scientific uncertainty the ['anal outcome than in informal rulemaking through procedural mechanisms patterned after either (McGarity, 1979). Nonetheless, any particuhr ruling the legislature or the courts. The formality of the is treated more as a prec.~ent than as a general rule ageney's deliberations, its gathering and use of and can thus be more easily modified or withdrawn evidence, and its connection to judicial oversight will than can the standards coming out of an informal typically follow one of these two institutional designs. rulemaking procedure. In short, the 'bias" toward how Along with their procedural mechanism, agencies are uncertainty is resolved is quite likely to reseTable a likely to find their particuhr 'bias" toward the 7-6 Health Effects of F~posure to Powerline Frequency Electric and Magnetic Fields judicial one, with responses subject to case-by-case this type will typically mandate precautionary action in revision. the face of uncertainty; thc justification in these instances may be based principally on at presumption in In Texas, this more formalizal regulatory process falls favor of protective measures when public health and under the rubric of 'contested cases'. Not only are safety are at issue. Not surprisingly, the few states that regulators bound by strict procatural raluircments, but have formal field limits (Montana, Oregon, and their decisions atre subject to much closer scrutiny by Florida) are also those that vest atuthority over the courts than for informal rulemaking. If a legall transmission line certification or operation in such challenge is made, the state court can review the catire atgencies or their offspring. process u though no agency decision had been rnadc; in othcr words, once at decision is challenged, no In conthat, public utility commissions tend to rely on dcfcreacc is paid to the agency 's deliberations or quasi-judicial proccdurr~ that bias them away from conclusions. The agcncy's ctccisinn must be replicated "false positives". Rules arc more likely to be statai in by thc court, othcnvisc, it will be modified or general terms as guidance in making claims before the overturned, commission and to be atpplied on a cue-by-case basis. Here, the presumption bchind the procos favors Examl)les Within Retaliatory Aitcncies. While much of existing levels and forms of precaution over the discussion so far has encouraged comparisons prospcctivc oncs. To thc cxtcnt that public opposition across diffcrent institutions, it is uscful to add somc or scicntific irmdings begin to challcngc thesc, thc tint cmpiricaJ detail by examining the implications of stcp is likely to bc cithcr a clearer atrticulation of structural and procaJural variety within a given guidance regarding thc treatmcnt of EMF in the institution. From this persp©ctivc, thc question is not certification process, as in North Dakota and Colorado, so much, what is likely to happen once at regulatory or an effort to establish the statns quo as at benclmmrk agency, as opposed to some other institution, considers for assessing the ncai for any subsalucnt precautions, EMF issues, but rather, which kind of regulatory as in New York. The effect of these two atpproaches is agency has assumed jurisdiction within a given statc, not to prohibit field strengths above a certain threshold, and what kind of judgmental bias does this imply? say, by making them illegal as in the case of codir~i ~cid limits, but rather to cxpand thc burden of If it is a regulatory atgency designed, art least in part, to persuasion incumbcnt on the proponcnts of new high address preblcms in terms of control strategies and voltagc lines or upgrades to incorporate EMF authorizat to promulgato mandatory standards through considcrations in their applications. informal rule-making, wc can cxpcct at set of such standards, ~cld limits or exposure limits, to cmcrgc as To datc, thc lcgislatures or administrative atgencies of thc fatvored ruponsc onco the agency decides to about 17 states have explicitly considered the possibic cxcreisc its predictire judgment. Following the samc health effects of EMF; thcir policy responscs range logic, a utility or siting commission designed for from thc lcgislature's dismissal of health conccrns for atdjudicatory proceedings and cue-by-caso rulings is lack of sufficient cvidcncc, as in Wyoming, to thc more likely to defer to thc praiictivc judgment of those codification of fornml EMF limits, or transmission outsidc thc agency and to reapend with gnidclines for lines, by rcgulatory authorities as in Florida (Banks, mitigating exposures in at particular context by a 1989a). If recent litigation over condcmnation and tom spcci~c party rathcr than with gcncral rules, This is whcre EMF health risk has been posai as an issue is not to suggest that once jurisdiction is settled thc results included (scvcral examples of which atplr, ar in soction arc determinate, but only that wc can cxpcct certain 7.1.4) then the number incrcascs to 25 statcs, Unlike institutional biases to come into play, given thc other policy issues, such as taxation, thcre has becn no uncertainty over what, if anything, should be done. particular pattern of leadership among the states, Furdicr, these biasea, on balance, are assumed to whcrcin thc innovative approaches of'a few *states inclinc participants to atccommo,~__,_-_-uncertainty in i diffusc to thc rest and create relativcly uniform characteristic wary that has much to do with the nature patterns. It may wcll bc that the lack of policy of thcirinstitution, agrccmcnt among thc states simply rc~ccts the unccrtainty within thc scientific community regarding Dcpartmcnts of Energy or Envirunmcntal Airfain, for the health risks from EMF, Based on the earlier cxamplc, are more likcly to rely on codi~ed field limits discussion of institutional responses, however, diversity since much ofthcir legal atuthority is framed in terms of is also consistent with the fact that the same institution the sating of standards. In other words, thcy atre does not have cxclusivc jurisdiction over EMF issuea in biased in favor of atvoiding 'false negatives' since thc cvcry state. Rcgulatory agenciea arc arctire in somo setting of standards that are binding across individual states, legislatures in others, a mix of institutions, or circumstance is more likcly to crr on thc side of over- none, in still othcrs. inclusiveness. The ambling statutes for agencies of Policy Issues and Options 7-7 7.1.4 The Courts In Klein, the jury, in effect, found that apprehension over EMF risks dictated caution and faulted the utility Criminal courU share the same aversion to false for not taking these concerns seriously. In San Dieeo positives that motival~ the scientific process; however, the jury compensated a property owner for losses due they also sitare the irony of aligning the incentives for to the apprehcnsiveness of buyers over potential health potential participants in their process in a way that rislu, given the proximity of his land to transmission expresses substantial concern for false negatives. The lines. The Miller case in New York State dealt with judgmental bias in civil courts depends in large part on the same issue, but added the burdens of proving that the kind of changes in the status quo sought by the such apprehensiveness had depressed land values in the claimants. Increasingly, the courts are open to put, and that apprehensivehess itself had a reasonable claimants willing to take on the burden of proving that basis. In all three cases, backing for the chims of EMF is potentially harmful either to their property specific health effects m-__-_cnxl far less than the doubts values or to their health. While these claims might these claims raised about public safety, doubts that serve as a barometer of public concern, and arc thus were kindled by the ambiguity and inconclusiveness of likely to attract media coverage, they also signify the the scientific cvidenco. In the end, the co urt's way some courts have chosen in practice to judgment hinged on the credibility of the expert accommodate cvidcnco regarding EMF health effects testimony and on whether the appreheasiveness borne along with the persistence of scientific uncertainty. of this testimony could or should have affected the particuhr bchaviors (utility siting or valuation) for The courts effectively reverse the logic of the public which a legal remedy was sought. utility commission, although the bias toward avoiding *false positives* (perhaps principally through an Since the OTA report (1989) cited the Klein case as aversion t,} change in the status quo) and case-by-case one of the "growing pressures for states to take consideration remains the same. In this forum, the regulatory action to protect citizens against possible burden of persuasion falls, not on the proponents of risks," it is useful to take a closer look. The Klein new lines or upgrades, but on their opponents, who Independent School District fried objections with a attemlX to prove that EMF considerations entitle them county trial cou~ to Houston Lighting and Power to compensation or relief of some sort. Here the court Company's condemnation of an easement across the roles fwst on the admissibility of such claims and then School District's property for the siting of a 345-kV upon whether the evidence behind these claims transmission line. Unlike most other condemnation provides sufficient grounds for a judgment in favor or cases where the amount of 'just compensation' is at against them. Unlike the public utility commissions, issue, in this case the School District alleged that the the courts will rarely pus judgment on the substance of utility had abused its discretion in taking the casement any EMF effects. Several ca.w~ can serve as and had subsequently erected the transmission line in prototypicalexamples. 'reckless disregard' for the school's use of the property. The trial court granted judgment against the Examples Within the Courts. By and hrge, most utility, voiding the condemnation and awarding punitive condemnation cases (chiming prnpcrty for public damages for trespass. The State Court of Appeals conveyance) have ruled out any consideration of EMF overturned the award of punitive damages and chims health risks, since the statutory language authorizing of trespus but found a credible basis for the jury's "takings" is typically very explicit about what may and finding that the utility had abused its discretion in may not bo admitted as grounds for compensation. taking the District's prnperty. Until this language is changed by the lcgishture, the courts are not very likely to expand the scope of The findings in this case did not address the claim "that conflict to incorporate these new elements with four there are potential health effects assooiated with notable exceptions, IKlcin Independent School District exposure to powerline fields," as the OTA report v. Houston LiL, hting & Power (Klein); San Diego Gas alleges. While Klein may affect how utilities & Electric v. Dalcy (San Diego): Miller v. State of accommodate health concerns in future siting New York (Miller): Zannavit, nav. State of New York applications bcfore the Public Utility Commission, its (Zaonavi~na)]. All of these cases involve the efforts of major import is procedural rather than substantive. different courts to address public concorn over the The Court of Appeals' acknowledgment that, based on health effects of EMF. None of the findings deals the expert testimony presented, the Klein jury could directly with whether EMF is the cause-in-fact of have believed that the utility exercised poor judgment, adverse health effects. Rather, they deal with peoplc's appears somewhat removed from the OTA rcport's apprehensions about such effects and, in two of the suggestion that Klein constitutes a call for rcguhtory four cases, with whether there are any reasonable action to protect citizens. Similarly, the OTA report grounds for this apprehensiveness. cited the "Csncerphobia" case ~ v. ,~tate of 7-8 Heahh Effects of E~cposure to Powerh'ne Frequency Electric and Magnetic Fields New York) as another dramatic incident of pressure for growth of research funds that typically accompanies the public regulation of field exposures. This was a class- transformation of a scientific controversy into a legal action suit by a group of land owners asking and political one. On balancc, when it comes to EMF compensation for losses duc to the apprchcnsivcness of issues, law and science seem to have a symbiotic buyca at the proximity of their land to a 345-kV relationship. While the questions surrounding EMF transmission lmc; the kcy issue was which one of two health cffecu may never be fully resolved to opposite holdings (the San Diego case, where damages cvcryone's satisfaction, it is unlikely that a stable, were awarded, or the Miller case, where they were public resolution will emerge from either research or denied) would be adoptai. The judge found the Miller litigation alone, but rather from their interplay, perhaps prccaicnt the more compelling of the two and rejcctal medistcd by other public institutions. the land owners' claims of buyer apprehension for lack of sufficient evidence. This case, if anything, Up to this point, th~ discussion ha relied on the undermines the warrant for regulation, assuming (as the concept of judgmental bias to epitomize the ways that OTA report seems to) that public concern is oftat a clifferr. at social and political institutions can manage prime cstalyst for reguhtory scdon. evidcacc and opinion on i~MF health effects. Emphasis ha been placed more on the concept's heuristic value The policy signifwanee of the 'Csnccrphobla' Cssc than on its predictlye validity, however, the intent is not porlmps ruides more in its scale than in its having to model institutional behavior, but to provide a upheld Miller on the matter of buyer apprehension. plausible framework for making sense of the diversity One reporting service notai that almost $3 million was of institutional responses to the EMF issue. Beyond spent on this litigation, with a substantial share going to the workings of bias, institutional diversity can be expert witness fees (Mien'ware News, 1989); this traced to more fundamental differences of interpretation figure cxccats the annul budget for EMF research of and meaning that will affect how the EMF issue is beth the U .S. Department of Energy. in cases with multiple understood and translatai into public policy. chimants and a corporate dcfenchnt, the stakes will typically be high enough to justify such large The next sub-section selects two institutions (science expenditures on both sides. Certainly. the law firms and the courts) that play a prominent role in casting the r~prcscnting these clients stand to gain financially but, public features of the EMF iasuc but, st the same time, in legal procecclings on EMF health effects, another represent extremes of enntruting interpretation, uses of group of professionals stands to gain as well, the language, and approachea to reasoning. While scientific cxpofi withessea. substantial literature addresses the fate of scientific evidence in the courB (O'Bricn, 1987; Huber, 1988; The hat scientists, however, do not necessarily make Smith and Wynne, 1989), the discussion here the hat export witnesses and vice versa, from the emphaizcs differences in the meaning assigned by litigstor's perspective, since the necessity of adopting each institution to certain terms, effectively providing legal conventions in addressing scientific matters the footing for disparate interpretations. Following requirea the skills of a translator and not princilmlly thesc comparisons, the focus shifts to thc more gencrtl those of a scientific investigator. Furthermore, tsking issues of intorprctation and meaning tied to the sides in an adverurial proceeding may require the kind definition of thc EMF issue as a public policy problan. of categorical ststcment that crosses the boundary from what an be scientifically supported to whet an be 7.2. Contrasting Interpretations by argued, quickly csca~ting from scientific claims to Science and the Couds mcta-scicntific ones (]V[crz, 1990). Statements by scicntific exports in thc Klcin easc, for example, holding that cbimants "would probably cxporicnce an Boffi science and law employ a complcx proccdural incrcued risk" or "a significantly increased risk" and logic as a means of justifying deeisions bes{xl on that such risk is "unncccssary or "indefensible", or uncertain evidence. Both rely on elaborate rules of convcrscly thM 'thcre are no significant biological cvidcnce to structure thc drawing of infcrencc=, and cffccts", clcarly havc ~bandoned scientific convcntions both erect high proccdural thresholds for proof to avoid in favor of lcgal or regulatory ones. Contesting fal.e positives. And while both embntcc advocacy scientific findings in an edvemHa] am dmt admits directed to onc's pocrs, science tests ~csc claims only two sides of the issue appeus andthctic41 to the through an iteretivc process of consensus formation, notions of reasonable doubt, unccrteinty, and porhaps law ~rough an edvcruria] onc. Despite their ambiguity, tlmt LrC prominent features of dcbstc within functional similarities, however, sevcrtl buic scicnce. differences cm a gap between them when it comes to thc handling of uncet, ainty. Still, ~c distortions th~ litig-tion can impose on scicntiSc questions may be more than offsct by the Policy Issues and Options 7-9 that, 'Statistically, this finding was of marginal 7.2.1 Meanings Assigned to significance, however' (BPA, 1989, p.43). A,s a Final · Significance' cxamplc, considcr the confusion crcatcd by confounding the two notions, u when an cxpcrt witness A pnmincnt obstacle to thc appropriate uac and in ]-Iouston ~ and powcr Company v, Klein intcrprctation of scicnti~c evidence by the law and |ndcpcndcnt School District inferred from statistically policy making institutions is the dis~rity in mcaning eignificant cclhlu cffccts, a 'significanUy incrcascd bctwecn ordinary and tcchnical usages of thc ~mc risk', Contrary to thesc uses, statistical signi~cancc tcrms, The tcrm 'significance' is notorious in this has no bearing on either thc 'chances of harm' or thc regard. Part of the problem can bc attributed to the strength of any postulated causc and cffcct rchtionship. attachment of science to a frequency interpretation of probability, wherein chance is deemed in terms of The only valid interpretation of statistical significance is repeated outcomes. Policymskcrs, in contrast, tend to as the outcome of a test designed to answer only one cmbrscc a subjective notion that ties chanco, not to question, 'is the observed result duc to chance as wc random processes, but to credibility and degree of have modeled it, or not?" if the chance model belief. A significant result to policymakcn then is a employed cannot sccount for a given result, then that cr~iiblc or belicvsblc result; in this sense, one result ruult is deemed 'statistically significant,' and likely may be robtartly more significant than another. A the product of something other than chance. This credibility interpretation of chance and uncertainty has designation has nothing whatsoever to do with the traditionally informed legal judgments about the probity magnitude or importance of the result and cannot be of evidence, and varying standards of legal proof arc used to calibrate its credibility; it is an aU or nothing often phrased in these terms. determination that admits no measure of degree. Further, the appropriateness of the chance model and A 'prcpondcranco of the evidence', the standard of its link to how the data wcrc generated must be proof in civil proceedings, for example, implies a stipulated in advance and then taken for granted in greater than 50 percent certainty regarding the truth of testing for significance. If the data wcrc not samplcd in the plaintiff's claims. Legal writings have also tied, ways consistent with the model of chance used to 'beyond a roasonablc doubt', the standard for proof of define statistical significance, then t~ts for significance guilt in criminal trials, to being at least 95 percent performed on these data could easily be inappropriate. cortain of thc truth of the prosccution's charges before guilt can be established (Large and Machie, 1981). By 7.2.2 MesninOs Assigned to 'Csuse' and coincidence, a 95 percent standard is also invoked in 'Confidence' scientific trcalxnents of c'vkienc~ but with a very different interpretation. As will be discussed below, the 95 percent level in science is not so much a truth As wc will scc below, much of the causal analysis standard for judging claims as it is a critcrion upon underlying judicial proceedings involves a search for which to bag predicfivc judgments. The former asks, th__c precipitating event, known as thc 'cause-in-fact'. Evidence about this cause is weighed relative to its 'how cotrain arc you about this clsim?' while the lsttcr ash, "what poregntagc of the time do you expect the crgdibility and to the probity of those testifying in its behalf. Proof of the event thcn is synonymous with evidence supporting it to recur?' Note that both admit bellcf in the cvidenco; the stronger the belief, the there is a chance of being wrong. For the law, higher the standard of proof that can be tact. Evidenco however, being wrong is cather a cognitive failing or a supporting a claim about cause-in-fact beyond a decc~on, while for sciencg, it is strictly controlled by rcssonablc doubt is cvidenco about which one can be a hypothetical model of future prospects (Latin, 1982). (approxinmtoly) 95 percent cortain; in other words, there is a 95 potcent chamcc that the alleged cause-in- Confusion of the two notions of signifiesnee plagued fact was responsible for the event in question (Scc Judge Poulton in Rausch v. The School Board of palm Brcnnan and Cartor, 1985). In contrast, scientific Beach County when hc interpreted the significance trcatmcnt of causal inference is very different. lcvcis of the cpidcmiologic cvidcncc on EMF health effects as a calibrafion of the chances of ham noting, '...[even] a 1 percent chamcc that there is substantial Causal inferences arc based .on a summary mcssurc of damget is unacceptable' (quoted in Banks, 1989b). the relationship between an alleged cause and the event Surprisingly, the scientific community is not immune to of interest. The measure itself ottcn appears as a range of values intended to reflect the limitations of the this confusion as the science intrudes on the policy chance model and the mount of information available rcslm. The Bonneville Power Administration EMF for calculating such a summary. Noncthclcss, the report, a model exposition in many respects, appears to make the same mistake as Judge Poulton when it frequency notion of probability dictates that the observation itself not be subject to chsncc or concludes from a review of community cancer studies uncertainty; only the sampling procedure can be. 7-10 HealtA E,~'ects of Expos-re to Powerline Frequency Electr~ a~ Mag~etic Fie~ A~hgly, s~men~ a~ut ~e m~u~ ~d i~ hve ~mpu~, s~ply cross over ~e ~unds of a~uncy must ~ ~nfid~ s~en~, not pmbab~ty a~ble sc~nfific ~fe~ ~d a~, for ~e most onfl. A ~nfiden~ level of 95 ~t ~di~ ht a ~a, unable. lmni~Hy, ~ven ~ of sum~ ~, ~und~ by ~e ofjud~ ~at ~vc ~c g~t legal ~. high~t ~d !o~t valu~ ~t ~c sum~ mi~t usumc, ~ cff~vcly ~ dm~ ~ ~dom ~m a ~ ~ mi~s ~m~ty ~H~on of such ~ ~d ~ ~n~ ~ "~c" ~fc~ ~cr or ~puhfion valuc m 95 out o~ I~ ~s~. ~ 5 ~v~fi~n, hw d~ ~c o~si~. ~w ~n~iv~ of not ~c Mmc ~ ~y~g ~ one ~ 9~ ~t ~m~ ~c ~gcly ~ Ic ~gc of sum~ ~ ~n~ ~c ~w cxp~o~ m ~c s~glc ~t cyst or ~pu~on valuc. N~ only ~ ~c ~gc of sum~ ~ndi~on ~ ~ ~t ch~ ~ usi~cnt of u~a~ chngc from Mmp~ ~ ~mpic, ~c~ ~ no ~nsib~ty; it adm~ no dc~ of ~u~ty or way s~Hy ~ dc~c ~c c~ ~t ~c ~puh~on ~ty ~ valuc ~ ~c ~ ~ ~al, cider it d~ or h do~n't. a ~r of ~cy ~ut ~c c~nt of a ~v~ ~y's ~volvm~t. Un~m~ty ~c lcgal p~s ~ not ~H-sui~ ~ ~d~g ~c pmci~Hy ~m do~t ~ut ~nclusiv~s of c~ ~cb, such ~ ~dom ~ ~c ph~ib~ty of ~c~ ~m~ ~n~g ump~g, ~Hy when ~ ~mu ~ ~b~h~g ~c s~m~B. p~bity of cx~ ~ony a~ut ~us~-fact. ~c cx~*s ~ of a ~h~onship, oy, a ~h~vc ~k ~ ~c~ ~vc p~uml lo~cs ~g~cr ~ iu high~t ~d ~w~t phusiblc value, ~sfcr of ~owl~gc ~ ~sh~ by ~c ml~ of cvidcn~ ~ ~c mo~ ~t, st~. On ~ a ~t gu~s ~, 1982). ~mc q~fi~on ~ ~b~h~ ~out bcnc~t of scicn~c ~nclusiv~s; ~s of ~c cx~'s ~h~vc ~ty a~ut it a~ whHc on ~c o~cr hnd, cv~ ~cn sci~ t~i~Hy ~u~t, but ~y ch~ va~b~ty ~ ~n~m~ ~nclusivc, it may ~ ~c obg~cr ~d not ~ w~t ~ obsc~. A ~va~iy 1~ ~nfid~ ~al ~cn ~ ~cly ~ ~nvcy ~c cvid~. To hold up ~gafion u a ~r of pub~ ~v~ga~r's ~n~d~ ht ~c "~c" ~ fa~ ~n~m ~d ~cn m ~H on gi~ for ~c ~vc ~ ~c ~ ~gc; a 9~ ~nt ~n~dcn~ level ap~cmcnt of ~ ~ ~s m~ ~c hi~t s~ of p~f ht su~t) ~ ~ m~d ~ ~ic ~. ~c legal sys~ ~s~. Unfo~u~ly, ~c ~nv~fio~l su~l fo~uhfion admi~ no 7.3. ~etoric and Public os~sm~t of ~n~ ~ ~c ~ of dcg~ of Interpretation ~cf or ~vc ~ty ~d ~n~m~ c~ va~b~y, n~ ~ ~c ~vu~r or ~ ~c ~hfionship ~g ~v~tig~, but cxclusivcly ~ ~ omp~g ~c d~g over EMF h~ p~u~. ~1o~ and rompbur ~at a ~cuhr ~on ~c s~t, "my ~h shoM ht ~c '~c' pm~s~ ac~n ~ wa~. C~cs of ~c civic ~h~vc ~k ~uld ~ ~ ~ ~ ~e value 'x' but ~ ~r ~du~, for c~plc, ~y highly un~cly ~ ~ ~s ~ ~c valuc 'y'", foHo~ sym~l of ~ ~d~'s ~cg~ d~g~ for by,*...~ ~ only a 1 ~ 20 c~ ~ I ~uid ~ pub~c's ~lfa~ or ~ng* ~ a ~cy ch~ and not a s~l uric. Bo~ sing of ~sm~sion ~. In ci~cr ~, EMF s~s~ of *~n~d~" a~ ~mb~ ~ a pmvid~ a usc~l way of ~u~g ~ c~cs ~md m~gl~s ~m~si~. S~Hy, it ~ ~c obj~o~blc ~ut ~c ~dus~ or i~ p~. ~nfidcn~ ~al i~clf (~c ~gc f~m ~c h~t value 'x' m ~c smaH~t value 'y') ht ~ ~c ~dom in a f~ ~, ~c romphot of "fields* ~ ~sfo~ ~m~ncnt sublet ~ c~cc va~b~ty and c~r. If a ~m "~v~iblc ~ys* ~ ~c adv~Ugc of ~c pubic hrgc humor of ~ab n~ ~mpu~ ~d~nd~tly f~n ~d ~i~on on ~ ~s~ns, ~ch ~ ~c ~mc ~n~dcn~ Ti~ such as lcvcl, uy, .95, ~cn ~ ~c long ~ of scvc~! hund~ ~m~s ~c modern ~mation of ~ ~cicnt a~ach or ~ ~, 95 ~t of ~ ~ab ~ ~vcr ~c m d~g ~ un~m~ty ~mu~ ~c m~ium of ~e '~' val~. ~c c~ p~s ~ ~ ~s~ ~ mon~ty phy (~ ~, ~ ~., 1~; B~cur, 1989, ~ ~mu~blc ~d ~~, ~ ~ ab~lu~ly no 1~). ~ of ~g on wh~cr ~c ~vu~ga~r should ~ ~s~ or ~p~cy of ~ ~ pmv~ ~ng ~ ~c ~t. Hunch~ a~ut tm~, cx~ for ~c vast cffom of a f~ ~b~, ~w ~ hc or she ~, or ~ut how ~cly ~ ~ ~t ~d ~ on. ~pi~ ~c ~dity of · c *~c' ~k a~Hy a~n ~ tc ~ffal ~cy m~c of ab~g ~ty ~ ~cnl ~v~m~ Policy Z~sues and Oplioas 7-11 over the others mentioned above. It simplifies and and adjustment costs (Morgan, et al., 1988). imposes a measure of order on complex developments Precautionary action thcn becomes justified as a cost- while appealing to a well-tested formula that evokes avoidance strategy on the part of the utilities, despite basic sentiments common to many of our cultural the flaws in the analogy between opposition to "fair myths. Few would argue, bowever, that its claims market" appraisals in condemnation hearings and provide an appropriate basis for making public policy. public concerns over EMF risk. 7.3.1 Annlo~y 7.3.2 Tho Political Contost Ovor Problom Moanin9 There is a great deal of eftoR devoted to making EMF familiar by comparing it to other hazardous agents. Fundamentally, the kind of language used to Such comparisons frame our understanding of the characterize a problem and delimit its solution is issues surrounding EMF's possible health effects and, contingent on how the issue was initially framed. For more subtly porhaps, evoke criteria and control the EMF issue, the initial assertions of claims and experiences that at least appear to be relevant. grievances were prompted primarily by utility industry Comparisons with everyday risks, for example, place efforts to site · new generation of higher-voltage possible health effects in the context of useful activities transmission lines. The organized opposition to these and establish a connection between EMF sources and siting plans typically grounded their claims in the those that are both familiar and widely tolerated. While language of rights and health, and adopted a moral on the surface, attention is focused on numerical vocabulary that would prove to be more potent than an comparisons to establish a reference for the magnitude assertion of simple self interest (Furry, et al., 1988). of risk at issue, the unspoken analogy works on the These defmitional activities appear to have proceeded qualitative aspects of the EMF risk, making it appear on two fronts for official recognition and intervention. less mysterious, less threatening, and tied by In the courts, claimants attempted to redefine siting association to other beneficial or voluntary activities disputes in terms of threats to health or the well-being (Douglas, 1988). of mothers and children. As noted in an section 5 most of these efforts were unsuccessful in winning official Analogies not only color our interpretation of the suction for this definition; howover, involving the available evidence, they can circumscribe the kinds of courts brought several kinds of professionals into the institutional responses deemed appropriate. defmitional process. The fwst and most significant Comparisons with environmental hazards, such as air were the scientist-entrepreneurs who discovered · and water pollution, 'evoke criteria that suggest a welcome opportunity for doing career enhancing work threshold for possible effects, thereby making EMF in the construction of supportive scientific evidence. amenable to conventional control strategies, such as The media were also introduced to the defmitional regulatory standard setting or emission fees. conflict and played a role in channeling it into a Moreover, once the analogy takes hold, the arguments controversy among competing spokespersons for the over alternative responses must be dealt with on their affected public and the utilities. merits. The question of whether such strategies are even appropriate is effectively displaced by the debate The second front was in the formal proceedings of the over which one is best. administrative agencies responsible for granting siting approval. !n this context, · second round of Analogy also operates at the level of institutional defmitionalactiivities grebegun. Rather than debating tradcoffs. Here the focus is on emphasizing the risks the issue as framed by those initiating the claims, of one kind of error and, perhaps at the same time, agencies invited · wide assortment of experU to minimizing the risla connecuxl with the other. Selected participate ia its re-definition as an empirical matter to examples of past errors and their consequences are be resolved, not through public debate, but through presented as precedents that bear on the EMF case and scientific investigation. In several states, most notably, ugd as the basis for justifying · given response. New York and California, the utilities wcr~ asked to Arguments for precautionary action of different kinds, underwrite the costs of this defmitional change, in some for example, typically place a ceiling on the economic instances, as the price of interim siting approval costs of · false-positive error (taking action later (Banks, 1990). The Electric Power Research Institute, proven needless or unjusti~ed) and · floor on the a consor~um of utilities, currently funds the largest political costs of a false negative (taking no action to share of these investigations while trying to guard its address EMF concerns when, in rcffospect, action neutrality in the awarding of funds to avoid even the should have been taken). Estimates of the latter costs appearance of · pro-utility bias. Nonetheless, it has are extrapolated from the utilities' experience with helped to create · community of experts who have a public opposition to transmission line siting or upgrades vested interest in EMF as · research problem and to and may be inflated to include prospective settlement 7-12 Health Effect~ of E~posure to Powerline Frequency ELectric and Magnetic Fields expand the scope of the original issue beyond the siting practice, it is instructive to consider which definition of transmission lines. predominates, especially in discussions of what might happen if the prcfcrrod p~licy options arc not adopted. So successful has been this effort that the administrative agencies in several states have tried to regain control The EMF "problem" promiXing action has been framed over the dcfmitional process by calling for a federal in several different ways, each with a distinctive program of funding that would emphasize rcscaroh into emphasis that suggests the kind of remedy needed. As interventions via control tcchnotogy (· farniliar suggested earlier in our discussion of the range of rcgulatory regime) and public information (Task Force institutional responses to uncertainty, each problem- on Transmission Line Health and Safety, 1989). As remedy pair conveys · particular *bias* regarding the long as research funds arc available, the scope of the more likely kind of caTor (false positive or false issue will probably continue to expand, tak~g in an negative) and its possible toll. Some policy advocates, increasingly wider range of exposures, including in common with some institutions, find the avoidance household appliances and worksito aluipmcnt, and of false positives in dealing with uncertainty to bc the potential health effects. In the absence of · conclusive primary concern; others fred false negatives to bo resolution ruling out any harmful health effects, the equally abhorrent. Similarly, each problem definition disputo over EMF may ultimately turn not so much on betrays · basic assumption about how scie. nco relates to persuasiveness of scientific evidence, as on the public policy making. Some fred sciattific consensus costliness and differential impact of the suggcstod on technical matters to bc authoritative and decisive (or interventions. However the scientific uncertainty is ought to be) in policy deliberations. Others fred such resolved, at some point the 'question of acceptability consensus compelling, but often not entirely relevant, must bc faced. As · practical matter, the available and thus more contributory than decisive in interventions and their impact will prompt yet another deliberations on public issues. Still others fred such rcdcfmition of the EMF issue, this time in terms of the consensus, especially on rcguhtory matters, to bo feasibility and acceptability of its control. hopelessly biased by its sponsors and partisan rather than contributory in its influence on public debate. ]'.4 Contending Definitions of the Public Policy Problem Extcnd~g the earlier argument about institutional biases and contrasting interpretations, the selection of policy options will ancrgc as · product of the way the Most of the EMF investigations that explicitly offer problcm-to-bc-rcmediod is framed. At least four public policy recommendations draw on a distinctive distlnctivc problem definitions can bo found in the dcf'mition of the key problem in nccd of remedy. For literature addressing policy options. Each represents many, the problem at issue is not the presence of coat·rating intctprctivc and valuativc commitments, adverse health effects per sc but trouble with how wc expressed hcrc largely in terms of institutional bias and accommodate the uncertainty surrounding such effects. · favoral set of institutional roles. From this The remedy to this kind of problem basically involves perspective, efforts to depict the problan of choosing either stepping up or changing this mode of among policy options as either · technical problem to accommodation. In other words, either wc arc not bc solved by selecting the proper analytical tool, such doing the right thing, or not enough of it, to ·void as cost/benefit analysis, or s context-free choice among trouble. Each problem dcf'mition and implied remedy disembodied rcgulatory regimes ·re misleading at best is associataJ, in turn, with a profcrrcd set of policy (Scc %V'trick, 1990). Instead, policy options must bc options. Not all policy properic·t· t'md it necessary to judged on the basis of the commitments to invoke principles, such as equity, acceptability, or interpretation and certain values they represent, as well prudence, to support their choice of policy options; as for the reasons offered in their favor. A summary of those that do, typically have · larger gap to close the four definitions and their respective commitments between the obvious remedy suggested by their appears in Tablc T.2. problem definition and the policy options they favor. Sovcral illustrations will 5c offered. While more than one definition and set of options is typically invoked in Policy ]ssu,s and Optiox,u 7-13 Table 7.2 - Problem Dcfmifions & Policy Options 'RISING TIDE OF h'dlm *PUBLIC CONCERN EXPOSURE CONTROLS "PUBLIC HEALTH "INCONCLUSIVE !)erailia [sSb]" Is&b]" EMERGENCY" SCIENCE" Remedy a. Information Amicipalion & Initiative Regulation Research b. Symbolic Action Principle a. Right-to-Know a. Prudent Avoidance Lowest Achievable Conclusiveneu b. Equity b. MitiSadon Exposure Ltvels Options a. Conmmnica~on a. Reduce Exposures Health=Based TargetP..d FuMing b. R-O-W Field Limits b. En~inec~n~ Fixes Standards Bias [Type of a. False Pos > False Neg False Neg > False Po$ False Negative False Pos > False Error to Avoid] b. False Nc~ > False POs Ncg Authority & Compelling & Authoritative &Decisive Biased & Pa~isan Authoritative & Policy Role of Comtibumry Decisive S~ence Slate F, xamplcsI a. New Jergy a. Colorado lllinoir3 California b. Florida b. New York2 2 New Yodt lms charged Og u:iliti~ ',, develop bo~h mi~mion me~ures and an impic,,,--t-don pkn by 1993 3 Am lllinok House ibnolution (11064, O,aober 30. 1989) adopa ~ dd'mition in mat napgas but falb s~ort of manda6ng k,.-m,-baaed ,Umdard$. political retribution of some sort from unrequited public concern that might eventually prove correct weighs heavily on thc choice of policy options (Scc Wirick, 7.4.1 Growin0 "Public Concorn" 1990). The notion of welfare-based, n/her than health- based, EMF limits u a policy option can best be The growing level of public concern poses either an understood in this light. The political logic behind educational or · political problem, depending on one's wclfare-bued limits, however, is not to improve tradeoff bctween false positive and false negative welfare in a material way, but to keep it, or more erron. Those finding thc prospect of · false positive preciscly, perceptions of it, 'from changing for the the rare objectionable of the two kinds of errors will worse. Limits intended to forestall any decline in tend to promote public awareness and the provision of physical or psychological well-being connected with information on both sides of the health effects question; EMF naturally cabrace the status quo as their the untoward consequences of the absence of public benchmark. The costs of complying with limits defmed information, however, seem to bc expressed more oRcn by the status quo typically lic in the future, while the in terms of public over-reaction to adc minimis risk symbolic benefits of taking seemingly dramatic action than u under-reaction to · substantial one (Sandman, to contain an 'unemuin threat to the well-being of the 1989). Accordingly, thc machinery of risk public' lie in the immediate pruent. Further, in communication should bc engaged, with its capbasis fashioning such limits, there is no need to sort out the on addressing scientifically-unfounded, public fears. complex technical issues of exposure and dosc, nor to Fur, ha, from this point of view it is the experts' upset the existing balance of political interests both calibration of actuarial risk that defmes thc content of insidc and outside of government. Science, in this the message, with few exceptions, and not the poblic's instance, makes a contribution to these deliberations. views of thc social elemcnts of risk (Juanoff, 1986). Yet, it appears to hold more relevance for assessing the Science, in this instance, is the communicator's odds of a false negative than for determining what primary authority and aspires to be decisive on ali policy option to punuc. n~___-_ers of environmental hazard. Besides their political cxpaliency, wclfare-bued limits Convmely, those finding · false negative more rest on a presumption in favor of treating existing troublesome tend to advocate some kind of symbolic lcvcis of exposures as de facto both safe and lcgitimate. actinn to allay public fears. Here, the prospects of Allusions to "acceptability' and "equity" u principles 7-14 Health Effects of Exposure to Powerfine Frequency Electric and Magnetic Fields linking theproblem of publlc concern to welfar~med intended to anticipate the went in the face of limits tend to reinforce this presumption (See OTA, uncertainty; if EMF proves to have significant health 1989). The moral weight of acceptability as a principle effects, an early preparatory r~ponse might not only is tied to the tenuous notion that the gatus quo most cushion the electric power producer against abrupt accurately captur~ public prefer~nce~ regarding policy changes, but also preempt more coercive exposures to risk; the pattern of cxisting exposures then measures. In the werds of one proponent, "Unless is presumed to be acceptable and can thus be used as · anticipatory engineering and economic studies are benchmark for establishing parity across all sources of undertaken now a lot of wasteful and ineffective effort exposure. To ·void the appearances of expediency, can be expected in · scramble to take protective this account must assume that: 1) risks unknown to the meuures if and when experimental evidence someday public at the fine preferences were revealed in the clearly links fields to he~.th risks" (Morgan, 1990, p. status quo ar~ effectively zero, 2) risks fxom different 123). sources are perfectly comparable and can be in terms of a common numerato such as mo~ality rites, Yet, the key to fashionsag an anticipatory remedy is to and 3) there is a safety threshold above which, more decide how much to do and when to do it. The model exposure always means greater risk (Schrader- of technical rationality behind most economic planning Frechetto, 1985). In addition thero must be · plausible often · simple guideline: keep the response argument linking the acceptability of the risks proportional to the expected risk. As a principle of embodied in the status quo to some notion of consent. social investment, this guideline holds that the amount None of the~e conditions is met. spent on exposure control should reflect the value derived from avoiding exposures. Under uncertainty The moral content of equity as · principle in this over the eftwAs of exposuro, however, the value of instance depends on how one interprets the claims of avoidance itself becomes unclear, and we are lea back those who feel they have been wronged by the status where we started with the notion of avoiding the quo. If the status quo is blameless (although it seldom consequences of a false negative. Nonetheless, the is) then future wrongs can be avoided by options connection between proportional responses and the grounded in this principle. Otherwise, equity must canons of rationality renutins. The term, prudence, function as a compensatory principle, shifting past commonly used in rate reviews as a criterion for burdens as well as reassigning new ones. Claims that judging the appropriateness of investments, has been welfare-based limits can be justified out of equity extended to cover policy options that, in proponents' considerations assume that additional future exposures judgment, constitute a proportional response. are the only relevant feature of exposure inequity. Ambiguity also surrounds the concept of prudence Furlher, once public action is taken to define existing applied to investments in avoidance, since one's sense levels of exposure as permissible levels, these levels of proportionalsty will vary with beliefs about the become legitimate by intent rather than romaining an certainty and level of the risk. Short of knowing the unintended byproduct of decisions mad~ with goals value of avoiding exposures, we are leR to anticipate other than equity in mind. Such action imparts a false the eventual outcome of the scientific debate and to sense of confidence in the protection that those hedge againsta falsenegative, accordingly. porrnissible levels afford. Moreover, any subsoquent action must then be phrased in terms of changing Depending on one's predictions about the science and existing policy and arguing its inadequacies. In short, aversion to false negatives, two levels of hedging are nut only is the rationale for welfare-based standards possible, each interpreting prodcoce in a different way. weak, but their use may ultimately prove The aggressive hedge interprets prudence expansively: counterproductive. prudence in the choice of avoidance options calls for investments in practicable engineering controls that will 7.4.2 "A Risinl] Tido of ~xposuro mitigate EMF (OTA, 1989). This view anticipates that Gol~trol$" science will eventually support the need for field reduction and that, unless pr~"mptive measures are The prospects of hasty control measures by public taken now, coercive measures imposed by government authorities and of numerous compensatory awards by are likely in the future. A more passive hedge the courts have focused attention within the electric interprets prudence more narrowly: prudence suggests power industry on how best to avoid the harsh that people be kept out of fields whenever it is consequences of a false negative error in judgment inexpensive to do so. Reiatodly, people should be (Morgan, 1989; DeCiceo, et al., 1989). Unlike encouraged to avoid fields when they can. This is a responses to public concern as · political problem, common strategy among advocates of lifestyle changes. however, anticipating pressures for exposure control From this perspective, engineering controls may appear raises the problem of planning for economic and out-of-proportion to the risk that EMF presents, but engineering changes. Remedies, in this instance, are Policy lssutJ and Options 7-15 marginal changes in personal behavior or in the siting fields. Here, the possibility of a false positive has been of new facilities seem to offer a reasonable hedge. ruled out sltogethcr, and those who ·dmit such a possibility arc viewed as obstructionist (Brodcur, For those who frame the EMF problem in this way, 1989a, 1989b). This perspective has a very different accepting one version of prudence over another view oft he policy role and status of science. There is (aggressive mitigation over expedient avoidance)comes no quest for scientific conclusivehess, no need to down to a priori beliefs and apprehensions. In either pre~dicate public intervention on the emergence of event, the in2nt is to stay · step ahead of any rising scientific oonsensus, and no reason to postpone action tide of control expectations without stepping too far out until more is known about what should be done in this in front or in the wrong dit~tion. In contrast to the particular instance. All that is neeclod is some sugges- policy option of field limits described abovc, prudcnce Live evidence and avenuc in which to advocate thc dictatos departures from thc status quo, but ones snail lowest achicvable cxposurc lcvcls. At the root of this enough to be corrected if htor they prove to have been impatience, rclative to the other definitions of the EMF taken in error. Further, it advises policy making problem, is a mistrust of the scientific community and institutions against acting too prccipitously or out-of- skepticism about thcir chim to objectivity. turn, in deference to thc decisive role that scientific information should phy. For thosc focused on public The scicnti~c community's aversion to false positives is concern as thc primary problem, however, prudence taken by thc advocates of the 'contrary" view as · ofters little constructivc guidancc. For Utility convenient disguise for a political and social Commissions that trsditionally havc employed siting conscrvatism horn of the need to protect the sponsors rules that call for the avoidance of population ccnters, of research. In other words, thc scientific community historical sites, and cxisting facilities, thc morc is not only corruptiblebut the findings ofrcsearch arc reasonable of the two hedges against EMF risk may especially prone to the influences of ideology. Under alrcady appear to be in phce. these circumstances, the influence of science on public policy is scen as essentially a partisan one, compcting In the contoxt of a myriad of well-documented risks on the same footing u self-serving chims of faced every day, and incrcasing sensitivity to thc health stakchoidcrs. The irony in this view is that if thc implications of each consumption dccision and results of science arc principally a product of cxtra behavior, a vague prescription to people to be prudent scientific influences and to be mistrusted, then therc is about EMF, just in case, sends · mixed messagc. On no special warrant for or against public action in any the one hand, the basic proportions of the risk that set of results. Thc evidence supporting chims of · EMF might posc to penonal health arc viewed as still public health emergency becomes dcvaluod in the same open to question; on the other hand, advice about way as thc cvidence rcfuting them. limiting certain kinds of personal exposures is fordscoming, as though some pragnmtic approximation Onc can, of course, usign pedigrcc to scicnti~c rcsults of the risk wcrc aH that was necessary. Thc thrcshold batsod, not on thcir methodological rigor, but on their of safe exposurc rcmains underrood, and ad hoc ideological purity, that is, on whether they support guidelines for field avoidance arc indefinite as to one's. point of view. Nonetheless, the marsh·ling of whcther the prescribed changes in behavior will do any evidence and its use in argument will follow very good and, if so, how much good. Counseling prudence different rules, according to this construction, and will may make sense in situations wherc forcsight and often bc subject to partisan tests of legitimacy. From hodgod investments can savc money. As · strategy for this perspcctivc, the sponsorship of · study and the addrcssing public concern, however, raising the specter ·fftliation of its investigators will matter morc than the of an illusive threat to be*countercd with · few changes quality of the design itsclf. In turn, the quality of thc around the house (changes that may or may not do any design will bc judged primarily on the partisan good) offcn a weak basis for · health promotion implications of its ruu lts- cithcr in support or in campaign. In health education and promotion efforts, opposition to a given point of view. the risk of premamr~ or misleading admonitions may vcry wcH outwcigh possible gains from guessing While the use of predictire judgment to avoid false corrt~y. negatives, especially when the public's safety appean to be at stake, is a bias common to public health 7.4.3 "Public Health Emergency' professionals, in this instance therc is a morc subtle bias at work. if we examine the history of efforts to Viewing EMF as a proven disease-causing agent control exposures to other everyday substanee., such emphasizes the moral and legal imperatives behind its as lead and asbestos, the definition of the problem, the control. Fundamental changes in personal behavior predictive interpretation of the scientific evidence, and and in industrial design and practices arc called for, the control options advocated, follow · very similar from this perspective, to minimize exposurc to harmful pattern. Erring on the side of safety appears to have a 7-16 Health Effects of F. xposurt to Powerlint Frequency Electric and Magnetic F~tlds gcncric quality to it, shapcd in large part by thc power potcntiil return (US Environmcntal Pmtcction Agcncy, of analogy, The f~ct that argumcnt~ for maximum 1990), Furthcr, the funding and targeting mu~t bc regulatory controls, in thc face of scicoti~c accomplished in such a way as to placate thosc who inconclusivcncsa in the casu of lead and asbestos, have less confidcnco in scicncc's ability to cithcr settle wcre cvcntually born out by later cvidcncc rcinforecs EMI= policy questions or ·void sponsor-induced biases. this pattcrn, as well u the bias against falsc negatives Thc proposed remcdy is typically to vcst thc (See Brodcur, 1974, 1985). In cffcct, thc problem responsibility for targeting and funding in govcrnmcntal definition transforms EMF exposure, via a fcw highly authorities with ex4zricnce in both sciencc policy and sclectivc analogice to earlicr succcss stories among thc resolution of tcchnical issues. Scvcral bills havc and-exposure campaigns, into yet another episode of been introduc_-a-'J in thc U.S. Congress to define a cnlightcned advocacyprotectingthchealthofthcpublic fodcral role in EMF ruearch and to allocatc against the prcdation of industry (See Murray, 1988). ruponsibility for funding and targeting (U .S. House of Representatives, 1990). Should no decisive findings Unlikc the ubestos and lead casu, where the debatc emcrgc alter an initial round of new federal invcsunent tumcd on the question of how much scientific cvidenco and rcsca~h direction, wc might cxpcct thc sarnc was neccssary to vindicatc thc call for drutically instiiutional biases oporating at thc statc levcl under reducing current cxposure limits, thc EMF debatc hu conditions of scientific inconclusiveness to slutpc thc not yet defined what exposure and dose mean. Where course of fedcral activity. Even if most smu stccr the former were arguing over diffcrent estimates of safe clcar of rescarch management functions, staying levels of exposure and proposals to ratchet down abreast of thc emerging scientific cvidenco renmins an existing regulatoty controls, the latter is still sorting out important state ruponsibility. what forms control might takc and whethcr safety will reliably increuc u exposure is reducod. Finally, thc 7,5 COnClUSiOn: Multiple shcer volume of the ruearch record on kad and Interpretations and ubestos exposures and the ovcraH magnitude of the ,~ve~ ~ eff~.~, do~umen~ therein ,hi. ~ Institutional Design credibility of any ~nlogy between these ca~es and EMF. Given the cun~nt stale of sciontiTle knowlcdgc, The princil~l question addressed in this analysis hu the definition of the EMF p~ohlem u · public health not b4~n EM!: health c~e.~ pot so hut r~hcr the emcrgcncy·ppe~n ill-found~l. emhiguRy surrounding lhcm and the diveairy of int~retations ~ resu!L In general, the 7.4.4 The "Inconclusive Science" indecisiveness bred by embiguity is · corrosive force on institutions, and on the public Lrust upon which thcy This fourth dcfini~inn of the proHem hogins ~ the de!~nd; it clouds with doubt the propcr public ~sponsc notion tint the science of EMF is ~ in ilx To pressing issues, representing · vacuum of soru that · dolesconce. Accordingly, until i~ r~ches maturity, comes To bc fdled ~ diver~ clams, but yielding no there will bc fdsc !c~ds, unroplicable ~su~, point of reference for their usessn~nt. ~ach of thc intr~nund coatlieu, and a large mcuure of bias., in~rpnnlions, and problem definitions pon~xl inconclnsiveness (California, 19119; hncl, 1990). A~ .bore represc4~L~ an ~r~npt to impose somo SLruc~Ure this point, i~ ls prennture to 'sek~ one set of r. ults on this situSion, to resolve somo mcuure of this over another in an nilerapt to antici!~Tc how things will ambiguity by means of usuml~ionn and recil~s. Once Tom out. The his undcr unccr~inty is to ·void false stal~l, however, these ~lt~rn~iv¢ consl~uetions merely positives; but ~c motive, in this insumcc, ix not so shift the ambiguity from one place to anothcr. Wc may much To escapo the cons~xiucnccs u it is to resist chim to know how I~st To weigh the ¢vidcne~ bearing precmiXing the full play of the scientific process. From on · problem, for exanple, hut arc lcl~ uking, 'which ~ pcrspe~Aiv¢, science when glvcn its full play can bc pr0blcm dcfmili0n is best?" an~horilativc, or at knst beneficial, and can exert · decisive influence over policy questions th~ hinge on Al~rn~ivcly, wc might r~considcr how scl¢c~l t~chnicnl issues. In the ideal, science can qmmtify lhc inslimlions ·e~ommod~c this ambiguity and find some cvMuation of options in risk-benefit terms. The rcn~xly reanurancc in nll~lding to u!~o~ of the problem they for inconclnsivencss then is twofold: ~ourc. to emphuizc. As noUxl ~rlicr, difforcnt ins~utions will underwrite the .e._-__--_sary mreh, nnd ~ to permit convey distinctive biases; some ins~itutinns, however, the developmental pxx~ss of sclentif~ i and crmr to will ~ommodat~ more than one of the ~bovc- mentinned, four typ~ of problem definition in ilx como to r,~u~y. clsin~ ·bout an appropri-tc public policy. Souling on a The key to resolving inconclusivehess, from this fourth bins leaves unresolvcd the question of propcr definition pcnpcctivc, in no~ more resetrob per sc but hoist- and policy stance. The bcst way to ·ddren ambiguity fundcd rescareh largeted to T~c a~ of g~lest in the ~bscncc of scientific sgrocmcnt may rest, not on Policy lauesandOptions 7-17 the arbitrary choice of one definition to the excltaion of If the primary criterion for determining which oth~ra, or on ~cl ho~ prioritizing by exp~rts, but on the institution should sponsor public discussion were its development of a public proems for discussion and deliberation to lay the foundation for converging on a judgmental bias, the Public Utility Commission (PUC) colle~:.tive interpretation and public defmitinn. In place might appear to have the advantage. It is a quasi- of a confrontation among advocate~, there might bca judicial agency accustomed to procedural protections forum for deliberation open to possibilities for learning and open proceedings, and it has cooperative relations and accommodation among advocates and uncommitted with the state's major public utilities and expertise in alike. Although some problem definitions can be addressing the legal and economic implications of viewed more favorably than others and some remedies policy issues. Judgmental bias aside, however, most of rejected as ill-conceived, u long as the proper its deliberations involving the public have been framed interpretation of EMF health effects remains in adverttrial rather than conversational terms. And ambiguous, these nmters will be contested and thus can while it is well-positioned to monitor the progress of EMF-related activities in other states and provide a benefit from public discussion, as well as expert advice. clearinghouse for technical information about EMF, the PUC in Textts, as in many states, is l~gally bound to limit its focus on electric power to transmission tines. To foster such discussion, an institution must have Advocates for defining the EMF problem as a public sufficient expertise to be capable of brokering the health emergency, moreover, might find the PUC's relevant science and yet able to balance the enunsel of relations with utilities and emphasis on economic technical experts with public values and concerns. Until the scientific community can offer a conclusive implications tq be inhibiting and exclusionary. Others who deEme the problem in terms of public concern, answer to the question of health effects, if such an might take exception to the absence of any mandate or answer is possible, the EMF debate will continue to experience on the part of the PUC for getting revolve ..round questions of interpretation. By information to the public or for handling field providing an arena within which viewpoints may be investigations of public complaints. Finally, those exchanged in a fair and open way, an institution may viewing scientific inconclusiveseas as the key EMF do much to stem the corrosive effects of ambiguity, problem might fred the PUC's -lack of scientific even if only on an interim basis. expertise in health research to be limiting. If we assume further that the interpretations of both The State Department of Health, while sharing only · science and the couas will continue to play as large a few of the PUC's procedural strengths, counters most role in framing the eventual resolution as they have in of its substantive weaknesses. The Health Department structuring the problem, then whatever institution has scientific and investigatory expertise in · wide provides the arena must also be well-equipped to accommodate the interplay of legal and scientific range of health matters. More imperamtly, it has extensive experience in conducting on-site community reasoning. While the courts have traditionally managed meetings to address local health concerns and the this interplay, they lave done so through highly capacity for educating and informing the public. structured adverurial proceedings that merely channel Nonetheless, until the scientific evidence on health rather than mitigate conflict. Procedural safeguards are effects is conclusive, the PUC should continue its surely impommt to the operation of any public forum initiative in monitoring new developments in EMF for dellbeating contentious issues. Nonetheless, the research. Given the balance of strengths and adversarial conditions accompanying the "case and weaknesses between the two agencies, the Health controversy" r~quirernents for judicial process can be Department should pursue · collabonttive arrangement divisive when applied to larger issues of public with the PUC for establishing a public forum on the concern, appearing to pit law against science or science EMF' issue. Such collaboration might involve the PUC against itself. What is needed then is an institution and its advisory committee on health effects in the familiar with ufegtmrds for ensuring the fairness of a development of information while the Health publi~ forum and yet one not bound to either an advcruritl construction or the formalities of Department assumes the lead in its dissemination and in evid~ntiary hearings. the conduct of the forum itself. Public discussion might continue through periodic forums until the ambiguity surrounding health effects has been resolved. 7-18 Health Efftcls of F. xposure lo Po~erline Frequency Eleclric and MaSnelic F'wlds References Administrative Procedure and Texas Rc~istcr Act [APTRA], Seaion 19A, Article 6252-13a Vernon's Texas Civil Statutes, Rcvised, 1999. Banks, R.S. The EMF Hcalth Effects issuc: An Overview of ~ 1988-1989. Robert S. Banks Associatcs, Inc. Minneapolis, MN: 1989a. ·Powcr Lines: Kids. Schools. and CallcCr- Robert S. Banks Associstcs, Inc. Minncapolis, MN: 1989b. · EMF ~ A Look at the prcnmt Status. Robert S. Banks Associates, Inc. Minneapolis, MN: 1990. Bonneville Power Administration [BPAI. Electrill and B~bsrical Effects of ~ Lin~. DOE/BP-945, Portland, OR: 1989. Brcnnan, T., and R. C, artcr. "Legal and Scientific Probability of Causation of Cancer and Other Environmental Discuc in Individuals," Journal of Hcalth p~litics, po~cy and Law. VoL 10, No. 1, Spring 1985. Brcyer, S. ]~e~ulation and its Reform. Cambridge, MA: Harvard University Press, 1982. Brodcur, P. Expendabic Americams. Ncw York, NY: Viking Press, 1974· · Outra~ Misconduct: thc Asbestos ]q~lustrv on Trial. New York: l~ntheon, 1985. · "The Hazards of Electromagnetic Fields, !-Power Litms," The ]q{~w Yorker June 12, 19~9, pp. 51-88. · "The Hmrds of E_-_b-J~o_magnctic Fields, H-Something is Happening," The H~pv Ygrker June 19, 1999, pp. 47-73. · Cur~nts of peath: Power L~es. Comvutcr Terminals, and the Attemvt to (;oWl' uv Their Threat to Your Health. New York, NY: Simon and Schuster, 1989. · "Danger in thc Schoolyard ," Family c~rcl~, scpt 2.5, 1990. Bupp, I., and J. Derian. Light watt: How thc HuCl~ar pn~am pissolvcd. New York: Basic Books, 1978. California Public Utilitiea Commission and Delmrtment of Health Services, Dntft Re~ort: potential Health Effects of Electric Power Facilities. San Francisco, CA: July, 1989. Ciaxkc, L. "Explaining Choiecs Among Tcclmological Risks", Social Problems, vo1.35, No. 1, FebnmrT, 1988, pp. 22- DeC:m, J.M., J. Boyea and S.S. Bertlow. "Environmental Concerns Rcgarding thc Siting of Transmission Lines." Paper prepared for Consumer Energy Council of Amcrica/Rcsearch Foundation, December 14, 1989. Douglas, M. Risk Acccmabilitv Accordin2 to the Social Sciences. New York, NY: Russell Sa~c Foundslion, 1988. Feller, I., et sl. "Scintti"""""""Tm and Tcelmological Information in Statc Legislatures" Ameri~ ~ Science, vol. 22, 1979, pp. 417-436. Finkel, A.M. "Is Risk Assessment Really Too Conservative?: Revising the Revisionists," COlumbia Journal of ~ Law, Vol. 14, No. 2, 1989. Policy Issues and Options 7-19 Fi~chhoff, B., et al. Acce~able Risk. New York, NY: Cambridge Univcnity Pros, 1981. Furby, L., P. Slovic. B. Fischhoff, and R. Gregory. 'Public P~rc=ption. of Ehx~tic Puw~r Tratmmia.ion Lim~', ~ournal of Environmental hvcholoEv Vol. 8, 1988, pp. 19-43. Gains, S. 'Science, Politics, and the Managemeat of Toxic Riala Through Law', Jurimeffios Journal. vol. 30, 1990, pp. 271-321. Gelpc, M., and A. Tarlock. 'The Usu of Scientific Information in Environmental D~cisinnmaking,' ~outhcrn California Law Review, Vol. 48, 1974, pp. 371-427. Houston ~ & Power Company v. Klein Independent ~chool District. 739 S .W.2d 508. Hub~r, P. l,~iabilitv: The ~ R~volution and Its Consequence. New York, NY: Buic Boola, 1988. haanoff, 8. Risk Management and political Culture. New York, NY: Ruigll Sage Foundation, 1986. · The Fifth Branch: Scienc~ Advison as policymakers. Cambridgc, MA: Harvard Univcnity Pros, 1990. Large, D., and P. Michic. 'Proving That the Strength of the British Navy Ik"pcnda on the Number of Old Maids in England: A Comparison of Scientific Proof with Legal Proof,' Environmental Law, Vol. 11, 1981, pp. 557- 638. Latin, H. 'The Significance of Toxic Hcaith Riaks: An hay on !,cga] Decisionmaking Under Uncertainty, "Ecology Law Quarterly, Vol. 10, No. 3, 1982. Lindcr, S, , and M, McBddc, 'Enforcunent Costs and Rcgulatory Reform: The Agency add Firm Response," ~lourlul of Environmcntal F, conomics and Manaecmcnt Vol. 11, 1984, pp, 327-346, McGtrity, T, "Substantive and Procedural Discretion in Adminiatrativc Rcsolution of Science Policy Qucstions: Rcgulating Carcinogcns in EPA and OSHA,' Gcorectown Law Journal Vol. 67, 1979, pp. 729-810. Mcrz, J. "Scicntific Uncertainty in thc Courtroom," Az~t~liance Eneinccr, Junc, 1990, pp, 94-99. ELF News, Microwavc News, Vol, 9 No. 5, Scptcmbcr/Octobcr, 1989, pp. Millcr v, Statc o_.f New York and Powcr Authority, 117 Misc, 2d, 444. Morgan, M,G. "Altcrnativc Rcsponscs That Utilitics Might Takc To Thc Possiblc Risks of 60-Hz Elcctronmgnctic Ficlds,' Unpublishcd paper, Dcpartment of Enginccring and Public Policy, Camcgic-Mcilon Univcnity, Pittsburgh, PA, 1989. · "Expos~ Treatrncnt Confounds Understanding of a Serious Public-Health issuc," Scientific Amcrican, April 1990, pp. 118-123, Morgan M,G., H,K. Florig, I. Nair, and G.L. Hcstcr. "Controlling Exposure to Transmission Linc Elcctromagnctic Fields: A Rcgnlatory Approach That is Compatible With the Availablc Science," Public lrltilitics March, 1988, pp. 49-58, Morgan, G. and I, Nair. 'Electromagnetic Ficlds: Thc Jury's Still Out,' IEEE Sz3ccrmm, August, 1990, pp. 23-35. Murray, T. "Regulating Asbcstos: Ethics, Politics, and the Values of Science." In Ronald 13aycr, Ed. The Hcalth and Safety o__f Workera New York: Oxford Univcnity Press, 1988. 7-20 Health Effects of F. zposure to Powerline Frequency Electric and Magnetic Field~ O'Brkm, D. ~ pToc~s Is Due: Courts and Science-Policy pisputcs. Ncw York, NY: Russell Sage Foundation, 1987. Office of Technology Asseatment IOTA], U.S..Congress, ~ Effocts of Power Freaucncv Electric & Maimale Fields, OTA-BP-E-S3, Washington, DC: US Government Printing Office, May, 1989. Page, T. *A Geaetie View of Toxic Chemicals sad Similar Risks," Ecology Law Q~arterlv 1978 (2), pp. 207-244. Panel on 60-Hz Electric and Magnetic Fields, Report on Human Health Effects From Exposure To 60-Hz Electric and Magnctic Fields From, High Voltage Power Linea, Preacatai to the Oregon Encrgy Facility Siling Council, January, 1990. Philips, A., !kst, S, and Coghill, R. "KilLing Ficlds," Elcctronics World +W*~rCl~ss world, Fcbruary 1990. POol, R. "Is Thcre an EMF-Canccr Connection?' Scicncc Vol. 249, pp 1096-1098, Scptcmbcr 1990. Sabatier, P., sad D. Whiteman. "Lcgislativc Decision Making sad SubsUmtivc Policy Informalion", ]gg, gi~/S~ Studice Ouartcrlv, voi. 10, 1985, pp. 395-421. San Die2o Gas & Electric Co. v. Donald Dalcy ctal., 205 Cal. App. 3d 1334. Sandman, P. "Hazard venus Outrage: Expcrt versus Public PcrcclXion of Food Risks." Prcscntation at Intacollcgiato Nutrition Consofiium, St. Paul, MN, Octobcr 31, 1989. Savitz, D., N. Pcarcc, and C. Poolc. "Mcthodological lssucs in the Epidcmiology of ElccUomagnctic Fields and Canccr," ~Rcvicws, Vol. 11, 1989, pp. 59-78. Schradcr-FrechcUc, K.S. Risk Analysis and Scicnti~c Mcthod. Ncw York, NY: D. Rckicl, 1985. Schrocdcr, C. "Rights Against Risks," Columbia Law Rcvicw, vol. 86, 3: 495-563, 1986. Smith, R. and B. Wynnc, cds. Exhort Evjdc~c~: lntcrnrctiniz Scicncc in thc taw. New York, NY: RoutJcdgc, 1989. Sykea, T., and L. Ping. Possiblc Health Effects of Electric sad Magnctic Ficlds from Powcr Lima: A Summary of Sc~ Studks. Wuhington State Institute for Public Policy, January 1990. Task Forcc on Transmission Linc Health sad Safcty, Elcctric sad Magnetic Field Issuea: Drait Resolution, National Association of Rcgulatory Utility Commissioners, July 24, 1989. US Environmental Protection Agency, Evaluation of the Potcntial Carcinogcnicity of ElecUomagnctlc Fields: Workshop Rcview Dratt, EPAJ6OO/6-90/OOSA, Junc, 1990. US Housc of Rcprcacntativcs, H.R. 4801, Thc Elcctric and Magnctic Ficld Rcscarch and Public Information Act of 1990, Introduccd by Rcp. Fnnk PaLlonc (D-NJ), May 10, 1990. Whipple, C.G. "Dealing W*~th Unccrtainty About Risk in Risk Managcmcnt.' in Ronnld Baycr, Ed. Thc Health sad Safctv of Workers. Ncw York, NY: Oxford Univcrsity Press, 1988. Vfddavsky, A. ScarchinE for Safctv. Ncw Brunswick, NJ: Transaction Books, 1989. W'nick, D. W. "PubLic Porccptions sad Scientific Asscsaments of Risk and Their Implications for thc Resolution of thc EMP Hcalth Effccts !ssuc," Thc National Rcgulatory Rcseareh Instituto Quarterly Bullctin, VoL 11, No. 1, March, 1990, pp. 45-55. Policy lssuts and Options 7-21 V~rxrick, D. W. ",Some Armlyric Tools and Policy OIxions for Dealing w~th thc Potenthi Adverse Health Effects of Eh:ctromagnctlc Fields," ~l~tiona| Reszuhtorv Research Institute Ouartcrlv Bullc~n Vol. 11, No. 2, April, 1990, pp. 141-151. Donald Zannavi~na v. State of New York and Pov~r Authority New York State Cou~ of Claims, Claim No. 74085, Fiksd September 29, 1959. G-1 GLOSSARY OF TERMS straight or branched chains rather than rings. ABNORMALITY - The quality or state of being Propane and butane are examples. abnormal, i.e, deviating from the normal or average; markedly irregular; characterized by ALL (ACLrFE LYMPHOID LEUKEMIA) - see deficiency or disorder. LYMPHATIC LEUKEMIA. ACCLIMATIZATION - adaptation of an organism to ALLOGENEIC - sufficiently unlike genetically to a new environmental condition, e.g., temperature, interact antigemcally; e.g., when a foreign altitude, climate, or situation. protein or carbohydrate substance (as a toxin or enzyme) introduced into the body stimulates the ACCURACY - the degree to which a measurement, production of an antibody. or an estimate based on measurements, represents the true value of the attribute that is measured. ALTERNATING CURRENT (AC) - an electsic current that reverses its direction at regularly ACQUISITION - to gain, acquire/possess something recurring intervals, (e.g., 50- 60 Hz). The through an active process of search, capture, abbreviation AC is commonly used to describe exchange, etc. periodically varying electrical quantities. ACUTE - sharp, severe; having a rapid onset, severe AC ELECTRIC FIELD - the electric field produced symptoms and a relatively short course. In by AC power systems defined by its' space toxicology, refers to single large exposure to a components along three orthogonal axes. The substance (acute exposure), or to the magnitudes of the components are expressed by development of symptoms of poisoning soon their root mean square (rms) values in V/m or aRer a single exposure (dose) to a substance kV/m. The phases in time of the components (acute toxjcity). need not be the same. ACUTE LEUKEMIA - leukemia characterized by AC MAGNETIC FIELD - the magnetic flux density sudden onset and rapid progression of the produced by AC power systems defined by its' disease. (see LEUKEMIA) space components along three orthogonal axes. The magnitude of the components are expressed ADRENAL - referring to the adrenal gland and its by their rms values in Gauss (G) or miiliGauss functions; complex endocrine organ(s) near the (raG). anterior border of the kidney consisting of a mesodermal cortex that produces steroids like ALTERNATIVE HYPOTHESIS - a numerical sex hormones, and hormones concerned statement concerning the parameters of one or especially with metabolic functions and an more distributions that is mutually exclusive to ectodermalmednllathatproducesandrenalin. the null hypothesis. Sometimes called the research hypothesis because it is, in most cases, ADREIqO-CORTICOTROPHIC HORMONE the hypothesis that the investigator would like to (ACTH) - a protein hormone of the anterior lobe of the pituitary gland that stimulates the adrcnal prove. gland cortex. AML (ACUTE MYELOGENOUS LEUKEMIA) - AGE-ADJUSTED * a statistical method used in rate See MYELOGENOUS LEUKEMIA. calculations to minimize the effects of different age distribution among study subjects or AMBIENT - encompassing or surrounding area. populations. Only properly adjusted rates can be compared with each other. AMPERE - the unit of electrical current which results from I coulomb of charge passing through ALIPHATIC HYDROCARBONS - a series of a conductor in one second. chemical compounds made exclusively of carbon and hydrogen in which the carbon atoms form G-2 Health Effects of Exposure to powerline Frequency Electric and Magnetic Fields usually produces an increued risk m the exposed AMPLITUDE - the maximum departure of the value population. of an alternating current or wave from the average value. 4. Consistency of the association: The repeated f'mding of the association in several studies of 7, 12 - DIMETHYLBElslZENE (a) ANTHRACENE - different populations lends to support causation. a substance known to be a potent cancer promoter, oRen used in cancer promotion 5. Time sequence: Clear antecedence of the research. exposure of interest to the outcome by a period of time is neceuary to judge causality to be ANALYSIS OF VARIANCE (ANOVA) - widely reasonable. used statistical methods that isolate and useu the conlribution of categorical malependent ASSAY - examination and determination u to variables to variation in the mean of a coutmuous characteristics (as weight, measure, or quality). dependent variable. The observations are Analysis to determine the presence/absence or classified according to their categories for each quantity of one or more components. of the independent variables, and the differences between the categories in their mean values on A~fROCYTOMA - a tumor intermingled with the the dependent variable are estimated and tested euential elements of nervous system tissue for statistical significance. These methods can be especially the brain, spinal cord, and ganglia, but used to determine equality of treatment means composed of the star-shsped cells (astrocytes) adjusted for cotactors, or they can be used to ;hat are part of the supporting tiuue of the estimate components of variance attributable to nervous system. Astrocytomas in children and each random source ofveriation. persons less than 20 years of age usually arise in a cerebeilar hemisphere, and in adults they AROMATIC HYDROCARBONS - a series of usually occur in the cerebrum, sometimes chemical compounds made exclusively of carbon growing rapidly and invading extensively. and hydrogen in which the carbon atoms form closed rings rather than straight or branched ATHERMAL - a state, condition, or reaction of a Chain.~. Be~?.e~e and naphthalene are examples. substance or organism which is independent of Many have a fragrant odor. and unaffected by temlMn'ature. ASSOCIATION - association refers to the statistical ATIRIBUTABLE RISK - the difference in disease dependence between two variables, that is, the rates between exposed and non-exposed groups. degree to which the rate of disease in persons It can serve as a measure of the proportion of a with a specific exposure is either higher or lower dim in a population that can be explained by than the rate of disease among those without that the exposure under study. exposure. ALrFOPSY COHORT - a defined group of persons There are several factors to be considered in who have been autopsied aRer a_e~. Since evaluating whether or not an association autopsies are done on non-randomly selected observed in an epideminlogic study is causal, persons in the population, findings from autopsy these include: cohorts should be generalized cautiously. 1. The strength of the association: A strong AVERSION - tending to avoid or causing avoidance association is more likely to be causal than a of a noxious or puniCinS stimulus, as in weak one. behavioral modification through the use of stimuli. 2. Speci~city of the association: The exposure to the supposed catnative agent in every diseased BACTERKJM - any of a class of microscopic plants patient tends to support causality. having round, rodlike, spiral, or filamentous smgle-celled or noncellular bodies of ten 3. Dose-response relationship: In a causal aggregated into colonies or motile by means of relationsthip, increased exposure to the agent flagella, living in soft, water, organic matter, or Glossary G-3 the bodies of plants and animals, and being BIOMARKER- an indication of variation in cellular autotrophic, saprophytic, or parasitic in nutrition or physiological components or processes, and important to man because of theft chemical structures or functions that are measurable in a effects and as pathogens. biological system or sample. BEHAVIOR - anything that an orgamsm does BLOOD CHEMISTRY - the assay/analysis of the involving action, ~nd response to stimulation; the various chemical components of blood, including response of an individual, group or species to its counts/ratios of various cell types, ionic environment. concentrations, and hormonal levels, etc. BENIGN - having a gentle disposition; of a mild BONE MARROW - a soft highly vascular modified character; non-cancerous/non-malignant. connective tissue that occupies the cavities and cancellous part of most bones. BIAS - Eviation of research results or inferences from the truth. Any trend in the collection, MEMBRANE BOUND the presence of analysis, interpretation, publication, or review of elements/molecules, (e.g., calcium) in the data which tend to produce results that differ structure of the cell membrane, capable of being systematically from the "true values" of the released in the event of an appropriate triggering population variables being studied (e.g. disease signal. rates). Many varieties of bias have been described. Unlike conventional usage, the term BOVINE - relating to oxen and/or cows, or a closely "bias" does not refer to a partisan point of view. related animal. BIAS, RESPONSE - a condition present in a study CANCER - a disease characterized by malignant, subject that affects the accurate recording of the uncontrolled growth of cells of body tissue; a response. Particularly, in epidemiologic malignant tumor of potentially unlimited growth questionnaire studies where recall of events is that expands locally by invasion, and frequently more intense in diseased individuals systemically by metastasis. than in non-di.seased controls, this may lead to an inaccurate evaluation of a response factor. CANCER CLUSTER - a series of cancer cases that occur close together in time and/or location. The BIAS, SELECTION - systematic error in research term is normally used to describe a grouping of results due to difference in characteristics relatively rare diseases, such as leukemia. between those who choose or volunteer to participate in a study and those who do not or CANCER INITIATOR - a chemical substance or systematic differences in characteristics between physical stimulus that causes or facilitates those who are selected for study and those who are not. (makes easier) the beginning of cancer. CANCER PROMOTER - a chemical substance or BIAS, OBSERVATIONAL - a condition present in physical stimulus that furthers the growth or the person recording the observation or in the development of cancer. measurement instrun~ent that prevents an accurate recording of the data point. CAPACITOR - a device made of lwo conducting surfaces separated by an insulator capable of BIOLOGICAL PLAUSIBILITY - a guideline used to storing electric charge. judge whether an observed association between an exposure and a disease is a causal one. An CARCINOGEN - a chemical, biological or physical association meets the slandard of biological agent capable of producing tumor growth. plausibility if it fits with existing biological or medical knowledge. CARCINOGENESIS - a series of stages at the cellular level culminating in the development of BIOPHYSICAL MECHANISMS - physical and/or cancer. chemical interactions of electric and magnetic fields with biologic systems. G-4 Health Effects of Exposure to Powerline Frequency Electric and Magnetic Fields CARDIOVASCULAR-relating to, or revolving the CHI-SQUARE (X2) TEST - a common statistical heart and blood vessels. method for comparing proportions or percentages of a characteristic m one or more groups. For CASE - in epidetniology, a person identified as example, a chi-square test might be used in a having the particular health endpoint (e.g., case- control study to determine whether the disease)under investigation. proportion of those exposed to a parficttlar enviromental agent diffen significantly between CASE-CONTROL STUDY - a type of epidemio- the case group and the control group. Two major logic investigation that begins with the chi-square tests, Pearson chi-uluare test and identification of beth a group of persons who Mantel-Haenszel test, are ugd frequently in have developed the disease under study, the epidemiologic studies. cases, and a group of persons who have not developed the disease, the controls. An attempt CHROMODACRYORRHEA - a gland secretion is then made to compare the previous exposure indicative of smm. experience of the cases with that of the controls to determine irthe two groups differ ~ignificantly CHROMOSOME - a very long molecule or DNA, in the frequency or level of a particular exposure. complexed with protein containing genetic information. CATEGORICAL DATA - obgrvations that are classes of distinct groups, usually recorded as CI-I~ONIC - a condition or situation marked by long discrete numerical values; e.g., classes of dead or duration or frequent recurrence. alive can be recorded as 0 or I respectively. Likewig, white, Hispanic, black, native CHRONIC LEUKEMIA - a type or leukemia which American, etc. can be assigned values 0,1,2,3, is not acute. CLL is an abbreviation of chronic etc., also known as qualitative data or digrete lymphoid leukemia. data. CIRCADIAN RHYTHIVl - biological precesse~ CAUSATION - a condition in which a situation, which occur in syuchronogical daily cycles of event, or agent produces an effect in an outcome approximately one day (24 hours). variable or study. CIRCUIT - A closed conducting path for the flow of CELL - a mall, usually microscopic mass of current. protoplasm be'anded externally by a semilgtmeable m~nbrane, usually including one CLINICALLY OVERT - a term describing an or more nuclei and various aonliving products, iliness in which the symptoms have become c, spable alone or interacting with other cells of obvious. performing all rig fundamental fum:tions of life, and forming the l~ast structural unit of living COHORT - in q~idgmiology an id~nti~exi group of matter capable of functioning nidqgnclently. porsons with some common point of reference, i.e. birth year, place of employment, who sre fr~ CELL MEMBRANE- the sexni-lgrmesble matreal of dis, but who have vsrious degr~s of forming the beundary of s cell that enclogs and exposure to the agent under study. The group is supports the cell, and controls effiux and influx followed over time to determine the occurrence of cell metabolit~s, nutrients, wastes, etc. of disease among members of the cohort. CENTRAL NERVOUS SYSTEM (CNS) - the part of COHORT STUDY - a type ofepide-~niologic study in fig ncrvous system which in vertebrams consists which fig frequency of morbidity or mortality of fig brsin and spinal cord, to which gnsory from a specific dis of inun'est in a group impulgs are transmitted and from which motor ~gposod to a susp~ted risk factor is compaml to impulses pass out, and which supexvises an that inagroupofune~q~osed!~oPle. coordinates the activity of the entir~ nervous system. COLON - the part of the large intestine that extends from the c~,'um to the rectum. Glossary G-5 CONCEPTION - the capacity, function, or process of Factor becoming pregnant; in sexual reproduction, the act of egg fertilization. + Total CONDUCTANCE - the ease with which a material Disease or + a b a+b conducts current. condition c d c+d CONDUCTOR - a material that allows the flow of charge. The wires on transmission lines ate a+c b+d n conductors. CONFIDENCE INTERVAL - A range of values bracketing a risk estimate which is calculated in a CONTROL - in case-control studies, an individual in such a way that the range has a 95% probability the group of people that have not developed the of including the true value of the risk. disease of interest. (See CASE- CONTROL STUDY) CONFOUNDING - a situation in which an observed association between an exposure and a disease is CONTROL GROUP - in experimentation, the group influenced or distorted by other variable(s)that of subjects that ate treated as in a parallel are associated with the exposure and affect experiment except for omission oft he procedure disease occurrence. or agent under test and which is used as a standard of comparison in judging experimental CONFOUNDING VARIABLE/CONFOUNDER - a effects. variable that can distort the real association betwoen exposure and outcome. An extraneous CORRELATION - the establishment of a mutual or variable that may explain an observed association reciprocal relation between/among phenomena; (or lack of an association) between an exposure to show a causal relationship between two or and a disease in an epidemiologic study. A more events; a relation of phenomena as confounder can create a spurious association invariableaccompaniments of each other. between an exposure and a disease, or it may mask, weaken, or exaggerate a real association. CORRELATION - a linear relationship between two The variable is not an intermediate between the or more sets of variables. A linear association. exposure and the outcome. It is a risk factor of Correlation, like association, does not imply the outcome, but it is not the variable of interest causality. under investigation. Confounding must be ruled out before confidence can be placed in any CORTICOSTERONE- a colorless crystalline steroid observed association. Such a variable must be hormone (C21H3004) of the adrenal cortex that controlled in order to obtain an undistorted is important in protein and carbohydrate estimate of the effect of the study factor on risk. metabolism. (See STRATIFICATION) COULOMB - the unit of electric charge (C). One CONTINGENCY TABLE - a cross classification of electron (or proton) has a charge of about 1.6 x categorical data arranged in a table (of 2 or more 10'19C. dimensions) such that each row and column expresses the frequency of occurrence in the CUMULATIVE EXPOSURE - the total exposure to sample of subjects possessing the characteristics an agent, (such as magnetic fields,) experienced defined by the row and column variables; e.g., a by a person during a specified time period, e.g., 2x2 contingency table of disease and exposure to one hour, one yeas, or a lifetime of work. a certain factor may be given by: CUM'JLATIVE INCIDENCE - the number or proportion of a group of people who experience the onset of disease during a specified time CONTINGENCY TABLE interval. G-6 Health Effects of Exposure to Powerline Frequency Electric and Magnetic Fields CURRENT the flow of electrically charged DIELECTRIC - an insulator or non-conductor. particles. The unit of the smper~ (A) expresses the force due to the magnetic field. DIELECTRIC STRENGTH - the maximum electric field strength that a material can withstand CYCLOTRON RESONANCE - for an ion, the without brpaldng down and conducting. condition for which the product of the charge-to- m ratio and the static magnetic field yield a DISTRIBUTION LINES - power lines that distribute value of periodic motion at the same frequency electricity from substations to individual as externally applied energy. industrial, commercial, public and private users/customers. CYTOLYSIS - the usually pathologic dissolution or disintegration of cells. DOSE - the mount of a physical or chemical agent interacting with/or absorbed by a person or CYTOTOXICITY - the ability of s chemical organism. substance and/or a physical stimulus to !cili (poison) cells. DOSE-RESPONSE STUDY - an investiplion that attempts to statistically define the functional DELETERIOUS - having a harmful effect. relationship between a response (usually a disease incidence) and the dose of a specific DEMOGRAPHIC INFORMATION the agent. characteristics of a population such as place of residence, age, sex and race, binIx and death DOUBLE-BLIND EXPERIMENT- an experimental rates, and sociceconomic conditions. procedure in which neither the subjects nor the experixnenter know the makeup of th~ test and DEOXYRIBONUCLEIC ACID (DNA) - the nucleic control groups during the actual cours~ of the acid molecule in chromosomes that contains the experiments. genetic information EFFECT MEASURE - a quantity that measures the DESYNCHRONIZATION - to upset, alestabilize or effect of a factor on the frequency or risk of a destroy the r~urrence, periodicity, or health outcome. coexistence in time of various functions/activities of a biological system. EFFLUX - the action or process of flowing out, as in the flow of ions through a semi-permeable cell DETECTION - to discover or determine the membrane. existence, or the presence of som~hinE. ELECTRIC DIPOLE - two separated electric DEVELOPMENTAL EFFECTS - effects in the charges; a molecule (or other stnlcture) having developing offspring due to ev. posure before the effective centers of positive and negative conception (either parent ), promtally, or charges separated. posmatally to the time of sexual maturation. Developmental eff~zts may be expressed at any ELECTRIC FIELD - a v~ctor field describing the time in the life span of the orgamsm. el¢c-trical force per unit charge in space. Developmental effects are a ssbset of Electrical charges are a source of electric fields. reproductive effects. The electric field from a power line is an alternating, 60-Hz field du~ to charges on the DIRECT CURRENT (DC) - an electric current conductors. The intensity of the electric field is flowing in one dir~tion only and substantially expressed in volts per meter (V/m) or kilovolts constant in value. per meter (kV/m). DIAGNOSTIC CRITERIA - information, nsually ELECTRIC AND MAGNETIC FIELDS (EMF) - clinical information such as physical symptoms components of the electromagnetic spectrum and laboratory test results, used to determine which describe how energy travels through space. whether a person has a suspected disease. Glossary G-7 For 60-hertz and other low frequencies, these of the bones. The red blood cell's red color is vectors may be considered separately. due to the presence of the pigment hemoglobin, which combines with oxygen from the lungs to ELF (EXTREMELY LOW FREQUENCY) - usually form oxyhemoglobin which carries oxygen to the taken to denote the frequency range below 300 cells ofthe body. Hz. ESTIMATION - the process by which unknown EMBRYO - an animal in the early stages of growth parameters of a function or distribution are given and development (differentiation) that are approximate values (called estimates) based on characterized by cleavage, the laying down of available information. fundamental tissues, and the formation of primitive organs and organ systems (a vertebrate EXPOSED GROUP - the experimental group of test at any state of development prior to birth or organisms receiving a dose of a substance, to hatching).For humans, this stage lasts between determine the effect(s) of the substance. the second through eighth weeks after conception. EXPOSURE - the joint occurrence in space and time of a person and the physical or chemical agent of EMBRYONIC - an animal in the early stages of concern, expressed in terms of the environmental growth and differentiation prior to birth or level of the agent. hatching. EXPOSURE ASSESSMENT - measurement or ENDOCRINE SYSTEM - the glandular system estimation of the magnitude, frequency, duration which produces secretions that are distributed in and route of exposure of an orgamsm to the body through the blood stream, and aid the environmental agents. The exposure assessment nervous system in controlling and coordinating may also describe the nature of exposure and the the body functions. size and nature of the exposed populations. ENDODERMAL SINUS TUMORS CEST) - a EXPOSUR~ METRIC the means by which frequently fatal tumor of germ cell origin, exposure to an agent of interest (e.g., magnetic generally found in the ovary. fields) is measured or estimated. For example, in an occupational mortality study, the job title ENDPOINT an observable or measurable "electrician" listed on the death certificate might biological, chemical or functional event used as serve as the "exposure metric" to estimate the an index of the effect of a chemical, physical or likelihood of exposure. biological agent on a cell, tissue, organ, orgamsms, etc. EXPOSURE-RESPONSE RELATION - a relationship between exposure and the effect ENZYME - a protein molecule that acts as a catalyst produced by exposure. Response can be in living organisms. expressed either as the severity of injury or proportion of exposed subjects affected. EPIDEMIOLOGY - the study of the occurrences and, distribution,of a disease or physiological EXTRAPOLATION - an estimate of response or condition. in hunan populations and the factors quantity at a point outside the range of the that influence this distribution. experimental data. EPINEPHRINE - a colorless crystalline slightly FARADAY'S LAW OF INDUCTION - the induce basic aarenal hormone (C6Hi3NO3) which acts voltage in a circuit is equal to the negative rate at in the sympathetic nervous system which is used which the magnetic flux through the circuit is medicinally as a heart stimulant, a changing. vasoconstrictor, and a muscle relaxant (also known as adrenaline). FATTY ACID - any of numerous saturated aliphatic monocarboxylic acids (CnH2n+ 1 COOH) ERYTHROCYTE - a red blood cell - tiny, disk- including many that occur naturally, usually in shaped cells, without nuclei, made in the marrow G-g Health Effects of Exposure to Powerline Frequency Electric and Magnetic Fields the form of estas in fats, waxes, and essential FREQUENCY - the number of complete cycles of a otis. periodic waveform per ~mit time. The units of fgqucncy are Hertz (Hz), which is one cycle per FERTILITY - ability of a livin[ plant or tonireal to second. grow or develop; capability to breed or reproduce the species. ~0-60 HZ FREQUENCIES - the number of complete cycles of a periodic waveform per unit time. HBROBLAST - a mesenchyme (early connective Units of frequency are the hertz (Hz) which is tissue) cell which is one of the building blocks equivalent to cycles per second. The frequencies for connective tissue (tissue that binds and of AC power systems in Europe and North supports the body parts). Fibroblasts are cells in America are 50-Hz and 60-Hz, respectively. the "extracellular matrix", or the network of molecules connecting the space outside the cells. GAUSS - the historical unit of magnetic flux Fibroblasts are responsible for producing density, often used to express the magnitude of a material such as fibrin and other colla~ens, which maguetic field. One Sam is 10-4 tesla. form a large fraction of the matrix. GENE - the simplest complete functional unit in a FIELD - a set of values of a physical quantity DNAmolecuie. A linear sequence of nucleotides occurring at different points in space. in DNA that is needed to synthesize a protein and/or regulate cell function. HELD INTENSITY (FIELD STRENGTH) CE & H) - a vector quantity which describes the forces of GEOMAGNETIC HELD - the earth's natural interaction of macroscopic electric currents. 1.) magnetic field. Electric Field intensity is measured in units of volts pa meta (V/m) 2.) !viagnetic Field GEOMETRIC MEAN - one of several measures of intensity is measured in amperes per meter central tendency. Used when a data set (A/M), but is often indicated by a related encompasses a very wide range of values. It is quantity called magnetic flux density which is the calculated by converting each observed value to number of field lines that cross a unit of surface its logarithm, determining the arithmetic mean of area. Umt of magnetic flux density most the logarithms, and then converting back to a commonly used is the Sauss (G). direct count by taking the antilogarithm. (See HSHER EXACT TEST - a statistical test for association based on the distribution of frequen- GESTATION - the period of intrauterine fetal cies within a 2x2 table. development from conception to birth, usually 38 weeks. FOLLICLE - a mall anatomical cavity or deep narrow-mouthed depression. As related to hair: GLIOMA - a type of cancer of the central nervous the tabular epithelial sheath that surrounds the system, or a specific tumor composed of the lower part of the hair shaft sad encloses a the supporting cells of the brain, spinal cord, or other bottom of vascular papilla supplying the Wowing nervous system tissue. basal part of the hair with nourishment. GLYCOPROTEIN - a conjugated protein in which FOLLOW-UP - a process in epidemiolo83' by which the non-protein group is a carbohydrate. study subjects are tracked sad observations of variables of interest are made over time. Follow- GROUNDING conhooting a conductor to up has two critical features: completeness sad something that will accept excess charge, for duration. Completeness refers to the proportion example, the earth. of the study sample followed. Duration refers to the length oftline the sample is followed. HALL-EFFECT - the generation of an electrical potential papendicular to both the electric FORAGE - to wanda in search of food; food for current flowing along a thin conductive lateral animals, usually secured by browsing or gra?ing. and sa external magnetic field applied at right Glossary G-9 angles to the current upon application of the health conditions by an international group of field. experts for the World Health Orgamzation (WHO). Every health condition is assigned a HARMONICS component frequencies of an specific numerical code. The complete list is electromagnetic wave that are integral multiples periodically revised , in the Manual of the of the fundamental frequency. International Statistical Classification of Diseases, Injuries and Causes of Death. The HEMATOPOIETIC CANCERS - cancers of the Ninth Revision of the Manual (ICD-9) was blood-making organs, especially the bone published by WHO in 1977 after ratification of marrow and lymph nodes. 1976. HERTZ -one cycle per second. IMMUNE SYSTEM - the body's primary defense against abnormal growth of cells (i.e., tumors) HORMONE - the secretions of endocrine glands and infectious agents such as bacteria, viruses which act as "chemical messengers", controlling and parasites. and regulating the body's life functions. IMMUNOGLOBULIN - a protein (such as an HOST ORGANISM - a living animal or plant antibody), which is made up ofamino acid chains affording subsistence or lodgemeat to another usually linked by disulfide bonds. orgamsm (usually a parasite); an organism into which a tissue or part is transplanted from INCIDENCE OF A DISEASE -the number of new another. cases (persons becoming ill) during a g~,ven time period in a specified population. HYDROCARBON - a compound made exclusively of carbon and hydrogen atoms. Hydrocarbons INCIDENCE RATE - the rate at which new events are commonly found in petroleum, natural gas occur in a population. The numerator is the and coal. number of new events that occur in a defined period; the denominator is the population at risk HYPERTENSION - a condition in which the patient of experiencing the event during this period, has a higher blood pressure than judged to be sometimes expressed as person-time,e.g., normal. (100,000 persons in a year) HYPOTHAIAM0-YPOPHYSEALADRENAL INCUBATION - the period between the infection of SYSTEM - neuroendocrine system that plays a an orgamsm by a pathogen and the manifestation part in controlling behavior, metabolism, of the disease it causes; the maintenance of maintenance ofbodytemperature, etc. controlled conditions for the cultivation of organisms or the housing of young or sick HYPOTHALAMUS a basal part of the organisms. diencephalon (posterior subdivision of forebrain) that lies beneath the thalamus on each side, forms INDUSTRIAL HYGIENIST - a professional whose the floor of the third ventricle, and is usually job is to recognize, evaluate, and control considered to include vital autonomic regulatory environmental factors or stresses in the work centers. place which may cause sickness, impaired health, or significant discomfort and inefficiency among HYPOTHESIS - a supposition. amved at from workers or among other exposed persons. observation or reflection, that leads to refutable predictions. INFERENCE - the act of making a decision or evaluation concerning one or more characteristics HYPOTHESIS TESTING the systematic or properties of a population based on verification or rejection of a scientific information obtained from a sample. Reasoning proposition or argutnent. from a part to the whole. Based on the assumption that the sample represents the whole ICD (INTERNATIONAL CLASSIFICATION OF population. DISEASES) the classification of specific G- 10 Health Effects of Exposure to powerline Frequency Electric and Magnetic Fields INITIATION - the first (initial) stage of LATENCY, LATENT PERIOD - the delay between carcinogenesis (on.~et of cancer) caused by exposure to a dis causing agent and the carcinogenic agents, (e.g., ionizing radiation, appearance the di~eaze. For example, after certain chemicals), wherein cellular genetic exposure to iotazing radiation there is an average material, (i.e., DNA) is irreversibly changed or latent period of five years before development of mutated. leukemia, and more than 20 years before development of certain other malignant INITIATOR EFFECTaNITIATION - the u'ansforma- cendition~. tion of a normal cell of the body to a neoplastic (cancer) cell by means of a permanent change or LEC~ - any of ~veral waxy hydro~copic mutation in the nuclear DNA. Initiation result~ phosphatide~ that are widely distributed in from a limited exposure to a carcinogen, is animals and plants, form colloidal solutions in accomplished rapidly, and is irreversible. water, and have emulsifying, wetting, and antioxidant properties. INSULATOR - a non-conductor of electrical charges. LEUKEMIA - an acute or chronic disease (cancer) in man and other warm-blooded animals INTRAUTERINE EXPOSURE - exposure characterized by an abnormal increue m the experienced by the fetu~ in the uterus during number of white blood cells. pregnancy, e.g through the placenta. LYMPHATIC CANCERS - acute and chronic IN VITRO - describes studies that are done in the c,~cer~ of all sU'uctutes involved in the laboratory, literally Nin glad", as distinct from conveyimce of lymph from the ti~sue~ to the thoae performed using living ~nirnn|s. blood streazn. The lymph system includes the lymph capillaries, lacteals, lymph nodes, lymph IN VIVO - experiments performed "in the living ve~els, and mare lymph ducts. body" of a plant or animal. LYMPHATIC LEUKEMIA/LYMPHOID ION EFFLUX - the movement of ions, charged LEUKEMIA leukemia in which there is atoms or molecules, from a urnpie into a marked increase in the size of the spleen and surrounding ~olution. lymph giand~ with great increaze in white blood cells in the blood; acute forms occur in children IONIZATION - the dissociation of compounds into and young adults. ion~ through lo~s of electron(s) LYMPHOCYTE - a colorle~ weakly motile cell IONIZING RADIATION - any electromagnetic or produced in lymphold tissue that is the typical pa~iculate radiation capable of producing ion~ cellular element of lymph, and constitutes 20 to directly or indirectly, in its pa~age through 30 percent of the leukocy~es of normal human mailer. Ionivin~ radiation po~sesse~ sufficient blood. energy to remove elec~ron~ from the atoms or molecules it encounte:~, and i~ capable of LYMPHOMA - a general term for an abnormal causing injury to living cell~. Examples of (neoplasm) growth in the lymphatic system. iOniTin~ radiation: Xrays, gamma rays and alpha Included in this general group are Hodgicin's and beta particles . disease, lymphosar~oma, and malignant lymphoma~ IRRADIATED - exposed to the emi~ion of radiant energy (energy that travel~ through space, even in MAGNETIC DIPOLE - two ~zrated magnetic the absence of matter). poles; an object such as a permanent magnet, particle or ~urrent loop, that gives ri~e to a ISCtiEMIC HEART DISEASE - localized and magnetic field. The object acts as if it consists of temporary lack of blood in the heart due to an two magnetic poles of opposite sign separated by obstruction of the flow of arterial blood. a 5mall distance. Glossary G- 11 set, divided by the number of values in the set. MAGNETIC HELD STRENGTH - a vector field (See also GEOMETRIC MEAN and MEDIAN) describing the force per unit charge experienced by magnetic objects or moving electrical charges MECHANISTIC - mechanically determined - natural in space. The unit is the ampere per meter processes (as of life) can be mechamcally (A/m). determined and capable of complete explanation MAGNETIC FLUX DENSITY - a vector field, by the laws of physics and chemistry; the fundamental physical or chemical processes which is related to the magnetic field by the involved in or responsible for an action, reaction, magnetic permeability of the medium. The SI or other natural phenomenon (as organic unit is the tesla CO. The historical unit is the gauss (G), which equals 10'4T. evolution). MALFORMATION - a permanent structural change MEDIATE - to occupy a middle position; to effect by action as an intermediary; to transmit as in a developing organism that may adversely intermediate mechanism or agent. To intervene affect survival, development or function. between conflicting processes/substances (in a chemical or biological process) to promote MANTEL-HAENSZEL TEST - a sumrn~lry chi- square test developed by Mantel and Haenszel for harmony or compromise. stratified data and used when controlling for MEDIAN - one of several measures of central confounding. (see CHI_SQUARE TEST) tendency of the distribution of a set of values. The median represents the middle figure when MALIGNANT - tending to produce death or the measurements are arranged in ascending deterioration through the process of infiltration, order. Half the measurements are below the metastasis (spreading throughout the body) and destruction of tissue. median value, half are above. MELATONIN - a vertebrate hormone of the pineal MALIGNANCY - a neoplasm or tumor that is invasive with a tendency to metastasize. gland that produces darkening of the skin by causing concentration of reelshin in pigment- MAMMARY (GLAND) - one of the large compound containin8 cells. Melatonm influences sleep, modified sebaceous glands that in female perception of pain, psychological depression and social behavior. mammals are modified to secrete milk, usually situated in pairs (ventrally), and usually terminating MENARCHE - the onset of menstruation, normally occurring between the 10th and 17th year. MATCHED CONTROLS - in a case-control study, METABOLISM - the biochemical reactions by controls selected so that they are similar to the which energy is made available for the use of an cases in specific characteristics such as age, sex, race, and socioeconomic status. (see CASE- organism from the time a nutrient substance enters, until it has been utilized and the waste CONTROL STUDY) products eliminated. MATCHING - the process of making a study group METASTASIS - movement of bacteria or body cells and a comparison group comparable with respect to extraneous or potentially confounding factors. (especially cancer cells) from one part of the body to another by means of the lymphatics or blood stream. In cancer cases the result of MATCHING VARIABLE - a characteristic such as metastasis is a secondary growth arising from the age or sex used to select matched controls, such that those characteristics are similar in cases and primary growth in a new location. controls. (See MATCHED CONTROLS) MICROGAUSS 0(uG) - one millionth of a gauss, or 10.6 G. MEAN - one of several measures of central tendency of the distribution of a set of values. MICROWAVE - a comparatively short electromag- The arithmetic mean is the sum of all values in a netic wave, between 100 centimeters and one G-12 Health Effects of Exposure to Powerline Frequency Electric and Magnetic Fields centimeter in wavelength and corresponding MULTIPLE MYELOMA - a neoplastic dim frequency from 1Ghz-300 G hz. A form of non- characterized by the infiltration of bone and bone ionizin~ radiation used in commumcations & marrow by myeloma cells forming multiple other technologies that can also cause significant tumor masses. It occurs commonly in the sixth tissue heating. dec_~_~ of life. and more frequently in males. MILLIGAUSS (raG) - one thou~ancith of a gauss. or 10'3 G. !vI'ULTIVARIATE A_NALYSIS - a set of statistical techniques used when the variation in several variables is studied simultaneously. MISCLASSIFICATION ERROR - the erroneous classification of an individual into a category MUTAGEN - any agent that causes genetic changes. other than that to which he or she could he assigned. In an epidemiologic study of EMF Many medicines, chemicals and physical agents, such as ionizing radiation and ulWaviolet light, exposure, for example, including electricians can he in,tsgens under the right conditions. who routinely work on dead circuits in the "exposed" group would result in misclassification MYELOGENOUS LEUKEMIA/MYELOID error. LEUKEMIA - leukemia involving the blood- M1TOGEN - a substance that induces mitosis making bone marrow, especially that of the ribs, sternum and vertebrae. (see LEUKEMIA) (replication of a cell's nuclear material). MYOCARDIAL INFARCT - necrosis or death of MITOGENESIS - the process by which a substance tissue in the heart following the c induces mitosis (replica~on of a cell*s nuclear material). NEGATIVE STUDY - a finding or study which confwms the null hypothesis, i.e., the association M1TOSIS - cellular and nuclear division that between exposure and disease is not different involves duplication of the chromosomes of a from a measure of no association. Example: a parent cell and formation of two daughter cells. study with a relative risk of one or a risk di fiefence o f zero. MODEL - (1) Mathematical modal. A mathemati- cal representation of a natural system intended to NEOPLASIA - the pathologic process that results in mimic the behavior of the real system, allowing the formation and growth of a neoplasm. The description of empirical data, and predictions development of new tissues or neoplums. (See about untested states of the system. NEOPLASM) (2) Biological model. A condition or dis~sse in NEOPLASM - a new and abnormal formation of animals simile/to the condition or dim in human being. tissue as a tumor or growth that serves no useful function, but continues to grow after the stimuli that initiated the nt'w growth cease, and at the MODULATOR - something that regulates a proceu expen~ of the healthy organism. or concentration according to measure or proportion; something that varies the amplitude, NEOPLASTIC - pertaining to, or of the nature of, frtxluency, or phase of a carrier or signal. new, abnormal tissue growth. MORBIDITY - any departure from a state of NET CURRENT - the vector sum of the currents in physiological or psychological well-being,used in public health data to describe disease states. all the wires (primaries, seconda:ies, and neutrals) of a set of conductors. MORTALITY - death; or the number of deaths in a NEUROBLASTOMA - a malignant, bleeding tumor given time or place; or th~ death rate. comprised principally of cells resembling MOTOR BEHAVIOR - movement of an organism in embryonic nente tissue. Neuroblastoma occurs response toastimulns. frequently in infants and children in the mediastinal and retroperitoneal regions and Glossary G- 13 metastasizes widely. It occurs chiefly in infants under study to the number that would be and children. expected on the basis of the disease experience of a reference group. NEUROENDOCRINE HORMONE - of or pertaining to hormones that influence the activity of nerves ODDS RATIO - a calculation used frequently in (neurosecretion). case-control studies to compare the exposure experience of diseased and non-diseased groups. NELrROLOGICAL of or pertaining to the The odds ratio can serve as an estimate of the ftmctioningofthenervoussystem. relative risk of disease associated with the exposure. NEUROTRANSM1TI'ER - a chemical substance that transmits nerve impulses across the space OHM -the unit ofelectricalresistance. between nerve endpoint called the synapse. ONCOGENE - a mutation of a naturally occurring NEUTRAL - the wire at ground potential carrying gene involved in growth regulation that results in the return current of energized wires. uncontrolled growth. Oncogenes are associated with the development of some forms of cancer. NON-HODGKIN'S LYMPHOMA see LYMPHOMA. ONE-SIDED TEST - a statistical significance test based on the assumption that the data have only NONIONIZING RADIATION - radiation which does one possible direction of variahility. not transfer sufficient energy to remove electrons, or break chemical bends, to form ions ORNITHINE DECARBOXYLASE (ODC) an in the material it encounters. Examples of enzyme found in cells, essential for cell growth nonionizing electromagnetic radiation include because it helps synthesize biochemicals that are ultraviolet and visible light, infrared and necessary for DNA and protein synthesis. microwave radiation, radio and television waves and power-frequency fields (60 I-h) also included OVARIAN - Concerning the two glands in the may be laser sources (coherent light) and female which produce the reproductive ceil, the ultrasound (sonic radiation). ovum, and two known hormones. NONPARAMETRIC METHODS (DISTRIBLFFION- PAH - see POLYCYCLIC AROMATIC FREE METHODS) - statistical procedures that HYDROCARBON. are derived in such a way that they do not rely on the form of the underlying distribution. PARAMETER - in mathematics, a constant in a formula or model; in statistics and epidemiology, NOREPINEPHRINE - a crystalline compound a measurable characteristic of a population. CgH1 iNO3 that occurs with eppinephrine, has a "True parameter values" are usually unknown strong vasoconstrictor action, and mediates and must be estimated from the data. transmission of sympathetic nerve impulses. PARITY - the status of a woman with respect to the NULL HYPOTHESIS- the statistical hypothesis that number of full- term children she has berne, one variable has no association with another excluding miscamages or abortions in early variable or set of variables, or that two or more pregnancy, but including stillbirths. population distributions do not differ from one another. In simplest terms, the null hypothesis PATHOLOGIST - a specialist in diagnosing disease- states that the results observed in a study, related changes in tissues removed at operations experiment, or test are no different from what and postmortem examinations. that might have occurred as a result of the operation of chance alone. PEAK INCIDENCE - the highest recorded incidence of a particular disease in a particular population. OBSERVED-TO-EXPECTED (O/E) RATIO - in an epidemiologic study, the ratio of the observed number of cases of a disease in the population G-14 Health Effects of Exposure to Powerline Frequency ElecUic and Magnetic Fields tissues, or cells), and of the physical and PEER REVIEW - a review, usually of a study chemical phenomena revolved; the organic report, by persons who have similar expertise as processes and phenomena of an organism or any the author(s) in the subject area of the study. of its parts or of a particular bodily process. PERIODICITY - the quality, state, or fact of being PHYTOHEMAGGLUTANIN - a protein extract of regularly recurrent, (i.e., returning or happening the red kidney bean that ]ass been used to timeafter time). agghitmate red blood cells, and to reduce structural changes, followed by mitosis, m white PERIPHERAL - located away from a center or blood cells in culture. central portion; involving the surface or external boundary of a body. PIR (PROPORTIONAL ]NCIDENCE RATIO) - the proportion of new cazes of a specific disease PEROXIDATION - to lieat something with a among new cases of all diseases in a given peroxide (a substance having a high proportion of population, compared to the proportion of new oxygen making it a strong oxidivjng/bleaching cases of that disease among new cases of all agent). diseases in a reference population. The PIR may be used, for example, to determine if, compared PERSON-YEAR - a unit of measurement obtained to the occurrence of other diseases, leukemia by summing the lengths of time (usually one occurs more frequently among electricians than year) for each person and used as a denominator among the general public. Calculated as follows: in incidence and mortality rate calculations. Percent new cases of disease due to leukemia in study population divided by percent of new cases PERSON-YEARS AT RISK - the sum of the yean of dim due to leukemia in reference that the persons in the study population have population. been exposed to the condition of interest. With this approach, each person contributes only as PLASMA - the fluid part ofblood, lymph, or milk as many years of observation to the study as he is distinguished from suspended material. actually observed at risk; if he leaves, contracts the disease under study or dies after one year, he PLASMA MEMBRANE - the membrane contributes one person,year; if after tell, te~ surrounding plant and animal cells. person-years. PMR (PROPORTIONAL MORTALFrY RATIO) - PHASE SHIFTS - modulation of the phase of a the proportion of deaths due to a particular cause periodic waveform in time or space from a in one population, compared to the proportion of chosen instant or position, or a shift in the time deaths due to that cause in another population. of occurrence of peak and levels of hormone The PMR is used, for example, to determine if concentrations in animals during a 24-hour the proportion of deaths due to leukemia is period. greater among electricians than among the general public. The comparison of the PMRs PHORBOL ESTERS - substances, (e.g., TPA, or 12- from different populations can give rise to 0-teliadecan{oy/pho]ol-13-acetate) which are misleading conclusions if the populations have potent promoters of cancer and offell I~ed as different distributions of causes of death. experimental standards in cancer promotion Calculated as follows: percent of deaths due to experiments. specific disease in study population divided by percent of deaths due to specific disease in PHOSPHOLIPID - a complex phosphoric ester lipid reference population. that is found in all living cells in association with stored fats. POCKELS EFFECT - is the change in refractive properlies of certain crystals in the presence of an PHOTON-a particle ofelectromagnetic energy. applied electric field and is proportional to the first power of the electric field strength. PHYSIOLOGICAL - dealing with the functions and activities of life or of living matter {as organs, Glossary G- 15 POLYCYCLIC AROMATIC HYDR0-CARBON - a PREVALENCE RATE - the total number of existing series of organic compounds that have two or cases of a disease or other condition in a specific more usually fused rings in the molecule. Some population at a specific point in time divided by polycyclic aromatic hydrocarbons are known the size of the population at that time. human carcinogens. PRIMARY - (of a distribution line) The set of wires POPULATION - the entire set of subjects that are at voltages higher ~an the residential service the object of a study or investigation. voltage (120 V) that are connected to distribution transformers. POSITIVE FTNDINGS/POSFHVE STUDY - a finding or study in which the association between PROBABILITY DISTRIBUTION - a mathematical exposure and disease is appreciably different function that assigns a probability to the from a measure of no association. occurrence of a specific value or range of specific values that can be assumed by the POST-NATAL - subsequent to birth; relating to all random variable. organism immediately after birth. PROGRESSION - a continuous and connected series POTENTIAL - electrical potential energy, def'med at (sequence) of events, actions, etc. In the three- a point by the work necessary to bring a unit stage model of carcinogenesis, the state where a positive charge to the point from an infinite benign tumor becomes malignant. distance. The difference in potential between two points is defined by the work necessary to PROLACTIN - a protein hormone of the anterior carry a unit positive charSe from one to the lobe of the pituitary that induces lactation (milk other. The unit is the volt (V). production). POWER - the time rate at which work is done. PROLIFERATION - production of new cells Electrical power is proportional to the product of through the process of cell division. current, voltage, and power factor. The unit is the watt(W). PROMOTER EFFECT - the facilitation of the growth of dormant cancer cells into tuxnors. POWER FACTOR - is the cosine of the angle of phase difference between the voltage and current. PROMOTION - The second hypothesized stage in a multistage process of cancer development. The POWER (STATISTICAL) the probability of conversion of initiated cells into tumorigemc rejecting the null hypotheses when it is false. cells. Power is equal to I minus the probability of a type H error. In epidemiolo~y, the prohability PROSPECTIVE STUDY - a type of epidemiologic that a real association between an exposure and a study in which study subjects free of disease are dim, if it exists, will be detected in a followed into the future to determine their statistical hypothesis test. morbidity or mortality experience. PRECISION - the closeness of an observation to the PROTEIN - any of numerous naturally-occurring mean derived from repeated sampling of the extremely complex combinationa of amino acids same population. Precision can be estimated by that contain the elements carbon, hydrogen, standard error or standard deviation. nitrogen, oxygen, usually sulfur, and occasionally other elements. Proteins are essential PREDISPOSING FACTOR - a condition or constituents of all living cells, and are characteristic that contributes to an individual's synthesized from raw materials by plants but susceptibility to a dim. assimilated as separate amino acids by animals. PREVALENCE - the total number of existing cases PULSED FIELD - an electric/magnetic field of a disease or other condition in a specific produced or modulated (as electromagnetic population at a specific point in time. waves) by brief duration-pulses of electric G- 16 Health Effects of Exposure to powerline Frequency Electric and Magnetic Fiel~ current or voltage. Pulsed fields are fields which findin8 the "best" mathematical model to are turned on quickly for only a brief period. describe y as a function of the x1, x2, x3 ..... xn. Types of regression include: simple linear PRR (PROPORTIONAL REGISTRATION RATIO) - regression that is a straight- line fit to data points the proportion of a specific kind of cancer among observed in a two-dimensional grid; multiple all new cases of cancer registered for one group regression that is basically Linear regression in n- of people, compared to the proportion of that space; polynomial regression that is a kind of cancer in another group. Similar to the polynomial-curve fit to data points observed in a PIR, the PRR is used to determine, for example, two-dimensional grid; nonlinear regression that whether the propo~ion of leukemia compared to fits functions other than polynomials to data; and other cancers is higher among electricians than it logistic regression that is a categorical data is among the general public. (See also method fitting functions to data adjusted by the REGISTRY) logistic uansformation. PUTATIVE - commonly accepted or $upposed. RELATIVE RISK - the ratio of the risk of dim or _de_~_~ among an exposed group of study subjects />-VALUE to the risk among an unexposed group. A risk measure based on disease or death rates that is See STATISTICAL SIGNIFICANCE. used frequen~y in cohort studies. The relative risk indicates the increased (or decreased) degree QUALFFATIVE DATA - information collected in a of risk among the exposed subjects compared to study which can be classified into categories; the non-exposed. A relative risk value of 1.0 such as sex, race, hair color and nationality. indicates no association between the exposure and the disease. A relative risk of 2 would QUANTITATIVE DATA - information collected in indicate that the exposed group is twice as likely a study which can be classified on some as the non-exposed group to experience the continuous scale, such as age, height, weight and health effect being studied (death or dim). blood pressure. RELIABILITY - the degree to which the results RANDOM VARIABLE - a variable whose values obtained by a measurement procedure can be follow a probability distribution. replicated. Lack of reliability may arise from divergences between observen or inatntments of RANDOMIZATION - a process by which study measurement, or from instability of the am'ibute subjects are assigned to experimental test being measured. conditions such that each subject has an equal chance of being selected for each condition. REPLICATION - the action or process of reproducing an exact replica (copy) of REFERENCE GROUP - a standard aEain-~t wlxich something. the disease experience of the population that is being studied is compared. REPRODUCTIVE EFFECTS effects on reproduction which may include, but not be REGISTRY - in epidemiology the term "registry" is limited to, alternations in sexual behavior onset applied to the f~e of all cases of a particular of puberty, fertility, gestation, parturition, disease in a defined population such that the lactation, pregnancy outcomes, premature cases can be related to a population base, e.g. all reproductive senescence, or modifications in cancer cases in the state of Iowa. With this other functions that are dependent on the information, incidence rates can be calculated. integrity of the reproductive system. REGRESSION -statistical procedures that allow the RESISTANCE - the ratio of the voltage across an selection of the best numerical relation~ip object to the current following through it. among two or more variables. Given data on a dependent variable y (disease outcome) and one RESISTANCE - the property of material that or more independent exposure-related variables, determines the current produced by a given xl, x2, x3 ..... xn, regression analysis involves difference of potential. The unit of measurement Glossary G- 17 is the ohm. A difference of potential of one volt Risk Characterization The description of will produce a current of one ampere in a circuit the nature and often the magmmde of human where the resistance is one ohm. risk, including attendant uncertainty. RESPIRATION - the physical and chemical RISK ESTIMATE - the quantitative estimate of the processes by which an organism supplies its cells likelihood of adverse effects resulting from a and tissues with the oxygen needed for specified exposure. The relative risk and odds metabolism and relieves them of the carbon ratio are examples of risk estimates. dioxide formed m energy-producing reactions. RISK FACTOR an aspect of lifestyle, RETROSPECTIVE STUDY - a research design that environmental exposures, or an inherited is used to test etiologic hypotheses in which characteristic, which on the basis of references about exposure to the putative causal epidemiologic evidence, is shown to be factors are derived from historical data relating to associated with adverse health effects. exposure characteristics of the persons under study. R. MS (Root Mean Square) - the square root of the average of the squares of individual values. For RIBONUCLEIC ACID (RNA) - Ribose and nucleic a sinnsoidal variable, such as the amplitude of acids (any of various acids composed of a sugar 60-Hz alternating current, the rms value equals or derivative of a sugar, phosphoric acid, and a the peak value divided by the square root of two. base and found both in cells and cell nuclei. Measured and calculated EMF levels are usually Vital to cellular processes, especially the coding rms values. o f proteins. SAMPLE - any selected subset from a population. RIGHT-OF-WAY - the legally def'med corridor of land on which a transmission line is located. SAMPLE SIZE - the numba selected (sampled) from a population to be the subjects of study. RISK - the probability that an event will occur, e.g., that an individual will become ill or die within a SECRETION - 1) The process of segregating, stated period of time. elaborating, and releasing some material either functionally specialized, (i.e., saliva), or isolated RISK ASSESSMEb~ - Activity of evaluating the for excretion, (i.e., urine). 2) A product of toxic properties of an environmental agent and secretion formed by an animal or plant, the conditions of human exposure to it in order to especially one performing a specific useful ascertain the likelihood that exposed humans will function in the urgamsm. be adversely affected, and to characterize the nature of the effects they may experience. Risk SELECTION BIAS - error in the results of a study uscssment may contain some or all of the due to systematic differences in characteristics following four steps: between those who arc selected for study and those who arc not. For example, selecting only Hazard Identification The determination of HMO patients as study subjects might exclude whether a particular agent is or is not causally persons who are not employed or employable. linked to particular health effect(s). Thus the study's results would not accurately reflect disease patterns m the general population Dose-Response/issessment The determination of which includes unemployed people. the relation between the magnitude of exposure and the probability of occurrence of the health SEROTONIN - A phenolic amine (CiOH12N20) that effects ihqucstion. is a powerful vasoconstrictor and is found especially in the blood serum and gastric Exposure ,4ssessment The determination of (stomach) mucosa of Ktammals. Serotonin the extent ofhuman exposure. stimulates or inhibits many of the nerves and muscles, depending on the amount and the phase of the organ in its function. It can stimulate or G-18 Health Effects of Exposure to Powerline Frequency Electric and Magnetic Fields depress heartbeat, contract blood vessels and members. The ratio is usuallymultipliedby 100, change blood pressure. It prevents clotting, and so an SMR of 100 would mean that the expected provides reflexes such as cou~hinlg or hyper and observed deaths are essentially equal in ventilation. In humans, serotonin also serves as a number, and no excess risk is evident in the study chemical transmitter in the brain. Serotonin and group. its product melatonjn influences sleep, perception of pain, psychological depression and social SOLVENT - a substance usually liquid capable of behavior. dissolving or dispersing one or more other substances. Most industrial solvents are volatile SEROTONIN N-ACETYL TRANSFERASE (SNAT) organic compounds such as xylene, toluene, and - an enzyme that affects the proper functioning of perchloroethylene. serotonin, especially in relation to vasoconstriction. SQUARE-WAVE PULSE - the rectang~tlar wave form of an electric/magnetic field that varies SHAM GROUP - the experimental group that is periodically and abruptly from one to the other of treated as in a parallel experixnent except for two uniform values. omission of the procedure or agent under test, and which is used as a standard of comparison in STANDARD DEVIATION - a measure of the judging experimental effects (also called Control variation in a set of observations. The mean of Group). the observed values indicates where the values for a group are centered. The standard deviation SHIELDING - the cutting off or blockage of an is a summary of how widely dispersed the values organism from exposure to a stimulus, (e.g., are around this central value. EMFs) by the physical presence of another organism between the shielded organism and the STATIC FIELDS - elecuic and magnetic fields stimulus. that do not vary in intensity or strength with time. SIBLINGS - children borne by the same mother. STATISTICAL POWER - see POWER. SIEMENS - the SI term for the tnho, the historical unit ofconductance. STATISTICAL SIGNIFICANCE - a finding of an epidemiologic study is considered to be SIGN TEST - a statistical test ihat is used to test statistically significant if, according to certain data independent of any assumption about the assumptions and based on a mathematical distribution of the data. This test is used to probability, the finding has a low likelihood of compare the locations of two distributions when being due to chance or random variation. A test the samples are dependent. It is one of the of statistical significance is a measure of whether disUibution-free test alternatives for the t-test. a difference observed between the exposed and non-exposed groups in a study is statistically SIGNIFICANCE LEVEL The probability different from a chance occurrence. The predetermined by the investigator) that the test probability of an observed difference being due statistic will assume a value that will lead to a to chance may be expressed as a "~v" value. rejection of the null hypothesis when the null hypothesis is true. STATISTICS the science of collecting, ~t~nnmarizing and nnnly'zln_o data that are subject SINUSOIDAL - a regularly alternating electric field. to random variation such that the uncertainty of inductive inferences may be evaluated. The term SIVIR (STANDARDIZED MORTALITY OR is also applied to the _a~t_~ themselves and to MORBIDITY RAT[O) - a risk estimate which summarizations of the data, e,g. a statistic may compares the numbers of deaths or illness be a value that is computed from sample data or observed in the study group with the number any function of arandom variable. expected based on a comparison group (typically national or regional rates)while adjusting for STEADY-STATE - a state or condition of a system differences in the age and sex of the study group or process that does not change in time. Glossary G- 19 STEREOTYPE - something conforming to a fLxed or SURVIVABILITY - 1 ) Resulting in or permitting general pattern (Ex. Stereotypical behavior-a survival. 2) The ability to remain alive or exist standardized way of acting held in common by despite negative and potentially life-threatemng mereben of a group). stimuli. STEROID - any of numerous compounds containing SYNERGISM synergism exists between two the carbon ring system of the sterols and environmental agents if the risk of disease that including the sterols and various hormones and results from exposure to both agents is greater glycosides. than the sum of the risks for each individual exposure. STILL BIRTH - the birth of a dead fetus (dead at birth). SYSTEM 1 ) A regularly interacting or interdependent group of items forming a unified STIMULUS = an agent (as an environmental change) whole; 2) A group of body organs that together that directly influences the activity of living perform one or more vital functions. protoplasm (as by exciting a sensory organ or evoking muscular contraction or glandular SYSTEMATIC ERROR - distortion of study results secretion). due to non-random events. (See BIAS) STRAIN - (1) A group of presumed common TENSION - the stress resulting from unrest, or ancestry with clear-cut physiological but usually imbalance caused by the imposition of stimuli, not morphological distinctions. 2) Excessive often with physiological manifestation. physical or mental tension. TERATOGENIC - tending to cause developmental STRATIFICATION - an analyxic t,~hnique that realformations and monstrosities (abnormal in separates a sample into several subsamples growth or structure). according to specific criteria such a~ age. sex. or socioeconomic status. The effect or confounding TERATOGENIC - a substance that produces variables may be controlled by stratifying the abnormalities in the embryo or fetus by data for analysis of results. For c,'~plc, lung disturbing the mother's health or by acting cancer is known to be associated with smoking. ~tly on the fetus in utero. To examine the possible association between atmospheric pollution and lung cancer while TES LA - the unit of magnetic flux density(T), controlling for smoking, the population may be equivalent to 104 gauss (G) or 1 Wb/m2. divided into strata according to srttoking status. The association between air polhition and cancer TESTOSTERONE - a male hormone that is can then be appraised separately within each produced by the testes or made synthetically, and smoking status stratum. is responsible for inducing and maintaining male secondary sex characteristics, and is a crystalline STRESS - constraining force or influence - a force hydroxy steroid ketonc {C19H2802.} exerted when a physical, chemical or emotional factor that causes bodily or mental tension and THRESHOLD 1) The point at which a may be a factor in disease causation. physiological or psychological effect begins to be produced; 2) A level, point, or value above which SUBSTATION - a subsidiary station in which something will take place, and below which it electric current is transformed. will not. SURVIVAL ANALYSIS - the development and THYROID - a large endocrine gland of cramate application of statistical models anal methods for vertebrates, lying at the base of the neck and analyzing data representing survival times, producing especially the hormone thyroxine. waiting times, or occurrence timel to selected events. TIME-VARYING FIELDS - electric and magnetic fields that change in intensity or strength with G-20 Health Effects of Exposure to Powerline Frequency Electric and Magnetic Fields time. Examples include 60 Hz, modulated, and TUMOR REGISTRY - see REGISTRY. transient fields. TYPE I ERROR - the rejection of the null TIME-WEIGHTED EXPOSURE - a way of hypothesis when it is true. aver~in~ an individuars exposure to an environmental agent over a specified period of TYPE H ERROR - the failure to reject the null time. Specifically, the integration of a monitoring hypothesis when it is false (alternatively, this is curve of exposure to an environmental agent, the rejection ofthe alternative hypothesis when it divided by the total time the person was exposed. is true). Time-weighted occupational exposures are generally standardized to an 8-hour workday. ULTRADIAN -A biological activity or function being, having, characterized by, or occurring m TISSUE - an aggregate of cells usually of a particular kind together with their intercellular periods or cycles ofleu than 24-hour duration. substance that form one of the structural VALIDITY - the degree of accuracy of a materials of a plant or an animal. measurement. The degree to which a measurement measures what it purports to TRANSCRIPTION - the process of constructing a measure. In epidemiology, validity may also messenger RNA molecule using a DNA molecule refer to the degree to which study results may be as a ternplate with resulting transfer of genetic information to the messengerRNA. generalized to populations beyond the study sample, i.e. external validity. TRANSDUCTION - 1) The action or process of transducing (to convert [as energy or a message] VARIABLE - any attribute, phenomenon, event, or measure that can auume different values. into another form); espec~y the transfer of genetic determinants from one microorganism to VARIANCE - a measure of dispersion of a set of another by a vital agent (as a bacteriophage). 2) observations defined for a sample as the sum of To bring about the transfer (u a gene) from one microorganism to another by means of a vital squares of deviations from the mcan divided by one lcu than the total number of observations. agent. Standard deviation is the square root of the variance. (See STANDARD DEVIATION) TRANSFORMER - a device for changing from one set of voltage and current lcvels to another. VASOCONSTRICTION - natTowing of the lumen TRANSMISSION LINES power linu/wiru Coore or opening) of blood vessels, especially as a result of vasomotor (nerves or centers mounted on metal or wooden structures up to 50 controlling the size of blood vessels) action. meters in height which transmit (carry) electricity at voltages up to 765 kV and cuffcuts VOLTAGE - 1) Electric potcntial or potential of up to 2000 Amperes, from the power station to difference expressed in volts. 2) Voltage is a load centers and substations. measure of the electric potential energy that makes electric charges flow throuSh a circuit. t-TEST - a statistical test used to evaluate whether two arithmetic averages differ when the samplc WILCOXON TEST - a distribution-free statistical size is small. It is one of several methods that test is used with matched pairs of observations to can be used to evaluatc whether or not the result test the null hypothesis of no difference in the of an cpidemiologic study is statistically matched populations. This test is COmmOnly significant. (See STATISTICAL used in a pre-post test type study. (See SIGN SIGNIFICANCE) TEST) TUMOR - an abnormal mau of tissue that is not WINDOW EFFECT - for particular values of inflammatory, arises without obvious cause from frequencies and intensities, some electric and cells of preexistent tissue, and possess no physiologic function. magnetic field intensities produce an effect on test organisms/tissues, but othen don't. Glossary G- 2 1 Conversely, if an effect is observed at a "WINDOWED"RESPONSE -effects found within particular value of field, it might be "turned out" bands or ranges of frequency or intensity by changingthe frequency of the field. separated by bands or ranges without effect; nonlinear exposure-response relation. A-1 APPENDIX A COMPUTER CALCULATION OF ELECTRIC AND MAGNETIC FIELDS A.1 345-KV Transmission Line Configuration ~ ~iI~ ~ --, ~" ~ I ~ ~u ~ ,~ ~ "' - A-2 Heallit Effects of ~xposurt toPot4,erline Frequency F. lectric and Magnetic Tfwldx A.2 Corona Electric Field Report ELECTRIC FIELD CALCULATIONS SOtFYMWE$I'EItlNI PUBLIC SEItVICE CO., 345KN TUCO/OKLAUNION TRAN'SML.~ION lINE 2-795MCM CONDUCTOI~/PiiASE, 2-3116' SHmI~"~ i 33' MINIMUM CI.,EAIIANCE DB'T. FROM MAXIMUM SUBCON. NO. OF PilAtE it~BKENCE HEIGHT GRADIENT DLAM. SUBCON. ANGLE PH. A -27.50 33.00 15 .?3 !. 11 2. 0, PH. B .00 33.00 16.71 1.1 i 2. -120. PH. C 27.50 33.00 15.73 1.11 2. 120. Gl,4I) -21.50 58.30 5.76 .38 1. 0. GNI) 21.50 58.30 5.76 .38 1. 0. SEN,~it HT. - 3.3 FT. DIST FROM SPACE REFERENCE F,-F~-! D T!!ETA Ey-F~v~ n T!IETAY EX-Fr~J) 'I'MErAX POTENTIAL ~ ~'V/ME'r!~.;~ pEGREF_.~) ~rMETER~ pEGREF.~} ~'KNt/klETER~ pEO!tEES~ .0 3.324 90.0 3.324 60.0 .769 -30.0 32.51.7 10.0 3.302 86.0 3.300 15.6 .906 -67.7 3264.1 20.0 4.258 ~4.2 4.237 -25.0 .716 -12.6 4198.8 30.0 5.159 86.9 5.158 42.5 .307 -113.2 5101.4 40.0 4 .~)0 93.3 4.792 .41.5 .294 150.1 4 765.7 50.0 3.729 95.7 3.710 -50.4 .370 133.0 3712.4 60.0 2.679 96.5 2.662 -50.8 .302 129.5 2671.7 70.0 1.892 96.5 ! .880 -50.7 .213 12B.6 1888.8 ~0.0 1.3 50 96.2 1.342 -50.4 .145 128.6 134B. 7 90.0 .9S3 95.7 .978 -50.1 .098 128.8 982.6 100.0 .733 95.3 .730 -49.8 .068 129.2 732.2 110.0 .558 94.9 .556 49.5 .04~. 129.6 557.7 120.0 .434 94.6 .432 -49.3 .035 130.0 433.5 130.0 .343 94.2 .342 -49.2 .025 130.4 343.1 140.0 .276 94.0 .275 -49.1 .019 130.7 275.9 150.0 .223 93.7 .225 -49. I .015 131.0 223.1 160.0 .186 93.5 .186 -49.2 .01 ! 131.2 186.0 170.0 .155 93.3 .155 -49.3 .009 131.4 155.5 180.0 .131 93.1 .131 -49.4 .007 131.5 131.3 190.0 .112 92.9 . ! 12 -49.6 .006 13 ! .6 111.8 200.0 .096 92.8 .096 -49.8 .005 13 ! .6 96.1 210.0 .083 92.6 .083 -50.0 .004 131.7 83.2 220.0 .073 92.5 .072 -50.2 .003 13 ! .7 72.5 230.0 .064 92.4 .064 -50.5 .003 131.7 63.6 240.0 .056 92.3 .056 -50.8 .002 131.6 56.1 2.50.0 .050 92.2 .050 -51.1 .002 ! 31.5 49.7 260.0 .044 92.1 .044 -51.4 .002 131.5 44.3 270.0 .040 92.0 .040 -51.7 .001 13 1.4 39.7 280.0 .036 92.0 .036 -52. I .001 13 1.2 35.7 290.0 .032 91.9 .032 -52.4 .001 13 1.1 32.2 300.0 .029 91.8 .029 -52.8 .001 130.9 29.2 310.0 .027 9 !. 8 .026 -53.2 .iX) 1 ! 30.8 26.5 320.0 .024 91.7 .024 -53.5 .001 130.6 24.2 330.0 .022 91.6 .022 -53.9 .00 ! ! 30.4 22.1 340.0 .020 91.6 .020 -54.3 .001 130.2 20.3 350.0 .019 91.5 .019 -54.7 .001 130.0 18.6 360.0 .017 91.5 .017 -55.1 .000 129.8 17.2 370.0 .0 ! 6 91.5 .016 -55.5 .000 129.6 15.9 380.0 .015 91.4 .015 -55.9 .000 129.4 14.7 390.0 .014 91.4 .014 -56.3 .000 129.2 13.7 400.0 .013 91.3 .013 -56.7 .000 128.9 12.7 410.0 .012 91.3 .012 -57.1 .000 128.7 11.8 Appendix A A-3 DtST FROM SPACE ~CE E,-FIF:a.D THETA Ey-F~-Ln THETAY EX-FIELD THETAX POTENTIAL 420.0 .011 91.3 .011 -57.5 .000 128.5 11.1 430.0 .010 91.2 .010 -57.9 .000 128.2 10.3 440.0 .010 91.2 .010 -58.3 .000 128.0 9.7 450.0 .009 91.2 .009 -58.7 .000 127.7 9.1 460.0 .009 91.1 .009 -59.1 .000 127.5 470.0 .008 91.1 .008 -59.5 .000 127.2 8.0 480.0 .008 91.1 .008 -59.9 .000 126.9 7.6 490.0 .007 91.1 .007 -60.3 .000 126.7 7.2 500.0 .007 91.0 .007 .-60.7 .000 126.4 6.8 A-4 Health Effects of ~,,rt toPower!~e Freqmft~'y Electric amt Magnetic F'wlds A.3 Corona Magnsti~ Field R~port FIELD EFFECTS PROGRAM SOUTHWESTERN PUBLIC SERVICE CO., 345KV TUCO/OKLAUNION TRANSMISSION LINE 2-795MCIVI CONDUCTORS/PHASE, 2-3/16' SHIELD WIRE, 33' MINIMUM CLEARANCE I 0 3 5 345.0 10.00 1.00 (ENGLISH UNITS OFFION) LINE GRADIENT~ COMPUTED BY PROGRAM PHYSICAL SYSTEM CONSISTS OF 5 CONDUCTORS, WHICH 3 ARE ENERGIZED PHASES 4.921 6.562. 9.842 .000 ! .000 75 .GO0 3.280 .000 PH. A A -27.50 33.00 2 1. I 1 18.00 199.2 .0 .00 PH. B A .00 33.00 2 1.11 18.00 199.2 -120.0 .00 PH. C A 27.50 33.00 2 1.11 18.00 199.2 120.0 .00 GND A -21.50 58.30 I .38 .00 ,0 .0 .00 GND A 21.50 5830 I .38 .00 .0 .0 .00 51 .0 10.0 0 .0 .0 Apptn8~ A A-5 MAGNETIC FIELD CALCULATIONS 1000 AMPS SOU~TERN PUBLIC SERVICE CO., 345KV TUCO/OKLAUNION TRANSMISSION LINE 2-795MCM CONDUCTORS/PHASE, 2-3/16" SHIELD WIRE. 33' MINIMUM CLEARANCE DiST. FROM L-N MAXIMUM SUBCON NO. OF REFERENCE HEIGHT VOLTAGE GRADIENT DIAM. $U~CON. PH. A -27.50 33.00 199.2 15.73 1.11 2. PH. B .00 33.00 199,2 16.71 1.11 2. PH. C 27.50 33.00 199.2 15.73 1.11 2. SENSOR HT. = 3.3 PT. D(FT) B-FIELD THETA BY-FIELD THETAY BX-FIELD THETAX ~GAUSS/10O0,~P~ ~GAUSS/1000,~tP~ ~GAUSS/1OO0~W,4P~ .0 .19058038 61.9 ,19058038 -30.0 .10181256 240.0 ! 0.0 .19248731 61.7 ,18694734 -49, 1 .10044863 197,2 20,0 .18819978 50.3 .15313294 -61.0 .12708905 157.7 30.0 .16792006 28.9 .08363277 -72. ! .15135049 141. I 40.0 .13577562 10.7 .02561700 244.9 .13550900 135.8 50.0 .10078126 16.0 .02892969 163.9 ,10060552 134,6 60.0 ,07850940 28.0 .03699331 148.8 .06969700 134.9 70.0 .06019853 37,3 .03660011 145.0 ,04802834 135.6 80.0 .04714161 44.4 ,03305174 143.7 .03374459 136.5 90.0 .03772133 49.9 .02888295 143.4 .02433943 137.5 100.0 .00078302 54.2 .02498904 143.4 ,01802359 138.3 110.0 .02555809 57.7 ,02161276 143.6 .01367249 139.1 120.0 .02153971 60.5 .01876447 143,9 .01059633 139.8 130.0 .01838965 62.9 .01638242 144.1 .00836824 140.4 140.0 .01587789 65,0 .01439091 144.4 .00671860 141.0 150.0 ,01384461 66.7 .01271991 144.7 .00547298 141.5 160.0 .01217648 68.2 .01131021 144.9 .00451575 142.0 170.0 .01079153 69.6 .01011356 145.2 .00376854 142.4 180,0 ,00962942 70.7 .00909124 145,4 .00317701 142.8 190.0 .00864495 71,8 .0082123 1 145.6 ,00270278 143.1 200,0 .00780079 72.7 .00745204 145.8 .00231825 143.4 210.0 .00707949 73.6 .00679057 145.9 .00200322 143.7 220.0 .00645141 74.3 .00621188 146.1 .00174271 144.0 230,0 .00590326 75.0 .00570000 146,2 .00152543 144.2 240.0 .00542204 75.7 .00525332 146.4 .00134278 144.5 250.0 .00499731 76.2 .00485416 146.5 .00118815 144.7 260.0 ,00462057 76.8 ,00449831 146.6 .00105636 144.9 270.0 .00428485 77.3 .00417981 146.8 .00094335 145.1 280.0 .00398442 77.7 .00389366 146.9 ,00084590 145.2 290.0 .00371449 78.2 .00363568 147.0 .00076141 ! 45.4 300.0 .00347109 78.6 .0034023 1 147,1 .00068781 145.5 310.0 .0032~084 78.9 .00319054 147, I .0(X)62.339 145.7 320,0 .00305090 79.3 .00299782 147.2 .00056677 145.8 330.0 .00286885 79.6 .00282194 147.3 .00051680 145.9 340.0 .00270261 79.9 .00266100 147.4 .00047254 146.0 350.0 .00255042 80,2 .00251338 147.5 .00043319 146.1 360.0 .00241072 80.5 ,00237764 147,5 ,00039808 146.3 370.0 .00228220 80.8 .00225256 147.6 .00036668 146.4 380.0 .00216368 81,0 .00213705 147.6 .00033849 146.4 390.0 .00205416 81.2 .00203017 147.7 ,00031311 146.5 400.0 .00195275 81.5 .00193107 147.8 ,00029021 146,6 410.0 .00185867 81.7 .00183903 147.8 .00026949 146.7 420.0 .00177122 81.9 ,00175340 147.9 .00025070 146.8 430.0 .00168981 82.1 .00167358 147.9 .00023361 146.8 440.0 .00161388 82.2 .00159908 148.0 .00021804 146.9 450,0 .00154295 82.4 .00152943 ! 48.0 .00020383 147.0 A-6 Health F~ects of F. rposure toPowerliat Frequtncy F_.ltctric and Magnttic F~lds D(FT) B-Fun n THETA By-lmu n THETAY BX-FIF..LD THETAX (OAUSS/1000AM!~ (OAUSS/10GOAMP) (0AUSS/1000AMP} 460.0 .00147660 12.6 .00146422 148.0 .00019082 147.1 470.0 .00141444 S2.7 .00140308 148.1 .00017890 147.1 480.0 .00135612 12.9 .00134568 148.1 .00016795 147.2 490.0 .00130134 83.0 .00129173 148.2 .0G015788 147.2 500.0 .00124981 83.2 .G0124094 148.2 .0G014859 147.3 Appendix A A-7 A.4 Transpac Electric Field Report ELECTRIC RELDS ANALYSIS PROGRAM OUTPUT: SUMMARY OF LINE DATA USED IN CALCULATIONS: CASE DESCRIPTION: SOUTHWESTERN PUBLIC SERVICE CO., 345KV TUCO/OKLAUNION TRANSMISSION LINE AT X AND Y COORDINATES: -27.5000 33.0000 BUNDLE IDENTIFICATION: PHASE A LINE HAS 2 ACSR 795 26/7 SUBCONDUCTORS SUBCOND. DIA. = 1.1080 INCHES BUNDLE SPACING: 18.0 INCHES NOMINAL VOLTAGE IS 345.0 kV NOMINAL PHASING lS 0.0 DEGREES AT X AND Y COORDINATES: 0.000000 33.0000 BUNDLE IDENTIHCATION: PHASE B LINE HAS 2 ACSR 795 26/7 SUBCONDUCTORS SUBCOND. DIA.: I. 1080 INCHES BUNDLE SPACING: 18.0 INCHES NOMINAL VOLTAGE IS 345.0 kV NOMINAL PHASING IS -120.0 DEGREES AT X AND Y COORDINATES: 27.5000 33.0000 BUNDLE IDENTIFICATION: PHASE C LINE HAS 2 ACSR 795 26/7 SUBCONDUCTORS SUBCOND. DIA. -- 1.1080 INCHES BUNDLE SPACING= 18.0 INCHES NOMINAL VOLTAGE IS 345.0 kV NOMINAL PHASING IS 120.0 DEGREES AT X AND Y COORDINATES: 21.5000 58.3000 BUNDLE IDENTIHCAT1ON: RIGHT SHIELD LINE HAS 1 EHS STEEL 3/8 CL A SUBCONDUCTORS SUBCOND. DIA. = 0.3600 INCHES CONDUCTORS ASSUMED GROUNDED, VOLTAGE SET TO ZERO. AT X AND Y COORDINATES: -21.5000 58.3000 BUNDLE IDENTIHCATION: LEFT SHIELD LINE HAS 1 EHS STEEL 3/8 CL A SUBCONDUCTORS SUBCOND. DIA. = 0.3600 INCHES CONDUCTORS ASSUMED GROUNDED, VOLTAGE SET TO ZERO. AVERAGE MAX SUBCONDUCTOR SURFACE GRADIENTS, kV/cm: PHASE A 15.7 kV/cm PHASE B 16.7 kV/cm PHASE C 15.7 kV/cm RIGHT SHIELD 6.0 kV/cm LEFT SHIELD 6.0 kV/cm A-8 Health Effects of F2cposure toPowerline Frequency Electric and Magnetic Fields FIELD 18 CALCULATED AT AN ELEVATION OF 3.28084 FEET MAX E FIELD 18 5.16 KV/M PLOT I8 IN PER-UNIT OF MAX FIELD X E kV/m 0.25 0.50 0.75 1.00 0.0 3.33 __, __ 10.0 3.30 * I 20.0 4.26 * 30.0 5.16 · 40.0 4.80 * 50.0 3.73 · 60.0 2.68 * 70.0 1.89 * 80.0 1.35 * 90.0 0.98 * 100.0 0.73 , 110.0 0.56 * 120.0 0.43 * 130.0 0.34 * 140.0 0.28 * 150.0 0.23 I_* 160.0 0.19 * 170.0 0.16 * 180.0 0.13 * 190.0 0.11 * 200.0 0.10 * 210.0 0.08 * 220.0 0.07 * 230.0 0.06 * 240.0 0.06 * 250.0 0.05 * 260.0 0.04 * 270.0 0.04* 280.0 0.04* 290.0 0.03* 300.0 0.03* 310.0 0.03* 320.0 0.02* 330.0 0.02* 340.0 0.02* 350.0 0.02* 360.0 0.02* 370.0 0.02* 380.0 0.01' 390.0 0.01' 400.0 0.01' 410.0 0.01' 420.0 0.01' 430.0 0.01' 440.0 0.01' 450.0 0.01' 460.0 0.01' 470.0 0.01' 480.0 0.01' 490.0 0.01' 500.0 0.01' Appendix A A-9 A.5 Transpac Magnetic Field Report MAGNETIC RELDS ANALYSIS PROGRAM OUTPUT: SUMMARY OF LLNE DATA USED IN CALCULATIONS: CASE DESCRIPTION: SOUTHWESTERN PUBLIC SERVICE CO., 345KV TUCOIOKLAUNION TRANSMISSION LINE EARTH RESIS~ SET TO 33.3 OHM-METERS EARTH PERMITTIVJTY SET TO 2.8 (P, ELATWE) EFFECTS OF SHIELD CURRENTS INCLUDED. AT X AND Y COORDINATES: -27.5000 33.0000 BUNDLE IDENTIFICATION: PHASE A LINE HAS 2 ACSR 795 26/7 SUBCONDUCTORS NOMINAL CURRENT IS 1000.0 AMPS NOMINAL PHASING IS 0.0 DEGREES AT X AND Y COORDINATES: 0.000000 33.0000 BUNDLE IDENTIFICATION: PHASE B LINE HAS 2 ACSR 795 26/7 SUBCONDUCTORS NOMINAL CURRENT 1S 1000.0 AMPS NOMINAL PHASING IS -120.0 DEGREES AT X AND Y COORDINATES: 27.5000 33.0000 BUNDLE IDENTIFICATION: PHASE C LINE HAS 2 ACSR 795 26/7 SUBCONDUCTORS NOMINAL CURRENT IS 1000.0 AMPS NOMINAL PHASING IS 120.0 DEGREES AT X AND Y COORDINATES: 21.5000 58.3000 BUNDLE IDENTIFICATION: RIGHT SHIELD LINE HAS 1 EHS STEEL 3/8 CL A SUBCONDUCTORS, SUBCOND. DIA. = 0,3600 INCHES SUBCOND. RESISTANCE= 7.1900 OHMS/IVll REAC. AT I FT SP. = 1.5200 OHMS/MI CONDUCTORS ASSUMED GROUNDED, VOLTAGE SET TO ZERO. EFFECTS OF SHIELD CURRENTS INCLUDED. AT X AND Y COORDINATES: -21.5000 58.3000 BUNDLE IDENTIFICATION: LEFT SHIELD LINE HAS 1 EHS STEEL 3/8 CL A SUBCONDUCTORS SUBCOND. DIA. = 0.3600 INCHES SUBCOND. RESISTANCE= 7.1900 OHMS/MI REAC. AT 1 FT SP, = 1.5200 OHMS/MI CONDUC'I~RS ASSUMED GROUNDED, VOLTAGE SET TO ZERO. EFFECTS OF SHIELD CURRENTS INCLUDED. CALCULATED SHIELD CURRENTS ARE: RIGHT SHIELD 11.0 AMPS LEFT SHIELD 10.6 AMPS A- 10 Hta~h E~ects of F.V~r e to Powertiat Frequeacy F, ltclric aad !~agaetic Fields FIELD 13 CALCULATI~D AT AN i~LEVATION OP 3.2J0~4 FF.L~ MAX B FIELD iS 192.57 mG PLOT I3 IN PER-UNIT OF MAX F!!~LD X B m0 0.25 0.50 0.7~ 1.(30 0.0 3.33 [ 10.0 3.30 ]I ] [ , 2o.o 4.26 ] ] * so.o 5.16 ] ] * 40.0 4.80 ] 5o.o 3.73 1_, ] 60.0 2.68 * 70.0 1.89 80.0 1.35! · 90.0 0.98 lO0.O 0.73 · 110.0 0.56 * 120.0 0.43 * 130.0 0.34 * 140.0 0.25 * 150.0 0.23 160.0 0.19 * 170.00. 16 * 180.0 0.13 * 190.0 0.11 200.0 0.10]_, 210.0 0.08] * 220.0 0.07 ] * 230.0 0.06 ] * 240.0 0.06 ] 250.0 0.05 ]_* 260.0 0.04 * 270.0 0.04 * 280.0 0.04 * 290.0 0.03! * 300.0 0.03 * 310.0 0.03 * 320.0 0.0'2 * 330.0 0.02 * 340.0 0.02 * 350.0 0.02 * 360.0 0.02 * 370.0 0.02 * 380.0 0.01 * 390.0 0.01 * 400.00.O1 * 410.0 0.01 420.0 0.01 430.0 0.01 440.0 0.01 * 450.00.O1 460.0 0.01' ] 470.0 0.01* I 480.0 0.01* 490.0 0.01' 500.0 0.01' ] Appendix A A- 11 A.6 Expocalc Electric Field Repod File: spi345ok Dat~: Fri 2/16/90 Time: 5:51 PM Electric Field Profiles - kV/m Sensor Height - 3.28 ft. Distance Gnund Cle. ann~=(s)-to Cen~r of Bundle or Cond~tor (~) from CL (It;) 33.0 43.0 53.0 63.0 0 3.324 1.633 0.853 0.460 10 3.302 1.936 1,186 0.759 20 4.2.59 2.673 1.756 1.196 30 5.159 3.249 2.182 1.528 40 4.800 3.276 2.315 1.684 50 3.729 2.864 2.180 1.669 60 2.679 2.302 1.896 1.537 70 1.892 1.780 1.574 1.349 80 1.350 1.359 1.275 1.148 90 0.983 1.041 1.024 0.962 100 0.733 0.806 0.823 0.800 110 0.558 0.631 0.664 0.664 120 0.434 O. 501 0.540 0.553 130 0.343 0.403 0.443 0.462 140 0.276 0.329 0.366 0.389 150 0.22.5 0.271 0.306 0.329 160 O. 186 0.226 0.2.5 7 0.280 170 0.156 0.190 0.218 0.240 1S0 0.131 0.162 0.187 0.207 190 0.112 0.138 0.161 0.179 200 0.096 0.119 0.140 0.156 210 0.1383 0.104 0.122 0.137 220 0.073 0.091 0,107 0.121 230 0.1364 0.0g0 0.094 0.107 240 0.056 0.071 0.084 0.095 230 0.050 0.063 0.075 0.085 260 0.044 0.056 0.067 0.076 270 0.040 0.050 0.060 0.069 280 0.036 0.045 0.054 0.062 290 0.032 0.041 0.049 0.056 300 0.029 0.037 0.044 0.051 310 0.027 0.034 0.040 0.047 320 0.024 0.03 1 0.037 0.043 330 0.022 0.028 0.034 0,039 340 0.020 0.026 0.03 1 0.036 350 0.019 0.024 0.029 0.033 360 0.017 0.022 0.026 0.031 370 0.016 0.020 0.024 0.028 380 0.015 0.019 0.023 0.026 390 0.014 0.017 0.021 0.025 400 0.013 0.016 0.020 0.023 410 0.012 0.015 0.018 0.021 Diman~ Onausl Ckaranc.,(s)-m Com~ of ~undk ot Condglot (fi) fmmCL 420 0.011 0.014 0.017 0.020 430 0.010 0.013 0.016 0.019 440 0.010 0.012 0.015 0.018 450 0.0139 0.012 0.014 0.016 460 0.009 0.011 0.013 0.015 470 O.GOO 0.010 0.012 0.015 480 0.OO~ 0.010 0.012 0.014 490 0.007 0.009 0.011 0.013 500 0.007 0.009 0.011 0.012 510 0.006 O.OlM O.OlO 0.012 App endk A A- 13 A- 14 Health F. ffects of D4x3s~r t toPowtrltse Frtqfftz~/ ~ ~7 ~~ Magne~ F~M ~ F~: ~k D~: F~ ~1~ Y~: 5:52 ~ Mn~ ~ ~ ~h -mG ~r H~ - 3.~ ~. (n} 33.0 ~ .0 53.0 ".0 0 1~,58 133 .U ~*~ ~,~ 10 l~,e 132,47 95~ 71,27 ~ IU.~ 1~.61 ~.~ ~-~ ~ 167 .~ 114.M U .~ ~ 135.~ 97.~ ~.61 ~ ~31 ~.~ 52.~ ~.~ ~ ~.~ 51., ~.15 37.61 ~ 47.14 41.~ ~.~ 32.~ ~ 37.~ ~9 ~.95 27.63 110 ~ .~ ~.~ ~35 ~.S9 1~ 21 ~ ~.~ 19.~ 17.~ I~ ItS9 17.61 16.~ 1~ 15.~ 15.~ 14.61 13.~ 1~ 13 .~ 13.~ 12,~ 1~ 12.18 11 .~ i I .~ 10.~ 1~ !0., 10.~3 I0.~ 9.~ 1~ 9.~ 9.42 9.16 1~ 8.~ 8.~ 8.27 ~ 7.~ 7.67 7.~ 210 7.~ 6.97 6.~ 6.~ ~ 6.45 6.~ 6.24 6.10 ~ ~ .~ 4.~ 4.87 4.~ ~ 4.62 4.57 4~ i 4.~ 2~ 4.~ 4.~ 4.19 4.13 ~ 3 .~ 3.95 3 .~ 3 ~ 3.71 3.~ 3.~ 3.~ ~ 3.47 3.~ 3.41 3~7 310 3.~ 3.B 3.~ 3.16 3~ 3 .~ 3 .~ 3 .~ 2.97 3~ 2.87 2.U 2.U 2.~ ~ 2.~ 2.~ 2.67 2.~ 3~ 2.55 2.~ 2.52 2~ ~ 2.41 2.~ 2.38 3~ 2.~ 2.27 2.~ 2.B 3~ 2.16 2.15 2.14 2.12 3~ 2.~ 2.~ 2.~ 2.~ ~ i .95 1 .~ i .~ I Appendix A A-15 Diata~e Ground Clearance(,)--to Cealer of Bundle or CoaduGlot (l~) from CL ~;~ ~s .o 43 .o s3.0 63 .o 410 1 .M 1.85 1.84 1.83 420 1.7'7 1.76 1.76 1 .'74 430 1.69 1.68 1.68 1.67 440 1.61 1.61 1.60 1 450 1 ..54 1.54 1 460 1.48 1.47 1.47 1 .,145 470 1.41 1.41 1.40 1.40 480 1.36 1.35 1.35 1.34 490 1.30 1.30 1.29 1.29 .500 ! .7,5 1.75 1.24 1.24 .510 1.20 1.20 1.19 1.19 Appendix A A- 17 A.8 Comparison of Programs' Calculated Results ELECTRIC FIELD STRENGTH ~4AGNETIC FIELD FLUX DENSflY E-field E-field~ E-field B~eld B~¢ld B~eld Distance BPA Expocalc: Tranapac BPA Expocalc Transpac (ft.) (v/m) 0 3324 3324 3330 190.6 190.6 190.4 2 3316 3316 3320 190.7 190.7 190.6 4 3294 3294 3290 191.1 191.1 191.0' 6 3271 3271 3270 191.6 191.6 191.6 8 3267 3267 3270 192.1 192.1 192.1 10 3302 3302 3300 192.5 192.5 192.6 12 3394 3394 3400 192.6 192.6 192.7 14 3548 3548 3550 192.3 192.3 192.5 16 3755 3755 3760 191.6 191.6 191.8 18 3999 3999 4000 190.2 190.2 190.4 20 4258 4259 4260 188.2 188.2 188.5 22 4512 4512 4510 185.5 185.5 185.8 24 4741 4741 4740 182. ! ~ 182.1 182.4 26 4932 4932 4930 178.0 178.0 178.3 28 5073 5073 5070 ! 73.3 173.3 173.6 30 5159 5159 5160 167.9 167.9 168.3 32 5187 5187 5190 162.1 162.1 162.4 34 5160 5160 5160 155.8 155.8 156.2 36 5082 5082 5080 149.3 149.3 149.6 38 4959 4959 4960 142.6 142.6 142.9 40 4800 4800 4800 135.8 135.8 136.1, 42 4613 4613' 4610 129.0 129.0 129.4 44 4405 4405 4410 ! 22.4 122.4 122.7 46 4185 4185 4190 115.9 116.0 116.3 48 3957 3957 3960 109.7 109.7 110. 1 50 3729 3729 3730 103.8 103.8 104.1 52 3503 3503 3500 98.1 98.1 98.5 54 3283 3283 3280 92.8 92.8 93.1 56 3072 3072 3070' 87.7 87.7 88.0 58 2870 2870 2870 83.0 83.0 83.3 60 2679 2679 2680 78.5 78.5 78.8 62 2499 2499 2500 74.3 74.3 74.6 64 2331 2331 2330 70.4 70.4 70.7 66 2174 2174 2170 66.8 66.8 67.1 68 2028 2028 2030 63.4 63.4 63.7 70 1892 1892 1890 60.2 60.2 60.5 72 1766 1766 i 770 57.2 57.2, 57.5 74 1650 1650 ! 650 54.4 54.5 54.7 76 1542 1542 1540 51.8 51.9 52.1 78 1442 1442 1440 49.4 49.4 49.7 A-18 Health Effecls of Exposure loPowerline Frtqutncy E~clric and Magntlic Fitlds !:I Fr~IUC F]EL.D STRE, NGTH MAGNETIC FIT:J..D FLUX DENSITY E,-fekl E.-f~id E,-6dd B6eJd B6eJd B6dd Diaw, e BPA Expor.~ Trauq,,c SPA Expocdc T~ (ft.) (V/m) (V/m) (V/m) (raG) (raG) (raG) 80 1350 13501 1350 47.1 47.1 47.4 82 1265 1265i 1270 45.0 45.0 45.3 84 1186 1186[ !190 43.0 43.0 43.3 86 !113 1113 1110 41.1 41.1 41.4 88 1046 1046 1050 39.4 39.4 39.6 90 983 983 980 37.7 37.7 37.9 92 926 926 930 36.2 36.2 36.4 94 872 872 870 34.7 34.7 34.9 96 822 822 820 33.3 33.3 33.5 98 776 776 780 32.0 32.0 32.2 !O0 733 733 730[ 30.8 30.8 31.0 110 558 558 560' 25.6 25.6 25.7 120 434 434 430 21.5 21.5 21.7 130 343 343 340 18.4 18.4 18.6 140 276 276 280 15.9 15.9 16.0 150 225 225 230 13.8! 13.8 14.0 160 186 186 190 12.2 12.2 12.3 170 155 156 160 10.8 10.8 10.9 180 131 131 130 9,6 9.6 9.8 190 112 112 110 8.6 8.6. 8.8 200 96 96 100 7.8 7.8~ 7.9 210 83 83 80 7.1 7.1 7.2 220 73 73 70 6.5 6.5 6.6 230 64 64 60 5.9 5,9 6.0 240 56 56 60 5.4 5.4 5.5 250 50 50 50 5.0 5.0 5.1 260 44 44 40 4.6 4.6 4.7 270 40 40 40 4.3 4.3 4.4 280 36 36 40 4.0 4.0 4. ! 290 32 32 30 3.7 3.7 3.8 300 29 29 30 3.5 3.5 3.6 310 27 27 30 3.3 3.3 3.3 320 24 24 20 3.1 3,1 3.1 330 22 22 20 2.9 2.9 3.0 I 3 I 20 20 20 2.7 2.7 2.8 350' 19 19 20 2.6 2.6 2.6 360 17 17 20 2.4 2.4 2.5 370 16 16 20 2.3 2.3 2.4 380 15 15 !0 2.2 2.2 2.3 390 14 14 !0 2.1 2.1 2.1 400 13 13 10 2.0 2.0 2.0 410 12 12 10 1.9 1.9 1.9 Appendix A A-19 ELECTRIC FIELD STRENGTH MAGNETIC FIELD FLUX DENSITY E-fi~ld E-fi~ld E-fi~ld B~ld} Bfi~id Bficid Disiaac~ BPA F. xpocaic Trmnspac BpA~ Rxpocalc Trampac (~.) if/m) if/m) if/m) (mq (n,q CmG) 420 11 1 i 10 1.8 1.8. 1 430 10 10 10 1.7 1.7 1.8 440 10 10 10 1.6 1.6 1 4~0 9 9 10 1.5 1.5 1.6 460 9 9 10 1.5 1.5 1.6 470 8 8 10 ! .4 1,4 1 480 8 8 10 1.4 i .4 i .5 490 7 7 10 1.3 1.3 1.4 500 7 7 10 1.2 1.3 1.3 B-1 APPENDIX B - FUNDAMENTALS OF EPIDEMIOLOGY1 "confounding" factors may lead to an incorrect B. 1 Epidemiologic Methods intemreauon of thc relationship bcvwccn thc agcnt and the discasc undcr study. Epidcmiology i= gencraUy defined as the study of thc distribution of disease in human populations and the Other examples of the methodology problems that can determinants of that distribution. Characteristics of alter or bias observational studies include: uncertainties people and their environment may be examined for in determining the actual exposure status of individuals, possible causal associations with the occurrence of variations in disease definitions and diagnoses in human disease. different geographic areas or in different hospitals, loss of study subjects who leave the area, unwillingness of Because epidemiology draws its conclusions from subjects to participate, and inaccuracies in frequently observations of the natural distribution of disease, it used data sources such as death certificates and clinical possesses both unique strengths and limitations. Since records. Practical solutions to many of these problems humans are the subjects of study, epidemiology avoids have been developed by epidemiologists, although the problem of extrapolating from animal experiments frequently these sources of bias are not adequately in which both the exposure conditions and the addressed. appropriateness of the animal model are often questioned. Although each epidemiologic investigation poses its own unique problems and solutions, the overall On the other hand, epidcmiologic research generally approach of a study generally follows one of a several provides less conclusive fmdings than laboratory basic study designs. The choice of a study design will research does. The inability of epidemiologic research depend on many factors such as time and cost to offer direct proof of a cause-and-effect relationship limitations, frequency of the disease(s) to be studied, results from its observational methodology. In a frequency of the exposure, intended use of the laboratory inveatigation of a suspected harmful agent, it information, and the availability of required data. is assumed that the animals under study differ only on Several commonly used designs are describcti below, the basis of their exposure regimen. Any ensuing along with a brief consideration of their particular differences that are found between exposed and non- advantages and limitations. exposed animals can then reasonably be attributed to the exposure itself. Since obvious ethical and practical The terms incidence and prevalence will be used in the prohibitions on experimentation with humans exist, data following descriptions of epidemiologic study designs. must be collected on the "natural' occurrence of the These two commonly used measures of disease disease and agent under study in human populations. occurrence have distinctly different meanings in However, human exposure to an agent is not a random epidemiology. in a population-based study, the phenomenon occurring among memben of a prevalence of disease is the proportion of individuals in homogeneous population. Exposed and non-exposed a population with the disease at a given point in time. groups will differ in terms of age, residence, For example, the number of penons with lung cancer occupation, gender, and many other facton. Some of in a population of 100,000 on December 31, 1989 these variables are known to influence disease might be 30. The prevalence of lung cancer in this occurrence and can be accounted for in the design or population at this point in time is 30/100,000 or, .0003. analysis of a study. This figure would include all persons with lung cancer on December 31, regardless of whether the person has Other factors associated with both the disease and the had the disease for one day or three years. Prevalence exposure may not be known to the investigator and, is dimensionless, i.e., it has no units. therefore, cannot be accounted for. Such In a case-control study, on the other hand, disme defines the two types of individuals to be studied, those with disease (cases) and those without (controls). l'Source: Bard on materials initially prepared by Robeft S. Banks, R.S. Baaka Anociates, Inc. for Beetfie Power Here, the concern is to compare these two types of Remeareh Institute S~minar on New EMF Epidemiologie individuals with respect to the proportion having a Remlte and Their Implleations, October 16-19, 1990. history of an exposure or characteristic of interest. For Submatial r~vi~ions and additiom have b~n made by Boji example, in a case-control study of persons with lung Humt, M.D., and P.A. Buffier, Ph.D., University of Tens cancer, it would be of interest to compare the Healffi S¢ienee Center at Houaton, School of Public Health, proportion having a history of cigarette smoking in case and R. A. !kauehamp, Tens Depafia~nt of Heal:h. and control subjects. Appendix B B-3 · Relatively rare exposures (or occupations) Th~ limitations of retrospective cohort studies include: can be investigated. * Past ¢xpolurcs cannot be defined as precisely Prosp~ive cohort studies also have · number of as current exposures. For example, it may be limitations: difficult or impossible to estimate exposures to workm that occurred 40 years ago if no · They ofi~ requir~ many years of follow-up industrial hygiene data arc available and work since many diseases have long htencics. practices have changed over time. · A very large cohort and/or · long follow-up · Littic information may be available on period would be required to investigate confounding factors, such Its smoking history. relatively rare discsacs, including most cancers. . It may difficult to select a suitable popuhtion to which the cohort can be compared. · Substantial effort and expense arc necessary Frequently, the study results will differ to follow a large number of people over · dciznding on whether national, regional, or long period of time. local disease rat. arc used as the comparison. The problcrn of · suitable There arc no known EMF prospective cohort studies. comparison population is avoided in hrgc cohort studies in which internal comparisons Retrospective (Historical or Nonconcurrcat) Cohort can be made, i.e., a particuhr category of Studies. Rclxospcctivc cohort studies differ in that both workers within the cohort can be compared to the exposure and the follow-up ;scriod have occurred the complcto cohort. prior to the onset of the study. These studies u'tdizc data from existing records such as occupational Some cohort studies involve both prospective and records, professional registries, and death ccrti~cat. retrospective components. For example, a cohort may to identify the cohort and conduct the follow-up. be defined through porsonncl and other records as Studies of occupational groups arc most oRcn everyone who worked at Company X at least one conducted using this approach. As with prospective month between 1945 and 1985. The mortality studies, the cohort is first defined on the basis of experience of this population as of 1986 (the time the exposure status. Fo~ example, the cohort may be study is undertaken) can then be determined using state defined as all the members of· particular occupation or and national mortality rw~ords. Additional follow-up of all employees at Company X as of some specified time this cohort might then be conducted in 1990 or 1995, in the past. The subsequent oc~urr~sc~ of dis~as¢ in for example. the cohort (up to the time of the study) is then asocrtaincd, go·orally by using death corti~catos. To Milham, 1988 is · cohort of amateur radio operators. miucc costs, an unoxposcd cohort is often not Lurid, 1985 is · modified cohort study in which electric identified for comparison; instead_, the mortality worker union members wcrc followed for six years to cxporicncc of the exposed group is usnaHy compared to determine their !cukcmla risk. This study was designed that exporicnccd by the general population in the state, only to be · screening tool however, and the rigorous region or country from which the study population was metbuds generally used in cohort studies to ascertain derived. and validate complctsmess of the cohort, cause of death, and exposure wcrc not employed. Two case,-control The advantages of rcffospcctive cohort studies include: studies of clw. lric. al workers have b4~n derived from large coho~ studies that were not conccrocd · Much less time and less cost is require. d to specifically with persons wol46ng in electrical completo th¢ study comtmmi to · proslx~.iv¢ occupations (Stem et sl., 1986 and Gilman et sl., study, since the disc·so outcome has ah'~ady 1985). oc~urr~l. Case-Control Studks · These studies arc widely used in occupationsl This is the most common study design used in settings, where porsonnel records, industrial hygiene data and other records can be used cpidcmiology. As outlined above, cohort studies first both to construct the cohort and establish identify the exposure status of non-diseased individuals, then determine the subsequent incidence of disc·so in sorac rough measures of exposure. the cohort. In ~ontrsst, case-control studies begin by first identifying individuals who have developed the Appendix B B-5 incidence ratio (PIR) shows the observed-to-expected McDowall, 1983; CAile and 5avitz, 1985; Milham, ratio of eases. The population of this type of study 1985; and Peatee et sl., 1985. consists of those with cancers newly identifi~ in a specified time period, and the study clata is derived Cluster Studies rnotUJity dan, the proportions of cancers of different A cluster of disease cases (e.g., ieukemla) is genendly types e~n be compared among different oc~ul~tional considered to be an unustmUy high number of eases pup (if o~upational data tre avtilable in the appearing in the same setting (e.g. , ncighhorhood, t~,ord). town, work place) over a limited podod of time. Considerable attention has been given to the study of PMR studies have the following advantages: disease clusters, particularly cancer clusters, and a number of statistical methods have been developed to · They can be conducted very quickly and analyze them. Nevertheless, no non-infectious agent inexpensively, especially if the relevant data has ever been consistently implicated in causation of tre tlte~dy computerazed (as they often are). any type of cancer by cluster analyses. · Many different occupations and causes of Cancer clustering is therefore generally thought to be death (or types of cancer) can be examined the result of random (i.e., non-uniform) distribution of simultaneously. disease cases in a population. Thus, even when a significant elevation of a disease is noted for a specified · They can provide many useful leads for time and place, it is very likely a statistical artifact and possibly relationships b~w~zn disease and not the result of the cases' exposure to an agent in their occupation that can be investigated by more environment. For example, three new eases of ieukemla might occur in a conununity of 4,000 person powerful study methods. in a five-year period, and this incidence might be PMR studies also have song severe limitations: significantly 0~ = 0.04) elevated relative to the incidence in a comparison population for the same time period. But such a cluster might not be unusual. The · Uttlike a cohort study, no infornuttion is calculations used to determine the statistical colleaed on the population 'at risk"; only on significance of the increased incidence of leukemia those alnmdy deceased. Therefore, no imply that if there are 1,000 communities of 4,000 disease rates can be caleulataxl. An elevation persons in the country, 20 or 40 communities in the PMR may be due either to an increase throughout the country (depending on the type of in mortality from the cause of concern or to a statistical test used) might have an elevated incidence of reduction in mortality from another cause. leukemia due to random statistical variation alone. For example, a PMR for leukemia may be Thus, such n cluster, while 'sta~y signifwamt," elevated in power linemen because they are would not be all that rare. It is also unlikely that actually at increased risk of ieukemia, or similar environmental agents, le, ukemia subtypes, or because they have lower risk of some other age or gender distributions of the eats would be found common diseaR such as het~ disease. in these 20 or 40 communities. Information on confounding factors (such as smoking) it is possible that a cluster of cancer cases may be is usually not available. related to the presence of an environmental carcinogen. The fact that no such relationship has been established · Accuracy and completeneu of information through cluster analysis may in apart be due to the fact obtained from death certifi,'_~,3_~ or medical that cancer clusters generally consist of so few eases ree, ords is variabl~ (see be, low). tlmt it is not possible to t~t a hypothesis relating an environmental car~inogen to cancer risk. Thus, the · ~tudy t~ults can sometimes differ quite sic of the population being studied is generally significantly when compared to r~ults of inaufficient to yield statistically significant r~ults. more definitive studies such as cohort studies, especially if the overall mortality rates differ An example of a cluster study is Aldrich et aL, 1984. substantially among the groups (occupations) being compa~. B.2 8ourcaa and Validity of Data Some examples of EMF proportional mortality studies In any scientific study, careful atttntion must be given at~ Wright et ~1., 1982; Colenum et al., 1983; to the validity and reliability of the data. Unfortunately, Appendix B B-7 and/or surveilhnce activities. However, the repo~',,ing records. Additionally, many gleefive factors are of such oommunicable or oceupational diseaR ears to olgrafivc in any hospital admission. TIn~fore, the the m or local health dclmmncnt is olin neglected population base associated with particular hospitals or by many health professionah, and the reports! cases with all the hospitals in a given sxe~ may be undefined. for some diseaRs often represent only 10 percent or Finally, hospital-g~neraUxl statistics are of~n difficult less of the actual cases occurring in th~ state. to co!le~t, in part because of the lack of computerizcd retricval systems. F~deral Agencies Sources and Limitations of Mort~ity Data. The most Federal agencies, Inch as the CDC or the National routinely collected and enmmonly used source of data Center for Health Statistics (NCHS), nmintain active for surveillance of disease has been mortality data and/or passive surveillance on · wide variety of (Ketsey, Thompson and Evans, 1986). Advantages infectious and chronic diseases and health status include the fact that these ,ht. ar~ readily accessible parameters for the various geographic areu of the U .S. and inexpensive to use. Since death certificates are required by law in the U.5. and many other countries Medical Records. Hospital or clinical records m · as weB, very nearly all deaths in these countries are frequent source of information for identifntinn or reported to the authorities and death certificates are confirnmtion of study subjects. These records can also fded. Finally, mortality data have been collected, provide information about a person's age, sex, race, tabulated. and published annually for many decades address and additional relevant medical data. and for many different countries of the world so that beth temporal and geographical trends may be studied. The information contained in medical records is highly Few other a_~_~_ bases exist which allow such variable in accessibility, accuracy and completeness. comparisons. Access to medical records, even for scientific research, has been restricted due to privacy and legal concerns in There are · number of problems and limitations many situations. Another practical problem involves associated with the use of mortality data for the many judgments that must be made in revi~-wing epidemiclogic studies or as an indicator of frequency of and abstract·rig the data from medical records. Also disease. Some of these involve biases in mortality important in numy cas~s is the considerable variation in reporting which are dependent on the particular the definition and diagnosis of certain discues. While disease. For example, the mortality ran for diseases some diseases can be routinely diagnosed with a higher which have high case fatality rates are more likely to degree of certainty (e.g., acute leuketnia), others are a~curately represent the status for that disease than the not readily or consistently diagnosed (e.g., Alzheimer's mortality rates for · disease that less oRen results in discue). Diagnostic criteria can vary according to the death. In addition, those diseases which are easily physician, the hospital, and the geographic region. diagnosed may be over-represented on the death certificates relative to diseases for which the diagnosis Both diagnostic criteria and acumen for a particular is more difficuR. disease can vary significantly over time. One common procedure for achieving some degree of standardization The cause of death may be inaccurately :~ported on the is to identify and code diseases according to the death certificate for · number of r~sons, including diagnostic classifications found in the International unfamillarity of the physician with the past medical Classification of Diseases (ICD). Since major revisions history of the case. Causes of death which bear some of the ICD occur about every ten years, it is important social stignm, such as suicide, sexually transmitted to speci~J which r~vision is followed (the 9th revision diseazes, or ·bert·on-related d~aths, may be of the ICD was published in 19~g, the 10th in 19~0). significantly under-repof,~d on the de~ ~ertificates, When medical records are utilized in a study, Fstilur~ to l~rform an autopsy nmy resuR in totally diagnostic criteria should be explicitly established at the missing the underlying cause of de~h. Revisions in the study outset, and these criteria should be clearly International Classification of Diseases ICD codes or n~orted. clmnges in the way diseases ire grouped together may lead to apparent sudden changes in the mortality rate Although hospital roedial records or hospital disease for · particuhtr cam of death code. Onionally, the indexes oecasinnally provide valuable sources of cause of death is recorded correctly but is then disease morbidity data, information based upon hospital misooded on the certificate. 5orecrimes, a diagnostic admissions is likely to contain significant biases. Many 'label" is more heavily used in one place than in diseases require no medical intervention or the another or diagnostic "fads" may occur which result in individuats are tryted by physici~,ns on an outpatient the disproportionate assignment of death to a few basis. Consequently, these cases will not appear in any particular disease code·. incidence or prevalence clara derived from hospital Appendix B B-9 provid~xi by sample surveys, known as morbidity overall findings arc not necessarily inval~d-)_-~4. A~ surveys. Such surveys may consist of a single, cross- long as the exposed group hu on ·vetage · seotional cuminslion or of long-~crm Oongitudinsl) signifr. antly higher exposure level than the non- studies in which rcspondcnts arc rcvisital periodically. cxposal group, · comparison can still bc made and · Although the most ambitious of these, the National correct conclusion re·chat cvcn though the precise Health Survey, coUccts data on sit discs·u, specific cxposure*rcsportsc effect will bc somewhat in error. discue states may bc survcyal. The NCI has conductal cancer surveys which have yicldal basic In general, occasional and random misclassification is data on the incidence of ca·cot by primary site, likely to reduce the strength of, but not eliminate or histologic type, and state of discue at diagnosis. Both reverse, a significant association between exposure and lcukcndas and brain caneors ·re included in these disease. Frequent or systematic misclassitication, on surveys. the othcr hand, can obscure or cvcn changc thc dLrcction of an association, or create an association Data on personal facton, exposures, and cvcn disc·so where none exists. status arc often obtained from in-person or tolcphonc interviews of study subjects or their relatives. The Sound cpidaniologic rcscaroh strives to identify, validity and reliability of these a_n~,_~ ·ro affcctat by uscss, and reduce the inevitable inacouracies in data many factors such as the training and cxporicnco of the collection. It is particularly important that frequent or intorvicwcr(s), the length of the interview, and rcccncy systematic errors bc rccognizal. This is accomplished of the events questioned. In short, the quality of data through the usual scientific means of providing detailed obtained from interviews depends on how carefully the descriptions of data sources, observing rigorous intorvicws ·ro designed and administer·i, as well as on mcthodologic standards, and confirming data whenever the type of information being sought. The information possible. obtainal from interviews is frequently co·tinned by other data ,ourc~ (e.g., roedial records) to improve B.3 Comparability and Bias its validity. A serious threat to the validity of any cpidcmiologic Mailal questionnaires ~rc another source of data on study is the possibility that its subject selection and data demographic letors, exposures and health status. As arc biased. In cpidcmioloffy, *bias' does not imply any with intcrviows, the quality of questionnaire d~8 is prejudice or prcjudgmcnt on the part of the study in~ucncai by many factors including length of the investigators. Rather, *bias' generally rofcrs to quest·otto·ire, specific wording of the questions, types systomatic errors, i.e., crron other than sampling of questions, motivation of the study subjects, and the variability that prcvcut the truc value of · disease rate porccived importance of-the study. Questionnaire data or other production variable from being obtained. The arc also fraluattly confirmed (vai~dn~*"d) by other data introduction of bias renders study groups sources such as medical records. no·comparable in some important way. bnpac,* of Daza Errors To understand how · study can become bias·i, it is nscful to recall the dcgrcc to which the laboratory As indicated in the previous discussion, the data scientist strives to achieve comparability between sources usa· by cl~idcmiologists arc each cimractcrizcd exposed and uncxpo sed organisms in his/her by potaltial inaccuracies. These inaccuracies arc experiments. This is accomplished by such means as usually wcli-rccognizai by cpidamiologists, and efforts using · single strain of test organism, randomly arc made to limit and/or quantify their extent and assigning each orgsrdsm to an exposure group, magnitude. Inevitably, however, some error remains maintaining uniform environmental and dicUtry and must bc examined in torms of its impact on overall conditions during the course of the study, and using · study validity. consistent protocol for examination. The examination is porformed with the investigator *blinded* to the Epidcmiologic studies usually involve the comparison subjcct's previous exposure history. F·iluro to achieve of large groups of subjects, often hundreds or any of these major comparability clement· can bias a thousands of individuals. Study subjecU arc generally study, and its conclusions must bc considered suspect. classifial according to whether or not they arc exposed to the ·gaat in question, as well as according to their In cpidcmiologic roscarch, utilizing an observational discue status. An crmr in nsscssing an individua!'s methodology, an equivalent dcgrcc of comparability dcgrcc of cxposuro may affect his relative ranking in cannot bc achieved. The goal of the cpidcmiologist is the exposed group, but will not necessarily remove him to select from an existing population cxposod and non- from the 'cxposa!* category. Even if an occasional cxposai (or discasal and non-discs~) groups that arc individual is comptctcly raise·ass·fled on exposure, the Appendix B B- 11 exaggerate) exposures experienced during pregnancy, study 's findings. Strict adherence to established comptred to the mother or a healthy child. This is · procedures and stsmdaxds can reduce many, but not all, serious potontial basis in c~e-~onttol studies. possibilities for bias. !n many well-report~ studies, the authors frequently discuss the possible sources of Miselassi~ation Bias. This does not refer to random bias in their study and attempt to show, through logic misclassi~cation of some small proportion of the study sad/or a_._,_., that their findings ire not likely to be due population, but to systema~ ntisclassification of to soma bias. But there still remains the possibility that disotse or exposure status. This could happen if, for soma unknown bias is operating. example, the job title 'electt~ian" is used to identify .ubj~ct. o~up~o~Uy expu.ed to nMF, ~d · U~e B.4 Association or Causation7 portion of these ele~ricians wntk only on dead circuits, where EMF exposure is low. On the contrary, some In laboratory research, a well-designed experiment that other wori~rs, who are identified as "non"electricians", results in · statistically significant effect (i.e., one not might be exposed to EMF. due to clutnco vaxiation) is usually interpr~___~_ as demon·rating a cause-and"effect relationship. The Berksonisn Bias. This bias, first described by Berksen existence of · cause-and-effect relationship cannot be in 1946, can axig when controls ate selected from so readily inferred from hospittls (usually the same hospitals from which the cases have bocn selected). Hospitalized controls ·re observational epidemiologic studies. The generally not representative of the overall population =pidemiologist, at best, can show that some association with respect to certain chantct~risties; e.g., smoking, or relationship exists botw~en an exlx>sure (e.g., · lcohol or eoffce consumption. In order to detect and chemical, radiation) and a physiological or health- prevent Berksonisn bias, multiple controls ·re related effect (e.g., blood chemistry, disease, death)- suggested. Therefore, some ~ntrol studies have Typically, the epidemiologist further attempts to used both hospital and neighborhood controls at the demonstrate that the association is unlikely to be due to same time to delete this bias. chance and is not due to some third (confounding) variable. Confounding Bias. Confounding bias occur· when there is a third variable which is not of interest to the Judging Positive Asao~iations study, but it is related both to the exposure and the disease under study. A study of lung cancer, for Scientific "proof~ of · cause-and-effect relationship example, will find · significant association between eatmot be obtained from an observational study. alcohol consumption and the development of lung However, as a practical matter, explicit or implicit cancer. In such a study smoking would be considered judgments of causality axe frequently derived from such · ennfounding variable since it is associated both with studies, and strongly influence public health policy. alcohol ennsumption and lung cancer. if the effect of Therefore, it is important to ennsider epidemiologic smoking is removed in the study Coy design strategies findings from · vnriety of perspectives, such that · such as nuttohing, or analytic t~hniques such as reasonable assessment can be made. Epidemiologists stratifiation or adjustment), the association between have not established hard sad fast roles for determining drinking sad lung aneer might not remain. when · positive association should be ennsidcred a cause-and-often reiationship. Different experts stress Most epidemiologic studies consider at least seventl different f·ctors in =vnluating associations, and not well-established confounding vatisbles such as ·go, ·$re~ that certain items axe particularly useful. sex, nee, and s0cioeeenomic class. Other variable~ However, ee~in guidelines ·tile frequently in may be examined in · particular study. Howover, not discussions of causal relationships, tnd these are all v·riables can be examined; confounding bias is dis~uss~! below. probably pr~ent in tH data. Strength of Association. The more strongly an Conclasiotu exposure is associated with · disease, the more likely it is that the exposure causes the disease. Many different Given the difficulty in rocognizing and controtling ·11 measure· of association ·re utiliz~ in epidemiologic potential sources of bias, no study shouldbee, onsidered studies, depending in part on the type of study completely free of it. For example, rarely does a study ennduaed. Most ·re expressed as some ratio satin 100% follow-up of subjoin. Critics can always enmparing or estimating disease risk in the exposed to sttribute study finding to soma form of bits sinc~ there disease risk in the unexposed. For example, the ax~ indeed so many potential sources of bias. "relative risk" maybedefined as: However, it is difficult to actually demonstrate that soma bias ·ocounts for or even materially ·fleets the Appendix B B-13 for all cues, and there was considerable doubt by some A cohort study is similar to a laboratory study in tcrrns (including the manufacturers) that the tsmpons could bc of the time sequence of cvcnts. A study group (cohort) causaily related to the discuc, However, the abscnco of healthy paoplc is identified and each individual is of an cxpla~n_..2on or rccognizcd mechanism did not thcn classified according to whether he/she is cxposc<l prcvcnt withdrawal of the tampon· from the market, or not exposed to the agent under study, At some later Sornc time later an explanation was found: the discnsc paint in time, which may ho many years later, the wss satally caused by a toxin from a relatively cohort is rc-~h~:ked (the "follow-up")to identify those common bacterium. The toxin is only produced under study participants how have occurred when the study is certain physical and biological conditions, conditions initiated depending on the study de, sign. Two risk which arc more likely to occur in young women using rncuurcs from cohort studies will bc dcscril~d: the high-absorbency tamp·n·. relative risk and the standardized mortality ratio. Spccificity. V,'hen an association links exposure to · Relative Risk. From the follow-up data, an actual single discue rather than to a broad spectrum of discnsc rate can ho tabulated solstately for both the disc·sea, at causal interpretation is favored. An cxpascd and non-exposed groups. Depending on the example of high spccificity is the usociation between specific disease in question, the rate may bc either a occupational exposures to vinyl chloride and mofoidity or a mortality rato. As an example, consider angiosarcoma (a rare form of liver cancer). The high that the exposure under study is cigarette smoking and spccificity, as well as the stnmgth of this association the discue in question is lung cancer. Data from an leaves little doubt as to its cans·tire naturo. actual study showed that the lung cancer mortality rate for · particular ·gc-group of non-smokers was 19 A lack of spcci~city, however, does not necessarily deaths per 100,000 par year. In contrast, the rate for argue against causality. For example, cigarette smokers in this same ·gc-group was approximately 190 smoking has been usociatcd with · wide range of per 100,000 per year. diseases. In fact, the smoking history of study subjects is almost always considered in well-designed studies of These rates can bc comparod in several ways such that other diseases. This lack of spccificity, ·lthough still a qmmtiftcd expression of risk can 5c obtained. The somctilnca raisod in arguments by the tobacco industry, most common measure of risk in this type of study is is not particularly troublesome to cpidcmiologists since the rdativt risk (RR). The relative risk indicates the a great numy compancnts have hocn idcotificd in increased (or decreased) dcgrcc of risk of disease tobacco smoke, and many of these compancnts can ho among the cxpasod compared to the non-cxpascd. It transpartcd through the body to diffcrcnt sites. This provides a mcuurc of the cans·rive impart·nee of the rather broad excel·ion to the concclx of spccifacity has exposure uudcr study. A relative risk with · value of led some cpidcrniologists to conskier this gnidclinc one (1.0) indicates no usociation between the exposure useIns in determining whether an usociation is likely and the disc·so. to bc causal or not. The relative risk is calculated u follows: Conclus~oas Relative Risk (RR) = rate in the exposed It is impartant to crnphasizc that none of the above rate in the non-exposed facton is sufficient either to prove or disprove that an association represents · truc cause-and-effect For smokers, the relative risk is: relationship. They do, however, offer some reasonable guidelines MAth which both cpidcmiologists and non- RR = ]~90/lO0000/vcar = 190 = 10 cpidcmiologists may judge whether · positive 19/100000/year 19 usoclation is likely to represent ·tnac cause-and-effect relationship. This indicates that smokers have ten times the risk of dying from lung cancer coral·rod with non-smokers. B.5 Statistics: Risk Estimates Confidence limits can (and should) bc enmputcd for The following discussion adds how rates and relative risks to determine if the risk is statistically ratios arc used to provide cstimatcs of ri~, with significant, that is, not duc to chartco. Confidence emphasis on the mcuurcs of risk derived fTom two limits arc the range of the risk estimate which takes into major study designs: the cohort study and the ca·c- account sample size and variability. If the range oft tic estimate does not include 1.0, it is recognized as being control study. statistically significant. Risk Measures from · Cohort Study Appendix B B-15 fh~t bir~ and parity arc themselves related and that one The odds-rstio is thcn calculstcd u: i~ · confounder. hstead of ·singlc 2x2 table, thc investigator might construa a sori~ of 2x2 tables OR = 85 x 60 = 8.5 showing the association of breast cancer and age at first 15 x 40 birth for different parity levels (c.g., one child, two children, and dxree or more children). In other words, The odds-ritio will be · reasormble estimate of the the investiggor 'stntifies" on the variable parity. The relstivc risk if the discue in question is relgivcly rar~ investigator might also look ~t breast cancer and parity. (e.g., cancon), the exposure is reiativcly common, str~fying on age st first birth (e.g., under 30 years, and, of courac. there arc no serious gudy biases. and 30 yean and okler). The investigator would then find that the !stc age st first birth shows an increased The odds-ritio is interpreu:d exactly the same as the risk ratio regardleas of parity level. Stratifying on age relative risk. If equal to one (1.0), i~ suggests no st tint birth, the invcstigstor would find that parity is usociatlon benveen the cxposure and disease. If no longer associated with breast cancer risk. greater than one, i~ indicates · positive association, and Strgifying hu thus eliminated the confounding of if less than one, · negative usociation or · proU~ivc parity and age st first birth by examining each variable effect. As with relative risks, confidence limits can be scparitely. computed to determine if the odds-ritio is significantly diffcrent fromavalueofonc. Where several 2x2 tables and risk ratios are constructed in an analysis, a summary odds ratio can In some case-control studies, the controls are still be calculstcd. The usual technkluc for calcuhting individually "matched" to the cascs during the selection · summary odds ratio is the Mantel- Haenszcl proceu. For each case identified, · systematic proccdure. This summary odds ratio can be thought of approach is used to sclea · control who is in the sam as a weighted avcragn of the individual odds ratios age bracket, of the same sex and rico. etc. This from the sclarite 2x2 tables. Actual techniques for maWhing process avoids having to account for these calculating the summary odds ratio and its con~dencc variables hter in the analysis. When this type of limits an be found in basic cpldeminlogy or biostatistics matching is used in · study, the odds-ritlo is calcuhted texts. differently than above. Mathematical Modcfing. Until relatively recently, .4dualyac CostrolofCoaUrousders stratification was thc main approach to analysis in cpldemiologic research. The limitation, however, is As previously described, cpidcmiologic studies involve that if more than · few variables are controlled for s/ examination of mamy different variables in addition to unc tlmc, too few subjects fall into each stratum, and exposure and disease since there are other facton that the ruulting risk esthnates become unreliable. In this may be relaU:d to the exposure and/or disease under situation, the addition or deletion of even one subject in study. The challengn then is to tease out the effect of · stratum could change the risk estimate dramatically. the exposure of interest in the presence of these other Mghemstical modering overcomes this limitation. factors or co-variates. In pric~cc, analysis and Many different msthcmatical modcts are ·v·iiabic interpretation of epidcmiologic studks involves much today, some of which are available on personal more than · single 2x2 table. If confounding variables computers. The most commonly used modcl is logistic are not controlled for in the design of · study, Ihen they regression. AJthough developed originally for use with must be controlled for in the analysis. Two appro·ches cohort studies, logistic regression is commonly for the analytic control of co-variatcs or confounding cmpioyed in casc-control studies. Discussion of the are stratification and mathemgical modeling. assumptions. use. and limitations of loptic regression and other modcts can bc found in most recent Siratificatlon. If the number of variabks to be epidemiology and statistics texts. controlled for is no~ too hrgc, and the range of values for these variables is not too wide. them stratification Am-/bu. ab/e R/.~k Measures provides · simple and powerful analytic technique. in stri~ficstion, the investigstor looks st the relstinnship Simple Attributable Risk. Disease n!~ can also be bctwecn exposure and disoasc among subseU of the compared in another fashion: by looking st the study population which have beea categorized difference in the rates between exposed and non- according to the lovcl of some other co-variate. In · exposed group. This musure is referred to as · study of breast cancer, for example, it may be found simple aRributable risk. The simple allributable risk is that · hte age at first birth has · positive usocistion used to quantify the risk of disease in the exposed with breut cancer. and also low parity has · positive group that can be considered stUlbuted to the exposure. association. The investigation may suspect that age st Using thc smoking rlatl presented above, Apptnd/x B B-17 Significance testing Consicier · cis~ntrol macly which is investigating whether the proportion of tho.e with a specific The vast majority of published epkicmiologic studies cxposuro differs significantly hawcon thc case group have statanents ckx:laring that the differ~nc.~ obxrv~d ~nd th~ control group. The common stat~tical method between the compari~n groups are (or are not) for comparing proportions (or _!gycentagel) is the X2 2 'VafistieAHy aignificant.' Unfommately, much (chi Kluaro) tut. To apply the X test, it is helpful to confusion exists, particularly in the non-4cicatifw arrange the data in the form of · 2x2 table. To community, about the meaning of · 'statistically demonstrate this, assume that the study involved 100 signifwant' result. Such · finding is one that, cases and 125 controls. Thirty-five (35) of the cases according to certain assumptions and based on · wcrg eatposed to the agent, while 25 of the control] mathematical probability, has · low likelihood of being wag exposed. The table would then be constructai as due to random sampling variation (chance). follows: The probability of this differgncc being duo to chance is expressed as · 'p-value' and may be given in either aired 3s 2s 60 decimal or pcrccntagc form. its magnitude will depend SlabSmsd ss t00 tss on · number of factors including the size of the Tins zoo s2s ~s diffcrgnce observed, the size of the groups being compared, and sometimes, the data variability. The X2 tnst statistic is defined u: In practicc, the most common p-value used to hbel · finding u statistically significant is p <.05. If · (O.E)2 difference in · partiouhr study is large enough that X2 = there is less than ·fivc percent chance that it is the E ruult of random variation, then this finding is considcrai significant. Another commonly cited where O = observed number and E = expected thruhold of significance is p < .01, less than ·onc percent chance that a'diffcrcnce is duc to random number, and the summation is over all cells of the table. Usually, a correction for variation. continuity is made by subtracting 0.5 from IO-EI as shown below: In some published reports, an exact p-value (e.g., p= .03) is provided, while in others, the iovcstigator only states that 'p < .05,' or that the findings aro significant at the .05 level. Exact values arc, of course, moro iofonnativc and allow the reader to apply his or ([O'El'0'5)2 her own standard of statistical significance. X2 = E The level of significance selected depends on how strongly the investigator wishes to avoid the error of To calculate the X2 statistic, the 'expected' number for concluding that there is · real diffcrgncc when none each cell in the table must be found. This is done by may actually exist. This is referrod to u · 'Type r usuming that there is no usociation betwcgn exposure error. The smaller the p-value, the smaller the chance and disease, and making use of the marginal totals in the able. The cxpectal value can then be caiculatai that · Type I error cxim. with the following qualion: Calculation of p-value E = (tow totsl)xienlumn total) There arc · number of procalurgs for calculating p- values depending on whether one is considering, for example, the difference between the means of · To illustrato, consider the uppcr-leit cell of the 2x2 continuous variable (e.g., age, years of employment) or table shown above. The row table is 60, the column · ca~cgnrical variable, such as the diffcrencc between total is 100, and the grand total is 225. Thus the the pwportions of those in each group who have some expected number of exposed cases is E = (60 x characteristic (e.g., rgsidcnce near high-cu.~,t 100)/'225 = 26.7. This calculation can be applied to distribution lines). These formulae and their each of thc other coils of the able to obtain the other applications can be found in great detail in any statistics expected values. The expected (E) numbas thus toxt. Only one common example is provided below. obtained arc shown in parentheses below: Appendix B B- 19 Standard Error (SP-) = vtriancc = 0.09396 = 0.3065 Study Conclusion n. True bintlonship Association No Auocialion The logarithm of the lower and upper 95 percent ~ I Type I confidcncc limits is thcn: Association Correct No Association Type II En~r ~ Correct in LL = in OR - (1.96 x SE) = 0.1B77 in UL = in OR + (1.96 x SE): 1.3892 A rchtiormhip exists between Typc I and Type II caTon in that thc morc closcly onc 2uards against a Typc I Convcrting from natural logs of thcsc cxponc~tial crror, thc motc likcly a Type 11 crx~r will occur. In valucs to obtain thc acttml inwcr and uppcr limits can othcr words, thc grcatcr thc diffcrcncc requircd for a casily bc doric on most scicntific calculators: conclusion of a real association, thc grcatcr thc chancc LL = c0'1877 - 1.21 The probability that a rcal association will bo dctectcd UL = c1.3892 = 4.01 in a study is rofcrxcd to as thc ~t~wer of the study. It is dctcrmincd by subtracting thc probability of a Type II Thus, the 95 pcrccnt confidence interval for thc odds caror, cxprcsscd as BCocta), from 1.0. Although not as ratio is 1.2-4.0. This interval shoM scvcral things. commonly discussod in study reports, thc powcr of a Onc is that thc obscrvcd odds ratio (2.2) is not a highly study to dctccta particular lcvcl of risk is cxtrcmcly prccisc cstimato. The actual risk could rangc from a iniportant both in thc planning of a study and in the vcry weak 1.2 to a modcratcly high 4.0. All wc know cvaluation of "ncgativc" studics. Most rcccnt is that thctc is a 95 Porocnt chance that the real risk cpidemioloSic and statistical tom providc dctailcd ratio falls within this intcrval. discussions of Power and Type H CaTOrs for various types of studies. Rather than roylow these tacthods for Mother important aspect of this intorval is that it does calculating powcr or sample size, this section will only not includc onc (1.0). This is cxpcctod sincc the briefly outline the major factors that dctcrminc powcr significance test (X2 tcst) had alrcady rcvcalcd that an and then show one cxamplc. association cxists 5ctwccn thc cxposurc and discasc. Four factors affcct the probability of a Typc II error Many cpidctniologists boiicvc that a confidcnce intcrval and, therefore, thc powcr. Thc first is the lcvcl convcys more information than a significance tcst confidcncc uscd to rcducc thc chance of a Typc I error. bocausc it indicatcs thc inwcst and highcst likcly truc If onc dcmands an cxtrcmcly high lcvci of con~dcoce rclativc risk (or other paramctcr). It also rcvcals (Lc., avcry small p-valuc), a 2rcatcr chance cxists that something about thc precision of an cstimatc. An atruc association will not bo rcco2nizcd. cxtrcmcly widc confidence interval, whcthcr or not it includes a risk ratio of 1.0, su22csts caution in Anothcr factor that affccts the powcr of a casc-control intorprcting thc rcsults of a study. study is thc prcvalcnce of thc cxposurc ·toorig the controls. Power is rcduccd whcrc the cxposurc is cithcr vcry rarc or vcry common among the controls. in a cohort or cross-sectional study, powcr is mainly 8.7 Statistics: Type II Errors and depcndcnt on the "cxpectcd" numbcr of discatcd cases in the uncxposcd population, which is in turn rclatcd to Power samplc sizc. (Samplc size is discusscd boinw.) Thc discussion in the previous scction only considered Obviously a study has a 2rcatcr probability (powcr) to toclmiqucs to avoid a mistaken conclusion rc2arding thc dctcct large rclativc risks (or largc diffcrcnccs in means prcacncc of an association (Typc I crror). Lcss or proportions) than small risks. A particular study frequcntly discussed is another typc of error that may may havca 90% chancc of dctccting ·rclativc risk of bc cqually important: typc II error is thc chancc of 3, but leas than a 2~% chancc of d~_--r_~ing a risk of producing a conclusion of no association whcn onc 1.5. actually docs exist. Just as chance can opcnttc to produce an aplarcnt association whcn ncnc actually Another factor that affccts thc power of a study is thc cxists, chancc can also obscure a diffcrcnce that docs samplc sizc of thc study -thc numbor of cascs and cxist. Failing to dotoct atruc association is rcfcrrcd to controls in a casc-control study, or thc numbor of as a Type Hcrror. The possible combinations of pcoplc and poriod of follow-up (pcrson-ycars) in a correct and incorrect conclusions arc: cohort study. Obviously, thc larger thc study, the Appendix B B-21 References Thc following work, wcrc used in prcpiring this primer. Thc reader is rcfcn~ to lhcm for further information on thc ,ubject of epidcmiologic methodology. Allnun DG, Gor~ AM, Gardncr M!, Poc~zk ~I. 5tatiatic. tl guidelimm for contributon to medical joumaln. Br/tish Medical Journal 1983; 286: 1489-93. Breslow NE, Day NE. $tatistical Mtthod~ in Cancer Research. Vot I: The Analysis of Cast- Control Studies. IARC Scientific Publication 32. New York: Oxford Univeraity Plea, 1980. Buechley R. Dunn JP-, Linden G, Bnslow L. Doth certificate statement of occupation: Its usefulne~ in comparing mortalities. Public Health Reports 1956; 71(11): 1105.11. Cole P. The evolving caze,-eontrol ~tudy. Journal of Chronic Diseases 1979; 32:15-27. Feinstein AR,Horwitz RI. Double ~ndard,, ,cientific methods, and epidemiologic ruearch. New England Journal of Medicine 1982; 307(26):1611-17. Fleiaa JL. $tatistical Methods for Rates and Proportions, 2nd edition. New York: John W'dey and Sona, 1981. Friedman GD. Planer ofF. pidemioloSy, 2nd edition. New York: McGntw-Hill, 1980. ISBN 0-70-022434-X. (;laden B, Rogtn WJ. Misclasstftcation and the design of experimental studies. American Journal of Epidemiology 1979; 109(5):607-16. Glasser JH. The quality and utility of death certificate data (editorial). American Journal of Public Health 1981; 71(3):231-33. (3r~enberg R5, Kleinbaum DO. Mathematical modeling Rratcgies for the analyai~ of epidemiologic research. Annual Review of public Health 1985; 6:22345. Hain~ T, Shannon H. Sample size in occupational mortality. Journal of Occupatioual Medicine 1983; 25(8):603-08. Hill AB. The e~vironment and digtse: kssociation or eaumion? Proceedings oftht Royal Society of Mtdicint 1965; 58:295.300. Horwitz RI, Feintrein AR. Metholodologic mandards and contradictory reauit~ in ease-control research. American Journal of Medicine 1979; 66:556-64. Kelaey $L, Thompson WD, EvaIra AS. Methods in Observational F.p/demio/ogy. New York: Oxford Univenity 1986. MacMthon B, Pugh TF. Epideaniology: Princ/p/ts and Methods. Bo,ton: Little, Brown and Cornpiny, 1970. Mantel N, Haenszel W. Statistical apeco of the analysis of din from retrospective studies of diseue. Journal of the National Cancer Institute 1959; 22(4):71948. McMichael, AJ. 5umdardized motUlity ratios and th~ healthy worker effect: Scratching beneath the surface. Journal of Occupational Medicine 1976; 18: 165-68. Monson RR. Occapationa/F.p/demio/ogy. Boca Raton: CRC Pre~s, 1980. C-1 APPENDIX C - RESULTS OF EMF SURVEY C. 1 Siting 1. Allegheny Power System Walkeravilie Loop 138-kV Line Siting approved. Company) 2. Appahch/an Power Co. Collod~n-Gsvin 765-kV Line Court found that there sre no known edvene biological eft·ca euociated with the fields from this 765-kV line. 3. Arizooa Public $srvice Almdssa 69-kV $ubslatioa Site wu approvsd by both Scoosdale Ranning ~,_~ Board snd ScoOsdale City Council. 4. Arizona Public brvics SDGaD 5(X}-kV ina~oonsctioa No direct ruling oa EkIF is but Cornhalite· granlsd Cerli~ca2 of Envlromnsmal Compstibility. 5. Associated Electric Coop. Miuoori, Iowa, Nebrub Project received 'Finding of No Significant Transmission Project (MINT), 345- Impact' from licensing authority; EMF issues kV Line (scheduled completion were raised by lendowners in public hearings; 1992) mbsequemly scquired 94% of ROW through nego6ation, remaining 6~g thmogh conde·urn·ion; EMF not en is in co·de·urn·ion hearings. 6. Alia·tic Ek~tric Co. Application to Coaslruct and Landowners ralgd/asus of EMF health effects; Opec·,* · 230-kV Transmission AEC submiucd rcpom on EMF health effects; Lies and Rsleted Sobsratio· hndowmsrs di~,e.d EMF claims and the 230 Fecilitks (1984) kV line wu spproved when · new mute was · arced upon. 7. Allstoic Electric Co. Mickleton 230-kV Line Line rapproved; Board found Ihat fields from this line would be similar to fields from existing lines of same voltag~ class end that field levels from this llns fell witldn field standards in New Jersey snd other states. 8. Allstoic Elsctric Co. Millville~4surice Pjver 138-kV Line approved; Board fonnd that the line 'will Lies not unduly effect the quality of the environment or the health of the public.' 9. Alia·tic EkcUic Co. Cardiff-Nsw Fnssdom 230-kV Lies EMF issus raised; no ,peci~c findings re EMF; 10. Csntral el. Pub. Service Ceni~cstlon for 138-kV Line Site approved and squest that utility provide indenud~cstlon dezded. l 1. Central Main Power State Board of Envlronn~ntel Utility 6led writ2· tqx~rt on EMF issue; line Company Pro2ctioa was built. 12. Csmml Powgr 4~c lj_~'he Wdicsr-Mstagorda 400-kV Lies, Commission directed utility to provide more Company No. 5023 infornmtion on possible health cffecu of DC ionsi utility withdrew the application. Appendix C - Results of EMF Survey C-3 Siting (Continucd) No. ~ Case/Project Descdp6oe Status 7,6. Jersey Cenual Poser & Manltou-Whltlng 230-kV Line EMF issues raised; llne approved; opposition L~t Co. toodon to my pending appeal denied April, 1990. 27. Jeresy Cemral Power & Aberdeen-Red Bank Appliesdon withdrawn by utility after Lipt Co. r~R~um~nt of need for line. 28. Lowor Colorsdo River Appliestiou for 345-kV Commiuiou sdopted !testing Examiner's ruling Amhodty Tranmmiseion L~ae and Auocieted mat mere were no proven healm effecu and Subgallon (I 979) added provision that the commiuion could amend or mvok~ certi~cat~ if future reuarch shows that exposure to electric fields causes adverse heslth effecuo 29. Minamma Power Co. Applic4fion for Exemption From Applieslion approved; Board found that Siting Requirements for !~aluth propond upgtmis wooid remit in lower Area magnetic field levels and that there is definitive evidence* flint EMF esuum health problems. 30. Minnkote Power Coop. Center-Maple River 230-kV to 345- EgdF wirecues appeared; Application granted. Ira:. kV Upgrade (1980) 31. Miulselppi Power Application for CPCN: U-4128 Comnfiseion found no evider~e of health effects Company from 500-kV line. 32. Mornaria Power Company Application for Certification of State agency found field levels acceptable. Laurei-BrldSer 100 -kV Line 33. Nuhvil~ F. Jocuic ~rvic~ Sharondale 161-kV Submarion Project has been placed on hold pending results of independent engineering mudy of syaem-wide electricity needs. 34. Nuhville Elm:tr~ S~rvi~e Headereonville 161-kV Substation Line approved; conmmctlon mbm~iuently oppoud by Sierra Club. Utility nm with lo~al citizens group and prupoud engineering alternatives to delay conSuction of project; alternatives adop~d and proja:t delayed indefinitely. 35° Nebraska Public Power Application for Certification of Application granted. Utility decided no~ to build District Corridor Compatibility: No. 9942 line. 36. Nebraaka Public Power Application for Permit to Construct Appliesfion granted. Utility decided nc~ to build Dialriot and Op~rete: No. F-3371 line. 37. Nevada Power Company Warm Springs Eastern Line EJ~F reised; witherues appeared; approval granted; appealed to Coomy Planning Commission; lost approval for certification; utility rcmuted line. 38. New England ElecUic Petition for Perminion to Build Department of Public Utilities found that health System 345-kV Line, No. DPU 19559 hazards were conjecture and had not been esublished. Supreme Judicial Court affirmed this finding and remanded the case on other iuues. Appendix C - Results of EMF Survey C-5 Siting (Continual) No. ~ CmJProjtct Dfscription Stalin 5 1. Puget Sound Po~er & Canadian T'~-in Pending. Light Company 252. ,qall River Project Pinnacle Peak-Papago Buttes 230- (2erfificate denied at least in part on basis of kV line. potential for EMF health effects. 53. San Diego Gas & Electric Application for Certification: Commission found that available information did Co. De~isioe #93785 not indicate that EMF causes advene he41th effects. Ordered utility to continue to fund EPRI studies and inform Commission of remits. 54. Soulhem Cal. Edison Co. Kremgr-Vi~tor 230-kV Line CPUC grained approval to construct ramsmission line; required utility to "minimize" apetic fields associated with the line. 55. Tucson Bgtric Power Bul Air Ranch Fates v. TF_.PC; Court feared that phimiff bad failed to prove Company No. C458986 adverse he41th effects by ·preponderance of the evidence. 56. Union Electric Apache Flau 161-kV Line Line approved; decision made no reference to EMF. 57. United Power Association Bumo-Milaca 230-kV Line: Docket Board found that evidence to date did not No. UPA-TR-I indicate tranmniuion lines as a health hazard. 58. United Power Association Coal Creek-$tsmon 230-kV: Docket Commission found that designated route would No. 9593 have acceptable health and environmental impacts. 59. United Power As~ciation W'dnmrth 345-kV Line: Docket No. Council found no substantial showing of advene CU-TR-2 biological effects. 60. United Power Association Certificate of Site Compatibility; Commission found no evidence ofadvene D~ket 9459 health effects. 61. United Power Aaociation Route Permit 400.kV IM2 Line: Commission found no detriment to human health Ca0e No. 9370 from the line. 62. United Power Aaociatioa Consreaction Permit 400.kV DC Council found the line would have no adverse Line: Docket No. CU-FR-I health effects. 63. United Power Association Corridor Designation: W'dmarth Council found no repom of advene health 345-kV Liner No. CU-TCI effects from EMF. 64. United Power Auociafion Corridor Designation: 400.kV DC, Council found that propo0ed line would have no No. CU-TC-I advene health effects. 65. Vermont Eiectric Power 450-kV DC !metconnection. 4763 Board found that potential for adverse health Company, effecte was minimal or nonexistem. 66. Virginia Power Laudoun- Gain~vifie 230-kV Line Slate Cotpore~on Commition decided to monitor the EMF health effects i,sue. 67. Virginia Power Ox-Pnamm Plant 500-kV Line Slate Coalmarion Commission decided to monitor the !~dF health effects issue. 68. Vitlinia Power Co. glmont-C'hickahominy 230-kV Line EMF experts appeared; Hearing examiner found (1990) no EMF health effectsi line approved. Appendix C - Results of EMF Survey C-7 Zoning (Continued) No. U~_ c--drmjm net,~,-~- stat~ 12. Pacific Power & Light China Hat40o~t loop 69/115-kV Public he~ring re line including EMF and propeay Company Line value issues. Hearing Officer found that causal relatiomhip b~tween EMF and health had not been demonmated, and ,hat transmission ~ would have mlnh~t~ if any impact on prop~y valuea. Line approvmi in late 194)0. 13. PI~PCO Brighton-High Ridge 500-kV ~ Howani and Montgomery County Boards are pre~mpte, d from mcuin~ EMF Mandards once State PSC has authorized lint. 14. Public Servic~ Gas & Eagle Point 230-kV Fttd to N.U.G. Heating b~fore WeMville Enviromnenul Planning F, kcttic Board; EMF witntuas; utility modified route; zoning approv~d. 15. Public Service Co. of New ! 15-kV ~ in Santa Fe EMF raim~l; approval pending. Mexico 16. $ierre Pacific Power Camon City 120-kV Lint Approved. EMF testimony prestnted. No mtmion C,,,,~,a,,y of EMF in Final Ordcr. 17. Tti-,~ae G&.T Newton, Joltneon & Kasnyski v. Utility withdrew application attar District Court Grand Co. Commis. & MIn. Pmt~ ordered new examination of the BiF issue. Utility Electric inc.t No. 86-CV-225 built distribution lin~. C.3 Condemnation No. ~ c,,~ No"'- Status !. Al~gh~ny Pov~r $yatm Ali~ny v. h~u~ (138-k~ ~F mi~ by ~w~r in co~e~on ~c ~;~- Co.) w;~ p~i~i ca~ ~. 2. ~ ~i ~m ~ v. ~d; No. C~ No ~e~; u~lity p~e~ bdef~mny (5~ k~ on ~c ~F ira; ju~ did c~m~n for fur of 3. a ~ ~e ~ v. ~vc bats C~a held it ~ny ~bvaa m a c~eafion action · e~f~ ~d~uilc. 4. ~t ~ C~m~c BB~ v. C. A. Co~n 13~kV ~ co~veaicl ~v~ de~y~ ~B~ Tern ~ H~ lu~ 1~2 m ~lve di~ove~ i~s. 5. B~ ~c ~er Bm E~ v. ~cl~ by Utility offend $42,~ for 4.2 acgs. ~ ~m~ p~y. ~u~ aw~ $~,~ for 4.2 ac~s. No ~on of ~F ~n~m, ~. ~ R.O.W. awa~ h~r $1 l,~. ~ ~, lu~ aw~ ~r $10,~. No ~on of ~F ~ ~ 7. B~s Ei~ Po~r BE~. v. Mc~um 138-kV EMF ~su~ ~; Com~sioncr's C~vc, ~c. ~PEC) L~c R~-W ~uR a~ ~do~cr $6,~. On a~l, ju~ a~ ~do~cr $3,3~. No mcn~on of EMF ~ F~I O~cr. ,4pptndix C - Results of EMF Survey C-9 Condannation (~ontinued) No. Utilit,/ CaN Nams Slims 7,6. Indiana & Michigan IB~ctric Ik24EC v. Pounds, No recovery; coon refuMd to decide Compan~ No. C-790235 C/65-kV Line) EMF issue. 27. Int~rmat~ Powff Co. 0owe) Comknmation Action 345-kV Rock Landowner raised adverg health effects; Crgk Tansmission Lira (1985) hearing b~fore iowa Board of Public Utilities; EMF experts appeared on both sid~s; ca~ was settled and volumary mem~nt gran~l b~fore opinion was issued by board. 2~. Iowa Power and Li~t Comlm~y IP&L v. So~r~nb,ck~r. ~t 01. No recovery; court excl,_,_A~__ expet% temimony chiming EidF health effects bocauu thcvg was insufficient data to ruch a conclusion that tkc ~ would cmuu health hazards. 29. LaPtats Electricity A~ociation LaPlali Elcc. Au'u v. Curemira Landowners submitlui public opinion survey (1986) re dcslre to live mr power lines duc to health hazards of EMF~ app~llali ruled that admiuion was not reversible error since no wiight given to it in trial indle', award. 30. Los Azailu D~lnmlnt of City of LAv. Fmk. LA 5uparior Jury made special finding that there was no Wmr &Power Ct. No. C307309 diminution in valu~ of property due to adverse biological effects of the 500-kV iin~. 31. Los Angslis Depamnsnt of City of LAv. Van Dooremoien, No special findings; unable to determine Watsr& Power No. C606776 whether the EMF ime affectzd award. 32. Louisiana Power & Light Louisiam Power & Light Co. v. Court of Appeal of LA held it was Mobky (1986) parmiseibl~ for trial judge to force LP&L pay for darnage to land if damag~ was caused by fear of EbiF adverse health effecli. 33. Louisiana Power & Light LP&LC v. Zeringu~, Coott did not dexide ismo of EMF health No. 53, 164=E effects but nott~l the landowners fear of such effects. 34. Louisinto Powsr & Light LP&LC v. Churchill Farms & No spscial findings; unable to deu~rmine Mar~ellow No. 184-546 whether the EMF isau~ affected award. 35-45. Lower Colorado River Authority Ek-vsn Condsnmation Actions Involved EI4F-r~lat~l iuues; three cases actually went to trial; LCRA prevailed in each; te4tlmony was liksn in each regarding EMF issues, but no EMF/health effects findings were made in the final orders. Oisre eiaht cases settled out of coua. B-1 APPENDIX B - FUNDAMENTALS OF EPIDEMIOLOGY1 "confounding" factors may lead to an incorrect B.1 Epidemiologic Methods Lnterprc,ation of the relationship bdween the agcnt and · c disease under study. Epidcmiology is generally dcfmod as the study of the distribution of disease in human populations and the Other examples of the methodology problems that can determinants of that distribution. Characteristics of alter or bias observational studies include: uncertainties people and their environment may be examined for in deterraining the actual exposure status of individuals, possible causal associations with the occurrence of variations in disease definitions and diagnoses in human disease. different geographic areas or in different hospitals, loss of study subjects who leave the area, unwillingness of Because cpidcmiology draws its conclusions from subjects to participate, and inaccuracies in frequently observations of the natural distribution of disease, it used data sources such as death certificates and clinical possesses both unique strengths and limitations. Since records. Practical solutions to many of these problems humans are the subjects of study, epidemiology avoids have been developed by epidemiologists, although the problem of extrapohting from animal experiments frequently these sources of bias are not adequately in which both the exposure conditions and the addressed. appropriateness of the animal model are oRen questioned. Although each cpideminiogic investigation poses its own unique problems and solutions, the overall On the other hand, epidemiologic research generally approach of a study generally follows one of a several provides less conclusive fmdings than laboratory basic study designs. The choice of a study design will research does. The inability of cpidemiologic research depend on many factors such as time and cost to offer direct proof of a cause-and-effect relationship limitations, frequency of the disease(s) to bc studied, results from its observational methodology. In a frequency of the exposure, intended use of the hboratory investigation of a suspected harmful agent, it information, and the availability of required data. is assumed that the animals under study differ only on Several commonly used designs are described below, the basis of their exposure regimen. Any ensuing along with a brief consideration of their particular differences that are found between exposed and non- advantages and limitations. exposed animals can then reasonably be attributed to the exposure itself. Since obvious ethical and practical The terms incidence and provalence will be used in the prohibitions on experimentation with humans exist, data following descriptions of epidcmiologic study designs. must bc collected on the "natural" occurrenen of the These two commonly used measures of disease disease and agent under study in human populations. occurrence have distinctly different meanings in However, human exposure to an agent is not a random epidemiology. In a population-based study, the phenomenon occurring among members of a prevalence of disease is the proportion of individuals in homogeneous population. Exposed and non-exposed a population with the disease at a given point in time. groups will differ in terms of age, residence, For example, the number of persons with lung cancer occupation, gender, and many other factors. Some of in a population of 100,000 on December 31, 1989 these variables arc known to influence disease might be 30. The prevalence of lung cancer in this occurrence and can be accounted for in the design or population at this point in time is 30/100,000 or, .0003. analysis of a study. This figure would include all potsons with lung cancer on December 31, regardless of whether the person has Other factors associated with both the discase and thc had the discasc for onc day or three years. Prevalence exposure may not bo known to the investigator and, is dimcnsionless, i.c., it has no units. thcrcforc, cannot be accounted for. Such In a casc-control study, on the othcr hand, discasc defines thc two typcs of individuals to hc studied, thosc l'Source: Based on mau:rials initially pr~psrcd by Robert S. with discasc (cascs) and thosc without (controls). Banks, R.S. !~qlr, Auociatcs, lac. for Electric puwcr Here, the concern is to compare these t~vo types of Ruearch locilute Seminar on New EMF Epidcmiologic individuals with respect to thc properdon having a Re,ulta snd Their Implications, October 16-19, 1990. history of an cxposurc or character~.stic of interest. For Submantisl revisions and sddidon$ have been made by Boji cxamplc, in a casc-control study of porsons with lung Huang, M.D., and P.A. Buffier, !~.D., University of Texas cancer, it would bo of interest to compare thc Health Science Comer at Hou,toa, School of Public Health, propor. inn having a history of cigareuc smoking in case and R. A. Beauchamp, Texas Department of Health. and control subjects. Appendix C - Results of F. MF Survey C-13 p~ m bu~d~ gh~ ~r 69-kV ~m~n ~ p~ ~ c~c~on of ~h~l. 2. ~ ~ l ~ ~ M~ ~-W~ ~ ~ ~C hcM-~ w~ld C~a~ kV ~ ~ on basis of ~!/ 3. C~ d ~ 345-kV ~ ~ C~ ~ ~ Ci~ C~ ~er ~ u~ty ~ ~y ~u ~t w~ ~ 4. C~ of ~ ~ - ~ ~H H~ City of ~cveh~ ~ hu~ ~fo~ ~c C~e~ City C~H, ~i~ ~ny. C~H v~ 184 ~ ~c I~ ~ city. Sub~m c~ de~ ~ May~ 1~1. 5. ~ ~b ~er ~. Hyd~ T~-h ~F was digu~ ~ hv~mi ~ct hm~m~ li~ was built. 6. C~ ~ ~ m ~ C~: ~ C~y C~H huh ~y ~ ~ ~F ~. ~ B but dH ~ m~r ~i~; ~m~ wu e~ ~ 1985. Appendix B B-3 , Relatively rare cxposuru (or occulation~) The limitations of retrospective cohort studies includc: can be investig~d. , hat exposures cannot be dcfmcd as precisely Prospeaivc cohort studies also have · number of u current cxposures. For example, it may be limitations: difficult or impossible to estimatt exposures to workers that occurred 40 years ago if no · They oeten t~luire many yeats of follow-up indnstrial hygiene data tu~ available and work since many diseases have long latencies, practices have changed over time. · A very large cohort and/or · long follow-up · Little information may be available on period would be required to investigate confouuding factors, such as smoking history. relatively are diseases, including most cancers. · It may difficult to select nsuitable population to which the cohort can be coml~tred. · Substantial effort and expense are necessary Frequently, the study results will differ to follow · large number of people over · depending on whether national, regional, or long period of time. local disease rates are used as the comparison. The problem of · suitable There are no known EMF prospective cohort studies. comparison population is avoided in large cohort studies in which internal comparisons Retrospective (Historical or Nonconcurrent) Cohort can be made, i.e., · particular category of Studies. Retrospective cohort studins differ in that both workers within the cohort can be compared to the exposure and the follow-up period have occurred the complete cohort. prior to the onset of the study. These studies utilize data from existing records such as o~eupational Some cohort studies involve both prospective and records, professional registries, and death certificates retrospeaive components. For example, a cohort may to identify the cohort and conduct the follow-up. be defined through personnel and other records as Studies of occupational groups ax~ most oeten everyone who worked ·t Company X at least one conducted using this approach. As with prospe.~tive month between 1945 and 1985. The mortality studies, the cohort is fast defined on the basis of experience of this population as of 1986 (the time the exposure status. Fo~ example, the cohort may be study is undertaken) can then be determined using state defined as all the members of a particular occupation or and national mortality records. Additional follow-up of all employee~ at Company X as of some specified time this cohort might then be conduct,'d in 1990 or 1995, in the past. The subsequent occurrence of disease in for example. the cohort (up to the time of the study) is then ascertained, genertlly by using death certificates. To Milham, 1988 is · cehort of amateur radio operators. reduce costs, an unoxposed cohort is ofitn not Lund, 1985 is a modified cohort study in which electric identified for comparison; instead, the mortality worker union members wore followed for six years to experience of the exposed group is usually compared to determine their leukemia risk. This study was designed that experienced by the general population in the state, only to be · screening tool however, and the rigorous region or country from which the study population was methods generally used in cohort studies to ascertain derived. and validate completeness of the cehort, cause of death, and exposure were not employed. Two ease-control The advantages of retrospective cohort studies include: studies of electrill workers have been derived from large cohort studie~ that wore not concerned · Much less time and less cost is requir~ to specifically with persons working in electrical complete the study compatod to · prospective occupations (Stem ct t!., 1986 and Gilman et study, sinc~ the disease outcome has aiRMy 1985). occurred. Case-Control Studies · These studies ax~ widely used in occupational This is the most common study design used in settings, where personnel x~ords, industrial hygiene data and other records can be used ~pidemiology. As outlined above, cohort studies first both to construct the cohort and establish identify the exposure status of non-diseased individuals, then determine the subsequent incidence of disease in some tough measures of exposure. the cohort. In contrast, case-control studies begin by first identifying individuals who have developed the B-4 Health Effects of F, tposure to Powerline Frequency Electric and Magnetic Fieid~ disease under study (eases) and individuals without the case=control studies in which the disease (controls). Cases can be selected through approprlsteness of the controls is questioned. hospitals, disease registries, health maintenance organizations, physician's practices, or even through , It is difficult or impossible to ascertain death certificates. Controls may be selected from accurate exposures that have occurred in the individuals who are "patients" at the same past. neighborhood as the cases, or who live in the same hospital as the eases, or from other sources. An , These studies are inefficient for studying rare attempt is then made to compare the previous exposure exposures. experience of the cases with that of the control subjects. Certain factors that can influence disoasc Many EMF studies arc of case-control design, ran (e.g., age) must be taken into account in the including Werthcimer and Leeper, 1979; Fulton et al., design oranalysis of these studies. 1980; Gilman et al., 1985; Savitz et al., 1988; and Nan et al., 1988. Case-control studies offer several advantages: Proponional Mortality Studies · They arc generally much faster and less expensive to undertake than prospective This study design is frequcn~y used in exploratox7 or cohort studies. "hypothesis-generating" investigations, usually in occupational settings. The cotire study is most often · Sample sizes can be much smaller than cohort based only on death certificate information: age, sex, studies, particularly in the case of relatively race, cause of death, residenco, and in most states, uncommon diseases. Whereas a cohort size usual industry and occulation. Proportional mortality of tens or hundred of thousands might be study is conducted, when only the humben and cases required to demonstrate some cancer risk, of deaths among the exposed group can be ascertained, several hundred or even fewer subjects in a but the structure of the population from which the ease-control study might be sufficient to numbers and cases of deaths is unknown. reveal the risk. For very rare diseases, the only practicable study design is a case-control In a proportional mortality study, the proportion of study. deaths from a specified cause relative to all deaths among the exposed group is enmparcd with the · A variety of previous exposure variables can corresponding proportion in the non-cxposod group or be (and usually are) examined in · single a gcp. cral population. This comparison is done study. independen~y of any relationship to the incidence rate of the disease in the exposed and unexposed groups, Case-control studies have a number of lhnitations as i.e., only numerator data are used to make the well: comparison. For example, · researcher might wish to investigate the hypothesis that EMF exposure is related · The cases solccted for inclusion in the study to leukemia occurrence. Assuming that the job ti~e may not actually be representative of all those "power lineman" is an adequate surrogate for exposure who develop the disease. For example, the to electric and magnetic fields on the job, the selected cases may represent only those rescaroher could plan · study to fred out whether individuals who entered particular hospitals, lcukcmht mortality in these workers was elevatc<l in and these cases may differ from non- 1970-79. A relatively quick way to do this would be to hospitalizcd cases. identify aH power linemen (from death certificates) who died in that time interval, determine what proportion · It is oaea difficult to select an appropriate died of leukcmia, and compare that proportion with the control group that is sufficiently comparable proportion of persons of similar age and sex in the to the eases as well as representative of the general population of deaths duc to kukemia. The general population from which the cases quotient of the proportion in all power linemen divided arise. Comparability and representativeness by the proportion in the general population is called the are both important yet sometimes mutually proportional mortality ratio (]>MR). exclusive goals in selecting controls. In some studies, two different control group have A similar type of study might also be done using been u'tdized (e.g., hospital controls and incidence data from a disease registry such as a neighborhood controls). Many of the regional cancer registry. Such · study is referred to as contmvcnies in epidcmiology arise from · proportional incidence study, and a proportional Appendix B B-5 incidcncc ratio (PIR) shorn the observed-to-expected McDowtli, 1983; CAllc and Savitz, 1985; Milham, · rio of cuu. The population of this type of study 1985; and Pearcc ct al., 1985. consists of those with cancon newly identified in a specified time period, and the study data is derived Cluster Studies mortality data, the proportions of canccn of diffcrcnt A cluster of disc·so cases (e.g., Icukcmia) is generally typu can ho compared among different occupational considered to bc an unosuaHy high number of cues groulM (if occupational data arc available in the appearing in the same setting (e.g., noighborhood, record). town, work place) over a limited poriod of time. Considerable attention hu been given to the study of P!viRstudies have the following advantages: discue clusters, particularly cancor clustcn, and a nombcr of statistical methods have bean developed to · They can bc conducted very quickly and analyT. c them. Nevertheless, no non-infectious agent inexpensively, especially if the relevant data hu cvcr bccn consistcnUy implicated in causation of arc alzudy computcrizcd (as they oltcn arc). any type of cancer by cluster analyses. · Many different occupations and causes of Cancor clustering is therefore generally thought to bc death (or typu of ca·cot) can bc examined the result of random (i.e., non-uniform) distribution of simultaneously. discue cues in apopulation. Thus, cvcn when a significant elevation of a disease is noted for a specified · They can provide many useful leads for time and place, it is very likely a statistical art·fact and possibly rclationlhipl between disease and not the result of the cues' exposure to an agent in their occupation that can bc investigated by more environment. For example, three new cases of pownrful study methods. lcukemia might occur in a community of 4,000 person in a five*year period, and this incidence might bc PMR studies also have some scvcrc limitations: significantly (p = 0.04) elevated relative to the incidence in a comparison population for the same time period. But such a cluster might not bc unusual. The · Unlike a cohort study, no information is calculations used to dotermine the statistical collected on the population "at risk"; only on significance of the increased incidence of lcukemla those already deceased. Therefore, no imply that if there arc 1,000 communities of 4,000 disnit rates can bc calculated. An elevation persons in the country, 20 or 40 communities in the PMR may 5c duc either to an increase throughout the country (dcpe, nding on the type of in mortality from the cause of co·con or to · statistical test used) might have an cloy·ted incidcnco of reduction in mortality from another cause. lcukemis duc to random statistical variation alone. For example, a PMR for lcukcmia my bc Thus, such a cluster, while 'statistically significant," cloy·ted in pownr linemen because they arc would not bc all that rare. It is also unllkcly that actla!ly at incrcalcd risk of lcukcmis, or similar environmental agents, lcukemla subtypes, or bocanit they have lower rilk of some other age or gender distributions of the cases would bc found common disc·so such I1 hcltrt distag. in these 20 or 40 communities. Information on confounding facton (Inch u smoking) It is possible that a cluster of cancer cases may bc is usually not available. related to the prose·co of an environmental carcinogcn. The fact that no such relationship has bccn established · Accuracy and complete·us of information through cluster analysis may in apart bc duc to the fact obtained from delth ccrti~.atcl or medical that cancer clusters generally consist of se few cases records is variaS!c (sec below). that it is not possible to test · hypothesis relating an environmental carcinogcn to cancer risk. Thus, the · Study results can somctimu differ quite size of the population being studied is generally significantly when compared to results of insufficient to yield statistically significant results. more definitive studies such u cohort studies, especially if the overall mortality ntu differ An example of · cluster study is Aldrich ct ·1., 1984. substantially among the groups (occupations) being compazed. B.2 Sources and Validity of Dat8 Some cxamptes of EMF proportional momtlity studies In any seicntific study, careful attention must bc given arc Wright ctal., 1982; Coleman ct al., 1983; to the validity and rcllability of the data. Unfortunately, B-6 Health Effects of Exposure to Powerline Frequency Electric and Magnetic Fields aH da~ collection methods involve some degree of Environmental Measurements. A variety of inaccuracy and variability. Thuc data quality problems environmental mcuuremcnts may be used to provide are addressed in ~cienti~c studic~ by such means u full ,-I.h. on exposure levels either in the community or in descriptions of data collection techniques, calculation of the work place. Data likcly or potential mcusuremcnt errors, validation procedure, and r~lication of measurements. arc sometimes avatilabic for spcci~c industries from industrial hygiene surveys. These environmental Because of the wide variety of epidemiologic data mcasuremcnts, coupled with detailed work histories sources, professionits outside the field may find it (usually available from personnel records) are difficult to judge the validity, utility, reliability, and frequcntly used to establish exposure categories for limitations of these data. Data on exposure of study individuals. Environmental levels can also be rncuured subjects, for example, may range from direct u part of the study. measuremenU of chemicals or their meUbolites within the body to the relatively imprecise information on a The use of environmental incasure·eats can involve decedent's exposure history tklccn from interviews with several potential problems. Complete data may not bc ncxt-of-kin. available on past exposure, particularly if it occurred many years ago as is oltcn the ease in occupational The following discussion briefly describes the major studies. Even when data ·re ·vailable on past or sources and limitations of cpidemiologic data, with curpent exposures, the measurement errors associated particular emphasis on determination of exposure and with a particular sampling method must bc considcrcd. disease status. Thc overall consequences of data inaccuracies are also briefly considered. Another problem with environmental measurcments in their relevance to personal exposure levels. Different Data Sources individuals who live in the same community or work at the same job may experience quite diffcnmt exposure Population Data. Census data are collected by the U.S. to the same agent. More relevant personal exposure Department of Commerce, Bureau of the Census, measurements may be obtained by the use of persorrel every 10 years, and provide age, race, and sex-specific dosimeters, such as radiation film badges. However, counts for porsons living in the various geographic even these "personal" mcusurements ·re not always an subdivisions of the U.S. plus certain additional accurate indication of the biologically4bsorbed dose of demographic, socioeconomic, and household an agent. In recent years, ennsidcrable attention has characteristics for those individuals. Census been given to identifying markers of exposure in information is collected, analyzed, and presented for individuals at thc molecular or genetic level. the country ·s · whole and for progrusively sivaHer subdivisions such as states, counties, citiu, and census Thus, environmental measurements of an exposure tracts within cities. These published data supply the variable are an indin.'ct index of personal exposures. 'denominator' (i.e., population at risk) for most Personal dosimetry provides a more direct meuure of conunonly repofied disease morbidity and morality exposure, but even here, some degree of error will rate·. result in classifying an individua!'s exposure. Fortunately, misclassi~eation of · comparatively small One of the major limitations of census data is that · propeRion of individuals does not invalidate a study. count taken once every ten years will not provide the Exposure can best be determined if some biological up-to-a-_,_-_ information necessary to accurately describe marker of exposure can be identified or if the agent (or a growing (or shrinking), highly mobtic population. an appropriate mctabelite) can be mcasured in the This problem is, to a degree, offset by the Curr~nt individual. For example. the presence of antibodles to Population Survey (CPS), which consists of monthly thc Hcpatitis B virus in the blood can serve us · sample assessments of ·bout 50,000 homes. However, biological marker of exposure to that virus. Lead there are many ~reas in the county in which the levels in serum or in ~__,~__t~ can give a measure of either population does not behave as expected and in which short-term or chronic exposure to lead, rupectivcly. population projeetious arc significantly in error by the end of the decade. Such errors in the denominator State or Local Health Departments. Nearly all states (,population at riak) of a rate calculation can cause the have some form of iofcctious disease reporting law rate to be significantly in error also. For example, if requiring physicians, hospitals, and/or schools to report population growth has been higher than expected, cases of specific infectious diseases which ·re morbidity or mortality rates may appear aRtificially considered to be of significant public health importance elevated for that area. to that state or to the Centers for Disease Control (CDC) in Atlanta, Georgia. A number of states also have some form of occupational disease reporting law Appendix B B-7 and/or aurvcillanc~ activities. However, the r~ox~g r~cords. Additionally, ~y glove fa~n ~ of s~ ~~ or ~~ d~ ~ m o~ve ~ ~y ho~ M~s~n. ~fo~, ~c ~ ~ or 1~ ~ d~t ~ o~ ~ ~on ~ ~ ~ ~u~ hospi~ or by ~y ~1~ profusion, ~d ~ ~ ~u ~ ~ ~e ~spi~ ~ a ~v~ ~ ~y ~ ~c~. for m~ di~ o~ ~t o~y 10 ~t or F~y, ~sp~l-g~ ~fies ~ o~ didcuR hs of~e a~ul ~ ~u~g ~ ~ s~. ~ ~H~, ~ ~ ~u~ of ~ ~ck of ~mpu~ ~M sys~. ~u~ ~d L~fions of Mo~ty Dn. ~e most F~e~ a~c~, such ~ ~ CDC or ~ N~o~ mu~=iy ~H~ ud ~mmo~y ~ ~u~ of ~ ~r for H~ ~s ~CHS), ~ ~ve ~r su~e~ of d~ ~ ~ money ~ ~or ~ive su~ on a ~ v~y of ~e~y, ~om~n ~ Ev~, 19~). Advu~g~ ~f~ ~ c~n~ d~ ~ h~ s~ ~1~ ~ ~ ~ ~ a.t. ~ ~y a~sible ~~ for ~e v~ g~p~c ~ of ~ U .S. ~ ~x~ive m ~. S~ ~ ~fic~. a~ ~u~ by ~w ~ ~e U.S. ~ ~y o~er ~ M~i~ R~. Hospiml or ~ ~ ~ a u ~H, ve~ ~ly ~ d~s ~ ~ ~ a~ ~u~t ~u~ of ~fo~on for ~mff~n or ~ m ~ au~ofifi~ ud ~ ~ a~ ~n~n of stay subj~. ~e ~s ~ a~ ~. F~Hy, mo~y ~ ~ve ~ ~H~, provide ~fomfion ~ut a ~n's age, sex, ~, ~bu~, ~d pub~h~ un~y for ~y d~ ~s ~ addi~o~ ~levut m~i~l ~. ud for many diffe~t ~un~ of ~e ~rld ~ ~t ~ ~1 ud g~phi~! ~ds ~y ~ st~i~. ~e ~fo~on ~na~ ~ m~i~l ~ ~ hilly F~ o~er ~ b~ e~t which ~ow such v~ ~ a~sib~, a~uncy ud ~mpl~s. ~m~ons- A~s m m~i~ ~, ~ for gi~c ~h, ~ ~ ~ due ~ privacy ud ~1 ~n~ms ~ ~ ~ a num~r of pmbl~s ud ~iUfions ~y si~ons. ~er pa~l pmblm ~volv~ ~ ~ ~e u~ of mo~ty ~ for ~ ~y jud~U ~t m~t ~ ~e ~ ~i~g ~iologic st~ or ~ u ~i~r of ~u~cy of ud ~sn~g ~e ~ ~m m~i~ ~. ~ d~. ~me of ~e ~volve b~ ~ mo~ty ~mt ~ ~y ~ ~ ~e ~ns~enble va~on ~ ~g ~ich a~ d~t on ~e ~eu~r · e de~n ud di~os~ of ~ d~. ~e d~e. For e~p~, ~e mon~ty ~ for d~ ~ d~ ~ ~ mutely di~o~ ~ a hi~ ~h ~ve hi~ ~ faUHty ~ ~ mo~ ~ely m de~ of ~ty (e.g., ac~ le~e~), o~en a~ a~ua~ly ~t ~e ntus for ~t d~e ~ ~e n~ ~y or ~fly dip~ (e.g., ~e~er's mnBty ~ for a d~ ~t !~s o~ ~ul~ ~ d~). Di~os~c e~ ~ v~ ~g m ~e d~. In addi~on, ~ dim~ ~ich a~ m~y ~yaie~, ~e ho~, ud ~e g~p~ ~n. di~o~ ~y ~ over-~~ on ~ d~ ~ ~ve m d~ for ~h ~e di~os~ ~ dipsfie e~ ud acum~ for a ~u~r ~ mo~ differ. d~ ~ v~ s~ay over ~e. One ~on p~u~ for ac~g ~me de~ of ~~on ~e a~e ofd~ ~y ~ ~u~ly ~ on ~e ~ m M~fi~ ud ~e d~ a~g m ~e d~ ~ for a nu~ of ~ns, ~clud~g di~os~c e~si~ons found ~ ~ ~n~o~l unffi~ty of ~ phys~ ~ ~e ~t m~i~l C~s~n ofDi~ aCD). S~ ~jor ~iio~ h~m~ of ~e ~. ~ of d~ ~ich ~r ~me of ~e ICD ~ur ~ut eve~ ~ y~, ~ i ~t ~ sfig~, s~ u su~e, ~x~Hy ~smi~ m s~ ~ch ~i~n ~ foHo~ (~ ~ ~ion d~, or a~n-~ d~s, ~y ~ of ~ ICD ~ pub~h~ ~ 1978, ~e 1~ ~ 1~). si~fly u~er-~ on ~e d~ ~. fip~c c~m shou~ ~ exp~eifly ~b~h~ at ~e ~s~g ~e und~ly~g ause of d~. R~ions ~ ~e ~y ou~, ud ~ c~m should ~ clarly in~n~l C~si~on of D~ ICD ~ or ~ou~ hospital m~il ~s or ~sp~l d~ for a ~cu~r ~ of~ ~e. ~io~y, ~e ~ ~io~y provide val~ble mu~ of ~e of d~ i ~ ~y but ~ ~en d~ m~ity ~, ~fo~n b~ u~n hosp~! ~ on ~e ~. ~m~, a di~o~c M~s~ ~ ~ely m ~n~ si~fi~t b~. Muy 'h~l' h mo~ h~v~y ~ ~ one p~ ~ ~ d~ ~u~ no m~l ~n or ~e u~er or di~osfic "~" ~y ~ur ~h ~uR ~ ~divM~ ~ ~ by phys~ on u ought ~e d~pm~o~ usi~t of d~ m a f~ ~. ~uenfly, ~ ~ ~ not a~r ~ uy ~culr d~ ~. ~ or p~ale~ ~ defiv~ from hospi~l B-8 Health Effects ofF, xposure to Powerfine Frequency Eltctric and Magnetic Fields Variations in the quality of medical care over time or Disease Registries. In same arc·s, disease registries from place to place may cause deceptive differences in have been established for the surveillance of diseases mortality r-_*__-*_. For example, improvements in which are of major concern to pubtic health. Cancer is diagnostic procedures for a certain disease may result one such disease entity. In 1972, the National Cancer in an apparent incrcue in the mortality for that disease Institute (NC!) established the Surveillance, when it is only the ratio of dlagnosed cases to Epidemiology, and End Result (SEER)Program which undlagnosed cases that has changed. consists of a number of population-based cancer registries from different geographical areas across the Lastly, elderly people may have several active disease country. These SEER registries systernaticaHy collect processes at time of death, and selection of the demographic, diagnostic, treatment, survival, and underlying cause of death may not be possible or follow-up infornmtion, on all patients in the aroa who feasible. For this last reason, more and more are diagnosed as having any of the various forms of authorities arc recommending the use of multiple.cause cancer. coding for death ceai~cates (Kelscy, Thompson and Evans, 1986). The term "population-based," as used above, means that the registry has made every effort to enroll every In conclusion, mortality data can, unquestionably, help case of cancer that occurred among the residents of a an epidcmiologist to understand fundamental disease particular geographical area such as a city, county, trends and relationships and to help generate region, or state, When aH (or very nearly nil) eases hypotheses about risk factors and ctiologic agents. from the geographical area have been ascertained, the However, because of the potential problems with 'population at risk" is the population of that area, and reliability, caution should be used in the interpretation only then is it possible to calculate a cancer incidence of the results of any mortality data analysis. Such rates. This is in contrast to a *hospital-base' cancer findings should be considered preliminary suggesting registry which may have enrolled every case of cancer further exploration through more def'mitive methods that was dlagnosed in (or admitted to) a particular 0Celsey, Thompson and Evans, 1996). hospital but which has insufficient information to generate cancer incidence rates because the "population Death certificate data are used frequently in at risk" is undefined. epidemiologic studies not only to ascertain the cause of death, but also it identify the decedent's occupation (a Registries, particularly those for cancer, are becoming possible indicator of occupational exposures) and · widely-usad resource for occupational and certain other demographic data (e.g., age at death, date environmental studies. The approximately 45 of death, sex). Several studies have shown that death population-based cancer registries in the United Sates certi~c~,_,_e data vary considerably in accuracy and collect d_n_,a on cancer cases in prescribed geographic reliability. Cordtin causes-of-death are more accurately areas. Some collect additional data on treatment and or more reliably reported on death certificates than are follow-up. These systems can be used to identify other causes. This variability is due to factors such as unusual clusters of cancer cases around a particular uncertain diagnostic criteria for same diseases, the environmental exposure, to study occupational hazards, existence of multiple diseases, whether the death to assess the magnitude of and trends in cancer rates, to occurred in a hospital, and whether an autopsy was identify cases for case=control studies, and to facilitate performed. the ascertainment of cancer among members of · study Occupational data from death certificates have even greater variability. Death certificates generally list the Despite the obvious usefulness of registries, the decedent's "usual occupation," as reported by family completeness, timeliness and accuntey of registry d_a_t_a_ members or other persons. Just as there is no routine vary substantially from one registry to another, and confirmation of cause.of-death (i.e., autopsy), there is these can be factors of concern in epidemiologic no verification of occupational information. Studies studies. In addition, limitations on access to porsonal show that certain occupations are systematically over- identities reduces the utility of these data for or under-repre. ented. For individuals with · multiple- epidemiologic studies requiring follow-up or record job history, the most recent occupation may be listed, linkage. rather than the usual occupation. Thus, except for · decedent who held a singic, well-defined occupation Special Disease Surveys or Epidemiologic Studies. and for whom this information is accurately entered on With the best of intentions, routinely collected _d~ta on the death certificate, the occupation shown on the death illness from hospitals, registries, and other sources do certificate may represent an incomplete and inaccurate not yield · complete picture of the illness and disability characterization of · pomon's actual occupational picture of· defined region. More comprehensive data history. for monitoring the health status of a region are Appendix B B-9 providcd by samplc survcys, known u morbidity overall findings arc not ncccsurily invsdkhtcd. As surveys. Such surveys may consist of a single, cross- long as the exposed group has on average · scctiontl examination or of long-torm Oongitudinal) significantly higher cxposur~ level than the non- studks in which rc~pondcnts trc x~visitcd periodically. exposed group, · comlmrison can still bc nmdc and a Although the most ambitious of these, the N~tional corroct conclusion reached evcn though the precise Health Survey, collects data on aH diseases, specific exposure-response effect will be somewht in error. disease states may be surveyed. The NCI Ms conducted cancer surveys which have yielded basic In general, occasionsl ~nd random mischssi~cation is a_,.t,. on the incidence of cancer by primary site, likely to rr. ducc the strength of, but not cl:unlnatc or histologic type, and state of disease s/diagnosis. Both reverse, a significant association between exposure and lcukcmias and brain cancers ·re included in these disease. Frequent or systematic ntisclasslflcation, on surveys. the other had, can obscure or even change the direction of an association, or create an association Data on personal facton, exposures, and even disease whcrc none exists. statns arc often obtsincd from in-person or telephone intctvic*ws of study subjects or thclr rcbttives. The Sound cpidcmiologic research stzlves to identify, validity and rcliabillty of these data arc affected by uscss, and reduce the inevitable inaccuracies in data many factors such as the training and experience of the collection. !t is particularly important that frequent or interviewer(s), the length of the interview, s~l rcccncy systematic crron be recognized. This is accomplished of the events questioned. In short, thc quality of data through the usual scientific means of providing dinlied obtained from intorvicws depends on how carcfuUy the descriptions of data sources, observing rigorous intorvicws arc designed and administcrcd, u wcll as on mcthodologic standards, and confuming data whencvcr the type of information being sought. The information possible. obtained from interviews is frequently confn'med by other data sources (e.g., medical records) to improve B,3 Comparability sad Bias its validity. A serious threat to thc validity of any cpidcmiologic Msilcd questiommit~ arc anothcr source of a_._~_._ on study is the possibility that its subject selection and data demographic factors, exposures and health status. As arc biased. In cpidcmiology, "bias" does not imply any with interviews, the quality of qucstionrmirc data is prejudice or projudgment on the part of the study influenced by many facton including length of the investigators. Rather, "bias" gencntUy refers to questiormairc, specific wording of thc questions, types systcmatic errors, i.c., crron other than sampling of questions, motivation of the study subjects, and the variability that prevent the truc value of · disease rate perceived imporUmcc of*the study. Qucstionmdrc data or other production variable from being obtained. Thc arc also frequently confirmed (val~_-_~_~) by other data introduction of bias rcndcrs study groups sources such as medical records. noncomparable in some important way. Impacf of Dcaa Errors To undcrsumd how a study can become biased, it is useful to recall the dcgrcc to which the hberatory As indica_t__M_ in the previous discussion, the data scientist strives to achieve comparabillty between sources uscd by cpidcmiologists arc cach characterized cxposcd and uncxposed organisms in his/her by potential inaccuracies. These itmccuracies arc experiments. This is accomplished by such means as usutlly woll-rccognizcd by cpidctniologists, and efforts using a single strain of test organism, randomly ar~ made to lin~it and/or quantify their extent and assigning cach organism to an exposure group, magnitude. Inevitably, however, some error remains maintaining uniform environmental and dictary and must be exstnined in terms of its impact on ovcratl conditions during the course of thc study, and using a study vslidity. consistent protocol for examination. Thc cxsxnination is performed with the investigator 'blinded" to thc Epidcmiologic studies usually involve the comparison subject's previous exposure history. Failure to achinvc · of hrgc groups of subjects, often hundreds or any of thcsc major comparability elements can bias a thousands of individuals. Study subjccta arc gnncntHy study, and its conclusions must be considered suspect. chssificd according to whether or not thcy asg cxposcd to the agent in question, as wcH as according to their in cpidcmiologic rucarch, utilizing an obscrvslional disease status. An error in usessing an individual's methodology, an equivalent dcgrcc of comparnbi]ity dcgrcc of exposure may affect his rchtivc rgnking in c~rmot be achieved. The goal of the cpidcmiologist is lk exposed group, but will not necessarily remove him to select from an existing population exposed and non- from the "exposed" c-_!_-gory. Even if an occasional exposed (or discssod and non-<liscsscd) groups tht arc individuad is complctcly misclassi~ed on exposure, thc B-10 Health Effects of Exposure to Powerline Frequency Eltctric and Magnetic Fields fundamentally comparable and from which ecluival~nt Healthy Worker Effect. Overall, the wodcing data can be obtained. Bias can be introduced in population is healthier than the general population. numerous ways, some of which cannot be known or Less healthy peopled ·re less likely to become or controlled by the investigator. Th~ following discussion re4nain employed; thus employment serves as · provides repose·tat·re examples of biases that can selective ·clot. In occupational studies, therefore, a occur in epidemiologic studies. clear possibility of bias arises if the general population is used u the basis for obtaining expected disease rates Blasts as is olten done in retrospective cohort studies. The use of general population rates can crea~ the f·lse Bias can noeur in virttmHy every aspect of ·ppcaranco that employment in an industry actually epidemiologic research. However, the following affords protection against mortality, or ·t least that examples illustrat~ only those biases that can arise (1) there is no excess mortality. This effect is more in the p~s used to specify or select study evident for some discues than others. For example, it partieip·nts (selection bias), (2) in the process of does not appear tlmt overaJi cancor mortality is collecting data on study participants (observation or typically lower than expected in occupational cohorts information bias), and (3) due to the existence of when compared to the general population. factors that are associated both with the cxlx~sure and C~rdiovucular disease on the other hand, is typicatlly the disease (confounding bias). Some types of bias can found to be lower in am occupational ~ohort than would only occur in a pa~cular study design, while othcrs be expected based on general population rates must be considered a possibility with any design. (McMiehael, 1976). Non-Response Bias. Some portion of thosc who arc Incidence-Prevalence Bias. Newly dlagnosed select~l or identified as study subjects cannot or will ("incident") cases of disease may differ in certain not participate in the study. Bias can occur when this characteristics from all existing ("prevalent") cases of group of non-respondents differs systematically from the disease. Studies that use prey·lent cases are more respondeats with respect to exposure or disme status. likely to include longer duration cases and exclude both For example, non-respondents may have more serious rapidly fatal and readfly cured cases. Consequently, health problems. To minimize this bias, considerable prevalence studies arc more likely to yield information effort must be expended to ·chieve a high participation relating to disease duration, rather than disease rate (e.g., 90 percent or better), or ·t least to obtain · development. ample of non-respondents to determine if, or how, they differ. Observation and Measurement Bias. Numerous biases involve non-comparable data or data collection Lost-To-Follow-Up-Bias. A similar type of bias can procedures. For example, exposure data obtained from occur when study participants arc "lost-to- follow-up" interviews will not be comparable to data obtained from in cohort studies. Those who are 'lost" may differ in environmental measurements. their disease outcome, removing from the study, for example, those who had more serious health problems. Observation bias can ·lso ·rise when interviews are Significant efforts must be made in cohort studies to used to obtain health outcome data. If the interviewer keep this proportion as low as possible. It is impotunt knows · subject's disease status as well as they to choose a propor study population, such as · group of hypothesis under study, he or she may subconsciously electrical workers in · company. The workers arc probe harder concerning past exposure. Furthermore, more easily located and followed-up. a subject who has been advised of a study's purpose may attempt to provide responses that ate poreeived as Dete~ction Bias, This bias refers to the situation in "favorable" or "helpful" to the interviewer. Thus which a discas~ (such tl a br~in tumor) is more interviews are fr~luently eonduetexi in a "double-blind" frequently dlagnosed and ascertained in a particular fashion where neither the interviewer nor the subject is population (such as an occupational group) than in the ·ware of the specific hypothesis under investigation. general population. This may result from better ·cceu to medical care. If the general population rates are Disease diagnosis may be influenced by the physician's used as · basis for comparison, it will appear that knowledge of· subject's exposure history. This bias, mcrnbers of the particular population are at increased which applies mainly to cohort studies, will tend to risk of the disease. In reality, they ·re ·t increased risk increase the association between exposure and disease. only of being dlagnosed with the disease. Widespread publicity about some ·gent may prompt exposed Recall Bias. A person who has developed · particular porsons to seek medical examination, ·g·in increasing disease may be better ·hie to recall prt~vious exposures, the possibility of detection emong members of that compared to control subjects. The mother of a child group. born with a realformation may better recall (or perhaps Appendix B B- 11 exaggerate) exposures experienced during pregnancy, study 's findings. Strict adherence to established compaxed to the mother or · healthy child. This is · procedures and standards can reduce many, but not all, serious potential basis in case-control studies. possibilities for bias. In many well-reported studies, the nutboa frequen~y discuss the possible sources of Misclsssifiation Bias. This does not refer to random bias in their study and attempt to show, through logic rniaciassification of some small proportion of the study and/or a._t_a_, that their findings arc not likely to be due lx>potatlon, but to systematic misciassi~cation of to some bias. But there still remains the possibility that disease or exposuro status. This could happen if, for some unknown bias is operating. example, the job title "electrician" is used to identify subjocts occupstionally exposed to EMF, and , large B.4 Association or Causation? portion of these electricians work only on dead circuits, where EMF exposure is low. On the contrary, some In laboratory research, a well-designed experiment that other workca, who arc identified u "non-electricians', results in a statistically significant effect (i.e., one not might bc exposed to EMF. due to chance variation) is usually intorprctcd as demonstrating · cause-and-effect relationship. The Berksonian Bias. This bias, first described by Berkson existence of · cause-and-effect relationship cannot be in 1946, can arise when controls arc solectcd from so readily inferred from hospitals (usually the same hospitals from which the cases have been selected). Hospitalized controls are observational epidemiologic studies. The generally not representative of the overall population cpidemiologist, at best, can show that some association with respect to cemtin characteristics; e.g., smoking, or reiationship exists between an exposure (e.g., alcohol or coffee consumption. In order to detect and chemical, radiation) and · physiological or health- prevent Betksonian bias, multiple controls are rchted effect (e.g. , blood chemistry, disease, death). suggested. Therefore, some case-control studies have Typically, the epidcmiologist further attempts to used both hospital and neighborhood controls at the demonstrate that the association is unlikely to be due to same time to delete this bias. chance and is not due to some third (confounding) v·riable. Confounding Bias. Confounding bias occurs when there is a third variable which is not of intcrcst to the Judging Positive Associations study, but it is related both to the exposure and the disease under study. A study of lung cancer, for Scientific "proof" of · cause-and-effect rchtionship example, will find a significant association between cannot be obtained from an observational study. alcohol consumption and the development of lung However, as · practical nmucr, explicit or implicit cancer. In such a study smoking would be considered a confounding variable since it is associated beth with judgments of causality arc frequently derived from such studies, and strongly influence public health policy. alcohol consumption and lung cancer. If the effect of Therefore, it is important to consider epidemiologic smoking is rcmoved in the study Coy design strategies findings from · variety of perspectives, such that · such as matching, or analytic techniques such as reasonable assessment can be made. Epidemiologists stratification or adjustment), the association between have not established hard and fast rules for determining drinking and lung cancer might not retnain. when · positive association should be considered · cause-and-effect rehtionship. Different experts stress Most epidcmiologic studies consider at least several different factors in evaluating associations, and not all well-established confounding vsriables such as ·gc, agree that certain items arc particuhrly useful. sex, race, and socioeconomic class. Other variables However, certain guidelines arise frequently in may be examined in · particular study. However, not discussions of causal relation·hi!M, and these arc all variables can be examined; confounding bias is discussed below. probably present in all dam. Conctusion~ Strength of Association. The more strongly an exposure is associated with · disease, the more likely it is that the exposure causes the disease. Many different Given the difficulty in recognizing and controlling all measures of association are utilized in epidemiologlc potential sources of bias, no study should be considered studies, depending in part on the type of study completoly frec of it. For cxamplo, rarely does · study conducted. Most are expressed as some ratio attain 100t follow=up of subjects. Critics can always comparing or estimating disease risk in the exposed to attributo study findings to some form of bias sinco there disoasc risk in the unexposod. For example, the arc indeed so many potential sources of bias. "relative risk" msy be dct"med as: However, it is difficult to ·ctuaHy demonstr·tc that some bias accounts for or even materially affects the B-12 Htahh Eftacts of Exposure to Powerline Frequency Electric and Magnetic Fields · Risk of disease among those exposed possible causa~ve agent. Prospective cohort studies most fm~ly establish this time sequence. In contrast, cross-sectional (prevalence) studies do not generally · Risk of disease among those not exposed permit determination of whether exposures preceded disease development. It may also be difficult to The 'strength" of the association refers to the establish the time sequence in same case-control studies magnitude of the risk ratio. Generally, the hrger the in which, for exarnple, a disease may have a very long magnitude of the relative risk, the stronger the prr~linical (non-symptomatic) phase (e.g., asbestos association. As an example, smokers arc ten times and lung cancot). more likely to develop lung cancer than non-smokers (relative risk = 10). This grea~y elevated relative risk Dose-Response Relationship. The existence of a dose- makes it much less likely that some other vaxiable response relationship between exposure level and (con founder) ovcrlookcd by the investigator(s) is disease incidence supports a causal interpretation. In actually responsible for the association. A very low other words, thosc who have the highest exposures relative risk, e.g., less than 2.0, would have a grcstcr should also have the highest disease risks. It has been probability of being due to some study bias or clearly demonstxn_,__,-.4_ that the number of cigarettes confounding factor. smoked corrolstes directly with the degree of lung cancer risk. Although · strong association is very suggestive of a causal relationship, it cannot be argued that · weak The absence of an apparent dose-response effect is not association is not causal. It might also be noted here considered evidence against causality. Exposures may that the strength of an association does not by itself not have been ascertained accurately enough in studies, reflect the overall public health impact or importance leading to misclassi~cation and bias that can obscure · due to that risk factor. risk gradient. It is also possible that some threshold of exposure is necessary for a given ·gent before an effect Other crucialvaxiables ·re the frequency oft he disease is observed. There is only weak evidence, for associated with the risk factor and the number of example, of a dose-response relationship between people commonly exposed to the risk factor. A risk occupational asbestos exposure and mesothelioma risk, factor associated with a high reletivc risk for · although there is no doubt about the causal nature of relatively rare disease may have less public health the relationship. impact than a risk factor associated with a lowor relative risk for · more common disease. For example, Cohercnce/Pleusibility. A causal hypothesis is occupational exposure to asbestos has caused many supported when an association is consistent with or more deaths through lung cancer than through supported by othcr knowo facts and observations. For mcsothclioma (a rare disease) although the relative risk example, a causativc hypothesis is favored if there is for mesothelioma is much higher than the risk for lung some demonstrated or potential biological mechanism cancer. And, of course, a risk factor commonly by which the effect can be explained. The cellular encountered by many people may result in more illness effects of ionizing radiation have long been recognized even though it is associated with a low relative risk for and offer a clear explanation for the health hazards of disease, than a rarely experienced risk factor with a radiation. In the case of cigarette smoking, laboratory high relative risk for disease. studies have identified a variety of organic compounds in inhaled smoke; a number of these compounds have Consistency. An association consistently found by been shown to cause cancer in animal studies. These different investigators, in different populations, and/or findings arc the consistent with the human in different geographical areas is more ilkely to be epidemiologie evidence. causal. Although the many studies that have examined lung cancot and smoking have involved many differeat The absence of a recognized biological mechanism investigators, study populations, locations, and study does not necessarily contradict · causative designs, all have shown a very strong positive interpretation. The lack of an apparent mechanism association. It is extremely unlikely that such may only reflect an early stage of investigation, · consistent findings can be the result of same situation well illustrated by the outbreak of toxic shock overlooked bias or can occur by chance alone. Some syndrome. Early epidemiologic findings clearly investigators attach special significance to consistent indicated that the highest risk group consisted of associations found in both retrospective and prospective young, menstruating women who used a high- studies. absorbency brand of tampon. However, it was found that some non-menstruating women and some men Temporal Relationship. Obviously, an exposure must were also disease victims. Thus, the use of high- precede a disease if it is even to be considered a absorbency tampons alone obviously could not account Appendix B B-13 for all cases, and there was considerable doubt by some A cohort study is similar to a laboratoi7 study in terms (including the manufacturcn) that the tampons could bc of the time sequenco of events. A study group (cohort) causally related to the disease. However, the absence of healthy people is identified and each individual is of an explanation or recognized mechanism did not then classified according to whether he/she is exposed prevent withdrawal of the tamports from the market. or not exposed to the agent under study. At some later Soma I~nc later an explanation was found: the discue point in time, which may be many years later, the was actually caused by a toxin from a relatively cohort is re-checked (the "follow-up")to identify those common bacterium. The toxin is only produced under study participants how have occurred when the study is certain physical and biological conditions, conditions imti""""""'~ depending on the study design. Two risk which are more likely to occur in young women using men,sum from cohort studies will be described: the high-absorbency tampons. relative risk and the standardized mortality ratio. Spccificity. When an association links exposure to a Relative Risk. From the follow-up dam, an actual single disease rather than to a broad spectrum of disease rate can be tabulated separatoly for beth the diseases, a causal interpretation is favored. An exposed and non-exposed groups. Depending on the example of high specificity is the usociation between specific disease in question, the rate may be either a occupational exposures to vinyl chloride and morbidity or a mortality rate. As an example, consider angiosarcoma (a rare form of liver cancer). The high that the exposure under study is cigarette smoking and specificity, as well u the strength of this association the disease in question is lung cancer. Data from an leaves little doubt as to its causativc nature. actual study showed that the lung cancer mortality rate for a particular age-group of non-smokers was 19 A lack of specificity, however, does not necessarily deaths por 100,000 par year. In contrast, the rate for argue against causality. For example, cigarette smoken in this sama age.group was approximately 190 smoking has been associated with a wide range of par 100,000 par year. diseasu. In fact, the smoking history of study subjects is almost always considered in well-designed studies of These rates can be compared in several ways such that other diseases. This lack of specificity, although still a quantifgd expression of risk can be obtained. The sometimes raised in arguments by the tobacco industry, most common measure of risk in this type of study is is not particularly troublesome to cpidemiologists sinco the relative ris/c (RR). The relative risk indicates the a great many components have been identified in increased (or decreased) degree of risk of disoasc tobacco smoke, and many of these components can be among the exposed compared to the non-exposed. it transported through the body to different sites. This provides a measure of the causativc importance of the rather broad exception to the coneopt of specificity has exposure under study. A relative risk with a value of led some epidemiologists to consider this guideline one (1.0) indicates no usoclation between the exposure useleas in determining whether an usoeiation is likely and the disease. to be eausal or not. The relative risk is calculated u follows: Conclusions Relative Risk (R.R) = rate in the cxnosed It is important to cmphasi,zc that none of the above rate in the non-exposed factors is sufficient either to prove or disprove that an association zipresents a true cause-and-effect For smokers, the relative risk is: rehtionship. They do, however, offer some reasonable guidelines with which both cpidemiologists and non- RR -- ]90/100000/y~ar = |90 = 10 epidemiologists may judge whether · positive 19/100000/year 19 association is likely to repruent · mac cause-and-effect relationship. This indicates that smoken have ten times the risk of dying from lung cancer compared with non-smokers. B.5 Statistics: Risk Estimates Confidence litnits can (and should) be computed for The following discussion addresses how rates and relative risks to determine if the risk is statistically ratios are used to provide estimates of r/~, with significant, that is, not due to chanco. Confidenco emphasis on the measures of risk derived from two limits are the range of the risk estimate which takes into major study designs: the cohort study and the cue- account sample size and variability. If the range of the control study. estimate does not include 1.0, it is recognized as being statistically significant. Risk Mcuures from a Cohort Study B-14 Health Effects of F;cposure to Powerfine Frtqutncy Eltctric and Magnetic Fields Standardized Mortality/Morbidity Ratios (SMR). In a PMR study, on th= ether hand, the data are otto· obtained completely from death certificates. Standardized mortality ratios (SMR) arc used for Consequently, the investigator only has data on people adjusting mortality ntu in order to compare health who have already died. Recall that death certificates outcomes between populations that tray have different also list the usual occupation and othcr porsonal data distribution~ of important variables such u age, sex, or such u age, race, and sex. The investigator does not race. Indirect standardization (adjustment) involves (and porhapa cannot) obtain denominator data, i.e., the applying mortality rates from some solected referrace total porson-ycars or porsons at risk (most of whom population, adjusted for age and possibly other facton, may even still be living). Thus, · mortality rato cannot to the study population. ~ procedure generates the bc determined; aU that can be done is to compare, for number of cleatha that would be 'expected' if the study example, the proportion of all deaths that were duc to population had cxpcriemccd the same dig·so incidence leukemia in one occupation with that proportion in as the reference population. Then, · common way to another (reference) group. compare ~ expected number with the actual observed number is to compute the standardizcd mortality ratio Although the PMR and SMR appear superficially ($MR). This is done by dividing the observed number similar, they tr~ quite different and arc derived from of_a,~__t_~ by the expected number, and then multiplying different types of data. Because they are both widely the quotient by 100 to eliminate decimals: used in occupational cpidcmiology, the distinctions need to be undcrscorod here. $MR = Observed deaths x 100 Expected deaths Risk Measures from a Case-Control Study An SMR of 100, then, means that the expected and In a centrut to cohort stodics which determine observed deaths are essentially equal in number, and subsequent disease rates between exposed and non- no cxcesa risk is evident. An SMR of 120 means that exposed people, me-control studies fwst identify there 20 percent more death, than expected, while an diseased and non-disused persons, then ascertain their SMR of 80 would mean that the observed deaths were pr,.'vious exposure history. This approach does not only 80 porccnt of the deaths cxpcctod based on the permit determination of ·ctual disme rates. Thus, · reference population. SMKs are used frequently in relative risk cannot be determined. One can, howover, occupational studies. compare 'exposure ratios' between disused and non- diseased groups. Under certain conditions, these In · typical SMR occupational_ study, the investigator exposure ratios can be used to estimate the relative t4Jk hu collected cxumsivc information on who hu worked by calculating an odds-ratio. in the industry, when, for how long, in what jobs, and if dcccag~i, the cause of death. Thus both numerator To illustrato the odds-ratio calculation, it is first useful data (dcaths) and denominator data Q~cnon-ycan or to catcgorizc ca·c-control study data in · two-by-two ponons at risk) are collected, and an actual mortality table according to discue and exposure status: rate for each dig~sc can be determined. This serves u the basis for the observed number of deaths. Exposed Unexposed Deceased pcnons (cases) A similar ratio can be calculatcd using morbidity data. Conuols b d For example, · standardized incidence ratio (SIR) can bc detcrmined using only incident cases. Thc letten represcnt thc numbcr of study subjects who ProportionalMortality Ratio fall into thc four catcgories. Omitting its dcrivation, the odds-ratio (OR) is then calculated u: An entircly different ratio, which appears frequently and aJmost cxclusivcly in occupational studies, is the proportional mortality ratio (PMR). As previously OR- · x d bxc ducribcd, proportional mortality expresses thc proportion of all dotths that are due to onc causc. For As an example, consider that the following data are cxamplc, of those who worked in · particular industry, obtained from · case-control study: 20 pcrcent of thc deaths may have been duc to cancer, whereu heart discasc may havc nccountcd for 35 Exposed Uncxposed Controls 60 Appendix B B-15 first bixth and parity ate themselves related and that one The odds-ratio is then calculated as: is a confounder. Instead of · single 2x2 table, thc investigator might con·react a series of 2x2 tables OR = 85 x 60 = 8.5 showing the association of breast cancer and age at first 15 x 40 birth for different parity levels (e.g., one child. two chiklren, and three or more children). In other words, The odds-ratio will be · reasormble estimat~ of the the investigator 'stratifies" on the variable parity. The relative risk if the disease in question is relatively rare investigator might also look at breast cancer and parity, (e.g., ~), the exposure is relatively common, stratifying on age at first birth (e.g., under 30 years, and, of course, there stre no serious study biases. and 30 years and older). The investigator would then fred that the late age at fn-st birth shows an increased The odds-ratio is intezpretod exactly the same as the risk ratio regardless of parity level. Stratifying on age relative risk. If equal to one (1.0), it suggests no at first birth, the investigator would find that parity is association between the exposure and disease. if no longer associated with bumst cancer risk. grc_~_!er than one, it indicates · positive association, and Stratifying has thus eliminated the confounding of if less than one, · negative association or · protective parity and age at farst birth by examining each v·riable effect. As with relative risks, confidence limits can be separately. computed to determine if the odds-ratio is significantly diffc'rentfrom·v·lueofone. Where several 2x2 tables and risk ratios are constructed in an analysis, a summary odds ratio can In some case-control studies, the controls ·re still be calculated. The usual technique for calculating individually "matched' tot hem during the selection · summary odds ratio is the Mantel- Hienszel process. For each case identified, · systematic procedure. This summary odds ratio can be thought of approach is used to select · control who is in the sam as · weighted ·verage of the individual odds ratios age bracket, of the same sex and ra~e, etc. This ficm the separate 2x2 tables. Actual techniques for matching process avoids having to account for these calculating the summary odds ratio and its confidence variables later in the analysis. When this type of limits an be found in basic epidemiology or biostatistics matching is used in · study, the odds-ratio is calculated tom. differently than above. Mathematical Modeling. Until relatively recently, AnoJytieControlofCo~foutulers stratification was the main approach to analysis in epidemiolegic mich. The limitation, however, is As previously described, epideminlogic studies involve that if more than · few variables ar~ centrolled for at examination of numy different variables in addition to one time, too few subjects fall into each stratum, and exposure and disease since there ar~ other factors that the resulting risk estimates become unreliable. In this may be related to the exposure and/or disease under situation, the addition or deletion of even one subject in study. The challenge then is to tease out the effect of · stratum could change the risk estimate dramatically. the exposur~ of intorest in the presence of these other Mathematical modeling overcomes this limitation. facton or co-variates. In practice, analysis and Many different mathematical models are available interpretation of epidemiologic studies involves much today, some of which ·re available on personal more than · single 2x2 able. If confounding variables computers. The most commonly used model is 10gistic are not controlled for in the design of· study, then they regression. Although developed originally for use with must be controlled for in tie analysis. Two approaches cohort studies, loptic regression is commonly for the analytic control of co-vat··tea or confounding employed in me-control studies. Discussion of tie are stratification and mathematical modeling. assumptions, use, and limitations of logistic regression and other models can be found in most recent Stratifieatlon. if the number of variables to be epideminlogy and statistics tom. controlled for is not too large, and the range of values for these variables is not too wide, then stratification Attributable Risk Measures provides · simple and powerful tnalytic technique. In ~n, the investigator looks at the relaliouship Simple Attributable Risk. Disease rates can also be between exposure and disease among subsets of the compar~cl in another fashion: by looking it the study population which have been categerized differ=nce in the rates between exposed and non- according to the level of some other co-v·riate. In · exposed groups. This measure is referred to as · study of breast cancer, for example, it may be found simple ·ttributable risk. The simple attributable risk is that a late age at lust birth has · positive association used to quantify the risk of disease in the exposed with breast cancer, and also low parity has · positive group that can be considered attributed to the exposure. association. The investigation may suspect that ·ge at Using the smoking data presented ·beve, B-16 Heallh Effects of F. xposure to Powerline Frequency Electric and MaSnetic at·lids Rate differ~z~ = (190/1000GO/year)- (19/10G(X~/y~r) dim. These data are then used to calculate rates - 171/IG000/yur which are compared between groups. In the case- control study, for example, the investigator identifies a For this age group, the cxccss lung can,T rate among group of people with the dim under study (cases) smokers attn'butable to smoking is 171/100000/year. and another, presumably comparable group without the Expressed as a pcrecntage (171/190 x 100%), it would disease (controls). The investigator then seeks to be said that 90 percent of lung cancer among smoke· assess and compare the exposure history of the cases can be attributable to the fact they smoked and 90 and controls. The goal is to determine whether there is percent lung cancer among smoke· could be a higher proportion of those who are exposed among e'hminated if they did not smoke. the cases than among the controls. The goal is to determine whether there is a higher proportion of those Population Attributable Risk. Another more useful who are exposed among the cases than among the menure of attributable risk is referred to u the controls. If such a difference is found, its significance population attributable risk. It provides ·uscful must then be evaluated. measure of the propeRion of disease which can be explained by the exposure under study. The population A difference might be found for several reasons: implications since it provides an estimate of the · There may actually be an association between potential impact of a preventive program. For the exposure and the disease; example, if · population attributable risk of · disease is 40 percent for a factor, it is implic~_,__,~_ that 40 percent * The difference (association) may be the result of the disease can be attributed to the factor, then of bias m the study; eradieation of that factor should result in the eventual e'lunination of 40 percent of the disease. * The difference may be due to confounding; or Although · number of methods have been proposed for · Thc difference may be the result of random calculating thc population attributable risk, thc most variation. useful ("Levin's formula for the population attributable risk") is as follows: Unfortunately, it can never be proven that confounding and bias do not exist in a study. The best that can be PAR = P(R-11 P(R-1)+ 1 done is to consider the potential sources of confounding and bias, and attempt to show that they are unlikely to exist or to have strongly influenced the study results. where, P = the propeRion of the population exposed to the factor, and R = the relative risk (or An association between some exposure and a disease odds-retio if derived from a case-control study). does not automatically demonstrate that the exposure causes the disease. It may be that the disease causes As an example, assume that 40 percent of the the exposure or that some other unknown factor causes popuhtion is exposed to some agent (this is beth the exposure and the disease. approximately the propoRion of smokers in thc U.5.) in which the x~lative risk (or odds-ratio) has been Before one attempts to judge the causal nature of an determined to be ten (10). Then: association, it is necessary first to determine whether the difference in rates (indicating an association) is PAR = .0.40(10-11 x 100~ = 78~ likely to be real or just due to chance. Because of 0.40(10-1)+ 1 random variations that arise among population samples, the case and control groups are unlikely to have exactly Thus, by eliminating this paRieular exposure, 7g the same proportion of exposed and non-exposed percent of the disease could potentially be eliminated. persons, even when there is no association between the exposure and the dim. The question, then, is how B.6 Statistics: p-Vaiuo$, large must · difference be to show convincingly that it C, ori~dOrics Iritorva[s arid is real and not due to chance. Two approaches are ~jgrii~Carics commonly used in epidetniology to determine whether a difference is likely to be the result of random chance: use of significance testing by the calculation of p- The epidemiologlst can use a variety of approaches and values, and use of confidence intervals. Both will be sources of information to obtain data on exposures and briefly described and an example shown. other factors that may be related to the development of ~vWsax B B-17 Significance testing Considcr· case-control study which is investigating whether the proportion of thosc with · specific The vast majority of published epidemiologic studies exposure differs significantly baween the case group have statcnnts declaring that the diffcrcnces observed and the control group. The common statistical method betwccn the comparison groups arc (or arc not) for comparing proportions (or _percentages)is thc X2 2 'mtisticaHy significant.~ Unfortunatcly, much (chi square) test. To apply the X test, it is helpful to confusion exists, particularly in thc non-scic~ arrangc the data in the form of a 2x2 table. To conununity, about the mcanlng of · "statisti~y demonstratc this, sssumc that thc study involved 100 signifacsnt" result. Such · finding is one that, cascs and 12~ controls. Thirty-~vc (3~) of thc cascs according to certain mumptions and basccl on · were exposed to the agent, white 7.5 of the controls mathematical probability, has · low likcllhood of being were exposed. The table would then be constructed as due to random sampling variation (chance). follows: ~__ cons The probability of this difference being due to chance is expressed as a "p-value" and my be given in either Inqmd 3S 2s 6o decimal or percentage form. Its magnitude will depend N,aP. aposd 6S X00 on · number of factors including the size of the TreSs ~00 aZS US diffcrence observed, the sizc of the groups being compared, and sometimes, thc data variability. The X2 tost statistic is deftned as: In practice, the most common p-value used to label a finding u statistically significant is p <.05. if a (O.E)2 differcnco in a particular study is large coough that X2 = there is less than a five percent chance that it is the E result of random variation, then this finding is considered significant. Anothcr commonly cited where O = observed number and E = expected threshold of significancc is p < .01, less than a one percent ehancc that a'diffcrcnco is due to tandem number, and the sannmation is over all cclla of the table. Usually, a correction for variation. continuity is made by subtracting 0.5 from IO-Ei as shown below: In somc published reports, an exact p-value (c.g., p= .03) is provided, whilo in othen, the investigator only stabs that "p < .05," or timt the findings arc significant at the .05 !cvcl. Exact valum arc, of coursc, more informativc and allow thc rcadcr to apply his or (IO'Er'0'5)2 hcr own standard of statistical signifr. anec. X2 = E The lovcl of significance sclocted depends on how strongly the investigator wishes to avoid thc crror of To calculatc the X2 statistic, the "cxpectcd" number for coneluding that there is · real difference when none cach ccll in thc tablc must be found. This is donc by may actually exist. This is referred to as ·"Typc I" usuming that thcre is no usociation ~ exposure crrer. Thc smallcr thc p-value, thc smailcr thc chancc and discnsc, and making use of thc marginal totals in thc tablc. The expcctcd valuc can then be calculated that · Type I crror exists. with the following cquation: Calculation of i>-value E = (low total)xfcolunm total} There are ·numbcr of procedures for calculating p- grand total values depending on whether one is considering, for examplc, thc difference bctwcen thc means of · To illustrato, considcr the uppcr-lctt cell of thc 2x2 continuous variablc (c.g., agc, yesn of cmploymcnt) or tablc shown above. Thc row tablc is 60, the column · catogorical variable, such as the difference bctwccn total is 100, and the grand total is 225. Thus the thc prepertions of those in cach group who havc somc expected number of exposed cases is E = (60 x charactcristic (c.g., residencc near high-cuif~n~ 100)/225 = 26.7. This calculation can be applied to distribution lines). These formulae and their each of the othcr cclls of the table to obtain the other applications can be found in great detail in any statistics expected values. The expeetod (E) numbers thus tcxt. Only oue common exaraplc is pmvidcd bclow. obtained are shown in perentheacs below: B-18 Health Effects of F. xposure to Powerline Frequency Eltctric and Magnetic Fields have the disease than those without the exposure. If the discue and the exposure arc not associated with Cas~s Controls Totals each other, the risk will be unity. Valuee lees than 1.0 indicate a negative association (protective effects), Ex~ 35 2~ 60 while values gre_~_t_e~ than 1.0 indicate a positive (26.7) (33.3) association. If the confidence interval for a relative Net Ex~ 6:$ 100 16:$ risk includes the value of 1.0, then no association is (73.3) (91.7) considered to exist between the exposure and the Totals 100 125 225 disease. As an example, consider the same following data from a case-control study: Using the simple formula for X2 gives: X2: 6.39. Usixlg the corroctcd fornlula gives a value of: X2 ~ Cases Controls Totals 5.65. Expo~d 3:$ 25 60 Net Expomml 65 1430 Once this statistic has bocn ealculated, the inveetigator Totals 100 125 225 can then refer to a table of X2 distribution, found in most statistics texts. These tables show the probability of obtaining a given X2 statlstic (the larger the value, the lower its probability or p-value). In this case, with In a case-control study, the odds ratio (OR) quantifies a simple 2x2 table with one degree of freedom, our the strength of the association as an approximation of corrected X2, which value is 5.65, yields a ~-valuc of the relative risk. In this example, it is calculated as approximately 0.0175. follows: Con~fi~tm~t ]~,a~ OR: 35 x 100: 2.2 25x65 A second, but relatai, approach for assessing the significance of epidcmiologic results is the use of Thus, thc~oi~ arfm~ of the rciativc risk is 2.2. confidence intcrvala. The principle of confidence intervals is very straightforward. A certain degree of The calculation of confidence limits will then show the chance variability occurs in sample data. The amount precision of this estinmte. Commonly, 95 porecat of this variability depends on such factors as the sample confidence limits are used. This means that there will size, the inherent subject-to-subject variability in a be a 95 percent chance that the true risk ~ be measured characteristic, and/or the prevalence of a contained in the interval. The first step of the characteristic. Therefore, the observed value of semc procedure is the calculation of the variance and the characteristic of the sample (e.g., means, risk ratio) is standard error of the risk ratio. For the above data, a only an estimate of the "true* value. However, more number of different formulas can be used. The credence can be placed in some estimates than in standard error is then used in calculating the lower others; i.e., estltnates range-in the vernacular-from a (LL) and upper (UL) confidence limits. In practice, *baH-park figure'to a "gnat's eyebrow." some of the simpler formulas provide results quite similar to the more complex formulas. A complete One way to express the precision of a sample eetlmate description of these calculations can be found in any is to calculate houndaries between which the true value standard epidcmiology textbook. is most likely to fall. These upper and lower limits, which bracket the estimatai value, arc referred to as In this example, the variance for the natural logarithm co~.~ i.~rv~l is the range of values from the (in) of the odds ratio dill be calculated as the sum of lower cenfidence limit to the upper confidence limit. the reciprocals of the uells in the 2x2 table. Confidence intervals are very useful in showing the precision of sample estimates. The larger the confidcoco interval the leas precise the eatimate is. Variance of In OR -- __1+ 1__+ _1_+ _1 In cpidcrniologic studies, confidence litnlts arc a b c d frequently calculated for risk ratios. A major purpose is to determine whether the interval around an estimate -- .J_+ __1+ _~-I- or the relative risk includee unity (1.0). Recall that a 35 25 65 100 relative risk of value "x" is interpretal as meaning that subjects with the exposure are x times more likely to ~ 0.09396 Appeadix B B-19 S~andard Error (SE) -- vkrianee -- 0.09396 -- 0.3065 Study Conclusion n. True Relationship Asa~iation No Association The log~'ithm of the lower and upper 95 percent con~tk:az limits is then: Aaaa:iation C. orr~t Type I Error No AMoslotion Type H Error Con~t in Lt- =: in OR - (1.96 x SE) = 0.1877 in UL = in OR + (1.96 x SE): 1.3892 A relationship exists between Type I and Type H errors in thaZ the more closely one guards against a Type l Converting from natural logs of these exponential error, the more likely · Type H error will occur. in values to obtain the actual lower and upper limits can other words, the greater the difference requires for a easily be done on most scientific calculators: conclusion of · ~ association, the greater the chance LL = ·0'1877 = 1,21 The probability that · real association will be detected UL = e1'3892 = 4.01 in a study is referral to as the power of the study. It is determined by subtracting the probability of · Type II Thus, the 95 percent confidence interval for the odds error, expressed as B(bcta), from 1.0. Although not as ratio is 1.2-4.0. This interval shows several things. commonly discussed in study reports, the pewor of · One is that the observed odds ratio (2.2) is not · highly study to detect a particular level of risk is extremely precise estinmte. The actual risk could range from · important beth in the planning of · study and in the very weak 1.2 to a moderately high 4.0. All we know evaluation of "negative" studie~,. Most recent is that there is · 95 percent chance that the real risk epidemiologic and statistical texts provide dinSled ratio falls within this interval. disousslons of power and Type I1 errors for various types of studies. Rather than review these methods for Another important aspect of this interval is that it does calculating power or sample size, this section will only not include one (1.0). This is expected since the briefly outline the major betors that determine power significance test (X2 teat) lad ·trendy revealed that an and then show one example. usecia~on exists between the exposure and disease. Four factors ·fleet the probability of · Type I1 error Many epidemiologists believe that · confidence intorval and, therefore, the power. The first is the level of conveys more information than · significance test confidence used to reduce the chance of · Type I error. because it indicate~ the lowest and highest likely true If one de·ands an extremely high level of confidence relative risk (or other panmeter). it ·!so reveals (i.e., · very small p-value), · greater chance exists that something ·bout the precisinn of an estimate. An · true association will not be recognized. extremely wide confidence interval, whether or not it includes a risk ratio of 1.0, suggests caution in Another factor that ·fleets the power of a case-control interpreting the results of· study. study is the pt~,-valence of the exposure among the controls. Power is reduced where the exposure is either very rare or very common among the controls. in · cohort or cross-sectional study, power is mainly B.7 Statistics: Type II Errors and dependent on the "expected" number of diseased casu in the unexposed population, which is in turn related to Power sample size. (Sample size is discu_-_,e~_ below.) The discussion in the previous section only considered Obviously · study has · greater probability (power) to techniques to avoid · mistaken conclusion regarding the detect large relative risks (or large differeneu in means presence of an association (Type i error). Less or proportions) than snail risks. A particular study frequently discussed is another type of error that may may have a 90% elmnee of detecting · relative risk of be equally important: type H error is the chance of 3, but less than a 7,5% chance of d~.~ting · risk of producing · conclusion of no association when one 1.5. actually does exist. Just as chance can operato to produce an apparent association when none actually Another factor that ·fleets the power of · study is the exists, chance can also obscure · difference that does sample size of the study - the number of cases and exist. Failing to detect · true association is referred to controls in · case-control study, or the number of as · Type H error. The possible combinations of people and period of follow-up (person-years) in · correct and incorrm conclusions ere: cohort study. Obviously, the larger the study, the B-20 Health Effects of Ecposurt to Powerl~ne Frtqutncy Eltctric and Magnetic Fields greater its power to detect a particular level of risk. Thus, depending on urnpie size, the exposure Sample size is also quite important in Type I errors. prevalence estimates, and use of 95 perca~t confidence Only by increuing sample size can the investigator level for significance testing, an 80 percent clanco simultaneously reduce the chance of both types of exists that this study would detect st relative risk as low error. Because of the relationship of sample size to the as 1.9 to 2.4. If the true risk wore higher than 2.0, the probability of these errors, and because sample size is power would be higher than 0.80. However, ff the true one of the conditions that is under some control of the risk were act·ally small (e.g., 1.5), the power would be investigator, considerable emphasis is placed on sample much less than 0.80. size in the design of studies. If the investigator considered it important to detect a Unfortunately, sample size questions cannot be risk of 1.5 with high probability, the study size would answered without consideration of many of the factors have to be increued. If the number of case, could not discussed above. For example, an approprige ample be further in~reued due to cost or availability, ample size will depend on: how small · difference (or bow size and power could be increased by inereuing the small · relative risk) the investigator wants to be able to numbor of controls. For each cue, 2, 3, or even 4 detect, the desired confidence level (p-value) to be used controls might be selected. Little further improvement in significance testing, the power desired to _ae~_._-~__ · in effmiency, bowever, can be gained beyond a ratio of given difference or risk, and the expected prevalence of about 4 to 1, controls to casu. the exposure in the general popohtion. The above calcuhtions could have been determined for In many situations, sample size may not be readily a power of 0.70 or 0.90 or any other desired level. under the control of the investigator. In an historical Conversely, one could have been selected various cohort study, for example, the sample size may be relative risks (e.g., 1.5, 2.0, 5.0, 10.0), and then fixed by the number of individuals employed and the determined the power to detect that risk. years of follow-up. In a ease-control study, sample size may be determined by the number of new cases An understanding of the rehtionships between power, diagnosed over a set period of time at a hospital or ample size, significance tests, etc., is quit-- important through · disease registry. Time and costs are always both to the design and evaluation of epidemiologic practical constraints. Power calculations can be rueareh. If a proposed study duign requiru a higher important in these situations by showing, for example, power to deRct a certain magnitude risk which is the power of a proposed study or a comp!~-,-vJ- study to considered important, then a hrger sample size should dctect specified levels of risk. be sought. Of course, the power of · study should To illustra2 some of these interrelationships, the always be taken into account in evaluating a report of following example is taken from an actual study negative findings, i.e. was the study sufficiently powerful to give the reader confidence that the author's proposal. The proposed projea would be a case- control study to examine the possible association claim of a negative study is valid? between leukemia and ruidential exposure to magnetic fields. Based on the population of · study area, expected cooperation rates, leukemia rates, etc., the investigators estimated that they ultimately would be able to compare 163 leukemis cases with 163 controls. Significance testing would be done at the 95 percent confidence level (p <.05). The investigators then determined the minimum detectable rehtive risks (odds ratios) that they would have an g0 percent chance of detecting (i.e., power = .80). Since they did not have reliable data on the prevalence of magnetic field exposure (however defined) in the general population, several estimatu wor~ considered bard on the findings of an earlier study. The following estimated minimum detectable risks were price·ted corresponding to exposure prevalence: Provaloneo of Expomaro 0.10 0.15 0.20 0.25 0.30 Minimum of 13~t~Utbi~ Risk 2.41 2.15 2.02 1.95 1.90 Appendix B 9-21 References The following works wcrc used in prepsring this primcr, Thc rcadcr is rcfcrrcd to thcm for furthcr information on thc subject of cpidcmiologic mehodolegy, AIrman DG, Gor~ AM, Gardncr MJ, Pocock SJ. Statistical guidelines for contributon to medical journals. Brit~h Medical J ounud 1983; 286:1489-93. Brcslew NE. Day NE. Stnlistical Methods in Cancer Research. VoL 1: The Analysis of Case- Conlrol Studies. IARC Scientific Publication 32. New York: Oxford Univcnity Press, 1980. Bucchley R. Dunn JE, Lindc~ G, Brcslow L. Dcath ccttificatc starcroons of occupation: Its uscfulncu in comparing mortalities. l~blic Htalth Reports 19S6; 71(11): 1105-11. Cole P. The evolving case-control ~tudy. Journal of Chronic Diseases 1979; 32:15-27. Fcinstcin AR,Horwitz RI. Double standards, acientific methods, and epidcmiologic research. New England Journal of Medicine 1982; 307(26):1611-17. Fleiss JL. Statistical Methods for Ralts and Proportions, 2rid edilion. New York: John W*dcy and Sons, 1981. Friedman GD. Frlmtr ofF. p~dtmiologyo 2nd ed/z/on. New York: McGraw-Hill, 1980. ISBN 0-70-022434-X. Gladcn B, Rogan WJ. Misclau~icalion and the design of txperimtnlal studies. Amcrican Journal of Epidcmiology 1979; 109(5):607-16. Gllslgr JM. The quality and utility of dcath ccfiifgatc data (editorial). American Journal of Public Health 1981; 71(3):231-33. Greenberg RS, Klcinbaum DG. Mathematical modchng stntogic~ for thc malysis of cpidcmiologic re, catch. Annual Review of public Health 1985; 6:223-45. Haines T, Shannon H. Sample sizc in occupational mortality. Journal of Occupalioual Medicine 1983; 25(8):603-08. Hill AB. The environment and discuc: Association or cauufion? Proceedings of the Royal ,Vocitty of Medicine 1965; 58:295-300. Horwitz RI, Fcinstcin AR. Mctholodologic vandaub and contradictory ruuits in cue-control ruearch. Amer/can Journal of Medicine 1979; 66:556-64. Kelscy JL, Thompson WD. Evum AS. Methods in Observa~ona/Ep/dtm/alogy. New York: Oxford University Press, 1986. MacMahon B, Pugh TF. Epidcmiolegy: Princevies and Methods. Boston: Little, Brown and Company, 1970. Maretel N, H~cn~zcl W. Statistical upem of thc analysis of data from rctnnpectivc studiu of discue. Journal of the National Cancer Institute 1959; 22(4):719-48. McMichacl, AJ. Standardized mortality ratios and the healthy workcr cffcct: Scratching bcncath the ,urfacc. Journal of Occupational Medicine 1976; 18:165-68. Mon~on RR. Occupational Epidemiology. Boca Raton: CRC Prut, 1980. B-22 Health Effects of F.~posurt to Powtrline Frequency Electric and Magnetic Fidd~ i~r~y C, $tanek E, 61o~x:kim' L. Aexura~y of c. tnc~r c~th ,:~ifiat~ and its ~ff~:~t on c. tn~-r morality varieties. /imtrican Journal of public Htalth 19B1; 71.'242-50. Rim AA. Bas/c B/ostat/st/cs/n Med/c/ne and F_..p/dem/o/ogy. N~-w York: Appleton-.C.~ury- C~roltt, 1979. Rothnan KJ. Modern F_.p/dem/o/ogy. Boston: Little, Brown and Cornpiny, 19S6. ~k~t DL. Bm in tnalyti~: r~gar~h. Jounml of Chronic Discasts 1979; 32:51-~. S~hlesselmsn JJ. Sample size rcquircmcnts in cohort and cag-control studies of disease. Anur/can Journal of F,p/dtm/o/ogy 1974; 99:381-84. SteinLand K, Betumont J. ~ ,~xur~y of occupation ~d indumy dm on de~th ~ific..v~. Journal of Occupational Med/c/ne 19S4;26:288-96. C-1 APPENDIX C - RESULTS OF EMF SURVEY No- ~ C~'~ect Dmc~,,,t~ !. ~ny ~r Sy~m W~ ~ 13~kV ~ Si~ ~- C~- 2. ~ ~er ~, ~ll~avi 765-kV ~ C~n f~ ~t ~e~ a~ ~ ~n adve~ bio~ ~ff~U a~i~ wiffi ~ fields from · is 765-kV ~. 3. ~ ~i ~ ~ ~-kV ~ Si~ wu ~ by ~ ~e C~y bd a~ ~e City 4. ~ ~i ~ ~ ~kV i~ No d~t ~ ~ ~ i~ ~ Co~ ~ Cs~ of t~ Co~a6bility. ~. ~i ~c ~. Mi~, ~a, N~n~ ~j~t ~eiv~ "Fi~ of No Si~cam Tn~ion ~j~t ~, ~5- ~act" from iic¢~ auffio~ty; ~F i~es kV ~ (g~d~ co~b~ we~ ni~ by ~ow~n i pubEc h~s; 1~2) i~ue~y acqui~ ~ of ROW ~gh ~g~on, ~B~ 6~ ~i~6on; IF ~ an co~on 6. / ~ ~. ~pl m ~t ~ ~n nl ~ of~F ~1~ eff~a; ~a~ a ~kV T~ ~ i~ ~ ~ IF ~ ~ff~u; F~ (19M) kV ~ wu ~ ~n a ~ ~ was a~ 7. ~ ~ ~. M~ ~kV ~ ~ ~mv~; ~ f~ ~t fields from ~is lira w~d ~ ~r m fields f~ e~ li~s of ~ v~m~ ~u ~ ~t field ~vels from ~ I~ fall wi~ field U~s ~ New Je~y a~ ~er urns. 8, ~ ~ ~. M~o-~u~o ~vsr 13~kV ~ ~ ~ f~ ~t ~ ~ u~ly aff~t tc q~llty of ~e ~nvi~m ~ ~e h~ of ~e p~iic." 9. ~ ~ ~. C~iff-N~ F~ ~V ~ ~ ~ n~; ~ ~i~ fi~i~s a~ ~ p~. 10. ~ ~. ~, ~ Ce~fic~ f~ 138-kV ~ Sire ~p~ ~ ~n ~t ~ity pmvi~ 11. Ce~ ~ ~w~ Sa~ ~ of ~~ U~y ~ ~n ~ ~ ~ ~ was built, 12. ~ ~ & lJ~t W~~ ~kV ~, Co~ d~ ~ty C~y No. 5~ ifo~ on ~bb ~ffi eff~u of ~i ~li~ wi/d~ ~e a~Eca~n. C-2 Health Effect~ of F.:cposure to Powerline Frequency Electric and Magnetic F'feids Sidng (Continucd) No. ~ r'~'/Proj,ct DelcH_:~.-- .~,--- 13. C4mtral Power & Light Co. Applk:adon fo~ Lon Hill - Cokto EMF ~ raisml; ca~ psnding; no trial dats Croe. k 345-kV i.im sot p~nding vuoaution of ~ndangevsd sp~i, ismes. 14. Chsy~nne Light, Fuel & CoriaO~ 115-kV Lins Commluion found no svidutce of adverse Power Co. health effects. Landowners appealmt dexision to Wyoming S. Ct., ~. Ct found mff~ient r~ord of no health risks. 15. Ckvdand E~tri~ Chair Township v. Powor Siting Oppoaition pmuntod EMF maimony 0wough Hluminating Co. Comm'n (1977) sngin~r, utility produced company gngin~r, lira approv~d~ S. Ct. of Ohio upheld. 16. Ckvsland !~zttic Ohio Powsr Siting Board: EMF iamg was pre~r, ntsd; Board roads no I~minating Company No. 02-00022 funding ~2anling EMF but d~nie~! c=~ficat= on certification Fanled in Febmaryt 1991. 17. Commonwealth Electric Commonwealth Electric Company lames aimed in this proceeding were resolved in Cornpiny v. Energy Facilities Siting Council: Cape Howitch - Dennis proceeding before the F. FSC 85-4A DPU, discussed below. 18. CommonwsalthEkclric Application for 345-kV Substation: Dept. of public Utililies found no evidence that Company No. DPU 86-257 EMF from Ihc substation would inct~asc health risk. 19. Commonwealth Electric ll5-kVTransmiuionljneCape Lin~approvcdon,~acmber28, 1990; ~y Howitch-Dennis Department of Public Utilities found that "[t]he evidcnc= presente41 by the in2rvcnors does not demonstrat= dmt the EMFs produced by dectric power lines caug adverg effects." 20. Connecticut Light & Power Braun V. CP&L Line apptoved; court found llmt tl~re was not "any0dng clog to proof of a reasonable ham one or mole of 0~ plaintiffs"; pcrmancnl injunction denial; landown=rs filed notice of int=nt Io sppeal o~ May 29, 1991; futOer hearinJ on dama2es possible. 21. F]Otid8 Power Corpotmlioe Applicatioo for Cetfificatiofi of Sil~ Board dolli~J corti~cafi<m bagd on failurc Lake Taq~on 5430-kV line, No. 85- of DER to pfi>vlde sundards; ~ replations 141 ! wer~ adop~d & line approved on t~mand. 22. Florida Power Co~ Applicalion for Cerd~calion Board found that ~r. ids from ~he line would havc cugntially no biololical effect. 23. Florida Power Coq, xalion Piedmom-Sorcmo 230-kV Lin~, Board found ~hat ~ from ~ ~ would No. 81-185~ heve no impact. 24. Hawaii Ekctric Lilht Puna-P0ho*d~i Ccni~cation for two Commisaion found insuf~cieng evidence m Company, Inc. 69-kV Lin~s conclude that a ~ risk exims and approvcd lin~. 2.5. Kauai F. kcttlc Co. 1989 Kauai Line Siting Action involving ma2 conaervation lands; Hearings held by Board of Land and Natural Resources Congrvation with expetu on issue of EMF health effects; found imuff~cient evidence of adverse health effecls. Appendi~ C - Results of EMF Survey C-3 Siting (Continucd) 7.6. lemy C4nmml Power & Manitou-Whiting 230=kV Lk~ EMF issues raisgl; ~ approve, d; opposition Light Co. nmlion to stay pendin~ appeal denied April, 1990. 27. Jermsy Csmral Power & Aberdeen-R~d Bank Application withdrawn by utility after Light Co. reas~ument of need for lin~. 28, Lowst Colorado River Application for 345-kV Commiuion adopted Hearing Examiner's ruling Authority Transmission ~ and Associated that there were no proven health effects and $ubslalion (1979) added provision that the commiuion could amend or revoke cefi~cate if fumr~ research shows that exposure to electric fields causss adverse health effecU. 29. Minnnota Power Co. Application for Exemption From Application approved; Board fonnd that Siting Requirements for Dululh proposed upgtade woold result in lower Aria magnetic field levels and that O~rc is "no definitive evidence" flat EMF causes health problems. 30. Minnkota Power Coop. Cemer-Maple River 230-kV to 345- EMF witness appeared; Application grentdgl. Inc, kV UpIrmde (1980) 31. MiuiuippiPower ApplicationforCPCN:U-4128 Commiuion foond no evidence of health effects Company ft~m 500-kV line, 32. Montana Power Company Application for Certification of State agency found field levels acceptable. Laurel-Bridget 100-kV Line 33. Nashviik Ekctric S~rvice Sharondale 161-kV Substation !~oject has been plEed on hold p~nding resulta of independent enginering study of sys2m-wide electricity needs. 34. Nuhvilk Electrk S~rvica Hendersonvitle 161-kV Substation Line approve, d; consreaction subsequently opposed by Sierra Club. Utility n~t with Ic~al citizens group and proposed engineering alternalives to dehy construction of project; alternatives adopted and project dehyed indefinitely. 35. N~braska Public Power Application for Cerlification of Application grunted. Utility decided noi to build Diatrict Corridor Compatibility: No. 9942 line. 36. Nebreska Public Power Application for Permit to Conemit Application gran2d. Utility decided not to build District and Op~rete: No. F-3371 line. 37. N~vada Power Company Warm Springs Eastern ~ EMF raig41; wirecues appeared; approval granted; appealed to County Planning Commission; lost approval for certification; utility rerooted line, 38. Nsw England Electrk Petition for Permission to Build !~parmuent of Public Utilities found that health $ys2m 345-kV Line, No. DPU 19559 hazards were conjecture and had not be~n established. Supron~ Judicial Court affirmed this finding and remanded the case on other issues. C-4 Heahh Effects of F. xposure to Powerline Frequency F.~ectric and Magnetic Fiekts Siting (Continued) No. Utat/ Cse/Pr~jsct i)meriplioa Status 39. New Eaglmzd Etscttic Csrlifrats to Install 450-kV Lins. Commiakm fouad that eftsets on Imblic health, SIslsm No. 1:~1~1-349 if soy, fall wilhin sc~sptabi= ran~. 40. Nsw England Powst CoUllo v. !)~mmns,v of Public Dsl~. of Public Utilitiss al~rovsd 345-kV linc; Coralany Utili6ss (1984) Costsilo challsnpd sti~toval; court dgision to allow ¢stion of ~ and a~eq~gi mility te~jmony that no adv~t,~ health 41. N~w York Power Commo~ Re. cord Hearings b~fot~ l~ndin$; utilities requir~d to fund t~ae. ar~h on Aufl~dty; Niagara Ih~ N~w York Publk ,q~rvi~ th~ EMF Issue. Mohawk Power CouV.; Cmnmiasion. No. 26529 N~w York ~at~ (}u & 42. Not~mst Utilifie4 Farmington-Noah ~oomf~ld 115- ~it~ approvml but de.4:iaion did not addif. as ~h~ kV Line EMF issue. 43. Northern ma~s Power Co. Applications Io Modify and Applications approvod in May 1991~ Board Upgrad~ Transmission ~ found Iiutt !h~t= is "no ¢vid~n~ that incrsascd mgnmic fisids avoclat~! with inctr. assd powsr flow in 500-kV iins will · risk to hunran hcallh." 44. hci~c Gas and ElscUic Cxsyssr Pmjsct, No. 79 AFC-5 Commission found limt th~rs is no p~suasivs svidsncc that advsrss health cffccts rsmlt from hip voics$c EMFs. 45. htcific i~x~nst & Light Eupss-Mcdfon] 500-kV Lins A~plication $ramsd; "COPE" group subsf. qucnay missd EMF isaus pdo~ Io coommctlon; ~ csni~cah. amsndnznts sought for 4 touts chap; EFSC appmv~d August 29, 1990. PP&L will use 'Ddta' Line con~$umtion for EMF mitlplion. 46. !~iladdphia F.~cttlc 230-kV Linc; No. E-81768003 Co~n found I!mt Ihc ptol~ssd iinc is nm Company dan~mus to humans or animals. 47. Philaddphia F,~tri~ Horsham-Middl~own 230-kV Line In May, 1991, $tat~ Court ordered PUC to company Ccnifw. ation reop=n its hearinga to take evidence on EMF issue4. 48. Potomac Ediaon Company Brighton-High Ridge 500=kV ~ Commission found no basis to conclude that power lin= figida ~auae adv=r~ hsallh Rejeeml prt~ah for field standarda which hay= no ,~i~mi~ basis. Stuff will monitor ongoing EMF r~v. ar~h and rq~on on ,emi- annual baai,. 49. Publi~ ,q~rvk= Company of Daniel, Park Tanniaion ~ Board of Commimongrs hdd Ihat Colondo undergrounding did not ~onmimt~ "Prudent Avoidam:e' under the ~in:unmam:e4. Overhead ~ approv=d wilh ~onditiona offgr~l by Utility. State Court ov~nutnai da:ision on grounds urn'elated to Ihg merits of !h~ EMF claim. 54}. Publk ,%rvic~ Company of $idn=y-No~h Yunm 230-kV Lin~ Pending. Colorado Appendix C' - Results of EMF Survey C-5 Siting (Continuexl) N0, ~ Cm./Project Dezripfon Status 51. Puget Sound Power & Canadian T'm-in Pending. Light Company 52. ,qalt River Pro~t Pinnacle Poak-Papago Buttes 230- Cerli~c,a,- thmi~d at I~ast in pmx on basis of kV line. potential for EMF health 53. ~an Di~go Gas & Ele.~lri~ Application for Certification: Commisskm found that available information did Co. D~ision t93785 not indicate that EMF cau~s adverM health eff~ts. Ovared utility to comim~ to fund EPRI studies and inform Commission of 54. Southern Cal. Edison Co. Kren=r-Vi~tor 230-kV Lins CPUC gran~ approval to eonstngt transmission line; required utility to 'minimize' mat, n~tie fields associated with the line. 55. Tucson E~:tri~ Power Boi Air Ranch Eatetes v. TEPC; Court found Ilat plaintiff bad failed to prove Company No. C458986 adverse health eff~ts by a prq~nderance of the evidence. 56. Union ~ Apache Flats 161-kV Litm Line approved; de~ision made no reference to EMF. 57. United Power AMo~iation Bonto-Milm:a 230-kV Line: Do~ke~ Board found that evidence to date did not No. UPA-TR-! indicate transmission lines as ·health hazard. 58. United Power Auo~iation Coal Creek-Stamen 230-kV: Docket Commission found that designated route would No. 9593 have acceptable health and environnmmal impacts. 59. United Power Asmglation W'dnmrth 345-kV Line: Docket No. Council found no substaraial showing of adverse CU-TR-2 biological effects. 60. United Power Association Certificate of Site Compatibility; CommiMion found no evidence of adverse Docket 9459 health effncts. 61. Unitmi Power Association Route P0rtnit 400-kV DC Lin0: Commission found no detzimam to human health Case No. 9370 from the line. 62. United Power Asso:iation Construction Permit 400.kV DC Council ~und the line would have no adverse Line: Do~k0t No. CU-FR-I health eff~ts. 63. United Power Auo~iation Corridor Designation: W'dmarth Council found no tq~otu of adverse health 345-kV Line~ No. CU-TCI effects from EMF. 64. United Power Asso:iation Corridor Designation: 4OO-kV DC, Council found that proposed line would have no No. CU-TC-I adverse health effects. 65. Vermont El~tric Power 450-kV DC Interconnection. 4763 Board found that potential for adverse health Company effects was minimal or nonexistent. 66. Virginia Power Loudoun- Oainsvillo 230-kV Litg Slate Corporation Commission de~ided to monitor tim EMF health effects issue. 67. Virginia Power Ox-Poseum P!m 500-kV Lins State Corporation Commission do~ided to monitor the EMF health effects ime. 68. Vh~linia Power Co. Elmont-Chi~kahominy 230-kV ~ EMF exlmrte aplmami; Hearing examiner found (1990) no EMF health effe~ts~ line appmved. 1. ~ ~ 115-iV N~ ~m~ ~ No~, H~ ~fo~ ~ ~; ~e~ ime ~F; ~A e~m ~ f~ utility, ~ f~ ~j~; ~o~ h~fi~ ~h~ul~ ~d~r 1~ on ~F h~l~ eff~. 2. ~e ~er ~ $pfi~eM ~ ~t H~ ~ ~ ofROW; ~F ~, mi~; Adminlm~ (5~kV t ~) ofe~ ROW f~ 5~kV ~ ~i~. 3. Ce~ H~ ~ ~m~ MaWr ~F ~ mi~ by ~ ~; ~p~al 4. Con. ~di~ ~ St. S~-~ Up~ ~i~. 5. ~ R~s ~n C~n~ ~t 69- H~ ~fo~ D. C. ~ of ~ ~is kV ~s s~!~ ~F ~ny; ~ ~val 6. ~o ~sr Sun V~ky Tn~ ~ Sit~ ~ a~ ~q C~zil, ci~q (1989) c~, ad~ "~1" ~ ~t ROW mug ~ ~ ~ 150 fm f~ 7. ~~s p~ 161-kV ~ ~t ~b~ cohere ~ ~ ~fo~ ~; u~lity v~um~l~ ~. 8. ~ l~ ~l ~. O~t N~k Sub~,~ F~-~ ~F imes mi~ a~mvai ~i~. .9. ~ ~ ~t Co. ~br ~, ~-~1 ~lc ~F ~ mi~; ~r ~ cable. 10. ~h~ ~c Co. ~liu~ m C~ct a ~idne witNt p~j~ice a~r hu~s dufi~ Tafo~r ~ M~u~, MA te ~ p~m we~ held ~ ~F ~il eff~a; c~ey is ~oa~c~ dcdp ~cr imcs. 11. h~ & R~ M~dc~m~ Ave. S~m~n ~pmv~. ~F ~y p~e~. No ~on U~i~, ~. of~F in ~ml ~er. Appendix C- Results of EMF Survey C-7 Zoning (Continued) No. Uttt/ Cu~Projeet Dese~ 12. Pacific Paywet & Light China Hat-south Loop 69/115-kV Public hearing re line including EMF and prope~y C0oVmny ~ value ilaues. Heating Officer foond that causal relationahip between EMF and health had not been demonstrated, and that ttansmisaion line would have minimal if any impact on property vslues. Linc approvsd in hte 1990. 13. PEPCO Btighton-ltigh Ridge 500-kV Line Howard and Montgomery County Boards are pv~mpmi from guing EMF standards once State PSC has suthorized iin~. 14. Public Service Gas & Eagle Point 230-kV Fe, t to N.U.G. Hearing before Wsstville Environmental Planning E~egtric Board; ~ withcases; utility modified zoning spproved. 15. Public Service Co. of New ! 15-kV Li~ in Santa Fe EMF reismi; approval i~nding. 16. Sierra Pacific Power Csreon City 120-kV Line Approved. E]~F leMhBotiy pfggtited. No Cc~-~vany of EMF in Fatal Order. 17. Tri-$tate GiT Newton, Johnson & Kamyski v. Utility withdrew application slter District Court Grand Co. Commis. & Me. Parks ordered new examination of the EMF issue. Utility EJectriG !z.r No. 86-CV-225 built distribution line. C.3 Condemnation No. ~ c,.. No"" Ststm 1. Allegheny Power System Allegheny v. Ramsburg (138-kV) EJdF reigd by landowner in condcnumtion ~ r~li.oa Co.) t ;,,~ preceding; casa satfled. 2. Arizona Publk Setvies AP~ v. $happard; No. C-444340 No recovery; utility presanted brief testimony (500 kV) on the EMF ism~; jury did not award any ¢omp~naation for fear of adverg health 3. Arizona Publi~ Service AP~ v. Sel~tive Resources Court held Oat temimony on EMF is irreJevant to a condemnation action and therefore inadmissible. 4. B!udxmn~t Ejo~tric Coopsretire BBEC v. C. A. Colhoun 138-kV Digevery controversies have delayed trial. 01BEC) T0xaa ~ Hearing June 1992 to re0olve digevery issues. 5. Brezos Elexu~ Powex Brazes EJe,~tri~ v. Thelee Ray Utility offered $42,000 for 4.2 acres. Cooperative, Inc. Landowner claimed damages of $1,000,003 to remaining property. Jury awarded $77,000 for 4.2 acres. No mention of EMF in fmal order. 6. Brazes Ej~tri~ Power BEPC v. Maddie 138-kV ~ EMF isau~a reisod; Commisaioner's Court Coopsretire, in~. 01EPC) R.O.W. awarded landowner $11,800. On appeal, Jury awarded landowner $10,000. No mention of E~F in Final Order. 7. Brazes Electric Power BEPC. v. MeAllure 13g-kV EMF issues raised; Commissioner's Cooperative, Inc. (BPEC) Line R-0-W Court awarded landowner $6,900. On appeal, jury ,warded landowner $3,325. No mention of EMF in Final Order. C-8 Htalth Effects of Exposure W Powerline Frequency Eltctric and Magnetic F~kts Cond~nnution (Continued) No. Utility Case Name Status 8. Control Eiectric Power Coop. Condenmtion Cue Appraiser repo~ raised EMF and fear imam½ case settled out of court. 9. Central Kanm F, kcUic Coop., Petition of Central IOnhum F, lectric Supreme Court of Kams ruled parties' Inc. Coop. tnc. (1978) electrical enginggring tcMimony of EMF health effects should have begn with opinion. 10. City Public Service, San Antonio Stone Oak cond~nmation; 345-kV Lowar Court found that property was Line worth $100,0(30. On appeal landowner raised the EMF issue. Appeals Court awarded $97t000 for the prt~erty. 11-17. City lkd~!~¢ Servicaa Sen ,t_n_,n_nlo Seven CO other Condemnation Cases Seuled out of court. 18. Colomdo-Ute Electric Aaan. Lainour v. Colorado-Ute (1986) EMF ime raised by Lainour, but never argued in court; no witn~s~s appear~l; settlement reached betwc~n parties on issue of valuation. 19. ~ Powar Co. ~ Powar v. Elizal~th Oran~r. !~IF witn~ues ap!~ared; no ~ Cond~nmation Finding; p~o~lin~ s~ttled by consent jud~n~nt. 20. Florida Powar & Light Co. Fla. P&L v. Jannin~s Court ruled that health affects t~etimony is irrelevant in cond~nmation action and thatafore inadmiseiblc. 21. C_~otgia Power Compuy C~oqia Power v. liraroy C. Baff~tt Landowner has raised i~!F as an issue with respect to land values of parcel hain~ ecquircd~ pending; no trial dat~ ghcduled. 22. Goorp Power Company Mystery Vallcy Subdivision v. EMF health cffccts ism~s raised by Georgia Powsr landowners around the subMarion; axpsns have hasn identlfmsd; EMF claim not buin] actively pursued. 23. Gulf Stare Utilities Company Counnsy v. Gulf Stars, No. 11058 Settled for amount much lau than that originally, claimed. 24. Hi Lighting & Power KJain v. HLAP, No. 395-755 Utility required to pay $104,275 in 25. Idaho Power Company IPC v. Lcmanich, No. 6828 (230-kV Landowner claimcd $230,000 in Line) dameget; Utility offered $14,700. Jury awarded $15,000. No special findings, unable to dctcrmine whcthcr tha EMF iuue affected award. App~lix C - Results of EMF Survty C-9 Condcmnation (Contints:d) No. Utility Case Nams Status 7.6. Zndiane & Michigan E:lectri~ [&MEC v. pO,,_nd% No recovet3r~ court refueed to decide the C_n~na~ny No. C-790235 (765-kV Line) EMF isme. 27. Iresmats Power Co. (Iowa) Consknmstion Action 345=kV Rock Landowner raiae~i adverg health effects; CrMk Tmmni__~,on ~ (1985) hearing b~fore lows Board of Public Utilities; EMF experts ap!~ared on both aides; case was Ruled and voluntary eases janted before opinion was issued by board. 7.8. lows Power and Light Company IP&L v. Stonsniss:imr, st al. No recovery; court exci,_,_A~__ exl~rt tinsimony chiming EMF health effects becaug there was insufficient data to reach · cosglusion that the ~ would caug health 29. LaPlata Electricity Amo~iation LaPlata Ekc. Ase'n v. Cummins Lsndown~n submined public opinion survey (1986) r~ desire to live naar power lines due to health hzsrds of EMF; appellats ruled that admission was not r~versible error since no weight given to it in trial judge's award. 30. Los Angeks Dspanmsnt of City of LAv. Fmk, LA Suparior !ury msd~ special finding that there was no Wstm' &Power Ct. No. C307309 dimimstion in valu~ of property due to sdverso biological effncta of the 500-kV Line. 31. Los An~ks lkparmmm of City of LAv. Van Dotwsmolsn, No spociel findings; usable to det~rmiue Watar &Power No. C606776 whether the EMF issue affected award. 32. Louisiana Power & Light Louisiana Power & Light Co. v. Court of Appeal of LA h~ld it was Mobiey (1986) petmiasible for trial judge to force LP&L to pay for danage to land if danage was caused by fear of EMF adverse health effects. 33. Louisiana Power & Liaht LP&LC v. Zeringus, Court did not d~id~ issue of EMF health No. 53, 164-E effects but noW. dthe hndown~rs fear of such effects. 34. Louisiana Prayer & Light LP&LC v. Churchill Farms & No stMcial findings; unable to determine Msrcello~ No. 184-546 whsther the EMF issue affected award. 35-45. Lower Colorado River Authority Eleven Cond~nmstion Actions Involved EMF-related issues; thrM cases actually went to trial; LCRA prevailed in each; testimony was taken in each r~gsrding EMF ismes, but no EMF/health effects fmdings were nade in the final orders. O~hers eilht cases settled out of court. C-10 Health Effects of F. xposurt w Powerline Frtqutt$cy Eltctric a~ Magneti~ Fwld~ Condemsm~on (Continued) No. Utilit~ Case Name Slams 46. Mississippi Power & Light MP&L v. Thomas, No. 3823 (500- No t~ov~ry; ~outt found no ~onvin~in~ kV Line) ovide, nc~ of harmful ~ffe~ts on 47. Miseouti !*ubli~ Servi~ Com~mnalion Case (131-kV EMF issuea raised in discovery; Transmission LiM) landowners ar~ pursuing !h~ory that fear of EMF will ff, duQe ~he im:on~ flow of the property; ease pendinl. 48. MoaUma !s~v~r Co. 161-kV ~ R~ia~i EMF isaues raised informally; 49. N~w York Power Au~otity Zappaviina v. PASNY, No. 74095 No teeovary; ~ou~ held there was no teaaonnble basis for f~ar of health efteels; bosh sid. appea~d; $ppol 50. Nsw York Power AuShor~ Jmas Condemnation R~laled ~ase to Zappavigna condemnation~ landowner etlempled to call Dr. David Carp~n~r as EMF wilne$$; Power Authority successf~Hy objeeled to inlroduclion of his lestimony. 5 |. NotO~m Virginia ~ C_,~dar Ot~v~Evs~ 115-kV EMF 'fsar* ism~s have been rai$~l; no Coolmmtlve Lino Co~S~m~oa ~ ~ m. 52. No~hsm S~gSss Powsr Company Hennapin Coomy Disui~t Ct. No. EMF ismgs raiaed in cood~nmntlon but CD-2G07 msbssqu~nlly wef~ dropped by 53. Philad~phla ~ ~y Goadby v. PEC No rzov~ry; plaintiffs claims could 54. Poton~ Ei~tti~ Power Co. Brighton-High Ridge Condenmation ea$e$ have begun; to (PEPCO) Condemnation C$g$ dale courls have excluclad evidence of EMF or. fear of EMF. 55. Potom~ Edison Co. __s3,~__~,ngStm-W.C. Rsid Line EMF i$$ue0 raised by landowner in condemnalion actionl case settled. 56. Publi~ S~rvi~e Co. of Colorado PSCo v. A~kemmn Ackerman Case settled when PSCo agreed to upgrade from 115-kV to 230- kV inst~d of 345-kV eliminating to tequirs additional ROW. 57. Publi~ ~rvie~ Co. of Colondo PSCo. v. MaHoney (1995) Award in Mahoney did not EMF $hough it was raised by 58. Public Servi~e Co. of Colorado Linn~ur v. PSCo (1986) Landowner $ooght to introduce t~atimony of heallh efteels expetU to support claim of damages; cona ruled haalsh effKle testimony inadmiaaibi~; no damales du~ to EMF. Appendix C - Results of EMF Survey C-11 Condcnmation (Conl~nued) No. Utilit~ Case Name Slams 59. Public Service Company of PSCo v. Judge Despha ~simony on th~ EMF issue, t'nln,~lo decision did not address the EMF issue. 60. Public S~rvic~ Company of PSCo v. Higgs Pending. 61. Public Service Company of Unknown. No r~overy; EMF evidence was t,a;,,,, exc!'--~-'~- by xrial jud2¢. 62. Public Service Co. of New 1980 Condemnation Case Electric field issues rei~w,,d; judge did not Mexico allow plaintiffs' wimeu to testify as EMF 63. PMblic Service Co. of Oklahoma Bud Clangon Condemnation 045- EMF raised through appraiser, landowner kV) (April 1990) c_ui,,,,~l $500,000; jury awarded $34,G00; no men6on of EMF in decision landowner appealed, but later druppcd the appeal. 64. Sen D/ego Gas & Elecuic SDG&E v. V&V Development Testimony on EMF was presented; no Company Company, No. 484042 special findings and no method of determining whether EMF affected amount of compensation. 65. San Diego Gas & Electric SDG&E v. Anderson, No. 487321 No recovery; coon rulad that EMF health Company effects ara mo speculative for the jury to consider in · denmass award. 66. San Diego Gas & EMcL,'ic SDG&E v. Tallafcrfo, No. 489761 No I!)¢ciflc filldiDgl; unable Io delefl!lhg whether the EMF issue affected award. 67. San Diego Gas & Electric SDG&E v. Ddey, No. 403834 The Court ruled that the utility could not present testimony on the EMF issue because it is irrelevant in · condenumtion action. 68. Santee Cooper MytOc Beach Condemnation Case No EMF witnesses; EMF first raised in closing argument by landowners; appeal 69. Soulhem Cal. Edison Company So. Cal. Edison v. Slaffofd, et al, No recovery; jury foond no proof No. INDIO 7.8701 property values had been diminiahed by fear of health effects. 70, Sou/wear Atkanms Utility SWAUC v. Black, No. E-83-8-G No tecovcry; coun found that 161-kV CosFt~f, iinc would not create a hcalth hazard. C-12 Health Effects of F. xposure to Powerline Frequency Electric and Magnetic Fields Condemnation (Continued) No. Unillt~ c.,, F*"" Slams 71. Sooths~tsm Publi~ ,q~rv~ Matador Cauk Co. v. SWPSC No ~xpli~it riding on health eff~ls of Company EMF; co~t~ m~d that no tr=spau damage had been caused by EMF. 72. Soulhwmsm Publi~ ~ Tu~o-Oklaunion 345-kV ~ EMF health off~ts issue raised but no Cc,,~,p~y sxpet~ wim~u t~stimony presented. 73. T~nneaa~ Valley Authority TVA v. An Eagn~nt No re~ov~ry; ~oort rui~l Oat EMF tutimooy is nm r~i~vant in a ~ond~mnation actlon. 74. Texas Utili6es ~ Company TESCO v. B~r~r, No. 85A-216 No r~overy; opposing anotney failed 75 Tu~stm ~ Power Company TF.,PC v. ~pis, No. 188972 No sp~if~ findings; unabls to d~rmin= wlM~:r EMF affe~tsd award. 76. Unitsd Power AMo~iatioo CPA/UPA v. Aaaami, No. C-1474 No ap~ifi~ findings t~garding lhe EMF iuu~, urmbl~ to determine whether EMF affected award. 7/. Virginia Power Co. Virginia Power Co. v. FendIcy (1982) Commiuion~ru ruled iandowner's ~xp~t~ withes did not qualify as an ~xpm~ on EMF; ~i~uit ~ou~ affirmed d~ision following iandown=r's filing of formal ~x~q~ion to ruling. 78. W'~onsin Publi~ ,~tvi~e Crop. Sturpon-Bay Condemnation Cags EMF raised; No finding of increase du~ to EMF; oo~ appeal md~l out of C.4 Tod 1. Clark Co. P.U.D. No. I Slt~ila Amierso~ v. P.U.D. Clark P~rsoml injury and wrongful death action by (%Vadainitoa Sla~) Co. n~r of deceased child (lcukemia); fields in mchool marewed; case di~mi~ed without prejudice. 2. FIo6da Power & Light Chan~y v. FP&L, Palm B~a~h Adver~ he, allh offers and numa:e from Cir~ult Coort ~ elaine!; plaintiff I~eenlly fi~ed mo6on to d=!ay trial; no de~ision on mo6on or trial 3. Florida Power & Light FantiJmui v. FP&L, Browaid Adverse health ~ff~cts, trcspau, nuiuncc C_na~y Circuit Coult ciaimcd~ action is pendinJ in discovery phase. 4. Florida Power & l.i~ht Mocciola v. FP&L Voluntary dimlaimi of case. 5. Roomon LiJhtlnI & Power Scott v. HL&P, No. 87-058967 Disroland. ippendi~ C - Rt~u~ of F. MF ~urvq C-13 Ton (Continucxl) N~. ~ Case Name Sa~s 6. Houmm Li~ & Pow~ Kkin v. HL&.P C.~u~ found that utiliW camm~ be held for tort ~y of trespass becaus~ it had acquiz~d pre~ofiy 7. ladimm & Mi, chilaa iaecuic RuBle v. laclimm Michilaa Power, Sub alk~le. permood iajur~ from EM~ and Co. No. 20D02-9000S-CP-3~I locks caused by po~r ~ on property; mit ~ May 24, 1990; didovery prne~dures curremly being !itintod. 8. Miss~pi Power Company Shaw v. MPC~ No. S87-0329(s~ Pe~clinJ. 9. !lhiladdplfia Ebctdc Power Hoch v. Philad~!pbla Electric Power Landow~r claimed trespus and nuisance as Co. Co. (1985) · rsmah of E]vlF; trial cou~ dismissed complaint stat~g Public Utility Commission was proper forum; on appol, Hoch 10. Utah Powsr & ~ (kreld Rantoo v. Utah P&L (farm in Aikged int=rf~rencs with farm animal/dairy Idaho) catt~ production; $8 million claim=d; dismi~d. I I. Kaum C~ Pow=r & Lilht Ov=dmi Park Subemioo Utility found new sitg for proporal subslstlon after local residehis expreMed coficern over proximity to local hiJh school. 12. Alabama Powsr Co. (APC) Alkn v. APC (Civil Action 91-321); Pending C'u~uit Coofit Jeffcraoo Co. Alabama. No. Utit,/ Case/Project Demcriplloa Slam 1. Arizona Public Service Contory SubMarion School Board cxamlncd hulth cffects iuuc prior to buiidixq school nur 69-kV roberation and procee~ed_ with conmructlon of s~hool. 2. Baltimore Gas & Ekctdc Oakland Mills Road-Wildc Lake 230- PSC bekl-line wouM nnt be undergrounded Compan]t kV Line on basis of health effects. 3. City of Auratin 345-kV Loop Around City to impoff City Council dcnled original plan and Power required utility to mudy dtormfivea that would minimize expomre; study is 4. City of Ck. vdand Inland - Lakeehore Approval Hearing City of Cleveland bed hearing before thc CleveLand City Council, including EMF tomlmony. Council voted 18-4 to approve fine in city. Subsequent appeal to state coo~ denied in MR),t 1991. 5. Cemeal Maine Power Co. Hydm-4;~bec Tle-ln EMF was diecussed in Eavimnmemal Impact Ststement~ line was built. 6. Consolidated Edboa Resolution to Halt Comtrm:tina: 40th City Council heard testimony on the EMF St. Submmioa issue but did nnt reoder decision; robfarina was enerlized in 1985. C-14 Health Effects of Exposure to Powerlint Frequency Electric and Magnetic Fields Other (Continual) 7. El Puo lE~Uic Company Springviltc-FJ Paso 345-kV Line ]EMF was discuaml in F..nvln,-~-'='='1 -'.~,~ct =~,.~..menti line was bulb. 8. Florida !:rimvet & l.~t Rau/ch v. School Board of Palm Limited access to playF'emnd. 9. Houstm~ Li~htin$ & ~ Rainwater v. HL&P, No. TI~ c. au has b~n dismiuod. Cc,~q,~ny 87-7058968 10. iacboewilk Electric Authority Sabre Palms Ekmenlary School School Board held hearloSs; local cancer (school located mr 230-kV Line) doctor asked by Board to testify; teMimony that no connection between EJdF and adverse health effects; no furdeer action taken by ~-k__,xd _non_ rd or parents. 11. Kansas City Power &Light Overland Park Substation Utility found stew site for i,i~ed substation after local resldesMs expressed concern over proximity to local hi.eh school. 12. LADWP Citizens for a Better Henderson v. Citizens challenged adequacy of EIS for Hodel (1985) proporal transmission line; federal judge found that HIS provided sufficient discussion of health effects; appeals coot1 upheld this ruling. 13. Louisiana Power & Light Jefferson Parish Council resolution FJectric fields wli~, ROW limited to 8 kV/m to limit EMF exposures in and st edge of ROW ! kV/m; maSheric transmission line ROW's & from exposure in ROW maximum substations hours per day. Resolution passed and will remain in effect until after 1991 ssuion of Lo~, ..... 14, Mmm~humts Electric Co, Planm~ conm'uction of a 345-kV Tomm of ~ast C,~nwlch considering line in the town of East Greenwich amrstorium on the conMmction of any new ~wcr lines above 115 kV; heariqs have been held; more may be bald; no ssheduled decision yet on the morstoflum. 1~. NarraSansett ElecLric Co. Warwick, R.I. Town Ordinance passed to place three year norstorium on facilities 60 kV and over, Narraganscu E-I°-'-s--'ic will ask PUC to overturn. 16. Nevadm Power Company Warm Sptiqs-Easlern 138-kV Usw County Board of Commissi5~i~ heard Appmvd Hcarin~ L-~'.,.,nony on EMF and approv=d liz=. 17. Public Service Co. of New 1984 ROW Widlh Hearings EMF raised t= width of ROWs; utility's ROW Mexico widl request 18. hffet Sound Power & Light Wattore Co. Wash. Initiative 4=90 iniliativc to prohibit constru~ii,,a of tnnemiuion lines above 115 kV in udan areas: l~=h_;~_,)_ve 4-90 passed by 2 to I majority. 19. Salt River Project Api~h~ Junction School Disarict School ~.n, 20. Wiscomin Public Service Co~. Wisconsin Public Service Township ordinance d~larod void let would Commission Cuc No. 9327-E1-100 have requirmt all lines above 20 kV to "reduce to the extent pouible' electromagnetic field effects. 21. Wissonsin Public Service Corp. Hiking Trail Ordinance City ordinance to declare deed for hiking trail under line void; PSC declared ordinance invalid. :7.2. Wiscomin Public Service Cor ,. Bianmml Planning Hearing Before Pendln;; Hearin8 on the EMF imme the WiK,~;-='=n PSC but report nm yet issued by Commission. Health EJ~ect~ of E~posure w Powerline Frequtncy F,~ctric and Magnetic F~elds In contrast to prevalence, incidence is · measure of the with short survival will not be available for n~v csgs of diseag occurring in · population in · study; thus, relx~aining cases may not be given tithe interval. !f 15 of the 30 persons with lung typical of all cases and a potential association cancer in the previous example were first diagnosed in between exposure and dim may be masked 1989, the incidence of lung c~ncer in this population or exaggerated. would be 15 per 100,000 per year (also written 15/100,000/yr or 15 per 100,000 person-yeaxs). The Cohort Studies dimension of time is the defining characteristic of incideno~, making this measure of dis4ase occurrence · Cohort Studies staxt with the seleaion of groups of rate. The term rate always implies that the disease-free individuah on the basis of some exposure measurement of diseaR or death in a population is ramble. Exposed and non-expom:d individuals are rehted to · specified period of time. Thus, although then followed-up to dettrmine subsequent development some scientists refer to prevalence as prevalence rs~, of diseag. There at~ two type, of ~ohort or follow-up the expression is a misnomer. studies: prospective (or concurrent) and retxospective (or nonconcurrent or historical). These two types Cross-Sectional or la~valence Studies differ in t~rms of when exposure and disease occur in re-ln_'_nnn to the onset of the study. Cross-seaional studies examine factors of interut in · defined population at · pattieuhr point in time. The Prospective (Concurrent) Cohort Studies. Prospective study group may represent · random sample of a cohort studies ar~ most ,imihr to the classic laboratory community, working at · paaieular occupation, or · study. The· studies first identify · group of persons sample chosen on the bash of some sociological or (cohort) who are currently free of disease, but who environmental variable. Through a questionnaire, differ in terms of exposure to the agent under study. physical examira·ion and/or other means, the preaence For example, the cohort nay be a specified group of or absence of the disease(s) in question i~ determined ~nxluetive~tge women, and the exposure variable for each individtml, along with other characteristics or may be the use of oral oontntceptivea. The cohort is exposures of interest (e.g., age, whether the person then "followed-up' at some future time to determine the smoke·, exercise level, blood pressure, diet). oexurrenee of disease(s) in the cohotx. How soon follow-up begs or the length of time it must be Some of the advantages of cmu-seetional stuclie, eondua~d, depends on the disease outcome(s) of include: interet and their c~ (e.g., induction or iat~n~y periods). Incidence can then be oomptrext in · They can generally be performed relatively exposed and non-~xposed group,. The~e ra~, which quickly and inexpensively. must be adjusted for differt~ees in age and other characteristics of the study subjects, are typically · They provide valuable de·trip·lye information expr~sed as · ratio or 'relative risk' for the exposed on the existing patterns of disease occurrence. group. · They can examine a variety of factors and The advantages of prospective cohort studies are diseases simultaneously. significant: Them ajorlimitationsincludc: * They allow thc direct detennisation of incidence among exposed and non-~xposed · The "snapshot" approach my not allow one groups. This peruaim calculation of thc to _,~e~__ermine whether exposure act·rally incnased disesse risk (relative risk) prcc~___,~_ development of thediseasc. usociated with the exposurc. They also permit calcuhtion of the "attn~>utablc risk," which is that portion of thc incidence of · · Diseascs that gencratly hvc · longer duration particular disease that is due to · specific arc more likely to be detccted than diseases with the same incidence rate but with a shortcr duration. Thus an association · They may yield more extensive and more betwcen an exposure and a discase of short duration nay be missed. reliable data on exposure levels, as well as on confounding factors (e.g., cigarette smoking). · Individuals who survive longer with a disease · Many different disease outcomes can be are more likely to be found than those with shorter survival times. Therefore, the cases investigated in · single study.