DRAFT
DO NOT QUOTE OR CITE
 EPA/600/6-90/005B
      October 1990
External Review Draft
               EVALUATION OF THE POTENTIAL CARCINOGENICITY OF
                           ELECTROMAGNETIC FIELDS
                                     NOTICE
THIS DOCUMENT IS A PRELIMINARY DRAFT. It has not been formally released by the U.S.
Environmental Protection Agency and should not at this stage be construed to represent
Agency policy. It is being circulated for comment on its technical accuracy and policy
implications.
                    Office of Health and Environmental Assessment
                         Office of Research and Development
                        U.S. Environmental Protection Agency
                                 Washington, D.C.
                                                                Printed on Recycled Paper

-------
                         DRAFT-DO NOT QUOTE OR CITE
                                  DISCLAIMER
      This document is an external draft for review purposes only and does not constitute
Agency policy. Mention of trade names or commercial products does not constitute
endorsement or recommendation for use.

-------
                          DRAFT-DO NOT QUOTE OR CITE
                                   CONTENTS

TABLES  .  .	vjjj

FIGURES	xii

PREFACE	xiij

AUTHORS AND REVIEWERS	xiv

1.   EXECUTIVE SUMMARY	1-1

2.   MECHANISMS OF ELECTROMAGNETIC INTERACTION	2-1

     2.1.  PHYSICAL PROPERTIES OF ELECTRIC AND MAGNETIC FIELDS   	2-1

          2.1.1.   Extremely Low Frequency Fields  	2-5
          2.1.2.   Radiofrequency Fields	 ' 2-6

     2.2.  COUPLING OF ELECTRIC AND MAGNETIC FIELDS WITH THE BODY	2-8

          2.2.1.   Extremely Low Frequency Fields  	2-10
          2.2.2.   Radiofrequency Fields	2-16

     2.3.  AMBIENT EXPOSURE   	2-20

     2.4.  PROPOSED MECHANISMS OF INTERACTION  	'.  .  . 2-22

          2.4.1.   Surface Compartment Model  	2-23
          2.4.2.   Ion Cyclotron Resonance	2-25
          2.4.3.   Cooperative Mechanisms	2-29

REFERENCES FOR CHAPTER 2  	2-33

3.   EPIDEMIOLOGIC STUDIES OF ELECTROMAGNETIC FIELDS AND CANCER	3-1

     3.1.  INTRODUCTION   	3-1

     3.2.  STUDIES OF CHILDREN  	'. . 3-4

          3.2.1.   50- or 60-Hertz Exposures   	•	3.4
          3.2.2.   Electromagnetic-Field Exposure at Unspecified Frequencies  . '.  '.  '. 3-33
          3.2.3.   Summary  	3.37

     3.3.  STUDIES OF ADULTS  	3.44

         3.3.1.   Radiofrequency Exposures  	3.44
         3.3.2.   50- or 60-Hertz Exposures or Electromagnetic-Field
                 Exposure to Unspecified Frequencies   	3-64

                 3.3.2.1.  Residential (50 or 60 Hertz)  	3-64
                 3.3.2.2.  Occupational	3-83

                        3.3.2.2.1.  Multiple Sites	3-83
                        3.3.2.2.2.  Cancer of the Hematopoietic System  	3-97
                        3.3.2.2.3.  Cancer of the Nervous System   	3-110
                        3.3.2.2.4.  Malignant Melanoma of the Skin	3-118
                        3.3.2.2.5.  Summary of Occupational Studies ....... .3-121

-------
                           DRAFT-DO NOT QUOTE OR CITE
     3.4.  SUMMARY	3-124

          3.4.1.   Introduction	3-124

          3.4.2.   Childhood Cancer	3-124

          3.4.3.   Adult Cancer .	3-126

                  3.4.3.1.  Radiofrequency Radiation	3-126
                  3.4.3.2.  Adults Residentially Exposed to Electromagnetic Fields   . . 3-142
                  3.4.3.3.  Occupational Exposure to Electromagnetic Radiation   .  . . 3-143

          3.4.4.   Conclusions	3-147

     3.5.  OTHER END POINTS	3-148

     3.6.  ONGOING RESEARCH	3-149

     REFERENCES FOR CHAPTER 3	3-154

4.    ANIMAL STUDIES	4-1

     4.1.  UNIVERSITY OF WASHINGTON LONG-TERM RAT STUDY	4-1

          4.1.1.   Description of Study	4-2
                  4.1.1.1.  Animal Facility  	4-2
                  4.1.1.2.  Animals	4-2
                  4.1.1.3.  Microwave Exposure	4-3
                  4.1.1.4.  Protocol of the Experiment  	4-5

          4.1.2.   Results of the Study	4-5

                  4.1.2.1.  Behavior and Corticosterone   	4-5
                  4.1.2.2.  Immune Competence  	4-6
                  4.1.2.3.  Blood  Chemistry and Hematologic Measurements	4-6
                  4.1.2.4.  Metabolism	4-6
                  4.1.2.5.  Survival	4-8
                  4.1.2.6.  Histopathologic Findings   	4-8

                          4.1.2.6.1.  Tumor Incidence	4-9

                                   4.1.2.6.1.1.    Adrenal medulla tumors  	4-15
                                   4.1.2.6.1.2.    Malignant tumors of all sites  ...  4-16
                                   4.1.2.6.1.3.    Glandular tumors	4-17
                                   4.1.2.6.1.4.    Exposure-induced changes
                                                in progression of tumors	4-19
                          4.1.2.6.2.  Historical Controls	4-20

                  4.1.2.7.  Summary of Results  	4-23
                  4.1.2.8.  Discussion of Results  	4-24

                          4.1.2.8.1.  Benign Adrenal Pheochromocytomas	4-24
                          4:1.2.8.2.  Combinations of Separate Tumor Sites  	4-26

     4.2.  PRAUSNITZ AND SUSSKIND (1962) STUDY	4-28

     4.3.  SPALDING ETAL (1971) STUDY .	4-30

     4.4.  SZMIGIELSKI ETAL. (1982) STUDY   	4-30

     4.5.  BAUM ETAL (1976) STUDY	4-32
                                         IV

-------
                           DRAFT-DO NOT QUOTE OR CITE
     4.6.  STUDIES IN PROGRESS OR PLANNED  	4.33

          4.6.1.   Ontario Hydroelectric Power Company	                4-33
          4.6.2.   U.S. Air Force  		4.^4
          4.6.3.   The University of Rochester	                 	4.34
          4.6.4.   Additional Studies	'.'.'.'.'.'.'.'.'.'.'.'.'. 4-34

     4.7.  SUMMARY OF LONG-TERM ANIMAL STUDIES   	4.35

REFERENCES FOR CHAPTER 4	4.39

5.    SUPPORTING EVIDENCE OF CARCINOGENICITY   	5-1

     5.1.  GENOTOXICITY OR NONGENOTOXICITY	5-1

          5.1.1.   Introduction  	       5_1
          5.1.2.   Effects on Nucleic Acids	'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'. 5-2

                  5.1.2.1.  Extremely Low Frequency Electromagnetic Fields   ..... .5-2
                  5.1.2.2.  Radiofrequency Electromagnetic Fields	5-4

          5.1.3.   Gene Mutations  .	5.5

                  5.1.3.1.  Extremely Low Frequency Electromagnetic Fields   .         5-6
                  5.1.3.2.  Radiofrequency Electromagnetic Fields	5-9
          5.1.4.   Chromosome Effects   	5_10

                  5.1.4.1.  Extremely Low Frequency Electromagnetic Fields   ..      5-10
                  5.1.4.2.  Radiofrequency Electromagnetic Fields	5-14
                  5.1.4.3.  Summary	5_19

          5.1.5.   Summary of Genetic Effects	5_1S

     5.2.  EFFECTS ON MITOSIS AND MEIOSIS	5.25

          5.2.1.   Extremely Low Frequency Electromagnetic Fields   . .               5-25
          5.2.2.   Radiofrequency Electromagnetic Fields   	   '     5.30
          5.2.3.   Summary	]    5_31

     5.3.  EFFECTS ON TRANSCRIPTION, TRANSLATION, AND CELL
          TRANSFORMATION	5.35

          5.3.1.    Extremely Low Frequency Electromagnetic Fields   	     5-35
          5.3.2.    Radiofrequency Electromagnetic Fields   	       5-38
          5.3.3.   Summary   	'.'.'.'.'.','.'.'.  5-39

     5.4.  CALCIUM EFFLUX FROM BRAIN TISSUE  	5-40

          5.4.1.   Extremely Low Frequency Fields  	                     5-40
          5.4.2.   Modulated Radiofrequency Fields   	          	5.44
          5.4.3.   Unmodulated Radiofrequency Fields	             '  5.43
          5.4.4.   Summary   	'    5.43

     5.5.   INTRACELLULAR ENZYME RESPONSES   .  .	5.49
                                            4
          5.5.1.   Protein Kinases	  5.49

                 5.5.1.1. Modulated Radiofrequency Fields	5-49

          5.5.2.   Ornithine Decarboxylase Activity	5-50

                 5.5.2.1. Extremely Low Frequency Fields   	5-51
                 5.5.2.2. Modulated Radiofrequency Fields	  5-52

-------
                         DRAFT-DO NOT QUOTE OR CITE
          5.5.3.   Summary	5-54

     5.6.  PARATHYROID HORMONE AND THE PLASMA MEMBRANE  	5-54

          5.6.1.   Extremely Low Frequency Fields  	5-55
          5.6.2.   Modulated Radiofrequency Fields	5-58
          5.6.3.   Unmodulated Radiofrequency Fields	5-59

          5.6.4.   Summary	  5-59

     5.7.  MELATONIN AND OTHER HORMONES  	5-59

          5.7.1.   Background: Melatonin and Cancer	5-59
          5.7.2.   Extremely Low Frequency Fields  	5-61
          5.7.3.   Modulated Radiofrequency Fields   	5-63
          5.7.4.   Unmodulated Radiofrequency Fields	5-63
          5.7.5.   Summary	5-64

     5.8.  GROWTH AND DIFFERENTIATION	5-64

          5.8.1.   Extremely Low Frequency Fields  	5-64
          5.8.2.   Modulated Radiofrequency Fields   	5-69
          5.8.3.   Radiofrequency Fields	5-69
          5.8.4.   Static Electric Fields	5-70
          5.8.5.   Summary   	5-70

     5.9.  IMMUNOLOGIC/HEMATOLOGIC EFFECTS	5-70

          5.9.1.   Extremely Low Frequency Fields  	5-71
          5.9.2.   Modulated Radiofrequency Fields   	5-75
          5.9.3.   Unmodulated Radiofrequency Fields	5-76
          5.9.4.   Summary   	5-79
                 5.9.4.1.  Extremely Low Frequency Fields   	5-79
                 5.9.4.2.  Modulated and Unmodulated Radiofrequency Fields ....  5-81

     5.10. CENTRAL NERVOUS SYSTEM EFFECTS  	5-82

          5.10.1.  Extremely Low Frequency Fields  	5-82
          5.10.2.  Modulated Radiofrequency Fields   	5-84
          5.10.3.  Unmodulated Radiofrequency Fields	5-85
          5.10.4.  Summary   	5-86

     5.11. SUMMARY AND CONCLUSIONS FOR SUPPORTING EVIDENCE OF
          CARCINOGENICITY	5-87
          5.11.1.  Summary   .	5-87
          5.11.2.  Conclusions	5-91

     REFERENCES FOR CHAPTER 5	5-93

6.    RESEARCH NEEDS	6-1

     6.1   INTRODUCTION   	6-1

     6.2.  INFORMATION NEEDS ARISING FROM THE EVALUATION IN THIS
          DOCUMENT	6-1

          6.2.1.   Epidemiology Research Needs	6-2
          6.2.2.   Laboratory Research Needs  	6-3

7.    SUMMARY AND CONCLUSIONS	7-1
                                       VI

-------
                     DRAFT-DO NOT QUOTE OR CITE
7.1.  INTRODUCTION
7.2.  MECHANISMS OF INTERACTION BETWEEN TISSUE AND
     ELECTROMAGNETIC FIELDS
7.3.  HUMAN EVIDENCE  ................................ 7.2

     7.3.1.   Studies of Children  ........                              72
     7.3.2.   Studies of Adults   ........... '.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'. 7-3

            7.3.2.1 .  Residential Exposure to Power Frequency Fields  ........ 7-3
            7.3.2.2.  Occupational Exposure to Extremely Low Frequency and
                    Mixed Frequency Fields  ..... ;  .....  ....  ..... 7-4
            7.3.2.3,  Radiofrequency Exposure ................  ' '  | 7.4

     7.3.3.   Summary of Human Evidence ..... ................. 7.5

7.4.  ANIMAL EVIDENCE  ..............  . ....... •  ......... 7.6

     7.4.1 .   Extremely Low Frequency Fields  ....  .......               7.5
     7.4.2.   Radiofrequency Radiation  ............. '.'.'.'.'.'.'.'.'.'.'. 7-6

            7.4.2.1 .  Unmodulated Radiofrequency Radiation  ........... 7-6
            7.4.2.2.  Modulated Radiofrequency Radiation  ....... ...... 7-7
7.5.  SUPPORTING EVIDENCE OF CARCINOGENICITY  ............... 7.7

7.6.  INTEGRATED DISCUSSION OF SEPARATE CHAPTERS   ............ 7-8
                                 VII

-------
                            DRAFT-DO NOT QUOTE OR CITE



                                        TABLES

Table 2-1.     Factors affecting internal exposure conditions  	2-9

Table 3-1.     Summary of studies of childhood cancer	3-2

Table 3-2.     Distribution of cases and controls according to various residences
              coded as high current configuration (HCC) or low/ current
              configuration (LCC) for total cancers  	3-6

Table 3-3.     Distribution of cases and controls according to various residences
              coded as HCC or LCC for specific cancer sites	3-7

Table 3-4.     Odds ratios for the distribution of the 2098 dwellings of cancer cases
              and controls by type of electrical structure visible within 150
              meters	3-14

Table 3-5.     Magnetic field measurements for case and control dwellings and odds
              ratios by magnetic field level and type of electrical construction	3-16

Table 3-6.     Risk ratios for specific cancer sites and magnetic field level  	3-18

Table 3-7.     Distribution of cases and controls by distance from overhead power
              lines and by cancer type	3-20

Table 3-8.     Distribution of cases and controls by estimated magnetic field level in
              milligauss (mG) and by cancer type	3-21

Table 3-9.     Eligibility, response, and losses for cancer cases, case subgroups,
              and controls:  Denver standard metropolitan statistical area	3-25

Table 3-10.    Distribution of electric and magnetic field measurements and wire
              codes for cancer cases, case subgroups, and controls:  Denver
              standard metropolitan statistical area	3-25

Table 3-11.    Cancer risk in relation to measured  magnetic fields and electric
              fields, under low- or high-power use conditions, in residences
              occupied at diagnosis:  Denver standard metropolitan statistical
              area  	3-26

Table 3-12.    Cancer risk (odds ratios with 95% confidence intervals in
              parentheses) in relation to magnetic fields and electric fields,
              categorized into two exposure groups and measured under low-
              or high-power use conditions, in residences occupied at
              diagnosis:  Denver  standard metropolitan statistical area   	3-27

Table 3-13.    Cancer risk (odds ratios for all sites combined) in relation to a
              five-level wiring configuration code for residences occupied at
              the time of diagnosis or 2 years before diagnosis: Denver
              standard metropolitan statistical area	3-29

Table 3-14.    Cancer risk (odds ratios) in relation to dichotomized wire codes (low
              vs. high and buried vs. very high)  for residences occupied at the
              time of diagnosis for all cancers and for specific sites:  Denver
              standard metropolitan statistical area	3-31
                                           VIII

-------
                           DRAFT-DO NOT QUOTE OR CITE
                                 TABLES (continued)
Table 3-15.   Microwave exposure levels at the U.S. Embassy in Moscow	3-45

Table 3-16.   Observed and expected numbers of deaths, standardized mortality
             ratios (SMR), and 95% confidence interval (Cl) by all causes of
             death, specified causes of death from cancer, and post for male
             and female and State and Nonstate Department employees '
             combined  	3-46

Table 3-17.   Number of deaths from disease and mortality ratios by hazard
             number:  U.S. enlisted naval personnel exposed to microwave
             radiation during the Korean War period	3-49

Table 3-18.   Mortality in Washington State and California:  U.S. Federal
             Communications Commission (FCC) amateur radio operator
             licensees, January 1,1979, to December 31,1984	3-51

Table 3-19.   Mortality of rad lab staff and physicians, through  December 31,1974,
             for all causes  of death and cancers	3-59

Table 3-20.   Age-adjusted  mortality rates per 1000 person-years for all causes of
             death and selected cancers for  rad lab white male staff members
             by exposure ranking group based on predominant division of
             work, followed through December 31,1986	3-61

Table 3-21.   Incidence rates per 100,000 per year (1971 -1980), for Polish military
             personnel grouped by work exposure to radiofrequency and
             microwave radiation and by age	3-63

Table 3-22.   Daytime 60-Hz magnetic fields measured next to  the part of the
             house nearest to distribution wires by wire code	3-66

Table 3-23.   Wiring configurations at the homes of cancer cases and controls	3-67

Table 3-24.   C-ratios for cancer in  Colorado adults for residences near higher
             wiring current configurations by town and various factors   	3-68

Table 3-25.   Losses/exclusions in a study of acute nonlymphocytic leukemia and
             residential exposure to power frequency magnetic fields in adults
             in western Washington State	3-77

Table 3-26.   Risk estimates of acute nonlymphocytic leukemia in adults in relation
             to exposure based on Wertheimer and Leeper's wiring
             classification scheme from fitted logistic regression models,
             western Washington State,  1981 -1984  	3-77

Table 3-27.   Risk estimates of acute nonlymphocytic leukemia in adults in relation
             to the weighted and unweighted mean magnetic field
             measurements made  in the kitchen, bedroom, and family
             gathering room of the subject's  residence at reference date,
             western Washington State,  1981 -1984  	3-79
                                          IX

-------
                           DRAFT-DO NOT QUOTE OR CITE
                                 TABLES (continued)
Table 3-28.   Electric blanket use among adults (aged 20-69 years) with acute and
             chronic myelogenous leukemia, residing in Los Angeles County,
             and their matched neighborhood controls  	3-81

Table 3-29.   Person-years and observed and expected all-cause mortality or
             cancer incidence by latency and exposure duration for selected
             Swedish male chloralkali workers, with expected values based on
             national and regional rates for males   	3-89

Table 3-30.   Standardized incidence ratios (and number of cases) for specific
             causes of cancer by type of work among male New York
             telephone employees 1976-1980; standard:  New York State
             male rates   	3-94

Table 3-31.   Standardized incidence ratios (and number of cases) for specific
             causes of cancer by type of work among male New York
             telephone employees 1976-1980, standard:  nonline workers  	3-95

Table 3-32.   Background information on shipyard workers at Portsmouth Naval
             Shipyard, New Hampshire	3-103

Table 3-33.   Results of Mantel-Haenszel analysis of leukemia for selected jobs
             and shops (with the  highest risk estimates) for shipyard workers  	3-105

Table 3-34.   Results from conditional logistic regression analysis for solvent,
             electrical, and welding work, treated either as a categorical or
             continuous variable  (years used as index of exposure), among
             shipyard workers    	3-105

Table 3-35.   Selected standardized incidence ratios (SIRs) for leukemias by major
             division of industry and  occupation, by general manufacturing
             industries, and by general craftsmen-tradesmen occupations,
             among Swedish males 1961-1979	3-106

Table 3-36.   Age-adjusted odds ratios for leukemia by type of electrical work
             among New Zealand males aged 20 years or older   	3-109

Table 3-37.   Age-adjusted odds ratios for leukemia among New Zealand male
             electrical workers, by age and leukemia subtype	3-110

Table 3-38.   Mortality odds  ratios (MORs) for brain cancer, leukemias, and acute
             myeloid leukemias among male electrical workers from 16 states	3-116

Table 3-39.   Studies of cancer in  adults   	3-127

Table 3-40.   Ongoing epidemiologic studies on electromagnetic fields and cancer  . .  .3-150

Table 4-1.    Frequency table for  cause of death: all animals	4-10

Table 4-2.    Single animal,  microscopic observations  	4-11

-------


Table 4-3.
Table 4-4.
Table 4-5.
Table 4-6.
Table 4-7.
Table 4-8.
Table 4-9.

Table 4-10.

Table 4-1 1 .

Table 4-1 2.

Table 5-1 .
Table 5-2a.
Table 5-2b.
Table 5-3.
Table 5-4.
Table 5-5.

Table 5-6.
Table 5-7.

Figure 2-1 .
Figure 2-2.

DRAFT-DO NOT QUOTE OR CITE
TABLES (continued)
Crude incidence and time of appearance of neoplastic lesions ....
Frequency of benign adrenal medulla tumors 	
Statistical significance of selected tumor types and sites 	
Multiplicity of tumors in glandular organs 	
Number of animals with multiple tumors (derived from Table 4-6) . . .
Assignment of ranks to stages of progression 	 	
Severity rankings of adrenal cortical preneoplastic and neoplastic
lesions 	 	
Results of tests for the statistical significance of graded tumor
response 	 	
Comparison of spontaneous endocrine tumors in Sprague-Dawley
rat stocks 	
Summary of chronic animal experiments with electromagnetic-field
exposure 	
Negative gene mutation reports 	 	
Chromosome effects, extremely low frequency fields 	
Chromosome effects, radiofrequency fields 	
Effects on mitosis and meiosis 	 	 	
Characteristics of electromagnetic fields tested 	
Effect of mw irradiation on transformation frequency (x 1 03) in
C3H/1 OT1/2 mouse embryo fibroblasts 	 	
Effects of fields on cAMP accumulation in bone cell monolayers . . . .
Summary of supporting evidence for carcinogenicity 	
FIGURES
The electromagnetic spectrum 	
External electric and magnetic field required to obtain an internal E
field of 10mV/m 	


4-12
4-14
4-17
4-18
4-18
4-19

4-20

4-20

4-21

4-36
5-7
5-19
5-22
5-32
5-37

5-38
5-55
5-88

2-5

. . 2-12
XI

-------
                            DRAFT-DO NOT QUOTE OR CITE
                                   FIGURES (continued)
Figure 2-3.
Figure 2-4.


Figure 2-5.

Figure 2-6.


Figure 4-1.



Figure 4-2.

Figure 5-1.
The minimum current densities observed to cause various biological
changes: (A) threshold for stimulation for muscle or nerve cells;
(B) threshold for fibrillation in a dog; (C) threshold for diastolic
stimulation in dog hearts; (H) threshold for phosphene release;
(Ji.Ja) noise current or current density fluctuations in nerve
impulses	 ,

Predicted frequency dependence of absorbed energy in spheroidal
models of biological bodies	
Plane wave induced current in humans vs. frequency
Variation in electric and magnetic field intensities at ground level as a
function of distance from the source  	
Modulation characteristics of the pulsed microwave source: fifty
10-microsecond-wide pulses per group, with a repetition rate of
800 pulses per second during the burst of 50 pulses	
2-15


2-17

2-18


2-21




. 4-4
Kaplan-Meier analysis of survival: exposed (A) vs. control (B)	4-9

Proposed mechanism by which chronic exposure to a 60-Hz electric
field may increase dimethylbenz[a]anthracene (DMBA)-induced
mammary carcinogenesis in rats   	5-62
                                          xii

-------
                           DRAFT-DO NOT QUOTE OR CITE

                                      PREFACE

      This document was prepared by the Human Health Assessment Group of the Office of
Health and Environmental Assessment (part of the Office of Research and Development, U.S.
Environmental Protection Agency) at the request of the Office of Radiation Programs. It
summarizes and evaluates the available literature relating to the potential carcinogenicity of
electromagnetic fields in the frequency range from 3 Hz to 30 GHz (3x1010 Hz).  Topics
evaluated are human epidemiologic studies relating  in some way to carcinogenesis, chronic
animal studies, and short-term and in vitro studies related to the carcinogenic effects of these
fields. The literature search supporting this review was completed in mid-1989.
                                         XIII

-------
                          DRAFT-DO NOT QUOTE OR CITE




                            AUTHORS AND REVIEWERS




      The Human Health Assessment Group (HHAG) of the Office of Health and Environmental

Assessment (OHEA) was responsible for the preparation of this document.
AUTHORS
Robert E. McGaughy
Human Health Assessment Group
Office of Health and Environmental Assessment
U.S. Environmental Protection Agency

Paul C. Gailey and Clay E. Easterly*
Oak Ridge National Laboratory
Oak Ridge, TN

Doreen Hill
Office of Radiation Programs
U.S. Environmental Protection Agency

David Bayliss
Human Health Assessment Group
Office of Health and Environmental Assessment
U.S. Environmental Protection Agency

Robert E. McGaughy


Bradford L. Whitfield*
Oak Ridge National Laboratory
Oak Ridge, TN

Mary Lou Daugherty*
Oak Ridge National Laboratory
Oak Ridge, TN


Robert E. McGaughy
Chapter 1




Chapter 2



Chapter 3



Chapter 3




Chapter 4, Chapter 5 (Section 5.11)


Chapter 5 (Sections 5.1 through 5.3)



Chapter 5 (Sections 5.4 through 5.10)




Chapters 6 and 7
*Prepared under interagency agreement No. DW8932701.




PROJECT MANAGER

Robert E. McGaughy
                                       xiv

-------
                          DRAFT-DO NOT QUOTE OR CITE
REVIEWERS
      This document has been reviewed, either in whole or in part, by the following individuals
within the U.S. Environmental Protection Agency. In addition, reviews have been carried out by
interested individuals from the U.S. Department of Energy, Centers for Disease Control, and the
National Cancer Institute.
Michael Berry
Carl Blackman
Rebecca Calderon
Gunther Craun
Chao Chen
Arthur Chiu
Christopher DeRosa
Joseph Elder
Larry Glass
Martin  Halper
Norbert Hankin
David Kleffman
Edward Mantiply
Debdas Mukerjee
Patricia Murphy
Neal Nelson
Stephen Nesnow
David Reese
Sherry Selevan
Vicki Vaughan-Dellarco
James Walker
Paul White
Office of Health and Environmental Assessment
Environmental Criteria and Assessment Office-RTP
Health Effects Research Laboratory-RTP
Office of Health Research
Health Effects Research Laboratory-CIN
Office of Health and Environmental Assessment
Human Health Assessment Group
Office of Health and Environmental Assessment
Human Health Assessment Group
Office of Health and Environmental Assessment
Environmental Criteria and Assessment Office-CIN
Health Effects Research Laboratory-RTP
Office of Health and Environmental Assessment
Environmental Criteria and Assessment Office-CIN
Office of Radiation Programs
Office of Radiation Programs
Office of Health Research
Office of Radiation Programs-Las Vegas
Office of. Health and Environmental Assessment
Environmental Criteria and Assessment Office-CIN
Health Effects Research Laboratory-CIN
Office of Radiation Programs
Health Effects Research Laboratory-RTP
Office of Health and Environmental Assessment
Human Health Assessment Group
Office of Health and Environmental Assessment
Human Health Assessment Group
Office of Health and Environmental Assessment
Human Health Assessment Group
Office of Radiation Programs
Office of Health and Environmental Assessment
Exposure Assessment Group
                                         xv

-------
                           DRAFT-DO NOT QUOTE OR CITE
      The document was reviewed by a panel of scientists at a Peer Review Workshop held in
Morrisville, North Carolina, on June 28,1990.  Comments and recommendations made by the
panel at the workshop and in written post-meeting comments were incorporated into this draft.
The peer panel members were:



Dr. Larry Anderson
Battelle Pacific Northwest Laboratories
Richland, WA

Dr. Richard Griesemer
National Toxicology Program
Research Triangle Park, NC

Dr. Lawrence Kunz
Neorex Corporation
Seattle, WA

Dr. Richard A. Luben
University of  California
Riverside, CA

Dr. Raymond Neutra
California Department of Health Services
Berkeley, CA

Dr. Richard Phillips
Spokane, WA

Dr. Charles Poole
Epidemiology Resources, Inc.
Chestnut Hni, MA

Dr. Asher Sheppard                                  Written comments only.
Pettis Memorial Veterans Hospital                      He did not attend.
Loma Linda,  CA

Dr. Richard Stevens
Battelle Pacific Northwest Laboratories
Richland, WA
                                        xvi

-------
                           DRAFT-DO NOT QUOTE OR CITE
                             1.  EXECUTIVE SUMMARY

   This review and evaluation of the potential carcinogenicity of electromagnetic (EM) fields
has been carried out by the Office of Health and Environmental Assessment within the Office
of Research and Development at the request of the Office of Air and Radiation (OAR) at the
U.S. Environmental Protection Agency (EPA). The Office of Radiation Programs within OAR is
responsible for the radiation protection activities of the Agency. The purpose of the document
is to evaluate the likelihood that exposure to nonionizing electromagnetic radiation (NIEMR)
poses a risk or is a risk factor for the development of cancer in humans.  Although the entire
NIEMR spectrum is of interest, the emphasis in this document is on time-varying electric and
magnetic fields in the extremely low frequency (ELF) range [approximately 3 to 3000 hertz
(Hz)] and on radiofrequency (RF) radiation  [approximately 0.003 to 30,000 megahertz (MHz)].
These two regions of the spectrum are emphasized because they are of regulatory concern to
the Agency and because the preponderance of information is in these regions.
   The evaluation of the likelihood of human cancer risk is based on a judgment as to the
overall weight of evidence that a carcinogenic response is causally related to specific levels or
types of exposure. Since the establishment of causality is often difficult, the
weight-of-evidence approach relies on the combination of empirical observations and
inferences founded in reasonable scientific judgment.  Under this approach, the evidence from
human studies is considered  most important, with lesser importance being attached,
respectively, to chronic lifetime animal studies and ancillary evidence, such as short-term tests
of genetic toxicity,  mechanistic studies, and evidence of carcinogenicity for chemical
analogues to the agent under study. Accordingly, this document considers human, animal,
and supporting  evidence in separate chapters.
   There are four  essential elements in a risk assessment: hazard identification,
dose-response assessment, exposure assessment, and risk characterization.  This document
deals largely with hazard identification, with a brief section (Section 2.3) devoted to exposure.
Dose-response  assessment is not attempted because the nature of the interaction between
the body and electric and magnetic fields is not well enough understood to be able to specify
the relevant aspects of exposure. In the absence of critical information about both exposure
and dose-response, an overall risk characterization is not developed in this document.  In its
entirety, this document represents an analysis of the state-of-the-science supporting a concern
for the potential carcinogenic hazard of EM fields.
                                         1-1
10/18/90

-------
                            DRAFT-DO NOT QUOTE OR CITE


    The two basic sources of information that furnish evidence of the relationship between
 exposure to EM fields and cancer are human evidence and laboratory studies. Human
 evidence is observational in nature and cannot account for or control all of the potentially
 relevant factors. Laboratory studies with biological models of the human disease search for
 explanations of the human findings by evaluating the effects of the various controllable factors.
 It is necessary to have both types of information  in order to be able to specify what measures
 are likely to reduce the hazard.  However, if the laboratory studies are not able to provide
 adequate explanations, the human  observations  still furnish a basis for concern.
    A large number of human studies are available in which the relationship between human
 cancer incidence or mortality and exposure to EM fields has been investigated.  From these
 studies the strongest relative evidence that exposure to EM fields is causally related to human
 cancer comes from case-control studies of cancer in children. Seven of these have examined
 residential exposure from electric power transmission and distribution lines and two others
 have examined cancer in children in relation to father's occupaition. These studies have
 consistently found modestly elevated risks (some statistically significant) of leukemia, cancer
 of the nervous system and, to a lesser extent, lymphomas.  These findings are associated with
 magnetic fields in homes where children reside which were estimated after the diagnosis with
 both magnetic field measurements and with surrogate  indicators of magnetic fields, i.e., wiring
 codes.  Electric fields were not found to be a critical factor thus far. In two studies in which
 magnetic fields measurements were made, significant elevated risks were observed in those
 exposed at or above 2 to 3 milligauss (mG) [0.2 to 0.3  microtesla (fiT)].  In contrast with
 adults, children probably have relatively few confounding factors other than EM fields that
 could explain the association because of their shorter lifespan and  lack of occupational
 exposure. In fact, potential confounders and biases that might have affected the results were
 examined by  one of the authors in some detail and found not to explain the results. As yet, no
 other agents have been identified that could explain this association.  Although a
 dose-response relationship with respect to surrogate measures of exposure is suggested in
two studies, reliable dose-response information is not available due to the use of dichotomized
 exposure categories and small numbers of cases within the exposure groupings.  Issues
 pertaining to personal exposure and latency have not been addressed.
    Additional, but weaker, evidence that there  is  an elevated risk of leukemia, cancer of the
nervous system, and perhaps other sites comes from occupational studies of EM-field
exposure. Although many of these studies have found an excess risk of these forms of cancer
to be associated with employment in certain jobs that have a high potential for exposure to EM
10/18/90
1-2

-------
                           DRAFT-DO NOT QUOTE OR CITE


fields, few or no measurements have actually been taken in those occupations. Furthermore,
the available sources of information concerning these occupations do not provide a reliable
indication of actual exposure to electromagnetic fields. The likelihood that misclassification or
information bias is present in these studies is high. However, exposure misclassification, if
random, tends to bias risks toward the null.  Despite these weaknesses, the occupational
studies tend to support the results of the childhood cancer studies, and excesses occur at the
same sites.
   The studies of residential adult exposures to EM fields provide mixed evidence of a risk of
leukemia, but due to a lack of statistical power and a lack of definite information on precise
EM-field exposures, these findings are not as strong as those for childhood cancer.  These
studies cannot be interpreted as evidence either for or against a causal association  between
cancer and EM-field exposures. On the other hand, the case-control study of cancer in
Colorado residents does support  an association of central nervous system cancer and
lymphoma if proximity to high-current electrical wiring configurations is assumed to be an
adequate surrogate for exposure.
   The studies of adults exposed to RF radiation produced mixed results, primarily  because of
limited sample size, inadequate length of follow-up, imprecise exposure data, and lack of
information on potential confounders.  These problems prevent conclusions to be made about
causal relationships with RF exposures.  However, the statistically significant excess risks of
leukemia in amateur radio operators requires further examination.
   There have been very few lifetime animal carcinogenicity studies of EM fields, and none at
power line frequencies. One study in mice of unmodulated 2450-MHz RF radiation at power
levels low enough to cause only moderate body heating showed an enhancement of the
growth rate of spontaneous mammary tumors  and of skin tumors initiated by benzo[a]pyrene,
a chemical carcinogen. One rat study of pulse-modulated 2450-MHz RF radiation designed to
simulate human exposure to medium-range radar showed the induction of benign adrenal
medulla tumors and an increased incidence of carcinomas at all tissue sites combined, with no
increase at any one site; the latter finding was not accompanied by the induction of benign
tumors.
   A large number of biological phenomena related in some way to known mechanisms of
carcinogenesis have been affected by EM fields under controlled laboratory exposure
situations.  ELF fields of relatively  high intensity producing induced tissue currents on the
order of  10 microamperes per square centimeter (/
-------
                            DRAFT-DO NOT QUOTE OR CITE


protein synthesis, delayed the mitotic cell cycle, induced chromosome aberrations, blocked
the action of parathyroid hormone at the site of its plasma membrane receptor, induced
enzymes normally active during cell proliferation,  inhibited differentiation and stimulated the
growth of carcinoma cell lines, inhibited the cytotoxicity of T-Iyrnphocytes (which indicates an
impairment of the immune system) in vitro but not in vivo, inhibited the synthesis of melatonin
(a hormone that suppresses the growth of several types of tumors), and caused alterations in
the binding of calcium to brain tissues. The large variety of exposure conditions and the lack
of detail on the geometry of the biological samples in these studies precludes a systematic
evaluation of the actual induced currents  and field strengths at the tissue and cellular levels
that are causing these effects. In  addition, the lack of reproducible results between
laboratories limits the interpretation of much of this literature.
    RF fields modulated at the same extremely low frequencies that cause some of the effects
noted above also result in the same responses, indicating that the ELF component may be
responsible for these effects. Unmodulated RF radiation has not caused any of the effects
noted above except for chromosome aberrations. None of the EM fields have caused gene
mutations, sister chromatid exchanges, or DNA damage (as measured by DNA breaks, DNA
repair, or differential killing of repair defective organisms) in a large number of studies.
    Only three ELF effects have been induced at field strengths comparable to the low
environmental exposures at which human cancer has putatively been caused: (1) the calcium
efflux from brain tissue preparations using 16-Hz electric and magnetic fields that were
perpendicular to each other; (2) calcium efflux from chick brain tissues after exposure of the
developing embryo to electric fields; and  (3) the inhibition of melatonin synthesis by the pineal
gland when a static magnetic field of approximately the strength of the earth's magnetic field is
changed through a small angle of rotation. The results of this first experiment are one of
several phenomena that show a complex dependence of frequency, intensity, and orientation
with respect to the earth's magnetic fields.
    In view of these laboratory studies, there is reason to believe that the findings of
carcinogenicity in humans are biologically plausible. However, the explanation of which
biological processes are involved and the way in which these  processes causally relate to
each other and to the induction of malignant tumors is not understood. Most of the effects
have been observed at field strengths that are many times higher than the ambient fields which
are the putative  cause of the childhood cancers in residential situations; as a consequence,
many of the candidate mechanisms may not be really involved in the response to low
environmental fields. The same issue arises in the evaluation  of chemical agents.
10/18/90
1-4

-------
                            DRAFT-DO NOT QUOTE OR CITE


    In conclusion, several studies showing leukemia, lymphoma, and cancer of the nervous
system in children exposed to magnetic fields from residential 60-Hz electrical power
distribution systems, supported by similar findings in adults in several occupational studies
also involving electrical power frequency exposures, show a consistent pattern of response
which suggests a causal link.  Frequency components higher than 60 Hz cannot be ruled out
as contributing factors. Evidence from a large number of biological test systems shows that
ELF electric and magnetic fields induce biological effects that are consistent with several
possible mechanisms of carcinogenesis.  However,  none of these processes has been
experimentally linked to the induction of tumors, either in animals or in humans by EM-field
exposure. The particular aspects of exposure to the ELF fields that cause these events are not
known.
    In evaluating the potential for carcinogenicity of chemical agents, the U.S. EPA has
developed an approach that attempts to integrate all of the available information into a
summary classification of the weight of evidence that the agent is carcinogenic in humans.  At
this time, such a characterization regarding the link between cancer and exposure to EM fields
is not appropriate because the basic nature of the interaction between EM fields and biological
processes leading to cancer is not understood.  For example, if induced electrical currents
were the causative factor, then exposure to electric as well as magnetic fields would be
important and the effect would be more severe as the frequency increases. But if the direct
magnetic field interaction were the  critical factor, then the ambient static magnetic field, as well
as the alternating  magnetic field, would be critical, and the effect may  be confined to specific
frequencies, resulting in an extremely complicated dose-response relationship. In addition, if
they were shown to be causative agents, these fields probably exert their effects via other
chemical and environmental factors rather than directly causing events known to be causally
related to carcinogenic processes, as with genotoxic chemical agents.
    Because of these uncertainties, it would be inappropriate at this time to classify the
carcinogenicity of EM fields in the same way as the Agency does for chemical carcinogens.
As additional studies with more definitive exposure assessment become completed, a better
understanding of the nature of the  hazard will be gained. With our current understanding, we
can identify 60-Hz magnetic fields from power lines and perhaps other sources in the home as
a possible, but not proven, cause of cancer in humans.  The absence of key information
summarized above makes it difficult to make quantitative estimates of risk. Such quantitative
estimates are necessary before judgments about the degree of safety or hazard of a given
exposure can be made.  This situation indicates the need to continue to evaluate the
                                         1-5
10/18/90

-------
                           DRAFT-DO NOT QUOTE OR CITE
 information from ongoing studies and to further evaluate the mechanisms of carcinogenic
 action and the characteristics of exposure that lead to these effects.
10/18/90
1-6

-------
                            DRAFT-DO NOT QUOTE OR CITE


                2.  MECHANISMS OF ELECTROMAGNETIC INTERACTION
    Electromagnetic (EM) fields with certain exposure parameters do affect biological systems,
 as demonstrated by the studies described in this report. These experimental findings raise a
 number of pressing questions. Primarily, how can the effects of an arbitrary EM-field exposure
 be predicted? In the present context, can environmental EM-field exposures be carcinogenic?
 Experimental data directly addressing the latter question are extremely limited. The biological
 effects described in this report have served mainly to indicate that present biological models
 are incomplete and to stimulate the development of new theories.  Bioeffects have been found
 to result from induced currents far weaker than normal physiological currents, and at imparted
 energy levels a fraction of the average thermal energy (per particle). Only by identifying the
 mechanisms through which EM fields interact with biological systems can the larger questions,
 such as the potential for carcinogenicity, be fully addressed. This section will briefly review
 several major interaction mechanism theories. In preparation for this review, a discussion of
 the properties of EM fields and field coupling with the body is presented.

 2.1. PHYSICAL PROPERTIES OF ELECTRIC AND MAGNETIC FIELDS
    Although electric and magnetic fields are used widely in modern technology, most persons
 outside of the scientific and engineering community have few reference points with which to
 understand them. The most common experiences include static electric fields produced
 indoors in dry climates and static  magnetic fields around permanent magnets.
 Electromagnetic radiation, or at least the effect of it, is familar to almost everyone in the form of
 visible light, radio transmission, and x-rays.  The purpose of this section is to provide a brief
 description of the properties of electric and magnetic fields and of how they interact with
 matter. This knowledge has been applied to human exposures to electric and magnetic fields
 for several decades, and many of the details of interaction at certain levels are now
 understood. Specific knowledge of how these interactions affect living organisms, on the
 other hand, is very limited and the subject of much investigation.
    Electric fields occur when electric charges are present.  The matter we experience
 commonly consists of almost equal numbers of positively charged and negatively charged
 particles, so even though intense  electric fields are occurring on a microscopic or atomic level,
the matter is practically neutral on the macroscopic scale of which we are aware (inches,  feet,
etc.). When some of the positive and negative charges are separated over distances on our
scale, we may experience obvious effects of an electric field. For example, the movement of
                                         2-1
10/18/90

-------
                            DRAFT-DO NOT QUOTE OR CITE


certain shoe materials across some carpet materials separates positive and negative charges
by a process called triboelectricity. The large separation of charges as one moves across a
room can produce a substantial static electric field. This field may cause the hair on the hands
or arms to stand up as the subject moves near a grounded conductive object and result in a
physical sensation. If the conductor is physically contacted, a shock may occur as electric
current (the movement of charges) flows, returning the separated electric charges to a more
neutral condition.
   This example demonstrates several of the principles of electric field theory. Separating
opposite charges results in a potential difference that is measured in the unit of volts (V). AH
units described in this chapter, unless otherwise specified, are in the Standard International
(SI) system, which is now almost universally accepted.  For a given system, the voltage
increases as more charges are separated,  in the above example, increased movement across
the carpet separates more charge and results in a larger potential difference. The electric field
is a description of the force that a  unit charge will experience at any point in space. Electric
charges attract more strongly as they are moved closer together in a way similar to bar
magnets.
   If a conducting path is provided between areas of separated charge, the charges, usually
in the form of electrons or ions (atoms with an unequal  number of positive and negative
electric charges), will flow between the two regions. This flow of charge is called electric
current and is measured in units of amperes (A). Electric charge is measured in coulombs (C).
Current is simply the number of coulombs of charge that flow through a given region per
second.
   The relationship between some of the above quantities is easier to describe using a
different example.  Imagine two large parallel sheets of a conductive material such as copper
separated by a small distance, d.  If charges from one sheet are removed and placed on the
other, a potential difference, V, will be produced. The strength or magnitude of the electric
field between the plates is given by,
                 E=V/d.
   The unit for electric fields is volts/meter. Note that either increasing the potential difference
or decreasing the separation between plates results in an increase in the electric field strength.
   Electric fields are known as vector fields, meaning that they specify a magnitude and a
direction for each point in space. The direction of an electric field at a given point is defined as
       10/18/90
2-2

-------
                             DRAFT-DO NOT QUOTE OR CITE

 the direction in which a small positive charge will move if placed at that point. For a parallel
 plate system, the electric field points from the positive plate towards the negative plate.  No
 current can flow between the plates because they are separated  by air, which is an insulator or
 dielectric.  Parallel plates are often used in laboratory experiments because they create a
 relatively large region where the direction of the electric field is uniform. In general, and often
 in environmental exposures, the direction of the electric field can  vary for each point in space.
    Magnetic fields occur whenever charges are in motion, as in the case of an electric current.
 Permanent magnets may appear to be an exception to this principle, but the field of these
 magnets is caused by the orbital motion of the electrons in the magnetic material. Magnetic
 fields exert a force on electric charges, but only charges that are  in motion.  Thus, a magnetic
 field will exert a force on a current carrying conductor, or on ions  or electrolytes moving
 through the bloodstream. A simple example can be used to describe the units of magnetic
 field. Imagine a long thin wire carrying a current, I. The magnitude of the magnetic field (H) at
 a distance, d, from the wire is

                   H = I/(2nd).    •

    Note that the quantity (2nd) is the circumference of the circle with a radius equal to the
 distance from the wire to the measurement point. The magnitude of the magnetic field is the
 magnitude of the current divided by the circumference of this circle. The field is directed in  a
 circle around the wire such that the field vector at any point is tangent to the circle.  As for the
 case of electric fields, the direction of a magnetic field can vary for each point in space, but
 special configurations are often used in the laboratory to produce fields with a relatively
 uniform direction throughout the exposure area. Standard International units for magnetic
 fields are amperes per meter (A/m).
    Besides free charges and currents,  electric and magnetic fields also act on certain
 materials. The electric and magnetic properties of materials  are described by their
 conductivity, a, permittivity, e, and permeability,^.  Dielectrics are  materials in which charge
 separation, or polarization, occurs when they are exposed to an electric field. Another vector
field, the electric flux density (or electric displacement), D, can be  used when dielectrics are
 present.
                    D=eE
                                          2-3
10/18/90

-------
                            DRAFT-DO NOT QUOTE OR CITE


   The permittivity of air is close to that of a vacuum, e0, so that when describing an exposure
in air, a fixed relationship exists between E and D.  In most bioeffects experiments, the
exposure field, E, is described before the subject is placed in the field.  Similarly, magnetic
materials have induced magnetic moments when exposed to a magnetic field. The magnetic
flux density is related to the magnetic field by
   The permittivity or dielectric constant of biological materials is significantly different from
that of a vacuum. Further complicating matters is the fact that different tissues and even
different microscopic structures within the tissues have widely different dielectric properties.
These properties have been studied in some detail (Schwan and Foster, 1980). The
permeability of most biological materials, on the other hand, is very close to that of air or a
vacuum. Thus, in these materials, B and H can be used somewhat interchangeably.  The SI
unit for B, the magnetic flux density, is tesla (T). Another commonly used unit for B is gauss
(G), which is a cgs unit.  The conversion from gauss to tesla is 1 tesla = 104 gauss.
   Normally, we think of an electric current confined to a thin conductor and simply specify a
value of the total current. Because of the structure and widely differing electrical properties of
the body, currents flowing through the body are not uniform. In this case, it is more
descriptive to specify the current density or current per unit area flowing through certain
regions of biological specimens.  When a conductive material, such as a biological substance,
is subjected to an electric field, electric currents are induced according to
                    J=aE
where J is the current density (A/m2 in SI units), and E is the electric field at the point of
interest inside the material.
    Field theory becomes more complicated when time-varying fields are considered. Electric
and magnetic fields which vary in time are described by a frequency, the number of times that
the field oscillates per second. The unit for frequency is cycles per second or hertz (Hz).
Phenomena such as radio waves, visible light, x-rays, gamma rays, etc., are all
electromagnetic fields of different frequencies (see Figure 2-1). The dielectric properties of
biological substances, for example, vary with the frequency of electric field exposure.
       10/18/90
2-4

-------
                            DRAFT-DO NOT QUOTE OR CITE


 Extremely low frequency (ELF) fields range from 3 Hz to 300 Hz.  In this range, the electric and
 magnetic fields are considered separately for purposes of biological effects experiments. The
 radiofrequency (RF) range is generally defined as 10,000 Hz to 300 gigahertz (GHz) (1 GHz =
 one billion Hz). In this range, a significant portion of field exposures may occur in the form of
 electromagnetic radiation, where a fixed relationship exists between the electric and magnetic
 fields. ELF and RF fields interact differently with the body and are discussed below in separate
 sections for clarity.

 2.1.1. Extremely Low Frequency Fields
   An ELF electric field can be imagined by referring to the parallel plate example and
 alternating the potential on the plates so that the field constantly reverses direction.  The most
 prevalent man-made ELF field in the environment is that resulting from power lines. In this
 case, the direction of the field oscillates (changes direction and returns to its beginning
 direction) 60 times per second. The electric field in the parallel plate system connected to a
 60-Hz voltage source will reach a maximum value, gradually drop to zero, climb to a maximum
 in the other direction, return to zero, and then climb to the maximum value in the starting
 direction. This cycle will occur in one-sixtieth of a second.
   Such a field is described as linearly polarized, indicating that, even though the field vector
changes in magnitude and polarity, it is always parallel to a certain direction. Power line fields
POWER AM
UNES RADIO TV RADAR
n
in i
MFRAREOJ
1 i
j ULTRA-
5 VIOLET
1X-RA

GAMMA
RAYS
YSJ 1
T t
               ,     ,
             10*   10B    10T
10"
                                                       1017   10*   10"
     I   I   I   I   I   I  I   I   I   I   I   I   I   I   I   I   I   I  I   I  I   I   I
          100    10"    106    10s    10"   10"   10"   10*   1018   10*  10ťHZ
       t      t   I  ft  t  I  ft!            f _    t
         ELF    VLF LF MF  HF VHFUHFSHF EHF
                                                       IONIZING RAOMTION
                         Figure 2-1. The electromagnetic spectrum.
                                         2-5
                                         10/18/90

-------
                            DRAFT-DO NOT QUOTE OR CITE


may have a more complex polarization.  High voltage lines, for example, often consist of three
lines with differing time phases. This term means that the voltages and currents in the different
lines reach their peak values at different times. The fields produced by the three lines add
together at each point in space and result in a field vector that moves in an ellipse rather than
back and forth along a single line. Under certain conditions, the field vector may trace a circle
in space during each cycle. This case is referred to as circular polarization.

2.1.2.  Radiofrequency Fields
    An interesting property of electromagnetic fields is that a changing electric field creates a
magnetic field and vice versa. Such effects along with the other properties of electromagnetic
fields are described by Maxwell's equations. One solution of Maxwell's equations shows that
this field creation will result in the transport of energy through space in the form of
electromagnetic radiation. The magnitude of the field produced by a change in the other field
is proportional to the time-rate-of-change. ELF fields that change slowly produce very little of
the other field. The parallel plate system described above can create a 60-Hz electric field, but
this field is changing so slowly that very little magnetic field component is created. If the same
system were driven by a 1 MHz (one million hertz) signal generator, the rapid change in the
electric field would create a substantial magnetic field and result in some of the energy
propagating away as electromagnetic radiation.  Electromagnetic radiation travels in a vacuum
at the speed of light, c, which is 3 x 108 meters/second.  This radiation may also be described
by the wavelength, A, which is related to the frequency, f, by Af=c. The unit for wavelength is
the meter (m).
    The electric and magnetic field strengths in a plane electromagnetic wave are related by

                 E/H=377 ohms.

The value 377 ohms is called the characteristic impedance of free space.  For this case, and to
a close approximation for some experimental cases, knowledge of either E or H implies both
the magnitude and direction of the other. Also for this case it is possible to define the power
density, S, of the wave. The power density is a vector quantity describing the rate at which
energy is transmitted through a unit area, and the magnitude is calculated for a free-space,
plane wave according to

                  S = E X H = E2/377.
        10/18/90
2-6

-------
                            DRAFT-DO NOT QUOTE OR CITE


    Power density in SI units is given in watts per square meter (W/m2), but RF bioeffects
exposures are often given in milliwatts per square centimeter (mW/cm2) as a more convenient
unit (1 Mw/cm2 = 10 W/m2). A power density of 1 Mw/cm2 for a free-space, plane wave,
translates to an electric field strength of 61.4 volts per meter (V/m) and a magnetic field
strength of 0.163 A/m or a flux density of 0.2 microtesla (ŤT). Power density can be a useful
measure because some bioeffects studies consider the amount of energy input to the
biological system from the electromagnetic field.
    A common error is to assume that power density can always be related to field strength in
the simple way shown above. In most real world cases, the relationship between field strength
and power density is complex and power density units are inappropriate.  One example is the
reactive near-field of an antenna, which generally extends from the antenna to a distance of
one wavelength. In  this region, a significant fraction of the energy stored  in the field collapses
back  into the antenna during each cycle rather than radiating away. Because little of this field
energy is transmitted as electromagnetic radiation, the power density is significantly lower than
would be predicted  by the above formula.  Reactive fields are common in  the environment,
especially at low frequencies. For example, the wavelength of power frequency fields (60 Hz)
is 5 million meters, indicating that all significant exposures occur in the reactive near-field of
the source (power lines, domestic wiring, appliances, etc.).
    In reactive fields, no fixed relationship exists between the electric and magnetic field
components for all cases. The parallel plate system described earlier, for  example, will
produce a ratio of electric to magnetic field strength much greater than 377 ohms at low
frequencies. A calculation of power density in this exposure field would show very low values.
This result is misleading because a dielectric or conductive object placed  in the field will
absorb more power  (energy/time) than is predicted to be incident on the object by the power
density calculation.  In such cases, the object is absorbing stored energy from the electric field
as the movement of charged particles or polarization of the dielectric produces thermal motion
in the object.  Similarly, a conductor or magnetic material will absorb energy from  a pure
(time-varying) magnetic field. Eddy currents (discussed later) in conductor and polarization
effects in magnetic materials result in heating.
   Nonplane wave conditions are common in environmental exposures and most laboratory
exposure systems.  Exceptions include transverse electromagnetic (TEM)  cells, which are
specifically designed to maintain plane wave  conditions, and anechoic chambers  in which
microwave frequencies are radiated from antennas and absorbed at the walls to eliminate
reflections.  (Waves  in which all of the electric and magnetic field components  are transverse
                                         2-7
10/18/90

-------
                           DRAFT-DO NOT QUOTE OR CITE


to the direction of propagation are called transverse electromagnetic waves). Many
environmental RF waves, which initially approximate plane waves, reflect off the ground and
other conductors resulting in standing waves.  EM radiation from a broadcast antenna, for
example, may reflect from the ground to create a standing wave which consists of alternating
regions of purely electric and purely magnetic fields in a stationary pattern. If the wavelength
Is short enough (<6 meters or >50 MHz), an individual standing in the field may be exposed
to both regions at different body locations. Power density calculations are not appropriate in
such conditions.
   A more meaningful quantity in many cases is the specific absorption rate (SAR). This
quantity is a measure of the power absorbed per unit mass of the biological specimen under
study.  The common unit is watt per kg (W/kg). SAR is difficult to predict analytically except in
geometrically simple objects with homogeneous electrical properties. Numerical techniques,
however, such as finite element analysis, have been used to calculate SAR's effectively in the
human body (Gandhi, 1982). SAR is particularly useful in bioeffects studies investigating
thermal effects. Until recent years, many researchers believed that all EM bioeffects could be
explained in terms of the total energy deposited or excess heating in the body resulting from
EM-field exposure.  Measurement and specification of SAR in a given experiment eliminate
ambiguities that may result from specification of external field strength or power density alone.
As described earlier, biological specimens can absorb energy from reactive fields for which
power density calculations are meaningless.  Clearly, power density cannot serve as a
surrogate for SAR.
    Energy absorbed from electromagnetic field exposures can be compared to the rate at
which  energy is produced by the body during  normal metabolism. The metabolic energy
generated by a 70-kg "standard" man, for example, ranges from a basal rate of about 1 W/kg
to a maximum of 21 W/kg during strenuous exercise (Polk and  Postow, 1986).  The 1982
American National Standards Institute (ANSI) standard for electromagnetic fields limits human
exposures to a maximum of 0.4 W/kg averaged over the entire body (ANSI, 1982).

2.2. COUPLING OF ELECTRIC AND MAGNETIC FIELDS WITH THE BODY
    The interaction of electric and magnetic fields with the body can be divided into
macroscopic and microscopic aspects.  Macroscopic interactions are those occurring as a
result of introducing a large dielectric object into the field. Viewed in this way, the body may
be thought of as an antenna that absorbs energy from the field. Microscopic interactions are
those occurring on the cellular or sub-cellular  level, such as induced membrane potentials,
       10/18/90
2-8

-------
                              DRAFT-DO NOT QUOTE OR CITE


 changes in ion transport, etc. Although the organism as a whole is exposed to the external
 fields, the tissues and cells experience a substantially different set of exposure parameters.
 Many details of macroscopic interactions are presently understood, including total energy
 absorption, induced surface charges, total currents, and others. Interaction mechanism
 theories generally address microscopic problems, assuming that local exposure conditions (at
 the tissue or cell level) have already been defined by evaluation of the macroscopic coupling
 effects. Table 2-1 illustrates that the transformation of external field  into internal field strengths
 and currents is affected by several factors. These factors are discussed briefly in the following
 sections, with ELF and RF effects separated  into separate sections for simplicity.  Detailed
 discussions of these relationships can be found in publications by the National Council on
 Radiation Protection (1981), Polk and Postow (1986), Tenforde and Kaune (1987), and Guy
 (1987).
    Internal exposure conditions can, in principle, be derived from the interaction of these
 fields with the organism. In practice, the precise internal field strengths and  current densities
 are extremely difficult to predict because of the complex structural variations in the body and
the highly heterogeneous electrical  properties of living tissues.  Further research is required to
determine the current distributions and internal field strengths at specific points in the body.
Such microdosimetry will be required to define the exposure parameters occurring at sites of
interest in the body in order to fully evaluate interaction mechanism theories.
                     TABLE 2-1. FACTORS AFFECTING INTERNAL EXPOSURE CONDITIONS
   Ambient Electric and
    Magnetic Field
Biological System
Internal Electric and Magnetic Fields
and Current Density
    Ei(t)
            Hx(t)
   External Exposure
Biological System
Internal Exposure
   Frequency
   Modulation
   Intensity
   Polarization
Geometry (size, shape)
Orientation
Heterogeneous tissue
Electrical properties
Coupling with other
 nearby conductors
Grounding condition
Internal fields
Current densities
Local variations due to structure
Microscopic variations due to membrane
 impedances, etc.
                                            2-9
                                                       10/18/90

-------
                            DRAFT-DO NOT QUOTE OR CITE


2.2.1. Extremely Low Frequency Fields
   ELF electric and magnetic fields vary so slowly that they can be considered quasi-static
when considering interactions with the body. This approach is possible because the
dimensions of the body are very small with respect to the wavelength of the field. When a
perfect conductor is placed in a static electric field, currents are induced that rearrange electric
charges. The new arrangement of charge produces another electric field that totally cancels
out the first field on the inside of the conductor.  A similar situation occurs when a human or
animal is exposed to a static electric field, except that the finite conductivity and dielectric
properties of the body prevent total cancellation of the field internally. The electric field is,
however, weaker inside the body than out by many orders of magnitude.
   At ELF frequencies, currents flow continuously as the direction of the electric field reverses
polarity. Internal electric fields are produced because current flowing through a finite
conductivity causes a potential drop. These internal electric field strengths are still orders of
magnitude smaller than external fields, but the ratio depends on the frequency of the field.  If
the exposure (external field before the subject is introduced into the field) electric field strength
is held constant, the ratio of internal to external field strength, as well as the total induced body
current, increase with increasing frequency of the field. At 60 Hz,  a 1  kilovolt per meter (kV/m)
exposure electric field will produce internal electric fields of only about 0.1  millivolt per meter
(mV/m) to less than 100 mV/m (Kaune and Gillis, 1981). Total short-circuit current (the current
passing between the subject and ground) for the same exposure  is predicted to be about 15.3
microamperes (aA) (Kaune et al., 1987).
    The charge redistribution caused by electric field exposure also results in variations in the
charge density on the surface of the body. These charges produce changes in the electric
field strength on body surfaces. For example, a grounded, erect human exposed to a vertical
60-Hz electric field strength of 1 kV/m will experience surface electric field strengths of up to 18
kV/m at the top of the head (Kaune and Miller, 1984).  A human or animal in an electric field
thus significantly disturbs both the magnitude and direction of the field in its vicinity.
    ELF currents can also be caused or enhanced by contact with conductive objects in the
field. A motor vehicle under a power line, for example, will be exposed to a 60-Hz electric field,
but current may not be able to flow to ground because of the insulating properties of the tires.
A grounded or partially grounded human touching  the vehicle will experience body currents
with a magnitude depending of the field strength, vehicle size, and other considerations. In
high field strength areas, such currents can  represent a shock hazard.
        10/18/90
2-10

-------
                            DRAFT-DO NOT QUOTE OR CITE


    Biological substances, with a few rare exceptions, are nonmagnetic, meaning that the
magnetic properties are similar to those of air or a vacuum.  Consequently, ELF magnetic
fields are practically the same inside the body as out, and the presence of the body does not
significantly affect the magnitude or direction of the magnetic field. This fact simplifies the
development of interaction mechanism theories based on magnetic field exposures because
the magnetic field strengths at the tissue or cell level are known.
    Changing magnetic fields creates electric fields, as pointed out earlier. These electric
fields occur in a circular pattern in planes perpendicular to the direction of the changing
magnetic field. If these induced electric fields are inside a conductive object, circular currents
known as eddy currents will flow. The intensity of the induced electric fields (and
consequently of the eddy currents) depends on the magnetic field strength, the time rate of
change of the magnetic field, and the radius of the circular path. A result of the path radius
dependence is that higher currents will be induced in larger cross-section areas of the body,
such as the torso, during human exposures.
     Eddy currents can  also be created by the movement of a conductive object in a static
magnetic field, such as the earth's field, if the movement causes the total magnetic flux
through the closed conductivity path to change.  In either case (i.e., either the magnetic field is
time-varying and the conductive path is fixed, or the magnetic field is static and the conductor
is moving), the phenomenon is referred to as "Faraday induction" of currents.  A typical
domestic 60-Hz exposure of 0.1 microtesla (wT) (Kaune et al., 1990) will produce in a human
with a "radius" of 0.1 meter, a current density of about 0.19 microamperes per square meter
(wA/m2) [1.9 x 10~5 microamperes per square centimeter (juftjcm2)].  For comparison
purposes, the external electric and magnetic field strengths required to produce an internal
electric field of 10 mV/m are shown in Figure 2-2. This figure is based on a simple spherical
model of a human. Geometrical enhancements of electric field strength occurring in humans
will reduce  the external electric field needed to produce the 10 mV/m internal E field. It has
been estimated that power line magnetic fields produce internal electric field strengths in
humans which are only about 15% as high as those produced by power line electric fields
(Tenforde and Kaune, 1987).
    Magnetic fields also produce a force on electric charges in motion as discussed earlier.
This force, known as the Lorentz force, acts only on charged particles moving with a
component of motion perpendicular to the magnetic field.  It acts in a direction perpendicular
to the direction of motion and can result in resonance effects which will be discussed later.
The Lorentz force can change the direction of motion of moving charged particles, but it does
                                         2-11
10/18/90

-------
                           DRAFT-DO NOT QUOTE OR CITE
                      101  102  10a  104 105  10
                              FREQUENCY  (Hz)
Figure 2-2. External electric and magnetic field required to obtain an internal E field of 10 mV/m.
SOURCE: Polk, 1986a.
not add energy to the system. In the case of moving body fluids, a magnetic field Lorentz
force can cause a separation of charged electrolytes of opposite polarities across the diameter
of the blood vessel (Persson and Stahlberg, 1989). This charge separation also occurs in a
variety of nonbiological substances and is known as the Hall effect. Such effects are very
weak with static fields having a flux density similar to the earth's field (less than 50 ^T), but are
more important in nuclear magnetic resonance (NMR) imaging devices, where static magnetic
fields of up to 2 tesla (T) are used.  The Lorentz force also leads to the magnetohydrodynamic
effect which reduces flow in moving electrolyte solutions like blood. A 2-T static magnetic field
is predicted to reduce axial flow in the human aorta by 1% to 2% (Easterly, 1982). Perspective
on the intensity of such exposures can be gained by comparison with common exposures. A
2-T magnetic field is over 20,000 times greater than the earth's magnetic field and more than
two million times greater than a typical domestic 60-Hz magnetic field exposure.
   Magnetomechanical effects are another class of direct magnetic interaction. One such
effect is the tendency to orient molecules with certain magnetic properties. It has been
observed experimentally in biological macromolecular assemblies such as retinal rod outer
segments exposed to a static magnetic field of 1T [World Health Organization (WHO), 1987].
      10/18/90
2-12

-------
                            DRAFT-DO NOT QUOTE OR CITE


Magnetic fields can also affect the electronic spin states of certain reaction intermediates and
alter the yield of reaction products. An example is the change induced in the charge transfer
reaction occurring during bacterial photosynthesis (Blankenship et al., 1977). This effect is
produced at static magnetic field exposures of 10 millitesla (mT).
    An interesting direct magnetic effect is the induction of magnetophosphenes. This term
describes the sensation of flickering light in the eye when it is exposed to ELF magnetic fields
greater than about 10 mT and with frequencies over 10 Hz.  Several types of evidence suggest
that the magnetic field interaction leading to magnetophosphenes occurs in the retina
(Tenforde, 1990).  The exact mechanism for this effect is not known at present, but it does
illustrate the principle that magnetic fields can interact with the body without substantial
heating or energy deposition. The other direct magnetic interactions described above have
been studied mostly in  regard to intense static magnetic fields and are mentioned here only to
introduce some of the known methods by which magnetic fields can affect living organisms.
Their importance in ELF bioeffects is not known.
    The International Radiation Protection Association (IRPA) recently published interim
guidelines on limits of exposure,to 50-Hz and 60-Hz electric and magnetic fields (IRPA, 1990).
These guidelines are based strictly on induced body currents and known biological effects
associated with these currents.  The IRPA committee which developed the guidelines reviewed
the interaction principles discussed here (WHO, 1987; WHO, 1984) and concluded that
insufficient evidence existed to base the guidelines on other types of interactions.  While it is
agreed that other  mechanisms may be important, it is also generally accepted that they are not
well enough established to use predictively.
    Coupling of ELF fields with the body is complicated by a number of factors. Changing the
orientation of the body, for example, can significantly alter coupling  and the resulting internal
field strengths  and current densities. Internal exposures are also affected by polarization of
the field, grounding condition, contact with other conductors, and body geometry. The latter
point is particularly important when evaluating laboratory bioeffects experiments.  Internal
electric fields and  current densities in a human and a rat exposed to the same external ELF
electric field are substantially different.  Exposure fields in laboratory experiments are
sometimes scaled to compensate for differences in body geometry,  but without a detailed
understanding of interaction mechanisms, it is not clear which parameter to scale for.
    There are fundamental differences between ELF interactions and either ionizing or RF
radiation interactions. ELF electric and magnetic fields couple poorly with the body.  In
contrast to some RF and microwave exposures, the body absorbs very little energy from ELF
                                         2-13
10/18/90

-------
                            DRAFT-DO NOT QUOTE OR CITE


fields at environmental exposure levels.  ELF bioeffects must therefore result from mechanisms
other than heating of body tissues.  Another important point is that ELF fields cannot break
chemical bonds or cause ionization, as can x-rays, which result in known damage to body
tissues and genetic material.
    Much attention has been focused on the cell membrane as a likely site of ELF bioeffects.
The dielectric constants of cell membranes are so high at ELF frequencies that induced ELF
body currents flow mainly around the outside of the cells.  Consequently, the interior of the cell
(cytoplasm) is shielded from applied fields and currents. Cell membranes have a natural
potential that is important in regulating ion concentrations. ELF field exposures impose
artificial membrane potentials, but the magnitude  of these induced potentials is at least 100
times less than the natural potential. It is not obvious how the diminutive induced potentials
might cause an effect, and many of the ELF interaction mechanism theories attempt to
address this problem (see Section 2.4).
    A 60-Hz magnetic field exposure of 0.1 fiY will produce a current density in a human of
about 0.19/4 A/m2, as discussed above.  An electric field exposure of 10 kV/m is predicted to
produce a maximum current density in a human of about 2.0pA/m2 (Kaune and Phillips, 1980).
These values are below physiological current densities and at or below predicted noise for
nerve impulses shown in Figure 2-3. Under normal circumstances, it is considered that
induced signals must exceed the  noise floor in  order to be detected.  In  engineering
terminology, the signal-to-noise ratio must be greater than one.
    Other typical body currents range from 5 x 10"2 to 4 A/m2 (5 to 400/^A/cm2), typically
around 10 Hz for the electroencephalogram (EEG) and up to 1 A/m2 (lOO^A/cm2) near the
firing heart cell (EKG)(Barnes, 1986). Note that the predicted induced current densities from
ambient electric and magnetic field sources are orders of magnitude lower than endogenous
currents associated with normal physiological processes.  Many investigators have found  it
difficult to understand how such small currents could cause adverse biological effects. It
should be noted, however, that the induced current predictions do not consider the extreme
complexity of body structure and the interfaces of materials with differing electrical properties.
Such heterogeneity may result in intensified currents or other effects at certain locations.
Further research will be required to address these questions.
    A recent analysis (Weaver and Astumian, 1990) has directly addressed the problem of the
signal-to-noise ratio of field-induced potentials to thermal noise in biological systems. For
example, the cell membrane is often considered a likely site for interactions of EM fields with
       10/18/90
2-14

-------
                            DRAFT-DO NOT QUOTE OR CITE


living organisms. Fundamental physical theory can be used to predict fluctuations in cell
membrane potentials resulting from thermal noise. The fact that certain bioeffects have been
reported to occur at field strengths that induce changes in membrane potential smaller than
predicted on the basis of thermally generated fluctuations is perplexing.  Some investigators
have even rejected the validity of the experimental findings because of this apparent
inconsistency.
   Weaver and Astumian (1990), as well as Bawin and Adey (1976), have noted that previous
analyses have considered noise contributions over a large frequency band. Thermal noise in
electrical resistance is proportional to the square root of the bandwidth, so a larger frequency
range implies more noise. If certain biological systems respond to signals only within a limited
                        10
                                  10  102  1Q3 104.105 106
                                         FREQUENCY
Figure 2-3. The minimum current densities observed to cause various biological changes: (A) threshold for
         stimulation for muscle or nerve cells; (B) threshold for fibrillation in a dog; (C) threshold for dias-
         tolic stimulation in dog hearts; (H) threshold for phosphene release; (Ji.Ja) noise current or cur-
         rent density fluctuations in nerve impulses.
SOURCE: Barnes and Seyed-Modani, 1987.
                                          2-15
10/18/90

-------
                            DRAFT-DO NOT QUOTE OR CITE


bandwidth, then the thermally generated noise within that effective frequency band is reduced.
Thus, much weaker signals from external sources may be detected. Another argument
presented for possible biological sensitivity to weak electric fields is that periodic signals may
be signal-averaged over some time period to improve the signal-to-noise ratio.

2.2.2. Radiofrequency Fields
    Radiofrequency fields oscillate thousands to billions of times more rapidly than ELF fields,
and their interactions with the body are more complex.  The body can absorb significant
amounts of energy from environmental RF exposures, and safety standards have been
established to limit the amount of heating from such exposures (by limiting field strengths).
For purposes of illustration, the body can be thought of as a lossy antenna; that is, it absorbs
energy from the electromagnetic field, re-radiates or reflects a portion of the energy, and
dissipates the rest through resistive and dielectric losses. The efficiency with which an
antenna absorbs energy from an electromagnetic field depends on its shape, orientation,
coupling with other conductors, and size relative to the wavelength of the field. Figure 2-4
illustrates this principle for spheroidal models of a human and a rat in free space.  The energy
absorption curves are labelled E, H, and K, depending on whether the long axis of the
spheroid is aligned with the electric field, magnetic field, or direction of propagation,
respectively.  Note that when the electric field is aligned with the long axis of the human model,
energy absorption peaks around 70 MHz (in the TV portion of the EM spectrum, see Figure
2-1). The peak is known as the frequency of whole-body resonance and is simply the
frequency at which the body dimensions appear as a tuned or resonant antenna to the field.
Energy absorption is greatest for the electric-field polarization as shown in Figure 2-4.
Resonant frequency is also affected by grounding conditions. A grounded subject will
resonate at about one-half the frequency of an ungrounded subject.
    It can be shown using electromagnetic theory that an external electric field does not
penetrate a perfect conductor (material with infinite conductivity). Exposed to an  external
electric field, currents will flow only on the surface of the conductor. These currents will create
electric fields that cancel the external field so that no internal electric field or internal current
densities exist. Real objects, like the human body, do not have infinite conductivity and as a
result, small internal electric fields exist.  In real objects, both the dielectric constant and
conductivity vary with frequency.  Schwan and Foster (1980) have characterized the complex
behavior of these electrical properties across the RF portion of the spectrum. When the body
is exposed to an external RF electric field, an internal electric field is produced which in turn
       10/18/90
2-16

-------
                             DRAFT-DO NOT QUOTE OR CITE
                                                                10a
                                     FREQUENCY (MHz)
Figure 2-4. Predicted frequency dependence of absorbed energy in spheroidal models of biological bodies.
The energy absorption curves are labelled E, H, and K, depending on whether the long axis of the spheroid
is aligned with the electric field,  magnetic field, or direction of propaganda, respectively.
 drives internal currents.  The distribution of current density inside the body may be even more
 complex than for the case of ELF exposures.
    Simplified models can be used to study variations in internal field strengths and currents as
 a function of frequency of the external electric field. Internal electric field strength varies in a
 fashion similar to the SAR curve of Figure 2-4. At whole-body resonance, the internal electric
 field strength is comparable in magnitude to the external electric field strength (Barber, 1977).
 This result is in sharp contrast to the case of ELF exposures where internal electric field
 strengths are orders of magnitude lower than the external fields.
    In general, internal electric field strengths and induced currents increase as the frequency
 of the external electric field strength is increased up to whole-body resonant frequency. Above
 the resonant frequency, these quantities,  as well as SAR, are reduced and level off, as seen in
 Figure 2-4. Figure 2-5 illustrates the variation in body current versus exposure frequency as
 predicted  numerically by a lumped impedance model of man.  Induced current also peaks at
 whole-body resonance. Recent studies have indicated that RF currents of 600 to 800
                                           2-17
10/18/90

-------
                            DRAFT-DO NOT QUOTE OR CITE


milliamperes (mA) can be induced in humans exposed to 1 mW/cm2 (61 V/m) fields near
whole-body resonance frequencies (Chen and Gandhi, 1989).
   Actual induced current densities vary substantially as a function of position within the body
due to contrasting tissue electrical properties and body geometry.  The distribution of current
will also change as the subject moves around in the field, altering body geometry, orientation
with respect to the field, and coupling with other objects and ground.  Current density, at any
frequency, is directly proportional in magnitude to external field strength if other exposure
parameters are held constant.
   The frequency of whole-body resonance, as well as the general shape of the induced
current vs. frequency curve varies with body shape, size, and grounding conditions. The
resonant frequency of  a small child is significantly higher than that of an adult and much lower
than that of most animals used in laboratory studies.  The resonant frequency for a rat is about
700 MHz in contrast with about 70 MHz for an adult human (see Figure 2-4). These principles
provide some insight into the question of frequency scaling in animal exposure systems. In
general, a given electric field exposure will produce a different SAR in a human than in a rat, as
                       0.03
                       0.00'
                                EXPOSURE E-FIELD -1 V/m
                                1.78m high,  180 Ibmole
                                Calculated by  NEC3Model
                           10'
10"       10        10"
   FREQUENCY (Hz)
10s
Figure 2-5.  Plane wave induced current in humans vs. frequency.
SOURCE: Adapted from Guy, 1987.
      10/18/90
    2-18

-------
                           DRAFT-DO NOT QUOTE OR CITE


illustrated in Figure 2-4. One way to equalize absorbed power in humans and animals is to
conduct the animal experiments at a higher frequency so that the SARs are the same.  While
this approach addresses the problem of total absorbed power, issues of body geometry and
local current densities are not as easily resolved.
   Differences in geometry between a rat and a human also dictate that different current
densities will be produced in a given tissue type (e.g., the liver) in the two species.  Further,
tissue electrical properties will differ at the two frequencies.  For thermal effects experiments, it
may be sufficient to equalize SARs, but nonthermal experimental results in animals or tissue
cultures cannot be accurately applied to humans until the quantities necessary for the
interaction are identified. Otherwise, one does not know whether to adjust for frequency, SAR,
current density, or internal field  strength. This point emphasizes the need for interaction
mechanism theories that will afford some criteria for sorting through the practically infinite
range of possible EM-field exposure conditions.
   Internal magnetic field strengths are not necessarily the same as those outside the body
for radiofrequency exposures.  As exposure frequency is increased, the magnitude of induced
eddy currents increases. These eddy currents produce a secondary magnetic field that tends
to cancel the incident magnetic field inside the body. This shielding capability of an object
with finite conductivity (such as the body) is quantified in terms of the skin depth. Skin depth
is the distance into the material at which both the electric and magnetic fields from an incident
electromagnetic wave are reduced to about 37%. Values of skin depth (or depth of
penetration) vary according to the electrical properties of the material. At 915 MHz, for
example, the skin depth in muscle tissue is about 3 cm while the skin depth for fat is about 18
cm (NCRP, 1981).  Skin depth decreases with increasing frequency, indicating that energy is
deposited closer to the surface of the body as exposure frequency increases through the
microwave range.
   The dielectric constant of cell membranes decreases with increasing frequency. As a
result,  electrical currents begin to pass through the cell membrane at frequencies of about 0.1
MHz and above, and the cytoplasm is no longer shielded as it is from ELF currents. This fact,
along with the slow response times of cellular processes (compared to radiofrequencies),
implies that mechanisms of RF and ELF biointeraction may be very different. A possibly
important effect demonstrated in the RF range is the rectification of induced currents by the
cell membrane (Barnes and Hu, 1977; Montaigne and Pickard, 1984). This process of greater
passage of current in one direction than the other results in lower frequency modulation
components being extracted from the RF signal. This effect is similar to the way the audio
                                         2-19
10/18/90 ,

-------
                            DRAFT-DO NOT QUOTE OR CITE


 information modulated on an AM radio signal is retrieved by rectification in a radio. Membrane
 rectification may be important in that it could result in low frequency signals at the cell level
 (which may be more biologically active) during an RF exposure.  Certain experiments have
 demonstrated biological responses from both ELF exposures and RF exposures which are
 modulated at ELF frequencies (Bawin et al., 1973; Blackman et al., 1979).

 2.3. AMBIENT EXPOSURE
    This section briefly reviews the magnitude of the ELF and RF fields which the general
 population experiences. The section points out that an adequate analysis of exposure needs
 to include several other factors in addition to the magnitude of the ambient fields.
    For 60-Hz electrical power, several good reviews are available [WHO, 1984;  New York
 Power Lines Project, 1987; Office of Science and Technology Policy (OTA), 1989; Florig et al.,
 1987]. A description of electric and magnetic field measurements associated with the electric
 power distribution system  in Seattle residences has been published by Kaune et al. (1987).
 The analysis of Florig et al. (1987) and the OTA (1989) has been summarized in graphs of
 electric field intensity and magnetic flux density as a function of distance from the source,
 reproduced here as Figure 2-6. In the graphs, the term "distribution" refers to 5-35 kV primary
 lines leading from the substation to the neighborhood step-down transformers and the
 115/230 V secondaries leading to individual buildings.  The term "appliances" refers to such
 items as electric shavers and blankets. The diagrams do not explicitly include 60-Hz ground
 currents carried by plumbing pipes, a factor that has recently received more attention.  Other
 sources of exposure, such as television receivers, video display terminals,  lighting dimmer
 switches, and wireless telephones, have significant components of power at higher
 frequencies, and are also excluded from the graphs. In addition, the earth's static magnetic
 field and its perturbations, strong direct currents, and proximity to ferromagnetic materials
 such as telephones, are other sources of magnetic field exposures. No information is
 available about personal exposures to these sources.
   Ambient RF exposures have been evaluated by scientists at the U.S. Environmental
 Protection Agency (1986).  They described population-weighted average power levels in urban
 areas calculated from measurements in 15 large cities. The estimated residential median
 exposure for people in these areas was 0.005 microwatts per square centimeter  (aW/cm2) at
 FM, radio, and television broadcast frequencies and 0.019^W/cm2 at AM broadcast
frequencies. This radiation has frequency components ranging from 30 Hz (audio
components of AM signals) to 806 MHz. The report concluded that there is negligible
       10/18/90
2-20

-------
                DRAFT-DO NOT QUOTE OR CITE
                     Distance from source (meters)
        10.000-
           0.1
                                          loo  *  idoo

                      Distance from source  (meters)
Figure 2-6. Variation in electric and magnetic field intensities at ground level
          as a function of distance from the source [1 milligauss (mG) =0.1
                               2-21
10/18/90

-------
                             DRAFT-DO NOT QUOTE OR CITE


 background exposure above 806 MHz. The audio components of the AM broadcast signals
 contribute negligible absorbed power.
    As Florig et al. (1987) point out, an adequate exposure assessment must consider many
 other factors in addition to the ambient fields described above. For electric fields, the body
 size, distance, and orientation relative to the source, the presence of shields such as trees,
 buildings, and vehicles, and the body's resistance to ground are other factors that determine
 the internal electric fields and currents from a given external field. For magnetic fields, body
 size is a major determinant of induced internal electric fields and currents.  Also, since the
 induced body currents are proportional to the rate of change of electric and magnetic fields,
 the product of frequency times field strength is an important factor for field effects.
    Since the basic mechanism of the adverse biological effects is not known, the relevant
 internal measure of exposure cannot be specified. Possible candidates are:  (a) internally
 induced current, spatially averaged over the body; (b) internal electric fields, spatially
 averaged; (c) current or electric fields existing at some critical organ or target site; (d) a certain
 critical frequency or set of critical frequencies for the currents or electric fields, averaged over
 either the entire body, critical organs (such as the brain), or critical cells; (e) a critical time of
 day when the induced currents or fields are capable of producing the effect; and (f)  a critical
 orientation of the external fields with respect to the earth's static magnetic field. Currently
 none of these possibilities can be ruled out, and the relevant external exposure parameters are
 dependent on which biological process is involved.

 2.4. PROPOSED MECHANISMS OF INTERACTION
    Efforts are now ongoing to develop theoretical models that can explain electric and
 magnetic field bioeffects. An interesting byproduct of research into the possible harmful
 effects of EM-field exposures is the recognition that present biological models must  have
fundamental deficiencies because of their inability to explain observed effects. Research into
the mechanisms by which EM fields interact with living systems is therefore accomplishing a
 dual purpose.  It is helping to evaluate the potential for adverse effects and at the same time
 providing new tools for investigating the basic processes of living systems. While a  much
deeper understanding of EM-field bioeffects has been developed, it is important to note that
 no single theory has provided a broad predictive ability or gained widespread acceptance in
the scientific community.  It is not yet possible, for example, to design epidemiologic studies
based on a set of exposure conditions predicted to cause certain bioeffects.  More
fundamental research will be required to reach this important level of understanding.
       10/18/90
2-22

-------
                           DRAFT-DO NOT QUOTE OR CITE


   The mechanism theories described below are attempts to model biological functions at the
cellular and subcellular level in a way that identifies processes that could be affected by EM
fields.  The primary obstacle for any such theory is to explain how weak fields, which deposit
tiny amounts of energy into living systems that are awash with much greater levels of random
thermal energy, can be detected by these systems.  Only a few of the many attempts to
achieve this goal are described here. More detailed information can be found in several
reviews (Wilson et al., 1990; Chiabrera et al.( 1985; Adey and Lawrence, 1984; Polk and
Postow, 1986; Blank and Findl, 1987).

2.4.1.  Surface Compartment Model
   A number of investigators have developed interaction mechanism theories based on
electric field driven alterations in ion transport across cell membranes. These theories attempt
to explain experimental findings such as calcium efflux and other effects which might be linked
to changes in ion concentrations or gradients. The difficulty encountered by such an
approach is that details of cell microstructure such as channel characteristics and operation
are only now being elucidated. Theoretical models must therefore either make assumptions
about these cell features, or use experimental results to infer their operation.
    One theory which has been developed extensively is the Surface Compartment Model
(SCM) reported by Blank and Findl (1987). The SCM provides an approximate solution to the
complex equations of electrical double layer theory permitting study of transient effects which
occur during electrically driven ion transport. In the SCM, the cell interior, inside surface,
membrane, outside surface, and exterior are each defined as compartments with specific
properties. A set of nonlinear, independent differential equations are then derived to describe
the system using principles of conservation of charge and mass along with chemical kinetics.
Channel function is modeled as a voltage-dependent permeability, and other properties such
as ion mobility in the  surface compartment and the rate constants for ion binding and release
are assumed.
    Cation channels are described physically as being cylindrical in shape with a constriction
at the cytoplasmic end in the resting state. The repulsive forces of bound surface charge keep
the outer portion of the channel open on the surface of the cell membrane. Gating, or opening
of the channel can result from any event that leads to changes in surface charge. For
example, mobile gating charges may result from electric field exposure, binding of ligands, or
certain enzymatic (e.g., ATPase) reactions. This depolarization shifts negative charges to the
inside of the cell membrane, which binds to the inside surface of the channel and opens the
                                         2-23
10/18/90

-------
                            DRAFT-DO NOT QUOTE OR CITE


 cytopiasmic end by electric repulsion.  Another effect of the charge shift is that concentrations
 of bound counter ions in the channel can be changed (especially if one species is
 preferentially adsorbed) such that a reversed local concentration gradient is produced.  Ion
 transport would then proceed in a direction opposite the overall gradient (i.e., ion pumping).
    The SCM offers a means for studying transient fluxes and concentration changes caused
 by time-varying currents and fields.  This capability is important because of the differences in
 time frame between electrical and chemical processes. While a membrane can be depolarized
 in less than 10 microseconds, chemical potentials change by diffusion which proceeds more
 slowly (e.g., about 1 millisecond over distances of a cell diameter).  Rapid electrical changes
 thus result in temporary electrochemical imbalances and transient fluxes. Evaluation of these
 transients requires knowledge of the surface properties (such as rate constants for binding
 and release) because of the influence of absolute concentrations on ion flux.
    Channel selectivity, for example, is explained by the SCM through the kinetics of channel
 opening rather than filtration  (channel size vs. ion size). The SCM considers the high
 concentration of sodium ions on the outside surface of the membrane and the enhanced
 potassium ion concentration on the inside surface of the membrane due to the membrane
 charge.  If the membrane is depolarized and the gate opened simultaneously, as described
 above, the ions will diffuse away from the membrane to establish electrochemical balance, and
 some of the diffusion will be through the channel. Selectivity is predicted by the SCM to arise
 from the differences in speeds of channel opening and diffusion rates. The more mobile
 sodium ions will pass through the channel in significant numbers only if the channel  opens
 before the diffusion process is complete. A slowly opening gate will miss the sodium diffusion
 but pass the more slowly diffusing potassium ions. The channels are assumed to pass both
 potassium and sodium ions, but appear selective due to gating time. Among the implications
 of this model is the inference that agents which appear to inhibit certain ion channels may
 actually be modifying the gating time.
   The ratios of ion species which carry charge are also important in the SCM. As described
 earlier, ratios of sodium and potassium ions in the layers on either side of the cell membrane
 differ because of the potential difference and concentration differences. An externally applied
 current passing through the membrane will carry more sodium if directed into the cell (positive
 charge inward), and more potassium  if directed out of the cell. The  SCM shows that these
 effects are not completely reversed when the direction of current is quickly reversed because
 of transient surface concentration changes that exist until the diffusion processes catch up
with the electrical changes. Applied alternating currents thus tend to reduce both of these
       10/18/90
2-24

-------
                            DRAFT-DO NOT QUOTE OR CITE


cation gradients as sodium ions are transported into the cell and potassium ions out. The
effects are additive up to some new equilibrium point and tend towards the same gradient
reduction on both sides of the cell.  Certain frequency dependent enhancements have been
predicted by the SCM (Blank and Blank, 1986).
   The effects described above also apply to other ions in the cell environment. Of particular
interest are predictions for calcium ions, especially since the influences of applied external
electric fields on calcium ion fluxes have been observed experimentally. Calcium is predicted
to carry about 1 % of the charge through the membrane, but because cytoplasmic
concentrations of this ion are so low, relatively small numbers of transported ions lead to large
changes in gradient.  The binding constant for calcium is high inside the cell and is associated
with a number of biological processes.
   Field-induced ion fluxes have also been described by other models. One theory describes
the external bilipid layer of the cell membrane as weakly cationic (containing bound cations),
while the inner layer is strongly anionic (Findl, 1987). The combined effect of these layers is to
act as the equivalent of a semiconductor P/N junction, passing current in one direction only.
Applied currents will transport ions out of the cell if the polarity is in one direction, and have no
effect in the opposite direction. This model predicts an overall decrease in intracellular ion
concentration when the cell is exposed to an externally applied alternating current.  Cellular
concentrations of calcium ion are normally low. During various signal transduction processes,
these levels may increase by an order of magnitude due primarily to calcium release from
mitochondria and other intracellular organelles. Any factor that prevents these signal
transients may inhibit normal cell processes and result in bioeffects.

2.4.2. Ion Cyclotron Resonance
   Another prominent interaction theory is centered around a phenomenon known as
cyclotron resonance. This well-known physical principle describes the motion  of a charged
particle in a static magnetic field absorbing energy from a time-varying electric  or magnetic
field.  The Lorentz force described earlier bends the path of charged particles moving with a
component of motion perpendicular to the direction of the magnetic field. As the particle's
direction is continuously altered, it will ultimately move in a circular or helical path around the
magnetic lines of force.  Outside the earth's atmosphere, where charged particles experience
few collisions with other particles, this circular motion can continue for long periods of time.
The charged particles in this  case are electrons and protons emitted by the sun which are
                                          2-25
10/18/90

-------
                             DRAFT-DO NOT QUOTE OR CITE
 trapped in the earth's magnetic field.  The frequency with which charged particles orbit in a
 uniform, static magnetic field is determined by
                                              m
                                                                           Equation 2-1
where  co =
       q=
       B =
       m =
              angular frequency
              charge of the particle
              magnitude of the component of the magnetic field perpendicular to the particle's direction
              of motion, and
              the mass of the particle.
    Note that the frequency is independent of the particle's velocity. The Lorentz force which
 bends the particle's path is proportional to velocity. A slowly moving particle will thus inscribe
 a smaller circle than a faster moving particle, but both will complete a revolution in the same
 amount of time.
    In situations where the charged particles lose energy by collisions with other particles such
 as in condensed media, this circular motion will eventually halt unless energy is added to the
 system. Energy can be added directly by the action of a time-varying electric field which is
 parallel to the plane of the orbit.  If the frequency of the electric field meets the requirement of
 Equation 2-1 , (i.e., 3) it will continuously add energy to the particle's motion as it orbits.  This
 condition is known as cyclotron resonance and has been used in many physics applications.
    Liboff (1985), Blackman et al. (1985), and Polk (1986b) have suggested that cyclotron
 resonance may play a role in the ELF bioeffects that have been established experimentally.
 The motivation for this approach was the pattern of frequency dependence in calcium efflux
 experiments reported by Blackman et al. (1985). The presence of discrete frequency windows
 argues against Faraday induction which would suggest increasing effectiveness of magnetic
 field exposure as frequency is increased.  Further, Blackman's results demonstrated the
 importance of the geomagnetic or other static magnetic field in producing the effects.
 Cyclotron resonance is attractive because it requires a static magnetic field and predicts that
 responses will occur at discrete frequencies. These frequencies are predicted from Equation
 2-1  using the charge and mass of various biologically active ions along with the value of the
 static magnetic field present during the experiment.  The electric field used to drive the ions is
 directed at a right angle to the static magnetic field,  and may be induced directly from an
 external electric field exposure or by Faraday induction from an oscillating magnetic field. In
order to impart energy to the ion, the frequency of the driving field must be equal to the
cyclotron resonance frequency, 3, for the given ion.
       10/18/90
                                         2-26

-------
                           DRAFT-DO NOT QUOTE OR CITE
   Although cyclotron resonance of charged particles is a well-known physical principle, it is
not clear what role this phenomenon may play in biological systems.  At this point, it is
possible only to search for bioeffects which occur at the predicted combinations of static and
alternating fields corresponding to the cyclotron resonance frequencies for biologically
important ionic species. This approach has been applied with limited success to the calcium
efflux results (Liboff, 1985; Blackman et al., 1988), but is not compatible for all the frequency
and field combinations shown to produce effects.
   The most intriguing results occurred in experiments designed specifically to investigate the
cyclotron resonance hypothesis.  In one experiment (Thomas et al., 1986; Liboff et al., 1989),
rats that had undergone prior operant conditioning were exposed to fields "tuned" to the
cyclotron resonance conditions for singly ionized lithium (60 Hz, 27.1 ^T)  in which the 60-Hz
magnetic fields were horizontal and in the plane of the earth's static magnetic field. Following
30-minute exposures, the rats  showed significant alterations in ability to perform a trained
activity (pushing a lever after a time delay).  The effect persisted at least one hour after
exposure, but less than 24 hours, and was reproducible over a several week period.  It
occurred only for the combination of static and alternating fields, and not for either field alone.
Both the static magnetic field (27.1 fiT) and the alternating magnetic field, which must be larger
than 27 [iT for the effect to occur, are within the range of normal domestic exposures (New
York Power Lines  Project, 1987).
   A second experiment designed to test the cyclotron resonance theory examined the
motility of benthic diatoms exposed to fields adjusted for cyclotron resonance of doubly
ionized calcium (Smith et al., 1987). In- this study, the motility of the diatoms across an agar
plate with a 0.25 to 0.5 millimolar (mM) calcium concentration was measured as the frequency
of the alternating magnetic field was varied.  With the  static magnetic field set at 21 /*T, diatom
motility was seen to peak at 1 6 Hz, which is a combination for cyclotron resonance in doubly
ionized calcium. Motility was also enhanced at the third, but not second or fourth harmonics.
The effect occurred only at calcium concentrations slightly less than required for normal
diatom movement. At higher concentrations (5 mM), the diatoms were fully motile and the
exposure fields produced no significant changes.  Because diatom movement depends on
transport of calcium ions across the cell membrane, this experiment implies enhanced
transport similar to that reported in the chick brain calcium efflux experiments (Bawin and
Adey, 1976; Blackman et al., 1988). A few other experimental findings have been used to
support the cyclotron resonance theory, but will not be described here (Liboff et al., 1987).
                                         2-27
10/18/90

-------
                            DRAFT-DO NOT QUOTE OR CITE
    The above findings clearly implicate the importance of the charge-to-mass ratio of
 biological ions such as appears in Equation 2-1. Efforts to develop a more detailed
 mechanistic theory based on ion cyclotron resonance, however, have been largely
 unsuccessful. The primary theoretical difficulty is that the thermal collision frequency for ions
 in solution is far greater than the cyclotron resonance frequency. Using the geomagnetic field,
 cyclotron resonance frequencies for biological ions fall in the ELF range under 100 Hz.  A
 resonating ion which therefore orbits in  10 to 100 milliseconds will experience some 1012
 collisions during this time frame (Halle, 1988).  Part of the appeal of resonance phenomena is
 that a single particle can continuously gain energy over several cycles, thus helping to explain
 how such low intensity fields can cause bioeffects. Clearly, the high collision frequency
 precludes completion of even a single orbit in a normal aqueous environment. This high
 collision frequency and consequential thermalization  of resonance-acquired energy argues
 against any significant buildup of energy by resonating ions.  A numerical model of cyclotron
 movement (Durney et al., 1988) predicts complex trajectories for free ions due to the additional
 forces exerted by the alternating magnetic field often  used to induce an internal electric field.
 The model indicates that no resonance  occurs if the collision frequency is significantly higher
 than the resonance frequency. Another issue of theoretical interest is that ions are normally
 hydrated in aqueous solutions. The additional mass of these attached water molecules alters
 the cyclotron resonance frequency predicted in Equation 2-1.
    Liboff and McLeod (1988) has attempted to respond to these questions by suggesting that
 cyclotron resonance affects ions while they are confined to helical membrane channels. The
 channel protects the ion from collisions  and provides a helical passageway corresponding to
 the predicted ion trajectory. Halle (1988),  however, has argued that the introduction of a
 helically constrained path eliminates the magnetic effect (Lorentz force  orbit). Also, there is as
 yet no convincing evidence that membrane ion channels are shaped like helical tubes.
    Recently, Lednev (1990) has proposed an alternate interaction mechanism theory related
 to cyclotron resonance which may overcome some of the problems. This theory considers
 calcium ions bound to oxygen ligands in calcium binding proteins as charged spatial
 oscillators. Such an oscillator will have a set of characteristic vibrational frequencies. A static
 magnetic field applied to this system will split excited energy levels into two sublevels
 separated in frequency by the cyclotron  resonance frequency. Application of an exposure
field at the cyclotron resonance frequency may result  in a transition between these states.
Such transitions could then affect calcium  binding rates.  This model has the interesting
characteristic of positioning the ion inside  a protein matrix, thus reducing the random thermal
       10/18/90
2-28

-------
                            DRAFT-DO NOT QUOTE OR CITE


motion that is the primary difficulty of the cyclotron resonance theory for free ions. However, it
will be necessary to verify that the vibrational states are not affected by thermal energy.
    It is clear that the multiple-collision objection must be answered in some way if the ion
cyclotron resonance hypothesis is to be plausible. The answer to these objections will
probably come in the study of ion-protein dynamics of the plasma membrane.  Nevertheless,
the attractive features of this class of hypotheses are: (a) it naturally explains
frequency-specific interactions and might ultimately be capable of explaining "intensity
windows." The frequency selectivity may be the "antenna" that is needed to "pull" weak
EM-field signals out of endogenous "noise" of background electrical activity; (b) it comes close
to explaining one of the few effects that have been experimentally induced at the low field
strengths comparable to environmental exposure (0.3fiT = 3.0 mG); namely, calcium release
from chick brain tissue; (c) since it deals with events at the plasma membrane, it has the
capability of linking ionic phenomena with  hormone-specific chemical signaling mechanisms;
(d) it is generic enough to include several different ions as candidates.

2.4.3. Cooperative Mechanisms
    An entirely different approach to explaining the mechanisms behind non-thermal effects
has been advanced by Frohlich  (1988).  It is motivated in part by assertions based on classical
linear physics, that the energy deposited in a biological system by nonthermal EM-field
exposure is insufficient to produce significant direct interactions with biological  molecules.
Instead, Frohlich and others have attempted to identify higher states of order or coherence in
biological systems which may be sensitive to weak interactions.
    The ideas described by Frohlich are difficult to comprehend in the present context because
they are based on concepts of theoretical physics rather than molecular or cell  biology.  Within
the framework of molecular biology, it is generally assumed that biological functions can be
derived by studying the chemical and structural properties at the microscopic and molecular
levels. The investigator considers average properties such as diffusion and reaction rates of
various substances.  From a physics viewpoint, however, another level of complexity exists.
Groups of particles may interact with the number of possible interactions depending
exponentially on the number of particles. Such systems are referred to as "many body
systems" and are not solved for exact motions, but described by state functions. These state
functions relate to the energy of the system and other properties.  Using this formalism, it is
found that various states of order or correlation of properties are possible at specific energy
levels.
                                         2-29
10/18/90

-------
                            DRAFT-DO NOT QUOTE OR CITE
   Examples are given of lasers and superfluid helium. In an optically pumped laser, a
substance is exposed to incoherent light (there is no phase correlation between waves). This
energy is absorbed by the system producing coherent light in which the phases of the light at
different locations are synchronized.  Thus, a high coherence is established over long
distances beginning with a more random supply of energy. Conservation of energy is not
violated by such a system because of the excess pumping energy, but this example does
illustrate the development of an ordered state (coherence) among many particles (the excited
states of the atoms) which is not apparent from study of a single particle.  In superfluid helium
cooled near absolute zero, the entropy approaches zero but the atoms exhibit apparent
disorder similar to higher temperature liquids. However, the macroscopic wave function
describes a correlation of the motion of the atoms. This subtle order reduces the entropy to
near zero in spite of the lack of spatial order one would expect.  The point here is that in
many-body systems,  other types or order and coherence are possible besides the obvious
spatial type found in crystalline structures.
   Ignoring for a moment the microscopic view which represents the basis for modern
biology, Frohlich looks at biological systems as a whole and seeks to describe the important
physical aspects. He finds that (a) they are relatively stable but far from equilibrium, (b) they
exhibit a nontrivial order, and (c) they have extraordinary dielectric properties. The third point
refers in part to cell membranes which support an electric field of 107 V/m. Because of the
lack of spatial order (at the level of ions), Froehlich hypothesizes that other types of order or
coherence play an important role in the description and properties of biological systems.  In
reference to the first point, he assumes that living systems consist of various excited states
that are stable or metastable.  It is these correlations and excited states through which weak
electromagnetic fields are able to interact even when sufficient energy is not available to affect
individual particles directly in a significant way.
   Much of the mathematical detail of this framework depends on nonlinear dynamics and
deterministic chaos. This branch of mathematics is relatively new and predicts highly complex,
but structured behavior from systems involving nonlinear forces. An introduction to this field
can be found in Davies (1988). Chaos is predicted even in simple systems meeting the
appropriate conditions, such as a two dimensional pendulum with a nonlinear driving force. It
can be shown mathematically that multiple, or even infinite states of motion can exist with
equal probability. Although the equations  of motion are deterministic in a strict sense, errors in
initial conditions are multiplied with each cycle so that exact motion is unpredictable in
practice.  Such systems display highly erratic and sometimes nonrepetitive behavior, but
       10/18/90
2-30

-------
                            DRAFT-DO NOT QUOTE OR CITE


exhibit certain predictable and common characteristics. Under certain conditions, partially
stable states can be achieved which are highly sensitive to external forces especially if these
forces occur at specific frequencies. Frohlich suggests that the high sensitivity and
functionality in living systems may result from ordered states within the apparently chaotic
motions and arrangements of biological particles. As in the case of a laser where coherent
light is produced from incoherent light, the higher order modes in living systems are driven by,
but decoupled from the surrounding random thermal motion. An example is self-excited
oscillations or limit cycles which are characteristic of certain nonlinear systems under specific
conditions, and may help explain threshold and saturation effects in biological experiments.
   A feature of this viewpoint is that ordered or coherent states can exist over large distances,
thus offering a mechanism by which cells may communicate in spite of the short range of
chemical forces. This long-range biological coherence may provide growth  control as exists in
healthy tissue but is absent in cancer. A number of investigators have expanded Frohlich's
approach, and sought to test predicted consequences experimentally (Frohlich, 1988; Adey
and Lawrence, 1984).  Rowlands (1988) has studied rouleaux formation or stacking of red
blood cells during clotting. From a strictly electrostatic viewpoint, one would expect repulsion
between the cells due to the negative electric charges on each cell. Instead, an attractive
force extending over microns of distance is observed which represents an order of magnitude
greater range than is associated with chemical forces.
   Conceptually, Frohlich's approach does not define or restrict the properties or physical
quantities which are coherently coupled.  However, attempts have been made to identify
specific cases. One of the earliest observations in this respect is the unique electrical
conditions existing in cell membranes.  Given the electric field (107 V/m), the membrane
thickness, and an estimated elastic constant, Frohlich calculates that the membrane should
oscillate in electric fields in the frequency range  of 10 GHz to 100 GHz.  Fields produced by
each cell might be coherently coupled. Other aspects of the theory indicate that interactions
with such excited coherent modes will be frequency and amplitude sensitive. Grundler et al.
(1988) have observed such effects in the growth of yeast cells near 42 GHz.  Enhancement
and reduction of growth rates were found at closely spaced frequency intervals for nonthermal
exposures.  Other experimental results supporting this theory can be found in Frohlich (1988).
   Lawrence and Adey (1982) have proposed a model by which weak signals could  be
transmitted through cell membranes. This model suggests that solitons or solitary waves may
carry weak signals  inside cells. Solitons are nonlinear waves that do not dissipate like
ordinary waves. In this case, solitons are thought to travel along intramembranous protein
                                         2-31
10/18/90

-------
                            DRAFT-DO NOT QUOTE OR CITE


particles (IMPs) that pass through the cell membranes. This process is summarized in three
steps: First, weak electrochemical events are sensed by glycoproteins on the cell surface.
The surface events are then amplified and transmitted to the cell interior along an IMP. Finally,
these signals are coupled to internal enzyme systems in the cell as well as the nucleus and
other organelles. It is suggested that this process may be affected by externally applied fields.
A detailed explanation of this theory can be found in Adey (1990).
   The concepts of biological coherence offer a new approach for studying the properties of
highly complex living systems, and may help explain the high sensitivity of certain biological
systems. There is a tendency to reject these ideas because they are difficult mathematically,
and because they deal with nonlinear properties which have often  been avoided in an effort to
reduce complex problems to simple terms. Further, the problems of coherence and
many-body interactions are seldom discussed in most disciplines and often cannot be solved
by direct means. The problems are theoretically solvable,  but very complex. Deterministic
chaos is also unfamiliar in most disciplines, but has been clearly demonstrated in many
physical processes including turbulent flow, nonlinear circuits, cavitation, and even in
biological  processes such as EEG and  stimulated heart beat (Kaiser, 1988). The role that
these principles play in other biological processes has yet to be established and will require
further investigation. Perturbation of metastable excited states  may be the mechanism through
which the  low energy of nonthermal EM exposures are able to interact with biological systems.
       10/18/90
2-32

-------
                           DRAFT-DO NOT QUOTE OR CITE
REFERENCES FOR CHAPTER 2

Adey, W.R. (1990) Electromagnetic fields, cell membrane amplification, and cancer promotion,
       in extremely low frequency electromagnetic fields. In: Wilson, B.W.; Stevens, R.G.;
       Anderson, LE. The question of cancer. Columbus, OH: Batelle Press, pp. 211-249.

Adey, W.R.; Lawrence, A.F. (1984)  Nonlinear dynamics in biological systems. New York, NY:
       Plenum Press.

Adey, W.R.; Lawrence, A.F., eds. (1984)  Nonlinear electrodynamics in biological systems.
       New York, NY: Plenum Press.

American National Standards Institute  (ANSI).  (1982) American National Standards Institute
       safety levels with respect to human exposure to radio frequency electromagnetic fields,
       300 kHz to 100 GHz. Institute of Electrical and Electronics Engineers, Inc., New York,
       NY: ANSI C95.1-1982.

Barber, P.W.  (1977)  Resonance electromagnetic absorption by nonspherical dielectric
       objects. IEEE Trans. Microwave Theory and Techniques. MTT-25(5): 373-381.

Barnes, F.S. (1986) Extremely low frequency (ELF) and very low frequency electric fields:
       rectrification, frequency sensitivity, noise, and related phenomena.  In: Polk, C.;
       Postow, E., eds. CRC handbook of biological effects of electromagnetic fields. Boca
       Raton, FL: CRC Press.

Barnes, F.S.;  Hu, C.L (1977)  Model for some non-thermal effects of radio and microwave
       fields on biological membranes. IEEE Trans. Microwave Theory Technol. 25:742.

Barnes, F.S.;  Seyed-Madani, M.  (1987) Some possible limits on the minimum electrical
       signals of biological significance. In:  Blank, M.; Findl, E., eds. Mechanistic
       approaches to interactions of electric and electromagnetic fields with living systems.
       New York, NY: Plenum Press.

Bawin, S.M.; Gavalas-Medici, R.J.; Adey, W.R. (1973) Effect of modulated very high frequency
       fields on specific brain rhythms in cats.  Brain Res. 58:365.

Blackman, C.F.; Elder, J.A.; Weil, C.M.; Benane, S.G.; Eichinger, D.C.; House, D.E.  (1979)
       Induction of calcium ion efflux from brain tissue by radiofrequency radiation: effects of
       modulation frequency and field strength. Radio Science 14(6S):93-98.

Blackman, C.F.; Benane, S.G.; House, D.E.; Jones, W.T. (1985) Effects of ELF (1-120 Hz) and
       modulated (50 Hz) fields on the efflux of calcium ions from brain tissue in vitro.
       Bioelectromagnetics 6(1):1-11.

Blackman, C.F.; Benane, S.G.; Elliot, D.J.; House, D.E.; Pollock, M.M.  (1988) Influence of
       electromagnetic fields on the efflux of calcium ions from brain tissue in vitro: a
       three-model analysis consistent with the frequency response up to 510 Hz.
       Bioelectromagnetics 9(3):215-227.
                                         2-33
10/18/90

-------
                            DRAFT-DO NOT QUOTE OR CITE
Blank, M.; Blank, J.N.  (1986)  Concentration changes at ion channels due to oscillating
       electric fields. J. Electro. Chem. Soc. 133:237-238.

Blank, M.; Findl, E. (1987) Mechanistic approaches to interactions of electric and
       electromagnetic fields with living systems.  New York, NY: Plenum Press.

Blankenship, R.E.; Schaafsma, T. J.; Parson, W.W.  (1977) Magnetic field effects on radical pair
       intermediates in bacterial photosynthesis. Biochim. Biophys. Acta 461:297-305.

Chen, J.-Y.; Gandhi, O.P.  (1989)  RF currents induced in an anatomically-based model of a
       human for plane-wave  exposures (20-100 MHz).  Health Phys. 57:89-98.

Chiabrera, A.; Nicolini, C.; Schwan,  H.P., eds. (1985) Interactions between electromagnetic
       fields and cells. New York, NY: Plenum Press.

Davies, P. (1988) The cosmic blueprint.  New York, NY: Simon and Schuster, Inc.

Durney, C.H.; Rushforth, C.K.; Anderson, A.A. (1988) Resonant AC-DC magnetic fields:
       calculated response. Bioelectromagnetics 9(4) :315-336.

Easterly, C.E. (1982) Cardiovascular risk from exposure to static magnetic fields. Am. Ind.
       Hyg. Assoc. J. (43)7:533-539.

Findl, E. (1987) Membrane transduction of low energy level fields and the Ca+ +  hypothesis.
       In:  Blank, M.; Findl, E., eds.  Mechanistic approaches to interactions of electric and
       electromagnetic fields with living systems.  New York, NY: Plenum Press.

Florig, H.K.; Nair, I.; Margan, M.G. (1987, March)  Briefing paper 1: sources and dosimetry of
       power-frequency fields. Prepared for Florida Department of Environmental Resources
       by Carnegie Mellon University.

Frohlich, H.  (1988) Biological coherence and response to external stimuli. New York, NY:
       Springer-Verlag.

Gandhi, O.P.  (1982)  Electromagnetic absorption in inhomogeneous model of man for realistic
       exposure conditions. Bioelectromagnetics 3(1) :81 -90.

Grundler, W.; Jentzsch, U.; Keilmann, F.; Putterlik, V. (1988) Resonant cellular effects of low
       intensity microwaves. In:  Frohlich, H., ed.  Biological coherence and  response to
       external stimuli.  New York, NY: Springer-Verlag.

Guy, A.W. (1987) Dosimetry associated with exposure to non-ionizing radiation: very low
      frequency to microwaves.  Health Phys. 53:569-584.

Halle, B. (1988) On the cyclotron resonance mechanism for magnetic field effects on
      transmembrane ion conductivity. Bioelectromagnetics 9(4):381-385.

International Radiation Protection Association (IRPA). (1990) Interim guidelines on limits of
      exposure to 50/60  Hz electric and magnetic fields.  Health Phys. 58:113-122.
       10/18/90
2-34

-------
                            DRAFT-DO NOT QUOTE OR CITE
Kaiser, F.  (1988)  Theory of non-linear excitations.  In: Frohlich, H., ed.  Biological coherence
       and response to external stimuli. New York, NY: Springer-Verlag.

Kaune, W.T.; Phillips, R.D. (1980) Comparison of the coupling of grounded humans, swine and
       rats to vertical, 60-Hz electric fields.  Bioelectromagnetics 1:17-129.

Kaune, W.T.; Gillis, M.F. (1981) General properties of the interaction between animals and ELF
       electric fields.  Bioelectromagnetics 2:1-11.

Kaune, W.T.; Miller, M.C. (1984)  Short-circuit currents, surface electric fields, and axial current
       densities for gunea pigs exposed to ELF electric fields. Bioelectromagnetics.
       5:361 -364.

Kaune, W.T.; Stevens, R.G.; Callahan, N.J.; Sorenson, R.K.; Thomas, D.B. (1987) Residential
       magnetic and electric fields. Bioelectromagnetics  8:315-335.

Kaune W.T.; Stevens, R.G.; Callahan, N.J.; Severson, R.K.; Thomas, D.B. (1990)  Residential
       magnetic and electric fields. Bioelectromagnetics (in press).

Lawrence, A.F.; Adey, W.R. (1982) Non-linear wave mechanisms in interaction between
       excitable tissue and electromagnetic fields.  Neurol. Res. 4:115-154.

Lednev, V.V. (1990)  Interaction of modulated electromagnetic radiation with biological
       systems: theoretical predictions and experimental data.  Presented at the Twelfth
       Annual Meeting of the Bioelectromagnetics Society; June 10-14; San Antonio, TX.

Liboff, A.R.; McLeod, B.R. (1988) Kinetics of channelized  membrane ions in magnetic fields.
       Bioelectromagnetics 9(1) :39-51.

Liboff, A.R.; Smith, S.D.;  McLeod, B.R. (1987)  Experimental evidence for ion cyclotron
       resonance mediation of membrane transport.  In: Blank, M.; Findl, E., eds.
       Mechanistic approaches to interactions of electric and electromagnetic fields with living
       systems. New York, NY: Plenum Press.

Liboff, A.R. (1985)  Cyclotron resonance in membrane transport.  In: Chiabrera, A.;  Nicolini, C.;
       Schwan, H.P. eds. Interactions between electromagnetic fields and cells.  New York,
       NY: Plenum Press.

Liboff, R.; Thomas, J.R.; Schrot, J. (1989) Intensity threshold for 60-Hz magnetically induced
       behavioral  changes in rats. Bioelectromagnetics 10:111 -113.

Montaigne, K.; Pickard, W.F. (1984) Offset of the vacuolar potential of characean cells in
       response to electromagnetic radiation over the range of 250 Hz to 250 kHz,
       Bioelectromagnetics 5:31-38.

National Council on Radiation  Protection and Measurements (NCRP).  (1981)  Radiofrequency
       electromagnetic fields.  Report No. 67. Bethesda, MD: National Council on Radiation
       Protection.
                                         2-35
10/18/90

-------
                            DRAFT-DO NOT QUOTE OR CITE
New York Power Lines Project. (1987)  Biological effects of power line fields. Scientific
       Advisory Panel final report. A. Ahlbom, E. N. Albert, A. C. Fraser-Smith, A. J.
       Grodzinsky, M. J. Marron, A. O.  Martin, M. A. Persinger, M. L. Shelamski, and E .R.
       Wolpow.  Albany, NY: Wadsworth Center for Laboratories and Research.

Office of Technology Assessment (OTA). (1989)  Biological effects of power frequency electric
       and magnetic fields, background paper.  OTA-BP-E-53.  U.S. Congress, Office of
       Technology Assessment. Washington, DC: U.S. Government Printing Office.

Persson, B.R.R.;  Stahlberg, F. (1989) Health and safety of clinical NMR examinations.  Boca
       Raton, FL: CRC Press.

Polk, C.  (1986a)  Introduction. In: Polk, C.;  Postow, E., eds. Handbook of biological effects
       of electromagnetic fields.  Boca Raton, FL: CRC Press.

Polk, C. (1986b)  Physical mechanisms by which low-frequency magnetic fields can affect the
       distribution of counterions on cylindrical biological cell surfaces. J. Biol. Phys. 14:3-8.

Polk C. (1989)  Nuclear precessional magnetic resonance as a cause for biological effects of
       time-varying electric or magnetic fields in the presence of an earth strength static
       magnetic field.  Bioelectromagnetics Society, Eleventh Annual Meeting Abstracts.

Polk C.; Postow E. (1986)  Handbook of biological effects of electromagnetic fields.  Boca
       Raton, FL: CRC Press.

Rowlands, S. (1988) The interactions of living red blood cells.  In:  Frolich, H., ed. Biological
       coherence and response to external stimuli. New York,  NY: Springer-Verlag.

Schwan, H.P.; Foster, K.R. (1980) Rf-field interactions with biological systems: electrical
       properties and biophysical mechanisms. Proc. IEEE 68(1 ):104-113.

Smith, S.D.; McLeod, B.R.; Liboff, A.R.;  Cooksey, K.  (1987)  Calcium cyclotron resonance  and
       diatom mobility. Bioelectromagnetics 8(3):215-227.

Tenforde, T.S. (1990) Biological interactions and human health effects of extremely low
       frequency magnetic fields.  In: Wilson, B.W.; Stevens, R.G.; Anderson, L.E., eds.
       Extremely low frequency electromagnetic fields: the  question of cancer.  Columbus,
       OH: Battelle Press, pp. 291-315.

Tenforde, T.S.; Kaune, W.T. (1987) Interaction of extremely low frequency electric and
       magnetic fields with humans. Health Phys. 53:585-606.

Thomas, J.R.; Schrot, J.; Liboff, A.R.  (1986)  Low-intensity magnetic fields alter operant
       behavior in rats. Bioelectromagnetics 7(4):349-357.

U.S. Environmental Protection Agency  (U.S.  EPA). (1986, July)  The radiofrequency radiation
       environment: environmental exposure levels and RF  radiation emitting sources. Office
       of Air and Radiation. Washington, DC. EPA 520/1 -85-014.
       10/18/90
2-36

-------
                           DRAFT-DO NOT QUOTE OR CITE
Weaver, J.C.; Astumian, R.D. (1990) The response of living cells to very weak electric fields:
       the thermal noise limit. Science  247:459-462.

Wilson, B.W.; Stevens, R.G.; Anderson, LE. (1990)  Extremely low frequency electromagnetic
       fields: the question of cancer.  Columbus, OH: Batelle Press.

World Health Organization (WHO). (1987)   Magnetic fields.  Environmental Health Criteria 69
       Geneva, Switzerland: WHO.

World Health Organization (WHO). (1984) Extremely low frequency fields. Environmental
       Health Criteria 35. Geneva, Switzerland: WHO.
                                       2-37

-------

-------
                            DRAFT-DO NOT QUOTE OR CITE

      3.  EPIDEMIOLOGIC STUDIES OF ELECTROMAGNETIC FIELDS AND CANCER
 3.1. INTRODUCTION
    The epidemiologic literature for radiofrequency (RF) radiation was previously reviewed by
 the U.S. EPA (1984) and by Elder (1987). But, in the past few years, there has been a marked
 increase in epidemiologic studies reporting an association between cancer and electric and
 magnetic fields.  Concern extends not just to RF radiation but also to exposure from electrical
 power transmission and usage. The studies generally fall into one of three categories:
 occupational, environmental, and residential exposures. Results have been reported for
 different geographic regions, countries, age groups, industries, and occupational
 classifications. Many of the studies focus on 50- or 60-hertz (Hz) fields, the frequencies used
 for power transmission. These studies have been described in various review articles
 (Ahlbom, 1988; Brown and Chattopadhyay, 1988; Coleman and  Beral, 1988; Easterly, 1981;
 Modan, 1988; Savitz and Calle, 1987; Sheikh, 1986).
   There have been studies on cancer in children and cancer in adults.  Childhood cancer
 studies have involved exposure to magnetic fields in the home, either measured or estimated
 by power line wiring configurations (Table 3-1).  The results of these studies bear directly on
 the risk of cancer in exposed persons. Studies of children and adults  are evaluated
 separately.  The presumed frequency of exposure is relevant. Studies that examined
 populations with  exposure to RF radiation are evaluated separately from studies of 50 or 60
 Hz, the frequencies of electrical power transmission. Many of the occupational studies use job
 titles or employment in industries or occupations with potential exposure to electromagnetic
 (EM) radiation, a  surrogate of exposure.  These studies may involve  exposure to different EM
frequencies, either singly or in combination, and so the most relevant frequency of exposure
 may be unclear.  Since it is likely that electric and magnetic fields from 50- or 60-Hz
frequencies or extremely low frequency (ELF) electromagnetic radiation predominate, the
occupational studies with unknown or mixed frequency exposure are described together with
50- or 60-Hz occupational studies.  It should be noted that higher frequency fields, specifically
 RF radiation, may be modulated at lower frequencies. Demarcation by frequency is somewhat
arbitrary but is important to any given study with respect to determining if a specific health
effect could explicitly be identified with a particular frequency of exposure or a given job title or
class of jobs.
3-1
                                                                             10/18/90

-------
                DRAFT-DO NOT QUOTE OR CITE
S-g
Is
2 .2

Ťť"$
O fC

CO CO

m g
   ,
 "D N
                                         
^ 2 Ť Ł
*n 3 *—
ffl ^ "0 Ł
•g ° jŁ E oi
13 •gŁ-°'H
Ł§Ł•5 8
c x'^ 'S Ł
lŤ||§
mil
                                                    I!*'
                               33S
                                                   c o>;
                                                   "§Ł:
                                           iSFJXsic  S w "O i-
                                           Ł.!>= o>Ť S-2-g-Ss
                                           ^"Ť^?TE^sgi
                                             Łg-
                         IIIllllllllf
                         •**> - o>
                                         *!
                                            !c
                                            e
    co f
    •- •§ ;
    S a>
     8
                                         DJ S S JJ S
                                               Ilia
w ^. -- —
03 O ^ *~
 tNiSS
  II O Ť


 of I"
  i_ C *- *D
   8O 3 5
   .c .a o
      i.5
   T3
  c c

 *•§!
 ,_ to a>
 
                                         ^j .t; s w m
                                         Q) O) m CO O
                                         =li&i
                                         lllll .
                                         ^iisli

                                     '     ^ SP S S a>
                                     ;~.-g   i^lll
                       iŁr
                  :l
                • *Ji (O •*—
                i^iiin
                       g


                       8
                       0)
                                     -"0
           o>

           ^
           

      II
        i
a>

o
                   8
                   S
                   in
                   oo
                   o>
                                   S.
ra

a>



I
 10/18/90
           3-2

-------
                       DRAFT-DO NOT QUOTE OR CITE
1
C

1

1

DC
UJ
O
Q

O

O


D
O

CO
UJ

Q
u.
O

tr
<
CO




CO


!iJ
00
~
Q.


1

CD
I



I
Q.

Ł

f

8)
C
at
"
     •So 2
     'Si _ ť
     SgS
     CO   C
     CD A g


     I E Ť-
           igŁ.-sl =
     x in .=  .

     Ž cl S
     (- ^ O O

     °>i II
     X S §.Ť
                1 g"^ 1 aJf
                Ť 2 * O 3 CO

                ls!||i.

              isl^lfi

              Mf Sl-Sl
              Ť•§ Ł•§.Ł Ł88
              8JťŤť-3 = S
              •S.-2
              •° Ť
              si
                    s||a
              3 'Ť
                  CD
      TJ
      C
     2 Ł
              Ł
              S
              v

              •5
    J2
     Ť

     S"-
      ^

     Ť? § 2

     1.8 g
              ^  S 2i>
              O) ^" CO >
              s 8 Ť> ••§

              ll^S.
              §§ŁS

              ~ e K? *"•
              •SS^E

              ?8?|
              S § "S
              •— Ta Q TJ

              l^.s

              1-iS?
              .g> a T- .1

              c § " ~
              o opc.0-
              2 T3 O CM
m uj   C

^o  Ł&E
O -C  .-g — br

m fl)  *
**— P  c

•S *  Ł
•o ™  2 eo |



"^ CO  O O - ŤB "o

•5 o  ťZ C CO ^ CO
CD .—  ^^ 3 C CO c
CO CO  T- S E CO E
                    |fl
                    Ť
    T3 .V—   Ť>

    co •*-• Q-

    g S S   Ť2
               J|fj|



                 c*|f
          : Q.
          ;Ť
              = 1
              •§ 01
                Ť 2 &;
                T3 "D O
           I
           §
           If
           si
                il?
                2 E S

                611
                                       O CD C8 =
                                       •t n Q. 2
                                      c
                                      o
O Ť c
f 58
T3 C C 3
o s i 8

fill!

v|lli
S°^il

nisi
o. -D S: 8 -S
                          IS
                                 •o ro
                                     1 a
                                  3-3
                                                                     10/18/90

-------
                           DRAFT-DO NOT QUOTE OR CITE

3.2.  STUDIES OF CHILDREN
3.2.1. 50- or 60-Hertz Exposures
   Wertheimer and Leeper (1979) reported the results of a case-control study that examined
the cancer mortality of children in relation to electrical wiring configurations. Cases were
children who died of cancer in Colorado before age 19 between 1950 and 1973. The
source(s) used to identify cases or the source(s) of cancer mortality data were not stated.
Cases were also required to have been in Colorado and have lived at Greater Denver area
addresses between 1946 and 1973. Two sets of controls were developed. The first, called
"file 1 controls," were next Denver-area birth certificates drawn from files organized by birth
month and county.  The second set, called "file 2 controls," were next Denver-area birth
certificates taken from alphabetical listings grouped by the years 1939-1958,1959-1969, and
1970-1974.  Birth addresses were taken from birth certificates.  "Death" addresses were
defined as the parents' address 2 years prior to diagnosis of the case and were obtained by
searching city directories. For cases that could not be traced in this manner, the address
listed on the death certificate was taken as the death address. If a "file 1 control" could not be
traced, a "file 2 control" with a birth date most similar to the case was selected.  The authors
stated that the proportion of controls with an address coded as of "high-current configuration
(HCC)"  (see below) was similar between used file 1 controls (21%), used file 2 controls (23%),
and the pool of controls available but not selected (25%). The author indicates that selection
bias is, thus, probably not an issue. The methods for this assessment were not, however,
stated.
    There were 344 cancer cases who met the study selection criteria. Of these, birth address
information was lacking for 72 cases or the birth occurred prior to 1946, and death address
information was lacking for 16 cases. As a result, analyses of birth address were made for 272
cases and their controls and of death address were made for 328 cases and their controls.
    Maps were made of the electrical wires and transformers in the vicinity of birth and  death
addresses for cases and controls. Primary [13-kiiovolt (kV)] wires were classified as
"large-gauge" or "thin." Large-gauge wires are designed to carry high currents. Homes were
classified as having either a "high-current configuration (HCC)" or "low-current configuration
 (LCC)."  HCC homes were (1) less than 40 meters from large-gauge primaries or an array of six
or more thin primaries, (2) less than 20 meters from arrays of 3-5 thin primaries or high tension
        10/18/90
3-4

-------
                           DRAFT-DO NOT QUOTE OR CITE

(50-230 kV) wires, (3) and less than 15 meters from first-span secondary wires (240 volts)
defined as secondaries issuing directly from the transformer without any loss in current
through a service drop beyond the pole.  LCC homes were all other configurations.  If
first-span wires carried current to only one or two single family homes, the homes so served
were also considered LCC. The lowest potential exposure was considered to be homes
beyond the pole at the end of a secondary line (called "endpoles"). They had no distribution
lines running past them.
   A significantly greater proportion of cancer cases had lived in homes classed as HCG.
This was  more pronounced for cases and controls who had lived in one residence from birth
to death.  The distribution of cases and controls in the dichotomous wire code classification of
HCC versus LCC was examined for three types of residences of the cases and controls, i.e.,
persons who had a single residence that was the same at birth and death, persons who had
moved and the code for their birth residence was evaluated, and persons who had moved and
the code  for their "death" residence was evaluated.  For subjects with the same death address
as birth address, i.e., "stable residence," 44% of the cases had lived in a home classified as
HCC vs. 20.3% of corresponding controls. For subjects who had moved, 32.5% of the cases
had lived in a home at the time of their birth which could be classified as HCC vs. 20.1% of
controls,  and 37% of the cases had lived in a home just prior to their death that could be
classified as HCC vs. 24.0% of controls. These differences were statistically significant (Table
3-2), with the highest significance in the "stable residence" group.
   Odds ratios were not given but were calculated  from the available data (Table 3-2).  For
subjects with a "stable" address, the odds ratio was 3.09, indicating an excess risk of cancer in
children living in HCC homes throughout their lifetimes. The relationship persists, albeit less
strongly,  when the children had moved during their lifetimes.
   The distribution of cases and controls by HCG and LCC residence at birth and at death
was presented for leukemias, lymphomas, nervous  system tumors, and  all other cancer
combined (Table 3-3).  For all cancer sites combined and for both types of addresses, the
proportion of cases who resided  in HCC homes was greater than the proportion of controls.
There was no difference in the risk of all other cancer between cases  and controls when
consideration was given to death address alone. No statistical tests were performed nor odds
ratios developed by the authors.  Again, these have been calculated from the reported data
(Table 3-3) by the reviewers. The excess proportion of leukemia cases who resided in
                                         3-5
10/18/90

-------
                                DRAFT-DO NOT QUOTE OR CITE
  TABLE 3-2. DISTRIBUTION OF CASES AND CONTROLS ACCORDING TO VARIOUS RESIDENCES CODED AS HIGH CURRENT
              CONFIGURATION (HCC) OR LOW CURRENT CONFIGURATION (LCC) FOR TOTAL CANCERS8
    A. Stable/one residence

                  Case         Controls
        HCC
        LCC
        Total
        %HCC
        x2
        P
        OR
        95% Cl
 48
 61
109
 44
 14.4
< 0.001
  3.09
  1.68-5.71
  26
 102
 128
  20.3
 Total

  74
 163
 237
    B1. Moved/birth residence only
        HCC
        LCC
        Total
        %HCC
        x2
        P
        OR
        95% Cl
   Case

 53
110
163
 32.5
  5.4
< 0.02
  1.91
  1.09-3.34
Controls

  29
 115
 144
  20.1
Total

 82
 225
 307
    C1. Moved/death residence only
B2.a All birth addresses

             Case      Controls

HCC       101            55
LCC       171           217
Total       272           272
%HCC       37.1
x2          18.2
p          < 0.0001
OR           2.33
95% Cl        1.56-3.49

C2.a All death addresses
Total

 156
 388
 544
                    Case
            Controls
                                         Total
                                                                  Case
                                                           Controls
                                                         Total
HCC
LCC
Total
%HCC
x2
P
OR
95% Cl
81
138
219
37.0
7.6
<0.01
1.86
1.19-2.91
48
152
200
24.0


129
290
419



HCC
LCC
Total
%HCC
x2
P
OR
95% Cl
129
199
328
39.3
18.2
< 0.0001
2.23
1.56-3.18
74
254
328
22.6


203
453
653



*B2 and C2 calculated for this report. Odds ratios (OR) calculated for this report. Cl = confidence interval.

SOURCE: Wertheimer and Leeper. 1979.
       10/18/90
                              3-6

-------
                                 DRAFT-DO NOT QUOTE OR CITE
 TABLE 3-3. DISTRIBUTION OF CASES AND CONTROLS ACCORDING TO VARIOUS RESIDENCES CODED AS HCC OR LCC FOR
                                         SPECIFIC CANCER SITES"
                    A. All Birth Addresses
                                                                    B. All Death Addresses
        Leukemias
HCC
LCC
Total
%HCC
x2
OR
95% Cl
Cases
 52
 84
136
 38.2
  8.5
  2.28
  1.29-4.05
Controls
   29
  107
  136
   21.3
 p=0.004
                                           Total
                                            81
                                           191
                                           272
HCC
LCC
Total
%HCC
x2
OR
95% Cl
                                                          Cases
                                                            63
                                                            92
                                                           155
                                                            40.6
                                                            16.8
                                                             2.98
                                                             1.72-5.15
                        Controls    Total
                                                                          29
                                                                         126
                                                                         155
                                                                          18.7
                                                                        p< 0.0001
                                                             92
                                                            218
                                                            310
        Lymphomas
HCC
LCC
Total
%HCC
x2
OR
95% Cl
Cases
10
21
31
32.3
1.4
2.48
0.64-10.00
Controls
5
26
31
16.1
p = 0.24


                                           Total
                                            15
                                            47
                                            62
HCC
LCC
Total
%HCC
x2
OR
95% Cl
                                    Cases
                                      18
                                      26
                                      44
                                      40.9
                                       1.9
                                       2.08
                                       0.76-5.71
                        Controls    Total
                                                                          11
                                                                          33
                                                                          44
                                                                          25.0
                                                                        p = 0.17
                                                                                   29
                                                                                   59
                                                                                   88
        Nervous System Tumors
HCC
LCC
Total
%HCC
x2
OR
95% Cl
          Cases
           22
           35
           57
           38.6
            3.4
            2.36
            0.95-5.89
Controls
   12
   45
   57
   21.1
 p = 0.24
                                           Total
                                            34
                                            80
                                           114
HC
LCC
Total
%HCC
x2
OR
95% Cl
                                                          Cases
                                                            30
                                                            36
                                                            66
                                                            45.5
                                                            4.8
                                                            2.40
                                                            1.08-5.36
                        Controls    Total
                                                                17
                                                                49
                                                                66
                                                                25.8
                                                              p = 0.03
                                                             47
                                                             85
                                                            132
        All other cancers

HCC
LCC
Total
%HCC
x2
OR
95% Cl
Cases
17
31
48
35.4
2.6
2.38
0.85-6.74
Controls
9
39
48
18.7
p = 0.11


Total
26
70
96





HCC
LCC
Total
%HCC
x2
OR
95% Cl
Cases
18
45
63
28.6
=0
1.08
0.46-2.54
Controls
17
46
63
27.0
p = 1.0


Total
35
91
126




aChi-squared, two tail p-value, and odds ratios (OR) calculated for this report. Cl

SOURCE: Wertheimer and Leeper, 1979.
          confidence interval.
                                                 3-7
                                     10/18/90

-------
                            DRAFT-DO NOT QUOTE OR CITE

HCC-coded homes was statistically significant at both "birth" residences as well as "death"
residences. The odds ratios were 2.28 and 2.98, respectively. The risk of cancer of the
nervous system was significantly elevated in children who had lived in HCC-coded homes
whether "birth" address or "death" address is used. Bearing in mind that the "death" address
represents the parents' address 2 years prior to case diagnosis, the authors argue that the
stronger positive association for death address may represent a late or promotional effect.
This is speculative, at best, since the strongest association for overall cancer is in children who
had lived at only one ("stable") address. The authors suggested that the observed broad
associations with different cancer types may indicate that an HCC-relationship may be indirect
or reflect some exposure effect on the physiologic processes of the child. The lack of cancer
site specificity is, however, not unusual. Many agents have been shown to produce cancers at
multiple sites.
    Certain demographic factors that could relate to the development of cancer were examined
to consider whether they could explain the observed association of excess cancer mortality in
children who had resided in HCC homes. These factors were urban-suburban differences,
socioeconomic class, family patterns of cancer, traffic congestion, and sex. There was  a slight
but nonsignificant excess of suburban addresses among controls.  A trend, albeit not
significant, toward higher socioeconomic class was seen in the cases.  The authors stated that
their tracing methods might have biased the selection of controls of lower socioeconomic
status. But an analysis of three class groupings (based on occupation) revealed no significant
difference in the percent of discarded and retained controls with HCC-coded  homes among
the three socioeconomic classes. There was also a trend toward firstborns and older mothers
among cases, but, again, the results for this factor were not statistically significant. The
authors stated that the association also held under analysis within maternal age and sibling
order categories, but the data were not provided.  Cases were found to generally live closer to
high traffic routes, but the HCC association held in an analysis of proximity to high traffic and
other routes. The excess of HCC residences was significant for both males and females, but
the relationship was observed to be stronger for males.
   The authors noted what they considered to be a striking finding, that is, six cases lived
near (less than 150 meters from) a substation, and all six cases were within 40 meters of large
primary wires.  However, none of the controls lived near a substation. This latter finding is
       10/18/90
3-8

-------
                           DRAFT-DO NOT QUOTE OR CITE

consistent with expectation since it was estimated that less than one home in 1000 in Denver
is near a substation.
   One drawback discussed by the authors involved mapping and coding. It was not done
blindly, that is, without knowledge of whether a case or control lived at the residence.  Blinding
is necessary to ensure that ascertainment of exposure (or outcome) for both cases and
controls is comparable and without bias, but it is often not possible for various practical
reasons. For this study, it would seem that identification and classification of physical
structures outdoors would be less subject to misclassification or manipulation by the coder
than would information from sources such as records or interviews that could  be more
amenable to interpretation or influence. As a check on the possible introduction of bias into
the procedures, a random sample of 140 addresses (70 cases and 70 controls) was receded
by an assistant unaware of the case or control status of the addresses. Agreement as to wire
configuration coding was found to be 91%. For the addresses (N=12) where  coding differed,
the split between supporting or not supporting an HOC association with excess  cancer was
about even.  A small blinded study was conducted for birth addresses in Colorado Springs
and Boulder. The same relationship with wire code was  observed, but the results were not
statistically significant which, according to the authors, was likely due to small numbers.
   A second problem with this study is a lack of an adequate discussion of latent effects with
respect to the various cancers. It is possible that certain site-specific cancers such as
lymphoma and/or central nervous system cancer may present an onset period longer than that
of leukemia and the net effect of lumping them all together as one group would be to mask
individual differences in length of latency. The fact that children who have always lived at the
same location have a higher risk of total cancer than  those who have moved would seem to
support this thesis. Differences in risk seen in "birth"  residency cases separately from "death"
residency cases may only be a reflection of the expression of different cancers, that is,
lymphoma in "birth" residency cases and leukemia in "death" residency cases. In any event,
this topic could have been more adequately discussed by the authors.
   This study did not provide measures or estimates of  magnetic field exposure, but a
seemingly reasonable surrogate was employed, i.e., classification by proximity to current
sources with differing potential for presenting high vs. low magnetic field exposure, namely,
wiring code  configurations.  Later studies have demonstrated a relationship between wiring
                                         3-9
10/18/90

-------
                           DRAFT-DO NOT QUOTE OR CITE

code and measured field values (Savitz et al., 1988; Kaune, 1987). Classification into two
groups (HCC and LCC) does not, however, readily permit examination of dose response.
   The authors stated that the observed increased risk is rarely greater than a factor of 2 or 3.
The authors noted that the magnitude of the risk, if true, may be less important than the
prevalence of the HCC exposure which is widespread; that is, a true but small risk could yield
an important increase in childhood cancer. This is a reasonably well-done study that supports
the possibility that exposure to EM fields (as estimated by wire code) is associated with cancer.
   Fulton et al. (1980) conducted a case-control study of childhood leukemia in Rhode Island
which attempted to replicate the study on electrical wire configurations done by Wertheimer
and Leeper (1979). Cases were identified from records of the Rhode Island Hospital and
consisted of 155 patients whose onset of leukemia occurred between 0 and 20 years of age
and between 1964 and 1978. The actual geographic area covered (patient base) for this
hospital is not given. It was also  not explicitly stated whether all cases were Rhode Island
births or not.  Thirty-six cases were excluded because they had lived out-of state for some time
during 8 years prior to  disease onset; 119 cases remained for study. Two controls of the same
birth year were randomly selected from Rhode Island birth certificate records to match with
cases on the basis of having the same year of birth.
   The basis for comparison was exposure to electric power lines at addresses.  The 119
cases produced 209 case addresses; all addresses prior to onset were included. Of the 119
cases, 66 had lived at only one address,  34 had lived at two addresses, and 19 had lived at
three or more addresses.  There were 240 control addresses identified from birth certificates.
The residences  and nearby power lines were sketched and the shortest distance between the
home and lines  was measured with a calibrated, split-image, optical range finder. Maps were
made for 198 case addresses and 225 control addresses. Addresses that could not be
mapped were not included in the analyses. Residence at both birth and diagnosis were
included for the cases; only birth residence was included for controls.  Approximating the
method used by Wertheimer and Leeper, four types of wires were identified, e.g., high tension,
large gauge primary, small gauge primary, and secondary bundle (including two hot lead and
one cold ground wire).
   Since power lines can have several wires, the wire information was grouped per address
into a summary  exposure value weighted by each type of wire present.  The weights used
were the relative median field strengths obtained by Wertheimer and Leeper (1979). The
       10/18/90
3-10

-------
                           DRAFT-DO NOT QUOTE OR CITE

maximum and median measurements were taken at a point described as 75 centimeters (cm)
above ground, under the wires. The source of this information from the Wertheimer and
Leeper study is not given. The summary exposures of controls were divided into quartiles to
assign four exposure categories for both cases and controls of very low, low, high, and very
high relative exposures.
   Across the four exposure groups, the control addresses were compared to the case
addresses in toto and to the case address by cell type (lymphoblastic or other), by stable
address, by age at onset (<7 years or >7 years), and by socioeconomic status (high vs. low).
In no analysis was the percentage of control addresses different from the percentage of cases
distributed across the four exposure groups. The authors saw no trend or tendency; the
frequency distributions in all situations were approximately equivalent.
   The greatest difference  (p<0.20) occurred when "stable" addresses (cases who had lived
in only one home) were compared to the control addresses. It should be noted that this may
be the most appropriate comparison because, overall, of the 119 cases, 55% had only one
address (or, alternatively, of the 198 mapped case addresses, about 29% had one address
prior to onset), and all the controls  had only one mapped (birth) address. Address histories
were not developed for controls. The mixing of multiple addresses for cases but not for
controls is inappropriate.
   Furthermore, the method of comparing "addresses" of cases and controls rather than
actual persons is methodologically flawed.  The authors noted in their paper that the 119
cases provided a total of 209 different "addresses" or a excess of 90 additional addresses. If
you assume that exposure to magnetic fields is associated with leukemia, then you would
have to assume that no exposure to fields is not associated with leukemia.  Multiple addresses
on the same case are not likely to be all located in relevant magnetic fields.  Hence, the 90
additional addresses included for the cases probably have a distribution of association of
exposure to magnetic fields similar  to that of the control addresses. The net effect then of
adding 90 essentially honexposed case addresses would be to dilute the effect and reduce
perhaps what might have been a positive finding to a nonpositive finding. Furthermore, the
authors made no effort to estimate the mean and median magnetic fields in the four categories
of exposure utilized. However, residential estimates of intensities by means of mathematical
formulation were accomplished. The authors could not substantiate that their wiring code
categories reflected "high" or "low" EM fields.
                                        3-11
10/18/90

-------
                           DRAFT-DO NOT QUOTE OR CITE

   As with Wertheimer and Leeper, odds ratios were not presented. The applied statistical
tests were not cited. The authors did note that matched-pair analysis was not used because
the controls and cases were not matched except for selection of same birth year.
   A communication between the authors (Fulton et al.) and Wertheimer revealed that her field
strength readings were for three-phase wires and so were representative of groups rather than
single wires.  Fulton et al. then reweighted and recalculated the summary exposure values.
They reported similar overall results, that is, no association; these data are not given in the
paper.
   The authors intended to repeat the substance of the Wertheimer and Leeper approach to
examining wiring and childhood leukemia in another geographic region. Although some
aspects of the study are similar to the Denver study, there are sufficient differences and
methodological problems with this study to negate its use as a validity check to the
Wertheimer and Leeper study. The methodological flaws of this study limit its usefulness in
assessing the risks of EM fields.
   Tomenius et al. (1982; Tomenius, 1986) conducted a case-control study on the incidence
of childhood cancer in the county of Stockholm, Sweden.  The county encompasses all of the
City of Stockholm plus 22 surrounding communities.  In Sweden, 50-Hz alternating current
(AC) is used, and most wires are buried.
   Cases were drawn from the Swedish Cancer Registry for 1958-1973 and included 716
individuals aged 0-18 years who were born and diagnosed in the county of Sweden. This
represents the exclusion of 175 cases from a total of 891 diagnosed and registered in the
system. Exclusions were primarily cases who did not meet the birth and diagnosis residency
requirements. Also, only primary tumors were included. The cases consisted of 660
malignant and 56 benign tumors.
   Controls were matched to cases for age, sex, and church district of birth and were selected
from birth registration records in parishes. Controls (N=716) also were limited to birth and
residence in the county of Stockholm.  There were  400 cases who still lived in the same church
district of birth at the time of their diagnosis; their controls also lived in the same church district
at the time of diagnosis. There were 316 cases who had moved from their church district of
birth by the time of their diagnosis; their controls could either still live inside or outside the birth
district. The matching procedures produced more  dwellings for cases (N=1172) than for
controls (N=1015).
       10/18/90
3-12

-------
                           DRAFT-DO NOT QUOTE OR CITE

   Exposure to EM fields at the residences of cases and controls was evaluated by
determining proximity to electrical sources and by taking single outdoor measurements.
Electrical structures indicating possible high-current flow were noted if they were within 150
meters of dwellings. These structures were high voltage wires  (6-200 kV), substations,
transformers, electric railroads,  and subways. Distance was measured by pacing off by foot.
There was little information on how long the electrical structures had been present, but
inquiries with residents living near 200 kV wires revealed that most wires had been in place at
the times for both birth and diagnosis.  Measurements were made in a blinded fashion.
Magnetic fields were measured at the entrance to single family homes and near both the
outside and individual  doors for apartment houses. Electric fields were not measured.
   There were 1172 dwellings among the cases and 1015 dwellings for the controls, i.e., 2187
total dwellings.  Eighty-nine dwellings were demolished or unoccupied; a total of 2098
addresses were visited (cases=1129, controls=969). Odds ratios (called relative risks by the
author) were developed using the total 2098 dwellings  (rather than the number of persons)
and were evaluated by the chi-square test.  Unfortunately, this study employs the same
methodology as the Fulton et al. (1980) study (inclusion of multiple and different case
addresses on the same case).
   Very few visible electrical constructions were noted for the sampled dwellings, i.e., 196 out
of 2098 dwellings or 9.3%. The structures examined were 200  kV wires, 6 to 200 kV wires,
substations, transformers, electric railroads, and subways. The most common structure was
electric railroads (N=58),  and the least common was substations (N=12). There was a
significant excess risk  of cancer in cases with dwellings within 150 meters of 200 kV wires
[odds ratio (OR) =2.1]  and a significant excess of case dwellings near all electrical structures
considered together (OR=1.3)  (Table 3-4). The distributions were also stratified by distance
from electrical constructions, i.e., 0-49, 50-99, and 100-150 meters (not shown here). However,
the risk tended to increase with increasing distance, an inverse relationship, from electrical
structures.  This result is the opposite of what might be anticipated, but numbers were small in
some strata so the results are probably unstable. The  mean magnetic field measurements
made for case and control dwellings grouped by electrical structures were generally similar
except for distance from 200 kV wires, where the mean value was 0.183 microtesla (^T) at
case dwellings and 0.329 //T at control dwellings. This result is contradictory.  But this
contradiction may be evidence of the difficulty of using "distance" from the source as a
                                         3-13
10/18/90

-------
                             DRAFT-DO NOT QUOTE OR CITE
   TABLE 3-4. ODDS RATIOS FOR THE DISTRIBUTION OF THE 2098 DWELLINGS OF CANCER CASES
        AND CONTROLS BY TYPE OF ELECTRICAL STRUCTURE VISIBLE WITHIN 150 METERS
     Electrical Structure
Cases
Controls
Odds Ratio
200 kV wires
6 to < 200 kV wires
Substations
Transformers
Electric railroads
Subway
Total
32
12
7
12
36
20
119
13
6
5
14
22
17
77
2.1a
1.7
1.2
0.7
1.4
1.0
1.3a
 *Per authors, p^O.05, chi-square test, and only given if expected numbers in each category were at least 5 and if odds ratios were
 different than 1.0.
 SOURCE: Tomenlus, 1986.
surrogate of dose. This study suffers from the same methodological flaws that pervade the
Fulton et at. study (i.e., inclusion of multiple case addresses on the same case.
    Birth and diagnosis dwellings were also examined separately. The observation of excess
risk for dwellings near 200 kV lines held for both birth and diagnosis dwellings, plus there were
significantly more case dwellings than control dwellings with fields <0.3/
-------
                            DRAFT-DO NOT QUOTE OR CITE
   In this situation, the odds ratio becomes 3.97, with a 95% confidence interval of 1.94-8.24.
This analysis obviates questions dealing with latency. This suggests that when length of
residence increases in the vicinity of such magnetic fields, the risk ratio tends to sharpen and
increase.
   The distribution of case dwellings and control dwellings with respect to one-time-only
measurements of magnetic fields at dwelling entryways was presented  (Table 3-5). The
average value of magnetic field measurements was similar in both case and control dwellings.
Yet, significantly more case (all causes) dwellings were found to have been located in
magnetic fields  >0.3/fT (OR=2.1).  For dwellings not near electrical constructions, a
significant excess of case dwellings with fields  >0.3/fT was observed (OR=2.1), but no
excess was seen for dwellings that were located near visible electrical structures.  Therefore,
this implies that the source of the fields and the estimated risk is from sources other than the
structures surveyed by the investigations. It is  evident that the identified structures in the study
are imperfect as a surrogate for magnetic field  levels. And, as noted, these measurements
were made outdoors.  This result also seems at odds with the differences that were noted in
the analysis  of individuals who live within 150 meters of 200 kV lines. The average magnetic
field measures for dwellings near 200 kV lines were significantly greater than the overall
average, and the levels increased with closeness to the lines. A gradient of potential exposure
was observed.  Yet, when dwellings were grouped as either >0.3/*T or <0.3 ^T, the odds
ratio was significantly increased for case dwellings for which measurements were  <0.3 /4~Y.
As seen with the inverse relation with distance, this apparent association with levels of
exposure  <0.3/* T does not fit a typical dose-response relationship.
   This reverse relationship may be an artifact. Proximity to "visible" 220 kV wires may not  be
a valid indicator of exposure without knowing how long the case or control lived in the dwelling
prior to the diagnosis. Because of the possibility that latency may have played a role in the
development of these cancers, it would have been more meaningful to have had a minimum
residency period before classifying a case or control according to a single measurement of the
magnetic field strength just outside of the doorway of the dwelling of the case or control.
Furthermore, it would have been much more relevant if all of the analyses done in this paper
dealt with individuals rather than addresses and multiple addresses on the same individual. It
                                         3-15
10/18/90

-------
                              DRAFT-DO NOT QUOTE OR CITE
  TABLE 3-5.  MAGNETIC FIELD MEASUREMENTS FOR CASE AND CONTROL DWELLINGS AND ODDS
          RATIOS BY MAGNETIC FIELD LEVEL AND TYPE OF ELECTRICAL CONSTRUCTION
Case
Dwellings
Control
Dwellings
Odds
Both Ratios
    Average Magnetic
     Field (ŤT)

    Range

    Average Field,
    200 kV lines nearby

    Total:
    > 0.3 XT
    < 0.3 ftf
     Total

    No construction:
    > 0.3 ftf
    < 0.3 ftT
     Total

    Any Electrical Construction:
    Ł: 0.3 /*T
    < 0.3 n-T
     Total

    200 kV wires:
    2: 0.3 (rt
    < 0.3 /
-------
                            DRAFT-DO NOT QUOTE OR CITE



appears that the significant odds ratios is mainly due to a deficit of control dwellings that are

classified in the category of exposure  <0.3 T. Perhaps this deficit resulted because some of

the 200 kV wires were not "visible" to the observer.  Misclassification because of the above

stated reasons may be the explanation for this anomaly in the data.

    Probably a finding of greater importance is the observation that the risk of cancer is much

greater when only the cases and controls who had no change in residency are considered, as

follows:
   Strength of
   Magnetic Field
Case
Control
Odds
Ratio
    >0.3/tT
    <0.3iaT
 10
242
  3
407
5.4*
     p<0.05

    However, another interesting finding is that when the analysis is properly restricted to just

"persons" and only residency at birth is considered, then again the risk of cancer is greater for

those persons residing in more intense magnetic fields, as follows:
   Strength of
   Magnetic Field
             Control
             Odds
             Ratio
    <0.3,wT

    *p<0.05
 21
661
  8
673
                                                      2.7*
    Likewise, in the same type of analysis but considering only residency at diagnosis, a

similar result is obtained as follows:
   Strength of
   Magnetic Field
             Control
             Odds
             Ratio
    <0.3<ŤT

    *p<0.05
  21
676
  9
689
                                                      2.4*
                                          3-17
                                                    10/18/90

-------
                             DRAFT-DO NOT QUOTE OR CITE

    When specific cancer sites, that is,  leukemia, lymphomas, nervous system, and/or other
 sites were analyzed, based upon case dwellings and control dwellings, only cancer of the
 nervous system is significant (Table 3-6) while that of lymphomas and/or other sites are
 nonsignificantly elevated. Leukemia is actually below expectation with a risk of 0.3. If persons
 were used as the units from which to calculate risk, all of these estimates would be elevated.  It
 is likely that the risk of lymphoma and risk of cancer of other sites would be significant.
 However, although the risk of leukemia would be raised slightly, it would still fall below one. It
 should be noted that the rate of leukemia seems to be unusually low in this Swedish
 population, regardless of the  EM fields issue.  This may reflect some anomaly  in the cancer
 registry. There is some evidence that leukemias may be underreported in the  Swedish cancer
 registry (Mattsson and Wallgren, 1984). Again, two problems emerge from these findings:
 first, the actual risks using persons rather than dwellings cannot be calculated precisely, and,
 second, length of residency at the address where the measurement of the magnetic strength
 was done is not known so that latency cannot be assessed. However, despite the limitations
 of this  study it does add to the evidence that EM fields are possibly associated with certain
 forms of cancer, i.e., nervous  system, lymphomas, and cancer of other sites, nonspecified. A
 general observation can be drawn from the varied multiple analyses, that is, positive
 associations were generally seen if residence was near 200 kV lines and if measured magnetic
 fields were >0.3/*T.

       TABLE 3-6. RISK RATIOS FOR SPECIFIC CANCER SITES AND MAGNETIC FIELD LEVEL
   All Visible Structures3
                                 <0.3fiT
Sites:
All
Leukemia
Lymphomas
Nervous System
Other Sites
All Malignant Neoplasms
Benign Neoplasms

2.1b
0.3
1.8.
3.7b
—
1.8
.

1.0
1.0
1.0
1.0
1.0
1.0
1.0
fDatafor 200 kV wires and other visible structures, considered separately, were given in the paper but are not given here.
°Per authors, pŁ0.05, chi-square test, and only given if expected numbers in each category were at least 5 and if odds ratios were
different than 1.0.
SOURCE: Tomenius, 1986.
       10/18/90
3-18

-------
                           DRAFT-DO NOT QUOTE OR CITE

    Myers et al. (1985) reported preliminary results of a case-control study that examined
childhood cancer in relation to 50-Hz magnetic fields from overhead power transmission and
distribution lines. A somewhat different approach for estimating magnetic field strengths was
taken in this study.  Exposure was estimated from records of load currents rather than from
spot measurements or coded wire configurations.
    Cases were defined as all children aged 15 years or less when diagnosed with cancer
between 1970-1979 and born within the Yorkshire Health  Region in England.  Cases were
identified from a childhood cancer register, formed from Health Service records. Additionally,
cases were included if their mother resided in the health district of study at the time of birth of
the subject child. A complete listing of incident cases was available for the period 1975-1979.
For the period 1970-1974, only children who had died by  1982 could be identified and
included plus a few survivors, Therefore, all cases diagnosed in 1970-1974 are not included in
the study. The. authors estimated that the underascertainment of incident cases was 15% of
the total.    ,
    Controls for deceased cases diagnosed between 1970-1974 were drawn from the local
birth registry and were chosen to have a similar birth date and nearby birth address to the
case. Control selection had been previously developed and made in the course of conducting
the National Oxford Childhood.Cancer Survey. Two controls per case were similarly selected
for cases diagnosed between 1975-1979 and for live cases diagnosed before  1975. It was
noted that control of local birth registers had been assumed by the Health Service after 1974;
thereby, the quality and completeness of the information collection differ in the two study
periods.
    A total of 376 cases and 590 controls were included in the study. Of these, 37 (9.8%)
cancer cases and 44 (7.5%) controls lived within 100 meters of an overhead power line.
    A master roster of birth addresses was assembled to mask the case vs. control status of
the subjects when coding exposure estimates. Magnetic  field exposures from overhead power
lines for case and control addresses were estimated by calculations that accounted for
maximum current load and by distance from the line. Load information was taken from
records of Electricity Boards. Maxiumum recorded loads  over 1974-1984 were used. Lines
not in use at the time of birth of a subject were excluded.  Because some relevant information
could not be derived from the electricity records, certain assumptions had to be made which
                                         3-19
10/18/90

-------
                            DRAFT-DO NOT QUOTE OR CITE

the authors believed tended to increase the calculated values. Distance was obtained from
measurements on Ordnance Survey maps.
    Statistical analyses were based on a stratified case-control design, and risk ratios were
calculated by the maximum likelihood method after Mantel-Haenszel with confidence limits
after Miettinen.
    Distance of residence from an overhead power line was grouped into intervals of 0-24,
25-49, 50-74, 74-99, and greater than 100 meters (Table 3-7).  The latter grouping was taken as
the reference group. For all cancers considered together, the odds ratios were greater than
unity at all distances less than 100 meters, but there was no trend of increased risk with
decreased  distance. The largest ratio was 1.6 at 24-49 meters.  No value was statistically
significant.  For analysis of lymphomas and leukemias, considered together, and of solid
tumors, there was no clear pattern of increased risk with decreased distance.  The risk ratios
were generally elevated, but no increase was statistically significant. The highest risk ratio
estimated for lymphomas and leukemias was 2.6 at 50-74 meters; statistical significance may
be viewed as marginal with a p-value of 0.05.
 TABLE 3-7. DISTRIBUTION OF CASES AND CONTROLS BY DISTANCE FROM OVERHEAD POWER LINES AND BY CANCER
                                          TYPE


Lymphomas/
Leukemias



Solid
Tumors



All
Cancers




Distance3
0-24
25-49
50-74
75-99
>100
0-24
25-49
50-74
75-99
>100
0-24
25-49
50-74
75-99
>100

Cases
1
8
9
3
169
3
4
5
4
170
4
12
14
7
339

Controls
3
8
6
5
269
3
4
10
5
277
6
12
16
10
546

RRa
0.5
1.6
2.4
1.0
1.0
1.6
1.6
0.8
1.3
1.0
1.1
1.6
1.4
1.1
1.0
95%
cia
0.06-5.0
0.6-4.3
0.9-6.6
0.2-4.1
—
0.3-8.1
0.4-6.5
0.3-2.4
0.35-4.9
—
0.3-3.8
0.7-3.6
0.7-2.9
0.4-3.0
—

p-value
0.29
0.18
0.05
0.48
-
0.27
0.25
0.26
0.35
_
0.46
0.12
0.13
0.41
-
^Distance is in meters. RR=risk ratio, Cl=confidence interval.
"Level of significance at 5%.
SOURCE:  Myers etal., 1985.
       10/18/90
3-20

-------
                            DRAFT-DO NOT QUOTE OR CITE

    In this analysis, the risk of lymphomas/leukemia or "solid" tumors was not associated with
 increasing distance (in four strata) from 50-Hz overhead power lines. However, when the
 distances are collapsed to only two categories, i.e., <100 meters versus >100 meters, then
 the risk ratios are 1.5,1.2, and 1.4 for lymphomas/leukemias, solid tumors, and all cancer,
 respectively. These slightly elevated risk estimates suggest that there may be an association
 with "closeness" to power lines, but this association may be sharpened by utilization of actual
 measurements of the magnetic field intensity rather than a poor surrogate such as distance.
    The next analysis by the authors was to take only those cases and controls who resided
 within 100 meters from the overhead power lines and assign an estimated magnetic field
 strength in milligauss to each person and then estimate risks based upon intensity of the fields
 (Table 3-8). Although most of the risks were elevated, none were significant, and no
 dose-response relationship surfaced.  Field values were grouped as less than 0.010,
 0.010-0.099, 0.10-0.99,1.00-9.99, and greater than 10.00 milligauss (mG). In units of
  TABLE 3-8. DISTRIBUTION OF CASES AND CONTROLS BY ESTIMATED MAGNETIC FIELD LEVEL IN
                          MILLIGAUSS (mG) AND BY CANCER TYPE

Lymphbmas/
Leukemias



Solid
Tumors



All
Cancers



Field (mG)
< 0.010
0.010-0.099
0.10-0.99
1.0-9.99
>10.0
< 0.010
0.010-0.099
0.10-0.99
1.0-9.99
>10.0
< 0.010
0.010-0.099
0.10-0.99
1.0-9.99
>10.0
Cases
6
7
4
2
2
3
6
5
2
0
9
13
9
4
2
Controls
4
5
8
4
1
4
8
7
1
2
8
13
15
5
3
RRa
2.4
2.2
0.8
0.8
3.2
1.2
1.2
1.2
3.3
NA
1.8
1.6
1.0
1.3
1.1
95%
Cla
0.7-8.3
0.7-7.0
0.2-2.7
0.1-4.4
0.3-31.2
0.3-5.5
0.4-3.6
0.4-2.2
0.3-31.8
NA
0.7-4.7
0.7-3.5
0.4-2.2
0.3-4.8
0.2-6.5
p-valueb
0.08
0.08
0.36
0.40
0.16
0.38
0.36
0.40
0.15
NA
0.11
0.11
0.47
0.35
0.47
f*RR=risk ratio. Cl = confidence interval, NA=not applicable.
 Level of significance at 5%.

SOURCE: Myers etal., 1985.
                                         3-21
10/18/90

-------
                            DRAFT-DO NOT QUOTE OR CITE

microtesla, these grouped ranges are less than 0.001, 0.001-0.0099, 0.01-0.099, 0.1-0.999, and
greater than 1.0//T. Again, subjects residing over 100 meters from a line were taken as the  •
reference group.
    However, this analysis suffers from several  limitations. Unfortunately, to calculate odds
ratios they used as a reference those cases and controls located beyond 100 meters from the
power lines on the implicit assumption that none of those cases and controls were exposed to
magnetic fields. Obviously, it would have been more appropriate to classify everyone
according to their respective magnetic field strengths and then analyse for dose-response
relationships. It is likely that persons living beyond the 100-meter point were subject to
magnetic field exposures from other sources. The character of these fields are probably not
different from those produced by overhead power lines.  Furthermore, the vast majority of the
few cases and controls that did fall within 100 meters were subject to magnetic fields
intensities that were very low, i.e., less than 1.0 mG (0.1 fiT). With the few cases and controls
that could be classified into five-field level strata (Table 3-8), the authors  did note an elevated
risk ratio for lymphomas/leukemias  (OR= 3.2) in the category with the highest magnetic field
intensity, i.e., >10.0 mG (1.0^T), based upon three persons.  The residual category of solid
tumors as well as all cancers also produced elevated risks for most of the field strength
subcategories.  But, again, the small numbers indicated little power to detect any moderately
elevated  odds ratios as significant if there is a true risk.
   There are other potential  problems that are not addressed by the authors. The accuracy of
the estimated magnetic field strengths, calculated from data on load, is an  issue.
Measurements per dwelling were not made.  Similarly, the calculations may not reflect
exposure received in the past.  Length of residency is not addressed.  A  recent residence
should not necessarily be counted as the residence of exposure. The estimated
underascertainment of cases (about 15%)  in the face of elevated but nonsignificant excesses
in risk raises the concern that the risk ratios and their precision may be underestimated
assuming a similar distribution pattern would be seen for additional cases. Consequently,
there are limits to the extent that this paper can help determine the  potential carcinogenicity of
electric and magnetic fields.
   Savitz et at.  (1988) reported on the results of a case-control study of childhood cancer and
exposure to 60-Hz magnetic fields in residences in Denver, Colorado.  These results were also
reported in Ahlbom et al. (1987) and Savitz (1987). All cases of cancer among children aged
       10/18/90
3-22

-------
                           DRAFT-DO NOT QUOTE OR CITE

14 years or younger, residing in the Denver Standard Metropolitan Statistical Area, and
diagnosed January 1,1976, through December 31,1983, were to be included in the study.
Cases were primarily identified from the Colorado Central Cancer Registry. Additional
ascertainment was completed using records of area hospitals. Over 98% of the cases were
confirmed either by microscopy, direct visualization, or radiography. Additionally, pediatric
oncotegists reviewed diagnostic accuracy.
    Controls were identified using random digit dialing methods and were matched to cases
on age (ą3 years), sex, and telephone exchange area.  Controls were restricted to those
children who had lived at their residence at the time of diagnosis of their matched case.  This
approach  excluded children who had recently moved into the area. The authors pointed out
that because the "control pool" did not include children who had left the area between the time
cases were diagnosed and controls were selected, bias could be introduced if migration was
due to magnetic field exposures and control versus case status.
    If located and if consent was given, case and control parents were interviewed using a
detailed questionnaire. The preferred interviewee was the biologic mother. Inquiries were
made about potential risk factors for childhood cancer, e.g., family demography, residential
history, family cancer history, in utero and postnatal exposure to x-rays and medications,
parental occupational history, and medical history of the child and of the mother during
pregnancy. Telephone interviews (11 % for cases, 2% for controls) were conducted if parents
had moved from the study region or if parents refused an in-home interview.
    Electric and magnetic field measurements were made at residences that were  occupied
prior to diagnosis and were taken near the front door, the child's bedroom, the parents'
bedroom,  and any room reported to have been occupied by the child on an average of one or
more hours per day. Measurements were taken near the center of the room to avoid proximity
to appliances or large metal objects. To attempt to isolate the outdoor contribution to
magnetic fields in the home, measurements were made  under both "low power" (home power
off)  and "high power" (all lights and selected appliances on) conditions. Because these were
short-term or spot measurements, potential confounders were noted, e.g., time of day, day of
week, outdoor temperature, and domestic electric load on the distribution  system.  For the
latter, data were made available from the Public Service  Company of Colorado.
    Consideration was made of how to summarize the measurement data to derive a summary
home value. Correlation between three approaches, e.g., a simple arithmetic mean of all
                                        3-23
10/18/90

-------
                           DRAFT-DO NOT QUOTE OR CITE

values, a mean of three locations measured in all homes, and a time-weighted average, was
very high (0.95).  As a result, simple arithmetic means were used. Magnetic field values were
then categorized by cutoff scores of 0.65,1.0, and 2.5 mG (0.065, 0.1, and 0.25^1).  Electric
field values were similarly categorized, i.e., 6.0, 9.0, and 14.0 volts per meter (V/m). The
rationale for these cutoff values was not stated. Dichotomous categories were also
developed, using 2.0 mG (0.2/fT) and 12.0 V/m as the cutoffs for magnetic and electric fields,
respectively, in order to isolate the high end of the distributions.
    In addition, wiring configuration codes were also developed. Procedures were used to
mask the case or control status of the occupants of coded residences. Coding methods were
based on those used by Wertheimer and Leeper (1979).
    The distribution of cases by selected cancer site is given in Table 3-9 which also indicates
the number and percentage response by cancer site.  Overall, the interview response rate for
cases (70.8%) was poorer than that for the controls (79.9%).  Table 3-10 shows the number
and percentage of cases and controls, for whom magnetic field measurements were  derived
and wire codes were never proscribed. Only 36% of the case residences were measured for
electric and magnetic fields, while 74.5% of the residences of the corresponding controls were
measured. This response rate is poor. Because the nonrespondents were mostly persons
who had moved after diagnosis, the "measured" homes were occupied by the most
residentially stable. Wire codes, on the other hand, were obtained for a noticeably higher
proportion of both cases (89.6%) and controls (93.2%), and these response rates are similar.
Wire codes could be developed for a similar proportion of case and control homes. However,
very few case homes could be measured.
    When consideration is given to magnetic field measurements (stratified into four levels:
<0.65, 0.65-<1.0,1,0-<2.5, and 2.5+ mG), the risk of cancer under low-power conditions is
elevated, albeit nonsignificantly, with  increased exposure. However, in contrast, the risk of
cancer under high-power conditions is weak, and when intensity of electric fields is elevated
under high-power conditions, the risk is equal to that expected (Table 3-11).  These results
must  be considered somewhat unreliable due to the rather large "nonresponse"  (almost 70% in
the cases) on measured  readings.
    These same data were used to evaluate site-specific cancer under low-power, high-power,
and high-voltage conditions with measured values dichotomized, i.e., >2.0 mG and <2.0 mG,
(Table 3-12).  Again, the results are equivocal. Nonsignificantly elevated risks were observed
       10/18/90
3-24

-------
                           DRAFT-DO NOT QUOTE OR CITE

  TABLE 3-9. ELIGIBILITY, RESPONSE, AND LOSSES FOR CANCER CASES, CASE SUBGROUPS AND
  	       CONTROLS: DENVER STANDARD METROPOLITAN STATISTICAL AREA
    Group
No. of
Eligibles
 Interviewed
No.       %
  Untraced
No.
            Refusals3
          No.       %
    Cancer Cases     356
     Leukemia        103
     Acute lymphocytic
      leukemia        83
     Brain            67
     Lymphoma       35
     Soft tissue        32
     Other cancers    119

    Controls          278
             252
             73

             59
             48
             26
             26
             79

             222
         70.8
         70.9

         71.1
         71.6
         74.3
         81.3
         66.4

         79.9
61
17.1
35
                            9.8
 "Includes individual refusals and physician refusals.

 SOURCE:  Savitzetal., 1988.
    TABLE 3-10. DISTRIBUTION OF ELECTRIC AND MAGNETIC FIELD MEASUREMENTS AND WIRE
     CODES FOR CANCER CASES, CASE SUBGROUPS, AND CONTROLS: DENVER STANDARD
                           METROPOLITAN STATISTICAL AREA
Field

Group
Cancer Cases
Leukemia
Acute lymphocytic
leukemia
Brain
Lymphoma
Soft tissue
Other cancers
No. of
Eligibles
356
103

83
67
35
32
119
Measurement
No.
128
36

26
25
1 3
14
40
%
36.0
35.0

31.3
37.3
37.1
43.8
33.6
Wire Code
No.
320
97

78
59
30
32
102
%
89.9
94.2

94.0
88.1
85.7
100.0
85.7
   Controls
      278
           207
                                                      74.5
                                               259
                                           93.2
SOURCE: Savitzetal., 1988.
                                       3-25
                                                       10/18/90

-------
                            DRAFT-DO NOT QUOTE OR CITE
 TABLE 3-11. CANCER RISK IN RELATION TO MEASURED MAGNETIC FIELDS AND ELECTRIC FIELDS,
   UNDER LOW- OR HIGH-POWER USE CONDITIONS, IN RESIDENCES OCCUPIED AT DIAGNOSIS:
                    DENVER STANDARD METROPOLITAN STATISTICAL AREA
   Exposure Level
     Odds Ratio
95% Confidence Interval
   Magnetic Fields (mG)a:
    low-power use

    0  -  <0.65
    0.65- <1.0
    1.0 - <2.5
    2.5 +

   Magnetic Fields (mG)a:
    high-power use

    0  -  <0.65
    0.65- <1.0
    1.0 - <2.5
    2.5 +

   Electric Fields (V/m)b:
    high-power use

     0 -  <6.0
     6.0- <9.0
     9.0-< 14.0
    14.0 +
      1.00
      1.28
      1.25
      1.49
       1.00
       1.13
       0.96
       1.17
       1.00
       0.88
       1.23
       0.90
0.67-2.42
0.68-2.28
0.62-3.60
0.61-2.11
0.56-1.65
0.54-2.57
0.49-1.58
0.68-2.22
0.43-1.88
frnG Ť milligauss.
^//m = volts per meter.
1.0 mG = 0.1 ftT

SOURCE: Savitzetal.. 1988.
       10/18/90
3-26

-------
                          DRAFT-DO NOT QUOTE OR CITE
 TABLE 3-12. CANCER RISK (ODDS RATIOS WITH 95% CONFIDENCE INTERVALS IN PARENTHESES) IN
   RELATION TO MAGNETIC FIELDS AND ELECTRIC FIELDS, CATEGORIZED INTO TWO EXPOSURE
  GROUPS3 AND MEASURED UNDER LOW- OR HIGH-POWER USE CONDITIONS, IN RESIDENCES OC-
        CUPIED AT DIAGNOSIS: DENVER STANDARD METROPOLITAN STATISTICAL AREA


Site

All Cancers

Leukemia

Acute lymphocytic
leukemia
Lymphoma

Brain

Soft tissue sarcomas

Other Cancers

Magnetic
fields:
low power
(>0.2AT)
1.35
(0.63-2.90)
1.93
(0.67-5.56)
1.56
(0.42-5.72)
2.17
(0.46-10.31)
1.04
(0.22-4.82)
3.26
(0.88-12.07)
0.31
(0.44-2.14)
Magnetic
fields:
high power
(>0.2/*T)
1.04
(0.56-1.95)
1.41
(0.57-3.50)
1.05
(0.34-3.26)
1.81
(0.48-6.88)
0.82
(0.23-2.93)
1.65
(0.44-6.20)
0.49
(0.14-1.66)
Electric
fields:
high power

0.93
(0.53-1.61)
0.75
(0.29-1.91)
0.67
(0.22-2.04)
0.70
(0.15-3.27)
0.53
(0.15-1.81)
0.64
(0.14-2.96)
1.65
(0.78-3.51)
aFor magnetic fields, <2.0 mG and 2.0+ mG. For electric fields, < 12.0 V/m and 12.0+ V/m.

SOURCE: Savitzetal., 1988.
                                      3-27
10/18/90

-------
                           DRAFT-DO NOT QUOTE OR CITE

for leukemia, lymphoma, brain, and soft tissue under low-power conditions.  But high-power
conditions produced somewhat reduced risks at these same sites.  It is problematic that the
risk estimates for magnetic fields are greater under low-power usage conditions than under
high power.  One might have expected that the greater the exposure, the greater the
magnitude of the response.  However, a possible explanation may be that while all measures
reflected present day single exposures, measures for high-power use reflect the most transient
and variable exposures, as is the nature of exposure from appliances.  Also, this seeming
reverse dose-response relationship must be viewed in light of the poor nonresponse of cases
on measured electromagnetic exposure. In addition, the author points out that electrical
consumption at the time of measurement could have influenced the results.  Electrical demand
on the system at the time of diagnosis and before diagnosis was not known. It must also be
remembered that these measurements were taken many years after the diagnosis, perhaps as
much as 9 years.  Savitz et al. (1988) concluded that, because long-term field exposures could
not be captured through measurement procedures, no further analysis would be done using
these values. The pitfall is that it is not known whether the missing data, if it were available,
would have produced  a different, perhaps unbiased result.
   The main point to be gained from this analysis is that, there appears to be a modest
increase in risk, up to 3.26, in several types of cancer in children who resided in homes where
magnetic fields were >2 mG. If 3.0 mG was used, similar to Tomenius, the authors noted an
increase in site-specific odds ratios. No numbers are given, but they were stated to be
imprecise.
    Stratified analyses for  low-power use magnetic field data were made to examine potential
confounding with maternal age, father's education, per capita income, maternal smoking
during pregnancy, and traffic density. The authors stated that the adjusted odds ratios did not
differ appreciably from the odds ratios developed with magnetic field cutoffs of 2.0 mG. The
adjusted risk ratios for total cancer ranged from 1.2 to 1.5 versus an unadjusted risk ratio of
1.4, and the risk ratios for all adjusted leukemias ranged from 1.8 to 2.4 versus an unadjusted
risk ratio of 1.9. The adjusted estimates were less precise with fewer subjects.  Only
lymphomas, when adjusted, seemed to be confounded for per capita  income since the risk
ratios changed from 2.2 to 3.2.
    Wire code configurations were also developed and categorized as buried, very low, low,
high, and very high (Table 3-13).  Considering the limitations of the measurement data, the
       10/18/90
3-28

-------
                           DRAFT-DO NOT QUOTE OR CITE
      TABLE 3-13. CANCER RISK (ODDS RATIOS FOR ALL SITES COMBINED) IN RELATION TO A
     FIVE-LEVEL WIRING CONFIGURATION CODE FOR RESIDENCES OCCUPIED AT THE TIME OF
   DIAGNOSIS OR 2 YEARS BEFORE DIAGNOSIS: DENVER STANDARD METROPOLITAN STATISTICAL
                                        AREA
A.








At Diagnosis:
Number3
Code Cases
Buriedb 95
Very low 29
Low 107
High 70
Very high 19
Missing 36


Controls
88
17
102
44
8
19

Odds
Ratio
1.00
1.58
0.97
1.47
2.20
_.

95% Confidence
Interval

0.81-3.07
0.65-1 .45
0.92-2.37
0.93-5.21
—
Mantel Chi-Square test for trend = 2.03, p = 0.02
B.








Mantel
Two Years Before Diagnosis:
Number3
Code Cases
Buriedb , 36
Very Low 1 1
Low 50
High 30
Very High 8
Missing 221
Chi-Square test for trend = 2.31, p = 0.01


Controls
47
15
56
28
2
130


Odds
Ratio
1.00
0.96
1.17
1.40
5.22
—


95% Confidence
Interval

0.39-2.34
0.65-2.08
0.71-2.75
1.18-23.09


|*The differences in the number of cases and controls per group reflect the availability of data
 Taken as the reference group.
SOURCE: Savitzetal., 1988.
wire code surrogate of exposure is probably a more useful estimate of relevant magnetic and
electric fields. Unlike measurements, which were made many years after diagnosis, wiring
configurations are historically more stable or permanent and so probably give a better
estimate of actual exposures at the time of and before diagnosis. In addition, about 90% of
both cases and control homes could be coded. That wiring code is probably a good indicator
of exposure is supported by the limited actual measurement data taken under low- and
high-power magnetic use conditions, as follows:
                                        3-29
10/18/90

-------
                           DRAFT-DO NOT QUOTE OR CITE
                                              Magnetic Fields (mG) by Wire Code
   High Power
   Low Power
Mean
Median
Mean
Median
Buried
0.60
0.39
0.49
0.30
Very Low
0.77
0.45
0.53
0.30
Low
1.09
0.73
0.71
0.51
Hiah
1.61
1.12
1.22
0.90
Very Hiah
2.92
2.01
2.12
2.16
A clear gradient of increasing intensity of exposure with increasing wire code is readily
apparent.
   Subjects who lived in homes with buried wires were assumed to be the no exposure group
(Table 3-13). The highest risk, which was not significant, for cancer at all sites was obtained
for very high wire code homes (OR=2.2). Although the confidence intervals spanned unity in
all code categories, the Mantel chi-square test for trend was significant (2.03, p=0.02). The
investigators also coded the wiring configurations,  where possible, of homes occupied 2 years
prior to diagnosis to allow for possible latency. The gradient was sharper here.  A significant
risk (OR=5.22) was seen for very high code homes. Again, the Mantel chi-square test for
trend was statistically significant (2.31, p=0.01), providing evidence of a dose-response
relationship for cancer of all sites combined.
   When consideration was given to the site-specific cancers (leukemia, lymphoma, brain, soft
tissues, and other cancers) by a dichotomized two-level wire code, i.e., low (buried, low, very
low) and high (high versus very high), it was found that the risk was elevated at all sites,
except lymphoma (Table 3-14). For brain cancer and total cancer (all sites combined), the risk
was significantly elevated at 2.04 and 1.53, respectively. Furthermore, when only the most
extreme codes were used, i.e., buried versus very high, the risks were even higher, but the
numbers were too small to detect an elevated risk as significant.  For leukemia, the risk was
2.75, for brain it was 1.94, and for lymphoma it was 3.3. Again, the authors stated that
stratified analyses to evaluate confounders did not demonstrate any changes in these risk
estimates.
   In summary, this paper provides evidence that exposure to EM fields, as represented by
wiring codes, is positively associated with certain site-specific cancers, most notably leukemia,
lymphoma, and brain cancer.  In addition, a significant dose-response relationship is evident
for total cancer based upon wire code configurations at the time of diagnosis as well as 2
years before (Table 3-13). Average and  median measurements of the intensity of the magnetic
       10/18/90
                     3-30

-------
                            DRAFT-DO NOT QUOTE OR CITE
  TABLE 3-14. CANCER RISK (ODDS RATIOS) IN RELATION TO DICHOTOMIZED WIRE CODES (LOW VS
  HIGH AND BURIED VS. VERY HIGH) FOR RESIDENCES OCCUPIED AT THE TIME OF DIAGNOSIS FOR'
  ALL CANCERS AND FOR SPECIFIC SITES: DENVER STANDARD METROPOLITAN STATISTICAL AREA
    A.
Two Level Wire Cndp

Controls
Total cases
Leukemia
Acute lympho-
cytic leukemia
Lymphoma
Brain
Soft tissue
Other cancers
B. Very High Versus Buried


Controls
Total cases
Leukemia
Acute lympho-
cytic leukemia
Lymphoma
Brain
Soft tissue
Other cancers

Low5
207
231
70
59

25
39
23
74
Wire Code

Buried3
88
95
28
24

10
17
13
27

High8-
52
89
27
19

5
. 20
9
28


Very High3
8
19
7
6

3
3
2
4
Odds
Ratio
1.53
1.28
1.28

0.80
2.04
1.56
1.51

Odds
Ratio
ť*.
2.20
2.75
2.75

3.30
1.94
1.69
1.63
95% Confidence
Interval
1 .04-2.26
0.90-2.63
0.70-2.34

0.29-2.18
1.11-3.76
0.68-3.55
0.89-2.56

95% Confidence
Interval

0.98-5.21
0.94-8.04
0.90-8.44

0.80-13.65
0.47-7.95
0.33-8.78
0.46-5.81
"Low = buried, very low, low; High = high, very high.
SOURCE: Savitzetal., 1988.
fields calculated for each of the five wiring codes support the use of wiring configuration codes
as surrogates for exposure to magnetic fields.  In the wire code category denoting highest
exposure, called "very high," the magnetic field intensities fall mostly in the range of 2 mG to 3
mG. Average and median measurements taken in the remaining four wire code categories fall
below 2 mG.
   There are two problems with the data that could influence detection of a significant
association of actual measured magnetic fields with cancer. First, there was a poor response
for making magnetic field measurements  (36% of the case residences were measured while
                                        3-31
                                                                  10/18/90

-------
                           DRAFT-DO NOT QUOTE OR CITE

74.5% of the control residences were measured). The authors suggested that the low and
differential response may introduce bias, and if there had been an "improved response for
magnetic fields," the observed odds ratios would have been elevated. Lack of such exposure
information would tend to bias the results toward the null. Second, the measurements were
taken many years after diagnosis, up to as much as 9 years later. Such measurements
probably do not adequately reflect long-term past magnetic field intensities prior to and during
the time of diagnosis.  The authors also pointed out that although wiring codes are not perfect
predictors of magnetic field intensities, they do tend to remain stable over long periods of time
and "better approximate historical field levels."
   In a very brief discussion appearing as a "project resume," Lin and Lu (1989) have reported
preliminary results of a case-control study of childhood cancer and residential exposure to EM
fields. Cases were diagnosed in the last 5 years (actual period of time is not stated) in the
Taipei Metropolitan Area in Taiwan.  The 216 identified cases were matched to 422 hospital
controls on the basis of age, sex, and date of admission to the same hospital of case
diagnosis.  EM-field exposure was classified as either low or high, with high exposure defined
as residing within 50 meters of either a high tension power line  (20 kV or greater), a
transformer, or a substation. Elevated nonsignificant risks of lymphomas, leukemias, and
cancers (all sites combined) were found.  However, these are based on small numbers and the
power to detect these risks as significant is low. The results were:
    All cancers
    Leukemias
    Lymphomas
    Brain Tumors
Odds
Ratio
1.30
1.31
2.0
1.09
Confidence
Interval
0.92 - 1 .84
0.78 - 2.21
0.62 - 6.50
0.50 - 2.37
    Additional information not available in the report is needed to evaluate this study more
 thoroughly. These brief results were presented at an annual Department of Energy (DOE)
 review of research on biological effects of 50- and 60-Hz electric and magnetic fields.
       10/18/90
3-32

-------
                            DRAFT-DO NOT QUOTE OR CITE

    For completeness, one case series reported in the literature should be described. Aldrich
 et al. (1984) identified an unusual cluster of five cases of endodermal sinus tumors in black
 children (all girls) from a hospital-based tumor registry in Jacksonville, Florida.  These tumors
 are very rare; the authors noted that little is known about their etiology, but they have been
 associated with twinning and are more frequently seen in black females.
    The authors discussed possible factors that might be related to the development of
 endodermal sinus tumors in these five cases, including possible  interaction between EM fields
 and various other environmental factors and interaction between genetic or familial factors and
 environmental agents. The area was primarily residential, but there were many large
 warehouses, an electromotive plant, and a lead smelter. The plants were stated to be in
 compliance with emission standards. Random blood leads were taken from residents in the
 area, including case family members, and were found to be similar to blood lead levels in
 persons residing  in other parts of Jacksonville. The area was also crossed by U.S. Highway 1,
 railroad lines, and electrical power lines.  The latter included primary distribution lines (26 kV)
 providing residential service via transformers and six primary transmission electrical lines each
 with 69 kV phase  to phase. These lines had been in place since  the 1950s, i.e., throughout
 the lifetime of the  cases. The distance that cases resided from the 69 kV lines ranged from 14
 to 592 feet.  Estimated magnetic fields at these distances were calculated and ranged from
 0.04 to 1.69 gauss (40 to 1690 mG).
   As a case series, this report provides little evidence that could bear on determining
 whether there is a causal relationship between EM fields and cancer. The authors discussed
 several factors, including possible exposure to EM fields, that might be related to this rare
 cancer. No factor was named as being causally related to endodermal sinus tumors. The
 authors suggested that further consideration be given to potential environmental exposures,
 including EM fields.

3.2.2.  Electromagnetic-Field Exposure at Unspecified Frequencies
   The following studies are  of cancer in children associated with exposures of fathers to
potential carcinogens encountered occupationally.
   In a case-control study, Spitz and Johnson (1985) examined neuroblastoma deaths in
children under age 15  in Texas between 1964 and 1978 in relation to paternal occupation.
Birth certificates corresponding to each case of neuroblastoma were obtained from the Texas
                                         3-33
10/18/90

-------
                            DRAFT-DO NOT QUOTE OR CITE

State Department of Health. If a birth certificate was not found, the case was excluded. These
birth certificates were then matched to one randomly selected control birth certificate within the
same birth year and to the 100th certificate within the same birth year from State birth rosters
grouped according to year of birth. There were 157 cases and 314 controls. Fathers'
occupation and industry at time of birth of the subject child were taken from the birth
certificates of both cases and controls and coded according to the Standard Industrial
Classification Manual (1972). These broad industrial code groups (agriculture, construction,
retail, etc.) were also regrouped into "clusters" based on shared exposures to different agents
and then further subdivided. This method was described in Hsieh et al. (1983).
    The risk of neuroblastoma for the broad industrial groups ranged from 0.55 to 1.35, and
none was significantly elevated. A significant increase in risk (OR=3.17, Cl=1.13-8.89) was
seen in the exposure cluster corresponding to aromatic and aliphatic hydrocarbons. This
cluster included workers in electric, electronic, and printing occupations; electricians;
insulation workers; and utility workers.  In an attempt to define this "cluster," occupations in the
cluster were reclassified into two groups  corresponding to differing levels of exposure to
electromagnetic radiation. The first group, whose exposure was presumed to be higher,
included electricians, electric and electronic workers,  linemen, utility employees, and welders.
The second group had lower presumed exposures and included (but was not limited to)
electric equipment salesmen and repairmen.  The risks were elevated in both groups, 2.14
(01=0.95-4.82) and 2.13 (Cl=1.05-4.35), respectively. The latter was significant. Of interest is
the fact that for children whose fathers were electronic workers, the risk was 11.75 and was
statistically significant (Cl=1.40-98.55). The authors noted that this job had potential EM-field
exposure.
    The major problems with these data involve use of birth certificate records for the
derivation of fathers' employment as well as lack of detailed exposure information on EM
fields.  Occupational designations from vital documents are not fully reliable. The occupation
of interest is not necessarily the occupation listed on the birth certificate at the time of birth but
should be that occupation the father was in before conception. There is little detail regarding
paternal employment before conception.  Furthermore, no data are presented regarding
maternal exposure before conception. This study suggests the need for further in-depth
research.
       10/18/90
3-34

-------
                             DRAFT-DO NOT QUOTE OR CITE

    Wilkins and Koutras (1988) reported on the results of a case-control study that examined
 mortality from brain cancer among Ohio-born children in relation to paternal occupation.
 Cases were identified as white children less than 20 years old who had died in Ohio from 1959
 through 1978 arid whose death certificate listed primary brain cancer as the underlying cause
 of death. Specific types of brain cancer could not be determined from death certificate data.
 Deaths occurring from 1959 through 1968 were coded according to the Seventh Revision of
 the International Classification of Diseases (ICD7:193.0), and deaths occurring from 1969
 through 1978 were coded according to the eighth revision (ICD8:191). Birth certificates were
 traced by linkage to name, parent's name, and date of birth as listed on the death certificate.
 Father's industry and occupation were recorded on Ohio birth certificates only for births before
 1968; therefore, decedents born after 1967 were excluded from the study.  Subjects were
 excluded if the  usual residence of the mother, as noted on the birth certificate, was not Ohio.
 After exclusions and  matching, there were 491  cases available for study (282 males and 209
 females).
    Controls were randomly selected from State listings of birth certificates for the period 1940
 to 1967. Each case was paired to a control of the same sex, race, and year of birth.  The same
 exclusion criteria were applied to the controls plus twins, siblings, or cases of stillbirths were
 excluded as controls.
    Fathers were classified with respect to both industry of employment and occupation.
 Industry coding followed the method described by Hoar et al. (1980) who used the Standard
 Industrial Classification Manual (1972) to establish a group of low exposure industries, i.e., low
 exposure to toxic substances.  For this study, the fathers classified as working in these low
 exposure industries were taken as the referent (or control) group in odds ratio analyses for
 industry. Fathers' occupation was coded according to the U.S. Department of Labor's
 Dictionary of Occupational Titles (1977). For analyses of occupation, the low exposure or
 referent group was considered by the authors to be a collection of occupations with
 presumably little or no exposure to toxic substances, i.e., professional, technical, or
 managerial occupations; clerical and sales occupations; and packaging and materials
 handling occupations. For other analyses, further refinements to the referent group were
achieved by the combination of low  occupational exposure and low industrial exposure into
industry-specific occupational categories.
                                         3-35
                                                                              10/18/90

-------
                            DRAFT-DO NOT QUOTE OR CITE

   Odds ratios were computed using a multiple linear-logistic regression model.  Odds ratios
were adjusted by adding various potential risk factors to the model, i.e., mother's and father's
age, birth weight, birth order, and proportion of farmland in mother's county of residence at the
time of the subject's birth (a surrogate for urban/rural residence). The authors noted that
matching could not be maintained when applying paternal low exposure reference
categorizations, and so sex and year of birth, the matching factors, were added to the model.
The cases and controls were not found to differ on the various birth characteristics examined.
   The risk of brain cancer was significantly elevated in 5 of the 12 industry classes examined,
i.e., agriculture, forestry, fishing (OR=2.4, 01=1.2-4.9); construction (OR=2.3,CI=1.3-4.1);
metal (OR=1.8, 01=1.1-2.9); machinery (OR=1.7, 01=1.1-2.7); and transportation (ship
building, motor vehicles, aircraft, and other transport methods) (OR=1.6, 01=1.0-2.40). The
latter result is similar to the result of Preston-Martin et al. (1982), discussed later.
    Examination of seven major occupational groupings only yielded significantly increased
risk for structural work (OR=2.1, Cl=1.4-3.1). The odds ratios for six subgroups within the
structural work class were all greater than 1.0, but only construction (OR=2.0, Cl=1.0-3.8) and
electrical assembling, installing, and repairing occupations (OR=2.7, 01=1.2-6.1) were
significantly elevated.  Risks were significantly elevated for paternal employment in  agriculture
(OR=2.0, 01=1.0-4.1).  The odds ratio for welders, cutters, and related occupations was
greater than  1.0 (OR=2.7, 01=0.9-8.1). Odds ratios were increased,  but not significantly, for
other major occupational groups and subgroups. Subgroups across several major groupings
seemed to involve work with or around metal and so were evaluated as a combined group. A
significantly increased risk for paternal work with metals resulted (OR=1.6, Cl=1.1-2.3).
    As previously mentioned, analyses were also made of industry- and occupation-specific
combinations, using subjects with fathers employed in low exposure industries with low
exposure occupations as the referent group. Unknown occupations (15 cases, 10 controls)
were included in the reference group [as were occupations that were known but could not be
classified (6 cases, 3 controls)].  This inclusion could represent a possible source of
 misclassification bias which, if present, could underestimate the risk for an affected subgroup.
    It is not clear why the authors designed the study the way they did.  With the individually
 matched cases and referents it seems they could have performed a conditional logistic
 regression analysis instead of breaking the matching to do a multiple logistic progression
 analysis. It is also not clear who was included in the industry- and occupation-specific
        10/18/90
3-36

-------
                            DRAFT-DO NOT QUOTE OR CITE

 analyses. Was it that particular industry and the "low exposure" group or did they include a lot
 of dummy variables?
    Significantly increased risks were seen for agriculture (OR=2.4, 01=1.1-5.4); construction
 (OR=2.0, 01=1.0-4.0); processing occupations in the metal industry (OR=5.3, 01=1.0-27.2);
 structural work in the  metal industry (OR=3.9, 01=1.2-12.8), which includes welders; and
 electrical assembling, installing, and repairing as part of structural work occupations within the
 machinery industry (OR=3.6, CI=1.3-10.0).
    The authors of this study (as well as the authors of the Spitz and Johnson [1985] study)
 concluded that paternal occupation might be a risk factor for childhood brain tumors,
 particularly in agriculture, construction, metal-related jobs, and electrical assembling, installing,
 and repairing occupations in the machinery industry. Although exposure to electricity, and
 presumably EM fields, is likely, the potential for exposure to chemicals, metals, and other
 agents is also likely. The authors specifically point out that exposure to a number of aromatic
 and aliphatic hydrocarbons (solvents), beryllium, nickel, lead, and zinc occurs to persons in
 these occupations. Limitations of this study include use of death certificates to identify cases
 and the use of birth certificates to identify father's occupation. Occupation, as given on a birth
 certificate, may be imprecise, subject to error, or not reflect the relevant causal occupational
 exposure before conception.  These difficulties may lead to random misclassification and will
 lead to a reduction of  the estimates of risk. The authors  made no attempt to identify
 categories having high potential for exposure to EM fields. They were more interested in
 identifying low exposure categories for toxic substances, in general.  The authors stated that
 this paper was exploratory. Hence, little can be gained from this paper that will  help in
 determining  whether EM fields are associated with cancer.

 3.2.3.  Summary
   There have been eight case-control studies on childhood cancer, and one case study of
 cancers in young girls concerning exposure to EM fields. Six of the studies have examined
 childhood cancer and  residential exposure to EM fields from power transmission and
 distribution sources (Wertheimer and Leeper, 1979; Fulton et al., 1980;  Myers et al., 1985;
Tomenius, 1986; Savitz et al., 1988; Lin and Lu, 1989).  AH six were of a case-control design.
Two other case-control studies examined cancer in children in relationship to fathers'
occupation (Spitz and  Johnson,  1985; Wilkins and Koutras, 1988).
                                         3-37
10/18/90

-------
                                        DRAFT-DO NOT QUOTE OR CITE

               The case series (Aldrich et al., 1984) reported on a cluster of endodermal sinus tumors in
            girls in Jacksonville, Florida, living near power lines.  By the nature of their design, case series
            usually do not provide the type of information needed to determine causality.
               Wertheimer and Leeper (1979) studied cancer mortality among children in the Denver area
            from 1950 through 1973. Control children were drawn from birth certificate files and were
            selected for similarity in birth month and county where born. For children who had only one
            address throughout their lifetime, a significantly greater proportion of cases (all cancer sites
            combined) had lived in homes that were classified as HGC. A significant risk (OR=3.09) of
            cancer was found for those children who had lived at only one residence classified as HCC.
            Unfortunately, analysis by cancer site for "stable"  residences was not provided. Additionally,
            both birth residence and death residence, considered separately, were also significantly
            elevated for total cancers although the magnitude of the risk was somewhat lower. Excess
            risks were also observed for leukemia, lymphoma, and nervous system tumors. Estimated
            risks ranged from 2.08 to 2.98.  Although some potentially biasing factors and  confounders
            may be present, the study reports significant risks of certain site-specific cancers as well as
            total cancer.
               One potentially limiting factor in this study is that wiring configuration codes may not be an
            accurate estimator for magnetic fields. A second issue is that coding was not conducted in a
            blinded fashion.  Coding physical structures such as wires may be less subject to
            misinterpretation and bias than other sources of exposure information. But, the potential for
            misclassification bias cannot be fully resolved without actual field measurements prior to
            diagnosis.
               In  the study by Fulton et al. (1980), Rhode Island residences of children with leukemia were
            compared to residences of controls. This study provided no significant findings. However,
            methodological flaws limit its usefulness. A major difficulty is the comparison of multiple case
            dwellings to control dwellings.  The multiple case dwellings probably included  homes with  little
            or no  exposure to electric and magnetic fields.  This, in effect, represents exposure
            misclassification and could, thereby, introduce a dilution of response. Furthermore, the four
            subjective categories representing levels of exposure are not defined well. This paper cannot
            be used to support or refute a carcinogenic effect.
               The study by Myers et al. (1985) compared children with cancer diagnosed between
            1970-1979 and born in Yorkshire, England, to controls with a similar birth date  and nearby
_
                   10/18/90
3-38

-------
                            DRAFT-DO NOT QUOTE OR CITE

 address.  Elevated nonsignificant risks were obtained in many of the analyses. The highest of
 these was 2.4 (p=0.05) for leukemia evaluated for residences at 50-74 meters in comparison
 to residences at or beyond 100 meters. Thus, this study showed slightly increased risks for
 lymphomas, leukemias (combined), solid tumors, and cancer (all sites) in children residing
 within 100 meters of overhead power lines. This suggests an association is present with
 closeness to overhead power lines.  This study is flawed in that the comparison children were
 those who lived beyond 100 meters from such lines on the assumption (erroneously) that they
 would not be exposed to magnetic fields. These reference children may have magnetic field
 exposures from other sources. This dichotomy also produced a second flaw, and that is the
 number of children living within 100 meters of lines is so small as to produce results with little
 chance to show a significant association. Only residence at birth was used as the marker for
 actual exposure. There is no information on duration of residency at the birth address. The
 observation of many elevated risk ratios without statistical significance raises issues that may
 be important in evaluating this study. For example, the small number of cases and controls
 and the low estimated magnetic field levels make it difficult to discern effects, if they exist. In
 addition, the authors stated that cases were underascertained by about 15%.  This study is
 limited in terms of its contributions toward resolving the  question of the potential
 carcinogenicity of EM  fields.
   Tomenius (1986) examined cancer among children (0-18 years  of age) diagnosed in the
 county of  Stockholm, Sweden, during 1958-1973. Again, like Fulton et al. (1980), residences
 of the cases were compared  with the dwellings of control children matched for age, sex, and
 church district of birth. Proximity to "visible" electrical structures was evaluated, and magnetic
fields were measured, primarily at entryways.  For all cancer sites considered together,
significantly elevated risk ratios were reported if dwellings were within 150 meters of 200 kV
lines (OR=2.1) or were near any visible electrical  structure (OR=1.3). When dwellings were
examined  with respect to measured magnetic fields alone, without consideration of whether
visible electrical structures were visible, excess risks for all cancers  (OR=2.1) were also seen
when magnetic field levels were >0.3/^T (or >3 mG). Similarly, risks were elevated for
lymphomas (OR=1.8), nervous system cancers (OR=3.7, statistically significant), and total
malignancies (OR=1.8).  However, within groups  of dwellings with nearby visible electrical
structures, especially 200 kV wires, excess risks were associated with magnetic field levels of
<0.3/fT (or <3.0 mG). This result is inconsistent with the overall results for residing near 200
                                         3-39
10/18/90

-------
                           DRAFT-DO NOT QUOTE OR CITE

kV wires, made without regard to measurement data. But it may be an artifact in that proximity
to visible 200 kV lines may not be a valid indicator of exposure without knowledge of how long
the case or control lived in the home prior to the diagnosis. A minimum residency requirement
may have been needed as well as measurements within the home. Another major problem
with this study is the use of (multiple) dwellings, rather than individuals, to form the basis of
comparison. But, the available data permit evaluation of the risk of cancer for children with
only one address with measured magnetic fields of >0.3//T and, in this situation, the risk of
cancer is significant (OR=5.4). Likewise, in one of the few analyses of "persons" where only
residency at birth is considered, the risk of cancer in children who are exposed to >0.3/*T, is
significant (OR=2.6).  In the same children, if residency at diagnosis is considered, the risk is
still significant (OR=2.4).  The fact that the risk of leukemia in children subject to measured
fields (>0.3/iT) is considerably below expected (OR=0.3) should be regarded as a suspect
finding (Table 3-6).
   In general, this paper support the finds of Wertheimer and Leeper (1979) despite the flaws
in its design and analysis. The nonpositive findings on leukemia should be considered
suspect without a review of study methods and protocols to determine why a significant deficit
should occur.  The small number of leukemia cases, in general, may reflect some problem with
case ascertainment or the cancer registry.  One study that evaluated Sweden's cancer registry
found underascertainment for leukemia to be about 18% (Mattsson and Wallgren, 1984).
   Savitz et al. (1988) conducted a case-control study to assess the relationship between
residential exposure to EM fields and childhood cancer.  Measured magnetic field values
grouped into two categories ( <2.0 mG and >2.0 mG) yielded moderate nonsignificant
increases in odds ratios  at most sites. However, the lack of a clear-cut significant association
of actual measured magnetic fields with cancer is probably a reflection of two problems with
the data. First, there was a poor response for obtaining magnetic field measurements (36% of
case residences, versus 74.5% of control residences). The authors suggested the possibility
that bias was introduced and that bias may have been reduced by an "improved response for
magnetic fields," thereby resulting in higher odds ratios.  Second, the measurements were
short-term and were taken many years after diagnosis and may not adequately reflect
long-term or past magnetic field levels prior to diagnosis.
       10/18/90
3-40

-------
                            DRAFT-DO NOT QUOTE OR CITE

    Significant associations were seen when evaluating the distribution of cases and controls
 with respect to five wiring code categories. Wiring codes were examined at the time of case
 diagnosis and, where possible, 2 years prior to diagnosis. The risk of cancer increased
 significantly with increased exposure, denoted by wiring codes. The test for trend with respect
 to wiring strata was statistically significant which indicates a dose-response relationship both
 at and 2 years prior to diagnosis. The risk of cancer in children who had lived in homes with
 the greatest exposure to magnetic fields coded as "very high" 2 years prior to diagnosis was
 significantly elevated (OR=5.22, 95% Cl =1.18-23.09). Regrouping wire codes into only two
 strata yielded more precise estimates and permitted evaluation by cancer site. Excess risks of
 leukemia, brain cancer, and other sites, were seen when the distribution of cases and controls
 were compared with respect to low versus high wire code. In these analyses, the 95%
 confidence intervals for the odds ratios for total cancers (OR=1.53) and brain cancer
 (OR=2.04) exceeded unity.
    One concern with the study's methods is that controls had to have lived at the same
 residence at the time of diagnosis of the corresponding case (to ensure comparability in time)
 as they did at the time of study selection.  Controls, thus, were more residentially stable than
 cases.  Families with cases of childhood cancer were not required to still live in the area. The
 direction of the potential bias introduced by this particular requirement is not known according
 to the authors. This lack of residential stability in the Savitz et al. study led to the limited
 response rate for field measurements.  This is a more serious problem. Missing data is
 intuitively disturbing. Analyses in the study that attempted to assess this issue suggest that
 improved response for measured magnetic fields would tend to increase the odds ratios to
 make them more representative of actual risks.  The risks, as given, are probably
 underestimated.  Several possible confounders were evaluated and found not to affect the
 estimates. This paper provides evidence that exposure to EM fields, as representated by wire
 codes, is positively associated with  certain site-specific cancers, most notably leukemia,
 lymphoma, and brain cancer.
    Lin and Lu (1989) have presented preliminary and sketchy data in a brief discussion of
 childhood cancer in Taiwan in relation to residence near power lines, transformers, or
substations. Elevated but nonsignificant risks were found for all cancers, leukemias,
 lymphomas, and brain tumors. Statistical power was low. The authors provide no discussion
of the findings.
                                         3-41
10/18/90

-------
                            DRAFT-DO NOT QUOTE OR CITE

   Spitz and Johnson (1985) examined neuroblastoma deaths in children under age 15 in
Texas between 1964 and 1978 in relation to paternal occupation, as recorded on birth
certificates.  Significantly increased risks were seen in a cluster of broad industrial groups,
grouped on the basis of presumed shared exposures to aromatic and aliphatic hydrocarbons,
which included electric and electronic workers and utility workers. The occupations in the
cluster were reclassified according to whether fathers were electricians, electric and electronic
workers, linemen, utility workers, and welders or whether fathers were electric equipment
salesmen or repairmen or others with some possible exposure to EM fields. These two groups
had presumed higher versus lower EM-fields exposure, respectively.  Elevated risks were seen
in both groups, the latter being significant. Interestingly, for one specific occupation, the
children of electronic workers had an excess risk (OR=11.75) that was statistically significant.
Because of major uncertainties in information derived from birth and death certificates, i.e.,
lack of information on occupation of the father before conception as well as no information
regarding material exposure, this paper suggests further in-depth research be done.
   Wilkins and Koutras (1988) examined brain cancer in Ohio children in relation to paternal
occupation, as recorded on birth certificates.  A collection of occupations with presumed
minimal exposure to toxic agents was used as the reference group.  Risks were significantly
elevated in children  of fathers employed in several broad industrial classes and occupational
groupings, one of which was electrical assembling, installing, and repairing occupations
(OR=2.7, Cl=1.2-6.1). This paper is more exploratory than supportive since the authors were
not specifically interested in any particular exposures, including EM fields. Several areas of
pursuit were suggested for future studies, including EM fields. Both of these studies use the
somewhat unreliable information  (on father's occupation) listed on birth certificates.
    In conclusion, nine case-control studies have examined childhood cancer. Two studies
involved cancer in children vis-a-vis their fathers' occupation. Seven studies examined
residential exposure to 60-Hz electric and magnetic fields, and six of these yielded elevated
risk estimates for leukemia, lymphoma, brain cancer, and total cancer. Two studies in
particular found significant risks of lymphoma,  leukemia, and nervous system cancer in
children exposed to magnetic fields estimated  by wiring configurations.  The estimates of risk
are modest, on the order of a 1.5 to threefold increase. These two studies by Wertheimer and
Leeper (1979)  and Savitz et al. (1988) present the fewest difficulties with respect to issues of
bias, confounding, or other methodological problems.  In the Savitz et al. study, wiring codes,
       10/18/90
3-42

-------
                            DRAFT-DO NOT QUOTE OR CITE
 the surrogate of exposure used, provided the best evidence of a dose-response relationship
 with respect to risk of childhood cancer. The greatest risk of childhood cancer occurred
 among children whose addresses prior to and at the time of diagnosis were determined to be
 from magnetic fields associated with the highest wire code.  Magnetic field measurement data,
 although of limited availability, were in the range of 2 mG to 3 mG in this highest wiring code
 category.
    The study by Tomenius (1986) also found significantly high risks of lymphoma and nervous
 system cancer in children exposed to similar magnetic field intensities (i.e., >3.0 mG) or
 >0.3^T. However, he found an exceptionally low risk of leukemia (OR=0.3) in children
 exposed to the same intensities. This rather low estimate for leukemia suggests an anomaly in
 the data, possibly in the selection of leukemia cases into the Swedish Cancer Registry during
 the period of study (1958 to 1973), and selection bias may be a possibility. There is some
 evidence that leukemia cases may be underreported to the Swedish Cancer Registry. In any
 event, before this finding can be accepted on face value, it should be reevaluated.
    The study by Meyer et al. (1985) reported a nonsignificant excess risk of leukemia/
 lymphoma, solid tumors, and total cancer in children located within 100 meters of 50-Hz power
 lines.  But this paper has methodological flaws that tend to reduce the estimated relative risk
 ratios toward the null, thereby obscuring any true differences  between groups.
    Only the Fulton et al. (1 980) study found no unusual distribution of exposure between his
 case dwellings versus his control dwellings based  upon wiring codes that were similar to, but
 not identical to, those of Wertheimer and Leeper.  However, this study has major flaws that
 preclude its usefulness in assessing the carcinogenicity of EM fields. These problems were
 discussed in detail in the individual study reviews.
    The Lin and Lu (1989) study, although demonstrating excess but nonsignificant elevated
 risks in Taiwanese children living within 50 meters  of high tension power lines (20 kV or
 greater), is very sketchy and contains few details involving the study methodology.
    Finally,  the two remaining studies suggest fathers' occupational exposure to EM fields may
 be a reason for the finding of a significantly high risk of neuroblastoma and brain cancer in
 children.  Although supportive, these two studies suggest future research in this area.
 Children with these cancers tended to have fathers whose occupations, as reported on the
 birth certificates, fell  into certain broad industrial categories thought to include exposure to EM
fields.  Actually, occupation before conception would have been more meaningful. These
                                         3-43
10/18/90

-------
                           DRAFT-DO NOT QUOTE OR CITE

categories also included potential exposure to aromatic and aliphatic hydrocarbons as well as
other toxins. All the residential studies, except that by Fulton et al. (1980), provide evidence of
a consistently positive but modest risk of leukemia, lymphoma, and nervous system cancer in
children under age 20 from exposure to power frequency magnetic fields estimated by wiring
codes.
   Measurement data that were available suggested that exposures were in excess of 2 mG.
Little is known about the factors important in the development of cancer in children. Bias and
confounding cannot be completely ruled out in these studies. Most biases in these studies
would tend to result in random exposure misclassification which tends to bias results toward
the null, resulting in underestimation. However, confounding or exposure to other
carcinogenic agents as an explanation is probably less valid in studies of children than in
studies of adults, because children experience a less diverse exposure environment over a
shorter period of time.  Savitz et al. (1988) evaluated possible confounding factors and found
none to explain the positive results.

3.3.  STUDIES OF ADULTS
3.3.1.  Radiofrequency Exposure
   The following two reports have been previously reviewed (U.S. EPA, 1984) but are
repeated here for completeness.
   Lilienfeld et al. (1978) completed a broad survey of the mortality and morbidity experience
of Foreign Service employees and their dependents to assess the potential health
consequences of microwave irradiation of the U.S. Embassy in Moscow. The health status of
Foreign Service employees and those from other agencies who had served in the U.S.
Embassy in Moscow from 1943 to 1976 was compared with that of employees at eight other
embassies or consulates in Eastern  Europe over the same time period.
   The microwave irradiation of the  U.S. Embassy in Moscow was first detected in 1953 and
subsequently varied in intensity, direction, and frequency over time.  The frequencies ranged
from 0.6 to 9.5 gigahertz (GHz) (Pollack, 1979; U.S. Senate, 1979). The measured average
power densities over time are given  in Table 3-15.
   Extensive efforts were launched to identify and trace the populations.  Information on
illnesses, conditions, or symptoms were sought from two major sources:  (1) employment
       10/18/90
3-44

-------
                           DRAFT-DO NOT QUOTE OR CITE
        TABLE 3-15. MICROWAVE EXPOSURE LEVELS AT THE U.S. EMBASSY IN MOSCOW
   Time Period
  Exposed
Area of Chancery
Power Density and
Exposure Duration
    1953 to May 1977
   June 1975 to Feb. 1976
   Since Feb. 7,1976
West Facade
South and East Facade
South and East Facade
Maximum of 5 jŤW/cm':
9 hrs/day
18 hrs/day
Fractions
1 8 hrs/day
SOURCE: Lilienfeld et al., 1978.
medical records, which were fairly extensive because of examination requirements for foreign
duty, and (2) a self-administered health history questionnaire.  Questionnaire responses were
validated for a stratified sample by review of hospital, physician, and clinic records.  Death
certificates were also sought, although other sources also were used to ascertain mortality
status.
   Standardized mortality ratios for various subgroups were calculated for each cause of
death, were standardized for age and calendar period, and were specific for sex.  Similar
procedures were used to develop summary indices of morbidity.
   A total of 4388 employees and 8283 dependents were studied. More than 1800 persons
with 3000 dependents were employed at the U.S. Embassy in Moscow, and 2500 persons with
more than 5000 dependents were employed at the comparison posts.  Ninety-five percent of
the employees were traced.  Receipt of completed questionnaires was less successful, with an
overall response rate of 52% for State Department personnel.
   Based on information in medical records, various health problems were generally similar,
with two exceptions. Moscow employees had a threefold greater risk of acquiring protozoal
infections than comparison post employees.  In general,  both sexes in the Moscow group had
somewhat higher frequencies of most of the common kinds of health conditions reported.
Lilienfeld et al. (1978) stated, "However, these most common conditions represented a very
heterogeneous collection, and it is difficult to conclude that they could have been related to
                                        3-45
                                         10/18/90

-------
                          DRAFT-DO NOT QUOTE OR CITE
exposure to microwave radiation since no consistent pattern of increased frequency in the
exposed group could be found."
   Some excesses were reported by Moscow employees in the health history questionnaire.
Both sexes reported more eye problems due to correctable refractive errors. More psoriasis
was reported by men and anemia by women. The Moscow employees, especially males,
reported more symptoms such as irritability, depression, difficulties in concentration, and loss
of memory.  It is possible, however, that a bias due to awareness of potential adverse effects is
operating, since the strongest differences were present in the subgroup with the least
exposure.
   The observed mortality was less in both male and female employees than expected, based
on U.S. mortality rates (Table 3-16). The male employees had lower mortality than did female
employees.  Cancer was the predominant cause of death in both sexes.  The risk of leukemia

    TABLE 3-16. OBSERVED AND EXPECTED NUMBER OF DEATHS, STANDARDIZED MORTALITY
   RATIOS (SMR), AND 95% CONFIDENCE INTERVALS (Cl) BY ALL CAUSES OF DEATH  SPECIFIED
   CAUSES OF DEATH FROM CANCER, AND POST FOR MALE AND FEMALE STATE AND NONSTATE
                         DEPARTMENT EMPLOYEES COMBINED
Moscow
Number of Deaths
Cause of Death
All causes (0.5, 0.7)
Malignant neoplasms
Digestive organs
Brain tumors/CNS
neoplasms
Pancreas
Lung
Leukemia
Hodgkin's disease
Breast
Uterus
Cervix
Observed
49
17
3
0

1
5
2
0
2
1
1
Expected
105.3
19.0
4.6
0.9

1.0
5.8
0.8
0.5
0.5
0.2
0.1
SMR
(95% Cl)
0.47 (0.4-0.6)
0.89(0.5-1.4)
0.65 (0.4-1 .9)
0.0

1.0(0.0-5.6)
0.86 (0.3-2.0)
2.5 (0.3-9.0)
0.0
4.0 (0.5-14.4)
5.0(0.1-27.9)
10.0 (0.3-55.7)
Other Posts
Number of Deaths
Observed
132
47
11
5

1
11
3
0
3
0
0
Expected
223.7
41.1
10.8
1.5

2.2
12.2
1.7
0.7
1.2
0.0
0.0
SMR
(95% Cl)
0.59
1.1 (0.8-1.5)
1.0(0.5-1.8)
3.3(1.1-7.7)

0.5 (0.0-2.5)
0.9(0.4-1.6)
1.8(0.4-5.3)
0.0
2.4 (0.5-7.0)
0.0
0.0
      10/18/90
3-46

-------
                           DRAFT-DO NOT QUOTE OR CITE

was elevated both at Moscow [standardized mortality ratio (SMR=2.5)] and at comparison
posts (SMR=1.8). Neither SMR was statistically significant. Comparison post employees had
a statistically significant excess risk (SMR=3.3) of nervous system tumors. In general, the
Moscow and comparison groups did not differ appreciably in overall and specific mortality.
However, the population was relatively young; it may have been too early to detect long-term
mortality effects.
   The authors concluded that no convincing evidence was discovered to implicate
microwaves in the development of adverse health effects at the time of the analysis. But they
also carefully discussed the limitations inherent in the study:  uncertainties associated with the
reconstruction of the employee populations and dependents, difficulties in obtaining death
certificates, the low response rate for the questionnaire, and the statistical power of the study.
An important limitation relates to ascertainment of exposure.  Problems relative to individual
mobility within the embassy and variation of field intensities within the building are present in
this study as in any other. No records were available on where employees lived or worked, so
one had to rely on questionnaire responses to estimate an individual's potential for exposure.
The highest exposure level [18 microwatts per square centimeter (aW/cm2)] was recorded for
only 6 months in  1975-1976; thus, the group exposed to the most intense fields had the
shortest cumulative time of exposure and of observation in the study.  These intensities are
considered to be very low.  It is also not clear what exposures may have been experienced by
employees at the comparison posts.
   Robinette and Silverman (1977) and Robinette et al. (1980) examined mortality and
morbidity among U.S. naval personnel occupationally exposed  to radar. Records of service
technical schools were used to select subjects for the study; the men graduated from technical
schools during the period from 1950 through 1954.  Exposure categorizations were made on
the basis of occupational specialty. The exposure group (probably highly exposed) consisted
of technicians involved in repair and maintenance of radar equipment. The "controls"
(probably minimally exposed) were involved in the operation of radar or radio equipment. It
was estimated from shipboard monitoring that radiomen and radar operators (in the
low-exposure group) were generally exposed at less than 1 milliwatt per square centimeter
(mW/cm2), and gunfire control and electronics technicians (in the high-exposure group) were
exposed to higher levels during their duties. Over 40,000 veterans were included in the study,
with about equal numbers in these two major exposure classifications.  The mean age in 1952
                                         3-47
10/18/90

-------
                            DRAFT-DO NOT QUOTE OR CITE

 of the low-exposure group was 20.7 years and of the high-exposure group, 22.1 years.  In
 conjunction with naval personnel, an effort was also made to develop an index of potential
 exposure, termed Hazard Number, for a limited portion of the population. This number was
 based on the duty months multiplied by the sum of the power ratings (equipment output
 power of gunfire-control radars (ship) or search radars (aircraft)) where technicians were
 assigned.
    Medical information was obtained through Navy and Veterans Administration records.
 Records were searched for information on four major end points: (1) mortality, (2) morbidity via
 in-service hospitalizations, (3) morbidity via Veterans Administration (VA) hospitalizations, and
 (4) disability compensation. Mortality was ascertained through the VA beneficiary system.
 Mortality ratios were calculated for both the low and the high group, standardized for year of
 birth and using the combined experience of both groups as the standard population.
    For the low-exposure groups, mortality ratios were only slightly elevated for diseases of the
 circulatory system (1.07); the cancer residual, other malignant neoplasms (1.19); and the total
 residual, other diseases (1.08).  Cancers of the digestive tract (1.14), respiratory system (1.14),
 and lymphatic and hematopoietic systems (1.19) were elevated for the high-exposure group,
 but none of the increases was statistically significant.  The differences in mortality from
 malignant neoplasms of the lymphatic and hematopoietic system, although elevated, were not
 statistically significant.
    As seen in Table 3-17, comparisons were also made within the high-exposure group
 across Hazard Number categories.  In this case, only two comparisons were statistically
 significant:  (1) the difference in respiratory tract cancer between those with a Hazard Number
 smaller than 5000 versus larger than 5000, and (2) the test for trend for all diseases combined.
These results may be fortuitous since one or two positive findings might be expected when
 many statistical comparisons are made. Furthermore, additional information  relative to the
 development of lung cancer, e.g., smoking histories, could not be obtained; the mortality data
were obtained from death certificates, and obtaining background information from next-of-kin
was not feasible. Among men whose work received the high hazard rating, elevated but
nonsignificant risks were also seen for all cancers comined (SMR=1.44), cancer of the
lymphatic and hematopoietic system (SMR=1.64), and the residual category of miscellaneous
cancers (SMR=1.17) as well as circulatory diseases (SMR=1.17).
       10/18/90
3-48

-------
                             DRAFT-DO NOT QUOTE OR CITE
  TABLE 3-17. NUMBER OF DEATHS FROM DISEASE AND MORTALITY RATIOS8 BY HAZARD NUMBER: U.S. ENLISTED NAVAL
              PERSONNEL EXPOSED TO MICROWAVE RADIATION DURING THE KOREAN WAR PERIOD
International
Classification
of Diseases
Cause of Death (8th Rev.)
All diseases 000-796

Malignant neoplasms 140-209

Digestive organs 150-159

Respiratory tract 160-163

Lymphatic and 200-209
hematopoietio
system
Other malignant Residue
neoplasms
Diseases of 390-458
circulatory system
Other diseases Residue

Number of Deaths
Hiah Exoosure
Low
Exposure
325
(1.04)
87
(0.96)
14
(0.85)
16
(0.85)
29
(0.83)

37
(1.19)
167
(1.07)
71
(1.08)
Total
309
(0.96)
96
(1.04)
20
(1.14)
24
(1.14)
26
(1.18)

26
(0.82)
150
(0.93)
63
(0.92)

0
63
(0.82)
22
(0.99)
6
(1.49)
4
(0.82)
6
(1.09)

6
(0.78)
36
(0.94)
5
(0.30)
Hazard Number
1-5000
160
(0.91)
45
(0.90)
11
(1.14)
10
(0.86)
12
(1-04)

12
(0.70)
73
(0.83)
42
(1.13)
5000
86
(1.23)
29
(1.44)
3
(0.78)
10
(2.20)
8
(1.64)

8
(1.17)
41
(1.17)
16 .
(1.08)
taken aL\rets°anndardrentheSeS) standardized for Vear of birtn:tne combined experience of the low and high exposure groups is
SOURCE: Robinette et al., 1980.
    Differential health risks with respect to hospitalized illness around the period of exposure
were not apparent. Subsequent VA hospitalizations and disability awards provided incomplete
information.  Because the study focused largely on the use of automated VA record systems, it
was not possible to determine non-Navy or non-VA hospitalizatfons, nonhospitalized
conditions, reproductive histories, or subsequent employment histories. Since actual
individual exposure could not be  reconstructed retrospectively, only an estimate of the
potential exposure of the individuals was possible.  Longer follow-up of the population would
be useful.
                                           3-49
10/18/90

-------
                           DRAFT-DO NOT QUOTE OR CITE

   Milham (1985b) investigated mortality among amateur radio operators who were members
of the American Radio Relay League and whose deaths were reported in the League's
magazine between 1971 and 1983.  Proportional Mortality Ratio (PMR) analyses were
performed.  Death certificates were obtained for 280 deaths in Washington State and for 1411
deaths in California.  Expected values were generated using 1976 U.S. age-specific white male
death frequencies. It was reported that PMRs were significantly elevated for all leukemias, and
acute and chronic myeloid leukemia considered separately and together. The author also
noted a strong association between League membership and an occupation with potential
exposure to EM fields.  For cases from the state of Washington, the PMR for amateur radio
operators who were also in electrical-exposure occupations was 2.64. The PMR for operators
whose  obituary did not mention such a job was 2.10. There may be biases in this study due to
ascertainment of the deaths from the League's magazine. It is not clear if all deaths of
members would be so reported. Deaths of ex-members may not be repprtable, and the
experience of ex-members may be different than that of continuing members. On the other
hand, this study illustrates use of an innovative and accessible source of survival information
useful for an exploratory study.
   Milham (1988) next investigated the mortality of 67,829 men licensed as amateur radio
operators between January 1,1979, and December 31,1984, with the Federal
Communications Commission and who resided in California and Washington State. Deaths
were sought in California and Washington State. SMRs were calculated  using U.S. death rates
to estimate expected deaths.
   Observed deaths (Table 3-18) were significantly lower than expected for all-cause mortality
and for mortality from all malignant neoplasms combined, pancreatic cancer, cancer of the
respiratory system, all circulatory diseases combined, all respiratory diseases combined, and
all accidents.  SMRs were elevated for several cancer sites, and statistically significant
excesses were found for specific sites in lymphatic and hematopoietic tissues, namely, acute
myeloid leukemias (SMR=1.76) and multiple myelomas and other neoplasms of the lymphoid
tissues, considered together (SMR=1.62). The latter rubric is the category for lymphomas
other than iymphosarcomas and reticulum-cell sarcomas and Hodgkin's disease.  It was not
stated why this heterogeneous group of lymphomas was considered with multiple myelomas;
separate analyses were not presented. It would seem to be more appropriate to evaluate
deaths from multiple myeloma separately. However,  Milham  noted that the observed deficit for
       10/18/90
3-50

-------
                         DRAFT-DO NOT QUOTE OR CITE
       TABLE 3-18. MORTALITY IN WASHINGTON STATE AND CALIFORNIA: U.S. FEDERAL
   COMMUNICATIONS COMMISSION (FCC) AMATEUR RADIO OPERATOR LICENSEES, JANUARY 1,
                            1979, TO DECEMBER 31, 1984

All causes
All malignant neoplasms
Esophagus
Stomach
Large intestine
Rectum
Liver
Pancreas
Respiratory system
Prostate
Urinary bladder
Kidney
Brain
Lymphatic and hematopoietic tissue
Lymphosarcoma/reticulosarcoma
Hodgkin's disease
Leukemia
Lymphatic
Acute
Chronic
Unspecified
Myeloid
Acute
Chronic
Unspecified
Monocytic
Unspecified
Acute
Unspecified
Other lymphatic tissues
All circulatory diseases
All respiratory diseases
All accidents
Observed
2485
741
22
30
88
14
11
27
209
78
16
19
29
89
5
5
36
9
3
6
0
18
15
3
0
0
9
6
3
43
1208
127
105
Expected
3478.9
836.9
19.4
29.6
79.0
18.2
16.8
41.9
315.6
67.6
24.1
20.1
20.8
72.1
10.6
4.1
29.0
8.7
2.5
5.5
0.8
12.9
8.5
3.5
0.9
0.6
6.7
3.4
2.5
26.6
1731.7
252.5
164.5
SMRa
71*
89*
113
102
111
77
65
64*
66
114
66
94
139
123
47
123
124
103
120
109
0
140
176*
86
0
0
134
176
120
162*
70*
50*
64*
aSMR=standardized mortality ratio
*p<0.05

SOURCE: Milham, 1988.
                                     3-51
10/18/90

-------
                           DRAFT-DO NOT QUOTE OR CITE

lymphosarcomas and reticulosarcomas (ICD8: 200) (SMR=0.47) nearly cancelled out the
excess observed for deaths from other neoplasms of the lymphoid tissue (ICD8: 202.0). The
risk of Hodgkin's disease (SMR=1.23) was also elevated but not significantly.  For the
leukemias, the SMRs were elevated at all sites for which deaths were observed, but, as
mentioned, only deaths from acute myeloid leukemias were found to be significantly in excess.
   This is a study with a large population (67,829 licensees; 232,499 accumulated
person-years; 2485 deaths). Excess risks are seen at several cancer sites but are especially
concentrated at tissues of the lymphatic and hematopoietic system, where certain excess risks
were found. The risk of acute myeloid leukemia was significantly elevated.  Leukemias
generally predominate in younger ages.  Chronic leukemias were low.
   Since licensing is required for amateur radio operators, enumeration of the population
should be reasonably complete although licensing per se does not provide information on
usage and exposure. Milham cites survey data that found that amateurs practice their hobby
about 6 hours per week.  It would seem that licensees would have exposure but the extent and
degree is not clear and is probably variable. However, the  potential for exposure
misclassification would tend to bias estimates towards the null.
   These results may or may not bear on risks from other frequencies, either of RF or ELF
radiation, or even on the operating frequency.  Amateur radios operate at the low end of the
RF band of the electromagnetic spectrum.  Modulations to  lower frequencies are known to
occur (Personal conversation with E. Mantiply, Office of Radiation Programs, U.S. EPA,
February, 1990).  Among Washington State licensees, Milham found that about 31% had jobs
that involved potential EM-field exposure. Exposure to many different frequencies among part
of the population is, thus, a possibility. Confounding exposures to chemicals are possible but
could not be evaluated in this study.
   This study points to excess risks for various cancers of the lymphatic and hematopoietic
system, especially acute myeloid leukemia, among amateur radio operators in Washington
State and California, some of whom also may have experienced occupational exposure to EM
fields. Further studies  of this source are warranted.
   Milham (1988) extended the above analysis by examining the mortality in the amateur radio
operators according to their Federal Communications Commission (FCC) license class.
Depending on one's level of expertise and experience, an individual will be granted one of five
specific licenses, i.e., novice, technician, general, advanced, or extra. These license classes
       10/18/90
3-52

-------
                            DRAFT-DO NOT QUOTE OR CITE

can, thus, serve to some extent as surrogates for duration of exposure. There were, as might
be expected, some differences in age by license class. The average age of persons per class
was: novice, 38.4 years; technician, 44.3 years; general, 49.5 years; advanced, 51.4 years;
and extra, 49.2 years.  It would have been interesting to know also the median ages and age
ranges by class.
    As in the previous analysis, SMRs were derived, and expected deaths were calculated by
applying age, sex, race, year of death, and cause-specific U.S. mortality rates to age-stratified
years at risk.  Results were presented for each license class for a limited number of causes of
death, i.e., all causes combined, all malignant neoplasms, brain cancer, all lymphatic and
hematopoietic neoplasms, all leukemias, myeloid leukemia, and multiple myeloma and other
lymphomas, considered together. The SMR for Hodgkin's disease was elevated in the earlier
analysis, but mortality from Hodgkin's disease by license class was not reported in this paper.
    All-cause mortality in all license classes was significantly lower than would be expected
based on general U.S. mortality. SMRs for the most advanced license class, extra, were the
lowest for all-cause mortality and for deaths from all malignant neoplasms. These results may
indicate that some survival, socioeconomic, or other sort of bias may be operative. Generally,
SMRs were lowest among the deaths of licensees in the novice class, the youngest group.
Milham states that, in a sense, the novice class provides an internal control group.  It would be
interesting to see results in any further analysis of amateur radio operators, if novice operators
are treated as a control subset of the population.
    Other than the lower SMRs in the novice class, there is no other gradient in the SMRs by
license class. In fact, SMRs are generally highest in the second level license class,
technicians. This group had a statistically significant excess of deaths from all lymphatic and
hematopoietic neoplasms, considered together.  Deaths in the general  license class resulted
in a significantly elevated SMR for multiple myeloma and other lymphomas (202-203). With the
exception of the novice class, SMRs were generally elevated, although  not significantly, for all
classes for the lymphatic and hematopoietic cancers reported.  This excess risk was also true
for deaths from brain cancer, but none of the observed increased SMRs was significantly
different than expectation.
    Since this study reanalyzes results given in an earlier report by specified strata, it, thus, is
an attempt to identify where risks are operative in a scheme that is a surrogate of duration of
exposure. The greatest risk seems to be centered in the technician license class which is the
                                         3-53
10/18/90

-------
                            DRAFT-DO NOT QUOTE OR CITE

class between novice and the more advanced license classes. This group was also
intermediate in average age. The similarity in age for the three upper classes may mask any
differences between those groups.  No clear or distinct gradient by class was observed. It
appears, therefore, that either license class may not be a good exposure surrogate (special
unknown conditions may be operative for technician class licensees or their operational
activity may differ), or class exposure has nothing to do with cancer.
    The Environmental Epidemiology Program in the State of Hawaii Department of Health
(1986) investigated cancer incidence in census tracts with and without broadcast towers in
Honolulu, Hawaii. This study was prepared for and reported to the City Council of the City and
County of Honolulu in 1986.
    Cancer incidence data was collected from the Hawaii Tumor Registry which registers all
newly diagnosed cases of cancer in the State, except cases of squamous and basal cell
carcinoma of the skin. Cases are identified from all hospitals, and private pathology
laboratories as well as from searches of death certificates. Less than 2% of cases are
identified solely from death certificates, and 94% of cases are  microscopically confirmed.
    Nine census tracts with broadcast towers and two without towers were examined.
Expected number of cases for the census tracts were calculated using age- and race-specific
rates for the State for the period 1979-1983. Standardized incidence ratios (SIRs),
representing the ratio of observed to expected values, were computed, and tested at a
significance level of p=0.01.  Confidence intervals were not presented.
    Age-adjusted rates and SIRs by sex were presented for each census tract and for tracts
categorized as having towers (N=9) and not having towers (N=2).  For all tracts with towers
combined, the summary SIR for all cancers was 1.45 for males and 1.27 for females, and both
values were significant at p=0.01.  The SIRs for all cancers for both sexes in the two tracts
without towers, taken together, did not differ from unity (male SIR=1.05, female SIR=0.85).
The SIR, for leukemias were also elevated, but statistical significance at the p=0.01  level was
not achieved. The SIRs for individual census tracts  were significantly elevated for all cancers
among males for eight of the nine tracts with towers, and not elevated for the two tracts
without towers. Among females, the SIRs for all cancers was significantly elevated in only two
census tracts, and these contained towers. Four other tracts with towers had elevated SIRs,
but the excess was not statistically significant. Age-adjusted rates and SIRs were also
examined in individual tracts for leukemias in both sexes, but the number of cases was very
       10/18/90
3-54

-------
                            DRAFT-DO NOT QUOTE OR CITE

small, making the results unstable.  Bearing this in mind, it was noted that there was a
tendency for elevated SIRs for both sexes in census tracts with towers.
    Because of its diverse ethnic populations, ethnicity and race are critical factors to consider
in health studies in Hawaii. To this end, race was another factor controlled in the study. The
authors stated that small numbers precluded simultaneous adjustment for age, race, and sex.
    After adjusting for race, the SIR for all cancers for tracts with towers was 1.88 (p <0.01).
The SIR for all cancers for the tracts without towers was 1.07 and was not significantly
elevated.  For individual tracts, the SIRs were again significantly elevated for eight of the nine
tracts with towers.  The one tract with towers that did not demonstrate an excess in all cancers
after adjustment for race was the same tract that did not show excess cancer after controlling
for age and sex. For the two tracts  without towers, the SIR for all cancers was significantly
elevated in one (SIR=1.31) but not in the other, after adjusting for race; this did not occur in
the analysis that controlled for age and sex although the one tract had consistently higher
rates and SIRs than the other. Again, analyses for leukemias yielded very small numbers of
cases in individual census tracts. No result was statistically significant, but tracts with towers,
individually and overall, presented elevated SIRs for leukemias.
    In summary, using State rates as the standard schedule of rates, the observed deaths in
tracts with towers for all cancers were significantly greater than expected values.  Tracts
without towers did not appreciably differ. Differences for leukemias could not be discerned.
    This study is of a type that can be called an ecological study or that is of an ecological
design,  meaning that relatively broad populations (here, from census tracts) are examined
rather than individuals.  It is difficult to determine causal relationships from such studies.  And
this difficulty is aggravated by typically weak measures of exposures.  In this case, broadcast
towers are sited in certain census tracts, but it is not known  if the cases of cancers were
exposed to RF radiation or, if so, at what levels. The dichotomous estimate of exposure is only
a crude proxy—some tracts have towers, others do not.  Studies of this design also cannot
usually address confounding factors except those that are broad demographic characteristics
such as age, sex, or race, readily available from vital and census records.
    The authors pointed out the limitations inherent in the study, including issues such as
personal exposure, latency, duration of exposure, confounders, and so forth. They do not
comment, however, on "urban/rural" differences that could be a factor in Oahu. The tracts with
towers largely constitute downtown Honolulu and Waikiki; the tracts without towers are more
                                         3-55
10/18/90

-------
                           DRAFT-DO NOT QUOTE OR CITE

centrally located on the island, have less dense population centers, and have some
agricultural lands and mountainous/valley areas. The study does indicate that further
investigation are warranted.
    It should be noted that this study followed an EPA (Office of Radiation Programs)
measurement study of potential public exposures near broadcast towers in populated areas in
Honolulu (U.S. EPA, 1984), where there was a concern over the health risks and ancillary
problems posed by broadcast transmitters.  Conservation and preservation concerns have led
to restrictions on the siting of commercial broadcast towers.  As a result, most tower sites
impinge on populated areas. The EPA study measured emissions close to towers and
generally found that levels of RF radiation tended to approach or exceed various voluntary
exposure guides.  The RF exposures in Honolulu may be unique given environmental
restrictions, topography, and population density patterns in relation to radiating sources.
Several military and navigational sources also exist on Oahu. The EPA data, plus the data
from this study, point out the potential  need for analytical studies of Hawaiian populations that,
if feasible, examine more analytically health indices with more detailed exposure assessment.
    Hill (1988) investigated the mortality of 1492 men in a radar research and development
project at the Massachusetts Institute of Technology (MIT) during World War II conducted in
what was known as the Radiation Laboratory (Rad Lab). The term radar is an acronym that
refers to the use of electromagnetic energy for detecting and locating reflecting objects.
Radars generally operate within the microwave portion (0.3 to 3.5 GHz) of the electromagnetic
spectrum.
    The cohort was scientific, technical, and senior management staff members ever employed
at the Rad Lab between October 1940  and January 1946. Support personnel, such as
technicians, clerks, or guards, were not included.  The population was ascertained from
records retained by MIT and was traced from World War II through 1986. Multiple sources of
follow-up information were used, especially alumni organizations and Social Security
Administration  (SSA) files. Only 4.6% of the subjects were not traced beyond the 1940s. Over
67% were traced into the 1980s, and about 80% were traced into the 1970s.  There were
52,805 person-years accumulated.
    Vital status of the cohort was determined from various sources, including state death
certificates or city/town clerk reports and SSA records of earnings reports or beneficiary
claims. Cause of death data were based on death certificates or city/town clerk reports.
       10/18/90
3-56

-------
                           DRAFT-DO NOT QUOTE OR CITE

Cause(s) were coded according to criteria developed by the National Center for Health
Statistics for coding with the International Classification of Diseases, Adapted for Use in the
United States (ICDA), 8th Revision. A trained nosologist supervised and reviewed coding.
    Comparisons were to U.S. white males and a population of white male physician
specialists. The latter were assumed to be similar in socioeconomic status.  Because of
certain characteristics of the populations, physicians were only compared to Rad Lab staff
members who were at least were 25 years old at first employment, and follow-up was
censored to end in 1974. Internal comparisons were also made.
    Two approaches were used to estimate the exposure environment at the Rad Lab. First,
summary estimates based on the parameters of two typical radar systems of 1943 vintage
were calculated.  It was estimated that the maxiumum power density in the near field of the
systems' antennas could be about 2 to 5 mW/cm2, corresponding to a specific absorption rate
(SAR) of 0.1 to 0.4 watts per kilogram (W/kg). Second, a surrogate estimate of exposure was
derived from the job patterns in Lab Divisions which were organized around the basic
equipment and systems being tested and developed. A three tier ranking system was used
(1 =little or none, 2=low or  moderate, and 3=highest) to represent a gradient of potential
exposure of the various jobs and projects in relation to each other, e.g., administrative (low)

-------
                            DRAFT-DO NOT QUOTE OR CITE

length of follow-up) and to estimate risk ratios. This is another approach to age adjustment.
The only independent variable evaluated was cohort (Rad Lab staff or physician); the cohorts
were already uniform with respect to race and sex. No other potentially relevant variables were
available on physicians to permit comparisons to Rad Lab cohort members.
   With respect to expected mortality based on U.S. rates, the overall mortality in the
population was lower than expected. This probably reflects a "healthy worker" effect and the
socioeconomic status of the population, and may be evidence that the protective benefit (or,
rather, the selection bias) of actively working is enhanced in professional occupations (here
physical scientists largely believed to be employed at universities, in government, or in high
technology industries).  Several specific causes of death were elevated, but no excess was
statistically significant. These increases occurred for certain cancers (skin, prostate, testis,
and cancers whose nature was unspecified) and mental disorders. With censoring to include
only men at least 25 years of age at entry to the study and to end follow-up in 1974, elevated
SMRs were again seen for selected cancers (pharynx, gallbladder and bile ducts, pancreas,
skin, prostate, Hodgkin's disease, and cancers whose  nature was unspecified) and  cirrhosis of
the liver. The SMR was greatest for Hodgkin's disease. No excess was statistically  significant.
These two SMR analyses also provided limited evidence of no differential risk by two different
periods of latency.  The strong healthy worker effect observed limits the value of these SMR
results, but they point to causes requiring further study.
   Mortality was also lower for Rad Lab staff members than for physicians (Table 3-19).
Rather than implicating nonionizing radiation (NIR) exposure, the study may have
demonstrated a difference in survival and all-cause mortality between two professions, that is,
physicians versus scientists, chiefly physicists and engineers. The physicians were  internists,
ophthalmologists, and otolaryngologists. For all causes combined, the staff members had a
lower rate of mortality than physicians. The risks for deaths from diseases of the circulatory
system and from external causes, e.g., accidents, were significantly lower for staff members
than for physicians. The deficit of deaths from these two causes was probably responsible for
better survival in the Rad Lab cohort.  The rates for all malignant neoplasms were the same.
The rates were also approximately the same for deaths from all other causes, considered
together.  But increased risks were seen for lymphomas, particularly Hodgkin's disease;
cancers of the nervous system; and cirrhosis of the liver. Given these excesses, the Rad  Lab
deaths may have an infectious, immunological, and/or behavioral component.
       10/18/90
3-58

-------
DRAFT-DO NOT QUOTE OR CITE
 o
111
^ UJ CO
UJ ^-O
52 cc Ł3

Oul§
il^Z
2|1
^ >•
k.sl
co < >
DC UJ W
0 ° W
UJ O 2
Q < u.
^_ . Ll
"*— l^ ^Ł —I
=) j? CO CO
xQE
l||
u_ — ' Q?

tofe^
GQ ^^ ^"*
 |—
Q Q <
< N D-
U- g 2jj
O < ^
t ^ O
~i < CO
TABLE 3-1 9. MORTAI
CANCERS. PART A. ST
BASED ON RATES IN PHY

^_^^
i!
^J .E Q.

ť |
|l |
^ ^* O"
(0 CO

o. ^^
0 g
CL CO
'o
OQ .2 'w
t 'S Ł
TO DC D.
CL ^ . .
~ CO
-D
Ł




O
OT &)
O O)
1
1 i
O CO
s
1 1
•o 8
C Q.
1 fi

<
t 1
o- 1
CO
.Q
O


CD
CO
C§

•Q
0"
g
o
d
~

p
•r-





1
d





CM
p
1
co
d

Ł5
d

T™
O
CD

CM



CO
CM



1

^^^
to
i
d


9
T—





S-
8
d





m
^i
g
d

O)
p

Co
05
CM





^
rf-



All Malignant Neoplasms


m co
in co
co in
co cO
o o


88
0 0





CO CD
CO CD
O O
1— T-







1 1


1 1

I |






t 1
1 1


CO
Neoplasms, Lymphatic &
Hematopoietic Tissues
Neoplasms/Lymphoid Tissu


So o CD
CO CM CD
S co m T-
o o o o


i- CM a> t~-
CM O CM T-
O O O CM





Tf f^ t~°) CD
f™^ CO ^-* CO
CO CO CM T—
co o r^ o
t- i- O •*



^J
CO
CM O CM_
in' cvio
CD co co
in o i-
d d CM

CM •ť •*•
i- CD CO
CM | d d

§•* i-
| i- CM
T^ co d





^ i CM CO
1


"o
Lymphomas
Lymphomas/Other Lymphi
Leukemias
Hodgkin's disease

.Q
O
d


co
p
^





co
CM
T—'



j3
en
CM
?
'
i—

CD
CM

^
d





CM




Gall Bladder/Bile Ducts


8
d


-
d





§
T-^





^"
9
CM
m
d

o
CO

CM
co





in




Pancreas


fe
}2
o


O)
o
d





O)
d





co

-------
                           DRAFT-DO NOT QUOTE OR CITE

    Internal comparisons were made with three broad exposure groups (Table 3-20). No
gradient in mortality on the basis of presumed exposure was seen. Age-adjusted death rates
were greatest among the group with the lowest presumed exposure for all causes combined
and in all major disease groupings; therefore, there was no evidence of dose response.  But,
the procedure used to identify and stratify exposure may have introduced confounding
negative factors that would appear to be stronger than the influence of exposure differences, if
such differences exist and affect outcome. Men in the low-exposure group mostly had
management, administrative, or support jobs rather than scientific and technical jobs; the
differences in mortality may thus reflect professional or socioeconomic differences among
exposure groups.  Despite this,  the results do not necessarily mean that an exposure gradient
does not exist.
    The two other exposure groups (medium and high) were more similar in profession and
age. The risk of death from all causes combined and in major disease groupings was greater
in the high-exposure group, but was not statistically significant. Numbers were small.  The
rates for all cancers combined were approximately the same. Excesses for Hodgkin's disease,
nervous system cancers, and cirrhosis of the liver were greatest for the medium-exposure
group. The reasons underlying the observations are not clear. The medium-exposure group
contains subjects exposed to several frequencies of NIR; about 25% worked with low
frequencies rather than microwaves.
    In summary, the overall survivorship of Rad Lab staff members was better than U.S. white
males and a group of physician specialists.  The study did not demonstrate significantly
increased risk for total mortality or mortality from specific causes to be associated with NIR
exposure, primarily from microwave frequencies. However, deaths from certain diseases were
elevated, e.g., Hodgkin's disease and digestive diseases, especially cirrhosis of the  liver.
Results for the cohort grouped by an exposure ranking scheme were equivocal.
Paradoxically, the mortality rates were highest in the group with the lowest presumed exposure
which may reflect some sort of socioeconomic bias. There is a slight gradient in the rates for
major causes of death between the groups with intermediate and highest exposure potential;
this may reflect frequency specificity.
    This study had a good tracing rate and a sufficient follow-up period to evaluate cancer, but
the necessary censoring of the data set to permit comparisons to physicians reduced  an
already small study population.  Statistical power was low for some cancers.  Exposures for
       10/18/90
3-60

-------
DRAFT-DO NOT QUOTE OR CITE
OQ
ii UJ
or§
O — '
^ O
c5 "-
Q CC
uj O
SŁ
yfc
uj .r
CO 2

^J r—
|i
o ^
UJ O
= Ł
cS z
|5 UJ $2
u_ CQ w
CO OL CC
CC 3 UJ
> CC -|
4-^0
g C5UJ
cc 5 x
UJ Z o
D- < 5
o cc o
ice i
s§*~
CO &
UJ QJ
DC CQ
f^ CC
13 UJ
< CQ
Eg
^ U-
s^
toŤ>
Q <
UJ UJ
< X
CQ <
^cc

c\
Ť c
"w <3
0 CC
1
CC T-
^ *Ł
cc 1
^ Q
To
O y
2 o
i
c
CC


X "
co
(0
cc %



E
3 CO
p
CM
v^
c5 *-
cc



ll
ť- ^
c
CC
01 *






ause of Death
o
CO f- Cť CO
T-' T-' 0 O



fe S S IS
o o o o


^ u> ^ Tf
^T rť eO CM
C3 C3 O T^




Is* ^ IO C5)
co •<* co T-
cd t>-' T-: cs

g CVj UJ CM




CO CO 0 ^
CM r*ť N- CO
U$ CO T^ C3



SO •* CM
CM



h- rŤ- CM m
O) 0 0 CM
T^ cť CM' o
*~

S'i- •* CM
CO



(O
1 s
o" S

"< < ^
V
o



CC
o


eo
e
T




§
0

i- CM T- i-




CD
CN
C9



T- i- i- eo



o


•>- CM 0 T-




co w Ł
I 11
Hodgkin's dis
Leukemias
Multiple myel
rain/CNS neor
m






























Q.
CO
1
^

"§5
Ł_
tn
t?
1 .
~ eo
ťs
Ml the confidei
OURCE: Hill, 1
CO
           3-61
10/18/90

-------
                           DRAFT-DO NOT QUOTE OR CITE

individuals could not be determined but were estimated by surrogates. There were no data on
other risk factors. Lack of smoking information is probably not a problem, because low
mortality for circulatory disease and lung cancer suggests smoking may not have been
prevalent in the cohort.
   Szmigielski et al. (1988) examined cancer incidence among Polish career military
personnel.  The authors stated that radar exposures predominated, but ELF exposures were
also noted. Large and consistent differences in cancer rates in exposed versus unexposed
personnel were found for all sites reported (Table 3-21). Most of the data is on cancer of the
lymphatic and hematopoietic system, and not all cancer sites were discussed. Generally, rates
in the exposed group were six times larger than rates in the unexposed group.
   The study was described as retrospective, but the exact design employed is difficult to
discern. In fact, the presentation of the procedures, results, and analyses is atypical. For
example, only rates per 100,000 are given; the numbers forming the basis of computed rates
are not given (Table 3-21). Furthermore,  most rates are only graphically displayed in figures
rather than listed in tables. The author noted that, due to limits, the number of cancer cases
and the size of the population, overall and within age groups, could not be presented.
Although not stated, it is possible these limits were governmentally imposed restrictions.
    In summary, mixed results are seen with studies that focused primarily on exposures to RF
radiation. Two  early studies (Lilienfeld  et al., 1978;  Robinette et al., 1980) showed only a slight
tendency for increased cancer risk in general and for several specific cancer sites. This
tendency was slightly stronger for cancers of the hematopoietic system.  These studies suffer
from either very low-exposure levels, crude exposure estimators, and/or limited follow-up
periods. The study of RF exposure and cancer, evaluated for census tracts in Hawaii
(Environmental Epidemiology Program, State of Hawaii, 1986), found significant excess risk for
leukemia if census tracts contained RF towers, but small numbers and the crude exposure
proxy, given the ecological design, limit this study's usefulness. The study by Hill (1988) found
some excess risks for some cancer sites, specifically lymphomas, Hodgkin's disease, cancers
of the nervous system (excluding brain cancer), and cancers of the digestive system, but
overall, the results are generally nonpositive or, if positive, usually not statistically significant.
Leukemias were not elevated.  Misclassification with respect to exposure was possible.  In
contrast, Milham conducted several increasingly more analytical studies of ham radio
        10/18/90
3-62

-------
                                DRAFT-DO NOT QUOTE OR CITE
      TABLE 3-21. INCIDENCE RATES PER 100,000 PER YEAR (1971-1980), FOR POLISH MILITARY
  PERSONNEL GROUPED BY WORK EXPOSURE TO RADIOFREQUENCY AND MICROWAVE RADIATION
                                            AND BY AGE
    Type of Cancer
                        E Group*
                                                                      NE Group
                             Total Population
    All Neoplasms


    Lung Cancer

    Hematolymphatic
    LGR*
    Ly Sa, Lymp*
    CLLa
    ALL8
    CMLa
    AMLa
    PLa
All ages
 40-49

All Ages

All Ages
 20-29
 30-39
 40-49
 50-59

All Ages
 20-29
 30-39
 40-49
 50-59

All Ages
 20-29
 30-39
 40-49
 50-59

All Ages
 20-29
 30-39
 40-49
 50-59

All Ages
 20-29
 30-39
 40-49
 50-59

All Ages
 20-29
 30-39
 40-49
 50-59

All Ages
20-29
30-39
 40-49
 50-59

All Ages
192.2
350.0

 33.2

 50.8
 26.3
 29.7
 81.3
117.6

  6.0
 18.8

 11.6
 18.3
 46.5
 58.8

  6.1

  9.9
 11.6
                                                         3.0
                                                         8.9
 12.2
  8.8
 19.8

 29.4

  6.1
                                                        11.6
                                                        29.4
 64.2
<50

 23.6

  7.4
  2.7
  3.0
  9.9
 29.6

  1.8
  2.1
  0.9
  2.5
  3.0

  2.2
  0.3
  0.3
  4.2
  8.9

  1.3

  0.3
  1.4
  8.9

  0.1
  0.3
  0.5

  1.2
  0.3
  1.1

  1.1

  0.3
  1.8
  2.2

  2.2
                                                                                            8.9
                                                                                            3.6
                                                                                            3.8
                                                                                           11.8
                                                                                           32.7
 E=exposed, ME=not exposed, LGR=malignant lymphogranulomatosis, Lv Sa, Lymp = lymphosarcomas and other lymphomas,
CLL=chronic lymphocytic leukemia, ALL=acute lymphoblastic leukemia, CML=chronic myelocytic leukemia, AML=acute mvelo-
blastic leukemia, PL = plasmacytoma or plasma cell leukemia.                                                 *
SOURCE: Szmigielski etal., 1987.
                                               3-63
                                                            10/18/90

-------
                           DRAFT-DO NOT QUOTE OR CITE

operators are found positive and statistically increased mortality risks for acute myeloid
leukemias (SMR=1.76) and, considered together, multiple myelomas and other neoplasms of
the lymphoid system (SMR=1.62) among FCC-licensed amateur radio operators in California
and Washington State.  For leukemias, all SMRs were elevated, but only the SMR for acute
myeloid leukemias was statistically significant. In follow-up analysis, Milham (1988) evaluated
ham operator mortality by five FCC license classes which served as surrogates for possible
higher and/or chronic exposure with increased experience and expertise. No clear gradient by
license class was found. The study by Szmigielski et al. (1988) reported increased rates of
cancer,  especially for speciific sites in the hematopoietic system, among Polish military
personnel exposed to RF and microwave radiation. Yet restrictions on the degree of detail for
reporting results limit evaluation and interpretation of this study.
   In general, epidemiologic studies of cancer and RF radiation show only a slight tendency
for excess risk, especially for the hematopoietic system sites, but they present several
methodological difficulties. The studies by Milham are the most persuasive in  demonstrating
positive and significant asssociations.   In should be noted that the exposures evaluated by
Milham  differ in frequency (lower end of the RF portion of the EM spectrum) from the other
studies  which chiefly examined microwave freqencies.

3.3.2. 50- or 60- Hertz Exposures or Electromagnetic-Field Exposure to Unspecified
Frequencies
3.3.2.1. Residential (50 or 60 Hertz)
   Wertheimer and Leeper (1982) extended their wiring analyses in a study of adult cancer in
four areas in Colorado,  e.g., Boulder, Longmont, the City of Denver, and the Denver suburbs.
In Boulder and Longmont, cancer cases were identified from  death certificates between 1967
and  1975 for residents of those towns.  A roster of potential controls was developed  by
drawing the next three certificates for noncancer deaths and for the same town.  Controls were
then selected for each case by matching on sex, age, and year of death within 5 years, and
socioeconomic level (if possible). The address history of both cases and controls was traced
and  developed for 10 years preceding diagnosis of the case.
   In addition, some living cases were drawn from the Colorado Cancer Registry. These
cases were defined as persons with a life-threatening form of cancer that had  been diagnosed
five years or more prior to the study and were alive without known recurrence  in 1979.
       10/22/90
3-64

-------
                            DRAFT-DO NOT QUOTE OR CITE

 Life-threatening cancers were those for which more than half of the registrants had died within
 5 years of diagnosis.  Controls for these living cases were drawn from a random sample of
 persons identified in a telephone survey and were matched for age, sex, and socioeconomic
 level of residential census tract.
    The approach for developing cases and controls was somewhat different for the Denver
 area. Death certificates for persons dying of cancer in 1977 in the Denver area were sampled.
 All cancer cases, except lung cancer, who had died before reaching 63 years of age, were
 included. Every other lung cancer death was included. Every other certificate was also drawn
 for deaths among persons older than 62 years.
    Neighborhood controls were selected for the cases in the Denver area. Taken from a 1970
 city directory, control addresses were randomly selected within two blocks in either direction
 of a case address. Apartment dwellers were matched.  This approach provided a similarity in
 housing stock or other characteristics, yet could tend to minimize possible wiring, hence,
 exposure differences.  Because of the latter, the method of categorizing  wiring configurations
 was somewhat changed from the approach taken in the earlier study of children.
    Cases and their controls were included only if address history could  be traced for at least
 four years prior to diagnosis. The address at which a case and control had lived the longest
 during a period of 3 to 10 years prior to diagnosis was the address used in wire coding and
 analysis.  There were 194 cases from Longmont, 321 cases from Boulder, 255 cases from the
 Denver suburbs, and 409 cases from Denver, totaling 1,179 cases.
    Five major classes of distribution wires were considered:  (1) Multiple (six or more
 nongrounded wires) or thick three-phase primary or high tension wires; (2) thin three-phase
 primary wires; (3) first-span secondary wires; (4a)  second-span secondary wires; (4b) "short"
first-span secondary wires; and (5) end-pole situations. Thin single-phase primaries were not
treated; the authors stated they are ubiquitous and carry very low current.  Wiring
configurations for homes were coded into four groups that accounted for the distribution
classes, noted above, plus distance. The codes were:  (1) very high current configurations
(VHCC) - Class 1 wires running past less than 15 meters (50 feet) and/or Class 2 wires
running past less than 7.5 meters (25 feet); (2) ordinary high current configuration (OHCC) -
Class 1 wires running past within 15-39.5 meters (50-129 feet, Class 2 wires running past
within 7.5-19.5 meters (25-64 feet), or Class 3 wires within  0-15 meters (0-50 feet); (3) ordinary
low current configuration (OLCC) - all other types of possible wiring configurations except
                                         3-65
10/22/90

-------
                           DRAFT-DO NOT QUOTE OR CITE

endpole configurations (homes beyond the endpole of a distribution line). Wires located more
than 40 meters (130 feet) from a house were not coded. This ordering of the codes represents
an estimate from high- to low-magnetic field level. Criteria were also established for
case-control pairs to determine which member of the pair had lived in a home with a higher
current configuration. Some pairs did have the same coded wiring configuration,  and these
pairs were excluded from matched pair analyses.
   Analysis was made on the basis of wiring configuration code, but some selective
measurements were made outside homes with side street access (Table 3-22).  These
measurements of magnetic fields generally supported the ordering of wire code current
configurations.
   From Table 3-22, it can be seen that even  in the very high current configurations (VHCC),
only 28.6% of the homes had actual magnetic field measurements over 3.0 mG (0.3 JU.T), which
suggest that these coding configuration categories may not be the best groupings for
exposure since most of the residences falling  in this very high category (VHCC) did not appear
to have magnetic fields at the levels found to have been associated with effects in studies of
children.
   The authors looked at only total cancer (all sites combined) in the analysis.  The four
geographic areas were analyzed separately. When the distribution of cases and controls with
respect to wiring code in four strata (VHCC, OHCC, OLCC,  and endpole) was examined, the
proportion of cases in each class generally decreased with  decreasing estimates  for
current/magnetic field, denoted by wiring code (Table 3-23). This suggests a dose-response
   TABLE 3-22. DAYTIME 60-Hz MAGNETIC FIELDS MEASURED NEXT TO THE PART OF THE HOUSE
                     NEAREST TO DISTRIBUTION WIRES BY WIRE CODE


Maximum
Median
% >3.0 mG
VHCC
(N=56)
10.0
2.5
28.6
OHCC
(N = 134)
8.0
1.2
10.4
OLCC
(N = 186)
3.0
<0.5
1.1
End-pole
(N=41)
1.4
<0.5
0.0
   SOURCE:
       10/22/90
3-66

-------
DRAFT-DO NOT QUOTE OR CITE












52
o
f
o
o
CASES AND
CANCER
fe
co
UJ
Ťj
O
X
UJ
1

^
co
O
5
cc
i
i
8
0
z
EC
*
CO
UJ
	 1
CD
^














0
o
V
.9 — Q.
{§i s




..
'c •Ł
O i: co
"ť 1 s "*
lid
2. §
8f s m
.i fc (0 05
Ł S S
O)
X





—1
Qj
•i s?
UJ

o
O g J
o "- ^-

.1
|o
i)O o co
>= X IT CM
1°
"5
I0
O 1^ CM
x *- *-





\-
o M
rn w P
y 0) —
Z en c






P g
0 0
V II
Q. Q.
CO  ?~
m o o in cb in

 ^ 5


m. in co
CM CVJ CD CO



cq N. co
^ tn cvj o o co c\i
in >p i- K 55 en in

co o eo
CO CO CO S § S S





_i
to ^ QC to 2? to *"
1
. .c
.O)
(0
f
6
O
X
O
-
.2
1
J?

8
1
3
O)
Ic
^
O
O
>
1
1 Ť
i" S
8 2
S- o g"

Ł 2 §>
Jo> ^
"c -n
ss|
If!
Q) 7J 3
III
.S CL Q
CQ ^1 Q)
is? c- *|
g.2^
|fl
f" " "c
1 1 s ^
If! ^
"O C 2 T3
SI 
-------
                            DRAFT-DO NOT QUOTE OR CITE

relationship; but, no statistical tests were reported.  No estimates of risk were given in these
tables. The authors presented other results in terms of a "C-ratio" ("C" stands for
"configuration") which seems to be similar to odd ratio estimates for matched pairs.  In this
situation, the C-ratio was described as the number of pairs in which the case lived in the
higher configuration home divided by the number of pairs in which the control lived in the
higher current configuration home, then multiplied by 100.  Significance was tested using the
Sign Test.  The results are like matched pair estimates of the odds ratios. The difference
relates to how the data are described or categorized; that is, rather than arraying cases and
their controls as with or without a factor, they are arrayed as to which has a higher or lower
wiring code for their residence.
    As shown in Table 3-24, some of the C-ratios were statistically significant by the Sign Test,
indicating an association between higher wiring codes and total cancer. This effect was
strongest in the youngest age group, i.e., persons under 55 years of age.  The cases and
controls were distributed with respect to urban versus nonurban residence, modified by age,
that is, older persons were treated as having lower exposure to "urban factors." In this case,
     TABLE 3-24. C-RATIOS FOR CANCER IN COLORADO ADULTS FOR RESIDENCE NEAR HIGHER WIRING CURRENT
                          CONRGURATIONS BY TOWN AND VARIOUS FACTORS

Total
Diagnosis Age:
19-54 Years
55-69 Years
70+ Years
Male
Female
Urban:
Total
Central Denver
Male Suburban,
19-69 Years Old
Low Urban:
Total
Male Suburban,
70+ Years Old
Female Suburban
Total
139*

201*
114
137*
NDR
NDR

113
121

81

162*

500*
151*
Denver Central
Longmount Boulder Suburbs Denver
164 143* 146* 121

170 264* 200* 188*
220* 129 97 191
136 129 225 130
NDR NDR 103 NDR
NDR NDR 194* NDR










 * p-value at least 0.05. NDR =- No difference reported.
 SOURCE: Wertheimer and Leeper, 1982
        10/22/90
3-68

-------
                            DRAFT-DO NOT QUOTE OR CITE

the risk was stronger in the "low urban" exposure group.  Although not given in Table 3-24, this
result was consistent across three age strata. The effect was also stronger in the highest
socioeconomic group.
    The authors considered all cancers together. But they noted that significantly high C-ratios
were found for cancer of the nervous system, uterus, breast, and lymphomas. Elevated but
nonsignificant ratios were found for cancer of the pancreas, bladder, kidney, and prostate.
Except for nervous system tumors, these sites are different than those associated with
magnetic field exposure in studies of children. Leukemia was not reported and presumably
was not associated with magnetic fields, as estimated by wiring code.
    The authors stated that they drew four main conclusions from their work, as follows:  (1) a
dose-response relationship was found; (2) the association was not an artifact of age,
urbanicity, neighborhood, or socioeconomic level; (3) the association was most demonstrated
where urban/industrial factors that may relate to cancer were least likely to obscure the effect;
and (4) there was a distinct pattern of latency consistent with a promotion effect of exposure to
alternating magnetic fields (AMFs). These broad generalizations may be unwarranted in view
of the unusual methods used to design and analyze this study.  The selection of cases and
controls for inclusion in the study group was somewhat convoluted. The unusual manner in
which certain subgroups (but not all) were included as cases, e.g., cancer survivors, as well as
others who were excluded as controls, e.g., lung cancer deaths, makes it difficult to assess the
direction and impact of selection factors on the results. Furthermore, the use of nonparametric
testing methods to obtain point estimates (C-ratios) and assess significance (the Sign Test)
instead of the more powerful odds ratio and Mantel-Haenszel testing procedures is strange.
The latter statistical tools are ideally suited to case-control studies such as this. Why they
were not used is not stated in the paper. The net result is to make it difficult to place  a value
on the usefulness of the conclusions in evaluating a carcinogenic effect.
    Coleman et al. (1985) presented preliminary results of a case-control study of leukemia in
London, England. Incident cases of leukemia registered in the Thames Cancer Registry for
1965-1980 were selected for study. Over 99% of the cases were histologically confirmed.
Cases were limited to residents of four adjacent South London boroughs. Why the study was
limited to these areas was not explicitly explained,  but the authors noted the boroughs were
densely populated, contained both urban and rural areas, and had no boundary changes in
the study period. Two controls per case were randomly selected from all registered cases of
                                         3-69
10/22/90

-------
                           DRAFT-DO NOT QUOTE OR CITE

solid tumors (except lymphoma).  If possible, a third control was also selected to hold in
reserve. Controls were matched to cases for 5-year age group, sex, year of diagnosis, and
borough of residence. The authors do not give the ages of the cases. There were 811 eligible
cases; 42 were excluded because they could not be located or there were no matching
controls, resulting in 769 final cases.  There were 1436 controls after exclusion for, primarily,
inability to locate either the control or the matched case. Unfortunately, 102 cases could only
be matched to one control.
    Exposure to EM fields at 50 Hz was estimated by mapping environmental electrical power
supply sources and computing distance from residence to sources. For overhead lines, the
distance used was the shortest distance between the residence and the line connecting
pylons. Ordinance Survey grid maps were used to plot subject addresses and substations
and overhead lines. Cases versus control status was blinded.  Underground cables were not
on the grid maps but data on their paths were obtained from utilities as well as was additional
information on substations.
    This paper presented only results for overhead power lines.  Less than 1 % of the
population lived within 100 meters of lines at the time of cancer registration. When 100 meters
was used as the demarcation distance for critical exposure, the odds ratio for leukemia was
1.31 (95% CI=0.50-3.45) but the result was not statistically significant; there were only 7 cases
and 10 controls who lived less than 100 meters from an overhead power line.  Distance to
overhead power lines was also stratified further into groups living 0-24, 25-49, 50-99, and 100
or more meters. Taking the last group as having no exposure, the odds ratios for leukemias
for these groups were 1.87,1.87, and 1.07, respectively. Although the odds ratios, to some
extent, increased with decreasing distance, the trend was not-significant (p=0.24). The odds
ratio remains 1.87 if the two low groups (0-24 and 25-49 meters) are combined. The lack of
significance may be due to the small numbers of cases and controls at these close-in
distances and a subsequent loss  of statistical power. The authors noted that almost the entire
local distribution network is underground. Since few people live near overhead power lines, it
could be very difficult to detect small increases in risk.
    The paper focused chiefly on proximity to overhead power lines which  are uncommon in
the region studied.  Other than this proxy measure, exposure estimates were not given.  Also,
proximity of residence to lines was taken at the time of tumor registration which may not
necessarily be the same residence as at onset of disease. Increased risk was seen for
       10/22/90
3-70

-------
                           DRAFT-DO NOT QUOTE OR CITE

leukemia for residences nearer to power lines. Statistical significance was not seen; this may
reflect the small number of observed cases.
   The authors updated this paper (Coleman et al., 1989) by enlarging the control group to
include two referents chosen from the same Thames Cancer Registry from which the cases
were chosen. The controls were to be picked from other registrants with solid tumors other
than lymphoma.  Using the same 769 plus 2 additional cases discussed in the earlier paper,
110 were matched to only one control because the address at registration could not be
located on the primary control and no reserve control was eligible.  The remaining 661 were
matched to 1322 controls. A total of 1432 controls resulted. The surrogate for exposure was
again distance from overhead power lines but now distance from substations was also
considered.  Distances were 0-24, 25-49, 50-99 and 100+ meters. Despite the large number of
cases and controls, very few individuals lived within 50 meters of the overhead lines. Only
three cases and three controls were found to live within 50 meters distance from power lines.
This was enough to produce a nonsignificantly increased risk of leukemia (OR=2.00). On the
other hand, for those residing within 25 meters of a substation the risk was nonsignificantly
elevated (OR= 1.26) based upon 35 cases and 51 controls. No increased risk was seen in
persons residing beyond 25 meters.
   Included as part of this case-control study were 84 cases in children under age 18 who
were matched with 141 controls. The authors performed a separate "nested" analysis of this
subgroup in which they noted a trend of increasing risk of leukemia with decreasing distance
from a substation, as follows:

RELATIVE RISK BY DISTANCE OF SUBJECT'S HOME FROM SUBSTATION IN CHILDREN UNDER AGE 18


Cases
Controls
RR

0-24
3
3
1.63

25-49
11
12
1.49
Meters
50-99
22
48
0.75

100 +
48
78
1.00

Total
84
141

                                        3-71
10/22/90

-------
                           DRAFT-DO NOT QUOTE OR CITE

None of these risks are significant, however, but they appear to be supportive of the findings in
the childhood cancer studies of an increased risk of leukemia from exposure to magnetic
fields.
   No actual measurements of magnetic fields were available to substantiate exposure in any
of the designated distance categories. In other studies where measurement data were
available, distances at and beyond 25 meters from most sources of magnetic fields have not
been found to harbor magnetic fields of strengths much greater than those at 100 meters.
Hence, it may not be unexpected that risks calculated in these locations are not elevated.
Furthermore, implicit in these risk calculations involving distance is the assumption that no
other sources of electromagnetic fields were producing magnetic fields at those same
locations.
   Another potential confounding effect can happen when referents are taken from the same
cancer registry as are the cases. If exposure to EM fields is associated with cancers other
than leukemia, then the referents with the associated cancers may be found to live closer to
the source of the magnetic fields, and the risk estimate for leukemia will be reduced as a result.
    And again as in most of these studies duration of exposure (length of residency in the
magnetic fields) and latent effects have not been addressed by the authors.
   However, despite the study's limitations, the observed relationship between proximity to
the line and excess leukemia risk is consistent with a weak positive effect.  However, the
authors also pointed out that residence near overhead power-lines is not a major leukemia
hazard for large numbers of people in the geographical area studied.
    Using a cohort study design, McDowall (1986) investigated the mortality rate of persons
who resided near electrical transmission facilities in East Anglia, England. The study period
extended from the 1971 Population Census until the end of 1983. Sample size calculations
(not provided) indicated approximately 8000 persons or 3000 households would be needed to
detect a twofold increase in risk for most major cancer sites.
   Subjects were entered into the study using a stratified random sampling scheme that
served to identify houses from National Grid Maps.  Houses (and, consequently their
occupants) were selected if they were within a 50-meter radius of a substation or within 30
meters either side of an over head power cable. A total of 2839 houses were selected.
   Census data were examined to identify and extract information about the resident
householders in the 2839 dwellings.  Abstracted information included name, address, date of
       10/22/90
3-72

-------
                           DRAFT-DO NOT QUOTE OR CITE

birth, sex, occupation, employment status, and address 1 year and 5 years prior to the census.
Census data on 7920 persons were linked to the National Health Service Central Register
(NHSCR), and 7631 persons were located, a tracing rate of about 96%.  Of these, 814 (409
men, 405 women) had died by December 31,1983, the end of the  study period. The sampled
population was generally similar to the whole East Anglia population with respect to sex and
social class (based on occupation and employment status), but the sample was somewhat
younger.
   Cause of death was obtained from death certificates and was coded to the Revision  of the
International Classification of Disease (ICD) in effect at the time death. National coding rules
were followed for comparability to the reference rates  used in analyses.  Standardized Mortality
Ratios (SMRs) were calculated for certain specified causes of death. Expected deaths were
calculated from East Anglia mortality rates.  National rates for England and Wales were used to
generate expected deaths due to hematopoietic neoplasms.
   For both men and women, mortality for major causes of death was generally similar to
what would be expected based on rates in East Anglia; SMRs for men were generally lower
than expectation, and SMRs for women were generally higher than expectation. The only
cause of death that significantly exceeded expectation was lung cancer among women.  SMRs
for leukemias and other hematopoietic neoplasms were elevated only for women (154 and
171, respectively), but the result did not significantly differ from expectation.
   SMRs were also calculated for subsets of the population, categorized by distance in
meters of the residence from electrical installations, e.g., 0-14,15-34, and 35-50 meters.  The
sexes were combined in this analysis to try to obviate  small numbers.  For  all groups, only  the
SMR for lung cancer (215) was significantly greater than expectation in the group who had
lived 14 meters or less from an electrical installation. Deaths from respiratory disease were
also slightly elevated (SMR 127).  Elevated but nonsignificant SMRs were noted for leukemia
(143) and other neoplasms of the hematopoietic systems (333) at this same distance based
upon only a few deaths. The highest SMR was for breast cancer (122), and that occurred in
the subgroup that resided 15-34 meters from an electrical installation.  But, generally, there
was little difference from expected values. The highest SMR for the subgroup that lived the
farthest away from electrical installations (35-40 meters) was for nonleukemia hematopoietic
neoplasms (144).  The SMR for leukemia was, at 120, the second highest.  The SMRs for
                                        3-73
10/22/90

-------
                           DRAFT-DO NOT QUOTE OR CITE

several other major cancer sites, including leukemia, exceeded unity, but overall cancer
mortality was less than unity.
   The sampled population was also examined with respect to address 5 years prior to the
1971 census, e.g., same address and different address. Again, the results for both sexes are
combined.  In both groups, only the deaths from circulatory disease significantly differed from
expectation, and were less than expected. Although not significant, the SMRs for malignant
neoplasms of the hematopoietic system, other than leukemia, were the highest in both
address groups, e.g., 125 for same address and 138 for different address.
   The two address groups were also further subdivided, and SMRs were calculated for
persons who lived within 25 meters of an electrical installation at the time of the 1971 census.
In this analysis, results for circulatory diseases remained significant only for the same address
group. Also, for this group, only deaths from lung cancer occurred in excess. In the group
who had resided at a different address 5 years prior to the census, many causes of death
occurred in excess of expectation, but SMRs did not differ significantly from unity.  The SMR
for nonleukemia lymphatic neoplasms was 300, but there were only three cases and the 95%
confidence interval was very wide.
   Overall, the mortality level of the sampled population for all and specific causes did not
differ from expected based on rates in East Anglia or on national rates.  All-cause mortality was
significantly lower than expected, primarily due to lower rates for death  from circulatory
disease. The author speculated that older housing units with older residents may have been
under-represented in the sample, and this is responsible for the observed lower mortality in the
study group.  Generally, for all analyses, SMRs for  leukemia and for other hematopoietic
neoplasms were elevated, based on regional and national rates. This result was strong for
females. These results were not significant,  but the number of cases was small. The highest
SMRs for these causes  occurred among persons living the closest (15 meters) to electrical
installations. Excess  lung cancer was seen among women and among both  men and women
who lived less than 15 meters from an electrical installation.
   In this study, residence near electrical installations (50 meters)  or overhead lines (30
meters) was taken as the indicator for exposure. Field measurements,  either single or
continuous, were not made. AH types of facilities were included regardless of type or current.
It is interesting that only 19 of the 7631 residents lived within 30 meters of an  overhead cable.
       10/22/90
3-74

-------
                           DRAFT-DO NOT QUOTE OR CITE

It seems that the methods employed yielded populations whose exposures were likely to be
low or very low.  Distance from a source was used as a proxy for field strength.
   Although the author stated that his findings do not support the previously reported
associations of exposure to EM fields with acute myeloid leukemia and other lymphatic
cancers, he did report elevated, albeit nonsignificant, risks of leukemia and cancer of the
hematologic and lymphatic tissue based on very small numbers in persons living less than 15
meters from electrical installations. This is a cohort study that basically had very little power to
detect risks of rare cancers such as leukemia; the conclusions of the author  must, therefore,
be viewed in the context of the study's limitations.  Electric and magnetic field measurements
are lacking and so the EM field environment that characterizes each distance category is
unknown. The author suggested that perhaps his study could not confirm associations noted
by other investigators because exposures characterizing the study area are probably too low,
or exposure was diluted by mobility, or there were inadequacies in the study design.
   Severson et al. (1988) investigated the occurrence of acute non-lymphocytic leukemia and
residential exposure to power frequency magnetic fields in adults in western  Washington State.
Cases were identified from the population-based Cancer Surveillance System of the Fred
Hutchinson Cancer Research System. Cases  were defined as all newly diagnosed cases,
aged 20-79 years, of acute nonlymphocytic  leukemia (of seven histological types) between
1981 and 1984 in three counties.  Living and deceased cases were included  in the study.
Controls in the same region were selected by using random-digit dealing methods and were
frequency matched for sex and age in 5-year groupings. Study subjects or their next-of-kin
were interviewed with respect to residence history, ionizing radiation exposures, occupational
history,  medical history, medications and drug use, smoking and alcohol history, pet
ownership, and various demographic characteristics.
   An arbitrary reference date was randomly assigned to each control to try to make the recall
period similar between cases and controls.  This resulted in a uniform distribution of reference
dates over the period of case diagnosis.
   Three different and extensive methods of exposure assessment were used. The first
method mapped external wiring configurations within 140 feet of each subject's residence
(except apartment dwellings) over a 15-year period prior to the reference date.  The technician
who developed these maps was kept unaware of whether an address was a  case or a control.
Each map was then  classified, according to  a coding scheme similar to that of Wertheimer and
                                         3-75
10/22/90

-------
                           DRAFT-DO NOT QUOTE OR CITE

Leeper (1979,1982), into one of four groups, e.g., endpoles, ordinary low current
configuration, ordinary high current configuration, and very high current configuration, that
presumably represent a gradient of increasing potential exposure.
   The second method of exposure assessment was to measure magnetic fields in and
outside the residences at the time of the interview for subjects who had lived in that residence
at least one year or more prior to the reference date.  Magnetic fields were calculated as root
mean square (rms) of three measurements per location and were expressed in milligauss (mG,
1 mG=0.1 microtesla./fT). Measurements were made in the kitchen, the subject's bedroom,
and the family room under conditions of low power use (most appliances off) and high power
use (all appliances turned on in the room being measured).  For both low- and high-power use
conditions, two exposures were calculated, i.e., the mean of the three room measurements
and the mean of the three room measurements weighted by the estimated time spent by the
subject in each room.
   The third method of exposure assessment was to measure magnetic fields over a 24-hour
period in residences where the one-time-only measurements had been made. Using an
assessment scheme developed by Kaune et al. (1987), the 24-hour meter readings  were
correlated to various aspects of the wiring configurations maps to estimate magnetic field
levels within each residence.
   Originally, 164 cases and 204 controls had been identified (Table 3-25). Refusals to
participate, either by physicians, hospitals, or subjects, or other reasons for losses or
nonparticipation resulted in 114 cases and 133 controls available for interview and analysis.
The interview response rate was, thus, 69.5% and 65.2% for cases and controls, respectively.
Although the level of response is similar for cases and controls, it is of concern that the rate is
low.
   The cases and controls were compared with respect to various demographic and other
factors. It was found that cases tended  toward lower socioeconomic status, measured
separately by education and family income, and more cigarette smoking. No increased risk
was found for pet ownership or estimates of ionizing radiation exposures.  The cases  and
controls were similar with respect to length of residence in mapped dwellings. Persons who
had lived in apartments for all 15 years prior to the reference date or had moved to the study
area just prior to the reference date were excluded from the analyses.
       10/22/90  ,
3-76

-------
                                DRAFT-DO NOT QUOTE OR CITE
 TABLE 3-25. LOSSES/EXCLUSIONS IN A STUDY OF ACUTE NONLYMPHOCYTIC LEUKEMIA AND RESIDENTIAL EXPOSURE
             TO POWER FREQUENCY MAGNETIC FIELDS IN ADULTS IN WESTERN WASHINGTON STATE
 Cases
No.
                                                          Controls
                                                        No.      %
        Number Identified
        Number Interviewed
        Exclusions/Refusals
164
114
 50
100
69.5
30.5
204     100

133     65.2

 71     34.8
SOURCE: Severson et al., 1988.
 TABLE 3-26. RISK ESTIMATES OF ACUTE NONLYMPHOCYTIC  LEUKEMIA IN ADULTS IN RELATION TO EXPOSURE BASED
  ON WERTHEIMER AND LEEPER'S WIRING CLASSIFICATION  SCHEME  FROM FITTED LOGISTIC REGRESSION MODELS,
                                  WESTERN WASHINGTON STATE, 1981-1984
        Wiring Configuration
        No. of
        Cases
            No. of
            Controls
            Odds
            Ratio8
 95% Cl b
        Longest Residence 3-10 Years before Reference Date
                Very Low (Endpole)
                Ordinary Low
                Ordinary High
                Very High

        Residence Closest to Reference Date

                Very Low (Endpole)
                Ordinary Low
                Ordinary High
                Very High
          42
          21
          21
           5
          42
          26
          24
           5
               44
               37
               23
                6
               52
               38
               19
                7
            1.00
            0.60
            0.77
            0.79
            1.00
            0.81
            1.36
            0.84
0.29-1.22
0.35-1.68
0.22-2.89
0.41-1.61
0.62-2.96
0.24-2.93
a Controlling for age, sex, cigarette smoking, family income, and race.
b Cl  = confidence interval.

SOURCE:  Severson et al., 1988.
                                                 3-77
                                                      10/22/90

-------
                            DRAFT-DO NOT QUOTE OR CITE

   With one apparent exception, there was no increased risk of acute nonlymphocytic
leukemia from exposure to magnetic fields under the exposure assessment schemes and
methods of analysis applied in the study.
    Two analyses of wiring configurations (based on residential mapping as previously
discussed) were reported, that is, analysis of wiring configuration for the place of longest
residence and for the residence closest to the "reference" date. Also, two analytic methods
were used, i.e., fitting the data to a logistic model and estimating relative risk for exposure
strata. Odds ratios were developed and controlled for age, sex, cigarette smoking, family
income, and race. With the logistic model and for both types of residences, odds ratios did
not generally exceed unity, and no increased risk was associated with residence in homes
classified with a higher current configuration, the presumed highest magnetic field exposure.
No estimate was, however, significantly different from expected. Results were fairly flat across
all configurations from very low to very high. Similar results were obtained when subjects with
underground residential wiring (called endpoles) were excluded from the analyses. Inclusion
of this subgroup in the lowest exposure category would tend to underestimate the risk if any
did (does) exist.  It was suggested in a critique of this paper by Wertheimer and Leeper (1989)
that if the two low exposure categories are combined and then contrasted with the two highest
categories, combined, a weak but modest increase in the risk of this cancer will be seen, i.e.,
1.2 for longest residence and 1.5 for "closest" residence. This is probably not an unreasonable
exercise in view of the small numbers that the authors dealt with.  The suggested
recategorization by collapsing into only two groups does add some stability to the numbers.
Savitz et al. (1988) did much the same thing to improve risk estimates and decrease variability
of those estimates.
   Analyses for both longest residence and most recent residence were also performed  for
exposure estimates in milligauss developed using the method of Kaune et al.  The exposure
estimates were categorized in three groups representing low (0.51-1.99 mG), and high (2.00
mG) exposure. With this approach, the odds ratios were again less than 1.0 and none were
significant. Of note is that the authors' table on this analysis lists 19 additional cases and 26
additional controls, in the category entitled "residence closest to reference date," which is 41
more subjects than the total interviewed.
   Risk estimates were also developed for the weighted and unweighted mean magnetic field
measurements, divided into three groups (0-0.5, 0.51-1.99, and 2.00+ mG), under both low-
       10/22/90
3-78

-------
                            DRAFT-DO NOT QUOTE OR CITE

and high-power use conditions, although there was little difference (0.174) between them. The
odds ratios were nonsignificantly higher in the exposure level 2.00+ mG. It was in this
analysis that the only dose-response result was obtained, that is, an increase in risk was seen
with increased mean exposure under the low power configuration (Table 3-27).  Also, the trend
was not observed when exposure was weighted by time spent in each measured room, based
upon interview data. Application of logistic models suggested measure fields were not related
to disease.
   There was a borderline significantly increased risk (OR=2.4) of acute nonlymphocytic
leukemia observed for use of electric blankets, water  bed  heaters, and electric mattress pads,
elicited in the interviews, in subjects with an annual family income under $15,000.
   This study has several drawbacks and weaknesses that need to be addressed before the
results can be considered definitive. First, the study deals only with one type of leukemia,
acute nonlymphocytic leukemia. Chronic nonlymphocytic as well as acute and chronic
lymphocytic leukemia were not considered; the results from this paper do not pertain to them.
   TABLE 3-27. RISK ESTIMATES OF ACUTE NONLYMPHOCYTIC LEUKEMIA IN ADULTS IN RELATION TO THE WEIGHTED
     AND UNWEIGHTED MEAN MAGNETIC FIELD MEASUREMENTS MADE IN THE KITCHEN, BEDROOM, AND FAMILY
  GATHERING ROOM OF THE SUBJECT'S RESIDENCE AT REFERENCE DATE, WESTERN WASHINGTON STATE, 1981-1984
Exposure Level
(milligauss)b
Mean Exposure
0-0.50
0.51 - 1.99
2.00 +
Weighted Mean Exposure
0-0.50
0.51 - 1.99
2.00 +
Odds Ratio
Low power
Configuration

1.00
1.16 (0.52-2.56)
1.50 (0.48-4.69)

1.00
1.17 (0.54-2.54)
1.03 (0.33-3.20)
(95% Cl a)
Hign Power
Configuration

1.00
0.55 (0.25-1.22)
1.56 (0.49-5.04)

1.00
0.91 (0.42-1.96)
1.25 (0.35-4.48)
 a Cl  = confidence interval.
 b 1 milligauss (mG) = 0.1 microtesla (fiT).
 SOURCE:  Severson et al., 1988.
                                          3-79
10/22/90

-------
                            DRAFT-DO NOT QUOTE OR CITE

 Second, a very small number of cases and controls were selected for inclusion into the study.
 In addition to the small sample size, a poor response rate was attained. Poor response
 among cases was chiefly due to physician and hospital refusals. No explanations were given
 regarding the reasons for the refusals. One must have some concern about the impact on the
 results of missing information from the approximately 35% of the population who were
 nonrespondents. The small sample size attained makes it somewhat questionable that the
 nonpositive results are indicative of the underlying risks for this type of leukemia. The study
 had little power.  Third, magnetic field measurements taken many years later are not
 necessarily representative of actual exposure received  by the cases prior to diagnosis. This
 problem is, of course, not unique to this study. Latency must be considered. Length of
 residency prior to the date of the measurement of electric and magnetic fields or the date
 when the next-of-kin was interviewed is not a proper measure of latency.  Fourth, a potential
 source of bias is the use of live controls matched to dead cases. Data provided from
 questions asked of live controls may differ from the data provided by the next-of-kin of dead
 cases. Lastly, no mean and/or median magnetic field measurements were taken within each
 wiring code category to substantiate any potential trend of increasing exposure from "very low"
 to 'Very high." The author's own analysis of his wiring code does not correlate well with his
 24-hour magnetic field measurements.
    Preston-Martin et al. (1988) reported the results of a study of 116 adult acute and 108
 chronic myelogenous leukemia (AML, CIVIL) cases that examined effects of exposure to EM
 fields from electric blanket use. This was a case-control study that was not originally designed
 to examine potential risks from EM fields.  But, because of high fields, prolonged exposure,
 and intimate contact from electric blankets, a source of EM fields, the authors stated they
 added questions on electric blanket use to ongoing studies
    Cases were identified from the University of Southern  California Cancer Surveillance
 Program, a population-based registry, and were Los Angeles County residents aged 20-69
 years with histologically confirmed AML or CML diagnosed from July 1979 to June 1985 (Table
 3-28). Cases had to be alive and able to be interviewed in English;  this criterion was also
 applied to controls. There were 859 registered cases but only 458 were alive. Further
 selections were made only among the live cases.  Exclusions were 122 patients or doctors
who refused to participate and 68 cases who could not be located.  After these exclusions
were made, there were 295 cases for whom questionnaires were completed. The controls
       10/22/90
3-80

-------
                            DRAFT-DO NOT QUOTE OR CITE
      TABLE 3-28. ELECTRIC BLANKET USE AMONG ADULTS (AGED 20-69 YEARS) WITH ACUTE AND CHRONIC
  MYELOGENOUS LEUKEMIA, RESIDING IN LOS ANGELES COUNTY, AND THEIR MATCHED-NEIGHBORHOOD CONTROLS
 Factor
                                                   Case      Controls    Odds Ratio
 Ever used electric
  blankets regularly
       AMLa
       CML
 Average usage duration,
  in years
       AML
       CML
 Years of first use
       AML
       CML
 Years since last use
       AML
       CML
40
38
8.5
9.0
1971
1970
3.2
2.7
43
42
7.0
10.0
1971
1970
4.1
2.8
0.9
0.8
 aAML=acute myelogeneous leukemia, CML=chronic myelogeneous leukemia.
 SOURCE:  Preston-Martin et al., 1988.
were individually matched neighborhood controls, matched on sex, race, and birth year ą5
years.  The selection procedures procedures resulted in 293 matched controls for whom all
questionnaires were completed.
   Telephone interviews were conducted and, because the process for identification and
selection was explained to the subjects, the interviewers were not blinded to case versus
control status. Information was obtained on events up to the time of the interview, but
analyses were restricted to events that took place at least 2 years prior to diagnosis. In
addition to electric blanket use, i.e., year first used, year last used, and duration of use, the
questionnaire solicited information on jobs, chemical use, radiation therapy, and diagnostic
radiography. Procedures for matched-pair analyses and multivariate methods were used.
Odds ratios were maximum likelihood estimates.
   Because the survey's question on electric blanket use was added to the ongoing study,
only 224 matched pairs answered the question on whether they had ever used an electric
blanket regularly. Cases (both AML and CML) did not differ from controls with respect to
regular usage of electric blankets (AML OR=0.9, CML OR=0.8). The cases and controls also
                                          3-81
                              10/22/90

-------
                            DRAFT-DO NOT QUOTE OR CITE

did not differ for patterns of blanket usage, as shown in Table 3-28. Adjustments for other risk
factors, i.e., diagnostic radiography, work as a welder, and farm residence, did not alter the
observed lack of association.
    This study has some merit as a general study of leukemia in adults in Los Angeles, but
several major problems limit its usefulness in judging the risks of EM fields. Although the
study provides no support for the hypothesis that electric blanket use is associated with
myelogenous leukemia, the authors pointed out that this hypothesis might be more effectively
tested in an area with a harsher climate than Los Angeles with its year round warm climate.  It
also says nothing about the risk of lymphatic leukemia, which mainly affects children. The
sample size just achieves the 80% power needed for a one-sided detection of a risk.
Furthermore, deceased myelogenous leukemias were excluded from this study. Only live
cases were considered.  Hence, the results apply only to survivors, since there may be
differential risks between surviving and deceased cases.  Only 53% of all registered
myelogenous leukemia cases survived to be included in the study. The use of only live cases
represents a substantial loss of information about all leukemia cases. On the other hand, the
accuracy of information on electric blanket use may be improved if the information comes from
the subject himself rather than the  next-of-kin. Another major problem is the very broad nature
of the interview questions (Table 3-27) on electric blanket use which only provided a very
crude surrogate of possible exposure. For example, detailed information on frequency of use
other than "years" of use as well as detailed questions on use patterns would  have been more
informative.  Also, as the authors point out, such blankets can generate fields from 12 to 50
mG (1.2 to 5.0/iT). Individuals vary according to how often and how much they depend upon
electric blankets for warmth and the intensity settings they prefer.  Furthermore, the questions
on electric blanket use were added, according to the authors, after 69 cases and 65 controls
had been interviewed. This implies that the authors had not originally intended to study
electric blanket use. The propriety of introducing a new hypothesis, midway through a
scientific study for testing, must be questioned. Because of these issues, this study cannot  be
considered conclusive and adds little to our understanding of the cancer risks posed by this
source of EM fields.  As the authors state, it might be better if future studies use populations
more dependent upon electric blankets than are residents of Los Angeles.
    In summary, the results obtained from studies of cancer incidence or mortality in adults in
relation to residential exposures to EM fields from 50- or 60-Hz power transmission frequencies
       10/22/90
3-82

-------
                           DRAFT-DO NOT QUOTE OR CITE

are not as strong as were the results obtained from studies of children. The only clearly
positive study was conducted by Wertheimer and Leeper (1982).  The other studies (Coleman
et al., 1985; McDowall, 1986; Severensbn et al., 1988) of power frequency fields are generally
nonpositive or only show weak associations for increased risks of leukemias or hematopoietic
cancers in toto from exposure to EM fields, estimated by proximity to lines or by wiring
configuration codes. There has been discussion in the literature that recategorization of
exposure groups in the Severenson et al. (1988) study may produce more strongly positive
results. The study also exhibited poor response rates, although rates were comparable for
cases and controls. The study by Preston-Martin et al.  (1988) is of a different type in that it
briefly examined electric blanket use rather than power transmission and distribution fields.

3.3.2.2. Occupational
   This section examines studies on occupations and workplace exposures. It is organized
by the cancer site, to the extent possible, that was the principal endpoint under study.

3.3.2.2.1.  Multiple Sites. The following studies examined cancer mortality or incidence in
general and for more than one site.
   Wiklund et al. (1981) reported on the first use of the Swedish Cancer-Environment Registry
to investigate cancer risk in an occupational group, that is, telephone  operators for the
Swedish Telecommunications Administration. The Registry represented linkage between the
1960 census and the National Cancer Registry for 1961 -1973. The development of this joint
registry was enabled by the assignment of individual personal identification numbers to
Swedish residents. About 1 % of registered cancer cases cannot be identified in the census.
About 0.5% of linked cancer case census matches are believed to be  inaccurate. Another
quality control assessment compared death certification to incident case registration and
found the total loss of cases to be approximately 3%-4%. But certain  malignancies, including
the leukemias (11.5%), exhibited a greater registry loss rate.
   Cases were defined as telephone operators at the Telecommunications Administration
identified as such in the 1960 Census and who were entered in the Cancer Registry with a
diagnosis of leukemia; only 12 cases resulted. Sex was not identified. The expected number
of cases was calculated by applying national rates specific for year of birth and sex; 11.7
cases were expected overall. Observed to expected ratios were evaluated by means of the
                                         3-83
10/22/90

-------
                            DRAFT-DO NOT QUOTE OR CITE

 Poisson distribution.  The observed cases were not found to differ significantly from
 expectation.
    The development and use of national cancer and other registries is very important and
 useful.  But the approach and results presented here do not seem to be fully satisfactory for
 examining carcinogenic risks among telephone operators, as discussed below. As the
 authors point out, many factors and different approaches should be applied to fully assess
 whether exposure to a suspected carcinogen may result in increased incidence of cancer.  For
 example, the underregistration of leukemia cases to the cancer registry, in general, is of
 concern. It seems surprising that the national linked databases yielded only 12 cases between
 1961 and 1973 (13 years) whereas the cluster of cases that prompted the study totaled 4
 cases in one city between 1969-1974 (6 years). Employment was ascertained from the
 census; this could possibly misrepresent the number of operators.  In any event, the census
 data is a point estimate for employment.  While this study is valid on its own merits, other
 approaches might be applied in the future to further examine the  Gothenburg leukemia cluster
 and/or to further examine whether any cancer risk is present for telephone operators or other
 telecommunications workers.  For example, a cohort study among workers at the
 Telecommunications Administration might be useful.
    Howe and Lindsay (1983) reported on a computerized system for linking records to
 monitor mortality among the Canadian labor force. Data were collected from all employers
 between 1965 -1969 and 1971 on an approximate 10% sample of the Canadian work force,
 using social insurance (SI) identification numbers as the sampling frame.  Occupations and
 industries were coded to 1961 (for 1965-1969 data) and 1971 (for 1971 data) census
 designations. Data from 1970 were not available; the reason was not stated.  A file of three
 million  records was generated. Because individuals could be identified over several sampled
 years, the records were sorted to develop a composite record per person. The file was also
 examined for errors. These procedures resulted in a final file of 700,335 records of individuals.
The records contained information on surname; first and second  given name; sex; day, month,
 and year of birth; mother's maiden name; and occupation(s) and  industry(ies) of employment.
    The data for the sampled population were linked to the national  mortality registration
system for the years 1965-1973. A scheme was developed to determine whether the linked
data were in agreement, and, thus, whether a match had been found. If a match was
      10/22/90
3-84

-------
                           DRAFT-DO NOT QUOTE OR CITE

 determined to have been made, data on date and cause of death, coded according to the
 operative ICD code, were added to the occupational file.
    In the analysis, the 1971 sampled employees were excluded. Females were also excluded.
 These exclusions resulted in 415,201 records of males who were ever employed between 1965
 and 1969 for which 19,374 deaths between 1965 and 1973 were identified.
    Standardized mortality ratios were calculated for all causes, all cancers, and for 19 specific
 cancer sites. Expected rates were calculated by applying age-specific rates for a given cause
 for a reference group to the person-years of observation of the "exposed" group, grouped in
 15-year age groups. Two  reference groups were used, namely, the Canadian population and
 the whole occupational cohort.
    Overall, the sampled population exhibited lower total mortality (SMR=0.83) than the
 Canadian population as might be expected given the "healthy worker" effect. The depression
 in mortality was less strong for death from cancer (SMR=0.88), from lung cancer (0.95), and
 from bladder cancer (SMR=0.89).
    An individual was coded as belonging to a given occupation or industry if he had been
 recorded as being employed in the same for at least one year between 1965 and 1969. SMRs
 for major occupational groups were derived using the total cohort as the reference group.  For
 13 major occupational groups, the group for transport and communication workers is most
 likely to encounter electric and magnetic fields. In this subset, SMRs were elevated for 12 of
 the major cancer sites presented (SMRs were less than 1.0 for cancer of the pancreas and of
 the buccal cavity and pharynx, except lip), for all malignant neoplasms, and for all causes of
 death combined.  Of the elevated SMRs, the SMR for leukemia and aleukemia (SMR=1.68,
 p<0.01) was significantly elevated as were the SMRs for deaths from all malignant neoplasms
 combined (SMR=1.12, <0.05) and for deaths from all causes (SMR=1.04, p <0.05).
    For the 12 major industry groups examined, the industry of transportation, communication,
 and other utilities is the grouping that can be most readily linked to electric and magnetic field
 exposure. In this group, the potential excess risk is less clear than it was for the similar
 occupational grouping. Here, the SMRs were largely greater than one but no increase was
 statistically significant.  The SMR for leukemia and aleukemia was 1.36.
    For presentation of the more detailed occupational/job codes, the group with the greatest
 apparent potential for exposure to electric and magnetic fields was linemen and servicemen for
telephone, telegraph, and power systems. The data presented only gave SMRs for those sites
                                        3-85
10/22/90

-------
                           DRAFT-DO NOT QUOTE OR CITE

for which an observed increase differed significantly from unity.  For linemen and servicemen,
significantly elevated SMRs were seen for cancer of the intestine (except rectum) (SMR=3.53,
p ^0.01) and of the stomach (SMR=2.33, p<0.05). These sites differ from those observed in
other studies, i.e., are not leukemia or other neoplasms of the hematopoietic system.
Unfortunately, the SMRs for these cancers are not given but were apparently not statistically
significant.  In a review, Coleman and Beral (1988) cite an SMR for all leukemias of 2.41
(Cl=0.97-4.97) among these power and telephone lineman (N=4583); these data were
obtained from a personal communication with Howe.
    For presentation of detailed industrial codes, the group with the seemingly greatest
potential for EM-field exposure is manufacturers of major appliances (both electric and
nonelectric). Again, only data on sites with significantly elevated SMRs were given. For this
industrial group, a significantly increased risk of death from neoplasms of lymphatic and
hematopoietic tissues was observed (SMR=5.28, p <0.01).
    OHn et al. (1985) evaluated the mortality levels of 1254 electrical engineers who had
graduated from Sweden's Royal Institute of Technology between 1930 and 1979 with a
Master's degree. The study group was compared to the general Swedish  male population.
SMRs were derived for the engineers and for male architects who also had graduated from the
same school over the same period of time. The mortality of the electrical engineers was lower
than expected based on the general population for all causes and for all major disease
groups; a very strong healthy worker effect seemed to  be operative.  The observed mortality in
architects was also lower than  expected on the basis of general population rates.  For the
most recent graduates (1979),  follow-up was, at most, only about 5 years.  The earliest
graduates could have been followed up to about 55 years. Considering the differing follow-up
periods plus the effects of secular changes on mortality, it would have been helpful if analyses
had also been performed in different calendar times or birth cohorts.
    Vagero  et al. (1985) conducted  a retrospective cohort study on cancer mortality among
workers in one large Swedish telecommunications company, established in 1910,  involved in
research, development, and manufacturing of telecommunications equipment. Study subjects
were all workers employed  for  at least 6 months during 1956-1960 at three work sites. It is not
clear if these three sites were inclusive or only a subset of all places  or employees of the
telecommunications company. The subjects were ascertained from a company employee
registry maintained by management at the three sites and consisted  of 2918 subjects, 2051
       10/22/90
3-86

-------
                           DRAFT-DO NOT QUOTE OR CITE

men and 867 women. The follow-up period was from 1958 to 1979, resulting in 62,028
person-years (PY) at risk.
    Company files were used to elaborate work histories throughout an individual's period of
employment. Work histories included information on the departments in which a person
worked by month and year. A group of senior engineers developed a scheme to describe the
type of work or work practice in each department.  The developed work information was then
used to specify subcohorts formed by workers in departments with specific types of work. The
potential  exposures or work practices of interest to the investigators were soldering (N=1514,
PY=30,940); oil mist (PY=1787); trichloroethylene (PY=7202); and grinding, polishing, and
degreasing (PY=5366).
    Tracing efforts were very successful (2916 out of 2918 total subjects).  Cause of death
could not be established for two persons who were, thus, excluded from analyses of the
resultant  2914 persons. The four losses were all male subjects.
    The data on the study population were linked to information in the Swedish Cancer
Registry to identify cases of cancer between 1958 and 1979. SMRs (here, standardized
morbidity ratios) and 95% confidence intervals were calculated, and expected cases were
developed by applying age-, sex-, and calendar year-specific Swedish incidence rates.  Cases
were coded according the Seventh Revision of the ICD.
    There were 102 cases of all cancers among men, and 37 cases of all cancers among
women. The number of cases for both sexes was close to expected with both SMRs
essentially at unity. No excess of lung cancer was observed in either sex. For men, SMRs for
several types of cancer were elevated, but only the SMRs for malignant melanoma of the skin
(SMR=25) and Brill-Symmer's (SMR=17.6) (nodular lymphoma) were statistically significant.
For the latter, two cases were diagnosed in the  population whereas only 0.1 cases were
expected based on Swedish incidence rates. Among women, cases of cancer of the small
intestine (SMR=16.0), corpus uteri (SMR=2.2), and malignant melanoma of the skin
(SMR=2.8) were greater than expected although only the SMR for intestinal cancer was
statistically significant.
    Given the observed excesses in malignant melanoma for both sexes, further analyses were
performed for this type of cancer. All cases were confirmed by biopsy. When the sexes were
combined, the calculated SMR was 2.6 with a 95% confidence interval of 1.3-4.5. An analysis
of all cases that assumed at least 3 years of exposure (work) and a 10-year latency period was
                                        3-87
10/22/90

-------
                           DRAFT-DO NOT QUOTE OR CITE

performed. The assumptions yielded 10 cases (7 men, 3 women) for any work and 6 cases (4
men, 2 women) for workers involved with soldering. The overall SMR was 2.8 (95%
Cl=1.3-5.1), but the SMR for soldering work was 3.9 (95% Cl=1.4-8.5). The cases were
clustered at work site A.
   This is another Swedish study that used cancer registry data to investigate the risk of
cancer in telecommunications workers. The focus of the study was chemical exposures rather
than EM fields. The study seemed to have good information on employment histories.
Observed excesses in malignant melanoma in both sexes were concentrated among workers
at one location who were involved with soldering.  However, more detailed evaluation of
soldering jobs was not done. Exposures from soldering can involve several potentially
important agents, including EM fields. The excess risk of cancer of the small intestine in
women was not subjected to a more detailed work history analysis. The age distribution of the
cohort was not given in the paper.
   A study conducted by Barregard et al. (1985) investigated mortality and cancer incidence
among workers exposed to static magnetic fields (Table 3-29). The subjects were employed at
a Swedish chloralkali plant.  Chlorine is produced by electrolysis using a direct current of
about 100 kA which can create strong magnetic fields. This work presumably took place in
what was called the "cell room." As identified by foremen, workers in the cell room were
selected for study if they had worked there at least one year between 1951 and 1983. Also
included were electricians who spent at least 25% of their work time for at least one year over
1951 and 1983. There were 157 subjects. Mortality was assessed by developing observed vs.
expected ratios (SMRs although not so stated) using calendar-year and age-specific (5-year
age groups) mortality rates of Swedish males as the standard. Cancer incidence among the
workers was compared to the incidence expected based on rates for Swedish males and for
men in the county where the plant was located. An accounting for a latency period (5 or 10
years) was incorporated into the analyses and was calculated from the first year of exposure.
For the examination of incidence, analyses were made for work >1 year and  >5 years; these
durations do not seem reflect a range.
   No association of exposure with either mortality or incidence was observed.  The results
are shown in Table 3-29.  The obtained ratios are similar to what is usually expected if a
healthy worker effect is operative.  Cancer mortality was not examined; there were three
cancer deaths. The source of the mortality and incidence data was not stated, and it is not
       10/22/90
3-88

-------
                            DRAFT-DO NOT QUOTE OR CITE
   TABLE 3-29. PERSON-YEARS AND OBSERVED AND EXPECTED ALL-CAUSE MORTALITY OR CANCER INCIDENCE  BY
    LATENCY AND EXPOSURE DURATION FOR SELECTED SWEDISH MALE CHLORALKALI WORKERS, WITH EXPECTED
                      VALUES BASED ON NATIONAL AND REGIONAL RATES FOR MALES
Duration9
Mortality —
>5
Incidence >1
>5
Latency PY
> 5 2142
>10 1272
>5 1990
>10 1157
O
17
14
6
5
EN
22
18
8.6
7.0
EC
_
—
7.7
6.3
0/E
0.8 (0.4-1 .2)b
0.8 (0.4-1.3)
0.8 (0.3-1 .6)c
0.8 (0.3-1.9)
 °Duration and latency in years, PY=person-years, 0=observed, E=expected, EN=expected values based on national rates,
 ER=expected values based on county rates.
 b95% confidence interval.
 CO/E ratio computed here used expected values based on county rates.
 SOURCE:  Bar-regard et al., 1985.
clear if the subjects, including the cases, were still actively working.  It is stated there were no
losses to follow-up.
    Static magnetic fields were measured in the factory in 1984, taken at 10,110, and 170 cm
above the floor. The fields ranged from 4 to 29 millitesla (mT) or 40 to 290 gauss (G). The cell
room averaged 14 mT (140 G). The control or rectifier room averaged 6 mT. The authors
noted that the concern of most other studies has been fields from AC. The chloralkali plant
used DC, but the authors indicated that currents could be induced by the movement of
people, and these currents could possibly be stronger than those induced by AC magnetic
fields.
    Detailed data on duration of cell room work, beyond the summary data presented, were
not presented and might have been helpful in determining the extent of potential exposure.
For example, it would be interesting to know  how many men worked in the cell room only one
year versus how many worked the most number of years.  Follow-up in this study may also
have been quite short since it is possible, from the definition of the cohort, that a worker may
have been employed beginning in 1982 and followed for only a year. There was no
information provided as to what the distribution of the length of follow-up was, i.e., how many
were followed one year, how many followed two years, and so forth.  Although electricians
have been identified in other studies as a potential high-risk group, in the study it is possible
that the inclusion of electricians may dilute any response, if one exists, because they spent
less time in the cell room. It would have been useful to have compared the foremen's
determinations as to who worked in the cell room with actual personnel records. The
                                          3-89
10/22/90

-------
                           DRAFT-DO NOT QUOTE OR CITE

foremen's determinations are certainly open to subjective bias, particularly considering that
they were asked to recall for a 32-year period. Finally, the size of the cohort was extremely
small (N=157) to be able to detect any significant effects.
   Tornqvist et al. (1986) investigated the incidence of cancer among a cohort of electric
power linesmen and power station operators.  The Swedish Cancer-Environment Registry was
used. Occupational data were derived from the 1960 census. Record linkage with the Cancer
Registry covered the period 1961-1979 which is an extension of the calendar period reported
in the earlier study Wiklund et al. (1981).  Cancers were coded according to the Seventh
Revision of the ICD (ICD7).
   There were 3358 power linesmen and 6703 power station operators in the electric power
industry identified in the 1960 census. The workers were Swedish born men aged 20-64 at the
time of the census. There were 236 cases of cancer recorded in the Cancer-Environment
Registry among linesmen and 463 among the station operators between 1961 -1979. After
stratification in 5-year age  groups and for county, expected values were "based on the
cumulative incidence for all 986,408 men classified as blue collar workers by the census."
   SMRs were calculated separately for the linemen and the station operators.  Results for 13
cancer sites and  all sites combined (ICD7:140-204) were given. For linemen, SMRs were
elevated for stomach cancer (SMR=1.2, 90% Cl=0.8-1.7), prostate cancer (SMR=1.2, 90%
Cl=0.9-1.5), kidney cancer (SMR=1.3, 90% Cl 0.8-2.0), cancer of the urinary organs, excluding
kidney (SMR=1.2, 90% Cl=0.8-1.8), skin cancer excluding melanoma (SMR=1.5, 90%
CI=0.7-2.6), cancer of the nervous system (SMR=1.5, 90% CI=0.9-2.4), and leukemia
(SMR=1.3, 90% Cl=0.7-2.1).  For all sites combined, the risk was slightly elevated with an
SMR of 1.1  (90% Cl=1.0-1.2). All of these increases were small and for no site did the lower
limit of the 90% confidence interval exceed unity. The authors did not state why they used
90% confidence limits rather than the more typically applied 95% limits.
   SMR values were generally lower for power station operators.  Again no increased SMR
had a 90% confidence interval for which the lower limit exceeded unity, although the results for
urinary cancer were marginal. Increased SMRs were seen only for laryngeal cancer
(SMR=1.6, 90% CI=0.9-2.7), testicular cancer (SMR=1.8, 90% CI=0.8-3.6), kidney cancer
(SMR=1.3, 90% Cl=0.8-1.7), and cancer of the urinary organs excluding kidney (SMR=1.3,
90% Cl=1.0-1.7). The overall SMR was 1.0 (90% Cl=0.9-1.0).
       10/22/90
3-90

-------
                             DRAFT-DO NOT QUOTE OR CITE

    This study has certain methodologic limitations. The calculation or estimation of the
 cumulative incidence of blue collar workers, used to derive expected values, was not
 explained. Although there were more than 10,000 individuals in the cohort with 19 years of
 follow-up, the work classification was made as of 1960; hence, the duration of employment for
 these individuals is not known. The age distribution of the cohort is not known.  No attempts
 to examine issues pertaining to latency or dose-response were reported. For these reasons,
 this study neither confirms nor refutes the carcinogenicity of EM fields.
    In summary, this study failed to show any statistically significant excess risks, although
 SMRs were modestly elevated for cancer at several specific sites. Generally, more excesses
 were seen for linemen than for power station operators, although this may be a function of age
 or other factors which are unknown. Cancers of the nervous system and leukemias, among
 other sites, were nonsignificant^ elevated for linemen.
    Guberan  et al. (1989) examined disability, mortality, and cancer incidence among painters
 and electricians who lived in the Canton of Geneva in Switzerland. The original purpose of the
 study was to compare the incidence of neuropsychiatric disabilities  (known as "painters'
 syndrome") among painters and electricians, but the study was extended to investigate cancer
 mortality and incidence in both groups.  The discussion here will focus on the cancer results
 for electricians.
    The two work cohorts of men were identified from 1970 census files, traced through
 various registers, and included 1993 painters and 1992 electricians.  There were 77 painters
 and 44 electricians excluded from this base population either because they were foreigners
 not included in population records, they had died or emigrated before follow-up was initiated,
 or errors were found in records due to misclassification of occupation, duplications, and so
 forth. After these exclusions, the study cohort was  reduced to 1916  painters and 1948
 electricians.
   The electricians included electricity installers and repairmen, telephone installers, linemen
 and cable jointers, and car electricians.  Electrical and electronic fitters and assemblers as well
 as radio and television repairmen were not included.
   The 14-year follow-up period included the calendar years from January 1,1971, to
 December 31,1984. The authors did not present detailed data on follow-up.  The two cohorts
 had a large proportion of foreigners, i.e., non-Swiss workers. The nationalities comprising
these subgroups were not described.  There were 763 foreign painters (39.8% of the painter
                                         3-91
10/22/90

-------
                           DRAFT-DO NOT QUOTE OR CITE

cohort). Of these, 230 (30%) left Switzerland and were lost to follow-up. This loss represents
12% of the total painter cohort.  There were 649 foreign electricians (33.3% of the electrician
cohort). Of these, 174 (27%) left Switzerland and were lost to follow-up. This loss represents
8.9% of the total electrician cohort. Only 13 Swiss painters and 31 Swiss electricians
emigrated; all but one electrician could not be traced and were included in the study.
   Coded death certificates were  provided by the national office in Switzerland responsible for
vital statistics.  Coding was according to the Eighth Revision of the ICD. Cause of death for 10
deaths occurring abroad was obtained from various informants in  addition to death
certificates. Incident cases were identified from the General Cancer Registry.
   Expected deaths and expected cases of cancer were calculated from regional rates, e.g.,
the Canton of Geneva. Standardized mortality ratios and standardized incidence ratios (SIRs)
were developed. One-sided statistical significance was tested under the Poisson distribution.
Ninety percent confidence intervals were calculated  for SMRs and SIRs.  Use of 90% limits is
appropriate but somewhat atypical. Expected death rates and incidence rates in electricians
were calculated based on 24,807 and 23,387 person-years, respectively.
   Among electricians, no cause  of death, including all causes together and specific types of
cancer, was found to be significantly greater than expected. Deaths from several cancers
were, however, elevated, i.e., all malignant neoplasms (SMR=114), esophagus (SMR=176),
stomach (SMR=130), intestine/rectum (SMR=140),  pancreas (SMR=143), larynx (SMR=273),
brain (SMR=154), Hodgkin's disease (SMR=250), and leukemias (SMR=143). All of these
excesses were based on less than six cases. A similar pattern was seen for the calculated
SIRs which were generally lower than the SMRs. The incidence and mortality of lung cancer
were not elevated.
   Although this study failed to show any significant excesses in  cancer mortality or cancer
incidence, it has a number of methodologic limitations: (1) follow-up yielded 24,807
person-years of observation, which may not be sensitive enough to observe excesses in
lymphomas and brain cancers; (2) 14 years of foilow-up may not be long enough for the
latency of these cancers; (3)  no age distribution of the study cohort was provided, thus, it is
not known whether this cohort had more younger workers or otherwise; (4) there were no data
available on actual exposures. The occupation was identified from the 1970 census.
Therefore, the length of employment was not known.  The above factors preclude
dose-response or latency analyses. Lastly, given potential ethnic, national, or regional
       10/22/90
3-92

-------
                            DRAFT-DO NOT QUOTE OR CITE

 differences in mortality and incidence, the relatively high withdrawal rates among foreign
 workers, and use of Geneva rates as the standard schedule of rates for certain analyses, it is
 questionable whether the non-Swiss workers should have been included in the cohorts. The
 benefit of increased sample size could be offset by the introduction of biases. Additional
 analyses for  nationality subcohorts or controlling for nationality might have indicated whether
 there was a problem.  However, it is interesting to note that higher cancer risks found in other
 studies were also seen in excess in this study.
    Matanoski et al. (1989) examined cancer incidence among New York Telephone Company
 employees.  By limiting the population to those actively working, the study was designed to
 examine risks in younger persons; the mean age was 40 years. Cases were identified through
 the New York State Cancer Registry. Person-years were accumulated and  events were
 counted if there was confirmation that an individual was still employed during the period of
 follow-up, 1976-1980.  There were 50,582 male employees who accumulated 206,067
 person-years.
    Two standard schedules of rates were used to develop SIRs. Age-specific New York
 cancer rates  and cancer rates of nonline workers for (comparison to line workers) applied to
 person-years to determine expected numbers of cable splicers, central office technicians,
 outside plant technicians, and workers on installation, maintenance, and repair. Statistical
 significance was evaluated at the 95% level under the Poisson distribution.
    In the analysis that used rates of New York State males as the standard schedule of rates,
 the SIRs for all cancers combined for all men and across job types were less than unity, except
 for cable splicers (see Table 3-30). This is likely demonstrative of a healthy worker effect.
 Similarly, for all men, SIRs at specific cancer sites were less than 1.0, except for acute and
 unspecified leukemia,  considered together, and for breast cancer. There were only two cases
 at both sites,  but the result for breast cancer in males is of interest.  Excesses do occur among
 specific job categories, although none are statistically significant. Of the 16 cancer end points
 examined, all  but four-multiple myeloma, breast cancer, bladder cancer, and stomach
 cancer- were in excess for cable splicers. With respect to these four end points, there were
 no cases of multiple myeloma, stomach cancer, or breast cancer, and the one bladder cancer
case was less than expected.
   Cancer among certain types of workers was also examined with respect to rates in nonline
workers to serve as an internal control (see Table 3-31).  Such an analysis enhances
                                         3-93
10/22/90

-------
                            DRAFT-DO NOT QUOTE OR CITE
 TABLE 3-30.  STANDARDIZED INCIDENCE  RATIOS (AND NUMBER OF CASES) FOR SPECIRC CAUSES OF CANCER BY
 TYPE OF WORK AMONG MALE NEW YORK TELEPHONE EMPLOYEES: 1976-1980, STANDARD: NEW YORK STATE MALE
                                            RATES

All cancers
Gastrointestinal
Oral
Stomach
Colon
Lung
Leukemia
Lymphold
Acute & Unspecified
Chronic
Myeloid
Brain
Lymphomas
Multiple Myeloma
Breast
Prostate
Bladder
All
Men
0.81* (391)
0.85 (107)
0.66 (20)
0.58(9)
0.63(36)
0.76*(85)
0.77 (12)
0.87 (5)
1.16 (2)
0.75 (3)
0.85(7)
1.00(13)
0.79 (25)
0.69 (4)
1.44 (2)
0.85 (21)
0.66 (21)
All
Line
0.83* (265)
0.83 (68)
0.71 (14)
0.88(9)
0.77 (20)
0.87 (63)
0.88 (9)
1.34 (5)
1.76 (2)
1.15 (3)
0.74 (4)
0.70 (6)
0.86 (18)
1.07(4)
-(0)
1.06 (17)
0.53 (11)
Cable
Splicer
1.27 (40)
1.30(10)
1.03 (2)
-(0)
2.11 (5)
1.20 (8)
2.65 (3)
5.18 (2)
7.14 (1)
4.17 (1)
1.59 (1)
2.00 (2)
1.96(5)
-(0)
-(0)
2.45 (3)
0.51 (1)
Central
Office
0.88(96)
1.19 (24)
1.36(9)
1.13 (4)
0.65 (6)
0.75 (19)
0.58 (2)
0.79 (1)
-(0)
1.11 (D
0.56 (1)
1.42(4)
1.18(8)
0.78 (1)
6.48 (2)
1.45(9) '
0.70 (5)
Outside
Plant
0.86(9)
1.92 (5)
1.61 (1)
3.03 (1)
2.40 (2)
0.44 (1)
-(0)
-(0)
-(0)
-(0)
-(0)
-(0)
-(0)
8.40 (1)
-(0)
1.95(1)
-(0)
Install
Repair
0.67 (86)
0.60 (20)
0.12*(1)
0.97 (4)
0.50 (6)
0.78 (23)
0.96 (4)
1.32(2)
2.14 (1)
0.95(1)
0.90 (2)
-(0)
0.35 (3)
1.30 (2)
-(0)
0.49 (3)
0.36 (3)
"Statistically significant at the 95% level, two tail, Poisson distribution.
       10/22/90
3-94

-------
                             DRAFT-DO NOT QUOTE OR CITE
TABLE 3-31. STANDARDIZED INCIDENCE RATIOS (AND NUMBER OF CASES) FOR SPECIFIC CAUSES OF CANCER BY
   TYPE OF WORK AMONG MALE NEW YORK TELEPHONE EMPLOYEES: 1976-1980, STANDARD: NONLINE WORKERS

All cancers
All gastrointestinal
Oral
Stomach
Colon
Lung
Leukemia
Lymphoid
Myeloid
Brain
Lymphomas
Multiple myeloma
Breast
Prostate
Bladder
Cable
Splicer
1.81*(40)
1.76 (11)
1.38(2)
-(0)
2.23 (5)
2.16 (8)
7.00*(3)
#(2)
2.33 (1)
1.79 (2)
3.59*(5)
-(0)
- (0)
4.38 (3)
0.60 (1)
Centr al
Office
1.15 (96)
1.02 (28)
2.45*(9)
#(4)
0.54 (6)
1.26 (19)
1.07(2)
#(D
0.53 (1)
0.90 (4)
1.94 (8)
#(D
#(2)
3.48*(9)
0.78 (5)
Outside
Plant
1.15(9)
2.05 (5)
2.31 (1)
#0)
2.14 (2)
0.73 (1)
-(0)
-(0)
-(0)
-(0)
-(0)
#(1)
-(0)
4.54 (1)
-(0)
Install
Repair
0.91 (86)
0.84 (24)
0.20 (1)
#(4)
0.55 (6)
1.41 (23)
1.77 (4)
#(2)
0.89 (2)
-(0)
0.53 (3)
#(2)
-(0)
1.02(3)
0.40 (3)
      Statistically significant at the 95% level, two tail, Poisson distribution.
     # No cases expected based on nonline rates.
                                           3-95
10/22/90

-------
                            DRAFT-DO NOT QUOTE OR CITE
comparability and minimizes the healthy worker effect that can be seen in working populations
when compared to general populations. In this analysis, SIRs for cable splicers were elevated
across many sites. Total cancers, leukemias, and lymphomas were significantly in excess.
SIRs for installers and repairers were generally not elevated. Few cancers were observed for
workers whose work was classified as "outside plant." Central office workers also
demonstrated an excess of cancer at most sites, and oral cancer was significantly in excess.
The breast cancers were all found in this group of employees. SIRs could not be calculated
for breast cancer among central office workers because expected values were so low as to be
virtually zero.
    Exposure was assessed for the different job functions by making measurements over the
workshift of a sample of U.S. telephone workers.  Of the five job groupings examined, the
time-weighted average magnetic flux density was highest for cable splicers (4.3 mG or 0.43
/tT). This group, with the highest average  magnetic field exposures,  exhibited the highest risk
ratios. The second highest time-weighted  average magnetic field levels were found for central
office employees (2.7 mG or 0.27 /*T). Risks for various sites were also elevated in this group,
although leukemias were not found to be in excess. The mode of exposure is different in
these central office workers; that is, the type of work (call switch rooms) involved brief but
continuous exposure to high spiked fields  (Personal communication with G. Matanoski at
DOE/EPRI contractor's  review, November 1990).  The two  breast cancer cases found in the.
study were in this group.  Breast cancer is extremely rare for males.  It is interesting to note
that Stevens (1 987) hypothesized an increased risk for breast cancer with exposure to EM
fields, based on laboratory results pointing to effects on the pineal gland and melatonin
production.  Average exposures for the other types of jobs ("outside plant," installation and
repair, and miscellaneous [supervisor, clerk, etc.]) were similar (1.6-1.7 mG or 0.16-0.17//T).
    This study was well-designed, appropriately analyzed,  and presented improvements in
exposure assessment.  Exposure measures for types of work are given. Although the work
categories are somewhat broad, an apparent dose-response relationship was seen, i.e., risks
are highest among the groups most  highly exposed. Also, different modes for delivery of
exposure seem to produce differential risks which, if not yet confirmed, serve to illustrate the
difficulties in evaluating exposure- and dose-response characteristics of EM fields. Mortality
comparisons using both external (New York State cancer rates) and internal (New York
Telephone Company nonlinesmen) comparison groups.  Excesses were heightened with the
       10/22/90
3-96

-------
                            DRAFT-DO NOT QUOTE OR CITE

 internal comparisons. The study has a relatively short follow-up period which may not be
 critical, however, if EM fields have a promotional effect on cancer. There are small numbers in
 some instances. The primary weakness of the study is that potentially confounding exposures
 could not be evaluated.

 3.3.2.2.2. Cancer of the Hematopoietic System.  As part of an update of occupational
 mortality for Washington State, Milham (1982,1985a) noted excess deaths from leukemia
 among men who worked with electric and magnetic fields. PMRs, standardized by age and
 year of death, were calculated for Washington State male residents aged 20 years or more,
 from 1950 to 1979, and for 158 cause-of-death categories and 218 occupational classes. Ten
 occupational classes with presumed exposure to electric and magnetic fields exhibited
 elevated PMRs for all leukemia [Seventh Revision of the ICD (ICD7):204], ranging from 111 to
 259.  PMRs for three of the these occupations (electricians, power-station operators, and
 aluminum workers) were statistically significant based on expected values for Washington
 State white males for all leukemia (ICD7:204) and for acute leukemias (ICD7:204.3). Also
 presumed to have electromagnetic exposures, the welders and flame cutters had a PMR of 67.
    Wright et a!. (1982) evaluated leukemia cases in men with jobs involving exposure to
 electric and magnetic fields, using data from a population-based registry, the Cancer
 Surveillance Program, in  Los Angeles County, California. Occupational information was
 obtained at the time of diagnosis. Categories for this analysis followed those used by Milham.
 Proportional incidence rates (PIRs)  were computed for all leukemias, acute leukemias, and
 AML for white males in Los Angeles County from 1972 to 1979. There was a general trend for
 increased PIRs in all the 12 occupational groups used, including all jobs combined. For all
 groups, PIRs were highest for AML Statistically significant PIRs were found for power linemen
 (acute leukemia and AML), for telephone linemen (AML), and for all jobs combined (acute
 leukemia and AML).
    McDowall (1983) evaluated the proportionate mortality of males, age 15-74 in England and
Wales, using 1970-1972 data from a report on occupational mortality. Electrical occupations
were broken down into 10 occupational subgroupings. PMRs were computed for all
leukemias (ICD8):204-207), lymphocytic leukemias (ICD8:204), acute lymphocytic leukemias
(ICD8:204.0), myeloid leukemias (ICD8:205), and acute myeloid leukemias (ICD8:204.0).
Considering all electrical occupations together, the PMRs were not significantly elevated for
any disease group.  Expected values were based on all Welsh and English men aged 15 to 74
                                        3-97
10/22/90

-------
                           DRAFT-DO NOT QUOTE OR CITE

in 1970 to 1972,  But McDowall noted consistently elevated PMRs among electrical engineers
(noted as "so described"), telegraph radio operators, electrical engineers (professional), and
electronic engineers (professional).  The PMRs were most increased for all myeloid leukemia
and acute myeloid leukemia.  Lymphocytic leukemia was elevated in electricians.  The results
of statistical tests of significance or confidence intervals were not given.
   In the same paper, the results of a case-control study were also given (McDowall, 1983).
Cases were the 537 deaths occurring in England in 1973 among males age 15 or more with
AML given as the underlying case of death.  Controls were randomly selected from males age
15 year or older dying from all .causes of death, excluding leukemia, in 1973 and then matched
within 5 years of  age. Two controls were matched per case (537 cases, 1074 controls).
   Seven occupational groups were defined, including "all electrical occupations" and "all
electrical occupations plus men in any occupation engaged in an electrical or
telecommunication industry." These were identified from death entries rather than from the
occupational data used in the PMR analyses. The results were stated to be "relative risks"
rather than "odds ratios."
   These risks were elevated in all seven occupational groups and ranged from 1.6 to 4.0.
Except for the two summary classifications, the number of cases was small, resulting in broad
95% confidence  limits that spanned 1.0.  Confounding exposures within the various electrical
occupations were not examined. In addition, some individuals were classified as  belonging to
more than one occupational group, but it is not stated whether deaths were counted in more
than one group.  If true, the alternatives of limiting deaths to one unique job category could
substantially change the risk ratios.
   Coleman et al. (1983) evaluated the incidence of leukemia among men aged 15 to 74 in 10
types of electrical occupations in southeast England for the period 1961 -1979.  Data were
taken from the South Thames Cancer Registry, and "proportional registration ratios" (PRRs)
were calculated.  Job classification codes were based on hospital record data and were similar
to the groups analyzed by McDowall (1983).  Since only 88% of tumor reports could be
assigned a job code, presumably the remaining 12% were excluded from the study. Cancer
sites were coded according to the 8th Revision of the ICD and were all leukemias
(ICD8:204-207), acute lymphocytic leukemia (ICD8:204.0), chronic lymphocytic leukemia,
acute myeloid  leukemia (ICD8:205.0), and chronic myeloid leukemia. The chronic leukemias
included all cases specified as chronic, plus cases unspecified as acute or chronic. In a matrix
       10/22/90
3-98

-------
                            DRAFT-DO NOT QUOTE OR CITE

 of 40 calculated PRRs, the vast majority exceeded unity; only seven ratios were less than one,
 and seven PRRs were significantly elevated, including the ratio for all leukemias for all
 occupational groups combined. Within the occupational groups, no pattern or particular
 group stands out as evidencing increased incidence.
    Another study was conducted in New Zealand by Pearce et al. (1985). A case-control
 design was used and drew on Cancer Registry data - 546 male leukemia cases registered
 during 1979-1983 aged 20 years or more.  Four male controls per case were matched on age
 registration year to yield 2184 controls. Eight occupational groups with presumed exposure to
 electric and magnetic fields were defined.  The authors believed the number of cases by
 occupation (18) were too small to warrant investigation of specific leukemia cell types. Over all
 occupations, an excess of leukemia was seen, but it was not statistically significant (OR =
 1.70, 95% Cl = 0.97-2.97). A significant excess was seen, however, for electronic equipment
 assemblers (OR = 8.17) and radio and television  repairers (OR = 4.75).  In a later report
 (Pearce et al., 1989),  it was noted that these job categories had been miscoded and should
 have been labelled as radio and television repairers and electricians, respectively. There were
 no cases among electronic equipment assemblers.
    Calle and Savitz (1985) examined mortality from leukemia among 10 occupational groups,
 using data from the state of Wisconsin. The occupational groups were those used by Milham
 and Wright, e.g., electrical engineers, radio and telegraph operators, electricians, linemen
 (power and telephone), television and radio repairmen, motion-picture projectionists, streetcar
 and subway motormen, power station operators, and welders and flamecutters. The source of
 the occupational information was not stated; one  might assume that the information was taken
 from death certificates. Deaths were Wisconsin white men, 20 years of age or older, who had
 died between 1963 and 1978. Proportional mortality ratios were calculated for each
 occupation using all deaths from 1963 to 1978 among Wisconsin white males aged 20 years
 or more.
    Mortality from all leukemia [41 cases; Seventh Revision of the ICD (1CD7):204.3] was
 presented. Excess mortality was not consistently observed over all 10 occupational groups.
 No cases were found for several groups. For all leukemias, the highest PMR (2.35) was seen
for radio and telegraph operators, and this result was statistically significant.  The PMR for
electrical engineers (1.86) was also significantly elevated. Two other occupational groups
exhibited elevated PMRs that were not statistically significant. The two groups were linemen
                                         3-99
10/22/90

-------
                           DRAFT-DO NOT QUOTE OR CITE

and electronics technicians. For acute leukemia, again, the largest PMR (3.0) was seen for
radio and telegraph operators, but the result was not statistically significant. The PMR for
electrical engineers was also again elevated (2.57) and was statistically significant.  Mortality
for acute leukemia was elevated among linemen (1.43) and welders and flamecutters (1.04);
neither was statistically significant.
   In a case-control study, Gilman et al. (1985) evaluated leukemia in a population of 19,000
U.S. male coal miners. This population  of workers was identified as potentially exposed to EM
fields because power distribution lines are strung overhead in mines, stepdown transformers
and converters provide power (600-2000V) to mining equipment, and transportation is
provided by electrically-operated trolleys. The miner population database had previously been
studied by the National Institute for Occupational Safety and Health (NIOSH) and had been
divided into four cohorts based on  length of follow-up for mortality, e.g., 5,10, and  15 years,
and the fourth was an autopsy cohort.
   Forty cases of leukemia [Eighth Revision of the ICD (ICD8:204-207)] were identified among
6066 death certificates.  Each case (white males) was matched to four controls within the
same cohort who died of causes other than cancer or accidents.  Matching was on age at
death and year of birth (ą 3 years). No information on EM-field exposure was known. Length
of employment in underground mining was used as a surrogate of exposure.
EM-field-exposed workers were defined as those miners with underground employment of 25
years or more.  Miners who had worked less than 25 years were classified as unexposed.
Cigarette smoking and work history information had been collected via questionnaires when
the cohorts were initially ascertained. Although not clearly stated, it also appears that
radiological examinations for coal dust deposition in lungs had been performed to  define coal
work pneumoconiosis.  Odds ratios were calculated, and statistical significance was assessed
either by the Chi-Square or the Fishers  Exact Test.
   Excess risk for death from leukemia for multiple cell types was found for underground
mining employment of 25 years or  more, taken as a surrogate of potential EM-field exposure.
Of observed excesses, statistical significance was achieved for all leukemias  (ICD8:204-207;
OR=2.53), chronic leukemias (ICD8:204.1, 205.1, 206.1, 207.1; OR=8.22), chronic lymphocytic
leukemia (ICD8:204.1; OR=6.33), and myelogenous leukemia (ICD8:205; OR=4.74). No risk
was observed with respect to cigarette smoking or pneumoconiosis for all leukemias
       10/22/90
3-100

-------
                            DRAFT-DO NOT QUOTE OR CITE

considered together. The only risk estimate that was depressed was for acute lymphocytic
leukemia (ICD8:204.0; OR= 0.63); only two cases were identified.
    In an extension of an earlier investigation, Flodin et al. (1986) have reported the results of a
case-control study in Sweden that examined the incidence of acute myeloid leukemia in
relation to various types of exposure. The chief exposure of interest was background levels of
ionizing radiation. Cases were persons of Swedish ethnicity aged 20 to 70 years, diagnosed
with acute myeloid leukemia between 1977 and 1982 from five hospitals. In addition, the
subjects had to be alive in the catchment area of the diagnosing hospital and be alive and able
to answer a questionnaire; therefore, very ill or deceased cases were excluded. A total of 59
cases met the study criteria.
    Two sets of control subjects were elaborated. The first set was drawn from the general
population and matched (four controls per case) with respect to sex, age within 5 years, and
parish residence.  There were 236 control subjects in this set. The second  set was randomly
selected from the general populations of the hospitals' catchment  areas and was restricted to
adults aged 20 to 70 years. This series was comprised of 118 individuals (two controls per
case).
    Information about various exposures was obtained via a questionnaire that was apparently
self-administered.  The questionnaire addressed medical care, especially the use of drugs,
x-ray treatment, and x-ray examinations; occupational exposures; chemical and solvent
exposures; smoking habits; house building material to evaluate background ionizing radiation
exposure by means of a gamma radiation index; and other environmental and leisure time
exposures. Approaches to account for dose and/or duration of exposure were applied.
    The authors stated that risk estimates were similar with respect to the two comparison
groups; therefore, they presented the results of pooled analyses.
    An increased risk for acute myeloid  leukemia was seen for a history of long-term residence
or work in concrete buildings, x-ray exposure and radiological work.  No excess risk was noted
for pesticide exposure, cigarette smoking, engine exhausts, plastics or rubber chemicals, the
occupation of painting, and contact with cats or cattle. However, weak but  nonsignificant
increases in risk were observed for some ,of the above agents when long-term exposure or
latency was considered, i.e., 25 years of contact with cattle, at least 5 years of exposure to
motor gasoline, and a 25-year induction-latency for pesticide exposure.
                                        3-101
10/22/90

-------
                           DRAFT-DO NOT QUOTE OR CITE

   Of chief interest to this review are the results with respect to potential exposure to EM
fields.  Electrical workers were found to have an elevated risk for the development of acute
myeloid leukemia. Electrical work was defined according to Swedish occupational codes and
comprised the following subgroups:  electrical technicians (code 002), electrical workers (code
755), and computer and telephone mechanics (code 764). The logistic rate ratio was 3.8 for
electrical work with 8 cases and 14 controls.
   Concern has been expressed in the literature that electrical work can also involve work
with or exposure to chemicals, including solvents and the observed risks from EM fields either
may represent an interaction with chemicals or may represent  an indirect measure of a true
underlying risk posed by chemicals alone. In this study, separate analyses for chemical
exposure found a significantly elevated risk for exposure to styrene (minimal subjects,
however) but not to any other solvents. Use of psychopharmacological drugs also posed an
increased risk. Joint analyses of multiple  exposures, except with respect to gamma radiation,
were not reported.
   Stern et al. (1986) conducted a case-control study of leukemia mortality among on-shore
workers at the Portsmouth Naval Shipyard in New Hampshire.  The cases and controls were
drawn from a previously developed cohort of 24,545 white  males employed between January
1,1952, and August 15,1977. The goal of the study was to determine if there was an
association between death from leukemia and occupational exposure to ionizing radiation or
organic solvents (Table 3-32).
   Cases were defined as all persons who had died by the end of 1980 with an underlying or
contributory cause of death of leukemia. Cases were not included if medical records could
not confirm death certificate coding for leukemia. Controls were selected from the rest of the
work force cohort after excluding persons who had died from hematopoietic or lymphatic
malignancies. Controls were matched to cases.  Controls had to have lived as  long as the
cases.  Matching criteria were the absolute difference in years  between birth dates, dates first
employed, and duration of employment. The differences in these three values between cases
and potential controls were summed and used as a score.  The four persons with the lowest
score were selected as controls to a given case. There were 53 leukemia deaths and 212
matched controls. The authors stated that the 4:1 matching ratio yielded 80% power for
detecting a true relative risk of 2.2.
       10/22/90
3-102

-------
                            DRAFT-DO NOT QUOTE OR CITE
                     TABLE 3-32. BACKGROUND INFORMATION ON SHIPYARD WORKERS
                          AT PORTSMOUTH NA\AL SHIPYARD, NEW HAMPSHIRE

Birth year
Year employed
Duration of
employment (yrs)
Average number
of jobs

Mean
Range
Mean
Range
Mean
Range

Cases
1904
1892-1939
1940
1914-1963
22
0-47
1.96
Contr ols
1904
1892-1940
1940
1914-1963
22
0-46
2.08
  SOURCE: Stern et al., 1986.
    Job history information was coded for cases and controls. Shops and jobs were classified
and grouped by the commonality of processes, materials handled, and occupational
exposures.  Information was derived from shipyard staff, personnel records, a NIOSH industrial
hygiene survey, and historical industrial hygiene data from the shipyard. Both shop and job
codes were developed per work assignment per person. There was no exposure monitoring
data for individuals except radiation film badges for certain workers.
    Mantel-Haenszel odds ratios were calculated for the evaluated exposure groups. All
leukemias and specific cell type were examined. To control for potentially confounding
exposures and to evaluate interactions, a conditional logistic regression model was also
employed.
    In the Mantel-Haenszel or univariate analyses for ionizing radiation and (likely) solvent
exposures, certain odds ratios were elevated, especially for solvent exposure, but no estimate
was statistically significant.  If at least three cases had ever worked in a given job or shop,
those job types and shops were also examined, with comparisons made to employees never
in the subject job or shop. In these analyses when all leukemias were considered together,
several job types and shops yielded odds ratios greater than 1.0, but the largest and only
significant odds ratios were  obtained for the job of electrician and work in the electrical shop.
The effect was not as strong when myeloid leukemia and lymphatic leukemia were examined
separately but the estimated risks remained elevated.  A statistically significant odds ratio
(13.0) for lymphatic leukemia for the class of supervisors, regardless of job, was observed.
                                        3-103
10/22/90

-------
                            DRAFT-DO NOT QUOTE OR CITE

The odds ratios for myeloid leukemia (1.53) and for all leukemia combined (2.36) were
elevated, but not significantly. The authors noted that 5 of the 13 supervisors who died of
leukemia had at some time worked as electricians or welders. (Elevated odds ratios were also
observed for welders and the welding shop.) Tables 3-33 and 3-34 give partial results for
some of the jobs and shops that exhibited the highest risk estimates.
   The data were also evaluated using conditional  logistic regression analyses.  Based on the
previous analyses, electrical and welding jobs were included in the model. Again, results for
ionizing radiation and solvent exposures were not strong. In analysis of categorical variables,
the jobs of electrician and welder again demonstrated significantly elevated odds ratios.
Exposure was also examined as a continuous variable using years as a welder or electrician.
Similar results were obtained although the magnitude of the risk estimates were not as large
as in the categorical analysis and the odds ratios for electricians only were significant for all
leukemias combined.
   Many jobs and shops were evaluated and so the results for electricians and welders could
be a chance occurrence.  Yet, the results were similar and consistent when examined from the
two perspectives of job type versus shop and across all methods of analysis.
   This study had well-developed work history information derived from various sources
although no measurement or exposure data were available except for ionizing radiation film
badges.  The work history information enabled evaluations of various types of workplace
exposures by job or shop, of potential confounders, and of interactions. While different jobs
may have multiple exposures that cannot be fully isolated, the analyses point to significantly
elevated risk of leukemia for electricians and electrical work and, to a lesser extent, welders.
Where predominating work with solvents could be identified, the risks were elevated but not
significantly, and the excesses were usually less pronounced than those for electricians and
welders.
   Another study has taken advantage of the information available from Sweden's system that
links cancer incidence data from the National Swedish Cancer Registry with census data,
including employment information. Incidence data for 1961 -1979 is linked to occupational
data from the 1960 census.  Linet et al. (1988) used the Swedish Cancer-Environment Registry
(CER) to investigate leukemia and occupation (Table 3-35). The CER uses the Seventh
Revision of the ICD. For purpose sof this analysis, the leukemia cases were receded into the
currently recognized categories of acute lymphocytic (ALL), chronic lymphocytic (CLL), acute
       10/22/90
3-104

-------
                                      DRAFT-DO NOT QUOTE OR CITE
                    TABLE 3-33. RESULTS OF MANTEL-HAENSZEL ANALYSIS OF LEUKEMIA FOR SELECTED
                      JOBS AND SHOPS (WITH THE HIGHEST RISK ESTIMATES) FOR SHIPYARD WORKERS
                               All Leukemias
                           Cases   OR      95%C1
                                       Myeloid Leukemia
                                    Cases   OR       95%C!
                                                            Lymphatic Leukemia
                                                          Cases   OR      95%C1
Jobs Ever Held

Electrician
Carpenter
Supervisor
Welder
Sheetmetal
Shipfitter
Engineer
      11
       7
      13
       7
       4
      11
       6
  3.00*
  2.50
  2.36
  2.25
  2.14
  1.54
  1.40
1.29-6.98
0.91-6.90
0.95-6.90
0.92-5.53
0.64-7.19
0.67-3.54
0.53-3.70
6
4
8
6
2
8
3
2.33
2.50
1.53
3.83*
3.50
1.71
1.00
0.77-7.06
0.71-8.83
0.54-4.40
1.28-11.46
0.49-24.96
0.62-4.72
5
3
5
0
2
3
3
 6.00*
 3.33
13.00*

 2.00
 2.00
 2.20
1.47-24.45
0.56-19.83
1.31-28.96

0.41-9.83
0.41-9.76
0.51-9.44
Shops Ever Worked
Electrician
Welding
Public Works
Sheetmetal
Shipfitting
Woodworking
Rigging
Electrical
10
7
3
4
14
4
6
7
2.57*
2.25
2.40
2.08
1.53
1.30
1.00
0.89
1.11-5.96
0.92-5.53
0.60-9.61
0.64-7.19
0.71-3.31
0.037-4.41
—
0.37-2.11
5
6
1
2
8
2
1
4
2.12
3.83*
4.00
3.50
0.21
1.00
0.22
1.44
0.64-7.10
1.28-11.46
0.31-51.79
0.49-24.69
0.46-3.18
—
0.02-1.75
0.38-5.55
5
0
2
2
4
2
4
2
3.80*
—
2.67
2.00
1.83
3.00
2.60
0.40
1.13-12.76
—
0.48-14.89
0.35-12.35
0.39-8.54
0.27-33.35
0.57-11.86
0.10-1.53
*=Significant.
OR=odds ratio, Cl=confidence interval.  There were 11 job types and 14 shops listed in the paper.

SOURCE:  Stern et al., 1986.
                 TABLE 3-34.  RESULTS FROM CONDITIONAL  LOGISTIC REGRESSION ANALYSIS FOR SOLVENT
            ELECTRICAL,  AND WELDING WORK, TREATED EITHER AS A CATEGORICAL  OR CONTINUOUS VARIABLE
                            (YEARS USED AS INDEX OF EXPOSURE), AMONG SHIPYARD WORKERS
Exposure
Categorical a:
Solvent job
Electrician
Welder
All leukemias
OR 95% Cl

2.32 0.85-6.29
3.39* 1.40-8.18
3.19* 1.09-9.37
Myeloid leukemias
OR 95% Cl

2.56 . 0.58-11.35
3.03 0.92-10.03
6.23* 1.64-23.64
Lymphatic leukemia
OR 95% Cl

1.99 0.46-8.67
6.11* 1.38-27.10
Continuous b:

Yrs in solvent job
Yrs as electrician
Yrs as welder
1.82
1.67*
2.86*
0.93-3.58
1.01-2.78
1.02-8.04
          2.16
          1.57
          5.53*
  0.79-5.95
  0.83-2.96
  1.52-20.09
                   1.53
                   1.90
                      0.56-4.22
                      0.89-4.07
aConsidered  if "ever worked" in these jobs.
bCalculated from average "exposure." Averages were 8.59 years for solvent jobs, 6.46 years for electricians, and 13.16 years for welders.
*=significant. OR=odds ratios, Cl=confidence interval.


SOURCE:  Stern et al., 1986
                                                      3-105
                                                                               10/22/90

-------
                                DRAFT-DO NOT QUOTE OR CITE
   TABLE 3-35. SELECTED STANDARDIZED INCIDENCE RATIOS (SIRs) FOR LEUKEMIAS BY MAJOR DIVISION OF INDUSTRY AND
   OCCUPATION, BY GENERAL MANUFACTURING INDUSTRIES, AND BY GENERAL CRAFTSMEN-TRADESMEN  OCCUPATIONS,
                                    AMONG SWEDISH MALES 1961-1979
ALL CLL ANLL CML
Cases SIR Cases SIR Cases SIR Cases SIR
Industry;
Electric, gas,
water, and
sanitary services
Transportation
and communications
Occupation;
Transport and
communication
workers
Manufacturing
Industry:
Machine and
electrical
Craftsmen-Tradesmen
Occupation:
Toolmakers,
machinists,
plumbers, and
welders
Electrical workers
Electrical line
workers

1 0.4 44 1.3 26 1.0 14 1.0
16 1.3 151 0.9 141 1.1 84 1.1

12 1.0 126 0.9 129 1.0 78 1.0

13 1.1 164 1.0 108 0.8 68 0.9

23 1.5 178 1.0 153 1.0 100 1.0
4 1.0 52 1.2 42 1.1 27 1.1
NG NG 13 1.9a 8 1.4 NG NG
ap < 0.05
NG Ť not given.
SOURCE: Unet et al., 1988.
          10/22/90
3-106

-------
                           DRAFT-DO NOT QUOTE OR CITE

nonlymphocytic (ANLL), and chronic myelocytic (CML).  Standardized cumulative incidence
ratios were computed. Expected values were derived by applying 5-year birth-cohort, sex-,
region-, and type-specific leukemia incidence rates for Swedish males for 1961 -1979 to 5-year
birth cohort-and sex-specific distributions of industrial and occupational groups. The 1960
Swedish census employment classifications, which followed international standards of the
United Nations International Labor Office, were used to define industrial and occupational
groups.
   There were 5351 leukemia cases identified (ALL=3%, CLL=45%, ANLL=33%, and
CML=19% of the total cases.) Although many results were presented, only the results of
possible pertinence to this document will  be discussed.  It was found that electrical line
workers had a significantly increased risk of CLL (OR=1.9,13 cases). The authors noted that
this result differs from other studies in that other studies had reported an excess risk from
ANLL, not CLL. Linet et al. (1988) also found ANLL to be elevated among electrical line
workers (OR=1.4, 8 cases),  but the excess was not statistically significant.  The authors noted
that, in Sweden, electrical line workers have exposure to many solvents and other chemicals
such as creosote, lead, isocyanates, and silicon.  Significantly increased risk was not seen in
the broader occupational category of electrical workers but all odds ratios were 1.0 or greater.
   The authors pointed out recent data indicating that underreporting of leukemia cases to
the CER may be 18%. They  also noted that the lack of significant association in the study by
Tornqvist et al. (1986) may be due to restricting the study population to workers 20 to 64 years
of age; CLL incidence is highest in the oldest age groups.  TOrnqvist et al. (1986) also used the
CER.
   Juutilainen et al. (1988) examined cancer incidence in Finnish males employed in electrical
occupations.  Data were taken from  the Finnish Cancer Registry and the Central Statistical
Office of Finland.  The eight  occupational groupings examined were electricians (indoor
installation), electric fitters and repairmen, telephone installers and repairmen, linemen and
cable jointers, electric and electronic equipment assemblers, and the residual, other electrical
occupations.  Standardized  incidence ratios (SIRs) were calculated using rates for all Finnish
males as the standard schedule of rates.  The authors stated that other occupational groups,
similar in socioeconomic status to electrical occupations, were also as a reference, but no
information on these groups was presented in the paper. Only results for leukemia were
                                         3-107
10/22/90

-------
                            DRAFT-DO NOT QUOTE OR CITE

reported.  Neither the applicable ICD codes nor the time period under study were stated. This
study seems to have been reasonably designed but is poorly reported.
    Only 17 leukemia cases among all electrical occupations were identified, and the SIR was
1.23; confidence limits or statistical test results were not given.  Bearing in mind that the
observed number of cases is very small, the SIRs for specific electrical occupations were
generally elevated, or else there were no observed cases. The highest SIR (3.13) was
determined for linemen and cable jointers, considered together.  There were four cases in this
group. The residual grouping of all other electrical occupations yielded an SIR of 2.63, but
there was only one case.
    Pearce et al. (1989) examined the incidence of cancer, especially leukemias, among New
Zealand electrical workers in a case-control study using a population-based cancer registry.
Cases were all male patients aged 20 years or older at the time of their registration in the New
Zealand Cancer Registry. For each person, the current or most recent occupation was
described in Registry files and coded according to the New Zealand Standard Classification of
Occupations.  Analyses were restricted to the 19,904 eligible registrants for whom there was
occupational information. This number is 80% of all eligible registrants (N=24,762). This
population overlaps with a previously studied population which was registered between 1979
and 1983 (Pearce et al., 1985). The percent of cases for whom occupation was available
ranged from 73% to 92%, but there did not seem to be any unusual underreporting of
occupation by cancer site. Reporting of occupational information was typically available for
80% to 84% of the cases within specific cancer types.  Occupational information was available
for 82% of all leukemia cases.  The occupations examined were electrical and electronic
engineers, electrical and electronic technicians, electrical fitters, electronic equipment
assemblers, radio and television repairers, electricians, telephone installers, linemen,  and
power station operators.
    Controls were males within the population of 19,904 who had a diagnosis of any cancer
site other than the particular site under test. This study, thus, represents a series of many
case-control studies conducted within the Registry population of adult males. The authors
stated that age-adjusted odds ratios were estimated by the Mantel-Haenszel method and
confidence intervals were calculated using Miettinen's approximate method.
    Odds ratios were elevated for  all electrical workers, considered together, for several cancer
sites. These sites were buccal cavity and pharynx (OR=1.25), stomach (OR=1.06), colon
       10/22/90
3-108

-------
                           DRAFT-DO NOT QUOTE OR CITE

(OR=1.06), rectum (OR=1.16), liver (OR=1.52), larynx (OR=1.23), soft tissue sarcoma
(OR=1.26), urinary bladder (OR=1.00), brain (OR=1.01), leukemia (OR=1.62), and a residual
category of miscellaneous sites (OR=1.28).  The odds ratio for leukemia was the only risk
estimate for which the 95% confidence interval exceeded unity (1.04-2.52). Brain cancer was
not significantly in excess in this composite group.
   More specific types of electrical jobs, previously mentioned, were also examined with
respect to leukemia (see Table 3-36). The odds ratios were elevated for all specific job types
except equipment assemblers (no cases), telephone installers (no cases), and a residual
category.  The 95% confidence  interval encompassed unity for the odds ratios for
radio/television repairers and, marginally, for power station operators. The jobs of motion
picture projectionists and welders were not specifically identified in the tables, but the authors
noted that there was one case of leukemia among seven projectionists (OR=7.8, 95%
Cl=1.23-49.46) and no cases among 44 welders. Brain cancer was not elevated in the total
group of electrical workers  but was elevated among electrical engineers (OR=4.74, 95%
01=1.65-13.63) and electricians (OR=1.91, 95% 01=0.84-4.33).
   Risks for specific types  of leukemia within two broad age groups (20-64 years and 65 years
or older) were also examined (see Table 3-37).  It should be cautioned that numbers are small.
For the older age group, odds ratios were elevated for all specific leukemia types, but not
significantly. But the 95% confidence interval for the summary odds ratio exceeded unity
(OR=1.85, 95% Cl=1.03-3.32).  The results are somewhat more variable for the younger age
   TABLE 3-36. AGE-ADJUSTED ODDS RATIOS FOR LEUKEMIA BY TYPE OF ELECTRICAL WORK AMONG NEW ZEALAND
                                MALES AGED 20 YEARS OR OLDER

Type of
Electrical Job
Engineers
Technicians
Fitters
Equipment assemblers
Radio/TV repair
Electricians
Telephone installers
Linemen
Power station operators
Other
Total

Exposed
Cases
2
1
1
0
2
6
0
5
2
2
21

Exposed
Controls
47
34
23
9
. 10
125
18
77
20
104
467

Odds
Ratio
1.54
1.03
1.56
7.86
1.68
2.35
3.89
0.70
1.62
95%
Confidence
Interval
0.38-6.27
0.14-7.32
0.21-11.41
2.20-28.09
0.75-3.79
0.97-5.70
1.00-15.22
0.18-2.78
1.04-2.52
 SOURCE:  Pearce et al., 1989.
                                        3-109
10/22/90

-------
                            DRAFT-DO NOT QUOTE OR CITE
         TABLE 3-37. AGE-ADJUSTED ODDS RATIOS FOR LEUKEMIA AMONG NEW ZEALAND MALE ELECTRICAL
                             WORKERS, BY AGE AND LEUKEMIA SUBTYPE
Age 20-64 Years


Acute lymphatic (ICD 204.0)
Chronic lymphatic (ICD 204.1)
Acute myelold (ICD 205.0)
Chronic myeloid (ICD 205.1)
Other
Total

Cases
0
4
3
1
1
9

OR

3.36
1.21
0.88
1.73
1.39
95%
Cl

1.27-8.89
0.38-3.85
0.12-6.38
0.24-12.67
0.71-2.71
Age >65 Years

Cases
1
4
2
2
3
12

OR
3.05
1.67
1.09
2.55
2.65
1.85
95%
Cl
0.44-20.97
0.62-4.53
0.27-4.44
0.64-10.10
0.86-8.18
1.03-3.32
  OR-odds ratio, CNconfidence interval.
  SOURCE: Pearce et al., 1989
 group, except the odds ratio for chronic lymphatic leukemia (CLL) was significantly elevated
 (OR=3.36, 95% Cl=1.27-8.89).  CLL is most typically a disease of older persons.  CLL is
 extremely rare under age 30; rates slowly increase between ages 30 and 50, and rates
 exponentially rise in older age groups (Linet, 1985).
    These results are generally consistent with previous studies, although risks are greater in
 older age groups, except for CLL.  Use of other cancer cases as  controls can present
 selection bias if exposure relates to other cancer sites. The authors did not believe any
 potential bias would be a major consideration because, based on other analyses of New
 Zealand electrical workers, age-adjusted total cancer mortality was 8% lower than for other
 employed males.

 3.3.2.2.3. Cancer of the Nervous System. Preston-Martin et al. (1982) reviewed the
 characteristics of 3215 cases of central nervous system (CMS) neoplasms diagnosed from
 1972 to 1977 and registered in the Los Angeles County (CA) Cancer Surveillance Program.
The Registry contains  information on age, sex, race, address, religion, birthplace, occupation
 and industry, and pathological diagnosis abstracted from hospital records. More than 95% of
the diagnoses are histologically confirmed.  For descriptive analyses, the authors calculated
 Incidence rates using the direct method for  10-year age groups with the U.S. 1970 population
taken as the standard. Employment of the cases at diagnoses was coded according to the
1970 U.S. Census Occupational Classification System. The analysis of industry and
occupation considered cases among white  males age 25 to 64 years, and PIRs were
       10/22/90
3-110

-------
                           DRAFT-DO NOT QUOTE OR CITE

calculated as the ratio of the number of observed cases to the number of cases expected in a
subgroup. Statistical significance was measured by use of a summary Mantel-Haenszel
chi-square test. Results were presented for various characteristics, but only the occupational
results are discussed here.
   The distribution of all CMS neoplasms, considered together, by industry of employment at
diagnosis was examined.  The industries for which at least five cases were seen, and the PIR
was significantly elevated were: beverage, lumber and wood products, apparel, printing and
publishing (excluding newspaper), transportation equipment, colleges and universities,
insurance, and petroleum  refining.  Of these industries, the lowest PIR was for the petroleum
refining industry (1.11,8 cases).  The highest PIR (2.63, 7 cases) was seen in the beverage
industry.  Most cases (N=85) were employed in the transportation equipment industry, and the
aircraft industry accounted for 73 of the 85 cases.  The PIRs for the transportation equipment
industry and its aircraft subset were 1.40 and 1.46, respectively.  For analyses by cell type (not
tabulated for presentation in the paper), the PIRs for both gliomas and meningiomas were also
significantly elevated in these industries.
   The job title at diagnosis of the cases was also examined. Two groups which contained at
least five cases and demonstrated significantly increased PIRs were electricians (PIR=1.42,11
cases) and engineers (PIR=1.28, 47 cases).  Engineers were noted to have been found to be
the predominant job title in the aircraft industry, accounting for 38% of those cases. The
authors stated that the excess incidence they observed for electricians corresponds to the
excess mortality among electricians and electronic workers observed in Britain (Registrar
General,  1971).
   Lin et al. (1985) evaluated a possible association between occupational exposure to EM
fields and death from brain tumors in white male Maryland residents who died between 1969
and 1982. Although 1043 cases were originally identified by ICD codes on death certificates,
analyses were based on the recorded histologic type to yield 951 cases — 370 gliomas, 149
astrocytomas, and 432 nonspecified brain tumors. Tumors known to be from metastasis were
excluded; nonspecified tumors were analyzed separately because of the possibility of
secondary tumors occurring in this group.  Occupations were those recorded on the death
certificates.
   The cases were evaluated in several different ways. Seventy-eight of the cases had
electrical or electronic or utility occupations.  As a  preliminary step, the number of deaths in
                                         3-111
10/22/90

-------
                            DRAFT-DO NOT QUOTE OR CITE

the three electrically-related occupational groups was compared to the number expected
based on the proportion of these occupations reported in the 1970 Maryland population
census.  In all comparisons, observed deaths always exceeded the number of deaths
expected for these occupations. Considering gliomas and astrocytomas together, the
expected values differed significantly from observed for all three job groups, separately and in
combination. For nonspecified brain tumors, significance was achieved only for the group that
Included electric or electronic engineering and technician occupations and for all occupations
in combination.
    Because of probable problems in the validity of the above approach, a case-control study
was also conducted and reported in the same paper. The cases were the same, i.e., 951 white
male Maryland residents who died of brain tumor between 1969 and 1982. Controls were
selected from a 10% computer-randomized roster of Maryland decedents.  Death certificates
were obtained for  white adult Maryland  residents who died of causes other than cancer.
Controls were matched to cases for age and date of death. There were equal numbers of
cases and controls. The report does not state whether the controls were only men. If not,  the
inclusion of women could seriously misrepresent the distribution of controls according to type
of occupation.
   The distribution of types of occupations was compared in glioma and astrocytoma cases
and controls. Occupations with potential exposure to electromagnetic fields occurred more
frequently in the case group. This was not seen for nonspecified brain tumors.  There were
fewer cases than controls in composite  groups, namely, "other occupations" and "unknown
occupation." The  only significant excess in glioma and astrocytoma cases compared to
controls occurred  for one job class, electrician or electronic engineer or technician
occupations.
   In a third approach, the occupations were classed according to the level of potential
exposure to electromagnetic fields. A panel of experts defined what kinds of occupations
would fall in the classes. These groups were definite, probable, possible, or none.  Cases and
controls were blindly assigned to the exposure categories. The calculated odds ratios for
gliomas and astrocytomas demonstrated a gradient across the exposure categories; the odds
ratios increased as exposure potential increased.  While measured or typical exposures are
not available for these broad job categories, these results provide some limited evidence of a
potential dose-response relationship. Odds ratios for the highest class —  definite EM
       10/22/90
3-112

-------
                           DRAFT-DO NOT QUOTE OR CITE

exposure and for the third group - possible EM exposure were statistically significant. The
results with nonspecified brain tumors were essentially negative.
   Lastly, the mean age at death was examined for the cases by EM exposure category
versus no exposure. Only one difference was noted. Glioma and astrocytoma cases classed
as having definite EM exposure died at significantly younger ages than did cases classed as
without EM exposure.
   Thomas et al. (1987) conducted a case-control study of deaths from brain tumors in
northern New Jersey; Philadelphia, Pennsylvania, and surrounding counties; and southern
Louisiana to investigate occupational risk factors. There were 435 cases and 386 controls;
both were identified from death certificates of residents in the three geographic areas.  Cases
were white  males age 30 years or older who died of brain or other central nervous system
tumors between January 1,1978, and December 31,1981.  Diagnoses were verified by
reviewing hospital records. One control  per case was selected from men who died of causes
other than brain tumor, stroke, suicide, or homicide.  Cases and controls were matched on
age at death, year of death, and area of usual residence.  Job histories for the study subjects
were obtained by interviewing next-of-kin; the response rate for cases was 74% and 63% for
controls. How eliminations occurred and how they were handled to result in 435 cases and
386 controls were not discussed in the report. Risk ratios were reported as relative risks  in the
paper, but are referred to as odds ratios here.
   Two methods for classifying exposure to EM radiation were used.  The first method defined
occupations with exposure to EM radiation after the classifications used by Lin and Milham.
These jobs were electronics and telecommunications engineers; electronics technicians and
teachers; radio, radar, and telegraph operators;  electricians; electrical linemen; electrical and
electronics equipment repairmen; aluminum production workers; welders; and motion picture
projectionists. Men never employed in these jobs were considered unexposed. The second
method involved classification by an industrial hygienist of jobs with presumed exposure to EM
radiation, with exposure to lead, and with exposure to soldering fumes.  This method
represented an attempt to sort out exposures to other agents also associated with
occupations involving EM radiation. Unexposed men were those classified as never working
with any of the three agents.
    Under the first method for classifying exposure, the odds ratio for subjects who ever had a
job with EM exposure was 1.6, but it was not significant. Significant excess risk (OR=2.3) was
                                         3-113
10/22/90

-------
                             DRAFT-DO NOT QUOTE OR CITE

 seen in the subset of subjects who had worked in a job involving the design, manufacture,
 installation, or maintenance of electrical or electronic equipment.
    The occupations were further evaluated by partitioning the jobs into two classes: (1) jobs
 involving the manufacture and repair of electronic equipment, including engineers, teachers,
 technicians, repairers, and assemblers and (2) jobs in the electrical trades, including
 electricians, power linemen and servicemen, and telephone linemen and servicemen. The
 former group was considered to have had mixed exposure to EM radiation and solvents and
 fumes from soldering. Excess risk was reported for the first class of jobs with a significant
 odds ratio of 3.9, as compared to an odds ratio of 1.9 in the tradesmen group.  Furthermore,
 the risk was concentrated among the electronics equipment repairers.
    Significantly increased risk (OR=1.7) for death from brain tumors was observed in
 occupations that had involved EM-radiation exposure, as classified under the second method.
 However, when subjects who also had exposure to lead and soldering fumes were removed
 from the analysis, the odds ratio dropped to 1.4 and was not significant.  Removal of additional
 subjects, those who had also worked with organic solvents, resulted in further reductions  in
 risk, with the odds ratio falling to 0.4. The methods for classifying jobs on the basis of
 exposure to organic solvents was not discussed in the report.  The results seem to indicate
 that the risk of death from brain tumors relates more to the other agents associated with jobs
 involving EM radiation, than to EM radiation alone.
    The excess risk associated with electronics manufacture and repair jobs was due to
 excess deaths from astrocytic tumors. When exposures to lead and soldering fumes in any
 job were evaluated separately, no risk was found for lead exposures.
    This study attempted to separate exposures to various agents seen in occupations
 involving work with EM-radiation sources and systems. However, the authors stated that,
 although their data showed elevated brain tumor mortality  risk among men occupationally
 exposed to soldering fumes, the effects of all the agents associated with electronics jobs still
 could not be separately evaluated. It is possible that electrical and electronic occupations may
 present risks that result from an interaction of exposure to various potential  carcinogenic
 agents, but further refinements in exposure assessment, if  possible, will be required to address
this question.
   In their report, Thomas et al. (1987) refer to exposure to EM radiation as exposure to
microwave and RF radiation. This may be misleading because the radiation frequencies
       10/22/90
3-114

-------
                           DRAFT-DO NOT QUOTE OR CITE

involved in the evaluated jobs probably encompass many portions of the spectrum. When
electrical tradesmen were separated, it was noted exposure would be to ELF radiation.  But
the other jobs may have involved exposures to emissions or modulations to frequencies other
than the primary or carrier frequency of a given set of equipment. In any event, the exact
frequency of radiation exposure is not always obvious for broad occupational classes.
   Speers et a!. (1988) conducted a case-control study of occupational exposures and brain
cancer among east Texas residents. Persons who had died of primary malignant brain tumor
(glioma, ICD8:191) between 1969 and 1978 were identified from death certificates of residents
in 40 counties in east Texas. The number of cases ascertained and included in the study
totaled 440; 202 male cases were the subject of analyses. All were white and between 35 and
79 years of age. One control was selected per case. The next death certificate listed in State
numerical files was selected as a control if the same criteria  of race, east Texas residence, age
35-79, and time period of death were met. Deaths from brain tumors were excluded as
controls. Male controls totaled 238. Descriptive information abstracted from death certificates
included usual occupation and industry which was coded according to the 1980 Census
classification system.
   Deaths were examined deaths in broad industry categories, classified according to the
system used by the U.S.  Bureau of Census.  Industrial subcategories and certain occupation
were also examined. Odds ratios and their confidence intervals were computed. A given
"exposed" industry was compared to the total sample (all other industries combined) less the
industry then under examination. Significantly increased risk of death (OR=2.26) from glioma
was observed in the broad census group of transportation, communication, and utility
employees. Although not stated in the report, this category  presumably represents a broad
industry grouping  that most likely contains occupations with potential exposure to EM fields.
The odds ratio for utility employees considered alone was 13.10 and was statistically
significant. The risk among workers employed in occupations (including electricians and
electronic workers) associated with electricity or EM fields was elevated (OR=2.11), but the
increase was not significant. It is not clear whether the  persons identified with these
occupations were  all drawn from the broad industrial category of transportation,
communication, and utilities. When  utility workers were considered with the group with
potential EM-field exposure, the odds ratio was 3.94, and this increase was statistically
significant.
                                        3-115
10/22/90

-------
                            DRAFT-DO NOT QUOTE OR CITE

    Loomis and Savitz (1989) examined 1985 mortality data from 16 states participating in the
National Center for Health Statistics industry and occupation coding program with respect to
EM-field exposure. A case-control design was used to study occupation among men and
mortality from malignant brain tumors (N=1095), leukemias (N=1694), and acute myeloid
leukemias (N=474).  Controls were drawn in a 10:1 ratio to cases from all other causes of
death. Adjusted mortality odds ratios were calculated.  For all electrical workers, excess risks
were found for brain cancer. The observed excess was concentrated among electrical and
electronic technicians and electric power repairers and installers. No increases in expected
deaths for leukemias, and specifically for acute myeloid leukemias, were found among the total
group of electrical workers, although some subgroups, i.e., electricians and electrical and
electronic technicians, had elevated mortality odds ratios.  This limited information (see Table
3-38) is taken from an abstract; a more detailed paper is under development. The authors
stated that the results were consistent with other data on brain cancer but not leukemia.
    Reif et al. (1989) examined occupational risks for brain  cancer in New Zealand, in a
case-control design using data from a national cancer registry. The criteria and methods for
case and control selection and methods of analysis are the same as those previously
described for the leukemia study conducted by the same authors (Pearce et al., 1989).  In this
          TABLE 3-38. MORTALITY ODDS RATIOS (MORs)a FOR BRAIN CANCER,  LEUKEMIAS, AND ACUTE
                 MYELOID LEUKEMIAS AMONG MALE ELECTRICAL WORKERS FROM 16 STATES
Occupation
Electrical Workers

Electrical & electronic
technicians
Electric power repairers
& installers
Electricians
Brain Cancer
1.5
(1.0-2.1)

3.1
2.4
1.8
Leukemias
0.9
(0.6-1.3)

NG
NG
NG
Acute Myeloid
Leukemias
0.9
(0.5-1.8)

1.9
NG
NG
 MORs and 95% confidence intervals (in parentheses) were not given for all subgroups.
 NG ť not given.
 SOURCE: Loomis and Savitz, 1989.
      10/22/90
3-116

-------
                           DRAFT-DO NOT QUOTE OR CITE

study however, cases were males, aged 20 years or more, registered with a diagnosis of brain
cancer [Ninth Revision of the ICD (ICD9:191)] or of cancer of other and unspecified parts of
the nervous system (ICD9:192). There were 506 total cases, and 452 cases (of both rubrics)
for whom occupational information was available. Again, controls were cases of other cancer
types. Risks among both broad occupational groups and more specific jobs titles were
examined.
   Elevated odds ratios were seen for several occupational groups and types of jobs, but only
those results that may relate to EM fields or electrical work will be given. Odds ratios were
elevated for men with professional and technical occupations (OR=1.32, 95% Cl=1.02-1.69).
Within this group, elevated odds ratios were found for physical  scientists (OR=1.37, 95%
Cl=0.33- 5.77), engineers and architects (OR=1.16, 95% 01=0.73-1.82), and aircraft and ship
officers (OR=1.76, 95% 01=0.64-4.83). Although not presented in the published tables, risks
were stated to be significantly elevated among electrical engineers (OR=4.74, 95%
01=1.65-13.63), which was a subset of engineers and architects. To examine the potential for
a bias from improved diagnosis (and access to medical care) among white collar workers,
risks were also estimated for specific professional and technical jobs using only all
professional technical workers as the reference group.  For this analysis, the magnitude of the
association for electrical engineers (and other jobs) only slightly decreased and the lower limit
of the confidence interval remained above 1.0 (OR=4.06, 95% 01=1.36-12.01).
    Elevated risk was not seen for the broad occupational grouping of laborers, production
workers, and transport workers. More specific job types within this broad group were
examined. The odds ratio for electrical workers was not elevated (OR=0.78, 95%
01=0.39-1.59), but risks for a further delineated subset of electricians were in excess
(OR=1.91, 95% Cl=0.84-4.33). The number of cases among electricians was not stated, but
was likely small since there were only eight cases among all electrical workers. It is interesting
to note that the risks among plumbers and welders were confined to plumbers (OR=2.02, 95%
01=0.99-4.12).
    Swerdlow (1983) presented data on "eye cancer" in adults in England and Wales.
Incidence data between 1962 and 1977 was obtained from  14 population-based cancer
registries which submit their information to a National Registry. The general goal of the study
was to investigate the epidemiology of "eye cancer," largely of melanoma histology, and
describe trends over time by sex, age, region, level of urbanization, and latitude.  Registration
                                         3-117
10/22/90

-------
                            DRAFT-DO NOT QUOTE OR CITE

 rates were directly age-standardized using the European population (Doll et al., 1970;
 Waterhouse et al., 1976) as the standard.
    For persons aged 15 years or older, there were 2159 cases of "eye cancer" registered for
 men and 2125 cases for women.  Linear regression analyses revealed that the age-adjusted
 incidence had significantly increased over the time period. Social class and occupation were
 also evaluated. Age-standardized proportional registration ratios were developed and tested
 as Poisson variables.
    Generally, "eye cancer" incidence was higher in higher social classes (which ran in six
 groups from professional to unskilled manual).  Occupation was examined with respect to the
 26 occupational orders or groupings established by the Registrar General. Significantly
 increased proportional registration ratios were observed in three occupational orders, e.g.,
 electrical and electronic workers; administrators and managers; and professional, technical
 workers and artists.
    The author stated that caution must be applied to interpretation of the results for social
 class and occupational group because of a small number of cases and the necessity of
 excluding cases for whom occupation was unknown or could not be classified. The possibility
 of errors in coding and occupational ascertainment was also noted. In contrast to the other
 occupational groups, the observed excess incidence among electrical and electronics workers
 is somewhat at odds with the results seen by social class since these occupations largely fall
 in the medium to low end of the social class groupings  (in skilled manual and partly skilled
 manual).  Thus, Swerdlow felt that these results were of particular interest.

3.3.2.2.4.  Malignant Melanoma of the Skin. De Guire et al. (1988) investigated the incidence
 of malignant melanoma of the skin in telecommunications workers in Canada, following
several case reports.  The study population consisted of all the workers (N=9590) in Montreal
 plants of one telecommunications company who were employed for at least 6 months between
January 1,1976, and December 31,1983.  Persons newly hired during this period were
included as well as persons who had been working before the study began. The start of the
study, thus, does not represent the data of first employment.
   Cases of malignant melanoma among employees were ascertained by linking employee
lists with name, sex, date of birth, and social insurance number to case lists of Montreal area
residents.  The malignant melanoma case lists had been developed by identifying all cases
newly diagnosed in the same time period from records in 30 local hospitals. This list was
       10/22/90
3-118

-------
                           DRAFT-DO NOT QUOTE OR CITE

cross-checked against the Quebec Tumor Registry. Pathological and histological reports were
examined, and, if equivocal, were reviewed by a consulting dematopathologist. Histological
types included were malignant melanoma not otherwise specified, superficial spreading
melanoma, nodular melanoma, lentigo malignant melanoma, amelanotic melanoma, and acral
lentiginous and "epitheloid cell type" melanoma. The underlying population was obtained from
national census data for 1976 and 1981, and was used to construct incidence rates for
malignant melanoma for the greater Montreal area for the study period.
   Person-years of observation were  calculated from date of first employment and date of
work termination, or other means of withdrawal from follow-up.  Montreal rates were applied to
the person-years for workers to develop expected values after standardization for age and sex.
The  ratio of observed cases among workers to expected cases formed an SIR, and 95%
confidence intervals and p values were calculated. The latter were not reported.
   Ten cases of malignant melanoma were diagnosed among the male telecommunications
workers during the study period; 3.68 cases were expected. The distribution of histological
types was similar to published data from various countries.  There were no cases among
women. There were 52,456 person-years of employment for men.  The SIR was 2.7
(Cl=1.31 -5.02) and was statistically significant.
   Risk with respect to latency period was also examined.  Here, latency period corresponds
to length of employment divided into two groups, years and 20 years.  Among the cases, the
minimum number of years worked was 5 years (two cases), and the maximum number of years
worked was 38 years (two cases). For persons who had worked less than 20 years, the SIR
was 5.0 [observed (Obs)=5, expected (Exp)=0.99, 01=1.63-11.75] and was statistically
significant. In contrast, the SIR for persons who had worked 20 years or more was 1.9
(Obs=5, Exp=2.69, 01=0.60-4.34). Length of any latency period and duration of employment
often reflect age.  This is true with this population, except for one case.  Four of the five cases
who had worked less than 20 years were age 42 years or less; two were in their twenties.
Therefore, excess risk is seen in the subset of the cohort that was younger and had worked in
the company for the least amount of time. The authors discussed the implications of these
results. The greater risk in the "short latency" subgroup argues against a causative
occupational exposure. The range of years employed in this group ran from 5 to 18.  A
causative exposure, requiring a given latency period, may have occurred prior to employment
or, for some,  during childhood. On the other hand, melanomas are generally diagnosed in
                                        3-119
10/22/90

-------
                         '   DRAFT-DO NOT QUOTE OR CITE

 middle or older ages. For example, the median age of diagnosis for Montreal males was 54.4
 years. This small case group seems to exhibit a slightly younger age at diagnosis than is
 typically expected. The authors noted that, in addition to change, the results could also be
 explained by either some highly potent carcinogenic agent or an additive agent, both of which
 could accelerate expression of cancer.
    Detailed job or exposure data were not available.  The title of the last job held and the
 longest held position were not available for all 10 cases, but, it appears that about half of the
 cases were involved with technical work for some period of time. Bearing in mind the small
 number of cases, the "low latency" group had somewhat more technical employees than the
 "high latency" group which had generally more office or unspecified jobs. The tumor location
 was also slightly different in the two latency groups.  Limbs predominated in the low latency
 group. The abdomen and back predominated  in the high latency group.
    Nine other cases and one case of choroidal melanoma were not included in the study
 because they did not meet the location and time criteria for inclusion. The authors pointed out
 that continued follow-up would  be necessary to determine if the risk of malignant melanoma
 continue among this company's employees or whether the observed risk disappears, perhaps
 indicating a cluster or elimination of the responsible risk factor.
    Some of the studies previously discussed also have identified excesses for skin cancer.
 These studies are Szmigielski, 1987; Vagero et al., 1985; and Tornqvist et al., 1986. The eye
 cancers reported by Swerdlow (1983) were stated to be of melanoma histology.
    There are several interesting features of the reported results for malignant melanomas of
 the skin.  The excesses occurred in occupational groups that are younger than typical cases.
 The location of the cancers included parts of the body not always exposed to sunlight or
 ultraviolet (UV) radiation, at least on a long-term basis, although short-term high level exposure
 is possible. The effect has been seen in various types of jobs, with various primary
 frequencies of putative exposure, and, as  a consequence, with various potentials for differing
 confounding exposures. In other words, the effect has been seen in different jobs with
 different primary EM-field exposures that may not all have the same potential for the same
 potential confounders. As such, no single agent or limited number of agents could necessarily
 explain the results. Except for servicing linemen, there is no reason to believe that workers
with some sort of exposure to EM fields would work  outdoors more than other types of
       10/22/90
3-120

-------
                           DRAFT-DO NOT QUOTE OR CITE

workers. Therefore, it is possible that exposure to EM fields or NIR radiation may present
some risk for developing malignant melanomas of the skin.
   3.3.2.2.5. Summary of Occupational Studies.  Many studies in the United States, New
Zealand, and Europe have shown cancer incidence or mortality among workers in electrical
and electronic occupations to exceed general population levels. Such workers are likely to
encounter exposures to various frequencies of nonionizing or electromagnetic radiation,
including electric and magnetic fields at the power frequencies of 50 or 60 Hz. All of the
studies have strengths and weaknesses, some of which are common to most epidemiologic
studies.
   Many of these studies are hypotheses-generating in nature for several reasons. For
example, some of the studies were not designed, nor were the populations originally formed,
to test the hypothesis of whether EM radiation relates to an increased  risk of cancer. Some
were re-examinations of existing studies or evaluations of vital records, registry, or
occupational databases, and so proportional mortality or incidence ratios (PMRs/PIRs) are
often the derived estimate of risk. Data on work histories and exposures are very limited;
consequently, some of the observed associations may be indirect. Most of the studies drew
occupational data from death certificates. Occupation as drawn from death certificates can be
very broad, raising the possibility of exposure to an array of agents including EM radiation.  It
should be noted that there are problems associated with the  reporting of occupation or job on
death certificates where the accuracy of the recorded occupation may be affected by
difficulties in recall or lack of information or understanding of the decedent's job by informants,
the recording physician, or others.  Although the standard certificate asks for usual
occupation, rather than last or current occupation, the usual occupation may not be recorded.
   Confounding by other occupational exposures has not been thoroughly addressed in the
occupational studies, and more research to improve exposure assessment would be an
important improvement. It is possible that many different chemicals could be present in the
occupational setting, and the types of chemicals may be different for different jobs. On the
other hand, it is not fully known what chemical exposures would be strongly associated with
leukemia or brain cancer.  While there may be exceptions, it should be noted that human data
on chemical carcinogenesis are limited, in the broad sense, and one cannot simply assume
that concomitant chemical exposures may explain these results. Similarly, it is difficult to
conceive that other possible explanatory factors, such as socioeconomic or behavioral factors,
                                        3-121
10/22/90

-------
                            DRAFT-DO NOT QUOTE OR CITE

would be consistently present to an important degree in all the occupations, jobs, and
populations that have been studied.  Mechanisms that could involve cancer promotion or
enhancement have been hypothesized.  If such theories can be confirmed, other exposures
may not be confounders but may be necessary.
    Several studies attempted to address multiple exposures in various ways in the face of
limited industrial hygiene information and no monitoring or measurement data on EM fields,
chemicals, or other factors. Two, in particular, made a fairly good effort of sorting out various
exposure by type of work performed. Stern et al. (1986) found significant excess risks of
leukemia among shipyard electricians and for shipyard electrical work. Results were not as
pronounced for work with solvents.  Thomas et al. (1987) found that brain cancer risks among
workers with jobs involving EM-field exposure were reduced when subjects with exposure to
lead, soldering fumes, and  solvents were removed from the analyses.  That approach does not
necessarily refute or support the potential carcinogenicity of EM fields (it may have merely
eliminated subjects with the clearest mixed workplace exposures), but it is useful to examine
potential differential risks.
    The potential for confounding exposures is a critical issue. An important point must be
kept in mind. For a factor to be a confounder, it must relate to exposure and to the  effect or
end point under study. Two perspectives can be explored:  what is known about the
epidemiology of the end point (specific cancer type or site) and what is known about other
putative causal agents in the environment of the exposed subject. Smoking is a major factor
for cancer in adults. Absence  of data on smoking histories is always problematic. However,
the cancers identified in these  studies are not strongly associated with smoking.  Occupations
involving RF radiation and EM  fields also involve exposure to other agents (chemicals), some
of which have evidence as to their carcinogenicity. Some of the putative confounding agents
are associated with the end points identified in these EM-field studies, while others are not.  It
is not yet clear what agent(s) is strongly associated with the specific cancers to serve to
explain the results.  Just as the extent of EM-field exposure is largely undefined in the
occupational studies, the extent and type of exposure to chemicals  in electrical jobs and
occupations is not yet clear.
    Although the many occupational studies have noted increases in various cancer sites,
three broad types of cancer predominate, i.e., neoplasms of the hematopoietic system,
especially leukemias, and, to a lesser extent, lymphomas; nervous system cancers,  including
       10/22/90
3-122

-------
                           DRAFT-DO NOT QUOTE OR CITE

brain cancer; and, to some extent, malignant melanoma of the skin. For those studies that
examined specific types of leukemia, acute myeloid leukemia (AMI) predominates, although a
few studies have noted increased risks for chronic lymphatic leukemia (CLL).  There is, thus,
some degree of site specificity.
   As mentioned previously, further research is needed to improve exposure  assessment and
to define relevant exposure parameters.  Few studies provided measurements. While it is clear
that the studied populations are exposed to NIR, it is not clear to what degree or what
exposures are biologically relevant.  To date, the study that best addresses the occupational
exposure environment was conducted by Matanoski et al. (1989). Although firm
dose-response gradients were not reported, excess risks, especially for leukemias, were found
in the group with the highest average magnetic field exposure. To some extent, risks became
less pronounced with lower average exposures.
   The occupational studies can involve exposures to multiple frequencies of NIR and
exposures to 50 or 60  Hertz. Excesses have been reported for many types of jobs, not only
electricians. Frequency of exposure cannot be readily discerned for job titles  or job
classifications. The predominating exposures are probably to extremely low frequencies, but
jobs involving radio, TV, radar, and so forth entail exposure to modulated RF radiation.
   The lack of supportive quantitative exposure data in the occupational studies makes it
difficult to determine harmful levels, if a true risk exists, and to test for dose-response
relationships. Therefore, dose-response has not yet been characterized.  However, many
studies have attempted to evaluate dose response in a qualitative fashion, e.g., by examining
risks across work or job subgroups that represent a gradient of potential relative exposures.
Admittedly, the approaches have been necessarily crude, and exposure misclassification is a
distinct possibility, but qualitative approaches have yielded some evidence of  greater risks with
greater potential exposure. Some results from laboratory-based studies have  identified
frequency and intensity windows for certain effects (see chapters on mechanisms and on
supporting  evidence) which point to the possibility of nonlinear dose-response relationships.  If
further animal and cellular level research continue to build support for that evidence, the lack
of quantitative dose-response evidence from epidemiologic studies may be of less importance
in weight-of-evidence assessments that draw upon classical criteria for causality.  Furthermore,
nonlinear dose-response may be difficult to discern in epidemiologic studies.
                                        3-123
10/22/90

-------
                            DRAFT-DO NOT QUOTE OR CITE

   Yet, despite limitations of the current data, the results of the occupational cancer studies
are remarkably consistent. Similar results in different geographic regions, countries, age
groups, industries, and occupational classifications have been reported.  Given this diversity,
plus the likelihood that, across studies, the work denoted by broad job titles is probably not
uniform, it is difficult to identify any single agent or group of compounding exposures that
could explain the consistently positive results.  Exposure misclassification is a distinct
possibility, but, if a true risk exists, that sort of bias forces risk estimates toward the null, and
observed risks would be underestimated. The magnitude of the risk estimates is similar
across studies. The specific cancer sites associated with excess risk are similar across
studies. In addition, similar results have also been obtained using different study designs.
   The strengths and weaknesses  of the many epidemiologic studies on jobs and
occupations involving work with EM-radiation sources and EM-field exposure have been
discussed above.  Despite their limitations and the clear need for more research, especially to
improve exposure assessment, the  consistency and specificity of the findings provide
evidence that EM-field exposure in the workplace may pose a carcinogenic risk for adults.  The
results from occupational studies are supportive of the positive results noted in most of the
studies of children exposed residentially.

3.4.  SUMMARY
3.4.1.  Introduction
   This chapter reviewed the available epidemiologic literature with respect to evaluating the
relationship between exposure to EM fields and cancer. More than 40 studies of EM fields
have appeared in the literature in the last 15 years. They can be divided into four basic
categories as follows: (1) studies of children exposed residentially to 50- or 60-Hz magnetic
fields, (2) studies of adults exposed to RF radiation,  (3) studies of adults residentially exposed
to EM fields from 50 or 60 Hz, and (4) adults potentially occupationally exposed to
electromagnetic radiation.

3.4.2.  Childhood Cancer
   Nine studies were reviewed, eight case-control studies of childhood cancer and one case
report involving adolescent girls.  Six of the case-control studies examined residential
exposure from power transmission  and distribution systems and the other two studies have
       10/22/90
3-124

-------
                            DRAFT-DO NOT QUOTE OR CITE
examined childhood cancer in relation to paternal occupation. Five of the six residential
exposure studies showed positive associations with exposure to EM fields.  Three of these
reported statistically significant results. One of these three dealt with wiring configurations that
were reasonably correlated with actual measurements of magnetic fields in the vicinity of the
residences of the children. The two remaining significant studies also reported similar findings
at similar levels as low as 2 mG, or 0.2 fiJ. Where different cancer sites were evaluated,
leukemia, lymphoma, and nervous system tumors were found to be in excess in the five
residential studies showing positive associations.  However, the author of one of the two
nonsignificant positive studies concluded that his findings did not support the results  of the
three significant studies based  upon his tests of significance. Failure to find significance may
be due to small numbers and the error of including children in his control group who were
most likely exposed to magnetic fields similar to his "exposed" children, i.e. within 100 meters
of overhead lines. The second nonsignificant positive  study from Taiwan did not have
sufficient power to detect a positive risk as significant.  The sixth nonpositive study based on
wiring configurations was methodologically flawed because of the use of multiple dwellings for
case children and single addresses for control children, which would tend to dilute exposures
and risk estimates.  The study also overestimates the degree of fall-off of field levels with
distance, which again would  dilute exposure and response. This device will, by necessity,
force risk estimates toward the  null.
   The two paternal occupational studies found statistically significant associations between
neuroblastoma and brain cancer among children whose fathers' job likely involved exposure
to electric and magnetic fields.  The case report of a cluster of endodermal sinus tumors in five
adolescent girls in close proximity to power distribution lines was not helpful. No
measurements of actual fields were done.  Although the girls lived near such power lines,
coincidental environmental factors that were  not identified could be responsible.
   The case-control studies of children residentially exposed to magnetic fields provide
evidence of a positive association of a risk of certain types of cancer, namely leukemia, central
nervous system cancers, and lymphoma. Unfortunately, detailed information regarding
specifics of these cancers is not available in these studies. Because these measured  risks are
low in all of these studies, the possibility that some unknown confounder is responsible cannot
be eliminated.  However, because of the consistent positive findings and suggested site
concordance, chance is not likely to be the explanation.  Furthermore, one of the positive
                                         3-125
10/22/90

-------
                            DRAFT-DO NOT QUOTE OR CITE

studies extensively examined several possible confounders and did not find another
explanation for the association between childhood cancer and wire code and magnetic fields.
Questions still remain regarding what length of residency would be required before one could
conclude that residency in a magnetic field is likely to be the cause of a cancer. Only a few of
these studies discuss this in a cursory fashion. Residency at time of diagnosis is not
necessarily a valid surrogate without information about how long the child lived there.

3.4.3. Adult Cancer
    There have been  many studies of adults potentially exposed to different frequencies of EM
radiation or in different exposure environments. These studies can be grouped into three
categories, that is, RF radiation, residential exposure to EM fields, and occupational exposure
to EM radiation (see Table 3-39).

3.4.3.1 Radiofrequency Radiation
    Among eight studies of radiofrequency radiation exposure in adults, five cohort studies
reported  statistically significant risks of lymphatic and hematopoietic neoplasms, especially
acute myeloid leukemia, in amateur radio operators based on proportional mortality ratios and
standardized  mortality ratios, and of total cancers in Hawaii residents living in  close proximity
to RF towers based on standardized incidence ratios, and of hematopoietic system cancer,
including leukemia, lymphoma and lymphosarcoma, as well as melanoma from exposure to RF
radiation, primarily from radar,  in Polish military officers.  For the Hawaiian study, leukemias
were nonsignificantly elevated, but small numbers made the leukemia estimates unstable.
Two studies (of U.S. Embassy  employees in Moscow exposed to microwave radiation and U.S.
Naval personnel exposed to radar) suggested a slight tendency toward an increased risk of
cancer in general and the hematopoietic system specifically.  Excesses were,  however, not
statistically significant in either study. The remaining cohort study of participants in a radar
research  and  development project at the Massachusetts Institute of Technology, where the
subjects were exposed to RF radiation estimated to be about 2 to 5 mW/cm2,  produced few
significant findings of mortality in the subjects. Some elevated risks of certain site-specific
cancers, most notably Hodgkin's disease, were found.
   AH of these studies, although suggestive of an elevated risk, suffer from design and
methodological problems, which in their entirety preclude any definitive statement regarding
       10/22/90
3-126

-------
                         DRAFT-DO NOT QUOTE OR CITE

          •O JJ

          S E
          o O  o) ui
          a> to  X o>
        fill!*

        5 1(0 ,gl g
        — -^^ In — §



        I 1 ll| S
        i CT ^ g SŤ
  * Q. CO >-
  2 P -0 .fr
  01 3 w =
Jg "^ LIJ nr ^






5-11 E Ł o c
  o> o Q. o
S- • — \j \j
   E o>
                                              f I
                                    0>  Ť
                                    I  4*
                                      fl> O "S
                                   o|e|^s
                                   l?^^-g§
                                   i e e s 5 "
                                   SISS

                                   llil
                                   <: o E <
            S a Ť
            tt> C C
            ais
            sgi
            -ISO
   o

   Q


   (O
        DC


        05
                                   cc
                                   i
                     E
                     10
                                   — 1

                                   Is
                                   -a
                                    55
   05

   CC
                     i
                     E  o
                                    3-127
                                                              10/22/90

-------
                 DRAFT-DO NOT QUOTE OR CITE


                      a •
                    ^ o
                              O -


                              HI
       I  I
       n  ?
                     l
                       Ť
                       flf
                                —

                               O) O) ^=

                              II1B
                                       V tf) Q.

                                       &^EoťSE0
                                       v 0..2M cc j; Q o
                              cc c
                                 1*8
 Ť   e**-'
 .8  SlllE
 i*?si|g
R|
                                       D _•

                                       *l

                                       !l
                                       c S
                                       o >
                                       te ^
                                       g&
                                       fi-
                                        "
                                                   ^§••0
                                                   9 ŁŤ

                                                   Ills
                                                   Ti -^
                                                   i Q
1st
C JO 03
c ? E

Ł h a. y>

S.E s?|
>ť "O 3 t8


|lle
2 DC S *:
                               ill
                                                   8
c.
E
                               E r?
                              E o
                              g E B
                               v u
                               n S
                               "5.S
10/22/90
                              uj uj 25




                            3-128
                                     1

-------
DRAFT-DO NOT QUOTE OR CITE




















I
1
S.
to
NCER IN ADULT
o
u_
O
ffi
5
i
oi
C1^
A
3
I






















f3
o
z
1! 1
GHTS/DEFICII
3
I
O
I





en
tj
=>
en
UJ
cc
Q_
3
O
cc
o
z
COMPARISO


z
o
1
Ł
>
o
fe
z
g
Q
V
Q
en

cc
H
§




CC

g
of

X
=>
"o
$
1 =
t; 0
_0> *3
ire based on <
;ode configur;
w _,

O. "C
UJ?

(A
O)
O
§
O -
3 o
Ť "5

1 b || §f
Łf!f if
l|lc?l|
Ililll
0) C -g - Q> ° ^5
CT 3 := ~  Ł S g Ł
s i.
^* X
CT^



ŤŁ
.
™ >,
8 1
TJ 1C
Non-cancer
and geograp

•o
i •§
w E
c 5
1 2
<3 >
CO 3
O W
c
"5 "^
o '°
0 ^
S "
S 5
"o*
1
N 0
cc en
X
s
en oi
So.
QJ
X Q
UJ ->
-J ofl
< i.
p g
z .=
It 1 QJ
9 § sr
iS ^§
>.-C g C
•^ >ť ~
"S o "^ .2



'"""


§
U)
a)
Q.
T5
QJ
j:
o
"S





.
0




















TJ
0 , g
6j^I|c^^ C ||
i-3S|Łl§,|a,So*S
liitHliiillli
iil!i!!!!llil!
csct-^3J a> S-0-0^ ^^.cciJ ^

c ^O^^^QOO ^+-*Or^**T)U)
!":?li§ililig
•Egc8BgcJ2J5n>JSra'5
|l|||i|llHl|
>.& — 3lJj!-|.gsig;s.EŤ Ť
15 ^oto_(i)tfJ*~t^WniQ. -^* 2 O
.g ra-^^^ra^-s w.co^ro-^^



'"--•c~c
? 0
s 5
S ™ o S S
D .2? jfj .y j^ m ro
Illti^
SŤl s Sg?
a-sfSlIS
x c p a) G> ro c

"° H- "o in

g^ESE-.^ |Sc
ililiolil^fl
^gEz^'^S^Jaj-D.Q;
liyifllllll
I
Ł
*J
0 (Ł
o to
CO
UJ^
I S
OI O)
Ł5




5?
a>
en
E'


o
o
S
           3-129
10/22/90

-------
                           DRAFT-DO NOT QUOTE OR CITE
I
T> O



ll
Ss
  •a u T b"5> s "

llllflSHs.
-TSSslssI-
          u> iD
                             ro *- S* to
                                  9.
                                 E -c Ł
                              sl
                  Ł•5 0> -wa ŁŁ"9^5
                                 Ť !S -K -^ -^
                                    ?• w
                                    is
                                      1-5
                                      if
                        B
      E-.
      18
                                          Ł   Ť

                                          ts  S 2
                                          0)  u Ł

                                          Ł  g 1
                                          0-0 S °
                                           Sra te ^
                                          3 2t5 g
                                          c -0 o Ť
                                          §=2 S
                                          "D Ł ^ (0
                                          i1  . .r
                                          O O) .ti

                                          S S 5
                                           *•
                                          JZ O . T!
                                          o -D ra Ł
                                           I E   S
                                           E .a o cr
                              01 =  I '
                            Ł"P g  ŤR

                            S ~ I1 ť g ^ <
                            0 'S o = ť •§•*
                            U) 01 tn Ť = ^ '
                            0} (0 •—  o T^

                            leM&li
                            < ^ 5 E o g '
                                                "S S
                                                S: o
                                                S-.S
                                                             •D en



                                                             S TJ
                                                                    ul
                                                             i  _
                                                             J3  o
                                                             S S 15
                                                             O
                                                             13
2  o^-c I
=  S jjj I jg
_y c c "Ł Ł^ ^

Ifll!!
      10/22/90
                                 3-130

-------
                         DRAFT-DO NOT QUOTE OR CITE
   10
   i
   52
   i
                                                            o ui

   O
    * S I Si
                                          lf^ll

                                          tIIII
           ' c Ť

            II.
p o o •" Ł'v s
iS 3 E n Ť S °
      cc
      S
                                        To
                                        S
                                        g




                                      3-131
                      10/22/90

-------
                   DRAFT-DO NOT QUOTE OR CITE
            y a
                                                            o  — Is
                                                         * g Ť  I !•Ł
                                                         g'sislEs
                                                         3'<33i§ g
                                                             01 ~ .a: Ji?
                                   E ^
                                   " 01.™ o
                                   S " §<•>
                                   5tf5
                                   i^!|
                                   i s 5 Ť
                                   !I?ii
                                   cc = tr g S
                                   5; ti 5 u) 3
                                   ŤEŤs I
                      ą: d> _c

                      T31.S I? I fc S-S
                      S ">  . o ^ -2 ^ Ť
                      > o> -o S Ť
                      " o 0> X fc
                         E Ł
                                                                 S .. S!
llsc
                     U) U> X -C
                     >i OJ QJ H—
                     W CD Ť O
                                   ||

                                   lla
                      ť 5 1 C, !2

                      Illll
                                                      E 2
          c s
          o g
                                   E •
                                   tf'
                      to
                      tf
          c.
          •a
I
                                                      SB
10/22/90
3-132

-------
DRAFT-DO NOT QUOTE OR CITE

ffi

HIGHUGHTS/DEFICIENC










RESULTS
?
c °-
1 S
S o
C §
;J 0)
1 1
2 S
a: 8
UJ Ť
CJ
1
o
U) ^
UJ 0
5 E
1 i
ABLE 3-39.
STUDY PC
1-
z
CD
s
Q
s
ID


,_
cc
1-
H
8






cr
>^
s
a:
0
<
(- Ť1 >L
S| It c
33 J2 . (fl • ,_ >
*= J= OJ -° fc "5,
S Ł "S T3 "O ^
to ^ 8 Ť c ° "S
— ~ C= O 3 flj X
tllplf
IHllii
i Iliii
a> fc 13 Ł: Ł t5 2
Dig ||E Ł = •Ł
oj o E ^ *ť- *•* o
_j a. = a. o to o
w
(0 C
IS
0) (0
_ •§ Ť Ł
Ť -Si te g
°IH
llM
S TJ g-a
S i Ť i
•Ł in -g w
S | " 0 Ł
E 1 -| 2 ^
liili
o
*~"
Ť ^
2 o c

1.1 1
to 'S °-
a. Q. (u
E 3 o
o o -S
o o S


"

in -a
c E S
2 1 01
Sj2 c
<ť C Q)
Q> "D 5
112

ll

O "(0
Q- "C
0 0
Ł E




E
^ t/>
"in ^
o
0)
o
a
o —.

Ł CD
QJ Ol
x ^
a) cJ
Ł03
O)
"o ^
5 E
o ^
8 i
DC
•^•5 . 5
| c § I ?

S|l s II
Il|lg ť
Its 1 il
I |s| SB
S E TJ Ł u a.
t-. 0 >- 2 °

c*^5 tj 1/jŁ w^

2(0 — "3"^ 5!5>E.2.o
C^*~y)2 "clW-1-^^
'°'§S'oS<"S^^'D^i'!i
g agsS^fa-SoS g
t_. in .^m. .
_ ,„
' lS|h2cli
C QJ '™ *
o *o !.. • -Q
to o 3 o to u
C? Q. w — ) o o

cc

t c
O 0)
II

.sf
c
o

1
tn
o





OJ
S
^-*
T5
S
1
5
lo^i=-
g ° -D OJ O O
•1 il 1 1

•§ 'g 3 ° S c
i) a* in *^ (-! o
Illlll
—  O> ^_
•g .c (5 t) Ł g
m tJ Ł 0 Ł (-
IT E 0 2 S <5
-D o oa

^w ^ŁJo "o^p

o-cE ^coc" "^^S
c*^§"^g^-o>* Ł*~^
E^aT^Sw^ ."Ť"§
(0-^S:io^^-|Ťol S
cc
f
O
1

           3-133
10/22/90

-------
                          DRAFT-DO NOT QUOTE OR CITE
S
cc
                                                                              " "55
5"'
^ "*, C OJ
                                     sii^
                                     m JS F 2

                                     ill I
                                     Uj _OJ O) Q,
<" oi
°- m
.03 ta

2§

|Ť •
d) O ,



= S •
                                      
-------
                  DRAFT-DO NOT QUOTE OR CITE
  UJ

  CJ
HIGHUG
    il!




RESU
             lilPliJSL !lsl!°g.-o!g|g
             *"        oS|^sj§.BSj Ť
    Ilslllllltlllllllll
1


cS
a

3
  Z

  O





  n
DY
                                           -s
fillls
3S-5S a"



  o

  55
  UJ
  o

                         °^B
                         " 5 2
                          E .E
               o



               o

               V
               
-------
                     DRAFT-DO NOT QUOTE OR CITE
          Ť .0   >n
          SS   S
2 S  g ŤŤE- .

   tilfffllf
   .2 o "C Ť*= Ł a. u> Ł•!
        =5 S^E §
   Ť .2 o "C
   c viz 8
 •gs-^a>.2.u!3-gB> . •§. u,


 Hllllllllll
                   . o
                  §Ł.
                   Qj 0>
E         —  on *"   -^
           y.y.— ^> B 10
ss:
_ E '
tu en .
 u. o E
                           .2 o
                          -•°i
                        •s s •= >
                        2,2" S
                        f S1 Ť• I ť
                        S^S ° S
                        Il518
                        Łi?ll
                        •O V) 0) 3 g
                          §f- -P Q) g-
                           E  "c

                        11-11
                       ti
                                          01 s
                                         T3 C
                                           O  3
                                         •(g W C ^ O

                                         5 o u .Ł o)
 si
 §s
                                                       02
                                                                 o o
-=IffeS
"CO -D Ł S jj O 5

Oo:3:3Ťoa)H
LL uJS.SJ.2 S"S*2
                                                      Etfl
                                                      0)
                                                     O "O
                                                      SO)
                                                     _
                                                     i
                                            f
  feťť  52 >,!
° 11 3 S -s I 8
•D.-oO'c'D-Sc
.a -| •§ -a § a E i
"Diisiccitn — s
                                                   ° 3 1
                                                   ^.E^^
                                                   5Iť13JW„



                                                   loB"" >,"^"5
                                                   S E 5.5. E 5.S a
                                                     o
                        I
                                                    S?
                                                    I
                                                    C
-------
DRAFT-DO NOT QUOTE OR CITE

w
1J
i
g

1
E
0
1
i

RESULTS
1 3;
fo
rr
_ §
(o Ť* ^
GJ k-^'S*" ^i*O S'W
ilillHll
Ocoo>t^o.ca.
•o
.= _ Ť in c
Examined electrically-related
occupations. For all electrics
workers considered together,
elevated ORs for several type
of cancer but only significant
for leukemia (OR-1.62). ORs
elevated for most of the
specific electrical occupation
but only significant for radio
and TV repairers (OR=7.86) ai
were borderline for linemen
— H
c

= 8 §1
1 SS fe
~ O Dl Ł
> Ť Ł o .2
_ Ł nj a> c
g Ť 0 0 0
Us si

i ^
O m m *^r -i
"^.^ 0 ° W
§, fe 1 "S SI
g 1 c -g §- E 3
Ł S.tl'olf

0
"Ł
o
o

o
S~
1
z
"c5
Q)
1










t_ O
o yj c yj c Ť_,


























                       ;•ť I

                 ili^iil
                 jJiliit
                 i Ł i Ł a 0 'S Ł
"i c ^-A
rf 8 3. a =c
^ U) r- T^ —

flill
                •D c



                | I

                S o
                |> 7t U) ' i != t
                o Ť c ^ -1 ro c
                sspl^sttl

                jlgllllpl
                T C Ci=^t)t/)^ O i_
                ssr
                U) CQ
                0) O)
            3-137
                10/22/90

-------
                     DRAFT-DO NOT QUOTE OR CITE
l^^i||||l^
                                       W Q)  D>
                                       •C ^  c

                                       Ł ~  Ť

                                   "o J2 Ł '
                                   •C S.'D .
                                   Its-
                                   g-Łs'
                                         | ^o
                                         i Ť in TJ
                                         ° Ł in 0)
                                         o .9 o Ł
                                   i- '> P O

                                   8sl§
                                   ŁŁ < S •
                                   : ^ 3 1
O
T3
^ ro o ^i
13 Ł -^ E
o i- w o
.c *
                                    B V V

                                    c 2-55

                                    S -3 I S
                                    ^ 10 C TJ
    CO'
w -  "5
||  #
?9SS

gliSw

iiir
(/> ^ W) ŁJ
                                             Ł 8
                                  .c 'g  g   B
                                  g =5 te |  5 •&
                                 S^3j:'o'SPT~ c'
                                 lis.ssŁ5'i;
                                 |||i§-Si:
                                 ft	
                                  I CO i_
                                  • .y -Q

                                61! i
                                ^3- 2 CJ ^
                                      o •
                                         o
                                 ^
                                E
                                o
10/22/90
                              3-138

-------
                   DRAFT-DO NOT QUOTE OR CITE
I
=>
D
<
DC
s
&
y)
E 3-
    (3  Q> yj QJ
                     p3 Sgi=5 >-S

                     5 8 Si i o §S

                     iiliiiiiiiiii
                     •:
MOR=2.4)

MOR=1.8)
           —
                     is e
                     t- <0
                     E Q)
                     iiHHli
                     lliiiiii!
                     Monť§,"'
     it
                     si
                              3-139
                                                          10/22/90

-------
                   DRAFT-DO NOT QUOTE OR CITE
10/22/90
3-140

-------
DRAFT-DO NOT QUOTE OR CITE


1
"c
O
O
ADULTS i
z
cc
111
o
•z.
0
fe
ffl
D
=)
to
S
CO
^J
<



















HIGHLIGHTS/DEFICIENCIES
RESULTS
D_
3
CD
PARISON
1




z
o
i=
3
=3
D,
2
S
=>
te
z
o
S
Q


to

CC
fe
8
o
S
C
0
cc
o
i
fe
Different from other studies
because is of eye cancer.
Proportional rates influenced by
rates of other diseases not under
study. Large study In toto of a
rare disease. No Information on
exposures. Results for electrical
& electronic workers unusual
given social class results, i.e.,
these workers tend to be classed
as skilled manual; author found
this to be of interest.
Significant Increase In incidence
over the time period. Higher
Incidence In higher social
classes. Examined 26
occupational groupings.
Significantly increased PRRs for
electrical and electronic
workers; administrators and
managers; and professionals,
technical workers, and artists.
T3
1
ť c i
1 ŁŤ^
Illl
if ii
LLJ O Q. U)



oi^ -o c:
„ 2 §J ? § E
oP ^ r" Ł "S *" 3
*- ° S O m 3 c
in Ł 5 -o -° "^ u>
Sc *jf Q) c en *
e Si ^= o  S2
co .E .a -^= — w g
,.„ fe Ł ^ •§ § tn
tc ^
_ . _ . U CŤJ




cc
cc
Q.
(0
_QJ
T3
S
0ť

rn"
to
CO

Q


Q>
frt
            3-141
10/22/90

-------
                             DRAFT-DO NOT QUOTE OR CITE

 carcinogenicity of RF radiation.  In some studies, follow-up was not sufficiently long.
 Identification of actual exposure was not possible in most cases; only estimates of microwave
 radiation were available. Proximity to or work with radar or microwave sources, census tract
 residency near RF sources, or membership in a ham radio club or listing on Federal amateur
 radio licensing rosters provided the basis for classifying subjects into cohorts of "exposed"
 persons. Hence, misclassification of nonexposed persons into these cohorts could have
 contributed substantially to the absence of the finding of a risk. The studies of amateur radio
 operators are the most persuasive in demonstrating positive significant associations. The
 frequencies to which amateur radio operators are exposed at the lower end of the RF portion
 of the EM spectrum, while the remaining studies in this group generally dealt with higher
 microwave frequencies; therefore, the potential risks among  amateur radio operators may or
 may not be directly applicable to higher frequency exposures.  Based upon the findings from
 these RF radiation studies, there appears to be little evidence to conclude or deny that a
 cancer risk exists from exposure to RF radiation.

 3.4.3.2. Adults Residentially Exposed to Electromagnetic Fields
    Five studies of adults residentially exposed to EM fields of 50 or 60 Hz were reviewed.
 Four of these were case-control studies;  the fifth was a cohort study. Three of the four
 case-control studies were studies of leukemia only.  The fourth  case-control study consisted of
 a series of adult cancer cases and deaths sequentially recorded in vital records offices and to
 a cancer register during a given period of time.  Two of the leukemia case-control studies
 found no elevated risk of acute nonlymphocytic leukemia in residents of western Washington
 State exposed to power frequency magnetic fields or risk of myelogenous leukemia with use of
 electric blankets. However, the third case-control study found a nonsignificant increase in the
 risk of leukemia to residents of South London in close proximity to overhead lines.  The fourth
 case-control study reported significant risks of CNS cancer, uterine cancer, breast cancer, and
 lymphoma in  Colorado residents living near power lines termed by the investigators as high
 current configuration. A cohort study of South London residents living in the vicinity of mainly
 low-voltage substations produced an excess risk of lung cancer in residents living close to the
power facilities, but not leukemia. Cohort studies are insensitive to the detection of elevated
risks of relatively low incidence diseases such as leukemia, unless large numbers of subjects
make up the cohort.
       10/22/90
3-142

-------
                           DRAFT-DO NOT QUOTE OR CITE

   The five studies discussed suffer from several problems that may explain the mixed results
that were evident. The two nonpositive case-control studies of leukemia were studies with
small numbers of cases, less than 120 in one and 225 in the second, for which the measured
exposure in the first was less than 2 mG in the large majority of the residences considered,
and in the second study of electric blanket exposure, it was suggested by even the author that
their hypothesis might better be tested in a colder, harsher climate than Los Angeles.
Presumably this would mean greater exposure to magnetic fields might be found elsewhere.
The third suggestive (for leukemia) case-control study of some 769 cases in South London,
still reported less than 1% of the population live within 100 meters of overhead power lines.
The vast majority of cases were not close enough to overhead power lines to be considered
exposed. The remaining case-control study of Colorado residents around Denver, although
including 1179 cases of various types of cancers, probably contained few leukemia cases in
accordance with normal chronologically recorded incidence and mortality patterns over a
specified period of time. Hence, with probably few leukemia cases identified with this method,
the study would have little  power to detect as significant an elevated risk of leukemia.
Similarly, the cohort study, notwithstanding 12 years of observation, only expected
approximately six cases of leukemia, and, as a consequence, had little power to detect a
significant risk.
    In short, these residential studies of adults, although showing mixed results, either lacked
 evidence of any substantial EM-field exposure to members of the study groups to which they
 belonged or  else lacked sufficient power to detect a significant risk of leukemia.  These studies
 cannot be used to support or refute a possible association of EM radiation with an elevated
 risk of leukemia, CMS cancer, lymphomas, or even possibly other types of cancer.

 3.4.3.3. Occupational Exposure to Electromagnetic Radiation
    Twenty-eight studies have been evaluated dealing with cancer incidence or mortality in
 workers in the electrical, electronic and similar occupations with a high potential for exposure
 to EM fields. The EM-field exposures in these occupations chiefly involve extremely low
 frequency fields including 50- to 60-Hz power frequency fields.
     Eight are cohort studies involving the investigation of mortality in the Canadian labor force,
 Swedish telecommunication workers, cell room workers in a Swedish chloralkali plant,
 Swedish male electrical engineers and architects, Swedish power linesmen and power station
                                          3-143
10/22/90

-------
                             DRAFT-DO NOT QUOTE OR CITE

  operators, Swedish telephone operators, New York telephone workers, and, lastly Geneva
  painters and electricians. In several of these studies, although they involve jobs that have a
  potential for exposure to EM fields, many of these workers probably had little exposure
  because of the ways in which they were chosen, e.g. census data or cancer registry data. The
  decision to include persons was not based upon actual exposure to EM fields, but rather was
  based on whether a given broad occupational category, as identified from a cancer registry or
  census record, includes persons who probably have exposure to EM fields. This means there
  may have been, and usually are, many persons erroneously included who were not exposed
  to magnetic fields. These studies border on being called ecological in nature. Two of these
  cohort studies (transport and communication workers and cable splicers) exhibited a
  significant excess risk of leukemia. In the former study telephone, telegraph, and power
  workers demonstrated a high risk of stomach cancer, but not leukemia. The study of New
  York telephone company employees found an elevated risk of leukemia in cable splicers
  compared to New York age-specific cancer rates, but when compared to other nonline
 workers in the same company, the risk was significant based on three cases. The risk of
 lymphoma was also significant compared to the nonline workers.
    However, most of the studies lacked the power necessary to detect even large risks of
 leukemia given small sample sizes and short observation intervals. Two of these studies
 demonstrated elevated risks of melanoma in telecommunication workers and in electrical
 engineers; one was significant. The single most important methodological problem in all of
 these cohort studies is the lack of definitive personal exposure information. It is probable that
 actual exposure to EM fields varied considerably within even those subgroups having the
 highest potential exposure.  Without such information and the likelihood of misclassification of
 exposure in all of the cohort studies, the results from these studies are equivocal.
    Ten case-control studies of adults who were diagnosed with either leukemia or central
 nervous system (brain) cancer were conducted to determine if prior exposure to EM fields  may
 have contributed to the risk of these two site-specific cancers. Six evaluated the risk of
 leukemia and five evaluated the risk of central nervous system (brain) cancer.
    Five of the six leukemia case-control studies reported significantly elevated risks
associated with jobs having a high potential for exposure to EM fields. These are: (1)
electronic equipment assemblers and radio/ television repairmen in New Zealand, (2) miners
working underground for 25 or more years, (3) electrical workers in Sweden, (4) electricians
       10/22/90
3-144

-------
                           DRAFT-DO NOT QUOTE OR CITE
and welders at U. S. naval nuclear shipyards, and (5) New Zealand electrical workers, namely
radio and television repairers, electricians, linemen, and power station operators.  No excess
risk of leukemia was found, however, in any groups of workers who were part of a 16-state
National Center for Health Statistics (NCHS) survey of industries and occupations.
   Five of the ten case-control studies were devoted to determining if an excessive risk of
central nervous system cancer was in any way associated with jobs having a high exposure to
EM fields. Four of the five noted significantly elevated risks of cancer in the following
categories of employment: (1) gliomas and astrocytomas in Maryland electricians, telephone
servicemen, linemen, railroad and telecommunication workers, engineers as well as electronic
engineers; (2) primary brain cancer in workers of Philadelphia, northern New Jersey, and south
Louisiana involved with design, manufacture, repair, or installation of electrical and electronic
equipment; (3) brain cancer in East Texas male workers involved in highly exposed (EM fields)
occupations in the transportation, communication, and the utilities industry; (4) brain cancer in
workers identified in a 16-state NCHS survey of industries and occupations. The remaining
case-control study did not exhibit any excess risk in New Zealand electrical workers in toto
with jobs coded to a Standard Classification of Occupations, but excess risks were found for
the more specific job of electricians, and statistically significant excesses were seen for
electrical engineers.
    All of the case-control studies discussed have problems with respect to who had
significant exposure to EM radiation. Simply identifying a job from a register, census listing,
 death certificate, or questionnaire as having potential exposure is not enough. No
 measurements of actual exposure have been done in any of these studies. However, there
 does appear to be some evidence of an increased risk associated with employment in job
 categories where there is a likelihood of EM-field exposure.
    There were four studies that could be characterized as either incidence or standardized
 incidence studies. Three of these were leukemia studies while the remaining one dealt with a
 study of melanoma. Of the three leukemia studies, only one reported a significantly elevated
 risk of chronic lymphocytic leukemia in Swedish electrical line workers, but not electrical
 workers as a general occupation, based upon census data and the Swedish Cancer Registry.
 The second reported a nonsignificant elevated risk of leukemia in a broad group of Finnish
 occupational categories that are assumed by the author to have a high potential for exposure
 to EM fields.  The highest of these was in linemen and cable splicers. The third, a
                                          3-145
                                                                                10/22/90

-------
                             DRAFT-DO NOT QUOTE OR CITE

  hypothesis-generating study, was an effort to identify occupations in the Portland-Vancouver
  Standard Metropolitan Statistical Association (SMSA) that appear to be at a high risk of
  leukemia based upon census bureau data, hospital records, and cases reported to a cancer
  registry. A fourth study reported a significant risk of melanoma in male Montreal
  telecommunication workers derived from hospital records reporting a diagnosis of melanoma.
  The mean age at diagnosis was younger than usual.
     There are four proportionate incidence studies, one dealing with central nervous system
  cancer (CNS), the second with leukemia, a third and fourth dealing with eye melanoma and
  leukemia, respectively, in England and Wales. The first found a significantly increased
  proportionate incidence rate  (PIR) of CNS cancer (all types) in Los Angeles electricians and
  engineers based upon data from the Los Angeles County Cancer Surveillance Program. This
  same database provided evidence of a significant risk of leukemia, chiefly acute myelogenous,
  in a second independent study of occupations with likely exposure to electrical and magnetic
 fields similar to those of Miiham.  Occupation was that given at the  time of diagnosis. The third
 noted a significantly increased risk of eye cancer in electrical and electronics workers
 identified as such by the Registrar General of Great Britain.  The fourth noted a significant risk
 of leukemia in 10 electrical occupations based upon data from the South Thames Cancer
 Registry.  These latter two studies the authors term proportionate registration ratio studies
 (PRR) because the diagnosis is "registered" in a cancer registry.
    All four of the proportionate mortality ratio (PMR) studies deal with leukemia as the cancer
 site. Three of these reported significant risks of leukemia in Washington State residents and in
 Wisconsin State residents who were occupationally exposed to electric and magnetic fields,
 that is, electricians, power station operators, and aluminum workers in Washington State and
 electrical engineers and radio and telegraph operators in Wisconsin.  One of these three
 studies also reported a significant risk of  non-Hodgkin's lymphoma in the same group of
 potentially exposed occupations. The remaining PMR study  reported increased risks of
 myeloid leukemia in specific electrical occupations, i.e., electrical and electronic engineers,
 and  telegraph and radio operators in England and Wales without providing any significance
tests.
   It should be kept in mind that PIR and PMR studies have inherent methodological
weaknesses. Estimates of site-specific cancer risks that are  derived from such studies are not
independent of each other, since the sum total of all observed events must equal the
       10/22/90
3-146

-------
                           DRAFT-DO NOT QUOTE OR CITE

expected. Furthermore, no measurements of actual exposure were available in any of these
studies. The authors are forced to rely on surrogates, such as employment in occupations
that have a potential for exposure to EM fields, without any proof to substantiate if any or how
much exposure actually took place.  Under these circumstances it is impossible to do
dose-response analyses.  This problem is further compounded by use of information on
occupation from cancer registers or death certificates. Random misclassification is probably
endemic in all of these studies. This particular form of bias will, in all likelihood, lead to a
reduction of the estimated risk toward the null.  The actual risk, if true, is likely to be higher.

3.4.4. Conclusions
    The strongest evidence that there is an association of certain forms of cancer (namely
leukemia, cancer of the CMS, and lymphoma) with exposure to magnetic fields comes from the
childhood cancer studies. Several studies have consistently found somewhat elevated,
statistically significant risks and elevated nonsignificant risks of these three site-specific
cancers in children whose exposure to magnetic fields has been estimated by the types of
wires near their homes or magnetic field measurements of 2 mG (0.2^T)or higher. Children
 do not have the same confounding influences that plague adults in the occupational studies.
 In fact, the potential confounders and biases that might have had an effect oh the data were
 examined by the authors in some detail and found not to be a serious problem.  No other
 agents have been identified to explain this association.  However, there are contradictory
 results within these same studies, and dose-response relationships could not be
 substantiated, except in Savitz et al. (1988), based upon limited information on wiring codes.
 Furthermore, there is little information on personal exposure and length of residency in areas
 exposed to EM fields.
     The studies of residential adult exposures to EM fields provide somewhat mixed evidence
 of a risk of leukemia, although they lack power and there is little evidence to substantiate
 exposure to levels of EM fields hypothesized as being associated with cancer. For these
 reasons these same studies cannot be used to argue that there is support for denying  that
 such an association exists.  On the other hand, the case-control study of cancer in Colorado
 residents does support an association of CNS cancer and lymphoma if proximity to
' high-current electrical wiring configurations is assumed to be a adequate surrogate for
 exposure.
                                          3-147
10/22/90

-------
                             DRAFT-DO NOT QUOTE OR CITE

     Additional, but weaker, evidence that there is an association of leukemia, cancer of the
 CNS, and perhaps other sites comes from the occupational studies of EM-field exposure.
 Although many of these studies have found an excess risk of these forms of cancer with
 employment in certain jobs that have a high potential for exposure to EM fields, but few or no
 measurements have actually been taken in those occupations. Furthermore, information
 about occupation has come generally from sources that could be characterized as sketchy.
 The likelihood that misclassification bias or information bias is present in these studies is a
 distinct possibility.
     The studies of RF radiation in adults exposed to microwave radiation or radar, particularly
 ham radio operators, produced mixed results. Again, the lack of definitive information about
 the kind, level, and length of exposure, as well as other confounders that may be responsible
 for the excess risks seen in the positive studies, precludes conclusions that a positive
 association exists with exposure to RF radiation.
     In conclusion, after an examination of the available epidemiologic data over the last 15
 years, there is evidence of a positive association of exposure to magnetic fields with certain
 forms of site-specific cancer, namely leukemia, cancer of the CNS, and, to a lesser extent,
 lymphomas.  This is supported by many studies of children and adults across many different
 populations and subgroups, and in jobs and conditions in which there is a high potential for
 exposure to EM fields. Attempts to identify potential confounders have been made, but no
 single confounder has been identified that could explain the positive results.  Much more work
 needs to be done to better refine exposure. This is more problematic for adults than for
 children. A dose-response relationship with magnetic field strength has not been firmly
 established, and potentially confounding influences have not entirely  been ruled out; therefore,
 a sufficiently strong case for causality has not yet been made. However, there is a link between
 exposure to EM fields and certain forms of site-specific cancer, namely leukemia,  CNS, and
 lymphoma.

 3.5.  OTHER END POINTS
   Several epidemiologic studies have evaluated end points other than cancer, chiefly
reproductive effects in work environments or in relation to specific  products such as video
display terminals, electric blankets, and heated water beds. These reports were not reviewed
here because the purpose of this review was to focus solely on cancer.
       10/22/90
3-148

-------
                          DRAFT-DO NOT QUOTE OR CITE
3.6. ONGOING RESEARCH
   Many epidemiologic studies are now being conducted and should be completed over the
next few years. These studies are being conducted in several countries and examine children
and adults and residential and occupational exposures.  This research is summarized in Table
3-40.
                                        3-149
10/22/90

-------
                   DRAFT-DO NOT QUOTE OR CITE







CANCER
Q
CO
3
LU
U.
o
tu
^
ELECTROMAGI
Z
0
CO
LU
5

CO
0
8
o
LU
Q
Q.
LU

Z
o
0
z
o

TABLE 3-40.


O CO
=ť ra
Q Q
CO
0)
^J
§
Q
I
O
4_>
05
3 C
|i
Q. o
LU co
*^


Study
Type:Cases/
Controls

_c
D_
1
LU


O
CO
Q.
CO





E

Institution

































CO
_a
CO
1. Occuoational

^lj^
1

^
^
[JL
GC
CC
S"


0





°0
o

o
_r
=r
_j-_l
< O





S
D_
LU

C
CO
i
m
•§.
Q)
CO
O

1
O
CO


i




-3


Q





§
CO
8
CM
8

_i
o
§•
UJ
O
*
<





cc
0.
LU

_o
'^.
Genevieve
Matanoski, Pat

JHU-SPH USA


1




—3
CD


Q





6
b
O




S
<


oc
Q.
LU
co"

LU
z

.0

Breysee
Genevieve
Matanoski, Pati

1
D.
CO
X
-)


1




cc








co"
C\J
O °


E
o
CL
1
J

To
o
CO
CO •_

11

CD

I
㥤.
S o
ll
03 O

Univ. of Bernd
Switzerland


c!



E
CO
0"
CO
6


— _





i
o
o




.0
o






z

D)
k_
Michael Salzbe

Monash MSe
Australia
Si
i
8 §


.C
o Ł
Ł§>Ł
|l o"
Emm
CO O 0_


Q ^
Q



o
CC:6000/1700
Co:215L, 186


C CO
2 Ł
<" o
0" § 0
to -5 m
_f E _T

y_
Q
LU
O _

o Ł
X LU



Giles Theriault
David Savitz
^
1
c <
111


CM




CO


_





§
1
o 6
m o




3
5




CL
LU
Z



Sergio Koifman
=
5
03
U.
8
X
D.
CO


S




CO
o"
CO
LU


_





O
o
CO
0
o
co
6
O

d
<
_r
Ťc
1
_r
<

CO
^>
o
c5
._
i

Iw
CO

Neill Pearce. Pe
Bethwaite

Wellington MS
New Zealand


1
CO
0"
CO
Q"
D_
U.
N~
CQ


=C
o




CC:200BC,
300L/1000

_f
t
O Ł
_f o


Birgitta Floderu

NIOH Sweden
10/22/90
3-150

-------
DRAFT-DO NOT QUOTE OR CITE





1
1
o
cc
LU
0
o
Q
Z
•^
CO
3
LU
Lt_
OMAGNETIC
cc
k
LU
LU

O
CO
LU
Q
Z>
CO
o
CD
0
EPIDEMIOL
CD
O
CD
O
o
CO
LU
CO
<






03 Ł
•^ CO
Q Q

2
CD
•o
0
8



Ł §
|i
X CO
LU 



c
'o
a.
c
LU




Q
CO
1
CO


EC




c
o
S
to


Cn CO ^P
CO Cn O5
en en en
^
CO
CO
D_
2
of
X 6 X



— Q —





CO CM CM
CO CO CO
Ł5 Ł{ o
• Z)
_ CO CO
CO "O "O
§_!<_!
•o ť_ co Ť•-
~ 0 13 0
ir .2 O >'
"- c co c
=3 z> z> z>

^
g



a>
o



—





en .
^f CD -
in -^ ?
?? 8" -
O a) r,
O a-e






O
CO
_r



co
Maastricht N


CO
.si
CO


CO
•n
CO C
2 45
S •
^ <~
o is
•c co
to 'Z-
u

•f-
"I

o
CO
O
Q.
Q"
a.
LlJ



Q





ih"~-
c^Ł?
^a^ci
-o 6Ť6o
3^0^ SS^

o
^f
O
"CO
co
O
CO
s



LU
CD
2*
g|
s-s
| s

c
o
S
2
0)
I



C/5
C •
S
s s
ll
5cl

.^.
S
o
o_
DC
0
< CO



Q





CM
CM
CD
CM
O
0





o
O
CO


CD
O
§
National Sci
Council


^
j
1-
DC

o
o
cc
c
CO
'co
1-

co


o"
a.
s>
Łto
Oof



Q





o
0
o
CO
8







0
m

CO
E
S
c5
•5.
0}
CO
CO


CO
c
C 0
CO CO
ť 2
CO 0-





<
CO
z>
0
CO

CO
g



o
Q.
DC



§





o
o
c5
o
O







^




X
o


J5
1
o
"c


o
c
2

•5
1
                                            CO
                                            CO

                                            O
                                            CO
                                            °0
                                            O
                                            8
                                            I
                                            O
                                            co


                                            I

                                            §
                                            13
                                            I
                                             CO
                                             (0
                                             CD

                                             "6
                                             >
             3-151
10/22/90

-------
                        DRAFT-DO NOT QUOTE OR CITE

o> .2
Q Q
CO
0
1
3
o
"c
o
O

if
If
o _w

O -
o5 o
o O
^
c
•5
Q.
1
UJ


Łr
S
O
Q.
CO




51


o
j





1
EQ -
. - CO
S a? p

-s Q"-
.G CO .ť
0_E-0-to
N ™ O C
CO ^Z ^^ CO
8
O
T—
8
o

§



jj
"IT







0


o
ŤSrŤ-S
Ł §|
= 3f
1 Sg
8 Ť S
•SiS

c CD
(U CO
2 O
03 O
CD ;<
^- o
^ "Z.

2? en
O> C7>
of

OL 3:. CL
o ,Ł o"
D. _ n
E Sfi
N NO
CD CD Q.

in m
O) T-
Ł3 CJ
in o
o) in
co *-
8" O
0




0 -J
0
§
X
!2f.
Q

< O
F %
Q. <;
ii i i
u.
o>
Richard Gallagh
Peter Boyle


A Canada
-ranee

^ o
^ 5
o <































. •
^r
CO
_o
^^
QJ
w
.5

ir
_Q
^

                                                                     s
                                                                      C
                                                                      e
                                                                      CO

                                                                   Ł  §
                                                                   3  OJ
                                                                   w CC
                                                                   o  ,_
                                                                   9-  C3
                                                                        ^.

                                                                             To
                                                                              cu


                                        I

                                        ^)
                                         CO

                                                                                      CO
                                                                                      Q)
                                                                                      JŁ
                                                   . . Q-
                                                                             D_
                                                                             CO
                                                                     .s

                                                                     I
                                                CO
                                          HI
                                          '•§, 9  5
                .
                g  "
CD
             Ť
     -
                                                                                     f
o =
•2 O
% c
o cu
O CT

•S °

El
                                     Ťs
                                        <<<
-------
                         DRAFT-DO NOT QUOTE OR CITE
             o
             I
             I
             05
o>
x:
Q- •>.
5 ff
* 1
*-- ";
CO TJ
.E C
b -B
S TO
P "5
>• Q.
o o
a. a.

CO
o
I
8
o
CO
4-*
1
co
a.



S
^
o
to
0>
D_


0)
S
&
O
1
1
Cu

Q
1
|
O)
cr
Ł
"6
li
cc


1
o
"5
_o
7S
D.
0)
CC



1
1
8
co



ra
o
CO



1
1
o
CO
cc
^}
.c
1
"5
>>
1
'c
•=>


B
o
CO
"5
>.
'1
1
D


QJ
1
i1
IS
•c
O
CO  .. ••  ..
O  D O  CO
a.  a. a.  a.
>  1 6 ....
O  LL O O E
a.  cc cc co co
O 6
Z CO

3
n Table 3-40 (contin
1
3
CO
c
o

"ro
1
^5
o
<


cn
1
a
"5
o
I
*~o

<^



^
2


I
8
"•5
0}
s



CO
s


0>
S
co
1
0
Z



3
Z

1
to
p
0}
111
"b
1
CO
"CO
o
CO
Z



lii
CO
Z


1 Cancer Institute
S
o
CO
Z



o
Z
o>
1
_g
1 Energy Administral
o
co
Z



OJ
Z
al Health Sciences
•E
CO
dgkin's lymphoma
1 Institute of Environ
O CO
? -2
O CO
Z Z



CO
X m
Z Z
1 Health
to
I
1 Institute of Occupa
CO
§
CO
Z



0
Z
Osvido Cruz Foundation
8
5
a.
•5
•o
o
o
CO
"5
.0
CO
Z
uJ
O
O
a.
CO
Z


1 Taiwan University
CO
_g
13
Z



g
Z


2
|
is -a
5 S
O -3



x ••
o Ł
CO


Q)

O
O



O
"^—

•D
0)
i*^
Q.
CO


1


§
CO
                                     o
                                    s
                                     0>
                                     o

                                     1
                                     0)
                                     Q.
                                     T3


                                     "c

                                     Q.
                                     0)
                                     CE
                                     §

                                     O
                                     CO
                                       3-153
                                                                   10/22/90

-------
                            DRAFT-DO NOT QUOTE OR CITE
REFERENCES FOR CHAPTER 3

Ahlbom, A. et al. (1987) Biological effects of power line fields, New York State Power Lines
       Project, Scientific Advisory Panel Final Report (July 1).

Ahlbom, A. (1988) A review of the epidemiologic literature on magnetic fields and cancer.
       Scand. J. Work Environ. Health 14:337-343.

Aldrich, T.E.; Glorieux, A.; Castro, S. (1984) Florida cluster of five children with endodermal
       sinus tumors. Oncology 41:233-238.

Barregard, L; Jarvholm, B.;  UngethCim, E.  (1985) Cancer among workers exposed to strong
       static magnetic fields. Lancet  892(October 19).

Bonnell, J. (1983) Leukemia and electrical workers. Lancet 1:1168.

Broadbent, D. et al.  (1985)  Health of workers exposed to electric fields.  Br. J. Ind. Med.
       42:75-84.

Brown, H.D.; Chattopadhyay, S.K.  (1988) Electromagnetic-field exposure and cancer.  Cancer
       Biochem. Biophys. 9:295-342.

Calle, E.; Savitz, D.A. (1985) Leukemia in occupational groups with presumed exposure to
       electrical and magnetic fields.  N. Eng. J. Med. 313:1476-1477.

Coleman, M.; Beral, V.  (1988) A review of epidemiological studies of the health effects of
       living near or working with electricity generation and transmission equipment. Int. J.
       Epidemiol. 17(1):1-13.

Coleman, M.; Bell, J.; Skeet, R. (1983) Leukemia incidence in electrical workers. Lancet
       982-983 (April 30).

Coleman, M.; Bell, C.M.J.; Taylor, H-L; Thornton-Jones, H. (1985, December) Leukemia and
       electromagnetic fields:  a case-control study.  In: Electric and magnetic fields in
       medicine and biology.  London, England: IEE Conf. Pub. No. 257, pp. 122-123.

Coleman, M.P.; Bell, C.M.J.; Taylor, H-L; Primig-Zakel, M. (1989) Leukemia and residence near
       electricity transmission equipment:  a case  control study. Br. J. Cancer 60:793-798.

De Guire,  L.; et al. (1988) Increased malignant melanoma of the skin in workers in a
       telecommunications  industry. Br. J. Ind. Med. 45:824-828.

Doll, R.; Muir, C.; Waterhouse, J., eds. (1970)  Cancer incidence in five centuries. Vol II.
       Berlin:Springer-Verlag (for the International  Union Against Cancer).

Easterly, C.E. (1981) Cancer link to magnetic field  exposure:  a hypothesis.  Am. J. Epidemiol.
       114 (2) :169-174.
       10/22/90
3-154

-------
                            DRAFT-DO NOT QUOTE OR CITE
Elder, J.  (1987, July) A reassessment of the biological effects of radiofrequency radiation:
       noncancer effects.  A report to the Office of Radiation Programs, U.S. Environmental
       Protection Agency, Washington, DC. (Unpublished)

Environmental Epidemiology Program, State of Hawaii Department of Health. (1986) Cancer
       incidence in census tracts with broadcasting towers in Honolulu, Hawaii. Report to the
       City Council, City and County of Honolulu, Hawaii.

Flodin, U. et al. (1986)  Background radiation, electrical work, and some other exposures
       associated with  acute myeloid leukemia in a case-referent study.  Arch. Environ. Health
       41(2):77-84.

Fulton, J. et al.  (1980)  Electrical wiring configurations and childhood leukemia in Rhode
       Island. Am  J. Epidemiol. 111:292-296.

Gilman, P.A.; Ames, R.G.; McCawley, M.A.  (1985) Leukemia risk among U.S. white male coal
       miners.  J. Occup. Med. 27(9):669-671.

Guberan, E., et al.  (1989)  Disability, mortality, and incidence of cancer among Geneva
       painters and electricians. Br. J. Ind. Med. 46:16-23.

Hill, D. (1988) A longitudinal study of a cohort with past exposure to radar: the MIT Radiation
       Laboratory follow-up study, [dissertation]. Ann Arbor, Ml:University of Michigan
       Dissertation Service.

Hoar, S.K.; Morrison, A.S.;  Cole, P.; et al. (1980) An occupation and exposure linkage system
       for the study of occupational carcinogenesis. J.  Occup. Med. 22:722-726.

Howe, G.R.; Lindsay, J.P. (1983) A follow-up study of a ten-percent sample of the Canadian
       Labour Force. 1. Cancer mortality in males.  1965-73.  JNCI 70:37-44.

Hsieh, C.C.; Walker, A.M.; Hoar, S.K.  (1983) Grouping occupations according to carcinogenic
       potential: occupation clusters from an exposure linkage system.  Am. J. Epidemiol.
       117:575-589.

Juutilainen, J.; Pukkala, E.; Laara, E. (1988) Results of an epidemiological cancer study among
       electrical workers in Finland. J. Bioelectrics 7(1 ):119-121.

Kallen, B.; Malmquist, B.; Moritz, U. (1982) Delivery outcome among physiotherapists in
       Sweden: is nonionizing radiation a fetal hazard? Arch. Environ. Health 37:81-85.

Kaune, W.T., et al.  (1987)  Residential magnetic and electric fields. Bioelectromagnetics
       8:315-335.

Lilienfeld, A.M., et al.  (1978)  Foreign Service health status study-Evaluation of health status
       of Foreign Service and other employees from selected Eastern European posts.  Final
       Report, Contract No. 6025-619073. Dept. of State, Washington, DC. Available from:
       NTIS, Springfield, VA, PB-288163.
                                        3-155
10/22/90

-------
                           DRAFT-DO NOT QUOTE OR CITE
Un, R.; Lu, P. (November 1989) Abstract of presentation at the annual Department of Energy
       (DOE)/EPRI Contractors review on biological effects from electric and magnetic fields,
       Portland, Oregon. Available through DOE, Washington, DC.

Un, R. et al.  (1985) Occupational exposure to electromagnetic fields and the occurrence of
       brain tumors. J. Occup. Med. 27(6):413-419.

Linet, M.S., et al.  (1988) Leukemias and occupation in Sweden: a registry-based analysis.
       Am. J. Ind. Med. 14:319-330.

Loomis, DP.; Savitz, D.A.  (1989) Brain cancer and  leukemia mortality among electrical
       workers. Am. J. Epidemiol. 130:814 (abstract).

Matanoski, G.; Elliott, E.; Breysse. P. (1989)  Poster presented at the annual Department of
       Energy (DOE)/EPRI Contractors review on biological effects from electric and magnetic
       fields; November 1989; Portland, OR.

Mattsson,  B.; Wallgren, A. (1984)  Completeness of the Swedish cancer register. Non-notified
       cancer cases recorded on death certificates in 1978. Acta Radiologica 23(5):305-313.

McDowall, M.N. (1983) Leukemia mortality in electrical workers in England and Wales.
       Lancet 1:246 (January 29).

McDowall, M.N. (1986) Mortality of persons resident in the vicinity of electricity transmission
       facilities. Br. J.  Cancer 53:271-279.

Microwave News.  November/December 1989.

Milham, S., Jr. (1982) Mortality from leukemia in workers exposed to electrical and magnetic
       fields. N. Eng. J. Med. 307:249.

Milham, S., Jr. (1985a) Mortality in workers exposed to electromagnetic fields.  Environ. Health
       Perspect. 62:297-300.

Milham, S., Jr. (1985b) Silent keys: Leukemia mortality in amateur radio operators. Lancet
       1:812 (April 6).

Milham, S., Jr. (1988a) Increased mortality in amateur radio operators due to lymphatic and
       hematopoietic malignancies. Am. J.  Epidemiol. 127(1):50-54.

Milham, S., Jr. (1988b) Mortality by license class in amateur radio operators. Am. J.
       Epidemiol. 128(5)1175-1176.

Modan, B. (1988) Exposure to electromagnetic fields and brain malignancy:  a newly
       discovered menace? Am. J.  Ind. Med. 13:625-627.

Myers, A.; Cartwright, R.A.; Bonnell,  J.A.; Male, J.C.; Cartwright, S.C. (1985) Overhead power
       lines and childhood cancer.
       10/22/90
3-156

-------
                            DRAFT-DO NOT QUOTE OR CITE
 Obrams, G. Iris.  (1988)  Leukemia in telephone linemen  [dissertation].  Ann Arbor, Ml:
       University of Michigan Dissertation Service.

 Olin, R.; Vagero, D.; S. Ahlbon, S. (1985)   Mortality experience of electrical engineers. Br. J.
       Ind. Med. 42:211-212.

 Pearce, N.E., et al. (1985) Leukemia in electrical workers in New Zealand.  Lancet 1:811-812
       (April 6).

 Pearce, N.; Reif, J.; Fraser, J.  (1984) Case-control studies of cancer in New Zealand electrical
       workers.  Int. J. Epidemiol. 18(1):55-59.

 Pollack, H. (1979) Epidemiologic data on American personnel in the Moscow embassy.  Bull.
       N.Y. Acad. Med. 55:1182-1186.

 Preston-Martin, S.; Henderson, B.E.; Peters, J.M. (1982)  Descriptive epidemiology of central
       nervous system neoplasms in Los Angeles County.  In: Selikoff, I.J.; Hammond, E.G.,
       eds. Brain tumors in the chemical industry,  Ann.  NY Acad. Sci. 381:202-208.

 Preston-Martin, S. et al.  (1988) Myelogenous leukemia and electric blanket use.
       Bioelectromagnetics 9:207-213.

 Registrar General (1971) Decennial supplement, England and Wales 1961: occupational
       mortality tables. London, England: Her Majesty's Stationery Office.

 Reif, U.S.; Pearce, N.; Fraser, J.  (1989) Occupational risks for brain cancer: a New Zealand
       cancer registry-based study.  J. Occup. Med. 31 (10):863-867.

 Robinette, C.D.; Silverman, C. (1977) Causes of death following occupational exposure to
       microwave radiation (radar) 1950-1974. In:  Hazzard, D.G., ed. Symposium on
       biological effects and measurements of radiofrequency/microwaves. Food and Drug
       Administration, Rockvilie, MD. HEW Publication (FDA) 77-8026,  pp. 338-344.

 Robinette, C.D.; Silverman, C.; Jablon, S. (1980)  Effects  upon health of occupational
       exposure to microwave radiation (radar). Am. J. Epidemiol. 112:39-53.

Savitz, D.A. (1987) Case-control study of childhood cancer and residential exposure to electric
       and magnetic fields.  Contractor's final report to the New York State (NSY) Power Lines
       Project, Contract #218217. Available from NYS Health Department, Albany, NY.

Savitz, D.A.; Calle, E. (1987)  Leukemia and occupational exposure to electromagnetic fields:
       review of epidemiological surveys. J. Occup. Med. 29(1):47-51.

Savitz, D.A. et al. (1988) Case-control study of childhood cancer  and exposure to 60-Hertz
       magnetic fields.  Am. J. Epidemiol. 128(1 ):21-38.

Severson, R.K., et al.  (1988) Acute nonlymphocytic leukemia and residential exposure to
       power frequency magnetic fields. Am. J. Epidemiol. 128(1 ):10-20.
                                        3-157
10/22/90

-------
                           DRAFT-DO NOT QUOTE OR CITE
Sheikh, K. (1986) Exposure to electromagnetic fields and the risk of leukemia.  Arch. Environ.
       Health 41 (1):56-63.

Speers, M.A.; Dobbins, J.G.; Miller, V.S. (1988)  Occupational exposures and brain cancer
       mortality: a preliminary study of East Texas residents.  Am. J. Ind. Med.  13:629-638.

Spitz, M.; Johnson, C. (1989)  Neuroblastoma and paternal occupation. A case-control
       analysis.  Am. J. Epidemiol. 121 (6):924-929.

Standard Industrial Classification Manual. (1972) Washington, DC, U.S. Government Printing
       Office.

Stern, F.B., et al. (1986) A case-control study of leukemia at a naval nuclear shipyard.  Am. J.
       Epidemiol. 123 (6):980-992.

Stevens, R.G. (1987)  Electric power use and breast cancer: a hypothesis.  Am. J. Epideliol.
       125(6):556-561.

Swerdlow, A.J.  (1983) Epidemiology of eye cancer in adults in England and Wales, 1962-77.
       Am. J. Epidemiol. 118 (2):294-300.

Szmigielski, S., et al.  (1988) Immunological and cancer-related aspects of exposure to
       low-level microwave and radiofrequency fields.  In:  Marino, A., ed. Modern electricity.
       Marcel-Dekker, Inc.

Thomas, T.L et al. (1987) Brain tumor mortality risk among men with electrical and electronic
       jobs: a case-control study.  JNCI 79(2):233-238.

Tomenius, L (1986)  50-Hz electromagnetic environment and the incidence of childhood
       tumors in Stockholm County. Bioelectromagnetics 7(2):191-207.

Tomenius, L; Hellstrom,  L; Enander, B. (1982) Electrical constructions and 50 Hz magnetic
       field at the dwellings of tumour cases (0-18 years of age) in the county of Stockholm.
       In: Proc. of the Inter. Symp. on Occupational Health and Safety in Mining and
       Tunnelling; June 21-25; Prague.

Tornqvist, S., et al. (1986) Cancer in the electric power industry.  Br. J. Ind. Med. 43:212-213.

U.S. Department of Labor. (1977) Dictionary of Occupational Titles. Washington, DC:  U.S.
       Government Printing Office.

U.S. Environmental Protection Agency (U.S. EPA). (1984) Biological effects of radiofrequency
       radiation.  Elder, J.A.; Cahill, D.F., eds.  EPA 600/8-83-026F. Available from: NTIS,
       Springfield, VA; PB 85-120-848.

U.S. Environmental Protection Agency, Offfice of Radiation Programs (1985) Radiofrequency
       (RF) radiation measurement survey, Honolulu, Hawaii, May 14-25,1984.  Letter from
       J.E. Cannon,  Assistant Administrator for Air and Radiation, to M.S. Fowler, Chairman,
       Federal  Communications Commission {January 22).
       10/22/90
3-158

-------
                           DRAFT-DO NOT QUOTE OR CITE
United States Senate Committee on Commerce, Science, and Transportation.  (1979)
       Microwave irradiation of the U.S. Embassy in Moscow. Washington, DC, U.S.
       Government Printing Office, 26 pp.

Vagero, D., et al. (1985) Cancer morbidity among workers in the telecommunications industry.
       Br. J. Ind. Med. 42:191-195.

Waterhouse, J.; Muir, C.; Cordea, D.; et al., eds. (1976)  Cancer incidence in five continents.
       Vol III. Lyon, France: International Agency for Research on Cancer.

Wertheimer, N.; Leeper, E. (1979)  Electrical wiring configurations and childhood cancer.  Am.
       J. Epidemiol. 109:273-284.

Wertheimer, N.; Leeper, E. (1982)  Adult cancer related to electrical wire near the home.  Int. J.
       Epidemiol. 11:345-355.

Wiklund, K.; Einhorn, J.; Eklund, G. (1981)  An application of the Swedish
       Cancer-Environment Registry. Leukaemia among telephone operators at the
       Telecommunications Administration in Sweden.  Int. J. Epidemiol. 10(4):373-376.

Wilkins, J.R.; Koutras, R.A. (1988)  Paternal occupation and  brain cancer in offspring:  a
       mortality-based case-control study.  Am. J. Ind. Med.   14:299-318.

Wright, W.; Peters, J.; Mack, T. (1982)  Leukemia in workers exposed to electric and magnetic
       fields.  Lancet 1160:61 (November 20).
                                       3-159
10/22/90

-------

-------
                           DRAFT-DO NOT QUOTE OR CITE


                                 4. ANIMAL STUDIES
    This section evaluates the published reports of lifetime bioassays that have studied
 radiofrequency (RF) and modulated RF radiation.  These studies are:
    •  University of Washington (Guy et al., 1985) rat study of pulsed 2450-megahertz (MHz)
       radiation.

    •  Prausnitz and Susskind (1962) mouse study of pulsed 9.270-MHz radiation.
    •  Spalding et al. (1971) mouse study of unmodulated 800-MHz radiation.
    •  Szmigielski et al. (1982) mouse study of 2450-MHz radiation in three specialized mouse
       assays: growth of lung cancer explant tissue; growth of spontaneous breast tumors;
       and growth of benzo[a]pyrene-induced skin cancer.
    •  Baum et al. (1976)  rat study of high-intensity pulsed radiation spanning the frequency
       range from 0 to 50 MHz.
    No reports of lifetime animal exposures to extremely low frequency (ELF) fields have been
 published.
    A detailed analysis of unpublished histopathological records from the Guy et al. (1985) rat
 study is described here. There is a greater emphasis on this study than on the other studies
 reviewed in this section because it has not yet been published in any other form.  The main
 features of the study are described in Section 4.1.1 and its subsections, 4.1.2.6.1, 4.1.2.7, and
 4.7.

 4.1. UNIVERSITY OF WASHINGTON LONG-TERM RAT STUDY
   This study was published as a series of nine reports dated from September 1983 through
August 1985 (Guy et al., 1983a, b, 1985; Chou et al., 1983; Johnson et al.,  1983,1984; Kunz et
al., 1983,1984,1985).  In a summary of these reports, Dr. Robert McGaughy of EPA's Office
of Health and Environmental Assessment (OHEA) concluded that there was a need to have a
more complete animal-by-animal description of the survival and histopathologic findings than
was provided by the published reports (McGaughy, 1987). Consequently,  he asked Dr.
Lawrence Kunz, the pathologist in the University of Washington study, to provide a description
of the histopathologic findings for each animal along with tables summarizing these findings.
His report (Kunz, 1988) formed the basis for the analysis of the survival and histopathologic
                                        4-1
10/22/90

-------
                           DRAFT-DO NOT QUOTE OR CITE


data given in this document. The data were analyzed by Dr. Robert McGaughy and Sharon
O'Boyle of the Computer Sciences Corporation (CSC).
   The University of Washington study was designed to simulate, in experimental animals, the
maximum absorbed power (0.4 watts per kilogram body weight) of 450-MHz radiation which is
permitted by the 1982 American National Standards Institute (ANSI) 95.1 standard under the
worst-case conditions when people with the body weight of children are being continuously
exposed (Guy et al., 1983b). The frequency of 450 MHz was chosen by the U.S. Air Force,
which sponsored the study, to represent a typical midrange radar system.  For this purpose, a
frequency of 2450 MHz was chosen for the rat exposure. The rationale for this choice was that
the ratio of wavelength to maximum body dimension is about the same for these two forms of
radiation.

4.1.1.  Description of Study
4.1.1.1. Animal Facility
   An animal exposure facility was designed and built especially for this study (Guy et al.,
1983a). It consisted of two rooms, each capable of housing 100 rats.  The animals were
housed individually in waveguide chambers which allowed unrestrained movement during
irradiation. Radiofrequency power was delivered to one-half of the chambers. The location
within each room of the exposed and nonexposed chambers (or animals) was random. In
each room, there were 50 exposed and 50 nonexposed animals.  People could enter the room
only from clean hallways which in turn could be entered only through a shower after donning
autoclaved garments. Soiled cages and waste collectors were taken out of the room into a
common "dirty" hallway adjacent to both rooms. An air pressure gradient was maintained in
which air flowed from the clean hallways to the ceiling of the exposure rooms into the dirty
hallway. The air flow rate in the exposure rooms was approximately 22 exchanges per hour
and the ambient temperature was held constant at 21 ą 1°C.  Microwave power was delivered
continuously for 21.5 hours per day, allowing  the remaining time for daily maintenance
procedures.

4.1.1.2. Animals
    Sprague-Dawley rats were chosen for the  experiment for several reasons according to the
authors:  (a) Since it is an outbred strain, its genetic heterogeneity would mimic the genetic
variation in human populations; (b) Males grow rapidly to their adult weight, lessening the
       10/22/90
4-2

-------
                           DRAFT-DO NOT QUOTE OR CITE

trauma of drawing blood samples; (c) They are docile and easily handled in the laboratory.
The rats were Cesarean-born and barrier-reared animals ordered in the summer of 1980 from
Camm Research Institute in Wayne, New Jersey. The colony was serologically tested at Yale
University and found to be free of specific pathogens. They were fed Purina Certified
Autoclavable Rodent Chow which was checked for acceptable levels of required nutrients after
being autoclaved.

4.1.1.3. Microwave Exposure
    The delivery of RF power to the animals was carefully calculated to simulate human
exposure conditions.  The details are given in two reports (Guy et al., 1983b; Chou et al.,
1983). Scale models of the human body with synthetic fluids simulating the RF power
absorption of tissues were constructed. The power absorbed by these models was measured,
and the results were scaled to full-sized humans. The spatial distribution of the absorbed
power in the models was also measured with a computerized thermographic system.  This
system generated maps of the distribution of absorbed power in different regions of the body.
These maps reveal inhomogeneities in the absorbed dose with complicated patterns that
depend on body posture and polarization of the radiation. Similar measurements were made
on models simulating rat exposure, and the parameters of the rat waveguide exposure
chamber radiation were adjusted to give average and peak-to-average ratios analogous to the
human exposure. The authors found that, because the human shape is more irregular than
the rat shape, the peak-to-average ratios in humans can approach 13, whereas with rats the
ratio is typically only 2 to 3. The parameters of the animal irradiation chosen after the analysis
of these measurements were as follows (Guy et al., 1983b): input power  to each waveguide:
0.14 watts (W); average power density in the chamber: 0.48 milliwatts per square centimeter
(mW/cm2); average specific absorption rate (SAR):  0.4 watts per kilogram (W/kg) body weight
for 200-gram (g) rats and 0.1 W/kg for 800-g rats; predicted range of hot-spot power for a
330-g rat: 0.63 to 1.33 W/kg; frequency: 2450 MHz.
   The 2450-MHz microwave power was modulated with the pattern shown in Figure 4-1 (Guy
et al.,  1983a, page 41). It was first modulated with 10 microsecond-wide pulses  occurring  at a
rate of one pulse per 1.25 milliseconds. Then a train of 50 such pulses, each train lasting 62.5
milliseconds, was switched on and off every 0.125 seconds, so that there were eight of these
pulse trains occurring each second.  Therefore, in terms of frequency, this signal has three
predominant frequencies: (1) the microwave component at 2450 MHz; (2) the pulse repetition
rate of 50/(0.0625 seconds) = 800 pulses per second (pps); and (3) the square wave
                                         4-3
10/22/90

-------
                             DRAFT-DO NOT QUOTE OR CITE
Figure 4-1. Modulation characteristics of the pulsed microwave source: fifty 10-microsecond-wide pulses
per group, with a repetition rate of 800 pulses per second during the burst of 50 pulses.
  modulation of 1/(0.125 seconds) = 8 pps. This radiation was delivered to the animal exposure
  chambers continuously for 21.5 hours per day.  Starting at 8:00 a.m. each morning, 7 days per
  week, the microwave power was turned off for 2.5 hours while the animals were taken out of
  their cages, the required measurements on the animals were made, the cages were cleaned,
  the water and food supply were replenished, and the animals were returned to their cages.
  These operations were completed for the entire colony of 200 animals between the hours of
  8:00 a.m. and 12:15 p.m. each day.  One could say that the animals were experiencing
  another diurnal frequency component in the radiation of 1/(24 hours)  or 1.2 x 10'5 Hz as well
  as the three components listed above.
     Since a specific absorption rate of 0.4 W/kg is as high as 25% of the resting and 15% of
  the average metabolic rate of an old lethargic 600-g rat, and about 10% of the resting and 5%
  of the average metabolic rate of a young 200-g rat, the experiment included several
  measurements of thermoregulation, energy metabolism, and growth (Guy et al., 1983a).
         10/22/90
4.4

-------
                             DRAFT-DO NOT QUOTE OR CITE



However, body temperature was not measured.. The cage temperature was kept at 21 °C, which is
below the "thermoneutrality" temperature range (28° C to 33° C) in which the metabolic rate is a

minimum. The lower ambient temperatures were necessary to avoid placing an extra metabolic load on
the animals receiving the radiation.

    This description of the microwave exposure obviously involves several parameters and, as we have

stated in Chapters 1 and 7, the parameter, or the combination of parameters which is relevant to
potential carcinogenicity, is not known.


4.1.1.4.  Protocol of the Experiment

    The following quote from Johnson et al. (1984) succinctly describes the protocol.

    Two hundred male rats at 3 weeks of age were obtained from a commercial, barrier-reared colony and
    randomly assigned to exposed and sham-exposed treatment conditions. Throughout this [exposure]
    period all surviving animals were bled at regular intervals, and blood samples were analyzed for a
    panel of serum chemistries, hematological values, protein electrophoresis patterns, and thyroxine (14)
    and serum corticosterone levels. In addition to daily measures of body mass and food and water con-
    sumption, oxygen consumption and carbon dioxide production were periodically measured on a sub-
    population of the exposed and sham-exposed groups. At regular intervals throughout the study,
    activity was assessed in an open-field apparatus.  After 13 months, 10 rats from each treatment group
    were killed for immunological competence testing, whole-body analysis, and gross and histopathologi-
    cal examinations. [The histopathologic examinations were done on all grossly evident lesions and on
    36 tissues of each animal.] The surviving 23 rats were killed at the end of 25 months [following the be-
    ginning of exposure], and similar analyses were made of 10 from each group.

4.1.2. Results of the Study

    During the study, measurements were made on a variety of physiologic functions. We

state here the conclusions of these measurements without describing the results in detail. The
separate volumes describing the results are as follows:  Volume 4, Open-field behavior
corticosterone; Volume 5, Evaluation of the immune system's response; Volume 6,
Hematological, serum chemistry, thyroxine, and protein electrophoresis evaluations; Volume 7,
Metabolism, growth, and development; Volume  8, Evaluation of longevity, cause of death, and
histopathologica! findings.


4.1.2.1.  Behavior and Corticosterone

    Observation of the  degree of activity in whole body movements  was made every 6 weeks
during the first year and every 12 weeks during the second year. No differences in the
exposed vs. control groups were observed except in the first session. Plasma corticosterone
was measured every 12 weeks during the first year and just before the termination of the
                                           4-5
10/22/90

-------
                           DRAFT-DO NOT QUOTE OR CITE


experiment.  No significant differences were found, indicating no cumulative effects of the
exposure on anxiety, fear, or stress.

4.1.2.2. Immune Competence
   At the 13-month interim kill and following the end of the experiment, the following three
tests were performed:  (1) numbers of B- and T-cell antigen-positive lymphocytes and
complement-bearing lymphocytes; (2) in vitro test of the proliferation of spleen lymphocytes
induced by five mitogenic agents; and (3) the ability of spleen cells to form antibodies and lyse
cells in response to immunization with sheep red blood cells, which are T-cell dependent
antigens.
   The first test showed that at 13 months the exposed group had a significantly larger
number of B- and T-cells than did controls,  but no difference was seen at 25 months. There
were no differences at either time in the percentage or total number of cells with complement
receptors in the spleen, which the authors interpret as an indication that the exposure does
not alter the maturation of lymphocytes.
   In the second test, the exposure significantly enhanced the effectiveness of three mitogens
and significantly decreased the  effectiveness of a different mitogen at the 13-month test.
   In the third test, the exposure had no significant effect on the response to the antigen.

4.1.2.3. Blood Chemistry and Hematologic Measurements
   A group of 11 hematologic measurements and 21  serum chemical measurements and
serum protein electrophoresis patterns were performed every 6 weeks during the first year and
every 12 weeks during the second year.  There were no significant differences between the
exposed and control group for any of the measurements. Thyroxine levels were determined
every 12 weeks, and no effect of exposure was detected, although the expected age-related
decrease in thyroxine was observed in both groups. The authors  interpret this finding as an
indication that the functioning of the entire  hypothalamic—pituitary—thyroid endocrine system
which regulates tissue metabolism was not adversely affected by the exposure.

4.1.2.4. Metabolism
   Daily measurements of body weight and food and water consumption showed no
difference between exposed and control groups.  Food intake in these animals was almost
twice that of animals kept under ambient temperature conditions; this difference is similar to
       10/22/90
4-6

-------
                            DRAFT-DO NOT QUOTE OR CITE
 4

earlier results in the laboratory and can be attributed to the lower temperature of the cages
(21° C) as compared to "thermoneutral" conditions (25° C to 33° C) under which the metabolic
rate vs. temperature is at a minimum.
    Weight measurements of heart, brain, liver, kidney, testicles, and adrenal organs were
made on 10 animals in each exposure group at 13 months and on 12 animals in the exposed
group and  11 animals in the control group at the termination of the experiment. No
exposure-related differences in any organ occurred at 13 months.  At the end of the
experiment, only the adrenal weights showed differences between treatment groups.  Data for
the mean adrenal weight were as follows (Johnson et al., 1984):
    The authors observed that at the end of the experiment there was a statistically significant
                                 Control
                           Adrenal
                       Weight, grams         Number
                        (mean ą SES        of Animals
                                                                     Exposed
  Adrenal
Weight, grams      Number
 (mean ą SE^      of Animals
Animals with
 adrenal tumors
Animals without
 adrenal tumors
All animals
0.069 ą 0.004
0.068 ą 0.005
4
7
0.132 ą 0.016
0.092 ą 0.017
7
5
                        0.068 ą 0.004
                                             11
                                                           0.116 ą 0.014
                                                                              12
increase in adrenal weight in all animals.  However, when animals with benign adrenal tumors
were separated from the remainder in the analysis, there was no statistically significant
difference in weights in animals without tumors. Therefore, the authors concluded that the
increased adrenal weight was related to the tumors and irrelevant to metabolic processes.
The slightly higher weight in exposed animals without tumors (compared to control animals
without tumors) was attributed to one animal with a hyperplastic adrenal cortex that was
secondary to a pituitary tumor.
    Several chemical measurements of the body carcass were made on animals at 13 months
and at the end of the experiment. These were total moisture, total ash, total crude fat,
protein-bound nitrogen, nonprotein-bound nitrogen, six fatty acids, and 27 minerals.  No
exposure-related differences occurred in any of these measurements.  The authors interpreted
this to mean that there is no evidence that the RF treatment irreversibly altered the body's
metabolic processes.
    Oxygen consumption and carbon dioxide production were measured under the same
waveguide conditions as in the long-term study but after termination of the experiment with a
                                          4-7
                   10/22/90

-------
                           DRAFT-DO NOT QUOTE OR CITE
                                                                                  ť
separate set of 36 animals obtained from the same suppliers. Measurements were performed
daily for 30 days encompassing both daytime and nighttime hours. No exposure effect was
observed in mature rats for either daytime or nighttime period.  However, in young animals the
nighttime COz production, Oa consumption, and the ratio of CO2 production to Oa
consumption were lower in the exposed group than the control group. The authors explained
this finding with the postulate that:  (a) under the slightly lowered temperature conditions of the
experiment, the absorbed radiation in the exposed group contributes to the maintenance of
the core temperature, whereas in the control group, increased metabolism supplies the heat
necessary to maintain the core temperature; and (b) in young smaller animals, the specific
absorption rate (W/kg) is larger than in older, larger animals.

4.1.2.5. Survival
   The times of death in the new pathology report (Kunz, 1988) were analyzed using the
"LIFETEST" procedure of the SAS set of statistical analysis programs (SAS, 1985). The results
of this test showed that the exposed group had a slightly longer time to 50% survival (669
days) than did the control group (653 days). However, the  difference in mean survival time for
the exposed group (649 ą17 days) compared to that of the control group (623 ą 19 days)
was not statistically significant. At the end of the 25-month  exposure period, 12 animals of the
exposed group and 11 animals of the control group were still alive (Figure 4-2).  The original
report (Kunz et al., 1985) came to the same conclusions qualitatively but did not give the
actual mean time to death or note the slightly longer survival time in exposed animals.
   The most frequent cause of death in both groups was kidney disease and urinary tract
blockage (Table 4-1).  Benign and malignant neoplasms were not frequent as a cause of
death, but pituitary adenomas (4 in exposed vs. 9 in controls), lymphosarcomas (4 vs. 2) and
pituitary carcinomas (2 vs. 0) each were the cause of more  than two deaths, and eight other
types of benign or malignant neoplasms were the cause of one death each. There were no
statistically significant differences between the groups for any of the causes of death.

4.1.2.6. Histopathologic Findings
   The Kunz et al. (1985) report tabulated, for each animal in the study, the cage location, age
at death, mode of death (spontaneous or killed), number of nonneoplastic lesions, and
number of neoplastic lesions, classed separately as benign, primary, or metastatic. It also
listed the total number of nonneoplastic and neoplastic lesions by type for the exposed and
       10/22/90
4-8

-------
                           DRAFT-DO NOT QUOTE OR CITE
        z
        o
        5
        2
        O
        P
        m
 SURVIVAL
  1+	
                                 SURVIVAL  ESTIMATES
              i
           0.8+
0.61

0.4-t
           0.2 +
        I
  0 +
......—ť
   B—.B._ .—.	s—ť
          B--B—B  ť-*-*	ťť
             B—B	B~BBŤflR	8
                     B--BBBB ftg
                        BB-ŤŤSA
                             '•A*
                             Be • AH
                              BBA-ť
                               BBBM
                                BB.RR
                                 BBRA
                                  BBŤA-SŤ
                                    BB ťRR
                                     B-BB AA
                                      B-BRfl
                                       BBtB
                                         R'B
                                         ***•
               0    100   200   300   4-00  500  600   700
                                                   800  900
                                                      FATEDAY
   Figure 4-2. Kaplan-Meier analysis of survival: exposed (A) vs. control (B).

control groups separately.  However, it did not have the information necessary to specify the
number of animals with any particular tumor type or combination of types or the times of death
for these animals.
   The new report (Kunz, 1988) has a complete description of histopathologic findings for
each of the 200 animals.  A description of the findings for a typical animal is given in Table 4-2.
It shows the animal number, time of death, cause of death,  and the histopathologic diagnosis
for each tissue.
   At the EPA, Dr. Robert McGaughy (OHEA) and Sharon O'Boyle (CSC), converted these
descriptions to a computer database coded so that the animals with any arbitrary specified
attribute or combination of attributes could be selected and examined for any other attribute.
   This database was the source for the analysis presented below.

4.1.2.6.1.  Tumor incidence. Table 4-3 shows the number of animals with tumors of various
organ sites and tumor types, the time of appearance of the first tumor at each site, and the
crude incidence of combined benign and malignant tumors of each site.
                                         4-9
                                                                  10/22/90

-------
                          DRAFT-DO NOT QUOTE OR CITE
           TABLE 4-1. FREQUENCY TABLE FOR CAUSE OF DEATH:  ALL ANIMALS
 Cause of Death
                                                         Group
         Control
Exposed
Chronic progressive nephropathy
Urinary tract blockage
Final sacrifice
Interim sacrifice
Atrial thrombosis
Anesthetic/bleeding
'Pituitary adenoma
Indeterminable
Lymphosarcoma
Cardiomyopathy
Degenerative vacuoiar encephalopathy
Asphyxiation
Pituitary carcinoma
Gastric squamous papilloma
Congestive heart failure
Nephroblastoma
Auditory sebaceous carcinoma
Gastric hyperkeratosis
Squamous cell carcinoma
Myocardial hypertrophy
Pyelonephritis
Liposarcoma
Hemangiosarcoma
Adrenal carcinoma
Pancreatic islet cell adenoma
Encephalopathy
Urinary tract blockage (calculus)
Adrenal cortical carcinoma
Hemopericardium
Glomerulonephropathy
Hemorrhagic cystitis
Transitional cell carcinoma
Cardiac neurinoma
Enteroliathisis
Chronic suppurative nephritis
Cerebral hemorrhage
Cerebral thrombosis
          15
          17
          11
          10
           9
           9
           9
           4
           2
           2
           1
 16
  9
 12
 10
  7
  5
  4
  5
  4
  4
  3
  3
  2
  1
  1
  1
  1
  1
  1

  1
  1
  1
SOURCE: Kunz, 1988.
    10/22/90
4-10

-------
                                     DRAFT-DO NOT QUOTE OR CITE
                        TABLE 4-2.  SINGLE ANIMAL, MICROSCOPIC OBSERVATIONS
    Animal number: E6     Sex:  Male    Group:  (1) Exposed   Fate: (Day = 661) Spontaneous death
    Cause of death
    Bone marrow
    Adrenal
    Thyroid

    Parathyroid
    Trachea
    Esophagus
    Brain
    Heart

    Skeletal muscle
    Spleen
    Liver

    Lung

    Salivary gland, parotid
    Salivary gland, sublingual
    Salivary gland, mandibular
    Lymph node, cervical

    Thymus
    Lymph node
    Kidney
    Urinary bladder
    Testis
    Epididymis
    Prostate
    Seminal vesicle
    Stomach
    Duodenum
    Jejunum
    lleum
    Cecum
    Colon
    Pancreas
    Skin
    Eye
    Harderian gland
    Ear, middle
    Nasal tissues
    Spinal cord
    Stifle joint
    Pituitary
    Zymbal's gland
    Preputial gland
-Chronic progressive nephropathy
-Within normal limits
-Adenoma, cortical, focal, moderate
-Perifollicular cell adenoma, focal, moderate
 Atrophy, diffuse, moderate
-Within normal limits
-Within normal limits
-Within normal limits
-Within normal limits
-Cardiomyopathy, ventricle, L., multifocal, moderate
 Cartilage at heart base, focal, mild
-Within normal limits
-Hemosiderosis, diffuse, mild
-Inflammation, chronic, perilobular, multifocal, minimal
 Hepatocellular adenoma, focal, moderate
-Hyperplasia, lymphoid, peribronchial, multifocal, minimal
 Congestion, diffuse, moderate
-Within normal limits
-Within normal limits
-Within normal limits
-Histiocytosis, diffuse, moderate
 Hemosiderosis, diffuse, moderate
-Within normal limits
-Within normal limits
-Chronic progressive nephropathy, diffuse, moderate
-Within normal limits
-Within normal limits
-Within normal limits
-Within normal limits
-Within normal limits
-Within normal limits
-Within normal limits
-Within normal limits
-Within normal limits
-Within normal limits
-Within normal limits
-Within normal limits
-Within normal limits
-Within normal limits
-Within normal limits
-Within normal limits
-Within normal limits
-Within normal limits
-Within normal limits
-Cyst, focal, moderate
-Inflammation, chronic, periductal, multifocal, mild
-Within normal limits
SOURCE:  Kunz, 1988.
                                                      4-11
                                                                    10/22/90

-------
TABLE 4-3. CRUDE
DRAFT-DO NOT QUOTE OR CITE
INCIDENCE AND TIME OF APPEARANCE OF NEOPLASTIC LESIONS
Number of Animals Time of First Crude Tumor
with Tumors Tumor8 Wavs^ Incidence15
Site/Type
Adrenal cortex
adenoma
carcinoma
Adrenal medulla
bonlgn pheochromocytoma
Thyroid
parafollicular cell adenoma
parafolllcular cell carcinoma
Uver
hepatocellular adenoma
hepatocellular carcinoma
Pituitary
adenoma
carcinoma
Testes
Interstitial cell tumor
squamous cell carcinoma
Epldldymls
squamous cell carciorna
Pancreas
Islet celt adenoma
squamous cell carcinoma
Urinary bladder
benign transitional cell
malignant transitional cell
Stomach
gastric squamous papilloma
squamous cell carcinoma
Duodenum
squamous cell carcinoma
Lymph node
transitional cell carcinoma
Soft tissues, thorax
fibroma
hemanglo sarcoma
Mesentery
llposarcoma
transitional cell carcinoma
Lymphosarcomad

Control
12


1

9


1


21


0


0

2


0


4


0

0

0


0


3


11
1

1

9
0

0
1

21
0

0
0

0

2
0

0
0

4
0

0

0

0
0

0
0


12


7

12


3


19


2


1

2


2


4


1

1

2





4

Exposed Control
619
9
3
811
7
448
10
2
791
3
0
457
17
2
_
1
1
-
1
781
1
1
-
1
1
248
3
1
-
1
-
1
~
1
1
2
1
1
448

Exposed Control
589 12/85
(14%)

673 1/73
(1.4%)°
661 9/85
(11%)

590 1/85
(1.2%)

540 21/85
(25%)

626 0/85
(0%)

650 0/85
(0%)
650 2/85
(2.4%)

590 0/85
(0%)

639 4/85
(4.7%)

569 0/85
(0%)
590 0/85
(0%)
505 0/85
(0%)

552 0/85
(0%)

438 3/85
(3.5%)
Exposed
12/76
(16%)

7/67
(10%)c
12/76
(16%)

3/76
(3.9%)

19/76
(25%)

2/76
(2.6%)

1/76
(1.3%)
2/76
(2.6%)

2/76
(2.6%)

4/76
(5.3%)

1/76
(1.3%)
1/76
(1.3%)
2/76
(2.6%)

2/76
(2.6%)

4/76
(5.3%)
10/22/90
4-12

-------
                            DRAFT-DO NOT QUOTE OR CITE
   TABLE 4-3. CRUDE INCIDENCE AND TIME OF APPEARANCE OF NEOPLASTIC LESIONS (continued))
Number of Animals Time of Rrst Crude Tumor
with Tumors Tumora Idavs} lnniripnrpb
Site/Type
Skin
benign
sebaceous adenomas
basal cell tumor
keratoacanthoma
lipoma
pilomatricoma
neurilemoma, hypodermis
malignant
auditory sebaceous carcinoma
basal cell tumor
fibrosarcoma, dermis
Control


0
0
1
0
0
1

0
0
0
Exposed Control Exposed Control Exposed

615 656
2
1
1
1

0

1
1
1
bAdenomas occurred before carcinomas at every site.
 Crude tumor incidence is defined as the number of animals with tumors divided by the number alive at the
 time of the first tumor in each treatment group minus the number of killed animals.
^The incidence of adrenal medulla tumors is presented in Table 4-4 and accompanying text.
 In the control group, the three animals had lymphosarcomas in 15 different tissues, and in the exposed
 group, the four animals had lymphosarcomas in 21 different tissues.

    For organ sites where adenomas are considered to progress to malignant tumors, which is
true for most of the sites in the table, the total number of animals is given in a separate
column.  For skin tumors, however, these totals are not given because there is no clear
justification for considering these very different tumors a single entity.  The animals with
lymphosarcoma had widespread metastases, and the separate tissues in which this occurred
are not listed.
    Table 4-3 shows the time of death of the animal which had the  first tumor at each organ
site. In all cases adenomas occurred before carcinomas. The crude incidence is defined here
as the number of animals with a tumor at a site divided by the number of animals alive at the
time of the first tumor and at risk for the development of tumors for the rest of the natural
lifespan.  It is called "crude" because it is unadjusted for mortality and it represents only an
approximation of the number of animals at risk to the tumor. Ten animals in each group were
intentionally killed at day 448. The first tumor in the control group was a stomach papilloma at
day 248, when 95 animals were alive, and the first tumor in the exposed group was a
lymphosarcoma, appearing at day 438 when 86 animals were alive.  The number of animals
considered to be at risk for the development of lifetime tumors is 10 less than this, or 85
                                        4-13
10/22/90

-------
                            DRAFT-DO NOT QUOTE OR CITE


 control animals and 76 exposed animals., These latter numbers were used as the
 denominators for the crude incidence column.
    In a more detailed examination of the adrenal tissues, Dr. Kunz noted, in a letter to Dr.
 Robert McGaughy (July 5,1990), that although the adrenal glands of all 200 animals in the
 experiment were examined, the adrenal medulla tissue of some animals was missing from the
 histopathological section.  He also noted that for other animals the adrenal medulla tissue from
 both left and right glands was present, whereas for some animals only one of the two adrenal
 medullas was present. Table 4-4 shows the number of animals in each of these four
 categories (exposed or control and one medulla or two medullas present) and the frequency
 of animals with tumors in these categories.
    For the exposed group, the number of animals at risk for the detection of adrenal medulla
 tumors is equal to the number of animals with medulla tissue (90), minus the number with
 medulla tissue that died before day 438 (13), minus the number of serial sacrificed animals
 with medullas present (10), which is equal to 67 animals. In this calculation we noted that,
 although 14 animals died before day 438, one of them had no  medulla present. Similarly, for
 the control group, the number of animals at risk for the detection of adrenal medulla tumors is
 88 - 5 -10 = 73 animals. Therefore, the crude incidence of these tumors is 7/67 = 10% in the
 exposed group versus 1/73 = 1.4% in the control group; p = 0.023 using the Fisher's exact
 one-tailed test.  These incidence figures are listed in Table 4-3.
    In the interpretation of adrenal medulla tumors (Table 4-4) one should check whether the
 presence of a variable number of adrenal medulla tissues per animal (0,1, and 2) has
 appreciable effect on the tumor incidence. This is not likely to be a problem for the following
 reasons:  (a) there is no appreciable difference in the number of animals with no adrenal
 medulla between the exposed group (10 animals) and the control group (12 animals), and (b)

               TABLE 4-4. FREQUENCY OF BENIGN ADRENAL MEDULLA TUMORS

Category
1
2
3
4

Treatment
Group
Exposed
Exposed
Control
Control
No. of
Medullas
Present
2
1
2
1
Animals
in
Category
70
20
60
28
Animals with
Benign
Tumors
6
1
1
0
SOURCE: Adapted from information provided by Kunz in a letter to McGaughy (July 5,1990).
       10/22/90
4-14

-------
                            DRAFT-DO NOT QUOTE OR CITE


in the exposed group, the incidence of tumors among animals with two tissues present
(6/70 = 8.5%) is not significantly different than among animals with one tissue present
(1/20 = 5%).
   Several observations about the data in Tables 4-3 and 4-4 can be made.

4.1.2.6.1.1. Adrenal medulla tumors. Benign pheochromocytoma of the adrenal medulla is
the only lesion in Table 4-3 that has a statistically significant elevation in the crude incidence of
tumors in the exposed group compared to the control group (7/67 in the exposed group vs.
1/73 in the control group; p=0.023 using the Fisher's exact one-tailed test). This site has the
largest group difference in the number of animals with tumors (seven animals in exposed
versus one animal in control), and it has a relatively large difference in the time of appearance
of the first tumor (138 days earlier in the exposed group than in the control group). The single
tumor in the control group was detected at terminal sacrifice.
   In view of the steadily decreasing survival of the animals in both the control and exposed
groups as the experiment progresses, a late appearing tumor inevitably is detectable with less
sensitivity because fewer animals are at risk late in the experiment.  Therefore, the smaller
incidence in controls could be considered another aspect of the late occurrence in controls.
To emphasize this point, another definition of "crude" incidence of tumors could be
constructed, called the site-specific crude incidence rate:  the number of animals with tumors
of a particular site divided by the number of animals alive at the time of the first tumor at that
site.  With this definition, the crude incidence of adrenal medullary tumors is as follows: (a)
control group: 1/9=0.011, since only nine animals with adrenal medulla tissue were alive at
the terminal sacrifice time of 811 days, and (b) exposed group: 7/38 =  0.184.  This difference
in proportions is not significant using the Fisher's exact test at a probability level of p<0.05.  If
the "site-specific crude incidence" is calculated for the other tumor sites, thyroid tumors (but no
other sites) become statistically significant (9/76 = 0.125 in control  animals versus 12/44 =
0.279 in exposed animals; p=0.031).  The first thyroid tumor appears later in the exposed
group than in controls (as opposed to earlier with the adrenal medulla) and the difference in
the times of first appearance (213 days) is larger than with the adrenal medulla.
   Therefore, we are faced with a contradiction because the large difference in the time to first
death with adrenal medulla tumors in the two treatment groups causes two apparently
reasonable definitions of crude incidence to lead to different conclusions. One solution to  this
dilemma is to disregard the concept of site-specific crude incidence rate because we are not
                                         4-15
10/22/90

-------
                           DRAFT-DO NOT QUOTE OR CITE


sure that the underlying assumption (namely, that tumors at various sites in the body occur
independently of each other) is correct. The crude incidence in Table 4-3 assumes that
animals in each of the separate experimental  groups are at risk to all tumors after the first
tumor at any site.
   Another solution to this dilemma is to use time-adjusted statistical significance tests, both
on single tumor sites  (which assumes independence of the sites) and on relevant
combinations of sites (which assumes that the selected combination of sites acts with a
common mechanism). This approach is preferable because it avoids the use of a statistically
extreme event (the time of appearance of the first of several tumors) as a measure of when the
animals are at risk and at the same time adjusts for any differing mortality patterns between
experimental groups.  The latter problem does not exist in this case, since the control and
exposed groups have similar mortality patterns (see Figure 4-2).
   The time adjusted tests employed here are the Cox tests for homogeneity of the two
groups and for dose-related trend. They are  calculated with the use of the microcomputer
version 2.1 of the program described by Thomas et. al. (1977). < Version 2.1 and its
documentation were kindly supplied to the author R.E. McGaughy by D. G. Thomas of the
National Cancer Institute.
   The results of these statistical significance tests for the adrenal medulla
pheochromocytomas and the combined thyroid adenomas and carcinomas are given in Table
4-5.  For the adrenal pheochromocytomas both the Fisher exact test of the crude incidence
and the time-adjusted trend tests are significant (p=0.023 and p=0.045, respectively), whereas
the incidence of thyroid tumors is not significant in  either test. These results indicate that the
crude incidence metric is not giving a misleading index of tumor incidence and the
"site-specific crude incidence" is  not appropriate for this data set. It also shows that the
excess incidence of pheochromocytomas is not likely due to chance  and  is therefore likely to
be an effect of the microwave exposure.

4.1.2.6.1.2.  Malignant tumors of all sites.  Another observation that can be made about the
data in Table 4-3 is that malignant tumors (carcinomas and sarcomas) of all sites occur in
more animals in the exposed than in the control group. This includes endocrine and exocrine
glands, gastrointestinal tract, skin and soft tissues, urinary bladder, and mesenchymal tissue.
The crude incidence, the Fisher exact test for significance of the crude incidence, and the
time-adjusted tests of significance of all sites listed in Table 4-3 are given in Table 4-5. For
carcinomas of all sites combined as well as for carcinomas and sarcomas of all sites
       10/22/90
4-16

-------
                            DRAFT-DO NOT QUOTE OR CITE
         TABLE 4-5. STATISTICAL SIGNIFICANCE OF SELECTED TUMOR TYPES AND SITES
Crude Tumor
Statistical Significance1"
Incidence Fisher Exact Test
Site
Adrenal Medulla
Thyroid
All Sites
Glandular Organs"

Lesion Control
Pheochromocytoma
(Benign)
Adenomas & Carcinomas
Malignant Tumors
Carcinomas
Sarcomas
Adenomas & Carcinomas
Carcinomas
Adenomas
1/73
9/85
5/85
2/85
3/85
31/85
1/85
31/85
Time-adjusted Tests0
Exposed of Crude Incidence Homogeneity Trend
7/67
12/76
18/76
11/76
7/76
35/76
8/76
34/76
0.023
0.23
0.0012
0.0049
0.12
0.14
0.018
0.18
0.091
0.71
0.014
0.026
0.38
0.81
0.048

0.045
0.35
0.0068
0.013
0.19
0.40
0.024
~
^Adrenal cortex, adrenal medulla, thyroid, liver, pituitary, testes, epididymis, and pancreas.
 One-tailed p-values.
 Cox tests, calculated with Thomas et al. (1977) computer program.

 combined, all three tests are significant at a level of p<0.05, but sarcomas of all sites
 combined are not statistically significant.  The rationale for combining malignant tumors of all
 sites is that all tissues are exposed to the radiofrequency and the ELF fields to approximately
 the same intensity and there is no a priori reason why one individual site should be either
 singled out or excluded from this combination.

 4.1.2.6.1.3. Glandular tumors. Several endocrine and exocrine glands in Table 4-3 have a
 slightly larger incidence of malignant tumors in the exposed group than in controls, although
 none of them is individually statistically significant.  These are adrenal cortex, adrenal medulla,
 thyroid, liver, pituitary, testes, epididymis, and pancreas.  This group of sites is obviously a
 subset of the group of tumors at all sites, which was discussed in the previous section. They
 are discussed separately here because of the possibility that, being sensitive to circulating
 hormones, the functioning of these tissues might be altered by the interaction between plasma
 membrane hormone receptors and electromagnetic fields or field-induced currents. In Section
 5.5 of this document, this interaction is discussed for ELF fields, and the pulse modulation in
 this experiment (at 8 pps and 800 pps) might supply the requisite energy to induce such an
 interaction.
                                          4-17
10/22/90

-------
                           DRAFT-DO NOT QUOTE OR CITE


   If animals with tumors of these glands are combined, the count shown in Table 4-5 is
obtained.  Note that the incidence of carcinomas of these sites is statistically significant,
whereas the incidence of benign endocrine tumors is not statistically significant.
   Many of the animals with carcinomas in one organ also had adenomas in another organ in
each treatment group. If Tables 4-3 and 4-5 are compared, one can see that some animals
have multiple tumors of these organs. For example, in the control group there are 44 animals
that have adenomas of one of the glandular organs mentioned above (Table 4-3), whereas
Table 4-5 shows that there are only 31 animals with adenomas in one or more of these  organs.
Table 4-6 compares incidence data for the separate organs (from Table 4-3) with those for the
combined organs (from Table 4-5).
   The difference between the sum of animals with separate tumors and the number of
animals with combined tumors is the number with multiple tumors. This is shown in Table 4-7.
It shows that multiple  adenomas occur more frequently than multiple carcinomas in both
treatment groups but that the microwave exposure does not appreciably change the
multiplicity of either carcinomas or adenomas.  An increase in  multiplicity would be expected in
the exposed group if the field exposure effectively removed a hormonal factor like melatonin
which normally inhibits the growth of several different glandular tumors.
                TABLE 4-6. MULTIPLICITY OF TUMORS IN GLANDULAR ORGANS
                        Control
                                  Exposed
        Sum of Animals with
          Separate Tumors
           (Table 4-3)
  Animals with
Combined Tumors
  (Table 4-5)
Sum of Animals with
 Separate Tumors
   (Table 4-3)
  Animals with
Combined Tumors
   (Table 4-5)
Adenomas
Carcinomas
44
2
31
1
48
10
34
8
        TABLE 4-7. NUMBER OF ANIMALS WITH MULTIPLE TUMORS (Derived from Table 4-6)
Control
Adenomas
Carcinomas
13
1
Exposed
14
2
       10/22/90
           4-18

-------
                            DRAFT-DO NOT QUOTE OR CITE


    Another issue that needs to be explored is whether there are associations between tumor
sites, such that the appearance of a tumor in one organ (e.g., pituitary) predisposes the animal
to an increased risk of tumors in another organ. This could be explored by calculating a matrix
of pairwise coefficients and examining the significance of these correlations.

4.1.2.6.1.4. Exposure-induced changes in progression of tumors. Since Dr. Kunz
evaluated the severity of benign and malignant lesions, it is possible to explore whether the
exposure affects the severity of the tumors. For this purpose, the rank-sum test recommended
by Gart et al. (1986, p. 163) and Snedecor and Cochran (1980) was used.  In this test, ranks
are assigned to the various degrees of severity.  In order to assign these ranks, a listing was
made of the terms used by Dr. Kunz in describing hyperplasia, benign tumors, and malignant
tumors. This list showed that he used the terms minimal, mild, moderate, and severe for
hyperplasia and for adenoma but only the terms moderate and severe for carcinomas.  In view
of this, it  is logical to rank the progression of neoplastic lesions from minimal hyperplasia to
severe carcinoma as shown in Table 4-8.
    This test was done for the adrenal medulla, adrenal cortex, thyroid, and pituitary. One
example  of the data set for a rank test is given in Table 4-9 for adrenal cortical tumors.  For
these data, the test statistic was not significant (p=0.31), indicating that the severity of the
tumors was not significantly different in the exposed group than in the controls. The result of
these tests (Table 4-10) is that the difference in severity is not statistically significant between
the exposed group and the control group for any of the individual glandular tumors except for
the adrenal medulla. Therefore, the microwave treatment is increasing the  severity of adrenal
medulla pheochromocytomas but not any other individual site. Of greater interest is whether
the exposure is increasing the severity of the glandular tumors as a group.  However, an index


               TABLE 4-8. ASSIGNMENT OF RANKS TO STAGES OF PROGRESSION
Stage Assigned Rank
No tumor
Hyperplasia, minimal
Hyperplasia, mild
Hyperplasia, moderate
Hyperplasia, severe

0
1
2
3
4

Stage
Benign, minimal
Benign, mild
Benign, moderate
Benign, severe
Malignant, mild
Malignant, severe
Assigned Rank
5
6
7
8
9
10
                                        4-19
10/22/90

-------
                          DRAFT-DO NOT QUOTE OR CITE
   TABLE 4-9.  SEVERITY RANKINGS OF ADRENAL CORTICAL PRENEOPLASTIC AND NEOPLASTIC
                                      LESIONS
Number of Animals
Description of Lesion
No lesions
Cellular alterations
minimal
mild
moderate
severe
Benign tumors
minimal
mild
moderate
severe
Malignant tumors
moderate
severe
Assigned Rank
0

1
2
3
4

5
6
7
8

9
10
Exposed
2

14
21
14
1

0
1
8
0

1
2
Control
2

15
28
9
0

0
6
4
1

0
1
   TABLE 4-10. RESULTS OF TESTS FOR THE STATISTICAL SIGNIFICANCE OF GRADED TUMOR
                                     RESPONSE
               Gland
 Significance Level (p-value)
            Adrenal medulla
            Adrenal cortex
            Thyroid
            Pituitary
          0.012
          0.31
          0.111
          0.641
of composite severity of a whole group of tumors is difficult to define, so this question will

remain unanswered.

4.1.2.6.2.  Historical Controls. In order to find out if the stock of animals in this study has a

different or unusual response pattern compared to endocrine tumors from other

Sprague-Dawley rat colonies, it is useful to compare the incidence of primary neoplastic

lesions in the endocrine organs in the control group of the University of Washington study with

the corresponding incidence in other published studies in this strain. This comparison is

summarized in Table 4-11.
       10/22/90
4-20

-------
                           DRAFT-DO NOT QUOTE OR CITE
   TABLE 4-11.  COMPARISON OF SPONTANEOUS ENDOCRINE TUMORS IN SPRAGUE-DAWLEY RAT
                                        STOCKS
Tumor Site/Type
Adrenal
cortical adenoma
cortical carcinoma
University
of Washington
Control Exposed
11/85
(13%)
1/85
(1.2%)
combined cortical tumors 12/85
(14%)
pheochromocytomas
Pituitary
adenocarcinoma
Thyroid
C-cell carcinoma
medullary tumor
parafollicular
(light cell)
parafollicular
cell adenoma
parafollicular
cell carcinoma
1/73
(1.4%)
0/85
-
-
9/85
(11%)
0/85
(0%)
9/76
(12%)
3/76
(3.9%)
12/76
(16%)
7/67
(10%)
2/76
(2.7%)
-
-
10/76
(13%)
2/76
(2.6%)
Charles'3 DiablobHoltzmanb Sprague" Sprague6
Hapa CDa River SD SD SD Dawley Dawley
15/70 18/88 - 2/216
(21%) (21%) (0.93%)
3/70 0/88 - 1/216
(4.3%) (0%) (0.46%)
0/448 7/209 2/229 16/226 -
(0%) (3.3%) (0.87%) (7.1%)
16/70 13/88 9/448 1/209 7/229 13/134d -
(23%) (15%) (2%) (0.5%) (3.1%) (10%)
0/56 2/93 - - - - 1/216
(0%) (2.2%) (0.46%)
------ 1/216
(0.46%)
6/71 5/98 -
(8.5%) (5.1%)
12/466 8/196 9/200 15/223 -
(2.6%) (4.1%) (4.5%) (6.7%)
_______
_ ~ — _ _ Ť. _
*Anveretal. (1982)
oMacKenzie and Garner (1973)
°Altmanetal. (1985)
 A second laboratory had significantly smaller incidence (values not given).
                  '(
   The most direct comparison possible is with the Anver et al. (1982) report. They examined
two particular stocks of barrier-bred Sprague-Dawley rats housed under standard laboratory
(not barrier) conditions for neoplastic, inflammatory, and degenerative diseases. The Hap
stock, bred originally at NIH and raised at Harlan  Laboratories, Indianapolis, Indiana, was
observed between 6 and 29 months of age, and the longer-lived CrI:COBS[R1 CD[R] SD stock
or CD rats from Charles River Breeding Laboratories, Wilmington, Massachusetts, was
observed from 12 to 38 months of age. The authors found that, although the life expectancy
of the two stocks was different, the cumulative incidence of the diseases was about the same
on a lifetime basis.
                                        4-21
10/22/90

-------
                           DRAFT-DO NOT QUOTE OR CITE


   MacKenzie and Garner (1973) in an earlier study compared the spontaneous tumor
occurrence in Sprague-Dawley rats from four different suppliers. This information was
generated in the course of a large series of studies on irradiated foods.
   The Altman et al. (1985) report is a summary tabulation of the historical control incidence in
seven different laboratories.  It came from the beginnings of the Laboratory Animal Data Base
Program of the National Library of Medicine. It is judged less useful than the other two studies
because the animals and pathologic diagnoses came from more heterogeneous sources.
   For adrenal cortical tumors, Table 4-11 shows no consistent pattern.  The adenomas in the
University of Washington study control group are less frequent than those in the two stocks
reported by Anver et al.  (1982) but much more frequent than the less-reliable data compiled by
Altman et al. (1985).  The combined adenoma and carcinoma reflect the same pattern. The
carcinoma incidence in the University of Washington controls is midway between that in the
two strains reported by Anver and somewhat higher than that reported by Altman et al. (1985).
Clearly, the University of Washington controls are not inconsistent with the wide variation
reported in this literature.
   For pheochromocytomas, the control incidence in the University of Washington study
control group is much smaller than  the Anver et al.  (1982) data and within the wide  range of
the stocks reported by MacKenzie and Garner (1973).
   For pituitary adenocarcinomas,  the University of Washington control incidence matches
the CD and not the Hap stock of Anver's report and is somewhat higher than Altman's data.
   For the thyroid, follicular tumors have been deleted from this comparison and the
parafollicular, medullary, and c-cell  tumors are assumed to represent the same entity. The
zero incidence of carcinomas in the University of Washington controls is consistent with
Altman's data but decidedly lower than that in the other two reports. This discrepancy may be
due to the inclusion of benign tumors along with  carcinomas in the other two studies.
   The conclusion that can be derived from examination of historical control data is that the
control animals used in the University of Washington study are not demonstrably different than
in other similar rat strains. Therefore there is no firm basis for attributing the statistical
significance of the adrenal pheochromocytomas  or the combined endocrine carcinoma
response to an abnormally low control incidence. Although Kunz et al. (1985) pointed out that
the Incidence of malignant endocrine tumors was not appreciably greater than the range
reported in the literature, we hesitate to draw conclusions about the effect of the microwave
exposure from a comparison of incidence in the exposed  group with historical control data
reported in the literature.  However, we agree with Kunz et al. (1985) that the incidence of
       10/22/90
4-22

-------
                           DRAFT-DO NOT QUOTE OR CITE


tumors at any one site is not remarkably larger than commonly observed in other animal
colonies of this strain.

4.1.2.7. Summary of Results
   This was a carefully designed, well executed study in rats of pulsed radiofrequency
radiation at power levels calibrated to simulate human exposure at the upper limit allowed by
current standards. The applied power was at most only 25% of the resting metabolic rate and
was apparently well tolerated by the animals. The number of animals (100 control and 100
exposed) was moderately large and the histopathologic examination of the tissues was
extensive.  In retrospect, the chief drawback of the study is that only one treatment level was
investigated. A room temperature control group and the addition of females to the
experimental design would have helped to assess the role of temperature  and sex-related
hormonal factors.
   The data analysis has developed the following conclusions:
   •  The incidence of benign pheochromocytomas of the adrenal medulla was significantly
      higher in the microwave-exposed group than in controls, but the incidence was not
      remarkably higher than that in control groups of other Sprague-Dawley rat colonies.
   •  No other single type of tumor was significantly increased by the pulsed microwave
      treatment.
   •  The incidence of carcinomas alone and combined carcinomas and sarcomas of all
      sites combined is  statistically significant and higher in the treated group than in
      controls.  The rationale for combining all sites is discussed in the next section.
   •  The incidence of malignant tumors in the endocrine and exocrine glands as a group
      was significantly higher in the microwave-exposed group than in controls. However,
      the incidence of benign tumors of these organs and the incidence of total benign and
      malignant tumors were not elevated in the exposed group. Glandular tumors here
      consist of adrenal cortex, adrenal medulla, thyroid, liver, pituitary, testes,  epididymis,
      and pancreas. The rationale for this combination of tumor types is  discussed in the
      next section.
                                        4-23
10/22/90

-------
                           DRAFT-DO NOT QUOTE OR CITE


    •  Multiple adenomas occur more frequently than multiple carcinomas in both treatment
       groups, but microwave exposure does not appreciably change the multiplicity of either
       carcinomas or adenomas.
    •  There is no exposure-related difference in the severity of adrenal cortical, adrenal
       medulla, thyroid, and pituitary tumors when ranked according to a scale of progression
       from minimal hyperplasia to severe carcinoma.

4.1.2.8. Discussion of Results
4.1.2.8.1.  Benign Adrenal Pheochromocytomas. The significance to humans of the adrenal
pheochromocytomas induced by the microwave exposure can be discussed in relation to two
issues. One issue is whether there is a difference between humans and animals in the
"functionality," or the endocrine activity of the gland with tumors. In human adrenal
pheochromocytoma, plasma and urinary catecholamine levels are elevated (Bravo et al.,
1979), whereas there is typically no evidence of secretory granules or hypertension in rats with
these tumors. Bosland and Bar (1984) have measured  blood pressure and urinary
catecholamine metabolites in aged Wistar rats with spontaneous adrenal medullary
hyperplasia and pheochromocytomas. They found that the correlation between blood
pressure and urinary catecholamine metabolites in animals with hyperplasia or tumors is not
appreciably different than for animals without these lesions. They also found that most of the
hyperplastic lesions and tumor cells show little or nonexistent staining for chromatin granules,
in contrast to the heavy staining of normal adrenal medullary cells. Presence of these
granules indicates secretory activity. They conclude that spontaneous tumors do not produce
excess catecholamines in the Wistar rats as is the case with humans. They also review other
animal studies where adrenal medullary tumors induced by radiation, chronic nicotine, chronic
thiouracil, chronic estrogen, and growth hormone, and  acute alloxan exposures have been
correlated with endocrine function. The results are inconsistent, with radiation and nicotine
inducing tumors with no chromaffin staining or weak chromatin staining and thiouracil and
alloxan inducing chromatin-containing tumors. With growth hormone, the results  are
conflicting, with one study showing an increase in blood pressure and the  other study showing
no blood pressure effect.
    On the other hand, there are several chemical agents that induce adrenal
pheochromocytomas. Gopinath et al. (1987) list the following agents as inducers of
       10/22/90
4-24

-------
                            DRAFT-DO NOT QUOTE OR CITE


 proliferative lesions in the medulla: nicotine, reserpine, synthetic retinoids, thiouracil,
 neuroleptics, growth hormone, 4-chloro-m-phenylenediamine, 1,1,2-trichloroethane, Zomepirac
 (an analgesic), Blocadren (a beta-androgenic blocker), lactose, and sugar alcohols such as
 mannitol, sorbitol, xylitol, and lactitol. Several different mechanisms have been postulated.
 Roe and Bar (1985) point out that medullary tumors are correlated with high food intake and
 increasing weight gain, and that they are commonly associated with pituitary adenomas,
 thyroid tumors, adrenal cortical tumors, islet cell tumors of the pancreas, and interstitial cell
 tumors of the testis.  They indicate that the sugar alcohols could induce adrenal medullary
 proliferative disease by increasing the absorption of calcium from the gut, and discuss several
 facts that indicate a link between increased calcium absorption and catecholamine release by
 the medulla. They also stress that the  induction of medulla neoplasia in response to
 carbohydrate intake is restricted to the rat and does not occur in humans.
    In the University of Washington study, no blood pressure or urinary catecholamine
 measurements were made, and no chromaffin staining was carried out. Therefore, the
 endocrine activity of the induced tumors is not known. However, the absence of any effect of
 exposure on corticosterone levels, whole-body activity, blood serum chemistry, and thyroxine
 levels indicates that if increased catecholamine levels are being induced by the radiation, the
 hormone balance is not seriously disrupted.
    One could postulate that the pheochromocytomas induced in the University of Washington
 study have some relationship to the abnormally high food  intake in these animals due to the
 lower than normal ambient temperature of both the control and microwave exposed groups.  If
 such  a relationship does exist, it has some indirect causation since both control and exposed
 groups  had double the normal food intake but the pheochromocytomas affected the exposed
 group preferentially.
    The second issue is whether the presence of benign pheochromocytomas in the animals
 indicates that humans may be susceptible to malignant tumors. Tischler and DeLillis (1988)
 point  out that nodules from proliferating adrenal medullary tissue can be transplanted into
 syngeneic rats and develop into metastasizing tumors. Also a cell line has been cultured from
 the medulla of an irradiated rat which develops tumors upon transplantation into rats and
 which secretes norepinephrine. Therefore these lesions undoubtedly have the capacity to
 metastasize. Tischler and DeLillis quote figures of 2% (4 of 213) lung metastasis in one large
 study of F344 rats and state that it is comparable to the less than 5% frequency of metastasis
from human adrenal medullary lesions.  They conclude that most medium and large sized
 medullary nodules in rats  are neoplasms but that they are probably not malignant as judged
                                         4-25
10/22/90

-------
                           DRAFT-DO NOT QUOTE OR CITE


by the criteria applied to human tumors. Hollander and Snell (1976) similarly regard them to
be of at least low-grade malignancy.  The National Toxicology Program (NTP) has used the
adrenal pheochromocytoma response in male F344 rats as an indication of carcinogenicity for
at least two agents, reserpine and 4-chloro-m-phenylenediamine. However, in the absence of
ancillary evidence of endocrine dysfunction, this lesion is generally regarded as relatively
Innocuous and relatively unlikely to result in malignancy.

4.1.2.8.2.  Combinations of Separate Tumor Sites. The biological significance of an excess of
malignant tumors of all sites and an excess of malignant tumors in endocrine and exocrine
glands as a particular subset of sites is uncertain.  Most chemical carcinogens affect only one
or a few tissues, and this generalization is broadly-enough recognized to be reflected in EPA's
Guidelines for Carcinogen Risk Assessment (U.S.  EPA, 1986) in which the following statement
appears:  "A statistically significant excess of tumors of all types in the aggregate, in the
absence of a statistically significant increase of any individual type, should be regarded as
minimal evidence of carcinogenic action unless there are persuasive reasons to the contrary."
In this discussion, we will present reasons for combining tumors and assess whether they are
persuasive, starting first with tumors  of all sites.
    There are two reasons for combining tumors of all sites in this study. First, the distribution
of the "toxic agent" to the various tissues with EM  radiation is likely to be more homogeneous
than for chemicals, so that all tissues are exposed more or less equally by EM fields. There
are two components to this toxic agent, radiofrequency fields and pulse modulation at 800 pps
and 8 pps. The authors have shown with thermograms (Chou et al., 1983) that the
radiofrequency energy deposition is uniform within a factor of 2 or 3, with the peak fields being
located at the nose and tail.  The lower frequency components would be completely uniform in
distribution if magnetic fields were to be shown as the biologically effective agent, but would
be as variable as the tissue electric conductivity if the biologically effective agent were
field-induced currents. By contrast, the tissue distribution of chemical agents depends upon
blood flow rates, tissue partition coefficients, and tissue clearance rates which are expected to
be more variable than the electrical parameters.  In the absence of a quantitative evaluation of
these factors, which could be done with some effort, our judgment is that the tissue
distribution of the EM field "toxic agent" is more uniform than that of a "typical" chemical agent.
 However, due to the uncertain outcome of a quantitative analysis that has not yet been
        10/22/90
4-26

-------
                            DRAFT-DO NOT QUOTE OR CITE


 performed, there are currently no firm persuasive reasons for asserting that the distribution of
 the toxic agent is significantly more homogeneous for EM radiation than for chemicals.
    Other reasons for the variability of tissue responses are biochemical differences (presence
 of enzymes) and differences in cell proliferation rates.  For both of these factors, the effects of
 EM fields and chemicals are expected to be the same, since EM field effects are probably
 ultimately mediated by the same cellular mechanisms as chemicals.
    Therefore, we find no compelling persuasive reason, given the current state of knowledge
 of these factors, for deviating from the Guidelines for Carcinogen Risk Assessment with
 respect to combining carcinomas of all sites.  The conclusion is that the finding of a
 statistically significant excess of malignant tumors of all sites furnishes only minimal evidence
 of carcinogenic action.
    For the finding of excess malignant tumors in the endocrine and exocrine glands as a
 group, a different rationale was used for the combination. It is based on the hypothetical
 similarity of effects induced by the low-frequency component of the pulsed microwave
 exposure (which was 8 pulse trains per second and 800 individual pulses per second during
 each pulse train, see Figure 4-1) and effects induced by power frequency (50-Hz or 60-Hz)
 electric and magnetic fields, reviewed in Chapter 5.  The rationale consists of a generalized
 argument and a more specific one. The generalized argument is that all tissues in the body
 are potentially affected by low-frequency EM fields and that either ion currents induced by
 these fields or the direct action of magnetic fields alter chemical reactions associated with
 plasma membranes. It has  been shown (Section 5.5) that low-frequency pulsed magnetic
 fields inhibit the function of parathyroid hormone at its cell membrane receptor. If this happens
 with other hormone  receptors, then one can postulate that other endocrine and exocrine
 glands, whose secretory activity is regulated by small concentrations of circulating hormones
 interacting with plasma membrane receptors, would be especially sensitive to fields or ion
 currents that can alter receptor function. The more specific rationale for combining the same
 tumor types is that nocturnal pineal melatonin activity is known to be inhibited by ELF electric
 fields (Wilson et al., 1986) and that pineal gland function is closely coupled to the function of
 other glands (see Section 5.7.1). Melatonin is known to inhibit tumor growth-enhancing
 hormones like prolactin and estrogen. The postulate has been made that when the blood
 melatonin concentration decreases because of the action of EM fields on the pineal gland, a
tumor growth inhibitor has been reduced or effectively removed, thereby causing a stimulation
 of tumor growth.  Although only breast and prostate tumors have been discussed  in this
                                         4-27
10/22/90

-------
                           DRAFT-DO NOT QUOTE OR CITE


connection, the same regulation by melatonin might hold for other hormonally-regulated
endocrine organs as well.
   This rationale for combining endocrine tumors is consistent with the information in Chapter
5 and is therefore a reasonable hypothesis, but it obviously contains several untested
assumptions. It could be extended to include other hormonally sensitive tissues besides the
endocrine glands. One aspect of the experimental findings is inconsistent with the proposed
mechanism. If hormonally-stimulated growth of tumors were occurring in this experiment, one
would expect the treated animals to have more frequent and more severe benign tumors than
the control group. This was examined and found not to be the case. The  effect seems to be
the induction of malignant tumors without the appearance of precursor benign tumors. We
expect that informed opinions will differ about how persuasive this argument is for combining
endocrine gland tumors. Indeed, more than one reviewer of previous drafts of this document
has found it to be excessively speculative. In view of these opinions, we conclude that with
our current knowledge there is no persuasive reason, based on established laboratory
findings, to combine the tumors of three glands into a single group. The conclusion is that the
finding of a statistically significant excess of malignant tumors in all endocrine and exocrine
glands as a group furnishes only minimal evidence of carcinogenicity.  It is possible that future
research on mechanisms of action will supply enough information to strengthen this rationale.
   In this connection, it should be noted that the human observations supply some support
for the postulate that ELF fields induce tumors in a variety of sites. Leukemias and  central
nervous system tumors have been observed in several studies with residential power
frequency exposures. However,  in male telephone workers, breast tumors were observed in
one study with a  different type of electromagnetic field exposure than power frequency.  In
some occupational studies, malignant melanoma occurred.

4.2. PRAUSNITZ AND SUSSK1ND (1962) STUDY
    In the first report of a long-term microwave irradiation study in experimental animals, the
authors exposed male Swiss albino mice to microwave power for 59 weeks (14 months), 5
days per week for 4.5 minutes per day.  The power source was a radar transmitter with a
frequency of 9270 MHz (3.2 cm wavelength) modulated with 2 microsecond pulses at a pulse
repetition frequency of 500 pulses per second. The irradiation was carried out in a walk-in
anechoic chamber via a horn antenna. Doses in the initial dose finding experiments were
adjusted by varying the distance between cages and the horn antenna. The power density
was calibrated with a precision 1/2 dipole antenna connected to a thermistor.
       10/22/90
4-28

-------
                            DRAFT-DO NOT QUOTE OR CITE


    To establish what power levels to use for the long-term experiment, individual animals with
 rectal thermometers were exposed to fixed power densities, and the rectal temperature was
 measured as a function of time. It was found that over a power density range of 0.068 to 0.20
 mW/cm2, half of the animals died when the temperature reached 6.7° C above normal and that
 the duration of exposure required to reach the LD50 temperature varied from 16 to 3 minutes
 over the power density range. Over this range the total LD50 dose (proportional to power
 density x duration) is approximately 1.1 (mW)/cm2-rninutes.  For the long-term experiment, the
 authors chose a power level of 0.1 mW/cm2 for 4.5 minutes, which amounts to a total dose of a
 little less than one-half the LDso. They found that this pattern (0.1 mW/cm2 for 4.5 minutes)
 causes an average temperature rise of 3.3° C.
    In the long-term experiment, the groups of 10 mice were placed in special plastic cages for
 the irradiation, and they were housed 10 to a cage for the rest of the time. A total of 300 mice
 were used; 100 were sham-exposed, and 200 were irradiated for 4.5 minutes per day, 5 days
 per week for 59 weeks (14 months).  During the experiment, the following measurements were
 made: blood counts;  spot checks of urine for glucose; weekly weighing of all mice;  and
 periodic recording of body temperature. When dead animals were found, a histologic analysis
 was made of all tissues that had not undergone extensive postmortem autolysis. Histologic
 examinations were done on all animals at three kill times: (1)10 irradiated and 5 control
 animals at 7 months, (2) 20 irradiated and 10 control  animals at 16 months, and (3) the
 remaining 67 irradiated and 19 control animals at 19 months.
    The authors found that the irradiated animals had better survival (65%) than  controls
 (50%). The difference was especially pronounced at  10 months and later.  There was no trend
 in body temperature throughout the experiment. The body weights in the two exposure
 groups were not significantly different. The red and white blood cell counts were the same in
 both groups at 4 months.
    In the 60 irradiated and 40 control mice that died spontaneously during the experiment,
there were two adverse effects that were more severe in the irradiated than control animals:
 (1) testicular degeneration (atrophy with no sperm) occurred in  23/57 (40%) of irradiated
animals and in only 3/37 (8.1%) of control animals, and (2) cancer of the white cells or leucosis
was seen in 21/60 (35%) of irradiated animals and 4/40 (10%) of control animals. This
condition was described as either monocytic or lymphatic organ tumors or myeloid leukemia
in the circulating blood.
   At the 7-month interim kill, which was half-way through the exposure period, both exposed
and untreated groups had the same histologic findings with respect to blood cell counts,
                                        4-29
                                                                             10/22/90

-------
                           DRAFT-DO NOT QUOTE OR CITE


kidney congestion, seminiferous tubules, and lymphoid infiltration in the brain, kidney, and
duodenum. At the 16-month interim kill, which was 1 month after cessation of the exposure,
6/20 (30%) of the irradiated animals had leucosis vs. 1/10 of the controls.  At the final 19-month
interim kill, which was 4 months after cessation of the exposure, testicular atrophy was seen in
14/67 (21%) of the irradiated animals vs. 1/19 (5%) of the control animals, and testicular
weights were less in the irradiated group.  However at that time, leucosis, as measured by
abdominal lymphoma, had no higher incidence in irradiated animals (12/67,18%) than in
controls (4/19, 21%).
   The authors did not comment on their inconsistent findings of leukosis. It occurred in
animals dying spontaneously, which could occur at any time in the experiment. It did not
occur midway through the exposure or 4 months after cessation of exposure, but it did occur
one month after cessation of exposure.  Therefore, it might be a transient phenomenon which
takes longer than 7 months of irradiation to induce but which regresses after cessation of
exposure. The testicular atrophy, on the other hand, is apparently permanent and  does  not
regress after exposure.

4.3.  SPALDING ET AL. (1971) STUDY
   The authors exposed female RFM mice to 800 MHz (wavelength of 37.5 cm) microwave
radiation placed in the center portion of a rectangular 10-foot long wave guide. There were 24
animals in both the exposed and the control groups. The exposures took place 2  hours per
day, 5 days per week for 35 weeks (8.1  months) and the power level at the mice's  location was
43 mW/cm2. They measured red and white blood cell counts, hemoglobin, hematocrit,
voluntary activity (with an exercise device), body weight, and survival. These measurements
were done periodically (approximately 23 times in 100 weeks) and analyzed by pair-wise
differences between exposed and control groups. No histopathologic observations were
made. The authors found no statistically significant differences in any of these observations
between the exposed and control groups at any time during the study, with one exception:
the body weight of animals older than 86 weeks in the exposed group was larger than that in
controls. The mean survival time was 655 days (93 weeks).

4.4. SZMIGIELSKI ET AL. (1982) STUDY
    These authors measured the effects of 2450-MHz microwave radiation at power density
levels of 5 and 15 mW/cm2 in an anechoic chamber in three different mouse bioassays:
       10/22/90
4-30

-------
                            DRAFT-DO NOT QUOTE OR CITE

    •  Lung cancer colony assay in Balb/C mice, in which lung sarcoma cells from donor
       mice are injected intravenously into recipient mice, and the number of lung cancer
       colonies was measured after 1, 2, and 3 months of irradiation (2 hours per day, 6 times
       per week in the morning hours from 8:00 a.m. to 12 p.m.). Induction of lung tumor
       colonies in this assay has been interpreted in earlier work by these authors as a
       lowering of cell-mediated immune reactions.

    •  Observation of the time of appearance of spontaneous breast tumors in C3H/HeA mice
       during irradiation periods up to 12 months.
    •  Induction of skin cancer in Balb/C mice in various protocols combining microwave
       radiation (1, 3, and 5 months), painting with benzo[a]pyrene (BaP), and stress from
       solitary confinement for 1 to 8 months. The skin painting was carried out with a
       concentration that causes histologically proven skin cancer in 7 to 10 months in more
       than 85% of the mice. A 7-grade scale of severity was used to measure the progress of
       the skin tumors.  Another report by Szudzinski et al. (1982) describes the skin tumor
       assay in greater detail.
   The radiation resulted in a specific absorption rate in the animal of 2 to 3 milliwatts per g
(mW/g) for the  5 mW/cm2 power level and 6 to 8 mW/g for the 15 mW/cm2 power level. At
both power levels, the animals  were able to maintain the body temperature within normal limits.
   The results of the experiment, briefly summarized, are as follows:
   •  The number of lung nodules (and the standard deviation) in an unspecified number of
       animals observed after 3 months of irradiation in the control, chronic stress, 5 mW/cm2,
       and 15 mW/cm2 groups were 3.6 ą 2.2, 7.7 ą 2.0, 6.1 ą 1.8 and 10.8 ą 2.1 .  Similar
       numbers were observed after 2 months of irradiation.  The authors concluded that lung
       modules were induced in a dose-dependent manner, and the magnitude of the effect
       induced by the lower power level was comparable to the effect induced by chronic
       stress confinement.
   •   For the control, chronic  stress, 5 mW/cm2 and 15 mW/cm2 group, the times when 50%
      of the initial 40 mice developed breast tumors were 322, 255, 261, and 219 days,
      respectively.  The times  for all three treatment groups are significantly less than
      controls. The authors concluded that both the time of breast tumor occurrence and the
                                        4-31
10/22/90

-------
                           DRAFT-DO NOT QUOTE OR CITE

      survival time is reduced in a field-strength dependent manner and, as with the lung
      nodule growth, the effect at 5 mW/cm2 power level is similar to the chronic stress
      confinement treatment.
   •  The time of appearance of skin tumors induced by BaP was shortened when radiation
      for 1 or 3 months preceded the BaP application and when the radiation and BaP were
      given at the same time. In both protocols, the higher power density resulted in earlier
      tumors than the lower power density. In another protocol where chronic confinement
      stress preceded  or was concurrent with BaP administration, the stress closely
      duplicated the effect of the 5 mW/cm2 level of irradiation noted earlier.
   In conclusion, the authors state that microwave radiation is a risk factor in the development
of neoplasms, since it enhances the development of tumors in ail systems tested.

4.5. BAUM ET AL. (1976) STUDY
   This study examined the effects of long-term electromagnetic pulse (EMP) irradiation on
Sprague-Dawley rats. The EMP irradiation consisted of high-intensity pulses of
electromagnetic energy at a repetition rate of 5 pulses per second. Each pulse had an electric
field intensity of 447 kV/m with rise and decay times of 5 nanoseconds and 550 nanoseconds,
respectively. Irradiation was carried out continuously, except for 1 hour per day for sampling
and animal care, for 94 weeks (22 months), for a total of 2.5 x 108 pulses.
   Several different experiments were done:
    •  Bone marrow cellularity was measured weekly for 94 weeks in 300 irradiated males and
       an equal number of non-irradiated males. One-half of these animals underwent
       necropsy; representative specimens of 22 tissues were saved for histologic evaluation;
       and tissues were selectively screened for microscopic lesions;
    •  Peripheral arterial blood was collected from 20 irradiated and 20 non-irradiated males
       on alternate weeks throughout the 94-week period;
    •  Twenty females were exposed for 94 weeks and observed grossly for mammary tumors
       and other lesions;
    •  Five pairs of rats were irradiated for 13 weeks and their progeny were examined for
       abnormalities;
       10/22/90
4-32

-------
                            DRAFT-DO NOT QUOTE OR CITE


     • The fertility of males irradiated for 94 weeks paired with females irradiated for 9 weeks
       was measured.
    The results of all of these tests showed no statistically significant differences between
 irradiated and control groups. An earlier report (Skidmore and Baum, 1974) of male AKR/J
 mice exposure to the same fields for 38 weeks showed no leukemia induction, but these
 results were not discussed in the later (1976) report.
    The conclusion that BMP irradiation has no apparent adverse effect in rodents is
 apparently valid, but the carcinogenicity evaluation does not meet current standards of
 complete histopathologic evaluation of all tissues and careful reporting of tumor incidence in
 each tissue. In addition, it is likely that the animals absorbed very little of the microwave
 radiation, since the maximum frequency (20 to 50 MHz) was very much less than the resonant
 frequency of the rats (letter from H. Bassen, Walter Reed Army Institute of Research,
 Washington, DC, to R. E. McGaughy, 1989).  Unfortunately, Baum et al. (1976) did not
 measure the specific absorption in their experiment.
    This study is reported here only for completeness, and no attempt is being made in this
 document to evaluate  more recent literature on high-energy pulse radiation.

 4.6. STUDIES IN PROGRESS OR PLANNED
 4.6.1. Ontario Hydroelectric Power Company
    As part of its program for assessing the human health effects of 60-Hz fields, the Ontario
 Hydroelectric Power Company is currently planning a lifetime animal cancer bioassay. It is
 being planned jointly with the Health and Welfare Agency of the Canadian Government. The
 study is being  designed to test whether 60-Hz magnetic fields can induce brain cancer and
 leukemia in experimental animals with an 80% power for detecting doubling of the background
 rate.  Magnetic flux densities up to 2000 microtesla (aT) will be  used, and the exposure
 duration is to be 10 hours per day for 80 weeks. Both cancer-initiated and noninitiated
 animals will be tested.  As of December 1988, a principal investigator had not been selected.

 4.6.2.  U.S. Air Force
   The U.S. Air Force is sponsoring a study of chronic microwave exposure of C3H/HJ mice,
which have a high spontaneous mammary tumor incidence. The irradiation is being carried
out continuously for 18 months, and the field has a frequency of 435 MHz (wavelength of 6.9
meters) modulated to simulate radar signals with an estimated specific absorption rate of 0.32
                                        4-33
10/22/90

-------
                           DRAFT-DO NOT QUOTE OR CITE


W/kg body weight.  A special facility was constructed at Georgia Tech to carry out the study.
The actual microwave exposures began in February 1988, and no interim reports are
contemplated (letter from J. H. Merritt, Radiation Sciences Division, U.S. Air Force School of
Aerospace Medicine, to R.E. McGaughy, December 10,1987).

4.6.3. The University of Rochester
   Under contract with the Electric Power Research Institute, the University of Rochester is
undertaking a series of experiments to see what factors influence the growth rate of tumor
cells in rats and in cultures of a breast tumor cell line. They will be investigating magnetic field
effects on the mitotic activity of various human malignant cell lines, the repair of x-ray damage,
and the timing of the cell cycle.  They are also planning to measure the effect of electric fields
on the growth rate of transplanted mammary carcinoma in rats and on hormonal levels (letter
from Shin-Tsu Lu, Biophysics Department, Univ. of Rochester, to R.E. McGaughy, May 26,
1989).

4.6.4.  Additional Studies
   As a follow-up to the peer review workshop for this document held on June 28,1990,
several additional planned studies were brought to the attention of the authors of this
document.
   A study of leukemia in C57BL/6 mice induced by an initial gamma ray dose and a daily
treatment with circularly polarized 60-Hz magnetic fields for 15 to 24 months at an intensity of
0, 0.1, and 10 gauss (G) is being carried out by T. Makinodan at UCLA. (Submitted by Richard
Phillips and Richard Griesemer.)
   The following information on EM-field animal experiments currently in progress or planned
was submitted by Richard Griesemer:

    •  USA UCLA: Mouse Leukemia Promotional model using 60-Hz fields up to 10 G.  Model
       uses C57BL/6 mice with or without ionizing radiation to determine initiating or
       promotion by magnetic fields. Sponsor: EPRI (P.I. Makinodan).
    •  France:  460-Hz pulsed magnetic fields on C3H/Bi mammary tumors, uses
       spontaneous viral-induced mammary tumors (P.I. Bellossi).
    •  Canada, Ontario: 60-Hz magnetic field skin tumor promotion in Sencar mice using
       DMBA and TPA  (P.I. Brinkmann).
    •  Germany, Hannover:  50-Hz magnetic fields and DMBA in Sprague-Dawley rat breast
       model. Negative study,  but milligram quantities of DMBA (P.I.  Brinkmann).
       10/22/90
4-34

-------
                           DRAFT-DO NOT QUOTE OR CITE
    •  Canada, Ontario: Rat (F344) Brain cancer promotion using END as an initiator.
       Sponsor: Ontario Hydro.  Initial studies to start this summer.  NPT will provide some
       animals for these experiments.
    •  Canada, Nova Scotia: Multigeneration mouse study and lymphoma, 250-G exposure.
       Experiment suffers from pathology quality and engineering problems.  At 250-G there
       should be significant noise, vibration, and heating.
    •  Sweden: 500 and 5^T magnetic fields in enzyme-altered rat liver foci model. P.I.
       Holmberg started in November 1988.
    •  Sweden: 500 and 5/*T skin study in NMRI mice, 19 hours per day for 12 and 24
       months. P.I. Holmberg started January 1989.
    •  Sweden: 500,  50, and 5//T in NMRI mice initiated with NMU.  Planned 2-year study to
       start this year  (P.I. Holmberg).
    •  Japan: study supposedly underway.  No details available.
    •  USA,  National Toxicology Program (NTP): Long-term study of magnetic field in mice
       and rats in final design  stages, to begin in 1991 (P.I. Boorman).
4.7. SUMMARY OF LONG-TERM ANIMAL STUDIES
    Of the five long-term animal exposures to electromagnetic radiation reviewed in this
document and summarized in Table 4-12, none were done using ELF frequencies alone.  Two
studies (Spalding et al., 1971; Szmigielski et al.,  1982) were done using unmodulated
microwaves, but the other three used pulsed microwaves consisting  of a range of frequencies.
    The study using the lowest frequencies (Baum et al., 1976) resulted in no effect even
though the intensity of the fields was much greater than that seen in the other studies.  Only
one of  these studies (Szmigielski et al., 1982) had more than one power level so that the
information on "dose-response" is limited. Neither the Spalding et al. (1971) nor the Baum et
al. (1976) studies reported the incidence of histopathological lesions, so that one has difficulty
drawing conclusions about their apparent lack of an observed response.
    In the Prausnitz and Susskind (1962) study in Swiss albino mice,  lymphoma and leukemia
occurred more frequently in exposed animals that died spontaneously than in controls dying
spontaneously. These tumors were also more frequent in the exposed group than in the
control group of animals killed one month after the end of the exposure, but the excess was
not observed in animals killed 5 months after the end of exposure.
    In the Szmigielski et al. (1982) study in Balb/c and C3H/HeA mice, a more clear-cut effect
was observed than in  any of the other studies. The effect observed was a dose-dependent
                                        4-35

-------
                                   DRAFT-DO NOT QUOTE OR CITE
      TABLE 4-12. SUMMARY OF CHRONIC ANIMAL EXPERIMENTS WITH ELECTROMAGNETIC-FIELD
                                                  EXPOSURE
 Study
Frequencies
                              Exposure
 Intensity
                                              Duration
                       Animal
             Study     Species
            Duration    (Number)  Results of Exposure
University     2450-MHz,
of Washington  800 pps.

               8pps.

             Individual
             animal
             waveguides.
Prausnitz
and Sussklnd
(1962)
 9270-MHz,
  500 pps.
Spaiding      800-MHz.
etal. (1971)    Animals in
              waveguide.
SzmigielskI
etal. (1982)
 2450-MHz.
 Horn antenna
 in anechoic
 chamber.
 Baum et al.
 (1976)
 0-20 MHz,
   5 pps.
                0 mW/crrr
                0.48 mW/cm2
                (0.4W/kgBW)
                (0.4W/kgBW)
0 mW/cm2
0.1 mW/cm2
                 0 mW/cm
                 43 mW/cm2
0 mW/cm
5 mW/cm2
(2-3W/kgBW)
15mW/cm2
(6-8 W/kg BW)
                21.5hr/day,
                7 days/wk,
                25 mo
                                                          25 mo
                        Sprague- Benign adrenal medulla
                        Dawley  pheochromocytomas are increased.
                                Carcinomas at all sites
Rats
(200)
4.5min/day,  19 mo
5 days/wk,
59 wk,
(14 mo)
2 hr/day,
5 days/wk,
35 wk

2 hr/day
6 days/wk
447 kV/m     22 hr/day
                             100wk
                             (23 mo)
                                                          3 mo
                                                           12 mo
                                                          5 mo
                                                           22 mo
Swiss-
Albino
Mice
(300)
RFM
Mice
are increased.
Malignant tumors of endocrine
and exocrine organs as a
group are increased.
No affect on survival.
No acceleration of glandular
tumors. No increase in severity
of tumors except in adrenal
medulla.
No increase in benign
tumors except in adrenal medulla.
No change in the multiplicity of
adenomas or carcinomas.

Better survival in exposed
group. Testicular atrophy
persisting after exposure.
Lymphoma and leukemia occurring
late during exposure period but not
persisting after exposure.

No effect on blood counts, voluntary
activity or survival.
Histopathology not examined.
Balb c    Lung cancer colonies
Mice     induced with positive
         dose trend.

C3H/HeA Time for development of
Mice     spontaneous breast tumors is
         shortened in a dose-dependent manner.

Balb/c    Time for development of
Mice     BaP-induced skin tumors is shortened both
         when exposure precedes or is concurrent
         with BaP skin painting.

Sprague- No mammary tumors in
Dawley   females.
Rats     No bone marrow proliferation.
         No blood cell alterations.
         No developmental abnormalities.
         No fertility effects.
          10/22/90
                                        4-36

-------
                           DRAFT-DO NOT QUOTE OR CITE


shortening of the development time of tumors already present, rather than an induction of
tumors de novo. This occurred for spontaneous mammary tumors in C3H/HeA mice as well as
for the development of skin tumors induced by benzo[a]pyrene in Balb/c mice. It is interesting
that microwave exposure before BaP skin painting as well as microwave exposure
concurrently with BaP treatment accelerated the induction time of these tumors. This suggests
that the exposure is producing some factor or change in the tissues that interacts with BaP
when it is given later.
    There is a possibility that heating  of the tissue produces one of these factors, but the
mechanism of this effect is not known.  At the higher intensity of the Szmigielski et al. studies,
the absorbed microwave power (6-8 mW/g body weight) is about one-half of the basal
metabolic rate of 30-g mice (about 15 mW/g body weight).  This  is a larger heat load for the
animals to dissipate than in the University of Washington study, where the absorbed power
was only 5% of the active metabolic rate of young 200-g rats.
    The clear positive findings of Szmigielski et al. (1982) show that radiofrequency fields
without low-frequency components stimulate the growth of tumors and indicate that they may
act as a tumor promoter, or a modifying factor in the development of tumors. The role of
tissue heating as a mechanism for this effect is not clear.
    The University of Washington study, done at a power level carefully calibrated to simulate
human exposure at the  maximum continuous level allowed by the ANSI standard, showed the
induction of benign adrenal pheochromocytomas in the exposed group and no statistically
significant elevation at any other site. There was a statistically significant elevation in the
incidence of carcinomas, but not sarcomas, at all sites combined. There was also an elevation
of carcinomas in each of several glandular organs (pituitary, thyroid, adrenal cortex, pancreas,
testes, and liver), which was statistically significant if they are all considered as a single group.
This apparent induction of malignant tumors occurred without an increase in benign tumors of
those sites.  Benign adrenal  pheochromocytomas in rats are not hormonally active as they are
in humans, and they are generally regarded as innocuous and unlikely to result in malignancy.
    The combining  of carcinomas of all sites is based on the rationale that all tissues are about
equally exposed to the same magnitude of fields and induced currents. On closer
examination, this is not a persuasive argument because a difference in exposure to various
tissues is only one  of many factors responsible for differences in tissue response and because
even tissue exposure is not clearly more uniform for EM fields than for chemical agents.
Therefore, there is no clear reason to depart from the convention for chemical agents that, in
                                         4-37
10/22/90

-------
                           DRAFT-DO NOT QUOTE OR CITE


the absence of a significant response at any one site, a finding of statistical significance at all
sites combined furnishes only minimal evidence of carcinogenicity.
   Combining tumors across these glandular organs is based on the generalized hypothesis
that electromagnetic fields affect all tissues in the body and that these glands, being
specialized to respond to small amounts of specific circulating hormones, have cell
membrane-bound receptors whose function could be affected by ion currents induced by the
external fields. Although this concept is consistent with recent research on the mechanism of
tissue interaction with low-frequency fields, there are at least two reasons why the combining
of glandular organ tumors of all sites may not be appropriate. These are:

   1.  The proposed mechanism is based on extremely low-frequency (ELF) field phenomena,
       and there is some uncertainty whether the low-frequency pulse component of the
       radiation can induce the same effects as ELF fields.

   2.  This proposed mechanism, though plausible, has not been empirically established.

   Therefore, with the current lack of empirical facts to confirm this proposed mechanism of
cancer induction, the rationale for combining endocrine tumors of all sites is not persuasive.
The study can be said to suggest, but not to demonstrate, a carcinogenic effect of pulsed RF
radiation.
       10/22/90
4-38

-------
                           DRAFT-DO NOT QUOTE OR CITE
REFERENCES FOR CHAPTER 4

Altman, P.L; McLane, K.; Brasseau, J. (1985)  Pathology of laboratory animals. Elmsford, NY:
       Pergamon Press.

Anver, M.R.; Cohen, B.J.; Lattuada, C.P.; Foster, S.J.  (1982) Age-associated lesions in
       barrier-reared male Sprague-Dawley rats:  a comparison between Hap: (SD) and
       CRL:COBS[R] CD[R] (SD) stocks.  Exper. Aging Res. 8:3-22.

Baum, S.J.; Ekstrom, M.E.; Skidmore, W.D.; Wyant, D.E.; Atkinson, J.L  (1976) Biological
      • measurements in rodents exposed continuously throughout their adult life to pulsed
       electromagnetic radiation.  Health Physics 30:161-166.

Bosland, M.C.; Bar, A.  (1984) Some functional characteristics of adrenal medullary tumors in
       aged male Wistar rats. Vet. Pathol. 21:129-140.

Bravo, E.L; Tarazi, R.C.; Gifford, R.W.; Stewart, B.H. (1979) Circulating and urinary
       catecholamines in pheochromocytoma: Diagnostic and pathophysiologic implications.
       N. Eng. J. Med. 301 (13) 682-686.

Chou, C.K.; Guy, A.W.; Johnson, R.B.  (1983, October)  Effects of long-term low-dose
       radiofrequency radiation on rats.  Volume III. SAR in rats exposed in 2450-MHz
       circularly polarized waveguide. University of Washington, USAFSAM-TR-83-19.

Gart, J.J.; Krewski, D.; Lee, P.N.; Tarone, R.E.;  Wahrendorf, J.  (1986) Statistical methods in
       cancer research. Volume 3. The design and analysis of long-term animal experiments.
       Lyon, France: International Agency for Research on Cancer, IARC Scientific Publication
       No. 79.

Gopinath, C.; Prentice, D.E.; Lewis, D.J. (1987) Atlas of experimental toxicological pathology.
       Chapter 7, The endocrine glands. Boston,  MA:  MTP Press, Ltd.

Guy, A.W.; Chou, C.K.; Johnson, R.B.  (1983a,  September) Effects of long-term low-level
       radiofrequency radiation exposure on rats.  Volume 1. Design, facilities, and
       procedures. University of Washington,  USAFSAM-TR-83-17.

Guy, A.W.; Chou, C.K.; Newhaus, B.  (1983b, September)  Effects of long-term low-level
       radiofrequency radiation exposure on rats.  Volume 2. Average SAR and SAR
       distribution in man exposed to 450 MHz RFR.  University of Washington,
       USAFSAM-TR-83-18.

Guy, A.W.; Chou, C.K.; Kunz, L.L.; Crowley, J.;  Krupp, J. (1985, August) Effects of long-term
       low-level radiofrequency radiation exposure on rats. Volume 9.  Summary.  University
       of Washington,  USAFSAM-TR-85-64.

Hollander, C.F; Snell, K.C. (1976) Tumors of the adrenal gland. Vol.6.  Lyon, France:
       International Agency for Research on Cancer, pp. 273-295.
                                        4-39
10/22/90

-------
                           DRAFT-DO NOT QUOTE OR CITE
Johnson, R.B.; Spackman, D.; Crowley, J.; Thompson, D.; Chou, O.K.; Kunz, L.L.; Guy, A.W.
       (1983, December) Effects of long-term low-level radiofrequency radiation exposure on
       rats. Volume 4.  Open-field behavior and corticosterone.  University of Washington,
       USAFSAM-TR-83-42.

Johnson, R.B.; Kunz, LL; Thompson, D.; Crowley, J.; Chou, C.K.; Guy, A.W.  (1984,
       September) Effects on long-term low-level radiofrequency radiation exposure on rats.
       Volume 7.  Metabolism, growth and development.  University of Washington,
       USAFSAM-TR-84-31.

Kunz, LL; Hellstrom, K.E.; Hellstrom, I.; Garriques, H.J.; Johnson, R.B.; Crowley, J.;
       Thompson, D.; Chou, C.K.; Guy, A.W.  (1983, December)  Effects of long-term low-level
       radiofrequency radiation exposure on  rats. Volume 5. Evaluation of the immune
       system's response.  University of Washington,  USAFSAM-TR-83-50.

Kunz, LL; Johnson, R.B.; Thompson, D.; Crowley, J.; Chou, C.K.; Guy, A.W.  (1984, March)
       Effects of long-term  low-level radiofrequency radiation exposure on rats. Volume 6.
       Hematological, serum chemistry, thyroxine and protein electrophoresis evaluations.
       University of Washington, USAFSAM-TR-84-2.

Kunz, LL; Johnson, R.B.; Thompson, D.; Crowley, J.; Chou, C.K.; Guy, A.W.  (1985, April)
       Effects of long-term  low-level radiofrequency radiation exposure on rats. Volume 8.
       Evaluation of longevity, cause of death and histopathological findings.  University of
       Washington,  USAFSAM-TR-85-11.

Kunz, LL (1988)  Supplementary data on the pathologic diagnosis of rats in the University of
       Washington long-term low-level radiofrequency study. Report prepared by
       the Biogenetics Research Laboratories, Inc., Ellensburg, WA, for R.E. McGaughy, Office
       of Health and Environmental Assessment, U.S.  Environmental Protection Agency,
       under Purchase  Order 7W-0633-NASX.

MacKenzie, W.F.;  Garner, F.M.  (1973) Comparison of neoplasms in six sources of rats. J.
       Natl. Cancer  Inst. 50:1243-1257.

McGaughy, R.E.  (1987, May 20) Preliminary  evaluation of issues associated with human
       carcinogenicity of radiofrequency radiation. Attachment to a memorandum from R.
       McGaughy, Office of Health and Environmental Assessment, U.S. Environmental
       Protection Agency, to Sheldon Myers, Office of Radiation  Programs, Office of Air and
       Radiation.

Prausnitz, S.; Susskind, C.  (1962)  Effects of  chronic microwave  irradiation on mice. IRE
       Trans, on Biomed. Electron. 9:104-108.

Roe, F.J.; Bar, A.  (1985) Enzootic and epizootic adrenal medullary proliferative disease of
       rats:  influence of dietary factors which affect calcium absorption.  Human Toxicol.
       4:27-52.

SAS.  (1985)  SAS user's guide: statistics, version 5 edition. Gary, North Carolina:  SAS
       Institute.
       10/22/90
4-40

-------
                            DRAFT-DO NOT QUOTE OR CITE
Skidmore, W.D.; Baum, S.J. (1974) Biological effects in rodents exposed to 108 pulses of
       electromagnetic radiation. Health Phys. 26:391 -398.

Snedecor, G.W.; Cochran, W.G. (1980)  Statistical methods. Ames, IO: Iowa State University
       Press.

Spalding, J.F.; Freyman, R.W.; Holland, LM. (1971)  Effects of 800-MHz electromagnetic
       radiation on body weight, activity, hematopoiesis and life span in mice. Health Physics
       20:421-424.

Szmigielski, S.; Szudzinski, A.; Pletraszek, A.; Bielec, M.; Janiak, M.; Wrembel, U.K. (1982)
       Accelerated development of spontaneous and benzopyrene-induced skin cancer in
       mice exposed to 2,450-MHz microwave radiation. Bioelectromagnetics 3:179-191.

Szudzinski, A.; Pietraszek, A.; Janiak, M.; Wrembel, J.; Kalczak, M.; Szmigielski, S.  (1982)
       Acceleration of the development of benzopyrene-induced skin cancer in mice by
       microwave radiation. Arch. Dermatol. Res. 274:303-312.

Tischler, A.S.; DeLillis, R.A. (1988) The rat adrenal medulla. II.  Proliferative lesions. J. Amer.
       Coll. Toxicol. 7(1) :23-44.

Thomas, D.G.; Breslow, N.; Gart, J.J. (1977) Trend and homogeneity analyses of proportions
       and life table data. Computers and Biomedical Research. 10:373-381.

U.S. Environmental Protection Agency. (1986)  Guidelines for carcinogen risk assessment.
       Federal Register 51:33992-34003.

Wilson, B.W.;  Chess, E.K.; Anderson, LE. (1986) 60-Hz electric field effects on pineal
       melatonin rhythms: time course for onset and recovery.  Bioelectromagnetics
       7:239-242.
                                         4-41
10/22/90

-------

-------
                           DRAFT-DO NOT QUOTE OR CITE

                  5. SUPPORTING EVIDENCE OF CARCINOGENICITY
   The objective of this chapter is to supplement the discussion of direct carcinogenicity
observations in humans and in rodent experiments with ancillary evidence that has some
bearing on the interpretation of those studies or gives some insight about the mechanisms of
carcinogenesis. For chemical agents, the U.S. EPA carcinogen risk assessment guidelines
suggests that physical and chemical properties of the agent, structure-activity relationships,
metabolic and pharmacokinetic properties, toxicologic effects and short-term tests of
genotoxic activity and promotion potential should be included in a hazard identification
document. For electromagnetic (EM) fields, this chapter will discuss biological phenomena
that have some relationship to the mechanisms of carcinogenesis. These are genetic effects,
including DMA interactions, chromosome effects and mitosis and meiosis, tissue growth and
differentiation, chemical signaling of growth induction stimuli across the cell membrane, and
the influence of EM fields on hormones, growth factors, cell transformation, the immune
system, and the central nervous system (CNS).
   The effects of EM fields on the various systems and cellular functions, such as calcium
transport,  hormone receptor activation, protein kinase and ornithine decarboxylase induction,
melatonin synthesis, and certain immunologic and CNS parameters, are indicative of effects
on cellular signal transduction processes. Also, it is likely that many cell growth and
differentiation processes in vivo represent an interaction between the genetic complement of
the cells and the effects of hormones, cytokines, and/or growth regulatory factors on signal
transduction processes (Luben, 1990). Although many of these findings do not deal directly
with cancer, they do indicate that  regulation of cellular growth and differentiation may be
directly modulated by EM fields (Luben, 1990). In addition, the processes of signal
transduction and oncogenesis are apparently closely linked, as discussed by Druker et al.
(1989).

5.1.  GENOTOXICITY OR NONGENOTOXICITY
5.1.1.  Introduction
   Genetic alterations are critical events in the carcinogenesis process. Thus, evidence on
the ability to produce a heritable genetic lesion (e.g., gene mutation, chromosomal aberration,
aneuploidy) can potentially provide useful mechanistic information for induced carcinogenesis
                                         5-1
10/22/90

-------
                           DRAFT-DO NOT QUOTE OR CITE

and is regarded as important qualitative information that reinforces the cancer concern for a
particular agent.  It should be emphasized, however, that genetic alterations inducing gene
mutations are only one component of carcinogenesis. Thus, the observations of the National
Toxicology Program that there are mutagens that are not carcinogens and carcinogens that
are not mutagens are to be expected. The use of results from short-term genotoxicity tests as
supporting evidence for or against carcinogenicity of an agent must be used with caution.
   This review addresses the biological effects of EM fields caused by direct action of the field
as opposed to secondary thermal effects. Genotoxicity studies in which thermal effects were
clearly involved are not included; however, it is virtually impossible to state with certainty that
the results observed following exposure to EM fields in the radiofrequency (RF) range were not
due to thermal effects. The following sections consider separately extremely low frequency
(ELF) electromagnetic fields (either magnetic fields alone, electric fields alone or both together)
and RF electromagnetic fields (predominantly in the microwave region). The genotoxicity
sections are organized by end point; therefore, a  given paper may be discussed in more than
one section if more than  one end point was examined and the demonstration of no effect on
one end point does not imply no effect on all end points.
   Because this review is intended to provide support for or against the carcinogenicity of EM
fields, studies on induction of dominant lethal mutations were not included as this  assay
system is not a useful  prescreen for carcinogenicity (Green et al., 1985).

5.1.2. Effects on Nucleic Acids
   Effects discussed in this section include DNA  damage and repair,  DNA synthesis and
related topics.

5.1.2.1. Extremely Low Frequency Electromagnetic Fields
   Strekova and Spitkovskii (1971), in an attempt to study a possible mechanism  of action of
magnetic fields on mitosis, examined the effect of a constant heterogeneous 12,000-oersted
(Oe) [1,200 millitesla (mT)] magnetic field on the Theological properties of
deoxyribonucleoprotein (DNP) strands. A graduated cylinder containing physiological saline
solution was placed between the poles of a permanent magnet and a drop of DNP solution
from calf thymus was placed on top of the saline solution.  Strands of DNP formed as the drop
fell. After a period of time the strands started to contract and the contraction of the strand and
       10/22/90
5-2

-------
                            DRAFT-DO NOT QUOTE OR CITE

the strand diameter were measured. The diameter of the strand was significantly greater than
control values after 15- or 60-minute exposures to the magnetic field (p<0.05). There were
also significant differences in kinetics of relative changes in length of DNP strands. Reaction
to the field was apparent during the first 15 seconds of exposure and was dependent on
protein content of the DNP. At high protein content (nitrogen/phosphorous ratio=4.6-4.9) the
relative contraction of the DNP in the magnetic field was lower than the control, while at a more
standard protein content (nitrogen/phosphorous ratio=3.7-4.2) it was higher (0.01 
-------
                           DRAFT-DO NOT QUOTE OR CITE

parallel metal plates for various times.  Delayed entry into the S phase was apparent by 3
hours after starting exposure and became significant by 24 hours (p<0.001).  The delay
remained relatively constant during 7 days of exposure. The authors also reported a
shortening of S phase and a stimulated incorporation of labeled thymidine during DMA
synthesis as compared to controls.  Liboff et al. (1984) exposed human foreskin fibroblasts in
culture to sinusoidally varying magnetic fields with frequencies of 15 Hz to 4 kilohertz (kHz)
and amplitudes of 2.3 x 10"3 to 5.6 x 10"1 mT generated by modified Helmholtz coils. They
found an enhancement of DMA synthesis with a threshold between 5 x 10"3 and 25 x 10"3
mT/second.  The enhancement was maximum during the middle of S phase and  was
independent of the time derivative of the magnetic field. Takahashi et al. (1986) looked at DMA
synthesis in Chinese hamster V79 cells in culture after exposure between Helmholtz coils to
pulsed magnetic fields. The effect of pulse width (pulse shape not given) was examined with a
100 Hz, 2 x 10"2 mT field (pulse width 6,10, 25, 50, 75, or 125 microsecond).  DMA synthesis
was significantly enhanced (about 30%, p<0.001) with a 25 microsecond pulse width but no
significant effects on DMA synthesis were found with the other pulse widths. The second
experiment held the pulse width constant at 25 microseconds and the magnitude constant at
2 x 10"2 mT and varied the frequency (5,10, 30,100, or 300 Hz). DMA synthesis was
enhanced about 13% at 10 Hz (p<0.01) and about 30% at 100 Hz (p< 0.001)  with no
significant effect at other frequencies.  The final experiment held the frequency and pulse width
constant at 100 Hz and 25 microseconds and varied the intensity. DMA synthesis was
significantly enhanced in the range of 2 to 8 x 10"2 mT (p<0.01 - p<0.001) but was significantly
inhibited at intensities greater than 2 x 10~1 mT (80% of control level at 4 x 10~1 mT, p<0.01).
The presence of windows of activity as demonstrated in this study obviously complicates
evaluation of biological effects of pulsed magnetic fields, particularly if the active  windows vary
with cell type.  The previous study (Liboff et al., 1984) using human fibroblasts found no
evidence for a window of activity for either frequency or amplitude.

5.1.2.2.  Radiofrequency Electromagnetic Fields
   DNA in aqueous solutions has been shown to absorb microwaves (Swicord et al., 1983,
and other  reports cited in this reference). Swicord et al. (1983) measured the absorption of 8
to 12 gigahertz (GHz) microwaves by DNA in saline solution following treatment of the DNA for
increasing times with  DNase Escherichia coli.  Temperature during measurements was
       10/22/90
5-4

-------
                            DRAFT-DO NOT QUOTE OR CITE


 maintained at 25 ą 0.2° C. They found that absorption increased with increasing DNase
 treatment time, implying that absorption increased with decreasing DMA chain-length.
 Absorption also increased with frequency. It remains to be determined whether this resonant
 absorption is relevant to any biological effects in vivo.  Additional data on absorption of
 microwaves in this frequency range by Escherichia coli DMA are given in Swicord and Davis
 (1983).
    An inhibitory effect of 1 GHz continuous wave (CW) and 1 GHz pulsed wave (PW)
 microwaves on DMA synthesis in cultured mouse L1210 leukemia cells was reported by Chang
 et al. (1980). This study was done to determine if low-power-density microwave radiation
 could enhance the antitumor activity of methotrexate.  The authors used incorporation of
 tritiated deoxyuridine into  DNA as a measure of DNA synthesis because methotrexate
 interferes with the conversion of deoxyuridine to thymidine which must occur before
 incorporation in DNA.  The power density of CW  microwaves ranged from 5 to 50 milliwatts per
 square centimeter (mW/cm2) with  an exposure time of 20 minutes.  A peak inhibition of
 incorporation of tritiated deoxyuridine of about 25% occurred between 15 and  25 mW/cm2 of
 CW but no inhibition was seen at 50 mW/cm2  A similar inhibition of incorporation resulted
 from a PW exposure at an average power density of 10 mW/cm2 The temperature of the
 culture increased from 23.5 ą 1° C before irradiation to 30.7 ą3.5° C at 25 mW/cm2 CW and
 37.4 ą 0.9° C at 50 mW/cm2 CW, the highest temperature reached being the normal growth
 temperature for these cells. CW exposure up to 25 mW/cm2 increased the cellular uptake of
 methotrexate about 80% compared to unirradiated controls while 50 mW/cm2 CW and 10
 mW/cm2 PW actually reduced uptake to below control level. Combined microwave exposure
 (15 mW/cm2) and methotrexate [0.2 molar (M)] reduced DNA synthesis slightly more than
 either alone. Thus microwave exposures as described above can, by themselves, inhibit DNA
 synthesis and can enhance the inhibition of DNA synthesis due to methotrexate exposure.
    Differential killing assays were employed in several investigations to determine if
 microwaves caused DNA damage. This assay compares the survival of wild type DNA
 repair-proficient cells to the survival of DNA repair-deficient cells following exposure to the
agent of interest. If an agent causes DNA damage, survival of the repair-deficient cells should
 be lower than that of wild type cells. Averbec et al. (1976) reported in abstract  form that no
differential killing occurred between wild type and excision and/or recombinational repair
defective strains of the bacteria or the yeast Saccharomyces cerevisiae. Exposures were for
                                         5-5
10/22/90

-------
                           DRAFT-DO NOT QUOTE OR CITE

30 minutes to 70- to 75-GHz microwaves with power densities up to 60 mW/cm2.  Cells were
irradiated on Millipore filters on agar plates to try to avoid thermal effects. The temperatures
are not reported but the results were negative in any case.
   Dardalhon et al. (1981) used a number of repair competent and deficient strains of E. coli,
including uvrA-, recA-, and multiple deficiency strains (uvrA- recA-) and (recA- recB- recC-),
and S. cerevisiae wild type and (rad 2-20 rad 9-4) strains. No significant effects were seen on
survival after exposures to 9.4,17, and 70-75 GHz microwaves. Power densities were varied
from 1 to 50 mW/cm2 at the two lower frequencies and from 1 to 60 mW/cm2 at the 70-75 GHz
range. Cells were irradiated on Millipore filter discs on solid agar plates for 30 minutes.
Heating the bacterial cells above 50° C conventionally for 30 or 60 minutes decreased survival,
particularly of the (uvrA- recA-) cells, indicating that these repair pathways are involved in
repairing heat-induced DNA damage. The conclusion reached is that microwave exposure as
reported here does not cause lethal DNA damage or DNA lesions repaired by the uvr or rec
pathways in bacteria or the rad 2 or rad 9 pathways in yeast.

5.1.3. Gene Mutations
   With few exceptions, attempts to induce gene mutations by exposure to EM fields have
been unsuccessful. The experiments have covered a range of frequencies and intensities and
have used as test objects bacteria, yeast, plants, insects, and mammalian cells in culture.
Because most studies were negative, this section will be presented in tabular form (Table 5-1)
with  some comments on the positive reports.

S.I.3.1. Extremely Low Frequency Electromagnetic Fields
   One publication available for this review reported a  positive response in the Drosophila
melanogaster sex-linked recessive lethal (SLRL) test after exposure to static electric fields
(Portnov et al., 1975). SLRL can result from gene mutations, small deletions, or chromosome
aberrations (Lee et al., 1983), depending on the  nature  of the chemical or physical insult. The
mechanism of action in this case is unknown. Female Canton-S flies were exposed for 24
hours to static electric fields of either 1500 or 3300 volts per centimeter (V/cm) (150 or 330
kV/m) and mated to MulIer-5 males immediately  after exposure.  SLRLs, determined in the F2
generation, were significantly more frequent in the exposed populations than in controls at
       10/22/90
5-6

-------
DRAFT-DO NOT QUOTE OR CITE
















CO
2
UTATION Rl
S
UJ
ill
CD
>
i
UJ
z
5
1-

















S
c
CD
•8
S






F

1
IS
8 Ť
all
_cz





^
JJ
CD
1




Genetic End Point




ts
CD
5"
O
w

c"
0!
1
5
CO
CD
c
'S
•ť=
2
3




2
3
O
f
§

^
X



o
c
Q
E
T3
1
•i
Ł
8



Reverse mutation to
E
1
i
a
TO
.21 1
1 1
<0 TO
ml co




en
en
0
ro 8
1 i
CO
3
O
.c
CO CO CO ^3-
§""33 w
o o o
.C C. ŁL ť•
sss 5-

h- 1- h-
JE E E
"o^L
xxx 1— 1—





1
TO
I"2
•I-I
TJ ^ O)
^ CO TO
^= d E
a>
|5
2
^ 0^-0 0
CX c- Q. r- c
2 p g o 5
§ || | |
S 13 |j? I
i s i u §• i
s §1 §s 1
CO (D CO (I) ^ o
Ic DT S DC TO U_

T-
.J .. _ w
E E c^ ^
§~ o ~ S _-
So 35 °2
< .i < E p >
^- 1 <°" I § ro"
i— coll— coin- _i


S
*-
•s
1
2
o>
^
2 2
CO 3 3

3 O
Q JI IO 1O
Nc3| 0
CO CM C\i cj

0)
S P o>
?|I|
m 10 o -*
CO CO OJ CO





ll
a> ^) o
> — c 'S
> 3 C3) o
N N E *CD
X X N N
C\j CO CM C\!


^
r>
arabinose resistance
Reverse mutation to
tryptophane auxotroi
g
3.
CM
Q.
O §
0 -
i I
w — ^
Q> ^i- j,™ _.
uj| § i3
                                     CD
                                     en
                                     jsŁ
to
en
                                          •&

                                          1
                                                TO

                                                %
                                      -2.  5
                                     i
                                     CO
                                                \
                                     s
                                     V
                                     in
                                     (v
                                     6
                                                  Q.
                                                o 2
                                                Lli
             5-7
     10/22/90

-------
                 DRAFT-DO NOT QUOTE OR CITE











1
ION REPOftTS <
ft
1
UJ
C3

> —
1
5
UJ
h-























Exoosure
Intensity





a- g
II
or o
CD "S



73
0.
Genetic End
Time



•5


* s i s

"3 -g Ť m fe
1 I 1 I rf
•5 c c w o
§ 1 X1 S 5

CO
0
(C
1 If § Is. IS
•2 S-S •= o'o^v
co r- -3, to ;= N. CM CM
•c
CM Ne .2
1 I &
1 * I Ť Jl
^ 0 > Q Q ^ ^
52 *~ o o o to
g v ° Sco 88!



o .0

to 'a to
ąJ CD ^ 4? CD
X N N C ^ g C
S 55 ^^ Ž !>
S 9 o 1 §> -S 1
JC "" "^ § CB CB §
oj S S x E 55 x
lo"
Trt Ž
1 f
O - •** ^-fc
°Z> 8 § P o
O O O O O CD O
TSofTS os"c5 nJ ^co*"
e|i 1| 1 l|" 1
"UsTJ ^o Q) coed ,co_
Irai*- S^- °" -"^ -1
Ł S Ł 8 ŁE iS". CTJM'W
11 ^
i 3 § 1 1
15 co c •Ł. §
to 3 TO CD 0)
E 2 ?! to to
ail cl El co to
Jj J | §.
_C |c S 3 3 _C O .E 0 ^~ S
SO^cjtMin Sin •D*.^.
*- \O v- v- *- t-T CD O OJ

CM CM CM ^ f- "D
EEE Ec t c lls ,_._ cE
-5i -ii -Si 5 -fe ^ -fe § § * PE •Ł >
i— CNCJ COCO ^ CD COIOCD Pť- CO^"



Q
'*' CD
"& 0) 1
•!ť Ť° * Ť *"
§ 3 g E ~
i 1| | i | || |
co *r c> Ł co *r .SJcDo
OJ T-CNlX CD CNJ LLIQ.CO
CD
f
'E
CO
3
SCO CO fl} O
o o o •=
N N N N Ł
CO CO CO CO 5* (/)
i 1^1 i 'cci o
CL CLt? Pi Q, "^ ~* ^
5 5|.5 55 s o'|
C/3 C/3 co CO CO CO O3 DC Ł
1 - Ł•
"SI "3 "7S C TS ^
CO CO CO C *u >
^ Ł .E .Ł E^ o|
C> 0)1 Q) & ^1 Ł **
CO 03 CO CO CO Cd O Ł ^~
D rr O rr C7rr O Or*" "-^ CO CO CD
o °r o H- o4- o oH- c^ercN ^'
cc cc cc c cc SJTO ..C
^5o J5o ^5o J5 ^So co o> o W,L
 (DO) tl> CD O) fl>CC 
-------
                            DRAFT-DO NOT QUOTE OR CITE

both exposure levels (p<0.05) and more frequent at the lower field level than at the higher
level but not significantly so (p>0.05).

5.1.3.2. Radiofrequency Electromagnetic Fields
   A number of positive results were reported after exposures to radiofrequencies. Danilenko
et al. (1974) reported an increase in the number of histidine revertants and morphological
variants following irradiation of an histidine-dependent strain of Candida tropicalis D-2 with a
37-GHz electromagnetic field.  This study also reported a synergistic interaction of the ultra-
high frequency (UHF) field with N-nitroso-N-methylurea and 1-methyl-3-nitro-1-
nitrosoguanidine.  The data in this paper are incomplete or unclear. The exposure was given
as 20 minutes at 1 milliwatt (mW) for the combined chemical and irradiation studies but no
data were given for irradiation alone.  The same exposure must be assumed.  Chemical
treatment alone was for 3 hours; however, in combined treatments the cells were treated for
1.5 hours, then irradiated, then treated again with chemical for "up to 3 hours." It may be
incorrect to conclude synergism when the exposure times are different.  The authors say that
no significant temperature change occurred during the course of the experiments, but they
provide no data.
   A number of studies by Harte (1972,1973, and others cited therein) reported the induction
of embryo lethal and morphological mutants in Oenothera hookeri (evening primrose) by 1.5-
or 3-meter (m) radio waves [200 or 100 megahertz (MHz)].  In Harte (1973) the plants were
grown for an entire growth cycle in the vicinity of an antenna radiating at 3 m (100 MHz) with
electric field intensities of 250, 235, and 145 millivolts per meter (mV/m) at 1, 0.5,  and 0.1 m
above the ground, respectively. M1 and M2 generations were analyzed. In Harte (1972)
pollen was irradiated with 1.5 m (200 MHz) radio waves at 1.4 or 1.8 mV/m for 4 or 12 hours
and the M1 and M2 generations were analyzed. Harte (1975) performed the same type of
experiments with Antirrhinum majus (snapdragon).  Pollen was irradiated with 1.5 m (200 MHz)
radio waves (field strength 1.5 V/m) for 4,12 or 43.75 hours. Again, both embryo lethal and
morphological variants were reported; however, serious infections occurred in all groups and
control data from the literature had to be used, making the results unreliable.  Both embryo
lethal and morphological variants can be caused by gene mutations or by various
chromosomal effects. Oenothera has quite atypical cytogenetic behavior (Steiner, 1975) so
the effects observed in these studies are most likely not due to gene mutations.
                                         5-9
10/22/90

-------
                           DRAFT-DO NOT QUOTE OR CITE

   A study by Blevins et al. (1980) reported induction of gene mutations in Salmonella
typhimurium strains TA98, TA100, TA1535, TA1537, and TA1538 by 2450-MHz microwaves.
This study is difficult to evaluate because no attempt was made to keep temperature constant
and at 23 seconds of exposure the temperature reached 100° C. The control was heating by
conventional oven and any difference between mutation induction at a given temperature by
the two heating methods (much greater in all strains with microwave heating) was attributed to
non-thermal effects of the microwaves.  No mutation induction occurred below about 70° C by
either heating method. This conclusion assumes equivalency of the two heating methods,
which because of possible effects such as differential absorption and thermal gradients, may
not be a valid assumption.
   A discussion of the Frohiich (1968) model in Biological Effects of Radiofrequency Radiation
(U.S. EPA, 1984) may be pertinent (see Section 2.4.3. for a discussion of the Frohiich model).
Frohlich's proposed mechanism "... is a possible mechanism for grouping individual photons
or phonons with energiesŤ kT (the average thermal energy per molecule at body
temperatures). This process results in the application of energy in a significant amount (>kT)
at a single locus.  Effects resulting from this process could not be duplicated by addition of the
same amount of energy to the system by a different process." In any event, the relevance of
the results with respect to carcinogenicity in humans is questionable since mutation induction
occurred only at 70° C and above. In conclusion, there is no unequivocal evidence for
induction of gene  mutations by EM fields.

5.1.4. Chromosome Effects
5.1.4.1. Extremely Low Frequency Electromagnetic Fields
   Two studies reported on induction of chromosomal aberrations in plant roots following
exposure to electric fields.  Dubrov et al. (1968) grew onion seedlings on filter paper for 43
hours, then placed the filter paper on a polymer film electrode in  a Petri dish, placed the dish
on top of another  electrode charged oppositely and allowed the seedlings to grow an
additional 24 hours. The electrodes "were connected to an alternating static potential of 4.5
kW [kilowatts] of industrial frequency" [sic], probably 50 Hz. The authors reported a
chromosomal aberration frequency in root tip cells that was more than twice the control
frequency when the seedlings were grown on a positively charged electrode and about 60%
       10/22/90
5-10

-------
                           DRAFT-DO NOT QUOTE OR CITE

 above control levels when grown on a negatively charged electrode.  No data were provided
 on types of chromosomal aberrations. Miller et al. (1976) exposed broad bean roots to 75-Hz
 electric fields at 10 V/m for 1 to 6 days and found no significant effects on chromosome
 aberration induction in root tip cells. The raw data are not given.
   The ability of magnetic fields to induce chromosomal aberrations in plants was also
 examined.  Miller et al. (1976) exposed broad bean root to 75-Hz magnetic fields of 0.5, 5, or
 17 gauss (G) (0.05, 0.5, or 1.7 mT) for up to 6 days and found no induction of chromosomal
 aberrations in root tip cells, but, as in their electric field exposure studies above, no raw data
 are presented. A higher frequency of chromosome breaks occurred in pea root tip cells
 exposed between the poles of an electromagnet to 8000 Oe (800 mT) than in control cells
 (Goswami and Dave, 1975). Other chromosomal anomalies were also observed. The time of
 exposure was not given. Shevchenko et al. (1978) found no induction of chromosomal
 aberrations in root tip cells of Crepis capillaris following exposure to constant magnetic fields.
 In one experiment, germinating seeds were exposed for 1 or 2 days to a constant magnetic
 field of average field strength  9000 Oe (900 mT) [gradient about 200 oersted per centimeter
 (Oe/cm), 20 millitesla per centimeter (mT/cm)] or average field strength of 12,000 Oe (1,200
 mT) (gradient of 300-400 Oe/cm, 30-40 mT).  In the second experiment, dry seeds were
 exposed to 9000 Oe (900 mT) as above for 2 months and then germinated in the same field for
 2 days. No increase in chromosomal aberrations frequency was induced by any exposure
 condition.
   Magnetic fields were shown to cause premature decondensation (despiralization) of
 chromosomes in pea and garlic root tips (Goswami, 1977).  It is not clear if the exposures were
for 10 or 15 minutes or were two 15-minute exposures separated  by 30 minutes in distilled
water.  Field strength was 7400 Oe  (740 mT).  Decondensation of chromosomes was also
seen in onion root tip cells after exposure to 6000 G (600 mT) for  140 hours starting from the
time dry seeds were placed on moist filter  paper (Herich, 1976). Goswami (1977) speculated
that removal of proteins from the chromosome and/or chelation of divalent cations was
responsible for the decondensation.
   The final plant paper considered in this section reported no induction of micronuclei in
Tradescantiajpo\\en mother cells after exposure of inflorescences  to magnetic field intensities
of 160, 760, or 780 mT for 6 or 7 days (Baum and Nauman,  1984). Both Tradescantia clone 02
and 4430 were used with similar results.
                                        5-11
10/22/90

-------
                           DRAFT-DO NOT QUOTE OR CITE

   The remaining papers discussed in this section used rodent or human cells (in vivo and
cell culture) as test objects.
   Two papers examined the effects of magnetic fields on chromosomal aberration induction
in mice. Strzhizhovskii et al. (1979) found no increase in frequency of chromosome
aberrations in corneal epithelial cells in mice exposed to 3 to 127 kOe (300 to 12,700 mT) for
one hour.  The magnetic field, produced by a water-cooled electromagnet with a cylindrical
active zone, had one constant component and one saw-tooth component with a period of 30
seconds. The authors state that "the intensity of the constant component coincided with the
amplitude of the saw-toothed component."  Additional data were not given.  No numerical data
were provided on aberrations.
   Mastryukova and Rudneva (1978)  exposed C57BL and CBA mice to a static magnetic field
of 1000 Oe (100 mT) for one day (presumably 24 hours) and examined  the duodenal
epithelium at 0,1, 3, and 24 hours and 6,10,12, 20, 24, and 30 days postexposure. The
authors report an increase in percent of pathological mitoses immediately after exposure [13.0
(0.78%) compared to 6.3 (0.86%) in controls] due mostly to chromosome damage, but they do
not give specifics of the damage. The peak time of chromosome damage was 24-hour
postexposure with the percent decreasing to control levels between 10 and 30 days. Eberle
and May (1982) found no effect on the rate of structural chromosome or chromatid aberrations
or micronuclei formation in bone marrow cells of Chinese hamsters exposed to 1000 mT for 3
or 24 hours, 7000 mT for 3 hours or 14,000 mT for 15 minutes and killed 26 hours later. There
were increases in hypo- and hyperploid cells, indicating a possible effect on chromosome
segregation.
   El Nahas and Oraby (1989) reported the induction of micronuclei in bone marrow
polychromatic erythrocytes (PCE) of male Swiss mice exposed to uniform 50-Hz electric fields
of 170,220, and 290 kV/m for 24 hours.  No statistically significant increase in micronuclei
occurred at a field intensity of 100 kV/m.  The number of micronuclei found increased with field
intensity and was dependent on time of the postexposure that elapsed  before the animal was
killed, reaching a maximum at 72 hours after exposure began. The mean number of
micronuclei per 500 PCE at 72 hours was 12.4 ą 4.67,19.6 ą 4.16, and 23.2 ą 12.27 at 170,
220, and 290  kV/m, respectively.  The micronuclei frequency decreased at 96 hours after
beginning exposure but was still significantly above control levels. Control levels varied in
different experiments from a mean of 3.3 ą 1.53 to 4.4 ą 1.52 per 500 cells.
       10/22/90
5-12

-------
                           DRAFT-DO NOT QUOTE OR CITE

   A number of studies in which human lymphocytes were examined following exposure to
ELF fields are available in the literature. Mileva et al. (1985) exposed human lymphocytes to a
magnetic field (permanent magnetic field of 3 kilooersted (kOe), 300 mT, for 30 minutes to 6
hours) and found no significant difference in chromosomal aberrations between exposed and
control cells (p>0.05).
   Cohen et al. (1986a, b) exposed human lymphocytes to 60-Hz electromagnetic fields (0.25
V/m and 1 or 2 G, 0.1 or 0.2 mT) or to just the electric or magnetic field alone for 69 hours.
Cohen et al. (1986b) also treated human lymphoid cell lines from patients with various
chromosomal instability syndromes.  No significant increase in chromosome breaks in any of
the cell lines or in normal  lymphocytes resulted from these exposures (p>0.05).
   Rosenthal and Obe (1989) exposed human peripheral lymphocytes to 50 Hz, 5 mT
magnetic fields for 48 hours and also found no effect on the number of chromosome breaks
compared to unexposed controls.  Tsoneva et al. (1975) had reported induction of
chromosome and chromatid breaks in cultured human lymphocytes after exposure to 1750,
3910, or 6000 G (175, 391, or 600 mT) for 30 seconds or to 1750 G (175 mT) for 30 or 60
minutes.  The cells were exposed before cultivation.
   Nordenson et al. (1984) found no increase in chromosomal aberrations in human
lymphocytes following exposure of whole blood to a 50-Hz electric field, current density of 1
milliampere per square centimeter (mA/cm2), for 3 hours; however, an exposure to ten
3-microsecond-long spark discharge pulses (peak field strength 350 kV/m,  about 5 seconds
between pulses) did cause a significant increase in chromosome breaks to a level comparable
to 0.75 Gray (Gy) of ionizing radiation (p<0.001).  Other abnormalities, including polyploidy,
endoploidy, and premature chromosome condensation, were also seen in exposed cells.  In
the same publication, these authors examined 72-hour cultures of peripheral lymphocytes from
men  who had worked at a 400-kilovoIt (kV) switchyard for 1 to 8 weeks immediately preceding
blood sampling. Significant increases in chromosomal aberrations were found compared to
unexposed controls (p<0.0005).
   In contrast to the in vivo results of Nordenson et al. (1984), Bauchinger  et al. (1981)
reported no increase in chromosomal aberrations in 48-hours cultures of lymphocytes from
workers exposed for more than 20 years to 50-Hz electric and magnetic fields in 380 kV
switchyards (p>0.05). This difference may be explained by the comment in the discussion of
Nordenson et al. (1984) that all except one worker in their study had acute exposure to high
                                        5-13
10/22/90

-------
                           DRAFT-DO NOT QUOTE OR CITE

strength electric fields and to spark discharges immediately before donating blood.
Lymphocytes from the one exception showed no chromosome or chromatid breaks.
   The sister chromatid exchange (SCE) assay is a method considered to be more sensitive
than analysis of chromosome breaks for detecting damage to DNA in chromosomes. The
actual mechanism of SCE formation is not fully understood but apparently involves DNA
breakage and reunion (Latt et al., 1981).  Five of the six papers to be considered in this section
reported no effect of ELF fields on SCE induction. Four of the negative papers, using human
lymphocytes and lymphoblastoid cells, were discussed above in the paragraph on
chromosomal aberrations and details of exposure are given there (Cohen et al., 1986a, b;
Bauchinger,  1981; Rosenthal and Obe, 1989). Rosenthal and Obe (1989) used a 72-hour
culture for SCE analysis in addition to the 48-hour culture mentioned in the previous section
on chromosome aberrations.  The fifth negative paper (Benz and Carsten, 1986) reported in
abstract form that exposure of male and female mice to 60-Hz, 15-kV/m - 3 G (0.3 mT) or 50
kV/m -10 G  (1 mT) fields for 1 to 28 weeks did not produce significant effects on SCE levels in
bone marrow cells.
   The only positive report, also discussed above in the section on chromosomal aberrations
(Eberle and May, 1982), found an increase in SCE induction in Chinese hamster bone marrow
cells after in  vivo exposure to  homogeneous static magnetic fields (1000 mT for 3 or 24 hours,
7000 mT for  3 hours or 14,000 mT for 15 minutes and killed 26 hours later). Increased SCE
levels were seen at all exposures with a maximum increase about twice the control level after 3
hours at 7000 mT.

5.1.4.2. Radiofrequency Electromagnetic Fields
    Radiofrequency EM fields have been shown to induce chromosomal aberrations  in a
number of test systems; however, it is very difficult to state unequivocally that the effects are
nonthermal.  A typical control for possible thermal effects is either to maintain the culture at the
starting temperature (a few tenths of a degree) or to compare the results with microwave
heating to those with conventional heating.  Neither approach may be acceptable because all
that can be measured is average culture temperature. Resonant absorption inside the cell
may cause local hot spots not representative of the average culture temperature. This caveat
will not be repeated with each experiment but should be kept in mind.
       10/22/90
5-14

-------
                           DRAFT-DO NOT QUOTE OR CITE

    Chen et al. (1974) exposed Chinese hamster cells in culture (unspecified tissue or cell line)
to 2450-MHz microwaves of various intensities for various times. From a starting temperature
of 22° C, the cells were exposed for 10 minutes to 50 mW/cm2 or for 4, 8, or 10 minutes to 85
mW/cm2. The culture temperature after irradiation was 37 to 41 ° C. A similar experiment was
done with a starting temperature of 37° C and exposures for 8 or 10 minutes to 20 mW/cm2, 4
minutes to 50 mW/cm2, or 2 or 3 minutes to 85 mW/cm2.  With a 37° C starting temperature,
culture temperatures following irradiation were 40° to 43° C. Conventional heating to 45° C
did not induce chromosomal aberrations. Data were presented separately for 10 types of
chromosomal damage plus polyploidy.  The authors state that "the difference in aberrations
observed between the control and the irradiated samples were not significant at the 5% level"
and then try to make a point that in some cases they are. The data are so varied with no
consistent pattern that conclusions are not really justified. Similar data are presented for
human amnion cells and the same comments apply.
    Alam et al. (1978) used Chinese hamster CHO-K1 cells in a study of chromosome
aberration induction by 2450 ą 25 MHz-microwaves under temperature controlled and
uncontrolled conditions. Exposures were given as 30 minutes at 25-watt (W) incident power
without temperature control (temperature of the culture reached 49° C) and 30 minutes at 75-,
125-, and 200-W incident power with the temperature maintained at 29° C. At the three higher
powers the  power density  at the surface of the culture medium exceeded 200 mW/cm2. The
culture without temperature control had significant increases in chromosome breaks
compared to control cells, and nuclear vacuoles and pycnotic and decondensed
chromosomes were seen.  Under temperature controlled conditions, none of these effects
were seen.
   A long-term exposure study was done by Yao (1982).  He cultured rat kangaroo RH5 and
RH16 cells for 50 passages (320 days) in incubators with the cell culture temperature
maintained at 37° C by either 2450-MHz microwave  heating or by conventional heating. The
power density necessary to maintain this temperature by microwave heating is not given;
however, the author states that the magnetron was operated at "about 10 volts" and that "the
energy absorption dose rate of the medium and cell culture was tested and  estimated to be
15.2 ą 1.82 mW per gram." A synthetic rubber cylinder at the center of the incubator
absorbed "most of the microwave radiation and was heated by it."  Heat released by the
cylinder was exhausted by a fan through an opening in the top of the incubator.  An
                                        5-15
10/22/90

-------
                          DRAFT-DO NOT QUOTE OR CITE

equilibrium was reached in 2 hours between the heat from the rubber cylinder and the culture
medium and the heat exhausted through the top air opening. The RH16 cell line has a normal
2n = 12 chromosome complement, but the RH5 line lost the number 5 chromosome and has a
2n = 11 chromosome complement. Cells were examined for chromosomal aberrations at 10,
20, 30, 40, and 50 passages. The microwave-exposed cells showed retarded growth, and
after 20 passages chromosome aberrations and polyploid cells appeared. By passage 50,
43% of the irradiated RH5 cells had at least one aberration and 35% were polyploid. Also,
there was a significant increase in number of chromosome breaks per cell in this cell line.
Many of the irradiated RH5 cells lost a chromosome (usually chromosome 4) during
irradiation, and after 30 passages, 41 % of the cells had 10 chromosomes. Only 6% of the
irradiated RH16 cells were aberrant after 40 passages, and no aberrant cells were found after
50 passages.  About 31 % of the RH16 cells were polyploid after 30 passages. Following
return to a conventional incubator, the number of chromosome breaks and polyploid cells
decreased. Because the RH5 line was hypoploid before irradiation  it might have been more
sensitive to microwave irradiation than the RH16 line.
   Yao had previously reported induction of chromosomal aberrations in corneal epithelium of
female Chinese hamsters after in vivo exposure to 2450-MHz microwaves (Yao, 1978). The
right eye of each animal was irradiated at 100 mW/cm2 for 5,10, 20, or 30 minutes or at 25
mW/cm2 for 10 or 20 minutes. The percent of abnormal cells and the number of chromosome
breaks per cell were significantly higher than control values (p=0.05) at 100 mW/cm2for 30
minutes. Dicentrics were the most common anomaly.  There was no mention of temperature
increases in this 1978 report.
    Chromosomal aberrations, polyploidy, and aneuploidy were found in rat bone marrow cells
after in vivo exposure  to 12-centimeter (cm) (about 2450 MHz) microwaves for 7 hours daily for
10 days at power densities of 50 or 500 W/cm2 (Kapustin et al., 1976). The types of
aberrations included chromatid deletions, acentric fragments, and chromatid breaks. Bone
marrow was examined after 18 hours, 2 weeks, or 2 weeks with hypoxia following the end of
exposure. At the lower exposure, the percent of abnormal cells increased from about 19% at
18 hours to about 34% at 2 weeks, while at the higher exposure the percent of abnormal cells
decreased from about 40% at 18 hours to 23% at 2 weeks.  This decrease probably reflects
elimination of cells severely damaged at the high exposure level. Control animals had about
4% abnormal cells in the bone marrow.  Postexposure hypoxia led to a lower percent of
       10/22/90
5-16

-------
                           DRAFT-DO NOT QUOTE OR CITE

abnormal cells at 2 weeks but still well above control levels. No increase in chromosomal
aberrations was seen in bone marrow cells of male (CBA x C57BL)F1 mice after exposure to
2400-MHz microwaves (Ramaiya et al., 1980). Mice were irradiated at 800 mW/cm2 for 21
seconds daily for 10 days. The percent of aberrant bone marrow metaphases was given as
1.37% in controls and 1.62% in irradiated mice. Beechey et al. (1986) irradiated male
(C3H/HeH x 101/H)F1 mice with 2450-MHz microwaves amplitude modulated at 100 Hz and
examined spermatocytes for chromosome aberrations. Exposures were to 0.1,1.0, or 40
mW/cm2 for 30 minutes/day, 6 days/week, for 2 weeks. Spermatocytes were isolated  2 to 3
days postexposure, when sampled germ cells were treated as spermatocytes and at 30 days
postexposure, when treated spermatogonia were sampled.  The rise in rectal temperature did
not exceed 1° C in sham-exposed and the two lower exposure mice and did not exceed  3° C
in the high dose mice.  There were no significant differences in chromosomal aberrations or
univalents at any exposure level compared to sham-exposed mice, although the highest
percent of cells with aberrations occurred at the highest dose level in cells treated as
spermatocytes.
    Human lymphocytes were irradiated with 2950-MHz [sic] pulsed microwaves (1200 Hz, 1
microsecond pulse width) following 66 hours of culture at 37° C (Stodolnik-Baranska,  1974).
Exposure was to 20 mW/cm2 for 5,10,15, or 20 minutes with 0.5° C rise in temperature at 15
minutes and 1 ° C rise after 20 minutes. Various abnormalities were seen, including
chromosome stickiness, aneuploidy, dicentrics, chromatid breaks, and possibly changes in
chromosome spiralization. The general trend was towards greater effects at longer exposure
times. The increases appear to be significant, but statistical data are not presented.
   Wolff et al. (1985) exposed human lymphocytes and Chinese hamster ovary (CHO) cells for
12.5 hours to the emissions of a magnetic resonance imaging instrument with a magnetic field
strength of 2350 mT. The 100-MHz signal was pulsed at 100 pulses of 330 microsecond
duration per minute. Both mitogen-stimulated (with phytohemagglutinin) and unstimulated
lymphocytes were used.  Unstimulated lymphocytes are in the G0 phase of the cell cycle and
do not leave this resting stage and progress through the cell cycle until stimulated. No
increase in chromosome aberrations occurred in any of the cultures.
   No significant increase in SCE levels was detected in any paper considered in this  section.
Wolff et al. (1985) used human lymphocytes and CHO cells under exposure  conditions just
described above.  Ciaravino et al. (1987b) exposed CHO cells for 2 hours to 2450-MHz
                                        5-17
10/22/90

-------
                           DRAFT-DO NOT QUOTE OR CITE
microwaves pulsed at 25,000 pulses per second, 10 microsecond pulse width.  The specific
absorption rate was given as 33.8 watts per kilogram (W/kg) resulting in a maximum culture
temperature of 39.2° C.  There were no interactive effects when cells were treated
simultaneously with microwaves and mitomycin C, a chemical agent that causes SCEs in CHO
cells.  Ciaravino et al. (1987a) reported in abstract form an identical study except adriamycin,
another SCE inducer, replaced mitomycin C. There were no interactive effects. Finally, McRee
et al. (1981) exposed female CD-1 mice to 2450-MHz CW microwaves at 20 mW/cm2 for 8
hours/day (two 4-hour periods separated by 1 hours) for 28 days.  Deep colonic temperature
was maintained at 0.7( to 1 ( C above the initial value during the 4-hour irradiation periods.
There was no difference in SCE frequency in bone marrow cells compared to sham-exposed
or normal controls.

5.1.4.3. Summary
    Chromosome aberrations have been reported in a number of test objects following
exposure to static and varying electric and magnetic fields as well as RF fields, but results are
not consistent.  Chromosome breaks are most commonly seen. Chromosome
decondensation occurred after exposure to static magnetic fields.  Other abnormalities such
as polyploidy, aneuploidy, chromosome stickiness, and dicentrics have been seen  in  some
experiments with no consistent pattern. One study reported the induction of SCEs  by static
magnetic fields; however, all other studies with either ELF or RF fields were negative.
    Table 5-2a summarizes the chromosome effects of static and ELF fields while Table 5-2b
summarizes the chromosome effects of RF fields.

5.1.5. Summary of Genetic Effects
    It is very difficult to draw general conclusions about the genetic activity (toxicity) of EM
fields. Because of the variables involved, there are virtually unlimited numbers of exposure
scenarios possible. Some  of the variables are:

    •  Frequency - 0 to many GHz
    •  Continuous wave or pulsed
    •  Pulse width
    •  Pulse shape
       10/22/90
5-18

-------
                                DRAFT-DO NOT QUOTE OR CITE
UJ

o
cc
O
tu
§
O
O
a






3
UJ
n:
1
UJ
O
UJ
SE
_i
>;
UJ
UJ

CO
.5
UJ
u_
u-
UJ
HI
S
o
en
a. CHROMO!
<*
in
UJ
m
<
i-














8
Ł
.CB
&

Ł
3
&




\
'§.
T3









 .? -0 Ť
"co Ł °
";_•_• co jc
•5 co co •- c
> o> Ť 5 5
X CB CD ^ >
a i i <§ Ł


+ 11 +




CO
CO
Ł

OOO O 5






co KT
** X
is "D 1O ^"
^ 9 ^ TJ "a g
"" *~ *~ !s i55 =6
%" -5 ° Ł E ° "ECCN
•* ^ in ^ o  ^* ť—


a!
c
O)
CO
E "g
^ ť
-S CO g^

10 in at *; ?
.9 .9- .9- § "0
1 Ť i i is
^ „ ť- ť_ w J5 O)
"1 S Ť Ť -9- Q-E
^3 Q O (D ,j (0 (0
^ *2l c *u *n O w "c
1. Q .2 g g Jj ^Ť
iJoJOmm a! o01



Ł
^
0


i







S








r static field
|
*—














                                                       N-
                                                       6
     I
     o
          -
         6
                                                           5
         •g
         I
          I
          CO
          m
0
O
g
3
0
o
g
3
0
o
g
3
0
                                                  2
                                                  1   IS
                                                  Ť S2  to 'c
                                                    Ł
                                                    c\l
                                                      O
         1!
                                                           oo
         1
         w
           „
         E&
         g?
                  o*2
                  CD CB
                                                  0 T3

                                                  2 i
                                                  co 0
                                                      a.
                                                      1
                                                      2
               TJ
               CB  ť-
               CO  CB
               O  '.g
               §•  |

               CO  c

               *  I
               s  §.
                                                           p
                                                           i
             11
                                        5-19
                                                                            10/22/90

-------
                DRAFT-DO NOT QUOTE OR CITE









j
CO
i
>




7^

iLl
pc
t

i
5
ffi
S
Ť
r ABLE 5-2












1
Ł
CO
1
s









ja

i

Exposure





Test System





S)
*—
"5
CD
S
I

1












g

SOOmT static field;
vv

3
S"" ~

Mammal, mammal cell
Mouse strain H, In vivo

























j=
CO
c
E
5
VJ




bone marrow


O)
O5
*n5
CO
1
O
1

1












g

!00-1 2700 mT static
ield; 1 hr





Mouse, in vivo
cornea! epithelium
co
O>
1
C
T3
3.
1
CO
1
1
to
I

+












g

1 00 mT static field;





Mouse, in vivo

























3-
T3





duodenal epithelium


CO
s ~
d- S
1 *
CO CO
& TJ
? CB
CO ^
N *^
CD .Q
m ui

I 1 + 1 + + 1 + +

CO CO CO
c c c
o o o
s s= •=
co CO co
hi h_ W
CD CD CD
.0 .a .a
rt cfl TO
(0 (tf (0
000
•c •= ~
CO 0) CD
111 g UJ Ł LLJ Ł UJ
CO O c CO O C CO O C CO
E*
*" IO
.0 g co 71
S 5 "° "° "5
CO ^Ł  i c3 t
•El ol
§s Ii
i* s*
f

i
T^_

ro
•o 1
§-:
CO
1 1
Qj S

+ 1









'53
o
c
i g

50 Hz; 100, 170, 220, 290
kV/m; 24 hour
SOOmT static field;
30 min- 6 hr


2
*>
c-
Mouse, Swiss, in vivo
Human lymphocytes, ii




t2
'-
"5
i
1

+ +










2. S.

Q) Q)
oo

175, 391, 600 mT static
fields; 30 sec 1 75 mT;
30 or 60 min

•o
CO
R
X
Human lymphocytes, e
before cultivation



JD
c§
OT
15
c
CD
g

1










J2
CO
JD
O CO

0.25 V/m; 0.1 or 0.2 mT;
60 Hz; 69 hr
69 hr


2
> 13
c S
Human lymphocytes, ii
or human lyhmphoid
lines in vitro










1 1 1










CO CO

Iglg
OcoO CO

Ł 8
1 Ł
.N 8
s I"
c W
•Ł o
K B
d d






10/22/90
5-20

-------
DRAFT-DO NOT QUOTE OR CITE
















Ł
0
DC
O
J5
0
Ł5
o
Q
1
t
DC
Q




















S"
a>
c
i
CO
3
UJ
E
UJ
Ul
DC
U.
|
UJ
Ul
oc
Ł

HROMOSOME EFF
o
S
Ul
CD













8
1
i



1

Q_



_Q
1
1




!§"








!
to
m
Ł
T-
5
8
"O
- c
a
75
1
1




1 1





12
(0
0)
*- UJ
Oco




50 Hz; 48 or 72 hr
iir
in


o
'>
c
1
o
.c
Q.
{
E
3
X

°>
75



S
1
1




1 + + +


i
1
Q>
<0
12 S
2 a>
S-° i O <
0 CQ.O


§8
x; u
? t|
N w
8 if s
• - f^ o CO
N_ r- ť J.
E .E ť c 7.
> ^ m" S ^
E ° c S



<5
>• w
Ť a Ł
.2 •? o
0 M ^
° 6" o.
11 I!
C Q) C Q}
go 5-5
c j: E o

X
oo
O)
75

V

0)
1
1




1 1






Ul
<0
O CO




1
A
Ł
8
^
1



(0

i}
1!
sŤ
C 0)
11
3 >
X S
         5-21
10/22/90

-------
                DRAFT-DO NOT QUOTE OR CITE







tn
3
Ł
>*
•J
111
8
TS", RADIOFRE
0
Bf
u.
UJ
r.2b. CHROMOSOME
in
g










Ł

"5
CO




End Point"



1
tn
g.
ff




Test System


co
o>

"3
In
E
CO






O)
V St

~-i
Ł ^
Ť j=
Chinese ham
corneal epit
1
2 2 5 * "ť
r ^ *" - a
~^ "<Ł *~- ~ci a
v . ~
1 f 8 | °
C CC Sr* ^*

TT"- - a T~i


CO OT
o o
1 !
JQ JQ
ca S Ť
^_ ť S ^ 5oT"
CSlN- OJ1'- CMCNJ OJ^OCO C\|T-IO"
T- Q
E > Ť c
--T -> OŤ
m 1 g x % S
!*• & •— n t
JQ co > I § 5*
O g c a H o
§^ CQo O^ CO0- E
> i °-'Q. ° 1 °-w- -
.eg 8 § 8" Ť S -i |
- g 3-> 3§ 3> E
g5 ^.c ^E ^.E j3
10/22/90
5-22

-------
                               DRAFT-DO NOT QUOTE OR CITE
UJ

O
cc
O
UJ

I

O
O
Q
DC
Q















^

|
1

CO
3
UJ
CE
>-
z
LLJ
a
rr
u_
O
Q
Ł
B-
co
O
Hi
ijj
UJ
*"^
s
BLE5-2b. CHROMO
p











8
Ł
o
1







^
CO
^






^
c
Ł
1








Exposure




Test System






*"^
13
Ł
|








= =










UJ




. - m
N W
li.
8 8
!-Ť-
MRI device; 2350 ml
1 .67 pulses/sec; 330
width


S
<
S
0)
Q>
to

1

§
.*:
CO
i.
§>
O)
o
'S
if any exposure scent
§
•_T3
the result is given as posil

"Summary only •
c
i
•o
•§
i
c
s
CO
CO
1
t
co
1
i
3
•z.
•a
S
1

-------
                           DRAFT-DO NOT QUOTE OR CITE


   •   Pulse amplitude
   •   Field intensity, flux density, etc.
   •   Field orientation
   •   Magnetic, electric, or combined fields
   •   Time of exposure
   •   Temperature
   •   Test object (whole animal, plant, cell culture, bacteria, etc.)
   The biological effects can depend on any or all of the above variables, and in some cases
very narrow "windows of activity" are seen, making extrapolation from one scenario to another
impossible. Although the sections on genetic activity do not constitute an exhaustive review,
the inclusion of all related papers would be unlikely to increase the reliability of any
conclusions drawn. Also, this review was to consider nonthermal effects only, but as
described previously, it is not always possible to eliminate thermal effects with exposures in
the microwave region.
   In spite of the problems mentioned, some conclusions (none of which are absolute) follow:

   •  In the absence of thermal effects, nonionizing EM fields do not cause DMA damage
       measurable by DMA breaks, DMA repair or differential killing of repair defective
       organisms.
   •  DMA can, under some exposure scenarios, absorb energy from EM fields.
   •  DNA synthesis can be affected by electric, magnetic, and EM fields and both
       enhancement and inhibition have been shown with no obvious pattern.
   •  Chromosomal aberrations have been reported in a number of test objects following
       exposure to static and varying electric and magnetic fields as well as RF fields, but
       results are not consistent. Chromosome breaks are most commonly seen.
       Chromosome decondensation occurred after exposure to static magnetic fields. Other
       abnormalities such as polyploidy, aneuploidy, chromosome stickiness, and dicentrics
       have been seen in some experiments with no consistent pattern.
    •  Sister chromatid exchanges are probably not induced by EM fields.
    •  Gene mutations are not induced by EM fields.
       10/22/90
5-24

-------
                            DRAFT-DO NOT QUOTE OR CITE

    In summary, EM fields clearly interact with genetic material, as detected by chromosomal
 aberrations.  Effects on transcription (gene induction) have also been reported (see Section
 5.3.1).  However, no obvious relationship between exposure parameters and effect is
 apparent, and it is premature to conclude from the genetic evidence available that EM fields
 are or are not likely to be carcinogenic.

 5.2. EFFECTS ON MITOSIS AND MEIOSIS
    Many of the publications considered in this section report an effect of EM fields, particularly
 the magnetic component, on cell cycle progression and/or mitotic index.  Although these end
 points are not genetic end points in the sense that gene mutations, DMA damage, and
 chromosomal aberrations are, they clearly demonstrate that EM fields affect DMA function.

 5.2.1. Extremely Low Frequency Electromagnetic Fields
    Plants have been commonly used in studies on the biological effect of ELF fields because
 the root meristem provides an easy to work with population of actively growing cells.
 Robertson et al. (1981) studied long-term exposure of pea roots in an aqueous inorganic
 nutrient medium [conductivity about 0.08 siemens/meter (S/m)] to 60-Hz electric fields of 140
 or 430 V/m and examined growth rate and mitotic index in the root tip cells. Mitotic index in
 this paper was defined as the number of cells in mitosis per 1,000 nuclei.  No significant effects
 were found at 140 V/m; however, at 430 V/m both growth rate and mitotic index were reduced.
 The peak reduction in mitotic index,  about 55% of control, occurred at 4 hours of exposure
 with gradual recovery at 6 and 8 hours of exposure. Reduction of growth rate was immediate
 and constant after exposure started and was about 40% at 2 days of exposure to 430 V/m.
 Growth  rate had almost returned to normal 5 days after exposure stopped. It is likely that the
 changes observed in this study and in the following two studies are due to membrane effects.
The induced membrane potentials reported in these studies of 3 to 7 mV in a 300-V/m field
 (which is considered the threshold for growth effects) and 6 to 12 mV in a 490-V/m field
"represent a significant fraction of the normal resting potential of most cells."
   Another report from the same laboratory (Brulfert et al., 1985) examined pea root growth
and mitotic index as described previously as well as cell cycle duration. Exposure in an
aqueous inorganic nutrient medium (conductivity about 0.08 S/m) for 48 hours to 60-Hz
electric fields of 430 V/m produced a reduction in root growth to  44% that of control roots,
                                        5-25
10/22/90

-------
                           DRAFT-DO NOT QUOTE OR CITE

comparable to the reduction observed in the previous experiment. The mitotic index,
determined after a 25-hour exposure, did not differ significantly from the control value.  This
result is not inconsistent with the previous study, where the reduction in mitotic index was
maximum at 4 hours with recovery by 8 hours. Duration of the cell cycle was determined after
a 25-hour exposure and found to be about 10% longer than that of control cells. The
conclusion reached was that reduced root growth was due to reduced cell elongation,
probably resulting from altered cell membrane function.
   A third report from this laboratory (Inoue et al., 1985) examined growth and mitotic index in
broad bean (Vicia faba) roots after exposure to 60-Hz electric fields of 200, 290, and 360 V/m.
As above, exposure was in an aqueous inorganic nutrient medium (conductivity 0.07 - 0.09
S/m) and lasted 4 days for root growth determinations and up to 30 hours for mitotic index
studies. Vicia faba roots were used because the cell size is greater than that of pea roots and,
for a comparable electric field exposure, a larger cell should have a larger induced
transmembrane potential and resulting greater sensitivity to electric fields. There was a
significantly reduced  rate of root growth proportional to field strength, and the rate of reduction
was significantly greater than that observed in pea roots. (A three-way analysis of variance
was performed on the data). Growth rate was almost back to normal by 4 days after exposure
stopped. There was  no significant difference in mitotic index (p>0.05) between exposed and
control roots  at any exposure, including an  exposure that reduced root growth to 35% of
control. The  authors say the results support the hypothesis that the ceil membrane is the site
of action of electric fields.
    One of the few studies that considered effects produced by the geomagnetic field was
done by Nemirovich-Danchenko and Chastokolenko (1976). They determined the mitotic
activity oiAllium fistulosum root meristems as a function of orientation of the seeds in the
geomagnetic field, morphological isomerism of the seeds, and age of the radicles when a 2000
Oe (200 mT)  magnetic field from a permanent magnet was applied.  Seeds of this onion were
characterized as right-hand or left-hand based on position of the embryonic radicle  in the seed
and positioned with the embryonic radicle in the direction of geographic  north or south. There
were therefore four possible experimental orientations. Age was represented by radicle length
which was 0.5,1.0, or 4.0 millimeters (mm) when the magnetic field was applied. The
magnetic field lines were perpendicular to the axis of the radicle and to the surface of the
earth. Without regard to morphological isomerism, the highest mitotic activity was seen when
       10/22/90
5-26

-------
                             DRAFT-DO NOT QUOTE OR CITE

 the radicle was oriented to the north and no increase occurred upon exposure to the 2000 Oe
 (200 mT) magnetic field for 3 hours. Activity was significantly lower with a southern orientation
 (p<0.01 was considered significant), but it increased to that of the northern orientation after a
 3- hour exposure to the 2000 Oe (200 mT) magnetic field.  The authors' conclusion is that with
 a northern orientation the maximum possible number of root tip cells are dividing, and the
 additional magnetic field  cannot increase the activity, whereas with a southern orientation, not
 all possible cells are dividing, and the additional magnetic field induces these cells to divide.
 There was no significant  difference in mitotic index with respect to isomerism. Apparently the
 magnetic field effect occurs before the genetically determined time of mass cell division in the
 meristem, and if the field  is applied after that time, there is  no effect on  mitotic activity.
    Herich  (1976) germinated onion seeds (Allium cepa) on wet filter paper in a 6000-G (600
 mT) static homogeneous magnetic field for 140 hours, then examined the root tips for mitotic
 activity and cell cycle progression. The mitotic index of exposed  cells was reduced to 73.3%
 of control cells due to an inhibiting effect of the magnetic field on  transition of cells from
 interphase to prophase.
    Greene (1983) grew onion bulbs (Allium cepa) in flats placed at various distances from an
 experimental 60-Hz high voltage (normally 895 kV line-to-ground,  no further data given)
 transmission line.  Flats were either on the ground, on 1 -meter redwood posts or on 1 -meter
 insulators.  The unperturbed electric fields measured at the exposure sites varied from  0
 (grounded  Faraday cage) to 15.5 kV/m and were about 50% higher than this on top of the
 posts.  Exposures lasted  5,10, or  15 days at which time root tips were examined and mitotic
 index reported as percent of cells in mitosis per 1000 cells.  One-way analysis of variance was
 performed on all data.  The author stated that there were many statistically significant
 differences between exposure stations; however, when the Duncan's multiple comparison test
was run on the data, there were no significant differences at the 5% level with respect to
exposure level.
    Gemishev (1976) studied the effects of constant magnetic fields on root growth and mitotic
index in sunflower (Hellianthus annus) roots. (This paper is in Russian with an English
summary and English headings for tables.) Seeds presoaked in distilled water were exposed
to a magnetic field of 450,1000, or 2000 G  (45,100, or 200 mT) for 1 hour and the roots were
examined at 24 and 96 hours for root and hypocotyl growth and at 48 hours for mitotic activity.
Growth of roots and hypocotyls was stimulated by the magnetic field, particularly at 450 G (45
                                         5-27
10/22/90

-------
                           DRAFT-DO NOT QUOTE OR CITE

mT). At 24 hours, the length of control roots averaged 8.55 ą14 mm, while at 450 G (45 mT)
the average length was 10.07 ą 0.16 mm. At 96 hours, the respective lengths were 177.78 ą
1.56 mm and 195.99 ą 0.89 mm.  Differences were less at higher field strengths.  Results were
similar with hypocotyl growth.  The mitotic index was about 2.4 times the control level at 450 G
(45 mT) and only about 0.3 times higher at 2000 G (200 mT).
    The effect of magnetic fields on meiosis  in lily (Lilium henry!) pollen mother cells was
reported by Linskens and Smeets (1978). They exposed anthers for 4 hours to a
"homogeneous magnetic field of 5000 G" (500 mT) (no additional information on the magnetic
field given) and found a significant number of abnormalities at anaphases I and telophases I
immediately after exposure (X2 = 6.38 and 4.39, respectively) and at anaphases II and
telophases II at 18 to 24 hours after exposure  (X2 = 12.61  and 4.12, respectively). Aberrations
observed  included bridges, fragments and chromosomes left at the equatorial plane. Also
normal synchrony of the meiotic divisions was reduced.  At 48 to 50 hours after exposure, the
number of abnormalities at telophases II was no longer significant (X2 = 3.71).
    Effects of ELF fields have been seen in test objects other than plants.  Mitotic delay was
induced in the slime mold Physarum polycephalum by 60- and 75-Hz electromagnetic fields
(Marron et-al., 1975).  The fields were produced by application of crossed electric and
magnetic  fields alternating in phase at 60 or 75 Hz at levels of 0.7 V/m and 2.0 G (0.2 mT).
Timing of the cell cycle can be done quite accurately in Physarum polycephalum because
mitosis is naturally synchronized in stationary cultures. Mitotic delay became significant (95%
confidence level) after 80 to 100 days exposure to the 60-Hz field and after 100 to 120 days
exposure to the 75-Hz field. The mitotic interval returned to the control interval during 30 to 60
days after the culture was removed from the field.  No significant delay occurred in cultures
exposed for 200 days to 75-Hz fields of 0.15 V/m and 0.4 G (0.04 mT). The same group of
authors also examined the effect of intermittent field exposure on cell cycle duration in
Physarum polycephalum (Goodman et al., 1984). A 76-Hz [sic] sinusoidal field of 1.0 V/m and
 1.0 G (0.1 mT) was applied to the cultures for 16 hours/day, 5 days/week, but the total
 exposure time was not given. The number  of hours to metaphase II increased from 15.97 in
 unexposed cultures to 16.40 in exposed cultures, an average difference of 0.43 % 0.03 hours
 (p<0.01 for this difference and all others given for this paper).  In comparison, continuous
 exposure to a 75-Hz field of 0.7 V/m and 2.0 G (0.2 mT) increased the duration of the mitotic
        10/22/90
5-28

-------
                            DRAFT-DO NOT QUOTE OR CITE

 cycle by 0.64 ą 0.03 hours.  Continuous exposure to a 75-Hz 2.0 G (0.2 mT) magnetic field
 alone increased the duration by 0.46 ą 0.03 hours, while continuous exposure to a 75-Hz 0.7
 V/m electric field alone increased the duration by 0.39 ą 0.03 hours.
    The ability of static electric fields to induce nondisjunction of the X chromosome in
 Drosophila melanogaster Canton-S flies was examined by Portnov et al. (1975). Exposure of
 females to 150 or 330 kV/m fields for 24 hours did not result in detectable X chromosome
 nondisjunction in F1 offspring (p>0.05).
    Barnothy and Sumegi (1969) kept male Swiss mice for 13 days in a vertical, homogeneous
 magnetic field of 9000 Oe (900 mT) with a gradient of less than 200 Oe/cm (20 mT/cm).
 Measurements were made to ensure that no high-pitched noise, which can be a stressor to
 mice, was present. Mitotic index determined in the liver was, per 400 cells, 11.84 ą0.80 in the
 exposed mice and 5.17 ą 0.35 in sham-exposed mice (p=0.0001).
    Mastryukova and Rudneva (1978) also found an effect of magnetic fields on mice as
 described in the section on chromosomal aberrations. They looked  at duodenal epithelial
 cells and in addition to chromosome damage they also reported changes in the mitotic index.
 C57BL and CBA mice were exposed for 1 day to a 1000 Oe (100 mT) static magnetic field
 between the poles of an electromagnet, and the mice were killed 0,1, 3, and 24 hours and 6,
 10,12, 20, 24, and 30 days postexposure. The mitotic index was slightly but significantly
 reduced in exposed animals at 0,1, and 3 hours (p<0.001, 0.05, and 0.05, respectively) and
 significantly increased at 6,10, and 12 days (p<0.01, 0.02, and 0.001, respectively).  The
 mitotic index of corneal epithelial cells of mice exposed for 1 hour to  magnetic fields of 3 to
 127 kOe (300 to 12,700 mT) was determined by Strzhizhovskii et al. (1979). The
 electromagnet had a cylindrical active zone and produced a two-component magnetic field.
 One component was constant and one had a saw-tooth waveform with a period of 30
 seconds. The authors state that "the intensity of the constant component coincided with the
 amplitude of the saw-toothed component."  No other information on the magnetic field was
 given  in the paper. The mitotic index decreased with field intensity, reaching 50% of the
control level at 8 kOe (800 mT).  The greatest rate of decrease (14.3% per kOe ,100 mT)
occurred between 3 and 8 kOe (300 and 800 mT), while the rate of decrease between 8 and
127 kOe (800 and 12,700 mT) was only 0.24% per kOe  (100 mT), reaching 20% of normal at
127 kOe (12,700 mT). One  day following cessation of exposure to  either 3 or 127 kOe (300 or
                                        5-29
10/22/90

-------
                           DRAFT-DO NOT QUOTE OR CITE

12,700 mT) the mitotic activity returned almost to normal and by 2 days had overshot to almost
twice normal. Return to normal from this point was somewhat different for the two exposures.
After 3 kOe (300 mT), the mitotic index declined to 10% above normal by day three and to
normal on day four. After the 127-kOe (12,700 mT) exposure, the mitotic index declined to
about 75% above normal by day three, about 15% above normal by day four with a slow
decline to normal by day 15.  The authors conclude that the magnetic field "causes  a
reversible retardation of the cells at late stages of the mitotic cycle, in particular, in the
premitotic G2 period." Part of the rise of mitotic activity above control levels is due to the
synchronous entry into mitosis of those cells blocked in G2.
   The effect of 50-Hz electric fields of 50 kV/m on the mitotic index of cultured human
embryo fibroblastoid cells was determined by Dyshlovoi et al. (1981).  The mitotic index of the
exposed cultures decreased with exposure time until at 24 hours the decrease was statistically
significant (p<0.001). In another series of experiments, the cultures were exposed to the same
intensity field for 48 hours at which time the mitotic index was reduced by about 30% (p<0.01).
Twelve hours following termination of exposure the difference was no longer significant
(p>0.05). The authors felt that the mitotic index reached its lowest value after a critical time
and no additional reduction occurred with longer exposure times.  To prove this they
maintained cultures in exponential growth in the field for 7 days and determined the mitotic
index at 48, 96,120,144, and 168 hours. As before, at 48 hours the mitotic index was
significantly reduced (p<0.01); however, at 96 hours the mitotic index in exposed cultures was
twice that of controls (p<0.001) and remained significantly above control levels until after 168
hours (p<0.05 at 168 hours). The fields also caused degenerative changes in the cultures.
The authors speculate that the degenerative changes were due to inhibition of synthetic
processes in the cells.

5.2.2. Radiofrequency Electromagnetic Fields
    McRee et al. (1981), as discussed in the section on SCE, exposed female CD-1 mice to
2450-MHz 20 mW/cm2, microwaves for 8 hours/day for 28 days and examined bone marrow
cells.  An average specific absorption rate (SAR)  was given as 21 milliwatts per gram (mW/g)
and deep colonic temperature increase as 0.7° to 1° C. There was no difference in the mitotic
index of bone marrow cells in control, sham-exposed or exposed mice.
       10/22/90
5-30

-------
                            DRAFT-DO NOT QUOTE OR CITE

   An effect of 9.4-GHz pulsed (1,000 Hz pulse rate, 0.5 microsecond pulse width)
microwaves on meiosis in male Balb/c mice was demonstrated by Manikowska et al. (1979).
The mice were exposed for 1 hour/day, 5 days/week, during 2 weeks to intensities of 0.1, 0.5,
1.0, or 10.0 mW/cm2, and spermatocyte and spermatogonial metaphases examined. The
number of metaphases I with univalents was significantly greater than controls at all exposures
at the p<0.001 level of significance. The number of translocations (metaphases with quadri-
or hexavalents) was significantly greater than controls at all exposure levels except 0.5
mW/cm2 (p<0.001). Metaphase I, metaphase II, and spermatogonial metaphase counts
(number of cells in meiosis per 1,000 cells examined) were significantly lower at 0.1 mW/cm2
exposure than in controls (p<0.05). Metaphase II counts were also significantly lower than
controls at 0.5 mW/cm2. Translocations occurred at  random with respect to chromosome
pairs, and no particular chromosome pair was more likely than another to have a translocation.
The authors speculate that the microwave exposure interfered with chiasma formation  and/or
behavior.
   Stodolnik-Baranska (1974) exposed human lymphocytes in culture to 2950-MHz [sic]
pulsed microwaves (1200 Hz pulse rate, 1  microsecond pulse width). At 20 mW/cm2 applied
after the cells had been in culture for 66 hours, the mitotic index increased with exposure time
to about twice the control value at 20 minutes (or after two 20-minute exposures separated by
30 minutes). Similar results were found after 3 or 5 hours exposure at 7 mW/cm2 when applied
after the 64th hour of incubation.  Little effect on mitotic index was seen when cells were
irradiated after 70 hours of incubation.  Irradiation of  unstimulated lymphocytes (no
phytohemagglutinin) induced blastoid forms and macrophage-like cells.

5.2.3.  Summary
   Table 5-3 is a summary of the section  on mitotic and meiotic effects of EM fields. As
shown, both static magnetic and electric fields can alter the mitotic index and cell cycle
progression of a number of cell types in a number of species. However, there  is no clear
pattern to the effect. Likewise ELF electric fields, at least in the 50- to 75-Hz range reported in
Table 5-3, cause perturbations of mitotic activity and  cell cycle progression.  Apparently an
inhibiting effect on mitotic activity, if it occurs, occurs early during exposure with recovery to or
beyond control levels at later times.  This effect was seen in plant systems and in a human cell
culture study. Therefore, a report of no effect in a test system at 24-hour exposure may mean
                                         5-31
10/22/90

-------
                        DRAFT-DO NOT QUOTE OR CITE
    s
    I

UJ

S
UJ
     §
     V)
     S.
    IT
    .a
    a
    3>
             T>
             S

             S
              : (O
              : o>
11
ei
Ł5
                      I

                      "
                                          o>

                                          75
                            g
                                                               I
11
yver activity with southern
entation 200 mT additional 1
sed activity of southern orie
3ds to that of northern
ented seeds
3 5 S S '§
o>
c c
tivity reduced due to inhibit!
ect on transition of cells fror
erphase to prophase
JtilS
15
totic index 2.4 times higher
mT and 1 .3 times higher at
OmT
^ in p
5 ir fa
13 =
iromosomal bridges and
gments, segregation errors
aphase and telophase 1 and
OSS

1. increased
S
^
\. increased at 0, 1 , 3, hr po;
posure
2* X
CD
M
1. increased at 6, 10, 12 day:
stexposure
58.
o E
totic index decreased with ii
Id intensity but not in a unifi
inner
IS E
                                               :=   -a
              <0
              o
  CO
  _3
  Q.
= P


•II
cS.
•§22
O O CD

f o o

si|
O c?H
         .Ł
         ;> c

         | |

         i I
5s
s •Ť
IS.
        o
        c
        J3
        CO
        O
index
X
it
II
if
                                          c
                                          8
                                          c
~ S
S'J
Si
S5 Q-
2 0>
                                                                 O>
                                                                 C
        X "
        1 =
        .9 o
        •5 Ł
                                                               31

•D
.0)

_g

~a>
c
0>
CO
o
s
vt
•g w
o o
S rr
o
•s
a>
c
D)
CO
E

o
3
•a
1

8
CM w
8-
^ TJ
in o>
T
•0
Q)
^=
O
•^
Ł
O)
CO
o
I
CO
1-
8
c magnetic field,
                                               2
                                               ca
                                               •f:
                                               _g
                                               c
                                               CD

                                               !
                                               o
                              TI _^
                              " T?"
  c J2
E-iS
ť ™-s-
115*
,g g O g

lifi
5-2S CD
c -c
o E
'E Ť
0 S
a c?.!2
M/um cep
(germinati
root tip eel
f
O CO
118
C " Q.
to o>ť=
Ł.. C *- 	
IPf
•a i -S 1
Isll
8.
^,
S^
Lillium her
mother eel
"3T
CO
CO
.2
l"
                                                   0

                                                   S
                                                        §
                                                        o
                                                               I
                             CD ^
                             'o co
                             E E
                             toŁ


                             11
                             I §
      10/22/90
                   5-32

-------
                               DRAFT-DO NOT QUOTE OR CITE
b
o
DC
O
LU

§

O
o
Q
 I

t
cc
o



c
CD
•S
.8.



IOSIS (continued)
Result
HI
5
Q
CO
I -
I i
& "Ť
LU
Ll_
LL
LU

TABLE 5-3.
Exposure







CO
s
IT
.0
*
co
O)

"7i5
*-
c~
o
2
CD

Maximum decrease of M.I. at 4 hr
exposure to 430 V/m recovery by
8 hr. Constant reduction in
growth rate from start of exposure



Ťl
.Ł 2
~€
§0
S§,


Hz, 140, 430 V/m chronic
g






CO
o
0
2! S
•5 IT
IE
m




^~
"5
a>
•e

m
No effect on M.I. (measured at 24
hr exposure)
Root growth reduced (measured
at 48 hr)
Cell cycle duration increase
(measured at 24 hr)


.CD
15
o
X
CD
1
O
1
5 .-


k.
c.
E
>
isT
S






1
s
CO
8.

3 T-
10 0> Cť
Sh- *^ *-
o> r
? co 1 0 '5
_T §> "Ť ť> o
CO T- *- C 'o
S ^ t O
CD <= o -o Ł
o i i 1 i
JE O 5 O S ť S § .E S S S
II &%* ° ŤiŤ
s.l ?.-p-s is §18.
i? 1-S5 |s ^§g
2Q. CDNCD ^CD SO^
CO TJXo JJE >-SS>
Ł6. - cgc 8.1 1| =
OTJX: o ""-"N OCD co^o
igSoS S"1 -SJr S"^
•"^S "> = e.K? 075 "S35
o-gŁ o cpg-^ |Ť wf Ť
2Łt> 2 WT3<0 5j3 Ť)ŁŁ


1.
I
S^ S S-_ g- S
?? "g Ťf Ť - 1
Ť— o •— *•* o ^5 Ť
o ~ o o ;>> o o
^ °> ť= "i= 0 S • -J
S B S 3=5 S S
5 2 5 is -2- 5 5
^.
TO 1— O
o Ť -a Ł o
§•Ť S Sg4 s.
.- o " o°$ * Ť
^ ^ CD S c 3e ^ "°
>Ł c gfw |- ">•
|l it ?si* s PE 1
8: 1? !ll| tf|l I
S.la ?5 ilii ss.E-B s
Ł^I ŁE ISSsS- Ł^^e Ł
So S o ^ o 2 o.t m > N. CM o
COO) CD^  §• §• s
O ^< ^ Ł. S
c wŤ §.^ §.§• -gS
I 1 1 i i i i ' 2 1
"O CCCOCD (OCD COn
§§.2ťg jog go
^ S C ^= ^= ^5
CQ 5OŁ42- C-H. xS

                                       5-33
10/22/90

-------
                 DRAFT-DO NOT QUOTE OR CITE









I
i
o
5
ť
_Ł
1
CIS ONI
UJ
u.
Ul
TABLE 5-3.










1
1 '






W
Ł





_
1
1



I
Ťi



W
2
S)
ir

'a
3)


T>
o
S=


23 S *-
i^i "*"*
S
nrj
0>
CO
C
1

o
^3
i§


Increase with both exposures





X
CB
•a
c
0
§
is


T- CM
2950 MHz, 1200 Hz pulse rate,
fisec pulse width, 20 mW/cm2,
min or 7 mW/cm2 3-5 hr
Ł
Q.
I
C
03

3
HI


s •
73
'S
1
o

"c
en
3 g
CO J3 5
Metaphase 1 with univalents
increased at all exposure level
increased translocations at all
lowest exposure levels,
metaphase 1, II, and
spermatogonial metaphase cc

CO
"° Ť
c S E
._ TO o)
x c 
-------
                            DRAFT-DO NOT QUOTE OR CITE

only that the author(s) didn't take a sample soon enough.  The three studies using RF fields
are not directly comparable as seen in Table 5-3, and no general conclusion is possible.

5.3. EFFECTS ON TRANSCRIPTION, TRANSLATION, AND CELL TRANSFORMATION
5.3.1.  Extremely Low Frequency Electromagnetic Fields
    Goodman and her collaborators have published a number of papers in which cellular
transcription was induced in gnat (Sciara coprophila) larval salivary glands following exposure
to ELF fields. Goodman et al. (1983) exposed salivary glands to pulsed magnetic fields in 0.5
milliliters (mL) of Schneider's Drosophila medium in petri dishes between a pair of 10-cm by
10-cm  Helmholtz aiding coils oriented vertically, producing a magnetic field parallel to the
bottom of the petri dish.  The glands (still attached to the larval bodies) were exposed to either
repetitive single pulses (single 380-microsecond pulses of 15 mV amplitude repeated at 72 Hz)
or to repetitive pulse trains [5-millisecond pulse trains of 200-microsecond pulses (also 15 mV
amplitude) repeated at 15 Hz]. The rate of change of the magnetic field was about 0.1
G/microsecond (0.01  mT/microsecond) for the pulse trains and 0.05 G/microsecond (0.005
mT/microsecond) for the single pulses. The authors  measured transcription  in three ways: (1)
nascent RNA chains attached to specific chromosome regions were identified by
autoradiography; (2) nick translation using deoxyribonuclease I to  identify  transcriptionally
active chromatin regions; and (3) RNAs of various size classes were isolated and analyzed for
changes in the pattern of tritiated uridine incorporation. At 15 and  45 minutes of exposure to
the single pulse field, there was a specific increase in RNA transcription in  most of the bands
and interband regions of the chromosomes.  At 30 minutes, exposure transcription was about
at the control rate.  Nick translation showed some "hot spots" of transcription at 45 minutes of
exposure. The pulse-train field led to a gradual increase in transcription up to 45 minutes of
exposure but not to the level reached by the single pulse exposure. Effects of both types of
fields decreased after 60 minutes of exposure.  Isolation of RNA on sucrose gradients showed
a fourfold increase in total RNA but an 11 -fold increase in the mRNA size class at 15- and
45-minute exposures to single pulse fields. All size classes of RNA were at control levels after
15 minutes of exposure to pulse-train fields, but after 45 minutes, the levels of all RNA size
classes had increased.
                                         5-35
10/22/90

-------
                            DRAFT-DO NOT QUOTE OR CITE

   Goodman and Henderson (1986), using the same test system just described, compared
the effect of wave shape on transcription induction. The magnetic fields were generated with
10 x 10 cm Helmholtz aiding coils orientated vertically. The authors used a 72-Hz sine wave
(0.8 mV positive amplitude, 1.15 mT peak magnetic field, rate of change of field was 0.5 T/sec,
induced electric field at a radius of 2 cm was 5 x 10"3 V/m) and a 72-Hz  repeating single pulse
(380-microsecond positive pulse width, 4.5 microseconds negative spike, 3.5 mT peak
magnetic field, rate of change of field was 9200 mT/second, induced electric field at a radius of
2 cm was 9.2 x 10"2 V/m). Autoradiographs of the salivary gland chromosomes, particularly
the X chromosome, showed similar induction of transcription for both EM field types.  The
grain patterns were consistent with induction of mRNA gene sites and suggested that
enhanced transcription was occurring at sites normally active at this stage of larval
development.  Analysis of RNA on sucrose gradients showed increased incorporation of
tritiated uridine into size classes consistent with processed and unprocessed mRNA (6-10 S
and 20-25 S).  When the sine wave frequency was raised to 222 Hz (0.37 mT) or 4400 Hz
(0.018  mT) a similar pattern  of induction resulted  but to a lesser degree. The relative
transcriptional activity of the 6-10 S size class was inversely correlated with frequency.
Goodman et al. (1987) did a more detailed analysis of grain count distribution over the X
chromosome of Sciara after exposures to 72-Hz single pulse, pulse train, or sinusoidal EM
waves  as described in the previous two papers.  In addition to enhanced transcription at
normally active sites, they also found transcription occurring at sites not detectably active in
control cells. The response was qualitatively the  same with the three different fields but the
sinusoidal and single pulse fields were' more effective than the pulse train field.
   Goodman and Henderson (1988), again using the S. coprophila salivary gland system,
reported altered polypeptide synthesis following exposure to ELF fields of various waveforms
and frequencies. Table 5-4 summarizes the exposures used.
   The polypeptide patterns obtained by two-dimensional gel electrophoresis were
qualitatively and quantitatively different for each type of exposure and different from control
and heat-shocked cells; however, conclusions about specific effects of  the different types of
exposure are not possible.  Individual proteins were not identified, and it is not possible to
ascribe the observed effects to either the electric field or the magnetic field.  No conclusions
could be made concerning any possible effect of frequency.
       10/22/90
5-36

-------
                            DRAFT-DO NOT QUOTE OR CITE
                    TABLE 5-4. CHARACTERISTICS OF ELECTROMAGNETIC FIELDS TESTED
Frequency
(Hz)


72
15
1.5
72
60
Positive
induced
amplitude
(mV)
15
14.5
2.5
0.8
0.8
Positive
duration

(Msec)
380
200
250


Burst
width

(msec)

5
30


Negative
space

(Msec)

28
10


Negative
spike

(Msec)
4500
24
4


Peak
magnetic
field
(mT)
3.5
1.9
0.38
1.1
1.5
Electric
field

(V/m)
9x10'3
9x10'3
1.5x10'3
5x10"*
5x10"*
   An effect of magnetic fields has also been shown on the lac operon system by Aarholt et
al. (1982). In this system the beta-galactosidase gene is under control of the lac operon and is
normally repressed by a represser protein.  If the dynamic equilibrium between synthesis and
degradation of the represser protein is changed, changes in rate of synthesis of beta-
galactosidase should be seen. The authors exposed the bacteria to a 50-Hz square wave
magnetic field varying from 0 to 0.7 mT and measured beta-galactosidase synthesis.  The rate
of synthesis was quite dependent on field strength; it started to decrease at 0.27 mT, was less
than one-third of the control rate at 0.30 mT, then increased to the control rate by 0.32 mT.
The rate remained at the control level until the field intensity reached about 0.51 mT when the
rate began to increase, reaching more than twice the control rate at 0.54 mT. The rate fell
sharply beyond 0.56 mT and returned to the control rate again at 0.58 mT.  The field strength
effect on synthesis rate was strongly dependent on cell concentration. The field strength
dependence was seen at 1.5 x 107 celis/mL (the lowest cell concentration reported), remained
constant until about 3 x 107, increased to its greatest level between 3.6 x 107 and 5.0 x 107
cells/mL, and disappeared above 1 x 108 cells/mL The effect is maximum when the
intercellular distance is about 30 m and is no longer present when the intercellular distance is
less than 20 ju.m. The mechanism of the observed effects in not understood; however, based
on the work reported in this  paper and on work by others cited in this paper, the authors
speculate that the effect involves the represser protein rather than the DMA.
   The following comment is not intended as additional proof of a transcription-inducing
potential of ELF fields but is  intended to show that this is a rapidly developing area and
additional peer-reviewed publications should soon be available. A number of meeting
abstracts and presentations have appeared in the past year showing transcriptional changes
                                         5-37
10/22/90

-------
                           DRAFT-DO NOT QUOTE OR CITE
in cells (including human and other mammalian cells) exposed to ELF electromagnetic and
magnetic fields.  Oncogenes are among the genes induced.

5.3.2. Radiofrequency Electromagnetic Fields
    Cell transformation has also been demonstrated following microwave exposure. This
subject is included here on the assumption that the transformed phenotype results ultimately
from altered gene expression.
    Balcer-Kubiczek and Harrison (1 985) presented what they considered to be evidence for
microwave carcinogenesis in vitro. Their conclusion was based on an observed synergistic
effect of pulsed 2.45-GHz microwaves (120 pulses/second, 83-microsecond pulse width,
SAR = 4.4 W/kg) and x-rays on the frequency of malignant transformation of C3H/10T1/2
mouse embryo cells.  The synergistic effect was seen only if the cells were treated with the
tumor promotor 12-O-tetradecanoylphorbol-13-acetate (TPA) following exposure. A similar
experiment was done with benzo[a]pyrene instead of x-rays, but no TPA treatment was given.
Temperature of the cell cultures during microwave exposure (37.2 ą 0.1° C) was controlled by
immersing the culture flasks in a constant-temperature water bath, but as discussed
previously, the effective temperature at localized intracellular sites may be higher than the
average culture temperature.  Cells were irradiated with microwaves for a total of 24 hours,
either continuously in the presence of 2.5 to 12.5 M benzo[a]pyrene or with an interruption
after 6 hours to allow for exposure to 1 .5 - 6 Gy of 1 00 kV (peak) x-rays. Following exposure
to x-rays, some cultures received 0.1 micrograms per milliliter (^ug/mL) TPA. Results of the
transformation studies are given in Table 5-5. No results are given on induction of cell
 TABLE 5-5. EFFECT OF MW IRRADIATION ON TRANSFORMATION FREQUENCY(x 103) IN C3H/10T1/2
                              MOUSE EMBRYO FIBROBLASTS
Treatment
X-ray only
X-ray + MW
X-ray + TPA
X-ray + MW + TPA
BP only
BP + MW
1.5Gy
0.31
0.40
1.80
6.0


4.5 Gy
2.9
2.9
5.1
8.2


OGy




10.0
10.3
       10/22/90
5-38

-------
                           DRAFT-DO NOT QUOTE OR CITE

transformation by microwaves alone or by microwaves plus TPA, obvious experiments when
the authors conclude that the microwave effect is at the initiation phase. Microwave exposure
alone reduced the plating efficiency by about one-half compared to sham-irradiated controls,
while TPA alone increased the plating efficiency by about 40%. TPA addition to
microwave-exposed cells raised the plating efficiency somewhat over that seen after
microwaves alone.  The authors speculate that the results were due to a membrane effect with
secondary DMA damage or to an effect on DMA repair, these effects being partially reversed by
TPA, allowing some cells to express the transformed phenotype that would otherwise have
died.
   A follow-up study was published by Balcer-Kubiczek and Harrison (1989) to further clarify
the suggestion made in the 1985 study that microwaves alone may act as an initiator of
neoplastic transformation in vitro or interfere with repair of damage caused by other
carcinogens.  Exposures of mouse C3H/1OT1/2 cells in culture to 2.45-GHz microwaves
pulsed at 120 pulses per second with an 83-microsecond pulse width for 24 hours, (SAR =
4.4 ą 0.8 W/kg at the cell monolayer) were as given in the previous experiment. Microwaves
alone, without post-irradiation TPA, or TPA alone, produced no transformed foci in the cell
cultures; however, post-irradiation treatment with 0.1 fig TPA/mL led to  a significant increase in
transformation frequency over the control level. Therefore, the conclusion is that microwaves,
as used in this experiment, act as an initiator in a two-stage transformation assay. Unlike the
previous study, there was no effect on plating efficiency by any treatment.

5.3.3. Summary
   Several experiments by Goodman and colleagues have shown that pulsed and sinusoidal
magnetic fields in the ELF frequency range between 1.5 and 222 Hz have the ability to affect
the transcription (or gene expression) of information from DNA to mRNA in gnat (Sciara
corprophila) larval salivary glands and  the translation of the mRNA message into protein
synthesis in the same system. The effect on gene expression is mainly one of enhancing the
activity of genes that are already active at that stage of larval development, but new sites were
also  induced by the field. The protein synthesis experiments showed the induction of different
patterns of molecular weight distribution for each of the waveforms used. Exposure of
Escherichia coli bacteria to a 50-Hz square wave magnetic field resulted in a complex
intensity-dependent enhancement and inhibition of the synthesis rate of a specific protein
                                         5-39
10/22/90

-------
                           DRAFT-DO NOT QUOTE OR CITE

known to be under the control of the lac operon and normally repressed by a represser
protein.
   Exposure to 2.45-GHz RF radiation pulsed at 120 pulses per second followed by treatment
with the phorbol ester TPA has induced the transformation to malignancy of the mouse
C3H/10T1/2 cell line (Balcer-Kubiczek and Harrison, 1989).  In this system the modulated RF
radiation is acting like an initiator in a traditional two-stage cancer promotion protocol in this
cell line.

5.4.  CALCIUM EFFLUX FROM  BRAIN TISSUE
   A rapid change in calcium concentration is essential in many physiologic, metabolic,  and
cellular processes (e.g., regulation of nerve membrane excitability, release of neurotransmitter
substances from presynaptic nerve terminals, mitochondrial function, the action ,of cyclic
nucleotides in controlling cellular activity, and the initiation of cell proliferation and tumor
promotion). According to Adey  (1988a), alterations in calcium efflux have been demonstrated
with low frequency EM fields, with low frequency electric fields, with combined low frequency
EM and static magnetic fields, and with RF fields amplitude-modulated at low frequencies.
   Calcium-ion (Ca++)efflux from brain tissue, sensitive to electric currents applied to brain
tissues in vitro, has been used as a biochemical marker to study the biological effects of EM
fields.
   Most studies on calcium efflux have used cerebral tissue.  Blackwell and Saunders (1986)
reviewed the effects of low-level  RF and  microwave radiation on brain tissue and animal
behavior and concluded that there is some evidence for effects of low level EM radiation on
Ca+ + exchange in nervous tissue, but they noted that many experiments reporting positive
effects have been criticized.  Examples of calcium efflux studies are presented in the following
section.

5.4.1.  Extremely Low Frequency Fields
   Bawin and Adey (1976) examined the effect of ELF fields on calcium efflux from chick
cerebral hemispheres, chick striated muscle, and cat cerebral cortex. The tissues, maintained
at 36° C during the experiment, were labelled in vitro with 45Ca++.  "Sets" of 10 brain tissues
were exposed for 20 minutes to  weak ELF sinusoidal electric fields of 1, 6,16, 32, or 75 Hz
with electric gradients at each frequency of 5,10, 56, or 100 V/m; 50 muscle samples were
       10/22/90
5-40

-------
                           DRAFT-DO NOT QUOTE OR CITE

exposed to a 16-Hz, 20 V/m field.  Controls consisted of sham-treated samples. The samples
were assayed for radioactivity and the data were smoothed by the removal of counts more
than 1.5 standard deviations away from the mean before statistical evaluation.
   The results of these experiments suggested that ELF fields inhibited calcium release from
cerebral tissue and the pattern of inhibition indicated the existence of frequency and amplitude
"windows." In the chick brain, the maximum reduction (p<0.01) occurred at frequencies of 6
and 16 Hz with field gradients of 10 V/m.  In the cat tissue, significant reduction in 45Ca+ +
efflux occurred at 6 Hz (p<0.05) and 16 Hz (p<0.01) with 56 V/m gradients.  Muscle tissues
were unaffected by field conditions that induced changes in 45Ca+ + efflux from brain tissue.
This study (Bawin and Adey, 1976) has been criticized, however, for the rejection of data that
were more than 1.5 standard deviations away from the mean of the exposed or sham data
before final analysis (Myers and Ross, 1981).
   In contrast to the inhibition of Ca++ efflux observed at 16 Hz by Bawin and Adey (1976),
Blackman et al. (1982) demonstrated an enhancement in Ca++  efflux at 16 Hz, but with
different exposure conditions.  Bawin and Adey had exposed the samples to an oscillating AC
electric field with only a small magnetic component, whereas Blackman et al. (1982) used an
AC electromagnetic field. Blackman et al. (1985a) tested the hypothesis that the differences in
calcium efflux were due to the AC magnetic component present in the system. The exposure
system consisted of a transmission line exposure chamber in which the electric and magnetic
fields were perpendicular to each other and oriented in the horizontal plane. To expose
samples to an AC electromagnetic field under altered local geomagnetic field (LGF)
conditions, a DC magnetic field was generated by a pair of Helmholtz coils which were placed
around the transmission line. The coils produced a  uniform magnetic field within the exposure
chamber that was parallel to the local vector of the geomagnetic field, which was inclined at
85° from the horizontal plane. One of the major findings of the study was that a 15-Hz signal,
effective in inducing a change in calcium  efflux when the  LGF was 38 tesla (T), was rendered
ineffective when the LGF was reduced to  19 T. Changes in the LGF also rendered ineffective
signals effective.  Blackman et al. (1985a) concluded that the AC magnetic component was
essential for the efflux enhancements observed in their laboratory.
   Blackman et al.  (1985b, 1988b) examined frequency-dependent exposure regions that had
been identified for the efflux of Ca+ + from brain tissue of newly hatched chickens (Gallus
domesticus). The frequency dependence of calcium efflux from  the brain preparations (32
                                        5-41
10/22/90

-------
                           DRAFT-DO NOT QUOTE OR CITE

chicks/exposure) was first tested using nine different EM fields, ranging from 1 to 120 Hz [LGF:
0.38 G (0.038 mT), 85° N (inclination of the magnetic field to the horizontal plane of the AC
electric and magnetic components)] (Blackman et al., 1985b). Two effective frequency
regions, one at 15 and 16 Hz and the other between 45 and 105 Hz, were identified. The two
frequency regions were not the same size, and the investigators decided to further
characterize the frequency dependence using higher frequencies (Blackman et al., 1988b).
Thirty-eight frequencies, ranging from 1 to 510 Hz (with a static magnetic field of 0.038 mT at
85° inclination), were tested using 28 to 32 chicks per exposure. The samples were exposed
in a transmission line exposure chamber to crossed electric [15.9 volts root-mean-square per
meter (Vrms/m)] and magnetic [73 nanotesla root-mean-square (nTrms)] fields. Cerebral
hemispheres from newly hatched chickens were removed, halved, and labelled in vitro with
45Ca++. Half of the halved hemispheres were exposed or sham-exposed to the field for 20
minutes. The other half (controls) were incubated for 20 minutes outside the exposure
chamber. The radioactivity in the control sample was used to normalize the radioactivity in the
paired-treated sample to adjust for possible influences caused by differences in sex, age, and
brain mass among the animals, and in the specific activity of the labelling solutions.
   When the differences in mean efflux values between exposed and sham-irradiated samples
were compared, there were no discernable patterns of response as a function of frequency.
However, calculation at each frequency of the p-value which combines the difference between
the means of the exposed and sham-exposed groups with the variance of each group
provided the investigators with a basis for hypothesizing the existence of three
frequency-dependent patterns of calcium efflux in the data. One pattern  occurred between 15
and 315 Hz, one occurred at 60, 90, and 180 Hz (but not at 300 Hz), and one occurred at 405
Hz. The authors speculated on mechanisms that could be responsible for EM field-induced
changes in calcium-ion efflux, focusing on the initial transduction of electromagnetic energy
into a small physicochemical change. Assuming that the LFG determines the frequencies that
are effective in the transduction step, magnetic resonance mechanisms, either nuclear
magnetic resonance or electron paramagnetic resonance, which operate through the
oscillating magnetic field and require an LGF, are the leading candidates. In nuclear magnetic
resonance, the oscillating magnetic field acts on nuclei with magnetic moments. The
interaction between the time-varying magnetic field and nuclear magnetic moments naturally
       10/22/90
5-42

-------
                            DRAFT-DO NOT QUOTE OR CITE
 found in biological systems provides the basis for nuclear magnetic resonance imaging
 (Blackman et al., 1988b).
    Another mechanism, the simple Lorentz-force interaction, in which an oscillating electric
 field, or an electric field induced from an oscillating magnetic field, causes charged species to
 move in an LGF, was considered as a possible explanation for the pattern of significant results
 observed at 60, 90, and 1 80 Hz. The authors state that the usefulness of this model is
 hampered by the absence of a known chemical entity that would explain the specific observed
 frequencies and LGF field strength.
    In a different type of study, Blackman et al.  (1 988a) tested the effects of ELF on
 field-induced Ca++ efflux in brain tissue of a developing organism. Fertilized eggs of Gallus
 domesticus were exposed in a parallel plate apparatus consisting of one ground plate
 between two energized plates.  During their 21 -day incubation period, the eggs were exposed
 continuously to either 50- or 60-Hz sinusoidal electric fields at an average intensity of 1 0
 Vrms/m. The LGF in the egg exposure apparatus was 40 microtesla (fiT) (0.04 mT) with an
 inclination of 55° N. The ambient 60-Hz magnetic field was less than 70 nanotesia (nT). The
 entire apparatus was mounted on a pivot which allowed the eggs to be automatically tilted
 through 66° once an hour.
    The chickens were removed from the exposure apparatus within 1 .5 days after hatching
 and their brains were removed, separated along the midline, and labelled for 30 minutes at 37°
 C with radioactive Ca+ +.  The assay consisted of a test and a control. A tube containing one
 of the brain halves of a pair was placed in the exposure chamber at 37° C (treated sample),
 and the tube containing the other brain half was placed in a water bath at 37° C for the
 20-minute exposure period (control sample). The samples that had been exposed to either 50
 or 60 Hz during incubation were exposed for 20 minutes to  either 50- or 60-Hz EM fields at
 average values of 15.9 Vrms/m and 73 nTrms (in an LGF of 38 fiT, 85°N to the horizontal plane
 of the AC electric and magnetic components).  Exposure took place in a transmission line
 exposure chamber. Efflux of radioactive Ca+ +  from the brains of the treated and control
 groups was then measured by standard procedures. The ratios of counts per minute in the
treated (exposed) samples were compared to those of the control samples.
   The brains from chicks exposed to 50-Hz fields during incubation and exposed to 60-Hz
fields in vitro exhibited increased calcium efflux  (40%, p<0.01 , Bonferroni-adjusted t-tests).
The brains from chicks exposed to 60 Hz during incubation and to 60  Hz in vitro did not. The
                                         5-43
10/22/90

-------
                           DRAFT-DO NOT QUOTE OR CITE

brains from chicks exposed to 60-Hz fields during incubation were not affected by either 50- or
60-Hz fields. The investigators concluded that exposure of the developing chick to ambient
power-line-frequency electric fields, at levels typically found inside homes, can alter the
response of brain tissue to field-induced calcium-ion efflux, but they stated that the
physiological significance of this finding for the intact organism or for other species is not
clear.

5.4.2. Modulated Radiofrequency Fields
    Bawin et al. (1975) demonstrated that weak VHF fields, amplitude-modulated at brain wave
frequencies, increased calcium efflux from the isolated brain of the neonatal chick. Ten
forebrains, preincubated with 45Ca++ for 30 minutes, were used for each field condition, and
each condition was tested at least three times. Fields of 147 MHz with field intensities of 1 to 2
mW/cm2 and amplitude-modulated at 0.5 to 35 Hz (modulation depths were kept between 80%
and 90%) were applied for 20 minutes.  One group of tissues was irradiated with an
unmodulated carrier wave, and controls were run in the  absence of fields. The counts of
radioactivity were normalized before statistical evaluation.
    Unmodulated radiations and fields modulated at 0.5 and 3 Hz produced no significant
changes in the 45Ca+ + efflux in comparison with the unirradiated controls. However, fields
modulated at frequencies ranging from 6 to 16 Hz produced a progressive increase  in the
45Ca+ + efflux frorrrthe brains (p<0.05 to p<0.01), then a gradual decline in efflux at higher
frequencies. This is indicative of a "windowed" effect, dependent on a narrow band  of slow
modulation frequencies.
    The Bawin et al. (1975) data evaluation has been criticized by Myers and  Ross (1981) for
the presentation of normalized control values, on the basis that normalization removes
important information about variation of the control data between different experimental tests
and between different experiments.
    Albert et al. (1987) examined 45Ca+ + efflux from cerebral cortex tissue slices and cerebral
 hemispheres that were prepared from Gallus domesticus chicks and exposed to 147-MHz RF
 radiation, amplitude-modulated at 16 Hz, and applied at a power density of 0.75 mW/cm2.  The
 data showed that exposure had no statistically significant effect on 45Ca+ + efflux. These
 results are in contrast to the increased Ca++ efflux observed by Blackman et al. (1980) using
 the same frequencies. However, Blackman (1987) notes that Albert et al. (1987) tested a
        10/22/90
5-44

-------
                            DRAFT-DO NOT QUOTE OR CITE

 different intensity (0.75 mW/cm2) from that tested by Blackman et al. (1980) (0.83 mW/cm2),
 suggesting the existence of an additional null effect intensity region. In a recent study,
 Blackman et al. (1989) confirmed the existence of narrow power density ranges for the
 enhancement of Ca+ + efflux from chick forebrain tissue exposed in vitro to 50-MHz RF
 electromagnetic radiation (magnetic field: 0.038 mT, 60°N), amplitude-modulated at 16 Hz
 utilizing a series of power densities. A statistically significant (p<0.001 to p<0.05)
 enhancement of calcium ion efflux was observed at 1.75, 3.85, 5.57, 6.82, 7.65, 7.77, and 8.82
 mW/cm2, while no change was observed at 0.75, 2.30, 4.50, 5.85, 7.08, 8.19, 8.66,10.6, and
 14.7 mW/cm2. In other words, six windows were observed with five being in a mathematical
 relationship to each other. Blackman et al. (1989) speculate that deterministic chaos is the
 potential cause of the multiple intensity windows observed. This probability would extend
 Fr?lich's model down to the ELF range.
    Lin-Liu and Adey (1982) examined the effect of weak sinusoidally modulated microwave
 fields on 45Ca++ efflux from synaptosomes  undergoing continuous perfusion. Synaptosomes
 are isolated subcellular neuronal elements that resemble synaptic terminals in situ. Their
 membrane properties can be more easily manipulated than whole brain or tissue slices. The
 synaptosomes,  prepared in duplicate from the cerebri of male Sprague-Dawley rats, were
 loaded with 45Ca+ + and applied to a millipore filter which was placed in the perfusion system
 and perfused with Ca-free medium for 45 minutes. One of the duplicate samples served as
 control and the other was exposed to the fields during perfusion. The perfusate was collected
 at one minute intervals and assayed for radioactivity. The temperature was maintained at 31°
 C to minimize synaptosomal autolysis. The 450-MHz field was either unmodulated or
 sinusoidally amplitude modulated at 16 or 60 Hz, the unmodulated signal had a power density
 of 0.5 mW/cm2 and the modulated signals had the same peak power with a modulation depth
 of 75%.  The electrical gradient produced by the unmodulated field was 43 V/m in air.  The field
 intensities used were not expected to cause significant changes in calcium efflux due to
thermal energy transfer. In the perfused, unirradiated samples, the rate of calcium efflux from
the synaptosomes showed a biphasic response, with a fast (half-time, 5 minutes) and slow
 (half-time, 40 minutes) phase. The 450-MHz field modulated at 16 Hz, applied for 10 minutes
during the second phase,  increased the rate constant for 45Ca+ + by 38% (p<0.01,
Mann-Whitney U test).  Unmodulated or 60-Hz modulated signals were not effective.
                                        5-45
10/22/90

-------
                           DRAFT-DO NOT QUOTE OR CITE

   These results support the experiments of Bawin et al. (1975,1978) that 16-Hz
amplitude-modulated signals can stimulate the release of preincubated 45Ca++ from isolated
brain tissue.
   Shelton and Merritt (1981) failed to demonstrate altered 45Ca++ efflux in rat cerebral tissue
preloaded with 45Ca++ and exposed to pulse-modulated (rather than amplitude-modulated)
microwave radiation (1-GHz carrier frequency) at various power densities (16 Hz at 0.5,1.0,
2.0, or 15 mW/cm2 or 32 Hz at 1.0 or 2.0 mW/cm2). Merritt et al. (1982) were also unable to
alter 45Ca+ + binding to brain tissue with pulse-modulated microwave radiation.  Rat brain
tissue was loaded in vivo with 45Ca+ + by intraventricular injection and exposed in vitro to
pulse-modulated 1-GHz radiation (SAR, 0.29 or 2.9 W/kg) or 2.45 GHz (SAR, 0.3 W/kg) and in
vivo to 2.06 GHz (SAR, 0.12 to 2.4 W/kg).  Merritt et al. (1982) suggest that because
pulse-modulated and amplitude-modulated signals [such as those used in experiments by
Bawin et al. (1975)] are quite different, the biologic effects induced by them may be different.
   Dutta et al. (1984) also reported enhanced Ca++ efflux from human neuroblastoma cell
cultures exposed to RF radiation at 915 MHz, amplitude modulated at 16 Hz at certain narrow
ranges of SAR (0.05 and 1.0 mW/g).  Subsequently, Dutta et al. (1989) conducted a similar
study using exposure conditions analogous to those used by Bawin et al. (1975) and
Blackman et al. (1980). Human neuroblastoma cells, labelled with 45Ca++, were exposed for
30 minutes to EM radiation at 147 MHz, sinusoidally amplitude-modulated at 16 Hz at various
SAR values (0.1, 0.05, 0.01, and 0.005 W/kg) (magnetic field: 0.016 mT,  53° inclination).
Calcium-ion efflux was also measured from the human neuroblastoma cells at 147 MHz and
various amplitude modulation frequencies (SAR, 0.05 W/kg) and from human neuroblastoma
cells and hybrid Chinese hamster-mouse neuroblastoma cell lines at 147 MHz, amplitude-
modulated at 16 Hz (SAR, 0.05 W/kg).  In all cases the results for the exposed groups were
compared with those for unexposed controls. The respective findings of the three studies
were as follows: significantly  enhanced 45Ca+ + efflux was observed at SAR values of 0.05
and 0.005 W/kg (p<0.001 and 0.003, respectively); significant 45Ca++ efflux from human cells
was observed at amplitude modulation frequencies of 16 and 57.5 Hz (p=0.008, and p<0.001,
respectively); and significant enhancement of 45Ca++ efflux was observed in both human and
nonhuman cell lines (p<0.001 for both). These studies show that cell lines derived from
tumors of the human central nervous system respond to modulated RF fields similar to normal
       10/22/90
5-46

-------
                            DRAFT-DO NOT QUOTE OR CITE

 nervous tissues or cell lines from nervous systems of chicks and cats. These results also
 confirm the findings of Bawin et at. and Blackman et al.
    In addition to the in vitro effects cited above, EM field-induced alterations in calcium efflux
 have been observed in intact animals. Adey et al. (1982) examined the effects of weak,
 amplitude-modulated microwave fields on calcium efflux from the cerebral cortex of 23 awake
 cats. The cerebral cortex was exposed while the animals were under ether anesthesia and a
 plastic cylinder was inserted and placed in contact with the pial surface to make a "cortical
 well"; at the conclusion of surgery, ether was  discontinued, and the animals were immobilized
 with gallium triethiodide for calcium efflux measurements. 45Ca+ + was placed in the wells for
 a 90-minute incubation, then the medium was replaced with nonradioactive solution. The
 solution was completely exchanged and samples were taken for scintillation counting every 10
 minutes for 3 to: 4 hours.  Field exposure, initiated  at intervals ranging from 80 to 120 minutes
 after incubation of the cortex with 45Ca+ +, lasted for 60 minutes. For the efflux experiments,
 the 450-MHz field (3.0 mW/cm2) modulated at 16 Hz (modulation depth of 85%) was applied in
 an anechoic chamber maintained at 28° C, with the cats oriented at right angles to the field. At
 the end of the experiment, the animals were killed  with an overdose of phenobarbital, and
 cortical samples were taken to measure the depth of 45Ca+ + diffusion. Sham controls were
 used in this study, but sham treatment was not detailed.
    The efflux of 45Ca+ + from preloaded cat cerebral cortex, not exposed to the experimental
 field, followed an exponential pattern with three phases:  an initial phase of rapidly declining
 efflux, lasting about 10 minutes;  an intermediate phase in which the slope of the efflux curve
 was somewhat  reduced (at 20 to 80 minutes); a third phase of gentle slope starting at about
 180 minutes from the beginning  of sampling and extending to about 210 minutes. The efflux
 curve for the test group followed the control curve  until field exposure was initiated 80 to 120
 minutes after incubation of the cortex with 45Ca+ + . Following field  exposure, the efflux curve
 of the exposed  group was interrupted by waves of increased 45Ca+ + efflux. The waves had
 periods of approximately 20 to 30 minutes and were irregular in amplitude, continuing into the
 postexposure period.  A comparison of the control and field exposure data with calculated
 predicted curves (binomial probability analysis) indicates that the field-exposed efflux curves
 comprise a different population from controls at a confidence level of 0.968.
   Tissue field measurements, performed in separate studies,  showed a field strength of 33
V/m (0.29 W/kg) for the interhemispheric fissure. Measurement of radioactivity in cortical
                                         5-47
10/22/90

-------
                           DRAFT-DO NOT QUOTE OR CITE

samples at the end of the study demonstrated that 45Ca+ + penetrated the tissue at the rate of
1.7 mm/hour.

5.4.3. Unmodulated Radiofrequency Fields
   Three of the studies mentioned in the preceding section under Modulated RF/microwave
fields included control groups that were exposed to unmodulated fields and examined for
alterations in calcium efflux or other membrane effects. No changes in calcium efflux were
observed at any of the unmodulated frequencies tested. These studies with brief descriptions
of field conditions are as follows:  Bawin et al. (1975) -147 MHz, 1 to 2 mW/cm2, 20 minutes;
Un-Liu and Adey (1982) - 450 MHz, 0.5 mW/cm2,10 minutes; Blackman et al. (1980) -147
MHz, 0.83 and 0 mW/cm2, 20 minutes.

5.4.4.  Summary
   The preceding studies demonstrate that calcium efflux from brain tissue was increased or
decreased by ELF fields, was decreased or not affected by modulated RF and microwave
fields, and was not affected by unmodulated RF radiation. These responses appear to be
highly dependent on specific field conditions: e.g., frequency, power density, and type of
modulation, emphasizing the complexities involved in comparing one study with another.
Although alterations in Ca+ + efflux indicate some interaction of ELF at the cell membrane, the
physiological significance of  this effect is still not fully understood (Blackwell and Saunders,
1986; Blackman et al., 1988a), and a direct relationship between the effect of EM radiation on
calcium efflux and tumor induction or promotion cannot be established at this time.
    The experiments of Blackman et al. (1988a) showing frequency selectivity in the calcium
 efflux from brain tissue at field strengths of 16 V/m and 0.07 fiJ (crossed  electric and magnetic
fields)  are the only laboratory studies showing ELF effects at levels comparable to ambient
fields in residential buildings. All  nonhuman effects measured in laboratory settings have
 occurred at field strengths much  higher than ambient levels.
    Adey (1988a) proposed that in the amplification stage of transductive coupling, the initial
 stimuli associated with EM oscillations elicit a highly cooperative modification of calcium
 binding to the glycoproteins  that  protrude from the membrane surface. Adey (1988a)
 suggests that the alteration in calcium binding could spread longitudinally, consistent with the
        10/22/90
5-48

-------
                            DRAFT-DO NOT QUOTE OR CITE

 direction of flow of extracellular currents associated with physiological activity and with
 imposed EM fields.

 5.5. INTRACELLULAR ENZYME RESPONSES
 5.5.1. Protein Kinases
 5.5.1.1. Modulated Radiofrequency Fields
    Protein kinases are enzymes that phosphorylate proteins on serine, threonine, or tyrosine
 residues.  These catalyzed phosphoryiations have profound effects on cellular protein activity
 and play a major role in the regulation of a wide range of cellular functions, including signal
 transduction and cell proliferation.
    Byus et al. (1984) examined the effects of modulated microwave fields on the endogenous
 activities of both cAMP-dependent and cAMP-independent protein kinases of cultured human
 tonsil lymphocytes. Cells were exposed in a Crawford cell exposure system to a 450-Hz field
 (peak intensity, 1.0 mW/cm2), sinusoidally amplitude-modulated at various frequencies
 between 3 and 100 Hz for up to 60 minutes; under the exposure conditions, no temperature
 rise was detected in the culture medium. Calf thymus histone was used as a substrate for
 monitoring kinase activity.
    At a modulated field of 16 Hz, no change (relative to controls) was observed  in cellular
 cAMP-dependent protein kinase activity following 15-, 30-, and 60-minute exposures. The
 same exposure condition, however, caused a 50% to 55% decrease in cAMP-independent
 protein kinase activity after 15- and 30-minute exposure periods. After longer exposure
 periods, 45 and 60 minutes, no detectable change in the enzyme activity was observed,
 suggesting that the decrease in enzyme activity at 15 and 30 minutes was transient and
 returned to control values even in the presence of continued exposure.
    The cAMP-independent kinase activity was also observed to decrease when the field was
 modulated at 60 Hz; however the decrease in activity, 15% at 15 minutes and 35% at 30
 minutes, was less than that observed at 16  Hz.  Again, as in  the case with 16-Hz fields, a
60-minute  exposure period at 60 Hz produced no decrease in activity,  suggesting a transient
effect.
    Experiments were also carried out to determine if other modulated frequencies caused
changes in lymphocyte kinase activity. Exposing lymphocytes for 30 minutes to modulated
                                        5-49
10/22/90

-------
                           DRAFT-DO NOT QUOTE OR CITE

frequencies of 3, 6, 80, and 100 Hz or unmodulated 450-Hz carrier caused no apparent
decrease in cAMP-independent cellular kinase activity compared to unexposed controls.
Decreases in kinase activity were observed only at 16, 40, and 60 Hz with the largest decrease
occurring at 16 Hz.
   This study clearly demonstrates a "windowed" effect for both time and modulation
frequency in the microwave-induced decrease of lymphocyte cAMP-independent protein
kinase activity. The significance of these results is unclear, however.  It is not known, for
example, what specific kinase (or kinases) was affected in this study.  The authors of the study
state that".. .our data offer no insight into the real biological effect of these fields upon
lymphocyte function in particular, or upon the general state of the  immune system." Although
the significance of the results are unclear, it should be emphasized that any effect on  protein
kinase activity of the magnitude observed in this study would be expected to have significant
effects on cellular activity because of the major role played by kinases in the control of cellular
function. For example, the attenuation of cellular responsiveness  (termed desensitization)
following plasma membrane receptor activation is known to be controlled  in some signal
transduction systems through receptor phosphorylation by specific kinases. Any unwarranted
decrease in the activity of one of these kinases, therefore, could cause perturbations  in the
signal system by allowing the receptor to be "on" for a longer period than appropriate. The
B-adrenergic receptor is an example of a plasma membrane receptor which is known to be
desensitized via phosphorylation by a specific cAMP-independent protein kinase (Benovic et
al., 1989).

5.5.2. Ornithine Decarboxylase Activity
   The enzyme ornithine decarboxylase (ODC) is the controlling enzyme  in polyamine
biosynthesis and is affected by a wide variety of hormones and growth factors active at the cell
surface (reviewed by Byus et ai., 1987); the activity of ODC can change rapidly and markedly
in response to extracellular signals.  ODC activity is elevated in all rapidly growing cells
including transformed or cancer cells, and is increased by phorbol ester tumor-promoting
compounds (reviewed by Byus et al., 1987). The following studies examined the effects of ELF
and modulated RF and microwave fields on ODC activity in established cell lines.
        10/22/90
5-50

-------
                             DRAFT-DO NOT QUOTE OR CITE

 5.5.2.1. Extremely Low Frequency Fields
    Byus et at. (1987) investigated the effects of a low-energy 60-Hz field on ODC activity in
 human lymphoma GEM ceils, mouse myeloma cells (P3), and Reuber H35 hepatoma cells.
 The exposure conditions were designed to allow the cell cultures to be physically isolated from
 any possible products of electrolysis at the carbon electrodes following field exposure.  GEM
 and P3 cells were tested in suspension culture in a series of tissue culture flasks connected by
 tubing filled with agar gel. During field exposure a small current [368 microamperes(^A)] was
 passed through the flasks.  This current produced an electric field of 10 millivolts per
 centimeter (mV/cm) (1 V/m) in the suspension.  This field was uniform over 80% of the area of
 the culture flask.  The cultures were exposed to the field for 1 hour, and ODC activities were
 compared to those of sham-exposed control cultures at time points ranging from 0 to 4 hours
 after exposure.
    A 1-hour exposure to the 60-Hz EM field with an intensity of 10 millivolts per centimeter
 [mV/cm (1 V/m) produced a threefold increase in ODC activity in human lymphoma GEM cells
 immediately after exposure. The activity continued to increase for 1  hour after exposure to a
 level fivefold greater than control cultures, then returned to control levels within 2 to 4 hours of
 exposure. P3 cells, exposed in the same apparatus to the 60-Hz field for 1 hour, exhibited no
 increase in ODC activity immediately after field exposure, but did show a two- to threefold
 increase in activity during the 1- to 2-hour period following exposure; the activity had returned
 to normal by 3 hours.
    Reuber H35 hepatoma cells were exposed to the 60-Hz field for one hour in monolayer
 culture in square Petri dishes connected by agar bridges. This system was developed to
 obtain greater uniformity in current distribution using low field intensities [0.1-10 mV/cm (0.01-1
 V/m)]. Since the resistivity of the culture medium is 50 ohm-cm, the current densities
 corresponding to these electric fields are 2 to 200^A/cm2.  ODC activity in the field-exposed
 cultures was compared to that of sham-exposed cultures 1, 2, and 3 hours after exposure.
   The response of the H35 cells was not a typical logarithmic dose-response relationship
 between field strength and degree of ODC induction.  Exposure of the cells at an intensity of
 10 mV/cm (1 V/m), to a 60-Hz field for 1 hour elicited an increase in ODC activity of nearly 50%,
which returned to control level by  1 hour after exposure. The ODC activity was only slightly
increased by the 5.0-mV/cm (0.5 V/m) field, was not increased by a 1-mV/cm (0.1 V/m) field,
                                         5-51
                                                                              10/22/90

-------
                           DRAFT-DO NOT QUOTE OR CITE

but was increased 30% by the much smaller field of 0.1 mV/cm (0.01 V/m). These data were
not analyzed statistically. In addition, the H35 cells were exposed continuously to the 60-Hz
[10 mV/cm (1.0 V/m)] field for 2 and 3 hours. The 2-hour exposure had no effect on enzyme
activity and the 3-hour exposure actually decreased activity in comparison to controls. At no
time during the 1- to 3-hour exposures did the temperature of the medium change by more
than 0.1° C.
   The authors' major conclusion from these experiments is that 60-Hz EM fields, similar in
type and intensity to those found in the environment, increase the activity of the enzyme ODC
inside the cell (Byus et al., 1987). The investigators could  not explain the heterogeneity of
response between the different cells. They also did not attempt to explain why the response
induced by 0.1 mV/cm was as large as at the highest field (10 mV/cm) while only background
rates were observed at 5.0 and 1.0 mV/cm.
   As we have shown in Section 1.2 of this document, typical domestic 60-Hz fields of 0.1 pT
and 33 V/m produce internal currents and electric fields on the order of 10'5 //A/cm2 and  10"6
V/m. The effects produced by Byus et al.  (1987) in these experiments were induced by cellular
currents  of 200^A/cm2 and electric fields  in the cell medium of 1 V/m. This is at least 1 million
times higher than internal currents and fields induced by ambient exposures. The authors, in
stating the similarity of their experimental conditions to background fields, referred to external
ambient fields rather than to the internal fields  experienced by the cells.

5.5.2.2.  Modulated Radiofrequency Fields
    Byus et al. (1988) examined the effects of frequency-modulated microwave radiation  and
TPA, a phorbol ester tumor-promoting agent, on ODC activity in cell cultures. The following
experiments were performed:

    •  Reuber H35 hepatoma cells, Chinese hamster ovary (CHO) cells, and 294T melanoma
       cells were exposed in circular Petri dishes in a Crawford cell exposure system, to fields
       of 450 MHz, 1.0 mW/cm2, sinusoidally amplitude-modulated at 16 Hz; controls were
       sham exposed;
    •  Reuber H35 hepatoma cells were  exposed to the 450-MHz fields modulated at 5,10,
       16, 20, 60, or 100 Hz; controls were sham exposed;
    •  Reuber H35 hepatoma cells and CHO  cells, either exposed to the 450-MHz field
       modulated to 16 Hz or sham-exposed for 1 hour, were treated immediately with TPA.
        10/22/90
5-52

-------
                             DRAFT-DO NOT QUOTE OR CITE

     The cultures were irradiated or sham-irradiated for 1 hour; modulation depth of the field
  was maintained at 75%-85%; there were no changes in the temperature of the culture medium
  between the beginning and end of exposure in any of the experiments. ODC activity was
  measured at various times ranging from 1 to 5 hours after exposure to the field or after
  addition of TPA. An additional experiment was performed in which DNA synthesis was
  determined in H35 cells that had been irradiated with a real or sham field for 1 hour and
  treated with TPA.
     In all three cell types a 1-hour exposure produced a notable (up to 50%) increase in ODC
  activity when compared with unexposed cultures. ODC activity remained elevated in the
  field-exposed Reuber H35 and CHO cell cultures for more than 3 hours following removal from
  the field, but persisted for only 1 hour in the 294T melanoma cell cultures.
     In the experiment in which the Reuber H35 hepatoma cells were exposed to the various
  modulated 450 MHz fields, modulation frequencies of 60 and 100 Hz failed to alter ODC
  activity in cultures to the field for 1 hour, whereas the modulation frequency of 16 Hz caused a
  50% increase in enzyme activity. ODC activity was also increased at modulation frequencies
 of 10 and 20 Hz, but to a lesser degree.
    The effect of the modulated microwave radiation on the induction of ODC by phorbol
 esters (TPA) was also tested in Reuber H35 cells and CHO cells. ODC activity, which is known
 to be stimulated by TPA, was further stimulated in H35 and CHO cells which were previously
 exposed for 1 hour to the 450-MHz field modulated at 16 Hz before treatment with TPA. This
 effect was most evident in the H35 cells 4 and 5 hours after addition of TPA (-13% increase),
 and in CHO cells 3,  4, and 5 hours after addition of TPA (-50% increase). A1-hour exposure
 to the same field did not alter either sham-control DNA synthesis or TPA-stimulated DNA
 synthesis, as measured by [3H]thymidine incorporation, indicating that the increase in ODC
 activity is not an effect secondary to the stimulation of cell division.
    Byus et al. (1988) noted that brief exposure of the cells to the EM field altered their
 responsiveness to TPA.  TPA has been shown to have a specific cellular  receptor which, when
 activated, becomes associated with the plasma membrane.  This phorbol ester receptor has
 been identified as a calcium and phospholipid-dependent protein kinase and given the
 designation, protein  kinase C. Protein kinase C, which is a cAMP-independent enzyme, has
been implicated in the regulation of a variety of cellular events, including modulation of
receptor functions for the major classes of hormones, adenylate cyclase activity, induction of
                                        5-53
                                                                            10/22/90

-------
                           DRAFT-DO NOT QUOTE OR CITE

ornithine decarboxylase, and the induction of cell proliferation (reviewed by Byus et al., 1988).
The results of this study are consistent with the idea that protein kinase C may be a target for
low energy EM fields, leading sequentially to a variety of altered intracellular events (Byus et
al., 1988).

5.5.3. Summary
    In an examination of modulated microwave fields on cellular protein kinase activity in
human lymphocytes, Byus et al. (1984) demonstrated a "windowed" effect (for both time and
frequency) for the microwave-induced decrease in cAMP-independent protein kinase activity.
(No effect was observed on  cAMP-dependent  protein kinase activity.) The maximum decrease
in activity (50%-55%) was observed at 16 Hz with lesser amounts of decrease at 40 and 60 Hz;
temporally, the decrease in activity was maximum after 15- to 30-minute exposures, and by 60
minutes the activity returned to normal. This microwave-mediated decrease in protein kinase
activity may have important  implications for hormone receptor function and the  regulation of
cell proliferation.
    The studies of Byus et al. (1987,1988) have also shown that ELF and  low energy
amplitude modulated microwave fields can increase ODC activity in various cell types.
Maximum ODC induction occurred in the range of 10 to 20 Hz, corresponding to the
frequency-dependent responses in brain tissue CA+ + efflux, both for low frequency fields and
for RF fields modulated at low frequencies (Bawin and Adey, 1976; Bawin et al., 1975).
    Byus et al. (1988) also demonstrated a potentiation by modulated fields of TPA-stimulated
 ODC activity in cultured cells.

 5.6.  PARATHYROID HORMONE AND THE PLASMA MEMBRANE
    Collagen is synthesized by osteoblasts and represents 90% of the organic matrix of bone
 (Rosen and Luben, 1983).  Collagen synthesis in cultured rat bones can be increased by
 treatment with insulin and glucocorticoids, and can be decreased by parathyroid hormone
 (PTH) and 1,25-dihydroxyvitamin D3 (reviewed by Rosen and Luben, 1983).  It appears that
 collagen may be important  in the calcification of bone matrix (Bloom and Fawcett, 1969). Low
 energy EM fields pulsed at  frequencies of 10 to 90 Hz are used to stimulate the healing of
 chronically ununited fractures in humans (Luben et al., 1982). The mechanisms of action of
 these fields, thought to be triggered at the cell membrane, have been studied in vitro with an
        10/22/90
5-54

-------
                             DRAFT-DO NOT QUOTE OR CITE

  isolated osteoblast-like cell line (MMB-1) and whole bone, using collagen synthesis and other
  end points as markers (Luben et al., 1982; Rosen and Luben, 1983; Cain and Luben, 1987).
     In bone cells, PTH and osteoclast-activating factor (OAF), both acting through cell
  membrane receptors, stimulate both the activation of adenylate cyclase to form cAMP and
  inhibit collagen synthesis (Cain and Luben, 1987; Rosen and Luben, 1983). Vitamin D3 also
  inhibits collagen synthesis, but via a cytoplasmic rather than a membrane receptor.

  5.6.1. Extremely Low Frequency Fields
     Luben et al. (1982) exposed cultured cranial bone from 3-day-old mice and MMB-1 cells (a
  line developed from primary cultures of mouse cranial bone cells) to EM fields similar to those
  used clinically to stimulate the healing of bone fractures, and then examined the responses of
 the cells to PTH, OAF, or vitamin D3 (Table 5-6). The tissues and cells were exposed, in
 incubators, to two pulsed fields of approximately 20 G (2.0 mT). One field, "single pulse,
 patient" (SPP), consisted of a continuous train of single pulses at a frequency of 72 Hz. Each
 pulse had an  initial  component 325 seconds long, with a drop of 20% between the peak of the
 rising phase and the onset of the falling  phase. The falling phase had an overshoot of
 opposite polarity with a typical peak amplitude 20% of the  initial deflection. The other field,
 "pulse train, patient" (PTP), consisted of bursts of pulses produced at a 4-kHz rate, each  burst
 lasting 5 microseconds and being repeated at a 15-Hz rate. The initial pulse was 200 seconds
 long, and it was followed by a deflection of opposite polarity lasting 18.5 microseconds and
 limited in amplitude to 20% of the initial deflection. The magnetic fields for both SPP and PTP
 induced electrical gradients of ~1.0 mV/cm (0.1 V/m) around a 1 -cm loop in the spatially
      TABLE 5-6. EFFECTS OF FIELDS ON cAMP ACCUMULATION IN BONE CELL MONOLAYERS
Agent
None
PTH
PTH
PTH
OAF
OAF
Dose
ng/mL

10
30
100
1
10
cAMP.
No Reid
2.1 ą 0.3
5.7 ą 0.8a
8.3 ą 1.0a
11.6ą 1.8a
9.3 ą 0.9a
13.8+ 2.1a
Dmol per 106 cells (% no field, aaent control*
SPP Reid
3.1 ą 0.6
3.3 ą 0.4 (58%)
4.1 ą 0.8 (49%)
5.3 ą 0.5a(46%)
4.2 ą 0.6 (45%)
5.6 ą 0.7(41%)

PTP Reid
2.3 •+• 0.5
2.6 ą 0.4(46%)
3.5 ą 0.6 (42%)
4.9 ą 0.8a(42%)
3.3 ą 0.5 (35%)
4.3 ą 0.8a(31%)
 Significantly different from control (no field), p<0.05.  Paired t tests indicated no significant differences be-
tween effects of SPP and PTP fields at any dose.
                                         5-55
10/22/90

-------
                           DRAFT-DO NOT QUOTE OR CITE

homogeneous portion of the field between the coils; the peak extracellular current density in
homogeneous conducting electrolytes would be ~1.0/Wcm2.
   The cells were exposed for 12, 72, or 90 hours in the fields, then were removed and treated
with either PTH, OAF, or 1,25-dihydroxyvitamin D3. Assays were performed for cAMP
accumulation, adenylate cyclase activity, and/or collagen synthesis.  In addition, total
adenylate cyclase catalytic units in the membrane were assessed by activation with fluoride.
(In this assay, MMB-1 cells were grown in the presence or absence of the SPP field for 72
hours and then removed; cell layers were disrupted to prepare membranes for adenylate
cyclase assay and the membranes were treated in the absence of fields with either PTH or 1
millimolar (mM) sodium fluoride, and the amount of cAMP formed was determined by
radioimmunoassay.)  The controls were treated with (1) no agent, no field (shielded); (2)
agents, but no field; (3) no agent, but fields. In the assays for cAMP accumulation, adenylate
cyclase activity, and collagen synthesis, no statistically significant differences were observed
between the no agent, no field controls (1) and the no agent, field exposed controls (3).
   The major findings of this study are as follows:

    •  The production of cAMP by the bone cell  monolayers was significantly increased
       (p<0.05) in the no-field controls by PTH and OAF. This increase did not occur when
       preparations were pretreated with each of the fields, particularly in the groups treated
       with 100 nanograms per milliliter (ng/mL) PTH and with 10 ng/mL OAF (Table 5-6).
    •  Neither basal nor fluoride-activated adenylate cyclase activity were altered in
       membranes from cells cultured  in the fields.
    •  The inhibitory effects of PTH on collagen synthesis, as measured by the incorporation
       of [3H]proline, were blocked in cells grown for 12 hours in the presence of the SPP
       field.  The cells were exposed only to the  SPP field and only for 12 hours; labelling with
       [3H]proline took place 42 to 48 hours after field exposure.
    •  The fields had no effect on the inhibitory effects of 1,25-dihydroxyvitamin D3) thought to
       act by a cytoplasmic, rather than by a membrane-dependent mechanism.
    The investigators suggest that because the fields inhibited the activities of PTH and OAF
 (which have membrane receptors), but did not change the activity of 1,25-dihydroxyvitamin DS
 (which has a cytoplasmic receptor), the cell membrane is probably the primary site of
 interaction with the EM field. The fact that PTH-activated plasma membrane adenylate cyclase
 was inhibited by the field, and the fact that adenylate cyclase catalytic units in the cell
 membrane and basal cyclase activity were not, suggests that the fields were not acting directly
        10/22/90
5-56

-------
                            DRAFT-DO NOT QUOTE OR CITE

 on the cyclase, but that they were interfering with the binding of hormone to receptor, the
 ability of the hormone-receptor complex to activate cyclase, or both. The other possible target
 for the fields is the coupling of the hormone-receptor complex to adenylate cyclase in the
 membrane. This type of effect could be mediated either directly, by effects on the intrinsic
 membrane coupling proteins or indirectly, by modification of other membrane functions (Luben
 etal., 1982).
    Cain and Luben (1987) conducted further in vitro studies to elucidate the biochemical
 mechanisms of EM-field effects on bone healing by examining the effects of exposure to
 pulsed fields on PTH-stimulated cAMP accumulation and bone resorption in mouse calvaria
 (the superior portion of the cranium). In contrast to the study of Luben et al. (1982) in which
 cells were exposed to the field for 12 or more hours, the cranial bones from newborn Swiss
 mice in the present study were placed in culture medium and exposed to pulsed EM fields
 (PEMF) of extremely low frequency for only 1 hour. The exposure system consisted of a 10-
 cm x 10-cm Helmholtz coil kept in a humidified incubator at 37° C and 5% carbon dioxide; the
 generator unit remained outside the incubator.  The waveform parameters used in the
 experiments were a positive pulse at 100 microseconds and a negative pulse of 2
 microseconds, repeated at a frequency of 15 Hz.  The induced magnetic field was
 approximately 8 G (0.8 ml) with an electric field strength of 0.6 mV/cm (60 mV/m) and a
 current density of 20/Wcm2 in the medium. The control bones were shielded during PEMF
 exposure.
   The investigators selected this system for its usefulness in observing early and late
 responses to PTH which acts on the cell through plasma membrane receptors. The early
 response to PTH can be detected as early as 1 minute after exposure to the hormone by
 monitoring cellular increases in the production of cAMP.  The late response, bone resorption
 (monitored by  release of extracellular Ca++ from the bone matrix), can be measured 72 hours
 after hormone treatment.
   For the cAMP assay, cranial bones were exposed to the PEMF for 1 hour and incubated
for 30 minutes  in medium with 5 mM  theophylline, a phosphodiesterase inhibitor, which inhibits
the breakdown of the accumulated cAMP. The bones were then treated with PTH and were
rapidly "killed" in a microwave oven.  cAMP levels were measured by radioimmunoassay.
   Following a 1-hour exposure to PEMF and a subsequent 30-minute incubation with
theophylline, the inhibition of PTH-stimulated cAMP accumulation was observed, similar to that
                                        5-57
                                                                            10/22/90

-------
                           DRAFT-DO NOT QUOTE OR CITE

reported by Luben et al. (1982), but more subtle. The observation was complicated by the fact
that exposure to PEMF accelerated the time course of cAMP response to PTH. cAMP levels in
field-exposed bones peaked after 3-minutes of exposure to PTH, whereas that in the
unexposed bones peaked after 5 minutes. As a result, the cAMP levels of the unexposed
bones were significantly lower (p<0.05) than those of the field-exposed bones at 3 minutes.
However, the cAMP accumulation in the field-exposed bones decreased within 5 minutes and
remained lower than the control levels throughout the 11 -minute observation period.
   For the bone resorption experiments, the neonates were injected subcutaneously with
45Ca+* 72 hours before the cranial bones were removed. Following incubations of 6 to 24
hours, to equilibrate exchangeable Ca++, the bones were irradiated for 1 to 5 hours. Within
30 minutes after removal from the field, treatment with various concentrations of PTH was
begun.  After preincubation and field exposure, some bones were "killed" by alternate freezing
and thawing. Those not frozen were called "live" bones. The bones were then cultured in
medium (37° C, 5% carbon dioxide) and were decalcified. The percentage of bone calcium
released during 72 hours of culture was determined by comparing 45Ca radioactivity in the
medium versus that remaining in the bones. The percent release for "dead" bones was
subtracted from the percent release of live bones.  The results of the experiments were based
on pooled samples.
   The main result of the bone resorption experiment was that a field exposure for 1 to 5
hours altered 45Ca++ release measured 72 hours after hormone treatment and field exposure.
At submaximal PTH doses [2.3 nanomolar (nM) and 6.9 nM], 1 - to 5-hour field exposures
 inhibited bone resorption (35% and 44%, respectively), but at the maximal dose of PTH, 23 nM,
field exposures did not inhibit resorption. In the absence of PTH, basal bone resorption of
 6.17% was increased to 8.66% (a 40% increase) after a  1-hour field exposure (a 5-hour field
 exposure was not tested in this part of the study), but basal cAMP levels were not affected.
The investigators noted that the data from this study showing that PEMF inhibited the
 hormonal action of PTH, cAMP accumulation, and bone resorption are consistent with the
 hypothesis that field  perturbation occurs at the membrane level.

 5.6.2. Modulated Radiofrequency Fields
    No data were found.
        10/22/90
5-58

-------
                            DRAFT-DO NOT QUOTE OR CITE

 5.6.3. Unmodulated Radiofrequency Fields
    No data were found.

 5.6.4. Summary
    The preceding studies show that pulsed magnetic fields inhibit the hormonal action of
 parathyroid hormone, which is to increase the concentration of cAMP, decrease the rate of
 collagen synthesis in bone cell cultures, and increase the rate of bone resorption. This action
 occurs at plasma membrane PTH receptors. These experiments indicate that pulsed magnetic
 fields interfere with the signal transduction system which is mediated by the binding of PTH to
 its plasma membrane receptor. Inasmuch as cell proliferation is also thought to be mediated
 through the activation of multiple signal transduction systems, it is possible that ELF also has
 the potential for causing changes in some of these systems and thus could have an effect on
 cell growth including the growth of preneoplastic lesions and tumors.

 5.7.  MELATONIN AND OTHER HORMONES
    In the previous section, the effects of ELF fields on parathyroid hormone-dependent
 aspects of collagen synthesis were examined.  Another endocrine gland, the pineal, and its
 hormones have been associated with certain forms of breast and prostate cancer in humans
 and with cancer induction in animals.

 5.7.1.  Background: Melatonin and Cancer
   Various investigators have reported an association of melatonin secretion with cancer in
 humans, particularly certain forms of prostate and breast cancer.  Fraschini et al. (1988)
 examined 254 cancer patients and found increased serum melatonin levels in 99 cases
 (38.9%), decreased levels in 15 cases (5.9%), and no change in 140 cases (55.2%).  Mean
 serum melatonin levels were significantly higher in cancer patients compared with 98 healthy
 controls (p<0.0001). Regardless of cancer type, serum melatonin levels were higher in cancer
 patients compared with controls: breast and lung cancer, p<0.001; colorectal and gastric
cancer, p<0.005; soft tissue sarcoma, p<0.01; and lymphoma, p<0.025. Fraschini et al.
 (1988) also observed that 66.7% of the patients whose tumors responded to chemotherapy
also exhibited increased serum melatonin levels following chemotherapy.
                                        5-59
10/22/90

-------
                           DRAFT-DO NOT QUOTE OR CITE

   Cohen et al. (1978) proposed that reduced pineal melatonin secretion may be a factor in
breast cancer risk.  Bartsch et al. (1981, cited in Wilson et al., 1988) reported that women with
breast cancer had reduced urinary melatonin levels. Danforth et al. (1982) noted altered
melatonin secretion in patients with estrogen-positive breast cancer. Bartsch et al. (1985)
reported that men with cancer of the prostate had lower nocturnal melatonin levels than men
without the disease. Stevens (1987) suggested that ELF field-induced exposure in rats may
result in loss of gonadal downregulation, resulting in increased circulating estrogen levels
which may in turn stimulate mammary tissue proliferation and hence increase breast cancer
risk.
   Tamarkin et al. (1981) reported that melatonin alters dimethylbenz[a]anthracene (DMBA)
mammary carcinogenicity.  Fifty-day-old rats were given  15 mg of DMBA and were divided into
four groups: (1) DMBA + vehicle; (2) DMBA  + daily melatonin injections (beginning at day
50); (3)  DMBA + pinealectomy (at day 20); and (4) DMBA + pinealectomy + melatonin.
Group 2 had significantly fewer mammary tumors than group 1 (controls), indicating that
melatonin inhibited carcinogenesis by DMBA; group 3 had more tumors than group 1,
indicating that removal of the pineal enhanced carcinogenesis; and group 4 had fewer tumors
than groups 1  or 3, indicating that melatonin  ameliorated the adverse effects of pinealectomy.
    From their studies, which demonstrated that rats constantly exposed to light had increased
DMBA-induced mammary tumors, Shah et al. (1984) and Mhatre et al. (1984) concluded that
constant light from birth effectively deprives female rats of melatonin and leads to a constant
availability of estrogen and elevated circulating prolactin, which increases the turnover of the
breast epithelial cells, thereby rendering the breast tissue more vulnerable to the
carcinogenicity of DMBA. Some experiments in rodents have shown an increase in mammary
cancer on administration of estrogen and of prolactin (Henderson and Pike, 1981).
    Immune and neuroendocrine functions cooperate closely to protect the organism from
external attacks (Maestroni et al., 1988). Maestroni et al. (1988) demonstrated in  experimental
studies with mice that melatonin has a general "up-regulatory" effect on the immune system.
Exogenous melatonin can  counteract the effect of acute stress and/or of pharmacologic
corticosterone on antibody production, thymus weight and antiviral resistance. Maestroni et
al. (1988) suggest that activation of T lymphocytes is necessary for the immuno-enhancing
and anti-stress action of melatonin.
       10/22/90
5-60

-------
                            DRAFT-DO NOT QUOTE OR CITE

    Melatonin can either stimulate or inhibit cell proliferation, apparently depending on dosage.
 Blask and Hill (1986) have shown that physiologic levels of melatonin inhibit cancer cell
 growth, while sub- and super-physiologic levels of melatonin do not. A melatonin-induced
 proliferation of the erythroid- and myeloid-bone-marrow cell compartments has been observed
 which apparently extends to all body cells (Di Bella et al., 1979). The growth of lung, stomach,
 and breast cancers; lymphoma; and bone sarcoma were depressed with melatonin treatment;
 the survival time of patients was increased and symptoms alleviated (Di Bella et al., 1979).
 This treatment is potentiated by simultaneously lowering the levels of circulating growth
 hormone. In vitro, melatonin exhibits oncostatic properties against certain cancer cell lines
 including carcinomas and breast cancer (Blask and Hill, 1986; Rodin, 1963). Melatonin has
 also been used in the treatment of leukopenia, in both chronic and acute lymphoblastic
 leukemia and during antiblastic chemotherapy (DiBella et al., 1979).  In contrast, there have
 been other reports indicating that the pineal gland either has no effect on or stimulates the
 growth of some tumors (Kachi et al., 1988).
    The inconsistent results of animal studies on the pineal gland and its hormones could be
 due to the dependence of pineal  response on the photoperiodic environment (Reiter, 1988).
 Of particular importance is the timing of the administration of melatonin, which is most effective
 in pineal-intact animals when given late in  the light period (Reiter, 1988).

 5.7.2. Extremely Low Frequency Fields
    Based on experimental evidence that shows an effect of light and ELF electric and/or
 magnetic fields on pineal melatonin production, and on the relationship of melatonin to
 mammary carcinogenesis, Stevens (1987) has  proposed a hypothesis that the use of electric
 power may increase the risk of breast cancer.
    Pineal production of the hormone melatonin, which shows a distinct circadian rhythm, is
 suppressed by light.  The circadian rhythm is evident in blood and pineal gland levels of
 melatonin: low levels in daylight and high levels at night (Tamarkin et al., 1985). Melatonin, in
turn, suppresses prolactin production by the pituitary and estrogen production by the ovary
 (Mhatre et al., 1984).
   If circulating levels of melatonin are reduced (by pinealectomy or constant-light exposure),
the growth of DMBA-induced mammary tumors in the rat is accelerated (see Section 5.7.1).
Stevens (1987) proposes  a scheme through which long-term exposure to ELF fields may act
                                         5-61
10/22/90

-------
                            DRAFT-DO NOT QUOTE OR CITE

as a "functional pinealectomy" and enhance mammary DMBA-induced carcinogenesis in rats
(Figure 5-1). The hypothesis is based on the idea that melatonin level affects production of
prolactin and estrogen, and that it is the action of these hormones that increases breast
cancer risk by increasing stem cell turnover.  In addition, Wilson et al. (1988) propose that ELF
may also have an effect on steroid hormone-promoted prostate cancers.
   Wilson et al. (1981,1986) demonstrated that melatonin production can be suppressed by a
60-Hz field.  Male Sprague-Dawley rats were acclimated to a daily 14-hour light: 10-hour dark
photoperiod at 21° C and 20% to 40% relative humidity (Wilson et al., 1981). At 56 days of
age, 20 animals in electrical contact with the reference ground  electrode were exposed to a
uniform, vertical 60-Hz field (field strength, 1.7 to 1.9 kV/m) in a parallel-plate exposure system.
The animals were exposed 20 hours per day for 30 days. At the end of exposure, animals
were killed in groups of 10 (5 exposed, 5 sham-exposed) at four different times during the
light/dark cycle (1400 [light], 2200 [dark], 0200 [dark], and 0800 [light] hours). All conditions
were the same for the exposed and sham-exposed animals except for the presence or
absence of the electric  field.  Pineal glands were removed and  quick-frozen usually within 2

                                 Chronic 60-Hz Electric Field

                          Pineal Gland:  Reduced Melatonin Production
                                  Ovary: Constant Estrogen
                                 Piituitary: Constant Prolactin
                                          I
                                          east
                                          I
Increased Turnover of Breast Epithelial Stem Cells at Risk
                           Increased DMBA Mammary Carcinogenicity
Figure 5-1.  Proposed mechanism by which chronic exposure to a 60-Hz electric field may increase
dimethylbenz[ajanthracene (DMBA) -induced mammary carcinogenesis in rats.
SOURCE: Adapted from Stevens, 1987.
       10/22/90
                     5-62

-------
                            DRAFT-DO NOT QUOTE OR CITE

 minutes of death.  Pineal melatonin was assayed by gas chromatography/mass spectrometry
 using an internal standard and the data were analyzed by analysis of variance.
    Exposed rats, killed at 0200 hours, showed a significant (p<0.05) reduction in melatonin
 levels compared with control rats. During the dark phase, there was a significant increase
 (p<0.01) in pineal melatonin of the sham-exposed animals, but no increase in the
 field-exposed animals (p<0.20), based on the internal standard.  In a duplicate experiment, the
 melatonin data followed the same pattern; however, there were large variances in the data, so
 the two sampling times in the dark period (2200 and 0200 hours) were combined and the
 light-period sampling times (1400 and 0800 hours) were combined. Melatonin levels for
 sham-exposed rats differed significantly (p<0.002) between light and  dark periods; in contrast,
 no significant differences were seen in melatonin levels of exposed animals between dark and
 light periods (p>0.05).
    In a similar study Wilson et al. (1983) reported that exposure to 60 kV/m also suppressed
 the nocturnal increases in melatonin.  The suppression of the normal nocturnal increase for
 melatonin was consistent with a reduction in serotonin N-acetyl transferase (SNAT) activity (the
 rate-limiting enzyme in the synthesis of melatonin from serotonin). Further studies, in which
 rats were exposed for 3 weeks to 60-Hz, 39 kV/m electric fields, demonstrated that the time
 required for recovery of the melatonin rhythm after cessation of field exposure was less than 3
 days, indicating the overall metabolic competence of the pineal is not permanently
 compromised by electric-field exposure (Wilson et al., 1986).

 5.7.3. Modulated  Radiofrequency Fields
   No data were found.

5.7.4. Unmodulated Radiofrequency Fields
   Elder et al. (1984) reviewed the effects of RF fields on endocrine gland function and
concluded that changes reported in hormonal activities and blood chemistry are similar to
those observed during increased thermoregulatory activity and heat stress and are generally
associated with SARs >1 W/kg. This conclusion is supported by Elder et al. (1987a, b) in an
update of the previous report (Elder et al., 1984). The endocrine effects reported by Elder et
al. (1984) appeared to have occurred in the presence of colonic temperature elevations of 0.3°
C or more.
                                        5-63
10/22/90

-------
                           DRAFT-DO NOT QUOTE OR CITE

5.7.5. Summary
   In the preceding sections, studies have been presented that demonstrate that exposure for
3 to 4 weeks to a 60-Hz ELF field suppresses the nocturnal production of melatonin in rats, but
that the overall metabolic competence of the pineal is not permanently compromised.
   Studies in humans have shown increased serum melatonin levels following chemotherapy
and decreased urinary levels in some cases of breast and prostate cancer.  In addition, animal
and in vitro studies have demonstrated that melatonin can inhibit tumor induction with
chemical carcinogens, can inhibit the growth of established tumors, and can enhance the
cellular immune response.  The results of these studies suggest that there is a relationship
between cancer and pineal gland function. In other studies, however, melatonin has had
either no or stimulatory effects on tumor growth. Some of the inconsistencies in these studies
could probably be resolved with improved techniques for dealing with the circadian aspects of
pineal gland function.
   The suppressive effects of ELF on  pineal melatonin production and the general oncostatic
properties of melatonin in several endocrine-stimulated tumors provide indirect evidence for
the hypothesis that ELF exposure may be a risk factor in the growth of these tumors. Studies
that incorporate all three parameters, ELF exposure, melatonin production, and breast cancer
induction,  are needed for further evaluation of this hypothesis.
   In other studies, pineal neurological activity and melatonin synthetic activity were inhibited
by static magnetic fields when the orientation of the field was changed by as little as 5
degrees, a change which is only a factor of 10 higher than ambient magnetic residential fields
(Welker et al., 1983; Semm et al., 1980). These studies and the role of the retina as the
magnetoreceptor (Olcese et al., 1985;  Reuss and Olcese, 1986) are discussed in greater detail
in Section 5.10.1.

5.8.  GROWTH AND DIFFERENTIATION
5.8.1. Extremely Low Frequency Fields
   Several studies have demonstrated a  growth-enhancing effect of ELF exposures on both
normal and neoplastic cells in vitro.  The stimulation of the growth of neoplastic cells is of
particular concern in the therapeutic treatment of bone fractures with ELF fields in cases where
a neoplastic lesion may be present.  Because osteogenesis is thought to occur as a result of
       10/22/90
5-64

-------
                            DRAFT-DO NOT QUOTE OR CITE

 differentiation of osteoblasts (Akamine et al., 1985), information on the effect of ELF on cellular
 differentiation is of interest.
    McLeod et al. (1987) demonstrated that protein biosynthesis in neonatal bovine fibroblasts
 (measured by [3H]proline incorporation into extracellular and intracellular protein) was reduced
 by low frequency sine wave electrical fields. Stable low-frequency current with DC flow limited
 to less than 0.1 % of the AC amplitude was provided by a programmable current source.
 Current was passed through the exposure samples for 12 hours via platinum electrodes that
 were separated from the bath by media bridges and convection barriers.  Experiments were
 performed over a range of current densities (0.1 ^wA/crn2 to 1 mA/cm2, root mean square) and
 frequencies (0.1 to 1000 Hz).  A frequency- and amplitude-dependent reduction in the rate of
 incorporation was observed. The data indicated an optimal frequency range for the alteration
 of extracellular matrix protein synthesis. Peak sensitivity was at 10 Hz, with a current density of
 only 0.5/Wcm2, at which a notable reduction in protein incorporation in the matrix component
 was produced. Furthermore, the response was dependent also on the orientation of the cells
 relative to the direction of the applied electric field.  The incorporation of radiolabel into
 intracellular protein reflected the pattern seen in the extracellular matrix (no other details were
 given). The investigators concluded that currents of physiological strength can stimulate a
 reduction in biosynthesis and thereby may influence tissue growth, remodeling, and repair.
   Whitson et al. (1986), on the other hand, applied 60-Hz, 1000 V/cm (100 kV/m) electric
 fields (the authors acknowledge that a magnetic field may also be produced) to human
 fibroblasts in vitro for up to 48 hours, and evaluated DNA repair and cell growth or survival. No
 effects were observed on any of the parameters examined.
   Akamine et al. (1985) examined the effects of a pulsed EM field (PEMF) on the growth and
 differentiation of F9 cells, a clonal line of embryonal carcinoma (EC) cells. EC cells are
 described as stem cells of teratocarcinoma that resemble undifferentiated cells of early
 embryos.  These cells can be induced to differentiate to endodermal cells in vitro by treatment
with retinoic acid.
   The current was produced by a generator  outside a carbon dioxide incubator connected to
two coils located inside the incubator. The coils were positioned so that the generated
magnetic component was normal to the culture surface and so that the induced electric field
becomes stronger as one moves out along a radius on the culture surface.  PEMF was begun
12 hours after plating the cells, and the stimulation was continued for 84 hours. The pulse
                                         5-65
10/22/90

-------
                           DRAFT-DO NOT QUOTE OR CITE

width was rectangular, the pulse width was 130 seconds, and the frequency was 100 Hz. The
magnetic field at the center of the field was either 1.0 G (0.1 mT) or 10.0 G (1 mT).  Immediately
after exposure the cells were counted and examined for morphological changes and
biochemical assays were conducted for cellular differentiation, based on the production of
plasminogen activator and the synthesis of glycopeptides.
   As evidenced by increased cell numbers, PEMF stimulated the growth of F9 cells with
retinoic acid (294% of control at 10 G, 176% of control at 1 G) and without (137% of control at
10 G, 150% of control at 1 G). Based on morphological observations, retinoic acid  stimulated
cellular differentiation in 90% of the cells not exposed to the field, in 58% of the ceils exposed
to 1 G, and in 46% of the cells exposed to 10 G. Retinoic acid-stimulated cellular
differentiation was also inhibited when based on production of plasminogen activator, but not
when based on glycopeptide profiles. Thus, the PEMF promoted the growth of the embryonal
carcinoma cells in the presence and absence of retinoic acid and inhibited retinoic
acid-induced differentiation based on morphological observations and on the production of
plasminogen activator.  Because field-stimulated growth of carcinoma cells was observed in
these studies, the investigators advised caution in the treatment of malignant tumors with
PEMF.
   The clonogenic capacity and surface properties of two human cancer cell lines  (Colo 205
and Colo 320, derived from adenocarcinomas of the colon) have been studied by Phillips et al.
(1986a, b) and Phillips and Winters (1987) using a standardized 60-Hz EM-field exposure
system. Changes in colony-formation (a measurement of proliferative capacity) of the cells
were assessed using a soft agar culture technique, and changes in surface properties were
evaluated using a monoclonal antibody binding assay, a transferrin binding assay,  and cell
lysis by human NK (natural killer) cells.
   Exposures to four different fields were performed concurrently in four exposure  chambers.
Exposures consisted of only an electric field [E+; current density = 300 microamperes per
square meter (aA/m2 )= SO^A/cm2)] a magnetic field [M = 1.0 G, 0.1  millitesla
root-mean-square (mTrms)], combined electric and magnetic fields (E+M+; at intensities
indicated), and unexposed control (E-M-). Electric fields were produced by transfer of a
uniform current density through closed system cylindrical cell-exposure chambers that had
been filled completely with the cell suspension.  A rotating magnetic field was produced using
two sets of Helmholtz coils, and the insulated glass incubator with plastic cell chamber holders
       10/22/90
5-66

-------
                            DRAFT-DO NOT QUOTE OR CITE

 was installed within the area of uniformity for the field.  The temperature of the chambers never
 varied more than 0.15° C and the temperatures of the exposure groups showed no consistent
 variation.
    Cells were exposed continuously for 24 hours. The cells were then removed from the
 chambers and prepared for the assays. Viability was greater than 90% following field
 exposure.  In five out of five experiments with Colo 320 DM and in four out of five experiments
 with Colo 205, cells exposed to M+ and E+M+ produced more colonies than the
 sham-exposed control cells, statistically significant at the p<0.01 or p<0.05 levels
 (Kruskal-Wallis nonparametric analysis of variance).  The cells exposed to E+ only
 demonstrated a mixed response (i.e., in some experiments, the numbers of colonies formed by
 E+ cells was significantly decreased, while in others the numbers of colonies formed by En-
 cells was either significantly increased or remained the same).
    The changes in cell-surface properties were assessed by quantifying the binding to the
 cells of monoclonal antibodies produced against Colo  205 and Colo 320 DM tumor-associated
 antigens. These properties were also altered by field exposure, as evidenced by a general
 increase in the expression of tumor-associated antigen in E+M+ and M+ cells. The
 alterations were not as dramatic as those observed in the cloning assay, but in several cases
 the increases in antibody binding were statistically significant (p<0.01 or p<0.05).  The effects
 on surface properties of the target cells were consistent with changes in plasma membrane
 structure and function following EM-field exposures described by other investigators (Luben et
 al., 1982).
    Transferrin is the major iron-transport protein in the body and is an obligatory growth factor
 for many cells when cultured in serum-free medium  (reviewed by Phillips et al., 1986b).
 Transferrin receptors are located on the cell surface, more so on the surfaces of malignant or
 proliferating normal cells, than on normal nondividing cells,  and the number of receptors on a
 cell is inversely related to cell density.  These may be the receptors for NK cells. In three
 separate experiments, the number of transferrin receptors quantitated on Colo 205 cells
 exposed to M+ and E+M+ fields were close to or exceeded the maximum theoretical number
 of receptors determined for the cell line and were independent  of ceil density. Thus, the cells
were no longer subject to the regulatory influence of cell density. E+ cells expressed fewer
transferrin receptors than were predicted on the basis of cell culture density.
                                         5-67
10/22/90

-------
                            DRAFT-DO NOT QUOTE OR CITE

   After the initial cloning assays were completed, the cells from the four groups were
transferred for maintenance in long-term culture.  After 4 months, increased numbers
(approximately twofold more than E-M- controls) of transferrin receptors were still present on
E+M+ and on M+ cells, while decreased numbers (approximately one-fourth of E-M- controls)
were present on E+ cells; lysis, measured by a standard 51Cr-release assay, was decreased
approximately 70% in both E+M+ and M+ cells but was increased about 53% for E+ cells.
After 8 months in culture, the cells still exhibited an increased reproductive capacity and
increased numbers of transferrin receptors.
   In summary, this study has shown that exposure of Colo 205 cells to EM fields resulted in
effects on cellular function, as evidenced by increased reproductive capacity of the cells,
particularly for cells exposed to combined electric and magnetic fields or to the magnetic field
alone. The increased proliferative response was correlated with increased numbers of
transferrin receptors (also indicative of growth potential) in ColoLT cells exposed to
electromagnetic and magnetic fields combined and to magnetic fields alone.  The EM-induced
increases in transferrin receptors were also consistent with changes in cell surface properties,
as were increases in the numbers of tumor-associated antigens, and changes in susceptibility
of the cells to lysis by NK cells, especially in cells exposed to E+M+ and M+ fields.  The
persistence of these effects suggest that under the conditions of these experiments, EM
exposures are capable of producing significant permanent changes in cellular structure and
function.
   It has been suggested that the mechanism for the healing of nonunion bone fractures with
PEMF is related to revascularization of injured osseous structures. To test this hypothesis,
Yen-Patton et al. (1988) examined the effects of PEMF on the rate of repopulation of denuded
areas of endothelial cell monolayers and the rate of endothelial  cell neovascularization in
culture.  Human umbilical vein endotheiial cells and bovine aortic endothelial cell cultures were
subjected to injury-simulating denuding of a vessel wall and were then exposed, along with
uninjured controls, to the ELF. The field was generated by 22.5-cm x 22.5-cm Helmholtz coils
(waveform 200 seconds wide, shaped as a burst of 20 to 21  closely spaced events); the burst,
5 milliseconds wide, was repeated at 15 Hz, resulting in a calculated induced voltage of 1.3
mV/cm (0.13 V/m) in the tissue culture dish. The magnetic field intensity at the center of the
Helmholtz coils was approximately 1 G, and the rate of change  of the magnetic field was 8.5 x
104 G/sec during field expansion and 4.3 x  105 G/sec during field collapse. Under these
       10/22/90
5-68

-------
                            DRAFT-DO NOT QUOTE OR CITE

 conditions, temperature changes in tissue culture media are <0.001(C (Bassett 1987, as cited
 in Yen-Patton et al., 1988).
    Reendothelialization was assessed by scintillation counting of live cultures that had been
 incubated with tritiated thymidine for up to 96 hours, and injury response (wound healing) was
 defined as the difference between thymidine incorporation of injured and uninjured cultures.
 Cultures were disrupted by passaging with EDTA-trypsin and monitored for tube formation
 (indicative of vascularization) at various intervals for up to 23 days.
    In the presence of EM fields, there was a small but statistically significant enhancement in
 growth rate of partially denuded endothelial cell monolayers as evidenced by increased
 tritiated thymidine incorporation (40% response in the human cells, and 20% response in the
 bovine cells). The cells exposed to the fields and entering the denuded regions were
 elongated and formed a sprouting pattern, while those outside the field had a more cuboidal
 morphology.
    Cells that were disrupted and passaged with ethylenediamine tetraacetate-trypsin
 reorganized into three-dimensional vessel-like structures after 5 to 8 hours of exposure to the
 EM fields and in the presence of endothelial cell growth factor, heparin, and a component
 fibronectin (protein) matrix.  In the absence of EM fields, vascularization of confluent layers of
 cells was observed only after long-term incubation (2-3 months). The investigators concluded
 that the discrete stages of neovascularization that were observed with field exposure were
 qualitatively comparable to stages of in vivo angiogenesis. With regard to carcinogenicity,
 angiogenesis (possibly promoted by EM fields)  is essential to neoplastic growth, as well as to
 the progression of benign to malignant tumors (Folkman, 1975).

 5.8.2.  Modulated  Radiofrequency Fields
    No data were found.

5.8.3.  Radiofrequency Fields
    No data were found for nonthermal effects of RF fields on growth or cellular differentiation
that could be related to cancer promotion.
                                         5-69
10/22/90

-------
                           DRAFT-DO NOT QUOTE OR CITE

5.8.4. Static Electric Fields
   Becker and Esper (1981) examined the effects of electrostimulation, at levels used to
stimulate osteogenesis, on the growth of human fibrosarcoma cells (HT 1080) in culture.
Stainless steel electrodes were inserted through the side walls of two chambers of a plastic
triwell culture dish.  The two wells were connected by a conducting agar bridge. The
chambers, containing coverslips, were seeded with the cells and incubated for 48 hours. A
current of 360 nanoamperes (nA) was transmitted between the two electrodes at an average
voltage of 1.1 Vfor 24 hours.  The third chamber, the control, received no current. The cells
on the coverslips were fixed for counting immediately after exposure.
   Cell counts revealed an approximately threefold increase in the cell population in  both
experimental chambers relative to the control chambers. Unlike the McLeod et al. (1987) study
in which an AC field reduced protein synthesis, Becker and Esper (1981) used a DC field with
no AC component. These preliminary results suggest that, because the growth of human
fibrosarcoma was stimulated in this study with currents and voltages used to stimulate
osteogenesis, it would be prudent not to administer electrostimulation to patients with
suspected premalignant or malignant lesions located within the current pathway (Becker and
Esper, 1981).

5.8.5. Summary
   The preceding studies show that currents of the strength used to stimulate bone repair can
stimulate alterations in biosynthesis and thereby may influence the growth, remodeling, and
repair of normal tissue. These fields also stimulated the growth of human fibrosarcoma cells
and embryonal carcinoma cells, and inhibited differentiation of embryonal carcinoma cells,
suggesting a tumor-promoting potential for ELF. The investigators in these studies advise that
caution should be exercised in the treatment of tumors with PEMF or the treatment of broken
bones with the fields in patients who also have tumors.

5.9.  IMMUNOLOGIC/HEMATOLOGIC EFFECTS
   The immune system is the physiological defense against a large spectrum of agents
including  bacteria, viruses, fungi, parasites, toxins from organisms, miscellaneous chemical
substances, and neoplasms. Two types of effects are possible:  immunosuppression and
immunopotentiation. Immunosuppression may result in an increased susceptibility to infection
       10/22/90
5-70

-------
                           DRAFT-DO NOT QUOTE OR CITE

 by microorganisms or to the development of tumors while immunopotentiation involves a
 generalized increase in immune responsiveness such as hypersensitivity (allergy) or
 autoimmunity. Impairment of the immune system could result in adverse health effects. The
 immune system consists of cells that are specialized for defense, broadly classified into
 lymphoid cells and phagocytic cells that produce humoral substances such as antibodies and
 complement.
   Assays of lymphocyte function, an important factor in cellular immunity, are used frequently
 to assess immune competence. Another, less specific, indicator of immune status is the
 peripheral blood cell count, particularly the lymphocyte fraction. Therefore, studies that assess
 hematological effects, as well as those that assess immune status, have been included in this
 section.

 5.9.1. Extremely Low Frequency Fields
   Fotopoulos et al. (1987), conducting an investigation of the effects on humans of exposure
 to 60-Hz electric and magnetic fields, reported preliminary results of blood chemistry,
 hematologic, and immunologic assays. Twelve subjects participated in four experimental
 sessions, spaced 1 week apart: two sessions involved 6 hours of exposure to a 60-Hz,
 9-kV/m, 16-A/m (2 x 10"2 mT) field, and two identical sessions involved exposure of humans to
 a sham field.  The subjects served as their own controls. The order of field presentation was
 counterbalanced under double-blind conditions.  Blood samples were collected in two
 sessions before the exposures to establish baseline conditions and then on the days of the
 experiments immediately before and after exposure to the real or sham field.  Details of
 exposure conditions were published elsewhere and are not available.
   Expected circadian variations were observed before and after the 6-hour exposure
 sessions among the hematologic, chemical,  and immunologic variables studied, and many of
the variables did not exhibit significant changes under field exposure conditions compared
with sham controls.  There were no significant differences in levels of calcium, glucose, uric
acid, albumin, potassium, or sodium between field-exposed and sham-exposed  subjects.
 However, lactic acid dehydrogenase (LDH) levels were significantly higher in the pre-exposure
compared with postexposure periods (p=0.004) on the first day but not the second day of field
exposure.
                                        5-71
10/22/90

-------
                           DRAFT-DO NOT QUOTE OR CITE

   No differences were observed in white blood cell (WBC) count, red blood cell (RBC) count,
hemoglobin (Hgb), hematocrit (Hct), mean corpuscular hemoglobin (MCH), mean corpuscular
hemoglobin concentration (MCHC), mean corpuscular volume (MCV), or platelets. The
differential cell count revealed a significant increase (p=0.043) in percent lymphocytes only on
the second day of field exposure.
   There were no field effects on total T-cells, B-cells, natural killer, and suppressor cells,
based on monoclonal antibody assays, and there was no effect on cell-mediated immunity,
based on a lymphocyte blastogenesis assay using T- and B-cell-specific mitogens.  However,
pre-exposure levels were significantly higher than postexposure  levels on the first day of field
exposures and not on the second day. This pattern was similar to the pattern with LDH. The
increases in lymphocyte counts on the second day of exposure and decreases in T-helper cell
counts on the first day of exposure suggest that the low frequency electric and magnetic fields
used in this study could enhance certain elements of the cell-mediated immune response in
humans. The investigators consider the findings in this study to be preliminary and defer
interpretation of the results until additional work is done.
   Seto et al. (1986) examined the hematologic effects of ELF fields on three generations of
Sprague-Dawley rats that were conceived, born, and raised in an electric field. The animals
(42 to 44 from each generation) were exposed to a 60-Hz, 80-kV/m, unperturbed vertical field
21 hours/day until they were approximately  120 days of age. Sham-exposed controls (42 to 44
from each generation) were conceived, born, and  raised under identical conditions, but were
not exposed to the field. The cell counts for the 135 field-exposed and 135 sham-exposed  rats
were analyzed statistically by multivariate analysis of variance, univariate analysis of variance,
and tests of simple effects.
    Red cell parameters, which included RBC, Hgb, Hct,  MCV, MCH and MCHC, were not
affected by field exposure. "Subtle" but statistically significant decreases in total white cell
counts (p=0.006), lymphocyte counts (p=0.027), and eosinophil counts (p=0.035) were
observed. The investigators commented that the observed hematologic variations related to
field exposure were similar to those observed in animals experiencing mild stress, implying
that the field (80-kV/m) could have induced  the effects through stress. They further stated that
such effects are not likely to be reliably replicated by an experiment in which sample sizes
smaller than the ones in this study are used.
       10/22/90
5-72

-------
                           DRAFT-DO NOT QUOTE OR CITE

    Ragan et al. (1983) examined numerous hematologic and serum chemistry variables in
female Sprague-Dawley rats (10-20/group) exposed to unperturbed 60-Hz electric fields at 100
kV/m for 15, 30, 60, or 120 days. The field strength delivered by the apparatus with no animals
or cages installed was 100 kV/m.  Each study was replicated once. The data underwent
rigorous statistical evaluation, and although no consistent effect of the field was detected,
statistically significant effects (using Student's t test) were observed in certain variables at
certain time points. For example, at 15 and  120 days of exposure duration the white count
was lower (p<0.02) and at 60 days it was higher (p<0.01) in the exposed as compared with
the sham-exposed groups. Lymphocyte values correlated with the total leukocyte counts.
There were no significant differences between exposed and sham-exposed rats in numbers of
neutrophils, monocytes, eosinophils, and basophils.  Platelet numbers were significantly
increased (p<0.02) after 60 days of exposure, but not at other time points. Bone marrow
cellularity was increased (p<0.05) in replicate 2 at 30 days of exposure. Similar occasional
changes were seen in RBC parameters (RBC counts, Hgb concentration,  RBC volume, MCV,
MCH, and MCHC) and in serum chemistry values (iron, triglycerides, alkaline phosphatase,
alpha and beta globulins). The investigators drew no definite conclusions but inferred that the
60-Hz field is of potentially low toxicity and emphasized the need for appropriate experimental
design and statistical analysis in studies of this type.
    A study was conducted in which ECR-SW strain mice were exposed to a 240-kV/m, 60-Hz
field 18 to 22 hours/day, 7 days/week for a total of 4500 hours of exposure (about 32 weeks)
before they were killed for tests (Fam 1980). Ten males and 10 females were used for
complete blood counts. Statistical analysis (analysis of variance and t statistics) indicated no
differences in the blood count values of the males at the a= 0.05 level, but there was a
significant difference between the exposed females and their controls in several parameters.
For example, exposed animals had a lower white blood cell count (a = 0.0021), a lower
hemoglobin (a= 0.0104), a lower mean corpuscular hemoglobin (a = 0.024), and a higher
percentage of bands (a = 0.0143). The investigators were not certain if the effects on the
females were due to field exposure or if they were the result of lower food and water
consumption. There were no statistically significant effects on lymphocyte counts in males or
females. At the end of the study, heart, lungs, liver, spleen, kidney, and ovary or testes were
                                        5-73
10/22/90

-------
                            DRAFT-DO NOT QUOTE OR CITE

examined microscopically. There were no histological findings that could be attributed to
exposure.
   Lyle et at. (1988) tested the effects of 60-Hz sinusoidal electric fields with intensities of 0.1,
1.0, or 10 mV/cm (0.01, 0.1, or 1.0 V/m) on T-lymphocyte cytotoxicity. The field exposure
apparatus consisted of six polystyrene culture flasks positioned vertically and joined in series
by short agar bridges, with carbon electrodes in the medium of the end flasks and cell-bearing
medium in the middle four flasks. Experimental conditions included field, sham, and plastic
control (no agar bridges). Cytotoxicity was measured using the 4-hour chromium release
assay. The effector (cytotoxic) cells used in the cytotoxicity assays were of the murine
T-lymphocyte line, CTLL-1.  These are normal T-Iymphocytes that are cultured in interleukin-2,
the T-cell growth factor, to stimulate growth.  The target cells were an allogeneic H-2d
B-lymphoma cell line, MPC-11. The assays were performed with electric fields present only
during the 4-hour assay, and using CTLL-1  cells pre-exposed for 48 hours to the electric field.
In addition, the effects of a 48-hour field exposure on growth of the CTLL-1 cells cultured in the
presence of "optimal" growth factor concentrations (1:2 dilution) or in the presence of
suboptimal concentrations (1:32) were determined.
   When the cells were exposed to the field only during the 4-hour assay, a nonsignificant 5%
decrease in cytotoxicity reaction of the CTLL-1 effectors against the target MCP cells was
observed in the field-exposed versus sham-exposed flasks. Forty-eight hours of exposure of
the effector cells to the 60-Hz, 10-mV/cm (1.0 V/m) field before the assay resulted in a  25%
inhibition of cytotoxicity (p<0.005, Student's T test), in comparison to the sham controls for
that experiment; the 1.0-mV/cm (0.1 V/m) field produced a 19% inhibition in comparison with
the sham controls for that experiment (p<0.0005); and the 0.1 -mV/cm  (0.01 V/m) field
produced a 7% (nonsignificant) inhibition. The field had no effect on the proliferation of the
CTLL-1  effector cells in the presence of interleukin-2, an indication that the inhibition seen in
the 48-hour cytotoxicity study was not due to the inhibition of cell proliferation but rather to an
alteration of the mechanism for cytotoxicity itself. The investigators concluded that, under the
conditions of the study, cytotoxicity shows a  dose response to a 60-Hz sinusoidal electric field
between 0.1 and 10 mV/cm, with a  detection  threshold that lies between 0.1 and 1.0 mV/cm.
   Morris and Phillips (1982) observed no effect on the primary antibody response to keyhole
limpet hemocyanin in mice following their exposure to 60-Hz, 0.15 to 0.25 kV/m fields for 30 or
60 days, or on the mitogen stimulation response of spleen cells to B- and T-cell mitogens in
       10/22/90
5-74

-------
                            DRAFT-DO NOT QUOTE OR CITE

 mice exposed to the same fields for 90 or 150 days. The reason for the difference in the
 results of this study and the study of Lyle et al. (1988) is not clear.  The main difference in the
 two studies was field strength which ranged from 0.01 to 1.0 V/m in the Lyle study and from
 0.15 to 0.25 kV/m in the Morris and Phillips study. In addition, in the Lyle study, cells were
 exposed in vitro, while in the Morris and Phillips study the whole animal was exposed.

 5.9.2. Modulated Radiofrequency Fields
    To test the effects of RF radiation on cells already challenged by a commonly encountered
 viral agent, Roberts et al. (1987) assayed mitogen responsiveness after exposure of influenza
 virus-infected human mononuciear leukocytes to continuous or pulse-modulated 2450-MHz RF
 radiation, specific absorption rate of 4 mW/mL (4 W/kg). Mononuciear leukocytes (MNL) were
 exposed or sham-exposed to influenza virus then exposed or sham-exposed to the RF
 radiation as continuous waves or pulse modulated at 60 or 16 Hz. The four groups of cells
 were then stimulated with the mitogen phytohemagglutinin (PHA). RF radiation exposure
 caused no changes in leukocyte viability or in mitogen-stimulated DMA synthesis by human
 mononuciear leukocytes infected in vitro with influenza virus when compared with sham-RF
 radiation-exposed ceils.
    Lyle et al. (1983) tested the effects of a 450-MHz microwave field, 1.5 mW/cm2, sinusoidally
 amplitude-modulated at 60 Hz on the same allogeneic cytotoxicity system as that described
 previously for the study of Lyle et al. (1988). An anechoic exposure system was used. The
 chamber temperature was kept to within 0.1 degrees of 35° C throughout the 4-hour exposure
 period, and the temperature within the sample well measured before and after the exposure
 did not differ significantly from the air temperature. The investigators characterized the
 cytotoxic response to the field with various manipulations of the procedure. The calculated
field strength  in the culture fluid was in the range of 10 to 30 mV/cm (1 to 3 V/m), somewhat
 higher than with the 60-Hz field tested by Lyle et al. (1988).
    Cytotoxicity to the target cell (MPC-11) by the CTLL-1 line was significantly inhibited (20%,
p<0.05 to p<0.0005 for five different experiments)  when the 4-hour cytotoxicity assay was
conducted in  the presence of the field; similar suppression was observed when the effector
cells (CTLL-1) were exposed to the field for 4 hours before the target cells were added for the
cytotoxicity assay.  The investigators attributed the suppression to an effect on the effector
cells, not the target cells. The results were similar when the cells were exposed to the field
                                         5-75
10/22/90

-------
                           DRAFT-DO NOT QUOTE OR CITE

during only the first 2 hours of the assay (exposure during the last 2 hours produced only
partial inhibition). This suggests a preferential effect on the recognition phase of cytotoxicity.
Cytotoxicity was also assessed in two experiments 1, 4, 9, and 12.5 hours after field exposure
of the effector cells.  In both experiments inhibition decreased as the time interval increased
(for experiment #1: 24% inhibition, p<0.05, during assay; 17%, p<0.025,1 hour; 14%,
p<0.025, 4 hours; 14%, p<0.05, 9 hours) and after 12.5 hours the cytotoxic inhibition was no
longer observed. The assay was then conducted after T-cells were exposed during the 4-hour
assay to various amplitude modulation frequencies (16, 40, 80, and 100 Hz).  All modulation
frequencies produced suppression, but it was maximal at 60 Hz (20%).
    Lyle et al. (1988) suggested that the differences in their experiments of 1988 and 1983 (the
60-Hz modulated 450-MHz microwave field of the 1983 study produced a rapid inhibition when
the 4-hour cytotoxicity assay was conducted in the presence of the field, whereas the 60-Hz
field in the 1988 study produced inhibition only after 48-hours of field exposure) could have
been due to either different characteristics of the two fields or different culture conditions
between the two sets of experiments (Lyle et al.,  1988).

5.9.3. Unmodulated Radiofrequency Fields
    A study of the long-term effects (including immunologic and hematologic parameters) of
RF radiation was conducted, using 14-week-old male Wistar-Furth rats (Smialowicz et al.,
1981). Sixteen animals were exposed to 970-MHz EM radiation [SAR = 2.5 mW/g (2.5 W/kg),
22 hours daily for 70 consecutive days].  The exposure system consisted of 16 individual
circularly polarized waveguides and similar nonenergized chambers served as sham-irradiated
controls. No differences were observed in the body weights, hematologic profile, or in vitro
lymphocyte responses to mitogens between the two groups. The only effects observed were
increased levels of triglycerides, albumin, and total protein which were thought by the
Investigators to be related to thermal stress.
    Chou et al. (1983) examined the effects of long-term exposure to continuous wave (CW)
microwave radiation on New Zealand rabbits.  Two groups of 16 animals each were exposed
to 2450-MHz fields in two experiments of 90 days each.  The incident power densities of the
first and second studies were 0.5 and 5 mW/cm2, respectively. After adapting to the anechoic
chamber exposure system, eight rabbits were  exposed for 7 hours daily, 5 days a week for 13
weeks,  and eight were sham exposed.  SARs were 5.5 W/kg in the head and 7 W/kg in the
       10/22/90
5-76

-------
                            DRAFT-DO NOT QUOTE OR CITE

 back at 5 mW/cm2. There were no changes in body weights, blood cell counts and
 morphology, clinical chemistry parameters, protein electrophoresis, lymphocyte blast
 transformation, and histology of the exposed animals in comparison with controls; and there
 were no changes in the eyes of the exposed animals. The only effect, decreased food
 consumption, was observed during exposure to the 5-mW/cm2 exposure.
    Wright et al. (1984) studied the effects of high frequency radiation in Wistar rats and
 Cynomologus and  Rhesus monkeys, exposed for 28 and 24 days, respectively, 23 hours/day.
 The rats were exposed to 28 MHz (125 mW/cm2) fields for 28 days and examined for
 histopathologic effects or were exposed to 220 mW/cm2 for 13 days and thyroid function was
 assessed. The Cynomologus  monkeys were exposed to 28 MHz (25 mW/cm2) for 24 days
 and examined for hematologic changes.  The Rhesus monkeys were exposed to 125 mW/cm2
 (28 MHz) radiation for 11 days and electrolytes were measured. There were no
 histopathologic changes in the rat or hematologic changes in the monkey that could be
 attributed to exposure. The rats did exhibit reduced uptake of iodine by the thyroid, reduced
 levels of plasma thyroid-stimulating hormone, and reduced ratio of protein bound to
 nonprotein bound iodine. However, these animals were exposed to a 220-mW/cm2 field (likely
 to induce thermal changes in man), and the thyroid effects were thought to be compensatory
 responses to an induced heat  load.
    Ottenbreit et al.  (1981) examined the effects of microwaves on the colony-forming capacity
 of human neutrophil precursor cells (CFU-C) in a methylcellulose culture system.  Bone
 marrow specimens  were aspirated from children with acute leukemia in remission, or from
 children with other disorders who had marrow aspirations performed for evaluation of clinical
 status, or for diagnostic determinations. The cells collected from the top layer of a
 Ficoll-Hypaque gradient were used for the assay. The cells were allowed to stand overnight,
were suspended  in  microcapillary tubes, and were exposed to 2450-MHz CW  microwaves for
 15 minutes in a fluid-filled waveguide irradiation system.  Irradiation of the cells for 15 minutes
at 31, 62,125, 250,  500, and 1000 mW/cm2 was conducted with  bath temperatures set at 37°
C for the two lowest power levels and at 7, 22, 37, and 41 ° C for the last four power ranges,
respectively. Sham-exposed controls were treated in the same way, but the microwave power
was not turned on.  The cells were then cultured with 20% fibroblast conditioned medium at
37° C, 7.5% humidity. Colonies were counted on days 6 or 7, and days 12-14.
                                        5-77
10/22/90

-------
                           DRAFT-DO NOT QUOTE OR CITE

   When colonies were scored on days 6 or 7, there was no reduction in the number of
colonies formed by field-exposed cells at power levels of 31 and 62 mW/cm2. As the power
level was increased to 1000 mW/cm2, there was a corresponding reduction in the number of
colonies formed by the microwave-exposed cells (p<0.05, Student's t test, at the four highest
levels).  The maximum reduction occurred at 500 and 1000 mW/cm2. At days 12-14, there was
a similar dose-dependent reduction in colony numbers. Additional experiments showed that
the effects observed were not related to temperature rise, or to the state of cell cycle, and were
irreversible. The investigators hypothesized that, because CFU-C require the addition of
exogenous stimulators for cell growth (fibroblast conditioned medium in this case), it is
possible that microwave irradiation alters the membrane receptors of CFU-C making some of
them unresponsive to the stimulation factors. In addition, the investigators suggested that the
cells from patients with leukemia or other disorders might react differently to EM radiation than
CFU-C from normal individuals.  Ottenbreit et al. are currently investigating this possibility.
   In a series of in vitro studies designed to characterize direct RF radiation effects on
immune function,  specifically the effects of RF radiation on the immunoglobulins in solution
when bound to the lymphocyte cell surface, Liburdy and Wyant (1984) observed that RF
radiation (10 MHz; 8500 V/m, applied electric field;  <0.134 W/kg, internal absorbed power; 20
V/m, internal electric field) altered the physical separation of immunoglobulin (Ig) and of
Ig-bearing T- and  B-lymphocytes during liquid gel chromatography and immunoaffinity cell
chromatography,  respectively.  Human serum was exposed to the 10-MHz electric field during
chromatography.  The gel column was placed perpendicular to the electric field surrounded by
a jacket of circulating water for conductive cooling. The temperature of the gel, maintained at
25° C, increased approximately 0.05° C during the first 15 minutes and then stabilized at
25 ą 0.01° C during the 18-hour exposure period.  To further ensure that changes in the
elution profile were not from temperature increases, two chromatography separations were
performed at 24 and  26° C in the absence of RF radiation field. No alterations in the elution
pattern were detected.
   The elution profiles for the three immunogolbulins examined (IgM, IgA, and IgG), were
altered by field exposure  during elution, as evidenced by accelerated elution of all three peaks.
This is thought to reflect an increase in the steric resistance of Ig molecules to the gel pores.
The investigators were able to almost completely rule out the possibility that RF
radiation-induced alterations in the gel matrix could have influenced the results.
       10/22/90
5-78

-------
                             DRAFT-DO NOT QUOTE OR CITE

     Effects on lymphocyte separation were investigated by performing immunoaffinity cell
 chromatography during exposure to 2500-MHz RF radiation (194 V/m,  < 0.117 W/kg, 10
 mW/cm2). This chromatography procedure separates lymphocytes based on antigen-antibody
 interactions at the cell surface.  Spleen cells from female Balb/c(H-2d) were fractionated at 4.0°
 C over Ig-derivatized agarose beads (derived with polyclonal antibody directed against all
 mouse lg+ classes) into Ig' and lg+ lymphocyte subpopulations. Because they were derived
 with polyclonal antibodies directed against all mouse Ig classes, all lg+ lymphocytes were
 expected to interact with the beads when no field is present.  RF radiation exposures, on the
 other hand, resulted in premature elution of 19% of the lg+ (B-cell) population. This premature
 elution resulted in a difference in distribution among cell fractions collected and indicates an
 RF radiation-induced alteration of specific Ig binding between B-lymphocyte cell surfaces and
 the column. Temperature fluctuations did not exceed  ą 0.03° C. Although these effects
 occurred at SARs well below the recommended safety limit for humans in the United States of
 0.4 W/kg, averaged over any 6-minute period, the investigators advise caution in extrapolating
 from in vitro to in vivo conditions. The results do suggest alterations in the structure of
 membrane-bound proteins that apparently affect receptor binding on the cell surface (Elder,
 1987b).
   There were two other in vitro studies in which thermal effects were carefully excluded.
 Szmigielski (1975) reported that radiation at 3 GHz with a power level of 5 mW/cm2 decreased
 the viability of rabbit granulocytes. Lin et al. (1979) reported that radiation of 2.45 GHz with  a
 power level of 60 to 1000 mW/cm2 reduced the numbers of granulocyte and macrophage
 colony-forming units from preparations of mouse bone marrow.

 5.9.4.  Summary
 5.9.4.1. Extremely Low Frequency Fields
   Preliminary data from the analysis of pre- and postexposure blood samples from humans
 exposed to ELF fields of 60 Hz, 16-A/m  (2 x 10'2 mT) and 9-kV/m for two sessions of 6 hours
 each resulted in the following observations (Fotopoulos et al., 1987): (1) no statistically
 significant differences in concentrations of calcium, glucose, uric acid, albumin,  potassium, or
sodium; (2) a significant decrease in lactic acid dehydrogeriase after one of the two exposure
sessions;  (3) no significant differences in red and white cell blood counts, hemoglobin,
                                         5-79
                                                                              10/22/90

-------
                           DRAFT-DO NOT QUOTE OR CITE

hematocrit, mean corpuscular volume, or platelets; (4) a significant increase in the peripheral
lymphocyte count after one of the exposure sessions; (5) no significant difference in total
T-cells, B-cells, natural killer cells, helper cells, or suppressor cells, based on monoclonal
antibody assays; (6) no significant effect on cell-mediated immunity; (7) although the level of
T-helper cells is typically higher in the morning than the afternoon, the difference was
significantly larger in exposure days than days of sham exposure. Although the authors
consider the data preliminary until more work is done, they point out that these field exposures
might be enhancing the cell-mediated immune response in humans.
    Seto et al. (1986) examined the hematologic effects of ELF on Sprague-Dawley rats that
were exposed to a 60-Hz, 80-kV/m, unperturbed vertical field 21  hours/day through three
generations.  "Subtle" but statistically significant decreases in total white cell counts,
lymphocyte counts, and eosinophil counts were observed. Red cell parameters were not
affected by field exposure. The investigators implied that the field could have induced the
effects through stress.
    In another study, no consistent effects were observed in rats exposed to unperturbed
60-Hz fields at 100 kV/m for 15, 30, 60, or 120 days (Ragan et al., 1983).  In two of eight
experiments, the white count was lower and in one it was higher in the exposed compared
with the sham-exposed groups. Platelets were significantly increased after 60 days of
exposure, but not at other time points. Bone marrow cellularity was increased only in replicate
2 at 30 days of exposure.  Occasional changes were also seen in red blood cell parameters
and in serum chemistry values. The authors drew no definite conclusions, but inferred that the
field had potentially a low level of toxicity.
    A study was conducted in which small numbers of ECR-SW strain mice were exposed to a
240-kV/m, 60-Hz field 18 to 22 hours/day, 7 days/week for a total of 4500 hours of exposure
 (about 32 weeks) before they were killed for tests (Fam,  1980).  No statistical differences were
 observed in the blood count values  of the males at the a= 0.05 level. However, exposed
 females had lower white blood cell counts, lower hemoglobin values, lower mean corpuscular
 hemoglobin values, and a higher percentages of bands, all of which were statistically
 significant The investigators indicated that the effects on the females could have been due to
 either field exposure or decreased food and water consumption.
     Morris and Phillips (1982) observed no effect on the  primary antibody response to keyhole
 limpet  hemocyanin in mice following their exposure to 60-Hz, 0.15- to 0.25-kV/m fields, for 30
        10/22/90
5-80

-------
                             DRAFT-DO NOT QUOTE OR CITE

 or 60 days, or on the mitogen stimulation response of spleen cells to B- and T-cell mitogens in
 mice exposed to the same fields for 90 or 150 days.
    An in vitro, field strength-related suppression of murine T-lymphocyte-mediated cytotoxicity
 was observed following a 48-hour exposure to 60-Hz sinusoidal electric fields at field strengths
 between 0.01 and 0.1 V/m, with a threshold for significance that lay between 0.01 and 0.1 V/m
 (Lyle et al., 1988). The inhibition was preferentially expressed during the early recognition
 phase of the immune response.  The field had no effect on the proliferation of the CTLL-1
 effector cells in the presence of interleukin-2, an indication that the inhibition of cytotoxicity
 was not due to the inhibition of cell proliferation but rather to an alteration of the mechanism
 for cytotoxicity itself.
    In conclusion, the effects of 60 Hz electric fields on immune function in vivo are small and
 inconsistent when present at all.  This is the overall conclusion from one human study and two
 studies each in rats and mice at field strengths ranging from 0.15 to 240 kV/m. However, in
 one in vitro study of a murine cytotoxic T-cell line, a 48-hour exposure to 60-Hz electric fields of
 only 0.1 and 1.0 V/m suppressed the cytotoxicity  of these T-cells; no effect was seen at 0.01
 V/m.
    T-cell cytotoxicity, important in the elimination of certain infectious agents, allograft
 rejection, and tumor immunity (reviewed in Lyle et al.,  1988), can be quantified in vitro with the
 4-hour chromium release assay.  The results of these assays have been shown to correlate
 directly with in vivo anti-tumor activity.  Immunotherapy trials have shown that when the
 T-lymphocyte  growth factor, interleukin-1, is administered with cultured cytotoxic
 T-lymphocytes, cures of cancer in animals and remissions in selected patients can occur
 (reviewed in Lyle et al., 1988). The suppression of T-lymphocyte-mediated cytotoxicity by ELF
 has been observed in vitro only, using only one clone  of cytotoxic murine rather than human
 lymphocytes.  These results, though limited, suggest the possibility that weak EM fields, in
 suppressing the effectiveness of cytotoxic lymphocytes, could provide a growing clone of
 antigenic tumor cells to develop the mass needed to overcome a continuing immune attack.

5.9.4.2. Modulated and Unmodulated Radiofrequency Fields
   Studies of the effects of modulated and unmodulated RF fields on immune function have
shown responses only at intensities large enough  to cause appreciable heating, with the five
exceptions noted below. This conclusion is based on  studies in virus-infected human
                                         5-81
                                                                               10/22/90

-------
                           DRAFT-DO NOT QUOTE OR CITE

mononuclear leukocytes, Wistar-Furth rats, New Zealand rabbits, rats of an unknown strain,
and monkeys of an unknown strain. The exceptions are:

    •  Lyle et al. (1988) showed that exposure to 60-Hz modulated 450-MHz RF fields inhibits
       the cytotoxicity of the same mouse T-cell line in which they found to be susceptible to
       60-Hz electric fields.
    •  Ottenbreit et al. (1981) found that the colony-forming properties of neutrophil precursor
       cells from childhood  leukemia patients in remission are inhibited by a 15-minute
       exposure to 2450-MHz unmodulated radiation at the higher power levels of 500 and
       1000mW/cm2.
    •  Liburdy and Wyant (1984) observed a change in the liquid gel elution patterns of
       human immunoglobins IgM, IgA, and IgG in solution after exposure to 10-MHz RF
       radiation (SAR = 0.13 W/kg). These investigators also measured changes in the
       separation patterns of lymphocyte-immunoglobin complexes in an  immunoaffinity cell
       chromatography assay; these changes were induced by RF radiation of 2500 Hz and a
       SAR of 0.12 W/kg. This study shows that the RF radiation alters the structure of
       membrane-bound proteins on the lymphocyte surface.
    •  Szmigielski (1975) showed that RF radiation decreased the viability of rabbit
       granulocytes.
    •  Lin et al. (1979) showed that RF radiation reduced the numbers of granulocyte and
       macrophage colony-forming units from preparations of mouse bone marrow.
 5.10.  CENTRAL NERVOUS SYSTEM EFFECTS
    This section examines some of the evidence that the central nervous system (CNS) is a
 target for ELF and RF interactions. Reports have been published showing that alterations in
 cellular morphology of brain tissue, changes in the electroencephalogram, and changes in
 pineal gland melatonin and  electrical activity occur in response to magnetic and electric fields.
 No attempt is made here to review the literature on circadian rhythms in behavioral activity,
 although the CNS is clearly involved in coordinating circadian variations in hormone levels.

 5.10.1.  Extremely Low Frequency Fields
    Welker et al. (1983) measured the melatonin and N-acetyltransferase content of the pineal
 gland in Sprague-Dawley rats before and after changes in the orientation of static magnetic
 fields. The strength of the earth's magnetic field was 0.62 Oe (0.062 mT) at a 63° inclination
 with respect to the horizontal plane. The animals were placed into cages with Helmholtz coils
 oriented so the static field experienced by the animals could be changed to various strengths.
        10/22/90
5-82

-------
                             DRAFT-DO NOT QUOTE OR CITE
  The investigators found that when the field was either inverted (-63°) or changed by small
  amounts (-5°, +5°, +15°), the pineal content of both substances decreased significantly. This
  occurred when the field was perturbed at night (when the pineal gland is active), but not
  during the day. The same group of investigators had shown earlier (Semm et al., 1980) that
  the firing rate of single pineal cells decreases gradually over periods of about 72 minutes when
  the vertical component of the field is increased by 0.5 Oe (0.05 mT).
     In a later paper from the same laboratory, Olcese et al. (1 985) showed that this response
  disappeared when the optic nerve was cut, indicating that the retina of the  eye is the
  magnetoreceptor. Later, they showed (Reuss and Olcese, 1986) that the response does not
  occur when the optic nerve is intact in total darkness and that dim red light must be present for
 the magnetic field stimulus to be effective.
    The inhibition of melatonin synthesis by a change of only 5( in the orientation of a static
 magnetic field of 0.062 mT = 62 ^T is  an interesting finding.  It is equivalent to the finding that
 the introduction of a small component either  perpendicular to the field of 62 sin (5) = 4.5 ju,T,
 or a change in the parallel component of 62 [1-cos (5)] = 0.24/^T is sufficient to inhibit
 melatonin synthesis. Ambient residential 60-Hz fields are within this order of magnitude.
    The effects of long-term exposure to ELF  fields on histology of nerve tissues of rabbits,
 pigs, rats, and mice were examined by light and electron microscopy. (Hansson 1981 a, b;
 Hansson etal.,1987).
    Rabbits were exposed under two different conditions:  (1) from conception to 6 months of
 age, outdoors,  in a 400-kV substation;  they were exposed continuously to a 50-Hz electric
 field, at approximately 14 kV/m; and (2) from conception to 8 months of age, in a laboratory
 with a controlled environment; 14 kV/m at 50  Hz, 23 hours/day, 7 days/week (Hansson, 1981
 a, b). In both situations, controls were  shielded from the field (sham-exposed) in Faraday
 cages or were unexposed.
    The most severe effects were observed  in the rabbits that were exposed outdoors  in a
 substation. Cerebellar tissues from exposed rabbits exhibited changes in almost all Purkinje
 nerve cells examined.  Nissl granules formed by clusters of granular  endoplasmic reticulum
 arranged in parallel that were seen in control animals had almost disappeared  from tissues or
exposed animals and were replaced by lamellar bodies. These structures were in continuity
with the endoplasmic reticulum, thus reducing the surface area of endoplasmic reticulum
                                         5-83
                                                                              10/22/90

-------
                           DRAFT-DO NOT QUOTE OR CITE

exposed to the cytoplasmic matrix; the changes extended to the dendrites.  The hypolemmal
cisterns normally studded along with the Purkinje cell dendrites, had vanished. In addition,
microtubules showed altered distribution and reduction in number.  Filaments and
membranous structures were observed in increased frequency.
   In rabbits that received 4 or more weeks of exposure postnatally, there were changes in
the glial cells of the cerebellum (and similar, but less definite ones, in the hippocampus) that
included replacement of normally slender processes with shorter, thicker more irregular ones;
the nuclei of the glial cells could be stained with S-100 antibodies, which was not observed in
the controls. (S-100 is a ligal cell  marker protein that can be demonstrated  ultrastructurally
with antibodies against S-100 using quantitative immunohistochemical methods.) The number
of large astrocytes in the granular layer was increased. Concentrations of S-100 in the
cerebellum and hippocampus were increased.  Changes somewhat similar  to these, but less
severe, were observed in the rabbits exposed under laboratory conditions,  rats exposed to
fields of 50 or 60 Hz, 14 kV/m for  22-23 hours/day, in mice exposed to fields (60 Hz,  10 kV/m)
during the first 4 weeks of age, and in minipigs exposed for almost a year and a half to electric
fields (60 Hz, 30 kV/m, 20 hours/day) (Hansson et al., 1987). The investigators concluded that
long-term exposure to power-frequency electric fields induces effects on the nervous system
of exposed animals, but they did  not discuss the physiological significance of the changes.
    The work of Hansson (1981 a, b) has been criticized on the basis that the animals were
 maintained out-of-doors under ill-controlled conditions and environmental factors other than
 the field [such as noise, ozone production, and vibrations that were related to high voltage
 installations (Kornberg, 1976; Michaeison, 1979, both cited in Portet and Cabanes, 1988)
 could have influenced the results (Portet and Cabenes, 1988).  In addition, a reduction  of
 growth observed in the rabbits exposed outdoors was not reported for animals exposed in the
 laboratory. In similar studies, Portet et al. (1984, cited in Portet and Cabanes, 1988) did not
 observe lesions in the cerebella of neonate rabbits exposed to electric fields. Experimental
 details were not available.

 5.10.2. Modulated Radiofrequency Fields
     Electroencephalogram (EEG) changes (which consisted of enhanced low frequency
 components and decreased high frequency activities) were observed in rabbits exposed to a
 RF field 2 hours/day for 6 weeks at 1.2 MHz (15-Hz modulation) at levels of 0.5 to 1 kV/m
        10/22/90
5-84

-------
                             DRAFT-DO NOT QUOTE OR CITE

  (~1.3-2.6 W/cm2) (Takashima et al., 1979). The effects were described as nonthermal,
  because the estimated current density of 0.082 mA/cm2 was below that needed to cause
  noticeable thermal effects.  The EEG signals were computer processed to obtain power
  spectra.

  5.10.3. Unmodulated Radiofrequency Fields
    Webber et al. (1980) observed ultrastructural damage, apparently of a nonthermal nature
  (SAR not known), in mouse neuroblastoma cells grown in culture and exposed to microwave
  pulses. The cells were exposed to short pulses at high fields [1.7 to  3.9 kV/cm (170 to 390
  kV/m)] from a magnetron radar transmitter with a radar modulator. The voltage pulses from
 the power supply and modulating system were converted to 1-second RF pulses in a 2.7-Hz
 band (330 pulses/second, 0.0335 duty cycle). The waveguide apparatus  termination was
 constructed for insertion of a glass slide horizontally between two section of S-band
 waveguide.  The coverslip carrying the cells was placed on the glass slide, with cells facing
 upward. A section of waveguide that terminated in a short circuit was mounted 4.16 cm above
 the slide.
    The cells were (1) exposed to 1.7 kV/cm (170 kV/m) for 30 seconds, (2) exposed to 3.0
 kV/cm (300 kV/m) for 60 seconds,  (3) exposed to 3.9 kV/cm (390 kV/m) for 60 seconds, or (4)
 not exposed to the field.  To test the effects of heat on the cells, coverslips carrying the cells
 were dipped into saline at the following temperatures for 30 seconds: 37,  41, 45, 50, 62, 70,
 and 80° C. The cells were examined by electron microscopy after field or  heat exposure.
    The most striking damage observed in cells exposed to 1.7 kV/cm (170 kV/m) was in the
 form of breaks in the cell and mitochondria! membranes. The cristae  lost their normal pattern
 and formed myelinated figures inside the mitochondria. Parts of the cell membrane were
 expelled and appeared as membrane-bound sacs outside the cell surface.  Exposure of the
 cells to this field for 30 seconds is expected to result in a temperature rise from 37° to 41° C.
To determine if the effect was heat-related, cells were exposed to temperatures corresponding
to temperature increases induced by microwave exposure. Cells exposed to heat alone (41° C
as well as 45° C) remained viable and maintained a normal structure without any cell damage.
Cells exposed to 50° C showed considerable damage;  however, the cell membranes remained
intact. Cells exposed to 3.9 kV/cm (390 kV/m) for 60 seconds were damaged severely. The
cell  content was totally disorganized and the membranes had completely broken down,
                                        5-85
                                                                            10/22/90

-------
                           DRAFT-DO NOT QUOTE OR CITE

appearing denatured. However, cells heated to 62° C, the corresponding temperature for this
exposure, exhibited greater integrity of the cell membranes (including the mitochondrial and
nuclear membranes). Because of the different nature of the damage to field-exposed
compared with temperature-exposed cells, the investigators suggested that the damage could
have been due to nonthermal effects.

5.10.4. Summary
    Long-term exposure of animals to moderate or high-intensity EM fields at 50 or 60 Hz
resulted in changes in the cerebellar Purkinje nerve cells that included rearrangement of the
endoplasmic reticulum and disappearance of the hypolemmal cisterns of the dendrites
(Hansson et al., 1987). Glial reactions that showed an increased concentration of S-100 in the
cerebellar hemispheres were the most consistent findings. These changes indicated that
disturbances had taken place in the interaction between plasma membrane structures and the
cytoskeletons of cells of the nervous system.
    Neuroblastoma cells exposed to microwave pulses in culture exhibited ultrastructural
damage  as evidenced by breaks in the cell and mitochondrial membranes (Webber et al.,
1980). The effects were apparently nonthermal.
    The studies of Semm et al. (1980), Olcese et al. (1985), and their colleagues show that
changes in magnetic fields are perceived by the retina. This stimulus decreases the firing rate
of neurons in the pineal gland and inhibits its melatonin content if applied at night when the
pineal is actively secreting melatonin. The finding that the retina is able to detect changes in
magnetic fields provides a mechanism whereby the CMS function is affected by these fields.
    Alterations in Ca++ efflux from nervous tissue have been described (see Section 5.3.2.1).
 Blackwell and Saunders (1986) reviewed the literature on CNS effects of RF and microwave
 exposure and concluded that although calcium ions play a critical role in many metabolic and
 physiological processes, the significance of changes in calcium ion exchange in brain tissue
 for the health and safety of people exposed to microwave and RF radiation is difficult to
 determine, and that furthermore, the evidence that calcium ion exchange in living nervous
 tissues is altered by amplitude-modulated RF and microwave and radiation is inconclusive
 (Blackwell and Saunders 1986).
    Michaelson and Lin (1987) also reviewed the effects of low-intensity microwaves on the
 CNS and concluded that to date there is no convincing evidence of the existence of
        10/22/90
5-86

-------
                            DRAFT-DO NOT QUOTE OR CITE

 low-intensity microwave effects on the human CNS. The animal studies that are the basis for
 reported effects suggest that the mechanisms for these effects involve microwave-induced
 nonuniform temperature distributions and/or thermal gradients (Michaelson and Lin, 1987).

 5.11.  SUMMARY AND CONCLUSIONS FOR SUPPORTING EVIDENCE OF
    CARCINOGENICITY
    The literature on effects of EM fields on biological processes relevant to carcinogenesis
 has been reviewed in this chapter. In this section, the summaries of other sections of Chapter
 5 are reviewed and overall conclusions are derived to the extent possible.

 5.11.1. Summary
    Table 5-7 summarizes in brief phrases the effects that have been observed for each
 biological process and each of the major categories of EM fields. These summary phrases
 have been derived from the section summaries and  text. The first observation apparent from
 Table 5-7 is that the evidence for any one process is widely scattered among different types of
 exposure. The scattering is even  more widespread than the table suggests, since it obviously
 cannot show the large variety of frequencies, intensities, and durations of exposure that have
 been used within each broad class of exposure.  The table demonstrates the gaps in the
current information; for example, transcription of genetic information into messenger RNA,
translation into proteins, and parathyroid hormone effects on bone cells have been studied
only with low-frequency pulsed magnetic fields.  This is understandable, since this work has its
origins in the successful use of pulsed magnetic fields for clinical healing of recalcitrant bone
fractures.
   The following conclusions can be made:

   •   None of the types of time-varying fields considered in this document cause DNA
       breaks, gene mutations, or sister chromatid exchanges.  (Static magnetic fields with
       high field strength have affected DNA in solution and have caused sex-linked recessive
       lethal mutations in Drosophiia, but the significance of this effect is not known). This
       lack of a DNA and gene mutation effect is expected, since these fields do not have
       enough energy to break chemical bonds.
   •   Effects on DNA synthesis have not been studied extensively enough to draw definite
      conclusions. Apparently RF fields cause inhibition and ELF magnetic fields cause an
      enhancement of DNA synthesis only within a limited range of frequency and intensity
       windows."
                                        5-87
                                                                            10/22/90

-------
                 DRAFT-DO NOT QUOTE OR CITE
  TABLE 5-7. SUMMARY OF SUPPORTING EVIDENCE FOR CARCINOGENICITY
Subject
DMA Damage:
Breaks
Repair
DMA Synthesis

Gene Mutations
Chromosome
Aberrations



Sister Chromatid
Exchange
Mitosis


Transcription
of Genes


Translation

Cell Transforma-
tion
Calcium Efflux,
Brain Tissue

Parathyroid
Hormone

Intracellular
Enzymes

Hormones




E-Relds
No breaks
No DNA repair
Delayed S phase

No effect
(1 study)
Breaks, aneuploidy
(inconsistent)



No effect
(4 studies)
Mitotic index
reduced early, but
recovers. Cell
cycle delayed
-


_

-
Inhibition Frequency
and Amplitude
"Windows"

-

ODC is induced.
Same effect
as TPA.

Inhibits night-
time melatonin
output of pineal
gland
ELF Fields
B-Relds
No breaks
—
Enhancement
(frequency and
intensity windows)
No effect
(2 studies)
Breaks, aneuploidy,
decondensation
(inconsistent)



No effect
(3 studies)
-


Enhanced, both
normal and newly-
induced sites (all
waveforms)
Altered pattern of
protein synthesis
—
-

Blocks action
of PTHatthe
cell membrane

Combined
Eand B
No breaks
~
-

-
Breaks



No effect
(2 studies)
Cell cycle
delayed


—


-

—
Enhancement
Amplitude and
frequency
"windows"
-


Modulated
RF Fields


Inhibition

No effect
(1 study)
Spark dis-
charges caused
breaks



No effect
(2 studies)
Inconsistent
(2 studies)


—


-

Initiation of
C3H/10T1/2
cells
Frequency
and intensity
"windows"

-

RF Fields
No damage

Inhibition

No effect
(10 studies)
Breaks,
uncoiling,
numerical
aberrations,
dicentrics
condensation
(inconsistent)
No effect
(1 study)
No effect
(1 study)


—


-


No effect
(3 studies)

—

_ — Protein kinase —
C inhibited.
Frequency "windows."

_




—



ODC is induced
-




-



10/22/90
5-88

-------
                             DRAFT-DO NOT QUOTE OR CITE
       TABLE 5-7. SUMMARY OF SUPPORTING EVIDENCE FOR CARCINOGENICITY (continued)
Subject
Growth and
Differentiation



Immunological
Systems


E-Relds
Inhibits protein
synthesis in
fibroblasts


No effect in vivo
(3 studies)
Inhibits cyto-
toxicity of T
lymphocytes in
culture.
H..FRBM*
B-Relds
Inhibits differ-
entiation and
stimulates growth
of embryonal
carcinoma cells
—


Combined
Eand B
Increases
reproductive
capacity of
carcinoma cells

No effect in
humans
(1 study)

Modulated
RF Reids





Cytotoxicity
same as ELF.
No effect on
leukocytes.
RF Reids





No effect
in rats, rabbits,
monkeys. Celi-
immuno-
globulin
binding is
tUAVA^J
Central Nervous
System
Inhibition of
pineal activity
via retinal
magneto-
receptors
Morpho-
logical
changes in
glial and
Purkinge
cells of
cerebellum.
EEG shifts to
lower frequencies
(ELF modulation).
Disruption of mito-
chondria and cell
membranes, different
than heat damage
(radar modulation).
       Chromosomal aberrations is a frequent finding for both RF and ELF fields, but it often
       does not occur. In one measurement of aberrations in peripheral lymphocytes in
       electrical switchyard workers, chromosome breaks occurred immediately after they
       were exposed to spark discharges, but in similar populations, with no spark discharge
       exposure, no aberrations occurred. This indicates that high frequencies may be more
       effective than low frequencies in causing aberrations, but these conclusions are only
       tentative, and specific studies are needed to address this important issue.

       ELF electric and combined electric and magnetic fields have delayed the cell cycle and
       caused transient reductions  in the rate of cell division, but RF fields have caused no
      consistent effect. ELF magnetic fields have not been tested for their effect on the
      mitotic cycle, and little testing of RF has been done.

      Pulsed magnetic fields of the type used for clinical bone healing have enhanced the
      transcription synthesis of mRNA at genetic sites that are normally active and have
      altered the molecular weight distribution of proteins synthesized with the fields present.
      The protein molecular weight distribution is different with different waveforms (pulsed
      versus sinusoidal) and frequencies (60 Hz vs. 72 Hz).  No other type of field has been
      investigated for this effect.

      Cultures of the NIH C3H/IOTI/2 cells, which are widely studied systems for investigating
      cell transformation from normal to malignant patterns of growth, have been shown to
      undergo transformation under special conditions. Microwave 2450-MHz power,
      modulated with pulses at a rate of 120 pulses per second, were administered to the
      culture at intensities low enough to cause no effect. Subsequently the cells were
                                         5-89
                                                                               10/22/90

-------
                        DRAFT-DO NOT QUOTE OR CITE
   treated with TPA, the phorbol ester which is known to cause promotion of
   chemically-initiated cells. The cells undergo transformation under these conditions,
   indicating that this pulsed microwave power has the properties of a chemical initiator of
   malignant transformation. Only one  experiment has shown this phenomenon, so no
   conclusions can be made about whether ELF or RF fields could produce the same
   effects.

•  The release of calcium ions from chick brain tissue into the medium surrounding the
   tissue has been affected by ELF electric fields, crossed ELF electric and magnetic fields
   and RF radiation modulated at the same frequencies which cause the ELF effects. The
   original investigations with ELF electric fields showed an inhibition of calcium release,
   whereas an enhancement of release occurred in the other experiments. This
   phenomenon has not occurred in three RF studies, but magnetic ELF fields have not
   been examined for this'effect. The conditions under which this occurs are very precise
   and not understood.  It occurs only at certain frequencies  (e.g., odd multiples of 15 Hz
   with some frequencies missing) with no response at frequencies between. At some
   fixed frequencies, it occurs only at certain field strengths with no response below and
   above this intensity "window." In at least at one of the "windows" the orientation of the
   alternating magnetic field must have a component perpendicular to the static earth's
   magnetic field.  The crossed electric and magnetic alternating fields producing this
   effect are extremely small (16 V/m electric and 73 nT magnetic field), and in the range
   of ambient magnetic ELF fields in residences. With the two exceptions discussed
   below all of the other effects reviewed in this document are induced by fields at least
   hundreds of times higher than this.  The biological significance of this phenomenon is
   not clear beyond the fact that brain tissue is somehow affected by these unique field
   conditions.

•  The effects of pulsed ELF magnetic  fields on the interaction between parathyroid
   hormone (PTH) and bone cells have been studied to elucidate the mechanisms of the
   clinically successful bone healing ability of these fields. These fields block the
   inhibitory effect of PTH on collagen  synthesis by the cells, and the action of the field
   occurs at the plasma membrane where the hormone binds with its membrane receptor.
   The significance of this is that pulsed magnetic ELF fields can alter the chemical
   signalling  process between  an exogenous hormone and the cellular activity induced by
   the hormone. No other type of field has been tested for this effect.

•  The intracellular enzyme ornithine decarboxylase, which is active during cell
   proliferation and DNA synthesis of most cells, is induced by ELF electric fields and by
   ELF modulated RF fields in three different cell lines. Information on the effect of other
   fields has not been found. The same enzyme is induced by the phorbol ester TPA, the
   most actively studied chemical promoting  agent.  cAMP-independent protein kinase,
   one of the chemical intermediates involved in this cell proliferation response, is
   inhibited in human lymphocytes by  modulated RF fields, with an apparent frequency
   "window."

•  Electric ELF field exposure to rats for 20 hours per day for 30 days causes an inhibition
   of the nocturnal synthesis of melatonin by the pineal gland.  Information on the effect of
   other fields has not been found.  This finding could have great significance in
   explaining the potential carcinogenicity of ELF fields, since there is a wealth of literature
   describing the oncostatic properties of melatonin, not only for chemically-induced
    10/22/90
5-90

-------
                            DRAFT-DO NOT QUOTE OR CITE
        mammary tumors in rats, but also (according to one report) in the treatment of human
        leukemia. Other experiments in rats have shown that pineal neurological activity and
        melatonin synthetic activity have been inhibited by changing the orientation of the static
        magnetic field through an angle of as small as 5(, a change which is only a factor of 10
        higher than ambient magnetic residential fields. This implies that exposure of humans
        to weak time-varying magnetic fields at night could have an inhibitory effect on pineal
        melatonin synthesis.  Experimental evidence to support this hypothesis is not available
        Information on the effect of ELF magnetic fields and RF fields on melatonin synthesis
        has not been found.  ELF electric, magnetic, and combined fields of the strength used
        to stimulate bone repair can also cause alterations in biosynthesis. Inhibition of protein
        biosynthesis in fibroblasts, inhibition of differentiation and stimulation of growth  of
        embryonal carcinoma cells, and increase in the reproductive capacity of colon
        carcinoma cells are all phenomenon that are induced by ELF fields and are
        characteristics of malignant growth. The extent to which they occur in the whole
        organism under realistic exposure conditions is not known, but these phenomena are
        consistent with the suggestion of carcinogenic effects in humans and animals.

    •   In one human study and three animal studies, exposure to ELF fields caused small but
        inconsistent changes in white blood cells. Exposure of rats, rabbits, and monkeys to
        unmodulated RF fields also caused small inconsistent changes. However, one
        investigator working with T-lymphocytes in culture found that both ELF electric fields
        and modulated RF radiation inhibit their ability to kill their normal target cells.  The
        reasons for this difference between the in vivo and in vitro response is not known, but it
        does call into question how directly one can infer that whole animal responses can be
        predicted from cell culture experiments for these effects. Information on the effect of
        magnetic fields on lymphocyte function or dynamics has not been found.

    •   Exposure to combined ELF electric and magnetic fields 23 hours/day for the first 6 to 8
        months of life caused a disappearance  of Missel granules and disruption of the
       endoplasmic reticulum in Purkinj'e cells  and a disruption of the morphology  of glial cells
       in the cerebellum. This occurred most severely in rabbits but also occurs to a lesser
       extent in rats, mice, and mini pigs exposed for various durations. Six-week exposures
       of rabbits to ELF modulated RF fields caused a downward shift in the frequency  of the
       electroencephalogram (EEG). The ELF modulation frequency of 15 Hz was in the
       same range as the EEG frequencies that were enhanced. The finding that the intact
       retina is needed for a functioning pineal gland response to magnetic fields implies that
       the  CNS can be a sensor for ELF fields and raises the possibility that other
       neuroendocrine functions of the CNS could be affected.

5.11.2.  Conclusions

   The finding that several biological phenomena, which are in some way related to

postulated  mechanisms of carcinogenesis, are  induced by time-varying  electric and magnetic

fields is far  from proof that these fields are carcinogenic by themselves or that exposure to

them are risk factors for humans.  There are reasons for both questioning and affirming the

relevance of each finding. One of the primary difficulties in accepting the relevance of most of
                                         5-91
                                                                              10/22/90

-------
                           DRAFT-DO NOT QUOTE OR CITE

these findings is the fact that they are induced by field strengths many times higher than the
ambient residential exposures which are hypothetically causing human cancer. Of the effects
summarized in this chapter, only three phenomena occurred under conditions similar to the
low ELF electric and magnetic fields characteristic of ambient exposure [10 V/m and 2
milligauss (mG) = 0.2^T]: (1) calcium efflux from chick brain tissue induced by crossed
electric and magnetic fields; (2) calcium efflux from chick brain tissue after exposure of the
developing embryo to electric fields; and (3) inhibition of nocturnal melatonin synthesis by
small changes in the orientation of static magnetic fields.
   The above statements are made under the assumption that human carcinogenicity is
indeed caused by 60-Hz fields with a field strength on the order of 2 mG. If the causative
agent really is the internal currents induced by these ambient fields, then it is conceivable that
the higher frequency components always accompanying ambient 60-Hz fields are the relevant
aspects of exposure. If the effects are really caused by high peak fields with high frequency
components, then the phenomena observed at higher field strengths would be relevant. In
this case more of these phenomena would be relevant, but precise quantitative evaluation is
difficult to carry out, given the current degree of knowledge.
        10/22/90
5-92

-------
                             DRAFT-DO NOT QUOTE OR CITE
  REFERENCES FOR CHAPTER 5
AdSy>

                                     electroma9netic environment and cancer promotion.
 Adey, W.R.; Bawin, S.M.; Lawrence, A.F. (1982) Effects of weak amplitude-modulated
        microwave fields on calcium efflux from awake cat cerebral cortex
        Bioelectromagnetics 3:295-307.

 Akamlne, T.; Muramatsu, H.; Hamada, H.; Sakou T. (1985) Effects of pulsed electromagnetic
        field on growth and differentiation of embryonal carcinoma cells. J. Cellular Physiol.
        I Łr\ť Łr\  ~
 Anderstam, B.; Hamnerius, Y.; Hussain, S; Ehrenberg, L (1983) Studies of possible genetic
        effects in bacteria of high frequency electromagnetic fields. Hereditas 98: 1 1 -32.

 Averbeck, D.; Dardalhon, M; Berteaud, A.-J. (1976) Microwaves action in procaryotic and
        eucaryotic cells and a possible interaction with x-rays.  J. Microw. Power 1 1 (2):143-144.

 Balcer-Kubiczek, E.K.; Harrison, G.H. (1985) Evidence for microwave carcinogenesis in vitro
        Carcmogenesis 6(6):859-864.

 Balcer-Kubiczek EK.; Harrison, G.H. (1989) Induction of neoplastic transformation in
        udH/i on/2 cells by 2.45 GHz microwaves and phorbol ester. Rad. Res. 1 1 7:531 -537.

 Baranski, S.; Debiec, H.; Kwarecki, K.; Mezykowski, T. (1976) Influence of microwaves on
        genetical processes of Aspergillus nidulans. J. Microw. Power 11 (2): 146-1 47.

                               Abn°rmalities in °rgans of mice induced bV * magnetic ^Id.
 Bartsch, C.; Bartsch, H.; Fluechter, S.; et al. (1 985) Evidence for modulation of melatonin
                                                     °f the pr°state-
Bassett, C.A.L. (1987) Low energy pulsing electromagnetic fields modify biomedical
       processes.  Bioessays 6:36-42.  (Cited in Yen-Patton et al., 1988.)

Bauchinger, M.; Hauf, R.; Schmid, E.; Dresp, J. (1981) Analysis of structural chromosome
       f°r,lan?nn Sw    >E after occuPational long-term exposure to electric and magnetic fields
       from 380 kV-systems. Radiat. Environ. Biophys. 19:235-238.

Baum, JW; Nauman, C.H. (1984) Influence of strong magnetic fields on genetic endpoints in
       Tradescantia tetrads and stamen hairs.  Environ. Mutagen. 6:49-58.

Bawin, S.M.; Adey, W.R. (1976) Sensitivity of calcium binding in cerebral tissue to weak
       environmental electric fields oscillating at low frequency.  Cell Biol. 73:1999-2003.
                                        5-93
                                                                             10/22/90

-------
                           DRAFT-DO NOT QUOTE OR CITE
Bawin, S.M.; Kaczmarek, L.K.; Adey, W.R. (1975) Effects of modulated VHP fields on the
       central nervous system. Ann. N.Y. Acad. Sci. 247:74-81.

Bawin, S.M.; Sheppard, A.; Adey, W.R. (1978) Possible mechanisms of weak electromagnetic
       field coupling in brain tissue.  Bioelectrochem. Bioenerget. 5:67-76.

Becker, R.O.; Esper, C. (1981) Electrostimulation and undetected malignant tumors. Clin.
       Orthop. Rel. Res. 161:336-339.

Beechey, C.V.; Brooker, D.; Kowalczuk, C.I.; Saunders, R.D.; Searle, A.G. (1986) Cytogenetic
       effects of microwave irradiation on male germ cells of the mouse.  Int. J. Radiat. Biol.
       50:909-918.

Benovic, J.L; DeBlasi, A.; Stone, W.C.; Caron, M.G.; Lebkowitz, RJ. (1989) B-androgenic
       receptor kinase: primary structure delineates a multigene family. Science 246:235-240.

Benz, R.D.; Carsten, A.L (1986) Effects on dominant lethal mutations, multigeneration
       prosperity and bone marrow sister chromatid exchanges and cell cycle time by 60 Hz,
       50 kV/m -10 gauss electric/magnetic fields in two strains of mice. Environ. Mutagen.
       8(suppl6):10-11.

Blackman, C.R. (1987)  Effects of modulated RFR.  In: Elder, J., ed. A reassessment of the
       biological effects of radiofrequency  radiation: non-cancer effects. Research Triangle
       Park, NC: Health Effects Research Laboratory, U.S. EPA.

Blackman, C.F.; Surles, M.C.; Benane, S.G. (1976) The effect of microwave exposure on
       bacteria: mutation induction. In: Johnson,  C.C.; Shore, M.L., eds., Biological effects of
       electromagnetic waves, Vol. I. Food and Drug Administration, Rockville, MD:  HEW
       Publication (FDA)77-8010, pp. 406-413.

 Blackman, C.F.; Benane, S.G.; Elder, J.A.;  House, D.E.; Lampe, J.A.; Faulk, J.M. (1980)
        Induction of calcium-ion efflux from  brain tissue by radiofrequency radiation:  efl
        sample number and modulation frequency on the power-density window.
        Bioelectromagnetics 1:35-43.

 Blackman, C.F.; Benane, S.G.; Kinney, L.S.; Joines, W.T.; House, D.E. (1982) Effects of  ELF
        fields on calcium-ion efflux from brain tissue in vitro. Radiat. Res. 92:510-520.

 Blackman, C.F.; Benane, S.G.; Rabinowitz, J.R.; House, D.E.; Joines, W.T. (1985a) A role for
        the magnetic field  in the radiation-induced efflux of calcium ions from brain tissue in
        vitro. Bioelectromagnetics 6:327-337.

 Blackman, C.F.; Benane, S.G.; House, D.E.; Joines, W.T. (1985b) Effects of ELF (1-120  Hz and
        modulated (50 Hz) RF fields on the efflux of calcium ions from brain tissue in vitro.
        Bioelectromagnetics 6:1 -11.

 Blackman, C.F.; House, D.E.;  Benane, S.G.; Joines, W.T.; Spiegel, RJ.  (1988a) Effect of
        ambient levels of power-line-frequency electric fields on  a developing vertebrate.
        Bioelectromagnetics 9:129-140.
                                    /
                                  effect of
        10/22/90
5-94

-------
                             DRAFT-DO NOT QUOTE OR CITE
 Blackman, C.F.; Benane, S.G.; Elliott, D.J.; House, D.E.; Pollock, M.M. (1988b) Influence of
        electromagnetic fields on the efflux of calcium ions from brain tissue in vitro: a
        three-model analysis consistent with the frequency response up to 510 Hz
        Bioelectromagnetics 9:215-227.

 Blackman, C.F.; Kinney, L.S.; House D.E.; Jones, W.T. (1989) Multiple power-density windows
        and their possible origin. Bioelectromagnetics 10:115-128.

 Blackwell, R.P.; Saunders, R.D. (1986) The effects of low-level radiofrequency and microwave
        radiation on brain tissue and animal behaviour. Int. J. Radiat.  Biol. 50:761 -787.

 Blask, D.;  Hill, S. (1986) Effects of melatonin on cancer: studies on MCF-7 human breast
        cancer cells in culture.  J. Neural Transm.  (Suppl.) 21:433-449. (Cited in Wilson et al
        1988.)

 Blevins, R.D.; Crenshaw, R.C.; Hougland, A.E.; Clark, C.E. (1980) The effects of microwave
        radiation and heat on specific mutants of Salmonella typhimurium LT2.  Radiat. Res. 82-
        511-517.

 Bloom, W.; Fawcett, D.W. (1969) A textbook of histology, 9th ed.  Philadelphia, PA:  W B
        Saunders Co., pp. 134-135..

 Brulfert, A.; Miller, M. W.; Robertson, D.; Dooley, D. A.; Economou, P.  (1985) A cytohistological
        analysis of roots whose growth is affected  by a 60-Hz electric field.
        Bioelectromagnetics 6:283-291.

 Byus, C.V.; Lundak, R.L.; Fletcher, R.M.; Adey, W.R. (1984) Alterations in protein kinase activity
       following exposure of cultured human lymphocytes to modulated microwave fields
       Bioelectromagnetics 5:341 -351.

 Byus, C.V.; Pieper, S.E.; Adey,  W.R. (1987) The effects of low-energy 60-Hz environmental
       electromagnetic fields upon the growth-related enzyme ornithine decarboxylase.
       Carcinogenesis 8:1385-1389.

 Byus, C.V.; Dartun, K.; Pieper, S.; Adey, W.R. (1988) Increased ornithine decarboxylase activity
       in cultured cells exposed to low energy modulated microwave fields and phorbol ester
       tumor promoters.  Cancer Res. 48:4222-4226.

 Cain, C.D.; Luben, R.A. (1987) Pulsed electromagnetic field effects on  PTH-stimulated cAMP
       accumulation and bone resorption in mouse calvaria.  In: Anderson, L.E.; Kelman, B.J.;
       Weigel, R.J.,  eds. Interaction of biological systems with static and ELF electric and
       magnetic fields: proceedings of the 23rd Hanford Life Sciences Symposium; October
       2-4,1984.  Richland, WA: Pacific Northwest Laboratory, pp. 269-278.

Chang,  B.K.; Huang, A.T.; Joines, W.T. (1980)  Inhibition of DNA synthesis and enhancement
       of the uptake and action of methotrexate by low-power-density microwave radiation in
       L1210 leukemia cells.  Cancer Res. 40:1002-1005.

Chen, K.M.; Samuel, A.; Hoopingarner, R. (1974) Chromosomal aberrations of living cells
       induced by microwave radiation.  Environ. Lett. 6(1):37-46.
                                         5-95
10/22/90

-------
                           DRAFT-DO NOT QUOTE OR CITE
Chou, C.-K.; Guy, A.W.; Borneman, L.E.; Kunz, L; Kramar, P. (1983) Chronic exposure of
       rabbits to 0.5 mW/sq cm 2450 MHz microwave radiation. Bioelectromagnetics 4:63-77.

Ciaravino, V.; Meltz, M.L.; Erwin, D.N. (1987a) SCE frequencies in CHO cells exposed
       simultaneously to adriamycin and radio frequency radiation. Environ. Mutagen. 9(suppl
       8):24-25.

Ciaravino, V.; Meltz, M.L; Erwin, D.N. (1987b) Effects of radio frequency radiation and
       simultaneous exposure with  mitomycin C on the frequency of sister chromatid
       exchanges in Chinese hamster ovary cells. Environ. Mutagen. 9:393-399.

Cohen, M.; Lippman, M.; Chabner, B. (1978) Role of pineal gland in aetiology and treatment of
       breast cancer.  Lancet 2:814-816.

Cohen, M.M.; Kunska, A.; Astemborski, J.A.; McCulloch, D.; Paskewitz, D.A. (1986a) Effect of
       low level, 60-Hz electromagnetic fields on human lymphoid cells: I. mitotic rate  and
       chromosome breakage in human peripheral lymphocytes.  Bioelectromagnetics 7:
       415-423.

Cohen, M.M.; Kunska, A.; Astemborski, J.A.; McCulloch, D. (1986b) The effect of low-level
       60-Hz electromagnetic fields on human lymphoid cells: II. sister-chromatid  exchanges
       in peripheral lymphocytes and lymphoblastoid cell lines. Mutat. Res. 172:177-184.

Danforth, D.; Tamarkin, L; Chabner, B.; Demoss, E.; Lichter, A; Lippman, M. (1982) Altered
       diurnal secretory pattern of melatonin in patients with estrogen receptor positive breast
       cancer. Clin. Res. 30:532A.

Danilenko, I.I.; Mirutenko, V.I.; Kudrenko, V.I. (1974). Effect of high-frequency electromagnetic
       fields on  mutagenic rate.  Appl. Electr. Phenom. (USSR) 4:85-87.

Dardalhon, M.; Averbeck, D.; Berteaud, A.J. (1981). Studies on possible genetic effects of
       microwaves in procaryotic and eucaryotic cells. Radiat. Environ. Biophys. 20:37-51.

Dahahi, S.J.; Habash, R.W.Y.; AI-Hafid,- H.T. (1982). Lack of mutagenic effects on conidia of
       Aspergillus amstelodami irradiated by 8.7175 GHz CW microwaves.  J. Microw. Power
       17(4):345-351.

 Di Bella, L; Rossi, M.T.; Sclalera, G. (1979) Perspectives in pineal functions. Prog. Brain Res.
       52:475-478.

 Diebolt, J.R. (1978). The influence of electrostatic and magnetic fields on mutation in
       Drosophila melanogaster spermatozoa. Mutat. Res. 57:169-174.

 Druker, B.J.; Marmon, H.J.; Roberts, T.M. (1989)  Oncogenes, growth factors, and signal
       transduction.  N. Eng. J. Med. 321 (20): 1383-1391.

 Dubrov, A.P.; Ostryakov, I.A.; Turkov, V.D. (1968).  Genetic effect of static electricity.  Dokl.
       Biol. Sci. 178:29-30.
        10/22/90
5-96

-------
                             DRAFT-DO NOT QUOTE OR CITE
 Dutta, S.K; Subramoniam, A.; Ghosh, B.; Parshad, R. (1984)  Microwave radiation-induced
        calcium ion efflux from human neuroblastoma cells in culture.  Bioelectromaqnetics 5-
        71-78.

 Dutta, S.K; Ghosh, B.; Blackman, C.F. (1989) Radiofrequency radiation-induced calcium ion
        efflux enhancement from human and other neuroblastoma cells in culture
        Bioelectromagnetics 10:197-202.

 Dyshlovoi, V.D.; Panchuk, A.S.;  Kachura, V.S. (1981). Effect of electromagnetic field of
        industrial frequency on the growth pattern and mitotic activity of cultured human
        fibroblastoid cells. Cytol. Genet. 15(3):6-9.

 Eberle, P.; May, C. (1982) SCE-frequencies following exposure to homogeneous high strength
        magnetic fields.  Sister-chromatid-exchange test, Workshop Gesellshaft fpr Unwelt-
        Mutationsforschung, 1980,17-78.

 El Nahas, S.M.; Oraby, H.A. (1989). Micronuclei formation in somatic cells of mice exposed to
        50-Hz electric fields. Environ. Mol. Mutagen. 13:107-111.

 Elder, J.A. (1987a) Radiofrequency radiation activities and issues: a 1986 perspective  Health
        Phys. 53:607-612.

 Elder, J.A. (1987b) A reassessment of the biological effects of radiofrequency radiation:
       non-cancer effects. Research Triangle Park, NC: Health Effects Research Laboratory
       U.S. Environmental Protection Agency. Unpublished.

 Elder, J.A.; Cahill, D.F., eds. (1984) Biological effects of radiofrequency radiation. Washington
       DC: U.S. Environmental Protection Agency; EPA-600/8-83-026F. Available from- NTIS
       Springfield, VA; PB-85/20848.

 Fam, W.Z. (1980) Long-term biological effects of very intense 60 Hz electric field on mice
       IEEE Trans. Biomed. Eng. 27:376-381.

 Folkman, J. (1975) Tumor angiogenesis. In: Becker, F.F., ed. Cancer 3. Biology of Tumors -
       Cellular Biology and Growth.  New York, NY: Plenum Press, pp. 355-388.

 Fotopoulos, S.S.; Cook,  M.R.; Gerkovich, M.M.; Koontz, E; Graham, C.; Cohen, H.D. (1987)
       60-Hz field effects on human neuroregulatory, immunologic, hematologic and target
       organ activity. In: Anderson, L.E.; Kelman, B.J.; Weigel, R.J.,  eds. Interaction of
       biological systems with static and ELF electric and magnetic fields: proceedings of the
       23rd Hanford Life Sciences Symposium; October 2-4, 1984. Richland, WA: Pacific
       Northwest Laboratory, pp. 455-469.

Fraschini, F.; Esposti, D.; Lissoni, P. et al. (1988)  The pineal function in cancer patients. In:
       Gupta, D.; Attanasio, A.; Reiter, R.J., eds.  The pineal gland and cancer.  Tubingen
       West Germany: Brain Research Promotion, pp. 157-166.
                                         5-97
                                                                              10/22/90

-------
                           DRAFT-DO NOT QUOTE OR CITE
Frazier, M.E.; Samuel, J.E.; Kaune, W.T. (1987) Viabilities and mutation frequencies of CHO-K1
      cells following exposure to 60-Hz electric fields.  In: Anderson,  L.E.; Kelman, B.J.;
      Weigle, R.J., eds. Interactions of biological systems with static and ELF electric and
      magnetic fields: proceedings of the 23rd Hanford Life Sciences Symposium; October
      2-4,1984. Richland, WA: Pacific Northwest Laboratory, pp. 255-267.

FrOhlich, H. (1968)  Long range coherence and energy storage in biological systems.  Int. J.
      Quant. Chem. 2:641-649.

Gemishev, Ts.M. (1976) Postmagnetic growth and mitotic effect at the action of the constant
      magnetic field on seeds of Hellianthus annus L  God. Sofii. Univ. Biol. 67:149-158.

Goodman, E.M.; Greenebaum, B.; Marron, M.T.; Carrick, K. (1984)  Effects of intermittent
      electromagnetic fields on mitosis and respiration.  J. Bioelectricity 3:57-66.

Goodman, R.; Henderson, A.S. (1986) Sine waves enhance cellular transcription.
      Bioelectromagnetics 7:23-29.

Goodman, R.; Henderson, A.S. (1988)  Exposure of salivary gland cells to low-frequency
      electromagnetic fields alters polypeptide synthesis. Proc. Natl. Acad. Sci. USA 85:
      3928-3932.

Goodman, R.; Bassett, C.A.L; Henderson, A. S. (1983) Pulsing electromagnetic fields induce
       cellular transcription. Science 220:1283-1285.

Goodman, R.; Abbott, I. Henderson, A. S. (1987) Transcriptional patterns in the x chromosome
       of Sciara coprophila following exposure to magnetic fields. Bioelectromagnetics 8:1-7.

Goswami, H.K.  (1977) Changes in chromosome morphology due to magnetism.  Cytologia 42:
       639-644.

Goswami, H.K.; Dave, N. (1975) Chromosomal aberrations in  Pisum sativum by magnetic field,
       x-rays and urea and their restitution in sucrose. Cytologia 40:53-60.

Green, S.; Auletta, A.; Fabricant, J.; Kapp, R.; Manandhar, M.; Sheu, C.; Springer, J.; Whitfield,
       B. (1985) Current status of bioassays in genetic toxicology - the dominant lethal assay.
       A  report of the U.S. Environmental Protection Agency  Gene-Tox Program. Mutat. Res.
       154:49-67.

Greene, R. W. (1983) 60 Hz ultra-high voltage electric field effect on onion root mitotic index: a
       lack of impact.  Environ. Exp. Bot. 23(4):355-360.

Hamnerius, Y.; Olofsson, H.; Rasmuson, A.;  Rasmuson, B. (1979) A negative test for
       mutagenic action of microwave radiation  in Drosophila melanogaster.  Mutat.  Res. 68:
       217-223.
        10/22/90
5-98

-------
                            DRAFT-DO NOT QUOTE OR CITE
 Hansson, H.A.; Rozell, B.; Stemme, S. (1987) Effects of experimental exposure to
       power-frequency electromagnetic fields on the nervous system. In: Anderson, L.E.;
       Kelman, B.J.; Weigel, R.J., eds. Interaction of biological systems with static and ELF
       electric and magnetic fields: proceedings of the 23rd Hanford Life Sciences
       Symposium; October 2-4,1984. Richland, WA: Pacific Northwest Laboratory  PP
       297-305.                                                              KK>

 Hansson, H.A. 1981 a.  Lamellar bodies in Purkinje nerve cells experimentally induced by
       electric field.  Brain Res. 216:187-190.

 Hansson, H.A. 1981b.  Purkinje nerve cell changes caused by electric fields: uitrastructure
       studies on long-term effect on rabbis. Med. Biol. 59:103-110.

 Harte, C. (1972) Mutations in Oenothera hookeri after treatment of the pollen with radio waves
       Theoret. Appl. Genet. 42:297-305.

 Harte, C. (1973)  Mutations in Oenothera hookeri after prolonged influence of radiowaves
       during one vegetation period.  Theor. Appl. Genet. 43:6-12.

 Harte, C. (1975)  Mutagenesis by radiowaves in Antirrhinum majus L Mutat. Res. 30:71-75.

 Henderson, B.E.; Pike,  M.C. (1981) Prolactin an important hormone in breast heoplasia?
       Banbury Report 8: hormones and breast cancer. 8:115-130.

 Herich, R. (1976) Influence of a magnetic field upon mitosis and upon the structure of the
       chromosomes. F.R.N. Univ. Com. Physiologia Plantarum 11:1-5.

 Inoue, M.; Miller, M.W.;  Cox, C.; Carstesen, EL (1985) Growth rate and mitotic index analysis
       of Vicia faba L. roots exposed to 60-Hz electric fields. Bioelectromagnetics 6:293-303.

 Juutilainen, J.; Liimatainen, A. (1986) Mutation frequency in Salmonella exposed to weak
       100-Hz magnetic fields.  Hereditas 104:145-147.

 Kachi, T.; Banerji, T.K.; Quay, W.B. (1988) Pineal-adrenomedullary relations: hormonal
       mechanisms affecting tumor growth.  In: Gupta, D.; Attanasio, A.; Reiter, R.J., eds. The
       pineal gland and cancer. Tubingen, West Germany: Brain Research Promotion  pp
       333-344.                                                                 KK<

 Kale, P. G.; Baum, J. W. (1979)  Genetic effects of strong magnetic fields in Drosophila
       melanogaster: I. homogeneous fields ranging from 13,000 to 37,000 Gauss Environ
       Mutagen. 1:371-374.

Kapustin, A.A.; Rudnev, M.I.; Leonskaya, G.I.; Konobeeva, G.I. (1976) Cytogenetic effects of a
       varying SHF electromagnetic field. Cytol. Genet. 10(5):16-18. (Eng. Trans.)

Kornberg, H.A. (1976) Biological effects of electric fields.  Special Issue, RGE-FRA, ISSN
       0035-3116, Revue Generate d'Electricite, Paris, France, 51-64.  (Cited in Portet and
       Cabanes, 1988.)
                                         5-99
10/22/90

-------
                           DRAFT-DO NOT QUOTE OR CITE
Latt, S.A.; Allen, J.; Bloom, S.E. et al. (1981) Sister-chromatid exchanges: a report of the
       Gene-Tox Program.  Mutat. Res. 87:17-62.

Lee, W.R.; Abrahamson, S.; Valencia, R.; Von Halle, E.S.; Wurgler, F.E.; Zimmering, S. (1983)
       The sex-linked recessive lethal test for mutagenesis in Drosophila melanogaster: a
       report of the U.S. Environmental Protection Agency Gene-Tox Program. Mutat. Res.
       123:183-279.

Liboff, A.R.; Williams, T, Jr.; Strong, D.M.; Wistar, R., Jr. (1984) Time-varying magnetic fields:
       effect on DNA synthesis. Science 223:818-820.

Liburdy, R.P.; Wyant, A. (1984) Radiofrequency radiation and the immune system. Part 3. In
       vitro effects on human immunoglobulin and on murine T- and B- lymphocytes. Int. J.
       Radiat Biol. 46:67-81.

Lin, J.C.; Ottenbreit, M.J.; Wang, S.-L; Inoue, S.; Bellinger, R.O.; Fracassa, M. (1979)
       Microwave effects on granulocyte and macrophage precursor cells of mice in vitro.
       Radiat. Res. 80:292-302.

Lin-Liu, S.; Adey, W.R. (1982) Low frequency amplitude modulated microwave fields change
       calcium efflux rates from synaptosomes. Bioelectromagnetics 3:309-322.

Linskens, H.F.; Smeets, P. S. G. M.  (1978) Influence of high magnetic fields on meiosis.
       Experientia 34:42.

Luben, R.A.; Cain, C.D.; Chen, M.C.; Rosen, D.M.; Adey, W.R.  (1982) Effects of
       electromagnetic stimuli on bone and bone cells in vitro: inhibition of responses to
       parathyroid hormone by low-energy low-frequency fields.  Proc. Natl. Acad. Sci. 79:
       4180-4184.

Luben, R.A. (1990) Peer review comments for potential carcinogenicity of electromagnetic
       fields, July 25,1990.

Lyle, D.B.; Schechter, P.;  Adey, W.R.; Lundak, R.L (1983) Suppression of T-lymphocyte
       cytotoxicity following exposure to sinusoidally amplitude-modulated fields.
       Bioelectromagnetics 4:281-292.

Lyle, D.B.; Ayotte, R.D.; Sheppard, A.R.; Adey, W.R.  (1988) Suppression of T-lymphocyte
       cytotoxicity following exposure to 60-Hz sunusoidal electric fields. Bioelectromagnetics
       9:303-313.

Maestroni, G.; Conti, A.; Pierpaoli, W. (1988) The immunoregulatory role of melatonin. In:
       Gupta, D.; Attanasio, A.; Reiter, R.J., eds. The pineal gland and cancer. Tubingen,
       West Germany: Brain Research Promotion, pp. 133-143.

Manikowska, E.;  Luciani, J. M.; Servantie, B.; Czerski, P.; Obrenovitch, J.; Stahl, A. (1979)
       Effects of 9.4 GHz microwave exposure on meiosis in mice.  Experientia 35:388-390.
        10/22/90
5-100

-------
                             DRAFT-DO NOT QUOTE OR CITE
 Marec, F.; Ondracek, J.; Brunnhofer, V. (1985)  The effect of repeated microwave irradiation on
        the frequency of sex-linked recessive lethal mutations in Drosophila melanoqaster
        Mutat. Res. 157:163-167.

 Marron, M. T.; Goodman, E. M.; Greenebaum, B. (1975) Mitotic delay in the slime mould
        Physarum polycephalum induced by low intensity 60 and 75 Hz electromagnetic fields
        Nature 254:66-67.

 Mastryukova, V. M.; Rudneva, S. V. (1978) Effect of a strong magnetostatic field on
        proliferation of duodenal epithelial cells in mice. Biol. Bull. Acad  Sci USSR 5(3V
        371-374.                                                      '        V ''

 McLeod, K.L.; Lee, R.C.; Ehrlich, H.P. (1987) Frequency dependence of electric field
        modulation of fibroblast protein synthesis.  Science 236:465-1468.

 McRee, D. I.; MacNichols, G.; Livingston, G. K. (1981)  Incidence of sister chromatid exchange
        in bone marrow cells of the mouse following microwave exposure.  Radiat Res
        85:340-348.

 Merritt, J.H.; Shelton, W.W.; Chamness, A.F. (1982) Attempts to alter 45Ca2+ binding to brain
       tissue with pulse-modulated microwave energy. Bioelectromagnetics 3:475-478.

 Mhatre, M.C.; Shah, P.N.; Juneja, H.S. (1984) Effect of varying  photoperiods on mammary
       morphology, DNA synthesis and hormonal profile in female rats. J. Natl. Cancer Inst
       72:1411-1416.

 Michaelson, S.M.; Lin, J.C. (1987) Neural effects of microwave/radiofrequency energies.  In:
       Biological effects and health implications of radiofrequency radiation. New York-
       Plenum Press, pp. 361 -411.

 Michaelson S.M. (1979) Human responses to power frequency exposures.  In:  Biological
       effects of extremely low electromagnetic fields.  DOE Symposium Series 50, Technical
       Information Center, Oak Ridge, TN, pp. 1-20. (Cited in Portet and Cabaes, 1988.)

 Mileva, M.;  Bulanova, M.; Ivanov, B. (1985)  Permanent magnetic field (PMF) exposure in vivo
       and in vitro to the genetic structures of somatic cells. Mutat. Res. 147:309.

 Miller, M.W.; Reddy, M.M.; Yettewich, G.R.; Kaufman, G.E. (1976) Lack of effect of extremely
       low frequency electric and magnetic fields on roots of Vicia faba. Environ. Exp Bot 16-
       83-88.                                                                     '   '

 Mittler, S. (1971) Failure of magnetism to influence  production of x-ray induced sex-linked
       recessive lethals. Mutat. Res. 13:287-288.

Mittler, S. (1976)  Failure of 2- and 10-meter radio waves to induce genetic damage in
       Drosophila melanogaster. Environ. Res. 11:326-330.

Mittler, S. (1977) Failure of chronic exposure to nonthermal FM  radio waves to mutate
       Drosophila. J. Hered. 68:257-258.
                                        5-101
10/22/90

-------
                            DRAFT-DO NOT QUOTE OR CITE
Moore, R. L (1979)  Biological effects of magnetic fields: studies with microorganisms. Can. J.
       Microbiol. 25:1 145-1 151.

Morris, J.E.; Phillips, R.D. (1 982) Effects of 60-Hz electric fields on specific humoral and
       cellular components of the immune system. Bioelectromagnetics 3:341 -347.

Myers, R.D.; Ross, D.H. (1981)  Radiation and brain calcium. A review and critique.  Neurosci.
       Biobehav. Rev. 5:503-543.

Nemirovich-Danchenko, E.N.; Chastokolenko, L. V. (1976) Effect of orientation in the
       geomagnetic field and age of radicles on mitotic activity of the meristem under the
       influence of an artificial magnetic field. Sov. Plant Physiol. 23(4) -.698-702.

Nordenson, I.; Hansson-Mild, K.; Nordstrom, S.; Sweins, A.; Birke, E. (1984) Clastogenic
       effects in human lymphocytes of power frequency electric fields: in vivo and in vitro
       studies. Radiat. Environ. Biophys. 23:191-201.

Olcese, J.; Reuss, S.; Volbrath, L (1985) Evidence for the involvement of the visual system in
       mediating magnetic field effects on pineal melatonin synthesis in the rat. Brain Res.
       333:382-384.

Ottenbreit, M.J.; Lin, J.C.; Inoue, S.; Peterson, W.D., Jr. (1981)  In vitro microwave effects on
       human neutrophil precursor cells (CFU-C).  Bioelectromagnetics 2:203-215.

Panchuk, A.S.; Kachura, V.S.; Dyshlovoi, V.D. (1 981 ) Effect of the electromagnetic field of 50
       Hz frequency on cell passage through certain mitotic cycle periods. Tsitol.  Genet.
 Pay.T.L; Beyer, E.G.; Reichelderfer, C.F. (1972) Microwave effects on reproductive capacity
       end genetic transmission in Drosophila melanogaster. J. Microw. Power 7:75-82.

 Phillips, J.L; Winters, W.D. (1987) Electromagnetic-field induced bioeffects in human cells in
       vitro. In: Anderson, L.E.; Kelman, B.J.; Weigel, R.J., eds. Interaction of biological
       systems with static and ELF electric and magnetic fields: proceedings of the 23rd
       Hanford Life Sciences Symposium; October 2-4, 1984.  Richland, WA: Pacific
       Northwest Laboratory, pp. 279-285.

 Phillips, J.L; Winters, W.D.; Rutledge, L (1986a) In vitro exposure to electromagnetic fields:
       changes in tumour cell properties.  Int. J. Radiat. Biol. 49:463-469.

 Phillips, J.L; Rutledge, L; Winters, W.D. (1986b) Transferrin binding to two human colon
       carcinoma cell lines: characterization and effect of 60-hz electromagnetic fields.
       Cancer Res. 46:239-244.

 Portet, R.; Cabanes, J.; Delost, H. (1984) Developpement du jeune lapin soumis a un champ
       electrique intense. C.R. Soc. Biol. 178:142-152. (Cited in Portet and Cabanes, 1988.)

 Portet, R.; Cabanes, J. (1 988)  Development of young rats and rabbits exposed to a strong
       electric field. Bioelectromagnetics  9:95-1 04.
        10/22/90
5-102

-------
                             DRAFT-DO NOT QUOTE OR CITE
 Portnov, F.G.; Shakarnis, V.F.; Maiore, D.Ya. (1975)  Study of the mutagenic effect of static
        electric fields on Drosophila melanogaster females. Sov. Genet. 11:797-799.

 Ragan, HA; Buschbom, R.L; Pipes, M.F.; Phillips, R.D.; Kaune, W.T. (1983) Hematologic and
        serum chemistry studies in rats exposed to 60-Hz electric fields. Bioelectromaanetics
        4:79-90.                                                               a

 Ramaiya, L.K.; Pomerantseva, M.D.; Vilkina, G.A.; Tikhonchuk, V.S. (1980)  Study of the action
        of UHF microwaves on mammalian germ and somatic cells. Cytol. Genet. 14(6):1-5.

 Reese, J.A.; Jostes, R.F.; Frazier,  M.E. (1988)  Exposure of mammalian cells to 60-Hz magnetic
        or electric fields: analysis for DNA single-strand breaks.  Bioelectromagnetics 9:237-247.

 Reiter, R.J.  (1988) Pineal gland, cellular proliferation and neoplastic growth: an historical
       account. In: Gupta, D.; Attanasio, A.; Reiter, R.J., eds. The pineal gland and cancer
       Tubingen, West Germany:  Brain Research Promotion, pp. 41-64.

 Reuss, S.; Olcese, J. (1986)  Magnetic field effects on the rat pineal gland: role of retinal
       activation by light.  Neuroscience Lett.  64:97-101.

 Roberts, N.J.; Michaelson, S.M.; Lu, S.-T.  (1987) Mitogen responsiveness after exposure of
       influenza virus-infected human mononuclear leukocytes to continuous or
       pulse-modulated radiofrequency radiation.  Radiat. Res. 110:353-361.

 Robertson,  D.; Miller, M. W.;  Cox,  C.; Davis, H.T. (1981)  Inhibition and recovery of growth
       processes in roots of Pisum sativum L. exposed to 60-Hz electric fields.
       Bioelectromagnetics 2:329-340.

 Rodin, A. (1963) The growth and spread of Walker 256 carcinoma in pinealectomized rats
       Cancer Res. 23:1545-1549.

 Rosen, D.M.; Luben, R.A. (1983) Multiple hormonal mechanisms for the control of collagen
       synthesis in an osteoblast-like cell  line, MMB-1.  Endocrinology 112:992-999.

 Rosenthal, M.; Obe, G. (1989) Effects of 50-Hertz electromagnetic fields on proliferation and
       on chromosomal alterations in human peripheral lymphocytes untreated or pretreated
       with  chemical mutagens. Mutat. Res. 210:329-335.

 Semm, P.; Schneider, T.; Volbrath, L. (1980) Effects of an earth-strength magnetic field on
       electrical activity of pineal cells. Nature 288:607-608.

Seto, Y.J.; Majeau-Chargois,  D.; Lymangrover, J.R. et al. (1986) Chronic 60-Hz electric field
       exposure-induced subtle bioeffects on hematology.  Environ. Res. 39:143-152.

Shah, P.M.; Mhatre, M.C.; Kothari, LS. (1984) Effect of melatonin on mammary carcinogenesis
       in intact and pinealectomized rats in varying photoperiods.  Cancer Res. 44:3403-3407.

Shelton, W.W.; Merritt, J.H. (1981) In vitro study of microwave effects on calcium efflux in rat
      brain tissue.  Bioelectromagnetics 2:161 -167.
                                        5-103
                                                                              10/22/90

-------
                           DRAFT-DO NOT QUOTE OR CITE
Shevchenko, V.V.; Grinikh, LI.; Strekova, V.Yu. (1978) Does a constant magnetic field induce
       chromosomal aberrations in plants?  Sov. Genetics 14(6):784-786.

Smialowicz, R.J.; Ali, J.S.; Bermen, E. et al. (1981) Chronic exposure of rats to 100 MHz (CW)
       radiofrequency radiation: assessment of biological effects.  Radiat. Res. 86:488-505.

Steiner, E. (1975)  Oenothera.  In:  King, R.C., ed. Handbook of Genetics, 2. New York, NY:
       Plenum Press, pp. 223-245.

Stevens, R.G. (1987)  Electric power use and breast cancer: a hypothesis.  Am. J. Epidemiol.
       125:556-561.

Stodolnik-Baranska, W. (1974) The effects of microwaves on human lymphocyte cultures.
       Biologic effects and health hazards of microwave radiation. Proc. Int. Symp., Warsaw,
       1973; pp. 189-195.

Strekova, V.Yu.; Spitkovskii, D.M.  (1971) Model of condensed supremolecular DNP-systems
       for analysis of possible structural damage to chromosomes in constant magnetic fields.
       Sov. Plant Physiol. 18(1):157-160.

Strzhizhovskii, A.D.; Galaktionova, G.V.; Chermnykh, P.A. (1979) Influence of
       infralow-frequency magnetic fields of high and ultrahigh intensity on the division of
       mammalian cells.  Biol. Bull. Acad. Sci. USSR 6(1):97-100.

Swicord, M.L; Davis, C.C. (1983)  An optical method for investigating the microwave
       absorption characteristics of DNA and other biomolecules in solution.
       Bioelectromagnetics 4:21 -42.

Swicord, M.L; Edwards, G.S.; Sagripanti, J.L.; Davis, C.C.  (1983) Chain-length-dependent
       microwave absorption of DNA. Biopolymers 22:2513-2516.

Szmigielski, S. (1975)  Effect of 10-cm (3-GHz) electromagnetic radiation (microwaves) on
       granulocytes in vitro. Ann. NY Acad. Sci. 247:275-280.

Takashima, S.; Onaral, B.; Schwan, H.P. (1979)  Effects of modulated RF energy on the EEG of
       mammalian brains. Radiat. Environ.  Biophys. 16:15-27.

Takahashi, K.; Kaneko, I.; Date, M.; Fukada, E. (1986) Effect of pulsing electromagnetic fields
       on DNA synthesis in mammalian cells in culture.  Experientia 42:185-186.

Tamarkin, L; Baird, C.J.; Almeida, O.F. (1985) Melatonin: a coordinating signal for mammalian
       reproduction. Science 227:714-720.

Tamarkin, L; Cohen, M.; Roselle, D.; Reichert, C.; Lippman, M.; Chabner, B. (1981) Melatonin
       inhibition and pinealectomy enhancement of 7,12-dimethylbenz(a)anthracene-induced
       mammary tumors in the rat.  Cancer Res. 41:4432-4436.

Tsoneva, M.T.; Penchev, P.R.; Karev, G.B.;  Gishin, S.S. (1975) Effect of magnetic fields on the
       chromosome set and cell  division. Sov. Genetics 11:398-401.
       10/22/90
5-104

-------
                             DRAFT-DO NOT QUOTE OR CITE
 U.S. Environmental Protection Agency (U.S. EPA) (1984) Biological effects of radiofrequency
        radiation.  Office of Research and Development.  EPA-600/8-83-026F.

 Webber, M.M.; Barnes, F.S.; Seltzer, LA.; Bouldin, T.R.; Prasad, K.N. (1980) Short microwave
        pulses cause ultrastructural membrane damage in neuroblastoma cells. J Ultrastruct
        Res. 71:321 -330.

 Welker, HA;  Semm, P.; Willig, R.P.; Commenty, J.C.; Wiltschko, W.; Volbrath,  L (1983) Effects
        of an artificial magnetic field on serotonin N-acetyltransferase activity and melatonin
        content of the rat pineal gland.  Exp. Brain Res. 50:426-432.

 Whitson, G.L.; Carrier, W.L; Francis, A.A. et al. (1986) Effects of extremely low frequency (ELF)
        electric fields on cell growth and DMA repair in human skin fibroblasts.  Cell Tissue
        Kinet.  19:39-47.

 Wilson, 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.

 Wilson, B.W.;  Anderson, L.E.; Hilton, D.I.; Phillips, R.D. (1983)  Chronic exposure to 60-Hz
        electric fields: effects on pineal function in the rat.  Bioelectromagnetics 4-293
        (Erratum to Wilson et al., 1981)

 Wilson, B.W.;  Chess,  E.K.; Anderson, LE. (1986)  60-Hz electric-field effects on pineal
        melatonin rhythms: time course for onset and recovery. Bioelectromagnetics 7:239-242.

 Wilson, B.W.;  Lueng, F.; Buschbom, R.; Stevens, R.G.; Anderson, L.E.; Reiter, R.J. (1988)
        Electric fields, the pineal gland, and cancer. In: Gupta, D.; Attanasio, A.; Reiter, R J
       eds. The pineal gland and cancer. Tubingen, West Germany: Brain Research  '
       Promotion, pp. 245-259.

 Wolff, S.; James, T.L; Young, G.B.; Margulis, A.R.; Bodycote, J.; Afzal, V. (1985) Magnetic
       resonance imaging: absence of in vitro cytogenetic damage. Radiology 155:163-165.

 Wright, N.A.; Borland, R.G.; Cookson, J.H. et al. (1984) Biological studies with
       continuous-wave radiofrequency (28 MHz) radiation.  Radiat. Res. 97:468-477.

 Yao, K.T.S. (1978)  Microwave radiation-induced chromosomal aberrations in corneal
       epithelium of Chinese hamsters. J. Hered. 69:409-412.

Yao, K.T.S. (1982)  Cytogenetic consequences of microwave irradiation on mammalian cells
       incubated in vitro. J.  Hered. 73:133-138.

Yen-Patton, G.P.A.; Patton, W.F.; Beer, D.M.; Jacobson, B.S. (1988)  Endothelial cell response
       to pulsed electromagnetic fields: stimulation of growth rate and angiogenesis in vitro
       J. Cell.  Physiol. 134:37-46.
                                         5-105
                                                                               10/22/90

-------

-------
                             DRAFT-DO NOT QUOTE OR CITE


                                 6. RESEARCH NEEDS
  6.1. INTRODUCTION

     In order for the Agency to evaluate the public health hazard of electromagnetic (EM) fields
  and to be in a position to recommend preventive measures, more information is needed in
  several areas. The research topics that need to be investigated to supply this information are
  summarized as follows:
     •  Health Hazards Evaluation
        -  Cancer
        -  Reproductive effects
        -  Central nervous system neuroendocrine, immunological effects
     •  Exposure Evaluation
        -  Characterization of high frequency transients and harmonics of fields from electrical
           power sources
       .-  Relative contribution of sources to total exposure
           -  High-voltage transmission lines
           -  Distribution lines to homes and industries
           -  Power distribution transformers
           -  Appliances
           -  Ground currents in households, industrial buildings, office buildings and
              schools
           -  Transportation systems (e.g., electric trains)
    •  Mitigation (exposure reduction)
       -   Reduction of ground currents
       -   Redesign of appliances
       -   Redesign of wiring and routing of distribution lines
       -  Supression of transients and high frequencies
       —  Avoidance of hazardous sources
   This section deals only with  research needed for evaluation of the cancer hazard, although
we recognize that other areas also need further research.

6.2.  INFORMATION NEEDS ARISING FROM THE EVALUATION IN THIS DOCUMENT
   The evaluation carried out in this document has raised several unanswered questions
about the carcinogenic potential of EM fields.  The major information needs that have been
identified are outlined in this section. Before these needs can be translated into a research
                                        6-1
                                                                            10/02/90

-------
                           DRAFT-DO NOT QUOTE OR CITE

program, the ongoing research that is now being carried out must be taken into account. This
is not being attempted in this section; rather, the emphasis here is on the issues about
population hazards that need to be dealt with before an evaluation of the public health hazards
and the determination of meaningful preventive measures can be made.

6.2.1.  Epidemiology Research Needs
   The association between cancer occurrence and exposure to either extremely low
frequency (ELF) or radiofrequency (RF) fields is not strong enough to constitute a proven
causal relationship, largely because the relative risks in the published reports have seldom •
exceeded 3.0 in both childhood .residential exposures and in occupational situations. Two
possible explanations for this are:  (1) our imprecise knowledge of the causal aspects of
exposure (field strength, frequency, time patterns of exposure, and synergistic factors)
prevents us from identifying exposure indices that distinguish exposed from unexposed
populations, or (2) the observed effects are actually caused by some other factor not related to
the EM-fields but which co-varies with EM-field strength. A third possible explanation is that
exposures have been too weak to produce an observable effect. This cannot be evaluated
because there is currently no reasonable basis for making  predictions of the expected human
response.
    To evaluate the first two possibilities, two things need to be done:  (1) define job categories
in electrically-related occupations to reflect actual exposure to EM fields. These definitions
would be used as the basis for selecting cohorts for study. These studies should be designed
to investigate a variety of exposure parameters, which need to be judiciously selected using
the most recent concepts of likely mechanisms of carcinogenesis; and (2) investigate, with
 improved study designs and exposure measurements, those populations that have already
 showed some excess risk from EM fields or have potentially high exposures, such as military
 communications and radar workers; amateur radio operators; telephone and electrical utility
 workers; electric-arc welders; aluminum smelter workers; engineers, scientists, and computer
 operators working with electrical equipment; people living near radio and TV broadcast
 towers; users of electric blankets; and Hodgkin's lymphoma patients. In designing these
 studies, special attention needs to be given to confounding variables  in order to avoid an
 incorrect attribution of the effects to EM fields simply as a consequence of the intense scrutiny
        10/02/90
                                           6-2

-------
                            DRAFT-DO NOT QUOTE OR CITE


 given to them. More information about the risks of RF and modulated RF exposure is needed
 before any conclusions can be made about either their safety or hazard.

 6.2.2. Laboratory Research Needs
    Both animal and in vitro studies are needed to discover the relevant exposure factors and
 their interaction, to gain some understanding of the mechanisms of action, and to extrapolate
 "effective doses" to the human exposure situation.
    For long-term animal studies, the most obvious need is to conduct a carcinogen bioassay
 with magnetic field ELF exposures. The experiment planned by the Ontario Hydroelectric
 Power Company (described in Section 4.6) will be the first animal carcinogen bioassay to be
 done with 60-Hz magnetic fields.  It may or may not confirm the human findings and will
 generate additional research topics regardless of the outcome.  Since the University of
 Rochester study is already examining the effect of ELF electric fields on the growth rate of
 mammary carcinomas in rats, there is not a high priority need at this time to initiate another
 chronic experiment with electric fields.  However, this same type of study is needed for ELF
 magnetic fields.  If these studies show that a carcinogenic response is induced in animals by
 these ELF exposures, then whether RF fields modulated at the same ELF frequencies will
 produce comparable effects is the next logical question, since modulated RF exposures are
 common and some laboratory phenomena (Table 5-7) have the same effect with or without the
 RF component.
   For in vitro studies, several biological phenomena, having some relationship to possible
 mechanisms of carcinogenesis, have been induced by EM fields. The discussion in Section
 5.11 summarizes the large amount of missing information relative to the biological effects of
 EM fields and forms the basis for this discussion.  The overall goal of the research program
would be to select the most promising candidate mechanisms of cancer induction at ambient
electric and magnetic field strengths and frequencies and to experimentally investigate, in
laboratory in vitro tests and in whole animals, the way in which each process depends upon
several field exposure parameters, such as type of field  (alternating electric and magnetic
fields, geomagnetic fields), field strength, and frequency and time patterns of exposure
(steady, intermittent, time of day).  The leading research areas identified in this review are:
                                         6-3
                                                                             10/02/90

-------
                     DRAFT-DO NOT QUOTE OR CITE


Biological response to weak fields
If ambient 60-Hz magnetic fields [which have strengths on the order of 5 milligauss
(mG) or less] are really causing a cancer response, then tissue currents on the order
of 1CT4 microamperes per square centimeter (uA/cm2) must be able to induce some
kind of change in cellular metabolism. Of the processes reviewed in this document,
there are only three that occur at field strengths comparable to this: (1) calcium release
from chick brain tissue in response to crossed electric and magnetic fields; (2) the
change in melatonin secretion in rats induced by a small change in the orientation of
the static magnetic field; and (3) calcium release from chick brain tissue induced by
electric fields.  Since the first phenomenon empirically exhibits frequency selectivity,
there is at least a theoretical possibility that weak currents of the correct frequency
could induce an effect even though they are within the range of thermal noise. There is
a need to establish a mechanism that would explain how magnetic fields could induce
such a calcium release from brain tissue. Ion cyclotron resonance and the quantum
beat models are two possibilities, and experiments are needed to test those hypothe-
ses.  However, nuclear magnetic resonance has been suggested as another possibility,
and that has not been examined.
The effect of low-strength alternating magnetic fields on melatonin secretion has not
been measured, and the single observation that a very small change in magnetic field
orientation has induced an inhibition of synthetic activity needs to be verified. Of partic-
ular interest is the ELF frequency and intensity dependence of this inhibition.  In addi-
tion, it would be of interest to ascertain whether modulated RF fields affect pineal
 melatonin synthesis.
 Chemical signalling pathways for controlling cell proliferation
 Since the transduction of hormone and other chemical signals into the cell is a neces-
 sary step in the hormonal control of cell proliferation, and since this transduction pro-
 cess is defective in transformed cells, the influence of EM fields on signal transduction
 has a potentially important role in the development of cancer. The experiments on or-
 nithine decarboxylase (ODC) (Section 5.6), gene transcription (Section 5.3), and para-
 thyroid hormone (Section 5.5) raise the possibility that electric and magnetic fields
 interfere with the normal functioning of signal transduction pathways.  The effects of
  EM fields on one of the major pathways, the receptor-mediated activation of  phos-
  10/02/90
6-4

-------
                      DRAFT-DO NOT QUOTE OR CITE


 phoinositide turnover, has not been explored but should be because this pathway is in-
 volved in several phenomena related to cell proliferation. This pathway involves
 diacylglycerol (DAG), which is a structural analogue of the tumor promoter TPA, both of
 which activate protein kinase C (PKC), and, ultimately, ODC.  It also involves inositol
 trisphosphate (IPS), which is known to release calcium from intracellular stores and the
 plasma membrane. Both DAG and IPS, through calcium release, activate PKC which
 phosphorylates growth factor receptors and proto-oncogenes and is  believed to play a
 role in the control of cell  proliferation. As our understanding of these pathways be-
 comes more detailed, we will be able to postulate chemical synergisms and  antago-
 nisms with EM fields.  Further study of parathyroid hormone as a mediator of bone
 healing response to magnetic fields will lead to a better appreciation of the interaction
 between hormones and EM fields in the control cellular proliferation.
 In addition to the chemical events associated with the cell membrane, the influence of
 EM fields on the dynamics of charged membrane-bound proteins needs to be studied.
 There is a possibility that the alternating electric field at the membrane or currents in-
 duced by these fields directly affect enzymatic reactions, such as ion transport reac-
 tions, carried out by these proteins. There is also the possibility that the fields modify
 the opening and closing of ion-conducting channels in the membrane, and they could
 affect the stability of membrane receptors.
 Gene expression
 The influence of EM fields on the expression of genetic activity, both with respect to
 transcription to mRNA's and in translation of genetic information into protein synthesis,
 is needed.  It is important to study how this process is connected to chemical signalling
 pathways.
 Cell transformation
The single experiment showing that modulated microwave radiation acts as an initiator
of the transformation of C3H10T1/2 cells to malignancy needs to be verified and re-
peated with ELF magnetic and electric fields. The interaction of EM-field exposure with
other factors known to influence transformation in this well-studied system may lead to
clues as to the mechanism of this transformation.
                                   6-5
                                                                       10/02/90

-------
                          DRAFT-DO NOT QUOTE OR CITE



   •  Meiatonln Activity
      Although it is well known that melatonin inhibits the growth of hormonally dependent tu-

      mors and that electric field exposure inhibits the synthesis of melatonin by the pineal

      gland, the effect of ELF fields on these end points has not been directly studied. In ad-

      dition, the question of whether ELF magnetic fields inhibit melatonin secretion is espe-

      cially important in view of the postulated human response to magnetic fields. This

      mechanism becomes especially important to the induction of leukemias and lympho-

      mas by EM fields in view of the successful treatment of clinical leukemia with melatonin

      (Section 5.7.1).

   •  Ion cyclotron resonance interaction

      Several questions need to be answered about the ion cyclotron resonance phenome-

      non:

      -  In artificial membranes impregnated with channel proteins of various ion
          specificities, does the ion cyclotron resonance field open an ionic conduction
          channel, as one experiment with cell suspensions suggests?
      -  Does the calcium release from brain tissue depend on an ion cyclotron  resonance
          process?
      -  Are other intracellular processes dependent on specific ions triggered by plasma
          membrane ion gating induced by ion cyclotron field conditions? This process has
          been implicated as a mechanism of diatom locomotion mediated by intracellular
          calcium ions and hypothetical^ for rat behavioral patterns mediated by lithium, but
          more examples are needed, perhaps with sensory cells and with cells in which
          field-sensitive biochemical reactions are taking place.

    In planning and executing this research, it is important to recognize that effects of fields on

the whole animal are likely to be unpredictable based on the results of laboratory in vitro

systems, so that there is a constant need to verify these effects in the whole animal before

being able to make valid inferences about effects in humans.
       10/02/90
6-6

-------
                            DRAFT-DO NOT QUOTE OR CITE
                          7. SUMMARY AND CONCLUSIONS
 7.1. INTRODUCTION
    In this chapter each of the major chapters in the document are summarized; a final section
 presents a discussion of the relationships among the individual chapters and the overall
 conclusions.

 7.2.  MECHANISMS OF INTERACTION BETWEEN TISSUE AND ELECTROMAGNETIC
    FIELDS
    The basic processes by which energy from electromagnetic (EM) fields of radiofrequency
 (RF) and extremely low frequency (ELF) frequencies is coupled to the body are described in
 this section. The frequency dependence of the RF power absorbed by an organism is
 dominated by the body size, so that mice, rats, and humans have different RF absorption
 characteristics. For ELF fields and the lower RF frequencies near the source, the relationship
 between the electric and magnetic fields is not fixed, as it is for RF fields, and they are
 evaluated separately in this document.  From the point of view of EM fields, the body is
 composed of a solution of ions; it is an electrical conductor and the penetration of electric
 fields into the body is very poor at ELF frequencies. Since the body is  composed of
 nonmagnetic materials, an external time-varying magnetic field permeates the body, inducing
 ionic  currents.
    The human evidence, as described in the next section, suggests that magnetic fields,
 rather than electric fields, are associated with cancer incidence, and mechanisms have been
 sought to explain how weak currents induced by ELF magnetic fields could interact with cells
 and body tissue in such a way as to induce a carcinogenic response. Three classes of
 models for this interaction are reviewed.  (1) The surface compartment model deals with the
 movement of ions towards and away from the inner and outer surfaces of the plasma
 membrane of the cell, and deals with ion-selective membrane channels, ionic pumps, and
 membrane ion  fluxes. The model describes the movement of ions in response to
 perturbations of electric fields and magnetically induced currents around the cell. (2) The ion
 cyclotron resonance hypothesis was developed in part to explain the frequency sensitivity of
 calcium  ion efflux studies of brain tissues. If the relationship among the frequency of
time-varying magnetic field, the strength of a parallel static magnetic field, and the ionic charge
to mass ratio of an ionic species is correct, then the ion will resonate, or synchronously follow
circular paths in a plane perpendicular to the field. In one experiment demonstrating this
                                        7-1
10/22/90

-------
                           DRAFT-DO NOT QUOTE OR CITE


effect, conditions were set up for calcium cyclotron resonance, and the movement of benthic
diatoms was measured. The authors interpreted the experiment as showing that calcium ions
entered into the cell under these specific conditions and stimulated the motion of the cells,
whereas the cell is normally impermeable to calcium.  This type of mechanism could be the
basis of an induced selective ion permeability of the plasma membrane and might ultimately
be capable of explaining both frequency selectivity of these effects and the sensitivity to small
induced currents. (3) Another class of models deals with cooperative motions of an ordered
array of lipid bilayer molecules and describes how a weak field affecting the motion of the
whole array could be transferred to just one site in the array. These theories have not yet
been tested in the context of ELF biological processes.  At the present time, these basic
models of tissue interaction with EM fields cannot be linked to the biochemical or cellular
processes involved in the development of malignant growth.

7.3. HUMAN EVIDENCE
    The effects of human exposure to EM fields from several sources have been reported.  This
document discusses ELF fields separately from higher frequency exposure where possible.
Children with residential exposure are more appropriate subjects than adults for evaluating the
effects of ELF fields, since children have relatively little exposure to higher frequency fields and
occupational chemicals as a consequence of their normal activity patterns. Consequently,
studies of childhood cancer associated with residential exposure to 60-Hz power frequency
fields are discussed separately from occupational exposure to adults,  which involves a mixture
of both ELF and RF fields.

7.3.1. Studies of Children
    There have been seven case-control studies of cancer in children  examining residential
exposure from power transmission and distribution systems. Two additional studies have
examined childhood cancer in relationship to father's occupation.  Six of the seven residential
exposure studies showed positive associations with  ELF field exposure; three were statistically
significant and the other three had odds ratios greater than one but not statistically significant.
Where different cancer sites were evaluated, leukemia, nervous system cancer, and, to a
 lesser extent, lymphomas were found to be in excess in the five residential studies showing
 positive associations.  Electric fields were not found to be a critical factor thus far. Surrogates
 of magnetic  field exposure differed among the seven studies. Wire code configurations and
 proximity to  distribution lines were used in six of the seven studies, and measurements were
        10/22/90
7-2

-------
                            DRAFT-DO NOT QUOTE OR CITE

 taken in two of the seven studies. There is a reasonably good, but not perfect, correlation
 between measured magnetic fields and wire code configurations. In two of the studies in
 which magnetic field measurements were made, cases were observed in those exposed at or
 above 2 to 3 milligauss (rnG) [0.2 to 0.3 microtesla (/
-------
                           DRAFT-DO NOT QUOTE OR CITE


7.3.2.2. Occupational Exposure to Extremely Low Frequency and Mixed
   Frequency Fields
   Twenty-eight reports dealing with cancer incidence or mortality in workers in electrical and
electronic occupations have been reviewed. These exposures have involved 50 or 60 Hz
power frequency fields as well as mixtures of higher frequency fields which are typically poorly
defined. The studies have been carried out in Europe, New Zealand, and the United States.
Many of them were re-examinations of previous studies or evaluations of vital records, cancer
registry, or occupational data bases, and thus the populations were not formed to test the
specific hypothesis of whether EM-field exposure is associated with increased cancer risk.
Most of them used death certificates as a source of occupational information; this information
furnishes only a very crude indicator of actual exposure to EM fields. Many of these are
proportional mortality studies, which are less informative than studies of cohort and
case-control designs because their results are affected by extraneous causes of death.
    In these studies three types of cancer predominate:  (1) hematopoietic system, especially
leukemia and specifically acute myeloid  leukemia;  (2) nervous system cancer, including brain
tumors; and (3) malignant melanoma of the skin. These cancer sites are found consistently
across different geographic regions, age groups, industries, occupational classifications, and
study designs.  Given this diversity of studies, in addition to the  likelihood that across broad
job categories the exposures to various  chemicals is not uniform, it is difficult to identify any
single agent or group of confounding exposures that could explain the consistent finding of
these same cancer sites.

7.3.2.3, Radiofrequency Exposure
    Reports that focused primarily on exposures to RF radiation have shown mixed results, but
most of the studies were difficult to interpret.  Two early reports  concerning microwave
exposure of U.S. embassy personnel in  Moscow and radar exposure of U.S. Navy personnel
showed only a slight tendency for increased cancer risk at all sites, and somewhat higher
odds ratios for hematopoietic system cancers. A study of personnel in a World War II radar
research and development laboratory found no convincing evidence of increased cancer
incidence, but errors of exposure misclassification are likely. A  series of reports of ham radio
operators found a statistically elevated incidence of acute myeloid leukemia and other
neoplasms of the lymphoid system, but no clear dose-response trend was seen with longer
exposure, where the degree of exposure was inferred by FCC operator license class. One
        10/22/90
7-4

-------
                            DRAFT-DO NOT QUOTE OR CITE


 report of military exposure to radar found increasing rates of hematopoietic cancer of specific
 sites, but a lack of detail limits the ability to interpret the results.

 7.3.3. Summary of Human Evidence
    The strongest evidence that there is a causal relationship between certain forms of cancer,
 namely leukemia, cancer of the nervous system, and, to a lesser extent, lymphoma and
 exposure to magnetic fields comes from the childhood cancer studies. Several studies have
 consistently found modestly elevated risks (some statistically significant) of these three
 site-specific cancers in children. In two of the studies in which magnetic field measurements
 were made, cases were observed  in those exposed above 2 to 3 mG (0.2 to 0.3/*T) but not in
 children exposed below that level. This is supported by the fact that children have relatively
 few confounding influences that could explain the association. In fact, the few potential
 confounders and biases that might have had an effect on the results were examined by one of
 the authors in some detail and found not to be a serious problem. No other agents have been
 identified to explain this association.  However, there are contradictory results within these
 same studies, and dose-response relationships could not be substantiated. Furthermore,
 there is little information on personal exposure and duration of residency in the EM fields.
    Additional, but weaker evidence that there is an elevated risk of leukemia, cancer of the
 nervous system, and perhaps other sites comes from occupational studies of EM-field
 exposure. Although many of these studies have found an  excess risk of these forms of cancer
 with employment in  certain jobs that have a high potential for exposure to EM fields, few or no
 measurements have actually been  taken in those occupations. Furthermore, information
 about occupation has come generally from sources that could be characterized as sketchy
 The likelihood that misclassification or information bias is present in these studies is high.
 However, exposure misclassification, if random, tends to bias relative risks toward the null.
 Despite these weaknesses, the occupational studies tend to support the results of the
 childhood studies, since the excess relative risks occur at the same sites.
   The studies of residential adult  exposures to EM fields provide little evidence of a risk of
 leukemia, mainly due to a lack of statistical power and/or probably little exposure to levels of
 EM fields that have been found to be associated with cancer in children. These studies cannot
be interpreted as evidence either for or against a causal association between cancer and
EM-field exposures.  On the other hand, the case-control study of cancer  in Colorado
residents does support an association of central nervous system cancer and lymphoma if
                                         7-5
10/22/90

-------
                           DRAFT-DO NOT QUOTE OR CITE


proximity to high-current electrical wiring configurations is assumed to be an adequate
surrogate for exposure.
   The studies of adults exposed to RF radiation produced mixed results, primarily because ol
limited sample size, inadequate length of follow-up, imprecise exposure data, and lack of
information on potential confounders. These problems prevent conclusions to be made about
causal relationships with RF exposure. However, the statistically significant excess risks of
leukemia in amateur radio operators requires further examination.

7.4. ANIMAL EVIDENCE
7.4.1.  Extremely Low Frequency Fields
    No lifetime animal carcinogen bioassay studies of ELF fields have been reported in the
literature.  Several studies currently in progress are designed to observe the induction of a
carcinogenic response to chronic magnetic field exposures.

7.4.2.  Radiofrequency Radiation
    Two chronic studies in mice have used unmodulated RF radiation at 800 megahertz (MHz)
and 2450 MHz, respectively.  Two studies in rats have used pulse modulated 2450 MHz of low
power density and pulsed RF of all frequencies from 0 to about 20 MHz of high power density,
respectively.  One mouse study used pulsed RF radiation of 9270 MHz.

7.4.2.1.   Unmodulated Radiofrequency Radiation
    For unmodulated RF radiation one of the mouse studies (Szmigeilski et al., 1982) shows
that the radiation enhances the growth rate of spontaneous mammary tumors and in a
separate  experiment enhances the growth rate of skin tumors initiated by a chemical
 carcinogen, benzo[a]pyrene. In a shorter test (3 months), the same authors showed that the
 radiation  also enhances the growth rate of transplanted lung carcinoma cells,  an effect
 attributed to the lowering of cell-mediated immunity. Unfortunately, histopathology was not
 reported  in the other mouse  study (Spalding et al., 1971), so conclusions about
 carcinogenicity from that study are difficult to make.
    The special nature of the response indicates that unmodulated RF radiation might be a
 promoter or cocarcinogen, since the growth rate of spontaneous breast tumors, BaP-induced
 skin tumors, and transplanted lung sarcoma cells is enhanced by the radiation. There is a
 remote possibility that body  heating could have contributed to this response,  since the
        10/22/90
7-6

-------
                            DRAFT-DO NOT QUOTE OR CITE


 absorbed RF power is estimated to be at least one-half of the basal metabolic rate of the
 animals.

 7.4.2.2. Modulated Radiofrequency Radiation
    For modulated RF radiation of relatively low power density [i.e., excluding the high power
 electromagnetic pulse (BMP) experiment of Baum et al. (1976)], the mouse experiment
 (Prausnitz and Susskind, 1962) showed a reversible pattern of lymphoma and leukemia which,
 in serial sacrifices, occurred toward the end of the 14-month exposure period but was not
 present in animals after a 5-month recovery period.  However, the short 4.5-minute daily
 exposure was intense enough to raise the body core temperature by 3° C, raising the
 possibility that thermal effects were a contributing factor in the response. The rat study (Guy
 et al., 1985) showed the induction of benign adrenal medulla pheochromocytomas and a
 statistically significant increase in carcinomas of all organ and tissue sites. There was also a
 statistically significant increase in glandular organ carcinomas which was unaccompanied by
 an increase in the incidence of benign tumors of these sites. Such an increase of tumors of all
 types in the aggregate without increase of tumors at any one of the sites is regarded as only
 minimal evidence of a carcinogenic response.

 7.5. SUPPORTING EVIDENCE OF CARCINOGENICITY
    Section 5.11. presents a summary of the effects of EM fields on a variety of basic biological
 phenomena relevant in some way to mechanisms of carcinogenesis; that information is not
 repeated here. ELF fields of relatively high intensity  [producing induced body currents on the
 order of 10 microamperes per square centimeter (^A/cm2)] have enhanced DNA synthesis,
 altered the transcription of DNA into mRNA, altered the molecular weight distribution during
 protein synthesis, delayed the mitotic cell cycle, induced chromosome aberrations, blocked
the action of parathyroid hormone at the site of its plasma membrane receptor, induced
enzymes normally active during cell proliferation, inhibited differentiation and stimulated the
growth of carcinoma cell lines, inhibited the cytotoxicity of T-lymphocytes (which indicates an
impairment of the immune system) in vitro but not in vivo, inhibited the synthesis of melatonin
(a hormone that suppresses the growth of several types of tumors), and caused alterations in
the binding of calcium to brain tissues.  The large variety of exposure conditions and the  lack
of detail on the geometry of the biological samples in these studies precludes a systematic
evaluation of the actual induced currents and field strengths at the tissue and cellular levels
                                         7-7
10/22/90

-------
                           DRAFT-DO NOT QUOTE OR CITE


that are causing these effects.  In addition,  the lack of reproducible results between
laboratories limits the interpretation of much of this literature.
    Radiofrequency fields modulated at the same ELF frequencies that cause some of the
effects noted above also result in the same responses, indicating that the ELF component may
be responsible for these effects.  Unmodulated RF radiation has not caused any of the effects
noted above except for chromosome aberrations. None of the EM fields have caused gene
mutations, sister chromatid exchanges, or DNA damage (as measured by DNA breaks, DNA
repair, or differential killing of repair defective organisms) in a large number of studies.
    Only three ELF effects have been induced at field strengths comparable to the low
environmental exposures at which human cancer has putatively been caused:  (1) the calcium
efflux from brain tissue preparations using 16-Hz electric and  magnetic fields that were
perpendicular to each other, (2) calcium efflux from chick brain tissue after exposure of the
developing embryo to electric fields, and (3) the inhibition of melatonin synthesis by the pineal
gland when a static magnetic field of approximately the strength of the earth's magnetic field is
changed through a small angle of rotation. The results of the first experiment are one of
several phenomena that show a complex dependence on frequency, intensity, and orientation
with respect to the earth's magnetic fields.
    In view of all of the laboratory studies referred to in this section, there is reason to believe
that the findings of carcinogenicity in humans are biologically plausible.  However, the
explanation of which of the biological processes is involved and the way in which these
processes causally relate to each other and to the induction of malignant tumors is not
understood. Most of the effects have been observed at field strengths that are many times
higher than the ambient fields which are the putative cause of the childhood cancers in
residential situations; as a consequence, many of the candidate mechanisms actually may not
be  involved in the response to low environmental fields. The same issue of low-dose
extrapolation arises in the evaluation of chemical agents.

7.6. INTEGRATED DISCUSSION OF SEPARATE CHAPTERS
    The occurrence of cancer in humans exposed to low frequency electric and magnetic
fields has been observed under several different conditions in different populations.
Residential exposure of children, but not adults, has been associated with leukemia,
lymphoma, and brain cancer, and the same sites occur in multiple studies of children. The EM
fields involved in these associations have been magnetic fields rather than electric fields, and
       10/22/90
7-8

-------
                            DRAFT-DO NOT QUOTE OR CITE


 their frequency is made up of primarily 60-Hz components but with inevitable high-frequency
 components introduced by electric motors and the switching of currents on and off. These
 effects have been observed jn children exposed to magnetic fields at or above 2 to 3 mG (O.2
 to 0.3 fiJ).  The types of EM-field exposures in the occupational studies are variable according
 to job category, with some jobs involving pulsed and modulated RF fields as well as 60-Hz
 power frequency components.  It is not possible to rule out the involvement of electric fields in
 these  studies. Exposure to electric fields is extremely variable under ambient conditions and is
 difficult to define.
    There is some, but not well-established, evidence that higher frequency components have
 different effects than 60-Hz components. Electrical switchyard workers exposed to spark
 discharges just before blood samples are taken have chromosome aberrations, whereas
 similar workers with no such exposure do not. Chromosome aberrations have been induced
 by unmodulated RF fields as well as by ELF fields.  A recent preliminary report of an
 epidemiologic study of telephone workers shows a different effect (rare breast tumors in
 males) in people working in the "central office," where switching equipment is typically
 concentrated, than in cable splicers (leukemia) who presumably are exposed to predominantly
 60-Hz  power frequencies. Both  electric and magnetic fields are more effective in inducing
 currents in the body if their frequency is higher, so that if induced currents are responsible for
 these effects, then the higher frequency components are expected to be more effective. If it is
 true that, as two studies indicate, the father's occupation in electrical jobs is a factor in the
 development of leukemia in their children, then the question is raised whether the effect could
 be transmitted via heritable genetic damage in sperm. This speculative hypothesis needs to
 be investigated.                     •
   Although there are several candidate EM field-induced biological phenomena (discussed in
 Chapter 5) that could explain how a cancer response  is caused in the whole organism by
these fields, none of these or any combination of them has been verified experimentally, either
 in laboratory animals or in humans.  Without understanding which combination of these is
relevant to the carcinogenic process, it is not possible to hypothesize what aspect of EM-field
exposure is responsible for biological effects; i.e., frequency, average peak field strength,
duration, time of day, whole-body average versus local critical site, electric versus magnetic
fields, orientation with respect to the earth's static magnetic field. The choice of which aspect
of the fields is the most relevant could be based on either knowledge of the correct
mechanism of action or on empirical epidemiology correlations, but,  given the current lack of
information, neither method can serve as a basis for a dose-response analysis.
                                         7-9
10/22/90

-------
                           DRAFT-DO NOT QUOTE OR CITE


   There are several indications that EM fields might contribute to the induction of cancer via
indirect mechanisms, in contrast to a direct mutagenic action of DMA as is the case with
nitrosamines, polycyclic aromatic hydrocarbons, or other DNA-alkylating agents.
   First, EM fields have not caused gene mutations in any of the large number of experiments
carried out with both ELF and RF fields (Section 5.1.3).
   Second, there is no indication from the animal studies that RF fields cause a de novo
induction of tumors (Section 4.1.2.7).  On the contrary, the mouse experiments by Szmigielski
et al. (1982) (see Section 4.4) indicate that unmodulated RF radiation acts as a growth
stimulator for pre-existing tumors. The same growth-stimulating or promotion characteristics
of RF fields could explain the induction of glandular tumors in the Guy et al. (1985) lifetime rat
study of modulated RF radiation (Section 4.1.2.7), since many of the glandular tumors in that
study had a naturally high spontaneous incidence.
   A third factor indicating that there may be multiple causes of carcinogenic action is that
120-Hz pulse-modulated 2450 MHz radiation can act as an initiator of phorbol ester-promoted
cell transformation in mouse embryo cell cultures (Section 5.3.3).
    Finally, there are possible cancer induction mechanisms mediated by the central nervous
system causing neuroendocrine  influence on cellular proliferation (Section 5.7.5).  These
mechanisms involve possible extremely sensitive detection of magnetic fields by the retina
(Section 5.10.4) with resulting neural control of pineal melatonin activity, which in turn
modulates estrogen and prolactin levels in the blood supply to the breast, prostate, and other
hormonally sensitive tissues (Section 5.7.1). Other speculative chains of events could be
fabricated from the existing information in this document; this one is mentioned here only as
an example that there are many possible explanations,  but no verified ones.
    In view of these indications, it is likely that if EM fields do contribute to the induction of
cancer, the causal relationship will probably turn out to be dependent on many chemical
factors and physiological conditions that are currently poorly understood.
    There are two issues in the hazard evaluation of chemical carcinogens that are analogous
to issues for EM fields.  It may be helpful to  explore whether the assumptions and conventions
 developed for chemicals are applicable to the EM fields problem.
    One analogy is that EM fields are mixtures consisting of several frequencies, intensities,
 and combinations of electric and magnetic fields that (for ELF frequencies) occur  in arbitrary
 proportions. One approach to the assessment of chemical mixtures is to identify  hazardous
 components of the mixture apd, assuming additivity of  components, consider the risk of the
 mixture to be proportional to the risk of the  hazardous components.  If this concept were
        10/22/90
7-10

-------
                             DRAFT-DO NOT QUOTE OR CITE

  applied to the EM fields problem, then magnetic fields from 60-Hz power usage in the home
  would be the only "hazardous component" identified, although there is some indication that
  occupational exposures of adults to mixed fields may cause the same effect. Laboratory
  studies under relatively controlled conditions of exposure have not been able to test the
  additivity assumption for EM-field components or for chemical  components except for a few
  rare cases, but one feels more comfortable with the latter.  With chemical agents, the basic
  phenomenon is ultimately some chemical reaction, which is expected to have additive
  properties at low enough concentrations, or at least to be monotonic in the sense that more
 chemical produces a greater effect. With EM fields, however, the ultimate causative interaction
 between fields and biological systems is unknown, and there is certainly no additivity with RF
 and ELF fields, or with ELF electric and ELF magnetic fields.  The consequence of not being
 able to add the risks for different exposures is that the effects for each combination must be
 investigated and assessed separately.
    Another analogy is the similarity between the "biologically effective dose" for chemical
 agents and the critical electrical measure of tissue "dose" which causes the effect for EM fields.
 For chemical agents the relationship between "administered dose" and "effective dose" has
 been studied occasionally, but only rarely. In the absence of this information, the default
 position for chemical agents has been to assume a linear relationship. Then there are several
 unresolved questions in determining whether the biological effect is proportional to the
 "effective dose." These questions  arise when, as is usually the case, the mechanism of action
 is not known. Here again the linearity assumption is made in the absence of knowledge, and
 the overall default position is that the adverse effect is proportional to the administered dose of
 the chemical agent. For EM fields, the "tissue doses" could be calculated, typically with great
 difficulty and uncertainty, but the same type of questions need to be answered about which of
 these dose metrics are relevant for EM-field exposure. As with chemical agents, the choice of
 a candidate mechanism of action dictates which tissue dose metric is appropriate, and there
 could be several mechanisms for each of the administered agents. For EM fields, the default
 linearity assumption may not be appropriate, basically because  there are frequency and
 intensity "windows" of activity for more than one EM field-induced biological effect, and such
 "window" interactions cannot be ruled out as contributory to cancer causation. On the other
 hand, ambient human exposure involves a wide range of static magnetic fields, frequencies,
 intensities, and intermittent exposures, so that narrow windows of response may not be the
dominant practical exposure consideration, and a higher average field may simply increase the
probability that a "windowed" condition will occur.
                                         7-11
                                                                             10/22/90

-------
                           DRAFT-DO NOT QUOTE OR CITE


   In conclusion, several studies showing leukemia, lymphoma, and cancer of the nervous
system in children exposed to magnetic fields from residential 60-Hz electrical power
distribution systems, supported by similar findings in adults in several occupational studies
also involving electrical power frequency exposures, show a consistent pattern of response
that suggests a causal link.  Frequency components higher than 60 Hz cannot be ruled out as
contributing factors. Evidence from a large number of biological test systems shows that
these fields induce biological effects that are consistent with several possible mechanisms of
carcinogenesis. However, none of these processes has been experimentally linked to the
induction of tumors, either in animals or humans, by EM fields. The particular aspects of
exposure to the EM fields that cause these events are not known.
    in evaluating the potential for carcinogenicity of chemical agents, the U.S. Environmental
Protection Agency has developed an approach that attempts to integrate all of the available
information into a summary classification of the overall weight of evidence that the agent is
carcinogenic in humans. At this time such a characterization regarding the link  between
cancer and exposure to EM fields is not appropriate because the basic nature of the
interaction between EM fields and biological processes leading to cancer is not understood.
For example, if induced electrical currents were the causative factor, then exposure to electric
as well as magnetic fields would be important and the effect would be more severe as the
frequency increases.  But if the direct magnetic field interaction were the critical factor, then the
 ambient static magnetic field as well as the alternating magnetic field would be  critical and the
 effect may be confined to specific frequencies, resulting in an extremely complicated
 dose-response relationship. In addition, if they were shown to be causative agents, these
 fields probably exert their effects via other chemical and environmental factors rather than
 directly causing events known to be causally related to the carcinogenic process, having the
 direct property of causing cancer, as with genotoxic chemical agents.
     Because of these uncertainties, it would be inappropriate to classify the carcinogenicity of
 EM fields in the same way as the agency does for chemical carcinogens. As additional studies
 with more definitive exposure assessment become completed, a better understanding of the
 nature of the hazard will be gained. With our current understanding we can identify 60 Hz
 magnetic fields from power lines and perhaps other sources in the home as a possible, but not
 proven, cause of cancer in people.  The absence of key information summarized above makes
 it difficult to make quantitative estimates of risk.  Such quantitative estimates are necessary
 before judgments about the degree of safety or hazard of a given exposure can be made.
 This situation indicates the need to continue to evaluate the information from ongoing studies
        10/22/90
7-12

-------
                            DRAFT-DO NOT QUOTE OR CITE
and to further evaluate the mechanisms of carcinogenic action and the characteristics of
exposure that lead to these effects.
                                                ti- U.S. GOVERNMENT PRINTING OFFCE: 1990-548-187/2050^
                                        7-13
                                                                             10/22/90

-------

-------