TABLE 24: RFR AND CANCER INCIDENCE
| Authors | Effects Sought or Examined | Exposure Modality | Effects Reported | Notes & Comments | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Lester and Moore (1982a) | Increased cancer incidence in areas surrounding U.S. Air Force bases (AFBs).
The authors found 92 counties that had 1 or more AFBs, and matched each county with one of about the same population that had no AFBs. | RFR from the radars at AFBs. (Specific characteristics not given.) | The authors used a 1-tailed test for correlated proportions. They concluded that AFB counties, when compared with population-matched nonAFB counties, had a significantly higher cancer mortality incidence for the period 1950-1969, and suggested that the higher incidence was due to RFR-exposure from the radars. | The authors indexed the data on cancer mortality in each AFB and nonAFB county in terms of severity relative to cancer mortality in the general U.S. population. Their findings (if valid) do not confirm that the increased cancer mortality is associated with RFR-exposure, but only that such mortality may be correlated with the presence of an operating AFB.
| Polson and Merritt (1985)
| These authors carried out an independent reanalysis of the raw data provided by Dr. Lester on this study. The review of the raw data showed several inconsistencies, mostly in AFB location and counting, which were corrected.
| See Lester and Moore (1982a).
| Statistical treatment of the corrected data base showed that the incidences of cancer mortality in counties with AFBs for the period 1950-1969 did not differ significantly from the incidences in the nonAFB counties most closely matched in population.
| In a response to Polson and Merritt (1985), Lester (1985) took issue with several aspects of the reanalysis and did not accept their finding of no significant differences in cancer-mortality incidence.
| Lester and Moore (1982b)
| The authors sought a possible geographic pattern of cancer incidence in Wichita, and whether specific RFR sources could be identified and related to any such pattern. Their analysis was based on morbidity data for first diagnosed cancer cases of Wichita residents for 1975, 1976, 1977.
| The authors suggested that the radars at Wichita Mid-Continent Airport and McConnell AFB were the primary RFR sources, and they presented a formula relating incidence of cancer to terrain and exposure via line-of-sight transmissions to relatively high locations.
| The authors selected 76 of the city's 94 census tracts, and assumed the incidence rates for the 76 census tracts to be the ratio of the number of cases in each tract to its population. They also obtained mortality data for all cancer deaths in those years, and analyzed incidence rates, age, economic status, male/female ratio, and race to obtain a correlation matrix for those 76 census tracts. Their finding was that cancer incidence in Wichita appears to be related to the probability of exposure to the RFR from the radars at the two airports.
| This paper contains a number of flaws, among them the assumption that the population is exposed only to the RFR from radars at the two airports adjacent to the city, without providing measurements to support the assumption or any indication that the scan sectors of such radars were considered. Moreover, their exposure formula was not based on the physical laws of RFR propagation, most particularly the inverse-square-law of attenuation with distance. Thus, little if any credence can be given to their finding.
| Milham (1982)
| Leukemia in men from presumed occupational exposure to electric and magnetic fields in Washington State for the years 1970-1979.
| The author determined the number of deaths from each of 158 causes in each of 218 coded occupations. No exposure data were given; assumption of field exposure was based on 11 of 218 occupational codes.
| The PMRs for 7 of the 11 assumed field-exposure occupations were marked by the author as not statistically significant. For the remaining 4 occupational codes, (electricians, television and radio repairmen, power-station operators, and aluminum workers), 3 of the PMRs for "all leukemia" were marked significant; the exception was for the television and radio repairmen. Three of the PMRs for "acute leukemia" in those 4 occupations were also marked significant, but the nonsignificant exception was for power-station operators.
| The author used the ratio of the number of deaths from each cause in each occupation to the expected number of deaths from that cause, multiplied by 100, called the "proportionate mortality ratio" (PMR). He surmised that the leukemia findings were associated primarily with exposures to large dc currents and high alternating electric and magnetic power fields.
| Liburdy (1982) disagreed with those findings, citing several references that do not support the leukemia association. Wright et al. (1982)
| Leukemia in men from presumed occupational exposure to electric and magnetic fields during 1972 to 1979, as determined from the Cancer Surveillance Program in Los Angeles County.
| No exposure data were given.
| A trend toward a higher risk of leukemia from occupational exposure to electric and magnetic fields was reported, with the greatest risk for acute myelogenous leukemia (AML).
| The authors used proportional incidence ratios (PIRs) for the data in the Los Angeles County Cancer Surveillance Program.
| The authors remarked that the subjects may also have been exposed to metal fumes, solvents, fluxes, chlorinated biphenyls, synthetic waxes, epoxy resins, and chlorinated naphthalenes, a point that render the findings highly questionable. McDowall (1983)
| PMRs for occupational mortality in England and Wales during 1970 to 1972 in males 15-72 years of age.
| A case-control study of the deaths of 537 males from AML in the year 1973, with 1074 male deaths from all causes except leukemia as controls. No exposure data were given.
| The PMR study showed that the mortality distribution for all leukemias did not differ significantly from expected, but the PMRs for AML were significantly higher than 100 for self-described electrical engineers, telegraph radio operators, professional electrical engineers, and professional electronic engineers. The results for the case-control study showed significantly elevated relative risks (RRs) for the electrical occupations, with the highest RR for telecommunications engineers.
| It is noteworthy that the PMRs for occupations that may have involved exposure to RFR (as defined herein) ranged from a low of 61 for radio and radar mechanics to a high of 249 for telegraph radio operators. However, as with other studies in which the PMR statistic was used, both the positive and negative findings of this study are open to question.
| Coleman et al. (1983)
| leukemia incidence in males of ages 15-74 years in 10 electrical occupations in South-East England during 1961-1979.
| The authors remarked that the South Thames Cancer Registry for 1961-1979 listed about 30,000 tumors per year in a population of 6.5 million; proportional registration ratios (PRRs) for leukemia were calculated in the males therein. No exposure data were presented.
| Found was a 17% excess of all leukemias in all electrical occupations (PRR=117, p<0.05). However, although 8 of the 10 occupations showed excesses, there was an all leukemia deficit (PRR=22, p=0.07) for "radio/radar mechanics". The PRRs for all of the electrical occupations also showed no overall excess of chronic myeloid leukemia, but a 46% excess of acute lymphoid leukemia, a 29% excess of chronic lymphoid leukemia, and a 23% excess of acute myeloid leukemia (AML). The excesses of AML were in 7 of the 10 occupations, but not all the excesses were significant, and there were no cases in the occupations "radio/radar mechanics" or "professional electronic engineers".
| The authors' null hypothesis was that the proportion of leukemic men in each 5-year-age group in each electrical occupation would be the same as the proportion of the men with leukemia in all occupations.
| The results show no evidence of excess leukemia associated with those occupations that may involve exposure to RFR. Milham (1983)
| The author gathered decedent data for the years 1974-1979 in Washington State, and analyzed the causes of death for 219 male and 51 female occupations.
| His report contained commentaries that described mortality patterns in each of the male and female occupation groupings. The author suggested that some occupations may have involved exposure to electric and/or magnetic fields.
| Some commentaries appear to be highly subjective and to reflect the personal biases of the author. On only one page of the report, 11 categories of workers presumed to be occupationally exposed to magnetic or electrical fields were grouped, and PMRs for two leukemia categories (acute leukemia, and all leukemia) were presented. Significantly high PMRs were reported for 5 of the 22 occupational/leukemia categories, but were probably significant because PMRs in some other categories were abnormally low. Thus, the association found between leukemia and exposure to electric and/or magnetic fields is questionable.
| Described were the occupation codes used in preparing death certificates, the occupation groupings for women, and the occupation indexes for men and women. The author used the "proportionate mortality ratio" (PMR), making their values interdependent. Use of the "standardized mortality ratio" (SMR) would have been more appropriate; it indicates the percentage of deaths for each cause to the expected number of deaths from that cause, independent of any other SMR.
| Milham (1985)
| Mortality data on the male members of the American Radio Relay League (operators of amateur radios) listed as decedents in QST, the monthly magazine of the League.
| No specific exposure data were presented. The author noted that some ARRL members may also have been in occupations involving RFR-exposure.
| The PMR was 281 for acute, chronic, and unspecified myelogenous leukemia; the PMR for monocytotic leukemia was only 77 (well below 100), and there were no cases of lymphatic leukemia. Thus, the PMR for all leukemias was 191 (24 deaths versus 12.6 deaths expected, p<0.01).
| Death certificates for 1971-1983 were obtained for 280 decedents in the State of Washington, and information on the age, date, and cause of death was obtained for 1411 decedents in California. PMRs were calculated for specific and all leukemias. As noted previously, the use of the PMR statistic is questionable.
| Wangler et al. (1985)
| Coleman (1985) For several reasons, Wangler et al. (1985) took issue with the Milham (1985) finding of a higher risk of leukemia in amateur radio operators.
| Coleman (1985) did not agree with most of the comments by Wangler et al. (1985). See Milham (1985).
| See Milham (1985).
| Wangler et al. (1985) cited a 1980 survey in Canada and the U.S. that showed that typical amateurs spent 6.1 hours per week on that activity, much of which was highly variable and most of which was listening, with the remainder involving intermittent transmissions at low average radiated powers. They also questioned the statistical treatment used.
| Milham (1988a)
| Mortality in amateur radio operators due to RFR-exposure from their transmitters.
| The author obtained the names of 67,829 males licensed in California and the State of Washington, and searched them for deaths during 1979-1984, which yielded 2,485 decedents taken to have had 232,499 person-years at risk. Specific frequencies were not indicated, but presumably were in the FCC-approved amateur bands.
| Unlike in his previous studies, Milham (1988a) used the SMR. For "all malignant neoplasms", the SMR was 89 with an 82-95 CI, a significantly lower death rate than for the general population. The SMR for the subcategory "other lymphatic tissue" of malignant neoplasms was 162 with a CI of 117-218, a significant excess. Also examined were 9 leukemia subdivisions or sub-subdivisions. Of 36 deaths, 15 were for "acute myeloid" leukemia versus 8.5 of the expected 29. The SMR was 176 with a CI of 103-285, also a significant excess.
| An SMR that exceeds 100 with a 95% confidence interval (CI) entirely above 100 is deemed a significant increase; an SMR less than 100 with a CI below 100 is a significant decrease.
| Little credence can be given to the significant excesses found, in view of the small numbers of deaths relative to the actual and expected totals. Milham (1988b)
| Mortality in amateur radio operators in the 5 FCC-license classes: novice, technician, general, advanced, and extra.
| Specific frequencies were not indicated, but presumably were in the FCC-approved amateur bands.
| The numbers of deaths and SMRs in each class from all causes, all malignant neoplasms, and various specific types of malignancies were tabulated. The SMRs for all death causes were significantly below 100. For deaths from all malignant neoplasms, the SMRs for all 5 license classes were below 100, but the collective SMR for the malignant-neoplasms category was 89. The SMRs for lymphatic and hematopoietic neoplasms exceeded 100, but the only significant excess was for the technician-license class. Also, except for the novice class, the SMRs for multiple myeloma and other lymphomas exceeded 100, but the excess was significant only for the general-license class.
| No confidence limits were shown, but the significant SMRs were so marked. In both the technician-license and general-licence classes, for which significant excesses were reported, the numbers of deaths were small: 18 of 409 deaths in the former class and 15 of 862 deaths in the latter class.
| Pearce et al. (1985)
| Pearce (1988) From a case-control study of 546 male leukemia patients for association with various agricultural occupations in New Zealand (NZ), Pearce et al. (1985) sought an association with several specific occupations that may have involved exposure to electric and/or magnetic fields.
| The patient data for the study were drawn from the NZ Cancer Registry. Tabulated from the case-control study were leukemia odds ratios (ORs) and 95% confidence limits (CIs) for patients in 10 occupational classes, a total of 18 of the 546 cases. The controls were 43 of the 2,184 controls in thelarger study. No data were given on actual exposures to electric or magnetic fields.
| Significant leukemia excesses were shown for electronic equipment assemblers, and in the radio/television repair category. The authors suggested that the excesses were due to exposure to nonionizing radiation or to metal fumes and other substances used in electrical components or their assembly. Several corrections were given in Pearce (1988), primarily that the excesses should have been ascribed to the radio/television repair and electricians occupations, and that there were no cases in the electronic equipment assemblers category.
| The findings are questionable from a statistical viewpoint, primarily because of the small numbers of cases.
| Thomas et al. (1987)
| Brain tumor mortality from occupational exposure of men who had died at age 30 or older from brain tumors or other tumors of the central nervous systems.
| Job categories or job codes were assigned to occupational histories to assess presumed exposure to: RFR, lead, and soldering fumes.
| Data were obtained about men who had died from tumors during a 3-year period. One control for each case was selected from men matched in age and year of death and area of residence, but who had died from other than brain tumors. Of 435 cases, 300 had astrocytic tumors, 90 had other types of tumor cell, and 45 had unknown types of tumor cell. Highest relative risk (RR) was for the combined categories of engineers, teachers, technicians, repairers, and assemblers. Significant RRs were also reported for those exposed to soldering fumes.
| The authors remarked that RFR was not the responsible agent because those exposed to RFR also may have been exposed to lead, solder fluxes, solvents, and other chemicals.
| Pearce et al. (1989)
| A case-control study of 488 patients in recorded electrical occupations out of 19,904 male patients with all types of cancer.
| No data were given on actual exposures to electric or magnetic fields.
| Of the 22 sites tabulated, 11 had ORs less than 1 and the other 11 had ORs in the range 1.0-1.62. However, all of the latter sites had CIs that spanned 1.0 (nonsignificant), except for leukemia, which had 21 cases versus 13 expected, OR=1.62, CI 1.04-2.52. In the specific occupations, the largest OR was for radio/television repair, in consonance with the previous finding for that occupation, but the OR for electricians was not significant, in contrast to the previous finding. For the five leukemia subtypes studied, the only significant excess was for chronic-lymphatic rather than acute-myeloid leukemia reported in other studies. Also noteworthy was no excess of brain cancer.
| The controls for the patients with cancer at any specific site or cancer type consisted of those with cancer at other sites.
| The findings of this study are questionable for the same reasons as for the previous studies by these authors. In this paper, moreover, they did note that many comparisons involved small numbers. Reif et al. (1989)
| Incidences of brain cancer by major occupational category, and in 3 occupational subgroups.
| The patient data were also drawn from the New Zealand Cancer Registry.
| Highest incidence of brain cancer was for "agricultural workers, forestry workers, and fishermen", with an overall OR of 1.38 and a CI of 1.08-1.77. Only within the third-highest-incidence subgroup, "production workers", were there 8 of 185 cases classified as "electrical workers"; with an OR of 0.78, and a CI of 0.39-1.59.
| As with the other studies of these authors, discussed above, little if any credence can be given to either the positive or negative findings.
| Archimbaud et al. (1989)
| Clinical examination of a cancer patient presumed to have been exposed to high RFR levels.
| From repair of 5 to 6 high-power microwave generators daily for 22 years and testing each by turning it on unshielded for 1 min (totaling about 5-6 min/day).
| The diagnosis was acute myelogenous leukemia (AML), and was ascribed to the RFR-exposure. Various treatments of the patient, including two courses of chemotherapy and a bone-marrow allograft, were ultimately unsuccessful.
| The association with RFR was questioned by Jauchem (1990) and by Hocking and Garson (1990), based on the lack of measurements of the RFR levels or any other details about the patient's exposure conditions.
| Jauchem (1990)
| Hocking and Garson (1990) Archimbaud (1990) Jauchem (1990) and Hocking and Garson (1990) questioned the findings of Archimbaud et al. (1989) for similar reasons. Archimbaud (1990) was a response to Jauchem (1990).
| See Archimbaud et al. (1989).
| See Archimbaud et al. (1989).
| In addition to remarking the absence of measurements of the RFR levels or of any other details about the patient's exposures, Jauchem (1990) also questioned the remarks about genetic activity of microwaves and the interpretation by the authors of the references they cited. In response to Jauchem, Archimbaud (1990) cited the findings of several studies, but those studies were questioned by others. Hocking and Garson (1990) noted the absence of ocular effects in the patient and the negative findings of their cytogenetic study of 38 radiolinemen [Garson et al. (1991)].
| Garson et al. (1991)
| Blood-sample analyses of 38 Australian radio linemen for chromosome damage, and of 38 controls matched in age and Australian State.
| The subjects erected and maintained towers for broadcasting, telecommunication, and satellite RFR transmission. The exposure frequencies ranged from 400 kHz to 20 GHz; the levels were below the ANSI 1982 guidelines and the Australian 1985 occupational exposure guidelines.
| From a statistical analysis based on the Poisson distribution, most of the ratios of exposed-to-control rates exceeded 1.0 slightly, but all of the 95% confidence intervals spanned 1.0, indicating that the differences between the exposed and control subjects were not significant (p>0.05).
| Duplicate blood samples from each subject were cultured, and 100 metaphases from each culture were counted and assessed for chromatid gaps, chromatid breaks, chromosome gaps, chromosome breaks, and "other" abnormalities. The results for the exposed and control groups were tabulated in terms of the rates per 100 cells for each abnormality, the ratios of the rates for the two groups, and the 95% confidence intervals thereof.
| Hayes et al. (1990)
| A case-control study of 271 patients in 3 medical centers who were newly diagnosed as having testicular cancer, to determine the occupational risk for that type of cancer relative to other types. The controls were 259 patients diagnosed at the same centers with cancer in other than the genital tract.
| In patient interviews about occupational activities, they were asked about exposure to or contact with various materials and equipments as sources of ionizing and/or nonionizing radiation and to pesticides or polycyclic aromatic hydrocarbons.
| A general finding was a higher risk for seminoma (OR=2.8) in all professional subcategories. No increased risk for seminoma was associated with the blue-collar occupations, and the risk was significantly lower for agricultural, forestry-and-fishery, and construction workers. Noteworthy are the results for "exposure to microwaves or other radio waves": OR=3.1 (1.4-6.9 CI) for all testicular tumors. Also, when only the subjects treated in the 2 military hospitals were considered, the OR for all testicular tumors was 3.5 with a CI of 1.3-10.2. Similar results were shown for seminomas and "other" germinal-cell carcinomas. However, for the category "radar equipment", the OR for all testicular tumors was only 1.1 with a CI of 0.7-1.9 (nonsignificant). It is not clear why the authors did not regard the categories "exposure to microwaves or other radio waves" and "radar equipment" as a single one.
| Two of the 3 medical centers were military hospitals, and many of the patients were military personnel on active duty. Results were tabulated by occupation in terms of the odds ratios (ORs), CIs, and numbers of controls, for all testicular cancers combined, and by subdividing the cases into seminomas and "other" germinal-cell carcinomas.
| The authors also presented an independent assessment by an industrial hygienist that did not support the higher risk for the category "exposure to microwaves or other radio waves". Thus, the results do not show an increase of risk of testicular cancer relative to the other types of cancer from occupational exposure to RFR, and do not provide either positive or negative evidence that the other types of cancer suffered by the patients was due to chronic RFR-exposure. |
This is a Department of Defense computer system for authorized use only. DoD computer systems may be monitored for all lawful purposes, including to ensure that their use is authorized, for management of the system, to facilitate against unauthorized access, and to verify security procedures, survivability, and operational security. Using this system constitutes consent to monitoring. All information, including personal information, placed on or sent over this system may be obtained during monitoring. Unauthorized use could result in criminal prosecution.
POC: AFRL/HEDM, (210)536-6816, DSN 240-6816