Epidemiological Studies of the Possible Adverse Health Effects of Pulsed Radar Emissions
In this chapter we discuss a limited range of studies that addressed potential health effects among populations exposed to pulsed-radiofrequency wavelengths. Ideally, the best studies would be carried out in the fields of radar stations. Unfortunately, such studies are very limited in number and quality. One of the reasons is the difficulty of determining the precise dosage of the radar waves received by the population over a long period of exposure. As a result, epidemiological studies in such populations are few in number. When such studies are undertaken, they usually occur after long exposures and often within the context of occupational exposures. A large number of such studies deal with implied exposures due to assignment on death certificates such as “electrician.”
We include in this chapter those studies which have been undertaken in situations where there have been serious concerns about possible health effects, such as in the case of the exposure in the Moscow Embassy (Lilienfeld and others 1978). In other instances there has been a defined cohort exposed to radar emissions of various types with follow-up for specific outcomes (Hill 1990; Robinette and others 1980; Groves and others 2002). We also include studies of Polish military personnel and of the region that was irradiated by the Skrunda radar facility, which report health effects of radar emissions not previously or subsequently confirmed by other observations.
EXPOSURES AT THE U.S. EMBASSY IN MOSCOW
A substantial RF exposure of employees working in the U.S. Embassy in Moscow (Lilienfeld and others 1978) occurred between 1953 and 1976. Those employees were exposed to microwave energy (500 MHz-3.0 GHz) at levels of 5
µW/cm2 to 15 µW/cm2 for periods of 9 to 18 hours per day. To determine the possible effects on morbidity and mortality, the health status of 1827 Department of State employees at the Moscow Embassy was compared with the health status of 2561 employees and their dependents at other U.S. Embassies in neighboring Eastern States using health records, health questionnaires, and death certificates. It was concluded that personnel working at the U.S. Embassy in Moscow from 1953 to 1976 suffered no ill effects from microwaves beamed at the building.
EXPOSURES AT THE MIT RADIATION LABORATORY
A cohort of 1592 white male physicists and engineers participating in a radar research and development project during Word War II was evaluated for mortality (Hill 1990). It was not possible to assign precise personal exposures although classes of exposure were assigned. Overall and cause-specific mortality, using standardized mortality ratios, was examined up until 1980 for an average survival of 36 years. Comparisons were made with all U.S. white males, and with a cohort of physicians. The MIT Radiation Laboratory experienced a lower than expected mortality based on a comparison with age-adjusted mortality in U.S. males, as well as with the comparison with the cohort of physicians. Among the different causes of death there was a slight elevation of rates for Hodgkin’s disease and cirrhosis of the liver, neither of which reached statistical significance.
TWO STUDIES OF A U.S. NAVY COHORT
A total of 40,890 U.S. Naval personnel, believed to be exposed to radar energy while aboard navy ships, was studied for potential adverse health effects. No RF measurements were taken of the cohort, consisting only of men, who were categorized into high exposure and lower exposure based only upon job classification. This cohort was the subject of two separate investigations, including the original report by Robinette and others (1980) and subsequently by Groves and others (2002), which included an extended follow-up period.
In the initial investigation (Robinette and others 1980), the individuals were from U.S. Navy technical schools (1950-1954). Follow-up in this study was via hospital records, the Social Security Administration, and the National Death Index. Cause of death was obtained from death certificates. In the first study, there was a finding of significantly increased trauma mortality in the group with significantly elevated microwave exposure. On further examination, this increased mortality was the result of military aircraft accidents. Total mortality was not significantly different in the two groups. While mortality from cancer, stroke, chronic nephritis, influenza and pneumonia, and liver disease was elevated among the group with the higher level of exposure, none reached statistical significance.
The subsequent study of this cohort by Groves and others (2002), including a follow-up period of approximately 40 years, provided an opportunity to identify
potential delayed health effects. In this study, comparisons of mortality were made between the two exposure groups and age-specific mortality rates in the U.S. white population. A total of 8393 deaths were identified for a cumulative crude mortality rate of 20.7%. The result for the comparison of the low radar exposure with that of the U.S. population was 0.80 (0.78-0.82) and for the high radar exposure rate it was 0.69 (0.67-0.71). This indicates that the whole cohort was in better health than the whole U.S. white male population of the same age distribution, an effect often referred to as the “healthy worker effect.” The only causes of death with a statistically significant elevated SMR were for war-related deaths and for accidents involving air transportation in the high-exposure cohort (49.6 [27.5-89.6]) and (4.74 [3.89-5.76]), and war injuries were also significantly elevated in the low-exposure cohort (9.13 [2.28-36.5]). Non-significant elevations in the SMR were observed for lymphocytic leukemia (1.12 [0.69-1.83]) and non-lymphocytic leukemia (1.24 [0.90-1.69]) in the subgroup defined as high-potential exposure. There were some weaknesses in the studies consisting of lack of exposure measurements and potential losses in the follow-up which would have a tendency to lower SMRs. The authors of the more recent study conclude, “Radar exposures had little effect on mortality in this cohort of U.S. Navy veterans”.
RADIO LOCATION STATION AT SKRUNDA
The Radio Location Station at Skrunda in Latvia was part of the Early Warning System of the Soviet Union (Romancuks 1996). The station had 2 pulsed-radar systems operating at 152 to 162 MHz, at a power of 1250 kW. The pulse duration was 0.8 msec with an interpulse interval of 41 msec and a pulse repetition rate of 24.5 Hz. The average-power level at nearby residential locations was less than 10 µW/cm2 and the peak power level at these locations was 500 µW/ cm2. The radar systems were equipped with a pair of antennas measuring 250 × 12 m, which were angled slightly towards each other.
Selected children in the vicinity of the Skrunda Station underwent psychophysical testing. A total of 609 children aged 9-18 living within 20 km of the Skundra station were assumed to be exposed. The test battery (Polytest 8802), requiring 70 minutes to complete, including 11 tests that measure tapping rates, reaction to visual and auditory stimuli, attention, and memory (Kalmins and others 1996), was administered to 609 children. Among participants, 224 children lived where they were directly exposed, 385 children lived nearby but were not directly exposed, and an additional 357 children lived in non-exposed areas and served as controls. The authors concluded that children living in the exposed area performed worse on the tests than children living in the nearby area and even worse than the children in the control group. Many of the differences in performance were not statistically significant and many of the differences were found to show correlation differences of 0.25 and were barely significantly different from 0.0. In some categories performance on psychophysical tests appeared to be
statistically impaired for children living in the radar line of sight compared to those living behind the radar and those living in a nearby community. The impairment occurred in all ages tested (9-18 years) in a tapping test, but in other cases the effect was only in one sex or in certain age groups. For instance, reaction times to auditory stimuli were significantly impared in ages 9-12 years, while ages 13-14 years showed no statistical difference on this test. The authors note that “evidence for a factor other than electromagnetic fields having caused the observed results was not found, but its existence cannot be ruled out …”. The small populations exposed in the Skrunda radio stations do not make it likely that further epidemiological studies in this population can be mounted.
POLISH MILITARY STUDIES
Over the years 1971-1985 there was a study of cancer morbidity (first diagnosis of cancer) in a population consisting of career personnel in the armed forces of Poland (Szmigielski 1996). This population varied over the years from 118,500 to 142,200 with a mean of 127,800 (±9620). Based on military records, the authors designated a fraction of this population, which varied between 3400 and 3700 persons (mean 3720 ± 360), as exposed to RF/microwave radiation. This population was divided into 4 age-groups (20-29, 30-39, 40-49, and 50-59 years). The study consisted of a series of annual cross-sectional assessments of morbidity in a series of populations. This approach would be sensitive to changes in age makeup in each of the population groups over the period of study, or to possible changes in the exposure characteristics, but it also makes it difficult to compare the results of this study with conventional cohort studies that follow a particular cohort over a period of years. The exposed part of the population was exposed to pulse-modulated 150-3500 MHz radiation with 80-85% not exceeding 0.2 mW/cm2 and 15-20% at 0.2-0.4 mW/cm2. Individual exposures could not be determined. The total incidence of cancer diagnoses was determined by age groups for the total population and for the exposed population. Cancer occurrence was determined for the following sites: oral cavity, pharynx, esophagus and stomach, colorectal, liver/pancreas, lung, bones, skin including melanoma, kidney and prostate, brain and nervous system, thyroid, and hematopoietic and lymphatic systems. Cancer incidence in the non-exposed population was used as the expected rate for the total population. The overall cancer morbidity rate was 57.6 per 100,000 per year and the cancer morbidity rate was 119.1 per 100,000 per year in the exposed population. The gross observed/expected ratio therefore was 2.07 (1.12-3.58). The study reported significantly increased cancer morbidity in all exposed age groups except in 50-59-year-old subjects exposed to radiation. The largest subclass increases were from chronic myelocytic leukemia, acute myeloblastic leukemia, and non-Hodgkin’s lymphoma (13.9, 8.62, and 5.82, respectively).
The results of the studies cited here are mixed in characteristics. On the one hand are the results of studies of populations that are reasonably described in terms of radar exposure such as the population in the U.S. Embassy (Lilienfeld and others 1978). They are reinforced by similar exposures in the MIT radiation laboratory (Hill 1990), which also indicate a lack of health effect. In the MIT study the exposure characteristics were doubtful, but the duration of the exposure was early in 1942 until 1980. In all these studies there was uncertainty in the exposure classification. The total exposure in terms of people was small, but the follow-up was adequate to determine that there were negligible effects of exposure.
The next investigation was of a large cohort in the U.S. Navy (40,890 people) that was followed-up by two study groups that were relatively independent of each other (Robinette and others 1980; Groves and others 2002). In the first study, follow-up was from 1950 to 1974. The group was divided up into two cohorts, the first of which comprised 20,109 individuals who, based on their job description, had a higher exposure than the second cohort of 20,780 individuals, who received a much lower exposure. There were no direct exposure measurements made in either group after the initial assignment of dose. Neither investigation identified statistically significant associations with exposure and adverse health-related outcomes.
The committee has reviewed two studies that were connected with the Skrunda Radio Location System. This system was operating at a slightly lower frequency than PAVE PAWS but at a higher power level of 1250 kW. The first study reported a lower annual tree-ring width following the startup of the facility in 1971 and is reported in a previous chapter (Chapter 6 on biological effects). A second report addressed performance in a psychophysical test administered to youths of 9-18 years. The annual tree-ring width appeared affected (diminished) after the Skrunda Radio Location System became operational. In some catagories performance on psychophysical tests appeared to be statistically impaired for children living in the radar line of sight compared to those living behind the radar and those living in a nearby community (all ages tested 9-18 year-olds in a tapping test and ages 9-12 in reaction time to auditory stimuli; ages 13-14 showed no statistical difference on this latter test). Other tests were not statistically conclusive and the overall health impact on the children living in the line of sight to the beam, if any, is unclear.
The last study we report on is by Szmigielski (1996). Szmigielski studied cancer morbidity (first-cancer morbidity) in a varying population of 118,500 to 142,200 with a mean of 127,800 (127,800 ± 9620). The authors designated a fraction of this population as exposed to radar 150-3,500 MHz with 80-85% not exceeding 0.2 mW/cm2 and 15-20% exposed to 0.2-0.6 mW/cm2. The exposed population consisted of 3400 to 4600 men with a mean of 3720 ± 360. The cancer
incidence in the non-exposed population was used as the background rate. For the whole population the SMR was 2.07 (1.12-3.58) for the observed/expected ratio, and the highest subclasses were for chronic myelocytic leukemia, acute myeloblastic leukemia, and non-Hodgkin’s lymphoma (13.9, 8.62, and 5.82). The interpretation of this study and its relevance to PAVE PAWS is problematic because the radar exposure is 0.17 × 3,720 (± 360) × 0.4 mW/cm2 or 250 µV/cm2, an amount considerably above the power densities experienced by the PAVE PAWS public. Because the 250 µV/cm2 is elevated above the comparison levels and because the measurement of SMR is different from most of the other studies we have, the committee concludes that this finding is sufficiently different from the PAVE PAWS findings in Chapter 9 of this report that additional data would be required to determine its significance with respect to possible health effects resulting from exposures to PAVE PAWS.
Groves, F.D., W.F. Page, G. Gridley, L. Lisimaque, P.A. Stewart, R.E. Tarone, M.H. Gail, J.D. Boice, and G.W. Beebe. 2002. Cancer in Korean War technicians: mortality survey after 40 years. Am J Epidemiol 155:810-818.
Hill, D. 1990. A longitudinal study of a cohort with past exposure to radar: the MIT radiation follow-up study (meeting abstract). Bioelectromagnetics Society, 12th Annual Meeting 10-14 June, San Antonio, TX Abstract No. A-2-6 p. 6.
Kalmins, T., R. Krizbergs, A.A. Kolodynski, and V.V. Kolodynska. 1996. Motor and psychological functions of school children living in the area of the Skrunda radio location station in Latvia. Sci Total Environ 180:87-93.
Lilienfeld, A.M., J. Tonascia, S. Tonascia, C.H. Libauer, G.M. Cauthen, J.A. Markowitz, and S. Weida. 1978. Foreign Service Study: Evaluation of Foreign Service and Other Employees from Selected Eastern European Posts. NTIS Document No. PB-28B 163/9GA p. 436.
Robinette, C.D., C. Silverman, and S. Jablon. 1980. Effects upon health of occupational exposure to microwave radiation (radar). Am J Epidemiol 112:39-53.
Romancuks, A. 1996. Measurement of the intensity of the electromagnetic radiation from the Skrunda Radio Location Station. Sci Total Environ 180 (1):51-56.
Szmigielski, S. 1996. Cancer morbidity in subjects occupationally exposed to high frequency (radio frequency and microwave) electromagnetic radiation. Sci Total Environ 180:9-17.