Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
LOW DOSE EPIDEMIOLOGIC STUDIES 371 7 Low Dose Epidemiologic Studies INTRODUCTION As pointed out in Chapter 1, studies of the imputed effects of irradiation at low doses and low dose rates fulfill an important function even though they do not provide sufficient information for calculating numerical estimates of radiation risks. They are the only means available now for determining that risk estimates based on data accumulated at higher doses and higher dose rates do not underestimate the effects of low-level radiation on human health. As also discussed in Chapter 1, there is good reason to postulate, on the basis of animal studies, that the carcinogenic effectiveness of low-LET radiations is reduced at low dose rates, although the available human data do not suffice to confirm this hypothesis. In its review of low dose studies reported since the BEIR III report (NRC80), this Committee considered populations exposed to radiation from a number of different sources: diagnostic radiography, fallout from nuclear weapons testing, nuclear installations, radiation in the work place, and high levels of natural background radiation. Studies of prenatal exposures to diagnostic x rays are discussed in Chapter 6. DIAGNOSTIC RADIOGRAPHY: ADULT-ONSET MYELOID LEUKEMIA A case-control study of patients with chronic myelogenous leukemia (CML) (Pr88) found that during the 3-20 years prior to their diagnosis, more cases than controls had x-ray examinations of the back, gastrointestinal (GI) tract, and kidneys; and cases more often had GI tract and radiographs of
LOW DOSE EPIDEMIOLOGIC STUDIES 372 the back taken on multiple occasions. A total of 5 cases and 0 controls had GI tract series done on four or more separate occasions, and 11 cases and 1 control had back x rays done on five or more occasions. The odds ratio for exposure to 0-0.99, 1.00-9.99, 10.00-19.99, and â¤ 20.0 Gray in the 3-20 years prior to diagnosis were 1.0, 1.4, 1.7, and 2.4, respectively (p for the highest exposure category, p < 0.05). The association was strongest for the period 6-10 years prior to diagnosis, and the effect of radiation exposure during this period remained significant after consideration of other risk factors in a logistic regression analysis. It was estimated that 23% of cases were attributable to exposure to diagnostic x rays during the period 3-20 years prior to the date of diagnosis of the case (17% during the 6-10 years prior to diagnosis). These recent findings support the association of adult-onset myelogenous leukemia (ML) with certain types of radiographic examinations and with multiple such examinations. The findings are similar to those of the case-control study in New Zealand which found that risk of ML increased with the frequency of x-ray examination of the back and GI tract (Gu64). In the earlier British and tri-state leukemia studies, it was also noted that patients with ML were more likely than controls to have had multiple radiographic examinations (St62, Gi72). A study that was without positive risk findings involved a smaller number of patients (63 patients with ML, including some children) and used nonleukemia patients as controls. The controls were matched to the cases by having visited the same clinic at two distinct times (the year when the patient was diagnosed with ML and the year when the patient first visited the clinic before diagnosis of ML) (Li80). This algorithm for control selection may have introduced a serious bias, since controls selected from among repeat clinical patients are likely to have received more medical attention (including more x- ray examinations) than the general population. Summary The issue as to how much adult-onset ML is attributable to diagnostic radiography is still unresolved. Questions that have been raised include: (1) whether the excess radiography may have been for preleukemic conditions; (2) whether the association between ML and radiography was due to confounding by the conditions for which x rays were taken; (3) whether there were possible sources of bias (selection, recall, etc.); (4) host susceptibility variables; and 5) dosimetry. Two studies that have attempted to evaluate questions 1 and 2 have found little to suggest that much of the observed association was attributable to these sorts of confounding (St62, Pr88); only in the period immediately preceding diagnosis did patients with ML have more x rays because of infections or vague illnesses, and the strongest
LOW DOSE EPIDEMIOLOGIC STUDIES 373 association of ML with radiography was seen not during this period but during the previous period. The positive studies found that the reasons for trunk x rays were distributed similarly in cases and controls, but that for any given reason prompting relatively high bone marrow doses, cases had more repeat exams. Potential bias is always a concern in case-control studies. Another concern and a major limitation of all case-control studies of ML associated with diagnostic radiography is that the dosimetry is uncertain. Doses for a typical examination are, therefore, usually assigned if dose estimates are made at all. A recent dosimetry survey of diagnostic radiographic procedures performed in the United Kingdom shows that the range of doses administered for each type of examination is wide (Sh86). FALLOUT FROM NUCLEAR WEAPONS TESTING In the late 1970s, several studies reported excess cancer, primarily leukemia, among persons who were exposed to fallout from nuclear weapons tests. These included residents of Utah and neighboring states downwind of the Nevada Test Site (NTS) and veterans who had participated in the tests. Estimates of the doses to most organs in both groups was reported to be sufficiently low (less than 50 milliGray for all tests combined) so that no detectable increase in risk would have been predicted on the basis of cancer risk estimates derived from high-dose studies. A possible exception was the dose to the thyroid, which exceeded 500 mGy in some individuals (studies of thyroid tumors are reviewed in Chapter 5). Cancer among Residents Downwind from NTS A survey of death rates from excess cases of childhood leukemia in Utah from 1944 to 1975 was reported in 1979 (Ly79). Based on preliminary data on fallout patterns, the state was divided into two parts; counties with above average and supposedly below average levels of exposure. The time periods considered were chosen so that there were two ''unexposed" cohorts (deaths occurring before 1951 or in children born after 1958) and one "exposed" cohort (those under age 15 at any time from 1951 to 1958). In the "low-exposure" counties, all three cohorts had mortality rates that were comparable to the rates for the U.S. population as a whole. In the "high-exposure" counties, the "unexposed" cohorts had rates that were lower than the rates for the United States as a whole, while the "exposed" cohort had rates slightly higher than U.S. rates and about 2.4 times higher than the rates for the "unexposed" cohorts. Land (La79) subsequently pointed out that death rates for all other childhood cancers in this study showed the opposite pattern; that is, a lower rate for the "exposed" cohort
LOW DOSE EPIDEMIOLOGIC STUDIES 374 in the "high exposure" area. This suggested that the apparent increase in leukemia rates might have been an artifact of diagnostic error. Land et al. (La84) later reexamined the association, using mortality data from the National Center for Health Statistics for 1950 to 1978, and found that while leukemia death rates were about 50% higher in the "exposed" than in the ''unexposed" cohorts, they were not significantly different at a 90% confidence level. Moreover, compared to the "unexposed cohort," rates for eastern Oregon, Iowa, and the total United States were also higher by about the same amount. Land et al. concluded that there was no pattern of excess leukemia mortality that supported a causal association with fallout exposure and that the excess reported reflected an anomalously low rate in southern Utah during the period 1944 to 1949. Both studies suffer from the fact that comparisons are based on aggregate groups and may not reflect any associations among individuals. Furthermore, Beck and Krey (Be83) have since shown that the levels of fallout were not, in fact, higher in the "high exposure" counties than in the "low exposure" counties, contrary to the original supposition. A case-control study is currently in progress to examine the association between leukemia and individual estimates of doses, which includes 1,179 patients with leukemia and 5,380 people who died from other causes among Mormon residents of Utah from 1952 to 1981. In 1984, Johnson (Jo84) reported on results of a retrospective cohort study of cancers in Mormon families who were listed in both the 1951 and 1961 telephone directories for towns in southwestern Utah and neighboring parts of Nevada and Arizona. Self-reports of cancer and other diseases among those that could be located in 1981 were obtained by volunteers. A total of 288 cases of cancer were reported in this group, compared with 179 cases of cancer expected on the basis of rates for all Utah Mormons. The major excesses (observed/ expected) were for leukemia (31/7.0), thyroid cancer (20/3.1), breast cancer (35/23.0), melanoma (12/4.5), bone cancer (8/0.7), and brain tumors (9/3.9). A subgroup of 239 persons who reported acute effects from fallout exposure showed even higher rates of cancer (33 cases of cancer at all sites observed compared with 7.1 expected cases). The cancers reported were not medically confirmed and were likely to have been overreported; Lyon and Schuman (Ly84) point out that the female:male ratio was about 70% higher in this study than nationally, suggesting overreporting of female cases, and that only 126 deaths from all causes were reported, whereas at least 192 deaths from cancer would have been expected. Johnson's reliance on data gathered by volunteers appears to be a weak point in his study. Machado et al. (Ma87) analyzed cancer rates from the National Center for Health Statistics for the three counties of southwestern Utah covered by the survey over the periods 1955-1980 for
LOW DOSE EPIDEMIOLOGIC STUDIES 375 leukemia and 1964-1980 for other cancers, and found no excesses of either single or grouped sites, with the exception of leukemia (62/42.8 for people of all ages, 9/3.2 for those from 0-14 years old). Cancer among Participants in Nuclear Weapons Tests U. S. Weapons Tests In 1980, Caldwell et al. (Ca80) reported that among the 3,224 participants of the nuclear test explosion Smoky, nine cases of leukemia occurred through 1977, compared with 3.5 expected cases. In a later report (Ca83), the number of cases of leukemia increased to 10/4.0 and data were provided on cancer at other sites through 1979. The total number of observed cases of cancer was 112, compared with 117.5 expected; there was a significant increase only in leukemia incidence and mortality. In 1984, four cases of polycythemia vera were observed, compared with 0.2 expected (Ca84). Robinette et al. (Ro85) expanded the study to include a cohort of 46,186 participants in one or more of five test series at the NTS or the Pacific Proving Ground (PPG). The excess cases of leukemia among the participants of the Smoky test were confirmed, but only 46 deaths from leukemia were observed in the participants of the other PPG tests, compared with 52.4 expected deaths. No one series showed a significant excess of leukemia, and there was also no consistent excess for any other cancer site. British Weapons Test Darby et al. (Da88) described a cohort study of 22,347 British participants in nuclear weapons tests and related experimental programs in Australia and the Pacific Ocean and 22,325 matched controls. For all causes of death RR = 1.01; for all cancers RR = 0.96. Leukemias and multiple myeloma occurred significantly more often in participants than controls; 22 versus 6 cases and 6 versus 0 cases, respectively. However, for participants at both test sites, the death rates were only slightly higher in participants than expected, based on national rates (SMR = 113 and 111 respectively), while the death rates were much lower than expected in the controls (SMR = 32 and 0, respectively). There was no association with the type or degree of radiation exposure. Canadian Studies Raman et al. (Ra87) carried out a cohort study of 954 Canadian military personnel who had been involved in clean-up operations after nuclear reactor accidents at Chalk River Nuclear Laboratories or who had observed nuclear weapons blasts in the United States or Australia; two matched controls were selected from military records for each exposed subject. No
LOW DOSE EPIDEMIOLOGIC STUDIES 376 differences in cause-specific mortality between cases and controls and no trends by degree of exposure were found; the study size was small, and only very large differences would have been detectable. Leukemia from Global Fallout Archer (Ar87) compared time trends in global fallout across the United States with trends in leukemia rates. Fallout activity was estimated from measurements of beta emissions, airborne particulates, precipitation, and 131I in milk. Fallout appeared to peak in 1957 and 1962. Death rates for acute and myeloid leukemia in children aged 5-9 years rose to an initial peak in 1962 and a secondary peak in 1968; no such pattern was observed for other types of leukemia. Leukemia death rates (for all ages and all cell types) peaked in the decade 1960-1969 and were consistently highest in states with high 90Sr levels in the diet, milk, and bones (based on surveys by the Public Health Services from 1957 to 1970) and lowest in states with low 90Sr levels. The excess of myeloid and acute leukemia deaths was estimated to be about 6.5 per 104 PYGy (based on an estimated average cumulative dose of 4 mGy). Darby and Doll (Da87) reviewed data on childhood leukemia incidence rates and fallout exposures in England and Wales, Norway, and Denmark. Fallout exposures rose rapidly between 1962 and 1965 and declined slowly thereafter. In England and Wales there was about a 10% increase in incidence rates up to 1979, possibly attributable to improvements in diagnosis, whereas incidence rates in Norway and Denmark declined slightly after 1960 during the period of highest population exposure from fallout. The data were thus interpreted to provide no convincing evidence of an increase in incidence that could be attributed to fallout. Summary There are several possible explanations for the cancer excesses that have been reported in the studies cited. The possibility that they may represent chance variations may explain the excess cases of leukemia associated with the Smoky nuclear test, although that test was unusual in ways that are discussed below. Chance may also explain the differences in results from the three studies of leukemia in Utah residents that are based on reported death rates. Although there appeared to be a small excess in southwestern Utah, the causality of fallout exposure cannot be assessed from these studies; the case-control study in progress may help resolve this uncertainty. On the other hand, associations may be real and reflect an underestimation either of the doses or of the risk per unit dose. This may be the case
LOW DOSE EPIDEMIOLOGIC STUDIES 377 for the Smoky nuclear test, which was the highest-yield tower detonation at the NTS. Fallout was particularly heavy, 10 to 20 times greater than at other detonations in this test series (Ha81). The leukemias occurred most frequently in two groups: those near the hypocenter and those ferried in by helicopters within hours of the test. Whether these doses could have been large enough to explain the excess is uncertain. Although there was a wide variation in individual doses among participants at nuclear tests (Ro85), the collective dose could not have been underestimated sufficiently to explain the excess if the risk coefficients derived from high-dose studies were correct and not underestimated. The most likely explanation is that the observed excess cases of leukemia are random overestimates of the risk coefficients. In view of the uncertainty in both sets of estimates, the discrepancy may be small; Archer's estimate of the risk coefficient for leukemia based on his data on global fallout is only slightly higher than that based on data for the atomic-bomb survivors (Ar87). CANCER AMONG INDIVIDUALS NEAR NUCLEAR INSTALLATIONS Nuclear Reactor Accidents It is still too early to assess whether any cancer excess will occur following the Three Mile Island or Chernobyl nuclear reactor accidents. The collective dose equivalent resulting from the radioactivity released in the Three Mile Island accident was so low that the estimated number of excess cancer cases to be expected, if any were to occur, would be negligible and undetectable (Fa81). For the Chernobyl accident, preliminary estimates suggest that up to 10,000 excess cancer deaths could occur over the next 70 years among the 75 million Soviet citizens exposed to the radioactivity released during the accident, against a background of 9.5 million cases of cancer that would occur spontaneously; hence the excess would not be detectable. However, among the 116,000 people evacuated from immediate high-exposure areas in the Ukraine and Byelorussia, there might be a detectable increase in the cases of leukemia and solid cancer (An88, No86). Leukemia among Individuals Near British Nuclear Reprocessing Plants In the district near the Sellafield nuclear reprocessing plant in northern England, 6 leukemia deaths in children aged 0-24 years occurred from 1968 to 1974, compared with 1.4 expected cases (Ga84), and 19 incident cases occurred (10.5 expected) (IAG84). Follow-up studies of two cohorts, one of children born to women resident in the Seascale Civil Parish during 1950-1983 (Ga87a) and one of children born elsewhere but attending schools in
LOW DOSE EPIDEMIOLOGIC STUDIES 378 Seascale were performed (Ga87b). There were five deaths from leukemia (0.53 expected) in the former cohort; no deaths were found in the latter cohort. Within 12.5 km of Dounreay, a nuclear reprocessing plant in northern Scotland, five cases of leukemia occurred (0.5 expected) in children of the same age from 1979 to 1984 (He86). Darby and Doll (Da87) reviewed the data on radiation exposures in Dounreay and concluded that the excess cases were not explainable by radioactive discharges from that nuclear installation. Recently, the influx of a large number of new workers with a concomitant increase in viral infections has been proposed as a causative factor for childhood leukemia in Dounreay and Sellafield (Ki88). Cancer among Individuals Near Other Nuclear Installations Roman et al. (Ro87) described a cluster of 29 cases of leukemia (14.4 expected) in children aged 0-4 years living within 10 km of one or more nuclear facilities in southern England. Hole and Gillis (Ho86) reported a cluster of 31 cases of leukemia (24.3 expected) in children aged 0-14 years living in regions adjacent to four nuclear facilities in western Scotland. These reports are difficult to interpret, owing to the bias due to first observing an apparent cluster and then defining the population at risk and the time period of risk. To avoid this bias, Baron (Ba84) examined cancer mortality in individuals living near 14 nuclear and 5 nonnuclear facilities in England and Wales and found no overall pattern of increasing cancer SMRs in individuals living around the nuclear facilities. A more comprehensive survey of cancer incidence and mortality near nuclear installations for the period 1959-1980 is reported by the United Kingdom Office of Population Censuses and Surveys (Co87a, Fo87). The investigators found significant overall excesses of cancer mortality due to lymphoid leukemia and brain cancer in children and due to liver cancer, lung cancer, Hodgkin's disease, all lymphomas, unspecified brain and central nervous system tumors, and all malignancies in adults; however, the mortality rates in the control areas were lower than expected, and there has not been a general increase in cancer rates in individuals living in the vicinity of nuclear installations. Moreover, there were no consistent, positive, or statistically significant trends in cancer rates with distance from the nuclear installations. Beral (Be87) noted that the incidence of leukemia and all cancers in children were significantly elevated in all exposed areas combined (excluding Sellafield) compared with those in control areas, whereas mortality was not. Cook-Mozaffari (Co87) confirmed the observation, but suggested that the differences may be due to a variation in case registration, possibly owing to social class differences. Recently, Openshaw et al. have demonstrated how cancer clusters
LOW DOSE EPIDEMIOLOGIC STUDIES 379 can be identified objectively using a Geographic Analysis Machine (Op88). This methodology was applied to mortality data for acute lymphoblastic leukemia in children living in the Northern and Northwestern regions of England. Again, Seascale, near Sellafield, in Cumbria was identified as an area having unusually high mortality. Although this type of analysis requires a large computational effort, it appears to free studies of cancer clusters from bias due to the selection of an arbitrary risk area and the effects of arbitrary administrative boundries. Clapp et al. (Cl87) reported excess cases of leukemia and other hematologic malignancies in five Massachusetts towns located near a nuclear reactor. There were 13 excess cases of myelogenous leukemia in males (5.2 expected); the excess cases were mainly in adults, and the possible confounding effect of occupational factors was not considered. No excess cases of cancer have been found around either the Rocky Flats nuclear reprocessing plant in Colorado (Cr87) or the San Onofre nuclear power plant in California (En83). Summary It is difficult to assess the significance of the reports of excess cancer cases near nuclear installations in Great Britain; it appears highly unlikely that all were caused by chance, although the anecdotal nature of some of the observations makes testing of significance impossible. Available radiation dosimetry information also makes it seem unlikely that the excesses or clusters could be explained by the very low radiation exposures. While there has not been a general increase in cancer rates in individuals living in the vicinity of nuclear installations (Co87a,b, Fo87), there does appear to be an excess in the number of cases of childhood leukemia, particularly in individuals living around installations before 1955 and among children born in the region. Whether the excesses will be found to be balanced by a comparable number of deficiences around other nuclear installations, or whether they will prove to occur more consistently than not, are questions calling for further study. EPIDEMIOLOGIC STUDIES OF WORKERS EXPOSED TO LOW DOSE, LOW-LET RADIATION A number of epidemiologic studies of individuals exposed occupationally to low levels of low-LET radiation have been reported. Although, because of limited size and exposure, such studies cannot contribute directly to the estimation of stable radiation risk estimates, they are of use for assessing whether such estimates are substantially in error. Occupational studies have several noteworthy advantages and disadvantages.
LOW DOSE EPIDEMIOLOGIC STUDIES 380 Occupational exposures are generally monitored, but there may still remain considerable uncertainty about exposures measured in early years (see for example In87). Also there may be multiple exposures both to external sources and internal emitters of radiation, and to many potentially carcinogenic chemicals, which may make any specific radiation effect difficult to isolate. Occupational cohorts are usually well defined and their individual members well identified, which facilitates follow-up, but the healthy worker effectâthe tendency for working populations to have lower rates of mortality than those of the general population, primarily because of selection factorsâmeans that comparisons between an occupational cohort and the general population can be difficult to interpret. Epidemiologic Studies of Workers Table 7-1 summarizes the available details on those occupational studies that have been published to date. In these studies, workers were monitored for their exposure to low-LET ionizing radiation. The power of such studies to detect a significant increase in risk depends on the number of observed deaths from the cause of interest. Several of these studies have yet to accumulate a sufficient number of deaths to reach any sensible conclusions relating to individual types of cancer. The most consistent result, observed to date from the studies shown in Table 7-1, is that the risk estimates for all types of cancer combined and for all types of leukemia combined are consistent with the risk estimates provided in the present report, since no studies have reported results which differ significantly from the null. In terms of individual cancers, a significant and dose-related effect has been observed for multiple myelomas in the Hanford study (Gi89) and in the British Nuclear Fuels study (Sm86); in the latter case, data from individuals who received the dose 15 years before death are excluded. A significant excess of prostate cancer has been observed in the United Kingdom Atomic Energy Authority study (Be85), but this excess seems to be associated, in part, with exposures to multiple forms of radiation, including tritium and other internal nuclides. Excesses of prostate cancer were also seen in the British Nuclear Fuels and the Oak Ridge National Laboratory studies (Ch85), but these excesses were not significant and were not dose related. In addition to multiple myelomas and prostate cancer, dose-response effects as a result of exposure to external gamma radiation have been reported for bladder cancer and all lymphatic and hematopoietic cancers by the British Nuclear Fuels study (doses received in the 15 years prior to death are excluded), and for lung cancer by the Oak Ridge Y-12 Plant Study (Ch88). In the latter study, part of the dose to the lung was due to alpha radiation.
TABLE 7-1 Some Epidemiologic Studies of Workers Monitored for External Gamma Radiation Study Type of Operation Years of Last Year of No. of No. of Doseb Total No. Employment Follow-up Individuals in Radiation (mSv) of Deaths Study Workersa United Kingdom Reactor research 1946-1979 1979 39,546 (29,173 20,382 32.4 3,373 Atomic Energy and development males, 10,373 (18,759 males, Authority, UK females) 1,623 females) (Be85) British Nuclear Plutonium 1946-1975 1983 14,000 (11,402 10,157 124.0 2,277 Fuels Limited production, fuel males, 2,598 (Windscale reprocessing, females) Plant, Sellafield waste treatment, Plant), UK fast reactor fuel LOW DOSE EPIDEMIOLOGIC STUDIES (Sm86) fabrication Hanford Site, Reactor research 1944-1978 1981c 44,100 (31,500 36,235 43.6 7,249d Washington and development males, 12,600 State, USA females) (Gi89) Oak Ridge Reactor research 1943-1972 1977 8,375 males 7,778 17.3 966 National and development, Laboratory, plutonium Tennessee, USA production, (Ch85) chemical processing, and separation of isotopes 381
Study Type of Operation Years of Last Year of No. of No. of Doseb (mSv) Total No. Employment Follow-up Individuals in Radiation of Deaths Study Workersa Oak Ridge Uranium 1947-1974 1979 6,781 males 5,278 9.6 862 Y-12 Plant, enrichment, Tennessee, weapon USA (Ch88) fabrication, isotope research Rocky Flats Plutonium 1952-1979 1979 5,413 males 41.3 409 Nuclear weapons Weapons Plant, fabrication USA (Wi87) Atomic Energy Reactor research 1950-1981 1981 13,570 (10,278 7,685 6,626 46.8 males; 946 LOW DOSE EPIDEMIOLOGIC STUDIES of Canada and development males, 3,292 males, 1,239 3.86 females Limited, females) females) Canada (Ho87) Ontario Hydro, Power reactor 1970-1985 1985 23,997 males 5,039 2,860 Canada (An86) operation a No. of individuals reported or estimated to be monitored. b Mean whole body gamma dose per radiation worker. c Deaths occurring in the State of Washington in the years 1982-1985 were also evaluated. d Observed deaths from 1945-1981. 382
LOW DOSE EPIDEMIOLOGIC STUDIES 383 Summary The studies have provided no evidence to date that risk estimates for leukemias and other types of cancer combined are in error, based on extrapolation from high-dose studies. For individual cancer sites, only for multiple myelomas and prostate cancer is there any suggestion that associations were seen in more than one study. In interpreting the latter associations, however, the potential biases discussed in Chapter 1 must be borne in mind. In particular, the problem of multiple comparisons and the tendency for both researchers and editors to focus on positive as opposed to null results. It must also be pointed out that the absence of any associations in a number of studies essentially offers no meaningful evidence, because of the very small numbers of observed deaths. Continued monitoring of these and other occupational cohorts in the future is highly desirable. When possible, standardization and pooling of study results should improve the interpretation and the overall significance of these studies. To date the evidence does not contradict or imply the possible inaccuracy of risk estimates derived from high-dose studies. HIGH NATURAL BACKGROUND RADIATION There are regions in the world where outdoor terrestrial background gamma radiation levels appreciably exceed the normal range (about 0.2-0.6 mGy per year). Such regions exist in Brazil, India, People's Republic of China, Italy, France, Iran, Madagascar, and Nigeria (UN82). Because the total dose rate of low-LET natural background radiation is low, and the lifetime dose of such radiation accumulated by any one person is small (<0.1 Gy), it is difficult to determine whether there are any variations in disease rates associated with changes in natural background radiation levels and, if so, whether such variations are consistent with the health effects estimated by extrapolation from the observed effects of high-dose and high dose-rate exposures. A cautious approach is warranted in the interpretation of geographically based mortality surveys. Although "beneficial" effects of radiation have been alleged on the basis of reduced mortality in high background areas in the United States (Hi81), analyses that include an adjustment for altitude indicate no "beneficial" effects (We86). While mortality rates for both cancer and cardiovascular disease are lower in areas of the United States having high levels of natural radiation, such areas are found primarily in high altitude locations. This apparently "beneficial" effect of radiation may, in fact, be an example of confounding, since conditions of reduced oxygen pressure stimulate a wide array of physiological adaptations, which could themselves be protective (Fr75).
LOW DOSE EPIDEMIOLOGIC STUDIES 384 Recently, childhood cancers have been analyzed in relation to natural radiation levels in England (Kn88), and although reported associations were observed, their interpretation is complicated by the general problems of correlational analyses (see Chapter 1, Epidemiological Principles). Guarapari, Brazil This village of approximately 12,000 inhabitants is located in an area where local soil contains monazite sands, which is the source of gamma and alpha radiation received by the townspeople. The radioactivity in monazite comes primarily from thorium. The average annual absorbed dose to an inhabitant of this area, based on lithium fluoride dosimetry, is about 6.4 mSv (640 mrem), which is roughly 6 times the global average background radiation dose level (excluding radon progeny in the lung) (Ba75). Studies of the health of this population are limited, but a cytogenetic study of 200 individuals, in comparison with a control group from a similar village, reported an increase in the total number of chromosome aberrations (Ba75). Kerala, India The population living along the Kerala Coast of India is exposed to about 4 times the normal level of natural background radiation (excluding radon progeny in the lung). Because of the presence of monazite in the soil (thorium concentration, 8.0-10.5%, by weight), the average absorbed dose rate for the 70,000 people living in the region has been estimated to be about 3.8 mGy/yr (380 mrad/yr) (Go71). The incidence of both Down syndrome and chromosome aberrations has been reported to be increased in this population (Ko76). Yanjiang County, Guangdong Province, People's Republic of China The most extensive observations on the health effects of high natural background radiation have been those made on the mortality experience of the population in Guangdong Province, People's Republic of China. In this area, which contains monazite with high levels of thorium, uranium and radium, individuals are exposed to about 3-4 mSv (300-400 mrem) of gamma radiation per year. The population of this region has been studied extensively for both genetic and carcinogenic effects (We86, Ta86). A sample of 70,000 individuals in this area and a geographically adjacent control area, receiving a normal background of radiation of 1 mSv/year (100 mrem/year), were followed for the period 1970-1985, with approximately 1 million person-years of follow-up in each area.
LOW DOSE EPIDEMIOLOGIC STUDIES 385 On analysis, site-specific, age-adjusted cancer mortality rates did not differ between the high natural background area and the control area. For total cancer mortality, the observed cancer rate was higher in the normal background area, although the difference was not statistically significant. Known risk factors affecting cancer mortality rates were generally comparable in the two areas, although there were some cultural and educational differences. Chromosome aberrations and a higher reactivity of T lymphocytes were found in individuals in the high natural background area. There were no differences for a large number of hereditary diseases or congenital defects in children. The prevalence of Down syndrome was greater in the high-background region, but this was discounted because the residents of the control area had a lower prevalence of Down syndrome than those of surrounding counties, who had rates similar to those living in the high natural background area. Summary In areas of high natural background radiation, an increased frequency of chromosome aberrations has been noted repeatedly. The increases are consistent with those seen in radiation workers and in persons exposed at high dose levels, although the magnitudes of the increases are somewhat larger than predicted. No increase in the frequency of cancer has been documented in populations residing in areas of high natural background radiation. REFERENCES An86 Anderson, T. W. 1986. Ontario Hydro Mortality 1970-1985. Ontario Hydro Report, Canada. An88 Anspaugh, L. R., R. J. Catlin, and M. Goldman. 1988. The global impact of the Chernobyl reactor accident. Science 242:1513-1519. Ar87 Archer, V. E. 1987. Association of nuclear fallout with leukemia in the United States. Arch. Environ. Health 42:263-271. Ay84 Ayesh, R., J. R. Idle, J. C. Ritchie, M. J. Crothers, and M. R. Hetzel. 1984. Metabolic oxidation phenotypes as markers for susceptibility to lung cancer. Nature 312:169-170. Ba75 Barcinski, M. A., M. D. C. A. Abreu, J. C. C. De Almeida, J. M. Naya, L. G. Fonseca, and L. E. Castro. 1975. Cytogenetic investigation in a Brazilian population living in an area of high natural radioactivity. Am. J. Hum. Genet. 27:802-806. Ba84 Baron, J. A. 1984. Cancer mortality in small areas around nuclear facilities in England and Wales. Br. J. Cancer 50:815-829. Be83 Beck, H. L., and P. W. Krey. 1983. Radiation exposure in Utah from Nevada nuclear tests. Science 220:18-24. Be85 Beral, V., H. Inskip, P. Fraser et al. 1985. Mortality of employees of the United Kingdom Atomic Energy Authority, 1946-1979. Br. Med. J. 291:440-447.
LOW DOSE EPIDEMIOLOGIC STUDIES 386 Be87 Beral, V. 1987. Cancer near nuclear installations (letter). Lancet i:556. Ca80 Caldwell, G. S., D. B. Kelley, and C. W. Heath. 1980. Leukemia among participants in military maneuvers at a nuclear bomb test: A preliminary report. J. Am. Med. Assoc. 244:1575-1578. Ca83 Caldwell G. G., D. Kelley, M. Zack, H. Falk, and C. W. Heath. 1983. Mortality and cancer frequency among military nuclear test (Smoky) participants, 1957 through 1979. J. Am. Med. Assoc. 250(5):620-624. Ca84 Caldwell, G. C., D. B. Kelley, C. W. Health, Jr., and M. Zack. 1984. Polycythemia vera among participants of a nuclear weapons test. J. Am. Med. Assoc. 252:662-664. Ch85 Checkoway, H., R. M. Mathew, C. M. Shy, et al. 1985. Radiation, work experience, and cause specific mortality among workers at an energy research laboratory. Br. J. Ind. Med. 42:525-533. Ch88 Checkoway H., N. Pearce, D. J. Crawford-Brown, et al. 1988. Radiation doses and cause specific mortality among workers at a nuclear materials fabrication plant. Am. J. Epidemiol. 127:255-266. Cl87 Clapp, R. W., S. Cobb, C. K. Chan, and B. Walker, Jr. 1987. Leukaemia near Massachusetts nuclear power plant. Lancet ii:1324-1325. Co87a Cook-Mozaffari, P., F. L. Ashwood, T. Vincent, et al. 1987. Cancer incidence and mortality in the vicinity of nuclear installations. England and Wales, 1950-1980. Studies on Medical and Population Subjects, No. 51. London: Her Majesty's Stationery Office. Co87b Cook-Mozaffari, P. 1987. Cancer near nuclear installations (letter). Lancet i:855-856. Cr87 Crump, K. S., T. H. Ng, and R. G. Cuddihy. 1987. Cancer incidence patterns in the Denver metropolitan area in relation to the Rocky Flats plant . Am. J. Epidemiol. 126(1):127-135. Da87 Darby, S. C., and R. Doll. 1987. Fallout, radiation doses near Dounreay, and childhood leukemia. Br. Med. J. 294:603-607. Da88 Darby, S. C., G. M. Kendall, T. P. Fell, et al. 1988. A summary of mortality and incidence of cancer in men from the United Kingdom who participated in the United Kingdom's atmospheric nuclear weapon tests and experimental programs. Br. Med. J. 296:332-338. En83 Enstrom, J. E. 1983. Cancer mortality patterns around the San Onofre nuclear power plant, 1960-1978. Am. J. Public Health 73(1):83-92. Ev86 Evans, J. S., J. E. Wennberg, and B. J. McNeil. 1986. The influence of diagnostic radiography on the incidence of breast cancer and leukemia. N. Engl. J. Med. 315:810-815. Fa81 Fabrikant, J. I. 1981. Health effects of the nuclear accident at Three Mile Island. Health Phys. 40:151-161. Fo87 Forman, D., P. Cook-Mozaffari, S. Darby et al. 1987. Cancer near nuclear installations . Nature 329:499-505. Fr75 Fraisancho, A. R. 1975. Functional adaptation to high altitude hypoxia. Science 187:313-319. FSG88 Fallout Study Group. 1988. A case-control study of leukemia in Utah and its relationship to fallout from nuclear weapons testing. Unpublished manuscript. Ga84 Gardner, M. J., and P. D. Winter. 1984. Mortality in Cumberland during 1959-78 with reference to cancer in young people around Windscale (letter). Lancet i:216-217. Ga87a Gardner, M. J., A. J. Hall, S. Downes, and J. D. Terrell. 1987. Follow up
LOW DOSE EPIDEMIOLOGIC STUDIES 387 study of children born to mothers resident in Seascale, West Cumbria (birth cohort). Br. Med. J. 295:822-827. Ga87b Gardner, M. J., A. J. Hall, S. Downes, and J. D. Terrell. 1987. Follow up study of children born elsewhere but attending schools in Seascale, West Cumbria (schools cohort). Br. Med. J. 295:819-822. Gi72 Gibson, R., S. Graham, A. Lilienfeld, L. Schuman, J. E. Dowd, and M. L. Levin. 1972. Irradiation in the epidemiology of leukemia among adults. J. Natl. Cancer Inst. 48:301-311. Gi89 Gilbert, E. S., J. G. R. Petersen, and J. A. Buchanan. 1989. Mortality of workers at the Hanford site: 1945-1981. Health Phys. (in press). Go71 Gopal-Ayengar, A. R., K. Sundaram, K. B. Mistry, C. M. Sunta, K. S. V. Nambi, S. P. Kathuria, A. S. Basu, and M. David. 1971. Evaluation of the long-term Effects of high background radiation on selected population groups on the Kerala Coast. Pp. 31-51 in Proceedings of the 4th International Conference on Peaceful Uses of Atomic Energy, Vol. 11. Gu64 Gunz, F., and H. Atkinson. 1964. Medical radiation and leukemia: A retrospective survey. Br. Med. J. 1:389. Ha81 Harris, P. S., C. Lowery, A. G. Nelson, S. Obermiller, W. J. Ozeroff, and E. Weary. 1981. Shot Smoky, a test of the Plumbob Series, 31 August, 1957. Technical Report DNA 6004F. Washington, D.C.: Defense Nuclear Agency. He86 Heasman, M. A., I. W. Kemp, J. D. Urquhart, and R. Black. 1986. Childhood leukaemia in northern Scotland (letter). Lancet i:266. Hi81 Hickey, R. J., E. J. Bowers, D. E. Spence, B. S. Zemel, A. B. Clelland, and R. C. Clelland. 1981. Low level ionizing radiation and human mortality: Multiregional epidemiological studies. Health Phys. 40:625-641. Hi87 Hickey, R. J., E. J. Bowers, D. E. Spence, B. S. Zeuill, A. B. Clelland, and R. C. Clelland. 1987. Low level ionizing radiation and human mortality: multiregional epidemiological studies. Health Phys. 40:625-641. Ho86 Hole, D. J., and C. R. Gillis. 1986. Childhood leukaemia in the west of Scotland. Lancet 2:525-525. Ho87 Howe, G. R., J. L. Weeks, A. B. Miller, et al. 1987. A follow-up study of radiation workers employed by Atomic Energy of Canada Limited: Mortality , 1950-1981. IAG84 Independent Advisory Group. 1984. Investigation of the possible increased incidence of cancer in West Cumbria. London: Her Majesty's Stationery Office. In86 Interlaboratory Task Group on Health and Environmental Aspects of the Soviet Nuclear Accident. June 1987. Health and Environmental Consequences of the Chernobyl Nuclear Power Plant Accident. DOE/ER-0332. Springfield, Va.: National Technical Information Service. In87 Inskip, H., V. Beral, P. Fraser, et al. 1987. Further assessment of the effects of occupational radiation exposure in the United Kingdom Atomic Energy Authority mortality study. Br. J. Ind. Med. 44:149-160. Jo84 Johnson, C. J. 1984. Cancer incidence in an area of radioactive fallout downwind from the Nevada test site. J. Am. Med. Assoc. 251:230-136. Ki88 Kinlen, L. 1988. Evidence for an infective cause of childhood leukemia: comparison of a Scottish new town with nuclear reprocessing sites in Britain. Lancet ii:1323-1326. Kn88 Knox, E. G., A. M. Stewart, E. A. Gilman, and G. W. Kneale. 1988. Background radiation and childhood cancer. J. Radiol. Prot. 8(1):9-18.
LOW DOSE EPIDEMIOLOGIC STUDIES 388 Ko76 Kochupillan, N., I. C. Verma, M. S. Grewal, and V. Ramalingaswami. 1976. Down's syndrome and related abnormalities in an area of high background radiation in coastal Kerala. Nature 262:60-61. La79 Land, C. E. 1979. The hazards of fallout or of epidemiologic research. N. Engl. J. Med. 300:431-432. La84 Land, C. E., F. W. McKay, and S. G. Machado. 1984. Childhood leukemia and fallout from the Nevada nuclear tests. Science 223:139-144. Li80 Linos, A., J. Gray, A. Orvis, et al. 1980. Low dose radiation and leukemia. N. Engl. J. Med. 302:1101. Ly79 Lyon, J. K., M. R. Klauber, J. W. Gardner, and K. S. Udall. 1979. Childhood leukemias associated with fallout from nuclear testing. N. Engl. J. Med. 300:397-402. Ly84 Lyon, J. L., and K. L. Schuman. 1984. Radioactive fallout and cancer (letter). J. Am. Med. Assoc. 252(14):1854-1855. Ma87 Machado, S. G., C. E. Land, and F. W. McKay. 1987. Cancer mortality and radioactive fallout in southwestern Utah. Am. J. Epidemiol. 125:44-61. NRC80 National Research Council, Committee on the Biological Effects to Ionizing Radiations. 1980. The Effect on Populations of Exposure to Low Levels of Ionizing Radiation (BEIR III). Washington, D.C.: National Academy Press. 524 pp. No86 Norman C., and D. Dickson. 1986. The aftermath of Chernobyl. Science 223:1141-1143. Op88 Openshaw, S., M. Charlton, A. W. Craft, and J. M. Birch. 1988. Investigation of leukemia clusters by use of a geographical analysis machine. Lancet i:272-273. Pi83 Pickle, L. W., L. M. Brown, and W. J. Blot. 1983. Information available from surrogate respondents in case-control interview studies. Am. J. Epidemiol. 118:99. Pr88 Preston-Martin, S., D. C. Thomas, M. C. Yu, and B. E. Henderson. Submitted. Diagnostic radiography as a risk factor for chronic myeloid leukemia. Ra87 Raman, S., G. S. Dulberg, R. A. Spasoff, and T. Scott. 1987. Mortality among Canadian military personnel exposed to low dose radiation. Can. Med. Assoc. J. 136:1051-1056. Ro84 Rowley, J. B. 1984. Biological implications of consistent chromosome rearrangements in leukemia and lymphoma. Cancer Res. 44:3159-3168. Ro85 Robinette, C. D., S. Jablon, and T. L. Preston. 1985. Studies of participants in nuclear tests. Washington, D.C.: National Research Council. Ro87 Roman, E., V. Beral, L. Carpenter, et al. 1987. Childhood leukemia in the West Berkshire and Basingstoke and North Hampshire District Health Authorities in relation to nuclear establishments in the vicinity. Br. Med. J. 294:597-602. Sh86 Shrimpton, P. C., B. F. Wall, D. G. Jones, E. S. Fisher, M. C. Hillier, G. M. Kendall, and R. M. Harrison. 1986. Doses to patients from routine diagnostic x-ray examinations in England. Br. J. Radiol. 59:749-758. Sm86 Smith, P. G., and A. J. Douglas. 1986. Mortality of workers at the Sellafield plant of British Nuclear Fuels. Br. Med. J. 293:845-854. St62 Stewart, A., W. Pennypacker, and R. Barber. 1962. Adult leukemia and diagnostic x-rays. Br. Med. J. 2:882. St73a Stewart, A. 1973. An epidemiologist takes a look at radiation risks. DHEW Publ. No. [FDA] 738024, BRH/DBE 73-2. Washington, D.C.: U.S. Department of Health, Education, and Welfare.
LOW DOSE EPIDEMIOLOGIC STUDIES 389 St73b Stewart, A. 1973. The carcinogenic effects of low level radiation. A re-appraisal of epidemiologic methods and observations. Health Phys. 24:223-240. Ta86 Tao, Z., and L. Wei. 1986. An epidemiological investigation of mutational diseases in the high background radiation area of Yangjiang, China. J. Radiat. Res. 27:141-150. UN82 United Nations Scientific Committee on the Effects of Atomic Radiations (UNSCEAR). 1982. Ionizing Radiation: Sources and Biological Effects. United Nations publication. E.82.IX.8 06300p. New York: United Nations, 773pp. We86 Wei, L. X., Y. R. Zha, Z. F. Tao, et al. 1986. Recent advances of health survey in high background areas in Yangjiang, China. Pp. 1-17 in Proceedings of the International Symposium on Biological Effects of Low Level Radiation. Wi87 Wilkinson, G.S., G. L. Tietjen, L. D. Wiggs, W. A. Galke, J. F. Acquavella, M. Reyes, G. L. Voelz, and R. J. Waxweiler. 1987. Mortality among plutonium and other radiation workers at a plutonium weapons facility. Am. J. Epidemiol. 125:231-250.
LOW DOSE EPIDEMIOLOGIC STUDIES 390