9
Environmental Radiation Studies
INTRODUCTION
A considerable number of epidemiologic studies have been reported that have attempted to determine whether persons exposed, or potentially exposed, to ionizing radiation from environmental sources are at an increased risk of developing cancer. All epidemiologic studies are inherently uncertain, because they are observational in nature rather than experimental. Nevertheless, not all study designs are equally informative regarding the estimation of radiation risk to humans, and not all epidemiologic studies are of the same quality. Therefore, in evaluating the evidence regarding the risk of exposure to environmental sources of radiation, it is important to consider carefully the specific methodological features of the study designs employed.
Studies of environmental radiation exposure are of three basic designs: (1) descriptive studies, often referred to as ecologic; (2) case-control studies; and (3) cohort or followup studies. The existing published literature consists primarily of reports that are descriptive in nature and ecologic in design. The preponderance of this type of study is due to the fact that they are relatively easy to carry out and are usually based on existing data. Such investigations have utilized incidence, mortality, and prevalence data to estimate disease rates and, typically, to evaluate whether rates of disease vary in a manner that might be related to radiation exposure. If these analyses are based on large numbers of cases or large population groups, such studies may give the appearance of very precise results. Most often, geopolitical boundaries or distance from a source of radiation are used as surrogate means to define radiation exposure. For example, cancer incidence rates might be evaluated as a function of distance from a nuclear facility, or specialized statistical techniques might be employed to determine whether cases of cancer cluster or aggregate in a particular region or time period characterized by potential radiation exposure more than would be expected to occur by chance (i.e., in the absence of any exposure).
Weaknesses associated with studies of this type make them of limited value in assessing risk. The primary limitation is that the unit of analysis is not the individual; thus, generally little or no information is available that is specific to the individual circumstances of the people under study. Of most concern in this regard is the definition of radiation exposure. Ecologic studies generally do not include estimates of individual exposure or radiation dose. Either aggregate population estimates are used to define population dose for groups of people, or surrogate indicators such as distance or geographic location are used to define the likelihood or potential for exposure or, in some cases, an approximate magnitude or level of exposure. This approach has serious limitations. It implies, for example, that residents who live within a fixed distance from a facility are assumed to have received higher radiation doses than those who live at greater distances or than individuals in the larger population as a whole who do not live in the vicinity of the facility. Further, it assumes that everyone within the boundary that defines exposure (or a given level of exposure) is equally exposed or has the same opportunity for exposure. In most situations, such assumptions are unlikely to be accurate, and variability in exposure of individuals within the population may be substantially greater than the exposure attributed on a population basis. The resulting almost certain misclassification of exposure can lead to a substantial overestimation or underestimation of the association of the exposure with the disease under study.
Similarly, there is usually no information available in ecologic studies regarding other factors that might influence the risk of developing the disease(s) under study (i.e., other risk factors). Thus, there is no way to evaluate the impact of such factors in relation to the potential effect of radiation exposure. This inability to evaluate or account for the potential confounding effect of other important factors, or the modifying effect of such factors on risk, makes the ecologic approach of limited use in deriving quantitative estimates of radiation risk.
A third limitation of the ecologic design is that disease outcome usually is not confirmed at the individual level. Most studies rely on routine reporting, either of mortality through death certificates or of cancer incidence through cancer registration and surveillance systems. Such sources of information vary in their degree of accuracy and completeness, and they can sometimes vary in relation to the surrogate measures being used to define exposure (e.g., geographic area). This can lead to the identification of spurious associations.
Fourth, ecologic studies seldom estimate or account for population migration or movement. This, too, can result in the appearance of spurious associations if aggregate or population measures of radiation exposure actually reflect underlying changes in population mobility with factors such as time, age, or geographic area.
Finally, descriptive studies are often based on a small number of cases of disease. Such studies have low statistical power to detect an association if it truly exists, and they are very sensitive to random fluctuations in the spatial and/or temporal distribution(s) of the disease(s) under study. This is especially true for diseases such as cancer, particularly childhood cancer, which are relatively uncommon on a population basis.
There have also been attempts to evaluate the effect of environmental radiation exposures using the two most common analytical study designs employed in epidemiology: the case-control and the cohort study. Such studies are almost always based on individual-level data and thus are not subject to many of the limitations summarized above for ecologic studies. Nevertheless, each of these study designs is subject to specific weaknesses and limitations. Of most concern in case-control studies is the potential bias that can result in relation to the selection of cases and controls, such that the two groups are differentially representative of the same underlying population. A second important source of bias can be differential recall of information about exposure for cases relative to controls. In cohort studies, a common limitation is the relatively small number of cases for uncommon disease outcomes and the resultant low statistical power. A second concern is the completeness of follow-up of the cohort under study, and equal follow-up and determination of disease status according to exposure. Such limitations of both types of analytic epidemiologic studies may be particularly problematic in investigations of low doses and relatively small increases in disease risk. Under such circumstances, the magnitude of the impact on risk estimates of small or modest biases may be as great or greater than the magnitude of the true disease risk.
In summary, most existing published studies of environmental radiation exposure are ecologic in design. Such studies are limited in their usefulness in defining the risk of disease in relation to radiation exposure or dose. They can sometimes be informative in generating new hypotheses or suggesting directions of study but seldom, if ever, are of value in testing specific hypotheses or providing quantitative estimates of risk in relation to specific sources of environmental radiation. Epidemiologic studies, in general, have limited ability to define the shape of the radiation dose-response curve and to provide quantitative estimates of risk in relation to radiation dose, especially for relatively low doses. To even attempt to do so, a study should (1) be based on accurate, individual dose estimates, preferably to the organ of interest; (2) contain substantial numbers of people in the dose range of interest; (3) have long enough follow-up to include adequate numbers of cases of the disease under study; and (4) have complete and unbiased follow-up. Unfortunately, the published literature on environmental radiation exposures is not characterized by studies with such features.
The accompanying tables provide a summary of the principal studies of environmental radiation exposure published since the BEIR V report (NRC 1990). Articles included in this summary were identified principally from searching the PubMed database of published articles from 1990 through July 2004. Searches were restricted to human studies and were broadly defined: key words included radiation; neoplasms; radiation-induced; radioactive fallout; and environmental radiation. Searches specific to the Chernobyl accident included Chernobyl, Russia, Ukraine, and Belarus as key words. Articles were also identified from UNSCEAR (2000b) and from the usual scientific interactions with other investigators. The tables are organized according to the type of exposure situation under study as follows: (1) populations living around nuclear facilities; (2) populations exposed from atmospheric testing, fallout, or other environmental releases of radiation; (3) populations exposed from the Chernobyl accident; (4) populations exposed from natural background; and (5) children of adults exposed to radiation. Within each type of exposure situation, the tables are further grouped according to study design: ecologic studies, case-control studies, and cohort studies. Each table contains a brief description of the principal design features and results of each study. The principal criteria used to assess the utility of each study in evaluating the risk of disease in relation to radiation exposure were the following: (1) Was there a quantitative estimate of radiation dose; (2) if so, was the estimate for individuals in the study (i.e., individual-level estimates of radiation dose received); and (3) was there a quantitative estimate of disease risk in relation to radiation dose?
POPULATIONS LIVING AROUND NUCLEAR FACILITIES
Table 9-1A lists 16 ecologic studies of populations living around nuclear facilities, 13 of the locations being outside the United States. Most define exposure, or potential for exposure, based on a measure of distance from the facility, although the two studies of exposures at Three Mile Island by Hatch (1992) utilized some information on measurements
TABLE 9-1A Populations Living Around Nuclear Facilities—Ecologic Studies
Reference |
Incidence/Mortality |
Population Studied |
Type of Exposure |
Dates of Accrual |
Type of Dosimetry |
Outcomes Studied |
Number of Cases |
Summary of Results |
Hatch and Susser (1990) |
Incidence and mortality |
Residents (ages 0–24) within 10 miles of Three Mile Island |
Background gamma |
1975–1985 |
Outdoor measurements taken in 1976 |
All cancer; leukemia |
49 (0–14) 104 (0–24) |
Increased risk for highest vs. lowest quartile; childhood cancer and leukemia |
Hatch and others (1990) |
Incidence |
Residents within 10 miles of Three Mile Island |
Xenon and iodine |
1975–1985 |
Dispersion modeling, based on monitoring data |
All cancer; childhood cancer (ages 0–14, 0–24); leukemia (ages 0–14, 0–24, 25); and all lymphoma |
5493 total |
No evidence of an effect on cancer incidence |
Jablon and others (1991) |
Mortality |
Residents of 107 counties in U.S. with or near nuclear installations |
Unspecified |
1950–1984 |
County with a nuclear facility that began operation before 1982, or an adjacent county if at least 20% of the county was within a 16 km radius |
15 cancer sites; benign and unspecified neoplasms |
900,000 deaths in 107 counties |
No evidence of excess mortality in study counties |
Sofer and others (1991) |
Incidence |
Children and young adults living near nuclear plant in Israel |
Unspecified |
1960–1985 |
Distance from Negev nuclear plant |
Leukemia |
192 |
No overall increase; some increase with time among 0–9 in Western Negev; increase in girls 0–4 from 1970 to 1979 |
Michaelis and others (1992) |
Incidence |
Children living near nuclear installations in Germany |
Unspecified |
1980–1990 |
Distance from nuclear facility |
Childhood cancer; acute leukemia |
81 within 5 km |
No increase for all cancer, acute leukemia; suggested increases in subgroups of early ages or close proximity |
McLaughlin and others (1993b) |
Incidence and mortality |
Children born to mothers residing near nuclear installations in Ontario, Canada |
Unspecified |
1950–1987 |
Distance from nuclear facility |
Leukemia in children |
Range by facility: 2–72 |
Suggestion of some excess over expected for some analyses; none significant |
Bithell and others (1994) |
Incidence |
Children in England and Wales |
Unspecified |
1966–1987 |
Distance from nuclear facilities based on ward |
Leukemia and non-Hodgkin’s lymphoma (NHL) |
Range for 25 km zones: 7–570 |
Linear risk score significantly elevated in Sellafield and Burghfield |
Black and others (1994a) |
Incidence |
Residents of Dalgety Bay, Scotland |
Particles of radium-226 |
1975–1990 |
Routine monitoring measurements |
All cancer; 18 specific sites |
211 (total) |
No evidence of increase over expected |
Black and others (1994a) |
Incidence |
Children and young adults in Dounreay, Scotland |
Contamination from nuclear reprocessing plant |
1968–1991 |
Distance from Dounreay |
Leukemia and NHL |
12 in nearest zone |
Evidence of increase over expected in nearest zone |
Reference |
Incidence/Mortality |
Population Studied |
Type of Exposure |
Dates of Accrual |
Type of Dosimetry |
Outcomes Studied |
Number of Cases |
Summary of Results |
Zaridze and others (1994) |
Incidence |
Children in Kazakhstan |
Unspecified |
1981–1990 |
Distance from nuclear testing sites |
All cancer; six specific sites |
Total: 1408; leukemia: 512 |
Increase in leukemia in areas closest to testing sites; some evidence of increase in brain tumors |
Viel and others (1995) |
Incidence |
People under age 25 living around La Hague reprocessing plant in France |
Unspecified |
1978–1992 |
Distance from the La Hague plant |
Leukemia |
25 |
Cluster of cases located close to La Hague plant |
Waller and others (1995) |
Incidence |
Children in 2594 parishes of Sweden |
Unspecified |
1980–1990 |
Distance from nuclear facility |
Acute lymphocytic leukemia (ALL) |
656 |
No significant clustering of cases found |
Gulis and Fitz (1998) |
Incidence |
Residents of Trnava, Slovakia |
Unspecified |
1986–1995 |
Distance from nuclear power plant |
13 cancer sites |
Range for zones: 0–323 |
Suggestion of increasing incidence closer to the site; nonsignificant |
Kaatsch and others (1998) |
Incidence |
Children living near nuclear facilities in Germany |
Unspecified |
1991–1995 |
Distance from a nuclear facility |
All cancer; leukemia; lymphoma; selected sites |
Total 550; leukemia 182 |
No evidence of an increase in incidence |
Guizard and others (2001) |
Incidence |
Residents under age 25 in areas around the La Hague plant in France |
Unspecified |
1978–1998 |
Distance from the La Hague plant |
Leukemia |
38 |
Increase over expected in area less than 10 km from site |
Boutou and others (2002) |
Incidence |
Nord Cotentin, France |
Population mixing—near nuclear power plant and reprocessing unit |
1979–1998 |
Population mixing index per geographic unit (commune), based on number of workers born outside department of La Manche |
Childhood leukemia in persons under age 25 |
|
Incidence rate ratio 2.7 in rural communes in highest tertile of mixing, relative to urban communes. Positive trend in leukemia with increasing mixing index. Risk stronger for ALL in children 1–6 |
taken around the site after the accident. All but one (Jablon and others 1991) are based on incidence data, and one study in Canada (McLaughlin and others 1993a) uses mortality data as well as incidence data. The focus of most of these investigations is leukemia and/or childhood cancer, although a few include all cancers as an outcome. The size of the studies, in terms of numbers of cases, ranges from very small (Black and others, 1994a; 12 cases in the most highly exposed zone) to extremely large (Jablon and others 1991). Notably, most of the studies do not specify the nature of the radiation exposure, and none of the 16 contains individual estimates of radiation dose. Although some of these studies report an increased occurrence of cancer that could potentially be related to environmental radiation exposures, none provides a direct quantitative estimate of risk in relation to radiation dose.
Table 9-1B summarizes three case-control studies of persons living around a nuclear facility. Two studies are of leu
TABLE 9-1B Populations Living Around Nuclear Facilities—Case-Control Studies
Reference |
Population Studied |
Number of Subjects |
Dates of Accrual |
Type of Exposure |
Type of Dosimetry |
Summary of Results |
||
Cases |
Controls |
Cases |
Controls |
|||||
Urquhart (1991) |
Leukemia and NHL in children under age 15 resident in Caithness |
Selected from birth register; matched by zone of residence at birth, date of birth, sex |
14 |
55 |
Diagnosis 1970–1986 |
Paternal preconception whole-body dose; antenatal X-ray |
Employment at Dounreay; recorded dose from employment records; questionnaire for X-ray |
No increased risk with employment at Dounreay, recorded radiation dose, antenatal X-ray; evidence of increased risk from playing on beaches within 50 km of Dounreay |
Shields and others (1992) |
Congenital and perinatal conditions, stillbirths, infant deaths |
Chronologically nearest normal single birth; matched by sex, mother’s age within 5 years, gravidity |
266 |
266 |
1964–1981 |
Environmental exposure from working or living near, or working in uranium mines |
Environment: time prior to child’s birth worked in uranium mine; residence within 0.5 mile of mine, dumps, or tailings; living in home made with mine rock. Workers: recorded WLM, estimated gonadal dose |
Only significant association with mother living near tailings or mine dumps. Overall, associations with measures of radiation exposure were weak |
Pobel and Viel (1997) |
Leukemia diagnosed in people <25 years of age living within 35 km of La Hague nuclear plant |
Sample of children cared for by general practitioners of the cases; matched to cases on sex, age, place of birth; and residence at diagnosis of case |
27 |
192 |
1978–1993 |
Antenatal and postnatal X-ray exposure; parental occupational exposures (including radiation); viral infections, life-style |
For parents employed in nuclear facility, whole-body external dose (mSv) was obtained from company records. Other information obtained by questionnaire |
No association with occupational radiation exposure of parents; increased risk for use of local beaches, consumption of local fish, length of residence in granitic area or house |
kemia, one in children under age 15 (Urquhart and others 1991) and the other in people under age 25 (Pobel and Viel 1997). Both studies are based on a small number of cases and focus primarily on parental radiation exposure and X-ray exposure of the child. Neither study found an increased risk associated with these types of radiation exposure. Both, however, did find an increased risk associated with playing on beaches near the nuclear facility. The third study (Shields and others 1992) focuses on congenital and perinatal conditions, stillbirths, and infant deaths in relation to exposures from uranium mines. Exposures include environmental exposures from living near a mine or mine dumps or tailings, or living in a home made from mine rock, as well as from working in a uranium mine. This study does not provide an estimate of radiation risk associated with any of the indicators of exposure.
In summary, most of the studies of populations living around nuclear facilities have not included individual esti-
TABLE 9-2A Populations Exposed from Atmospheric Testing, Fallout, or Other Environmental Release of Radiation—Ecologic Studies
Reference |
Incidence/Mortality |
Population Studied |
Type of Exposure |
Dates of Accrual |
Type of Dosimetry |
Outcomes Studied |
Number of Cases |
Summary of Results |
Darby and others (1992) |
Incidence |
Children under age 15 in Nordic countries |
Fallout from nuclear weapons tests |
Denmark (1948), Finland, Norway, Iceland (1958), Sweden (1961–1987) |
Estimates of bone marrow dose to fetus, 1-year-old, testes, received during fallout period: low, medium, high |
Leukemia |
Not given |
Little increase in high-fallout years; slightly elevated in high vs. medium group |
Gilbert (1998) |
Incidence and mortality |
United States |
Fallout from nuclear weapons tests in Nevada |
Deaths: 1957–1994; incident cases: 1973–1994 |
Mean thyroid dose by county, derived from measurements and environmental modeling |
Thyroid cancer |
4602 deaths; 12,657 incident cases |
No increased risk with cumulative dose or dose received at ages 1–15; suggested increase for those exposed under age 1 and those in 1950–1959 birth cohort |
mates of radiation dose and have therefore not provided an estimate of disease risk. The three case-control studies described above found no increased risk of disease associated with radiation exposure.
POPULATIONS EXPOSED FROM ATMOSPHERIC TESTING, FALLOUT, OR OTHER ENVIRONMENTAL RELEASE OF RADIATION
Table 9-2A describes two ecologic studies of populations exposed to fallout from atmospheric nuclear testing, fallout, or other sources of environmental release of radiation. The nature of the exposure is not specified beyond “fallout.” These studies utilize population-based measures of exposure rather than individual estimates of radiation dose. They address two separate outcomes (leukemia and thyroid cancer), but provide no quantitative estimates of risk associated with the exposure.
Table 9-2B summarizes two cohort studies of persons who participated in U.K. atmospheric nuclear weapons tests. The study by Darby and colleagues (1993) is an extension of an earlier analysis from this cohort and uses doses from film badges to characterize individual external whole-body radiation dose. It investigates all causes of mortality as well as all major forms of cancer. Overall, the study found no increased risk of developing cancer or other fatal diseases as a function of estimated dose received, based on follow-up through 1991 and relatively large numbers of cases. There was some evidence of an increase in leukemia, based on only 29 cases. The most recent update of this cohort (Muirhead and others 2003) found little increase in overall mortality or cancer incidence and no increase in other types of cancer, but continuing evidence of a small increased risk of nonchronic lymphocytic leukemia (CLL).
In contrast, a recent study of U.S. veterans (Dalager and others 2000) who participated in atmospheric nuclear weapons tests reported a significant increase in death from all causes, and for all lymphopoietic cancers combined, although the number of cases in the latter group was very small. This study focused on veterans whose external γ-radiation dose, as recorded on film badges, was 5 rem, and compared mortality in this group to veterans who participated in one nuclear test and whose dose was 0.25 rem. The mean dose among the 5 rem group was 7.8 rem and among the controls was 0.08.
Also included in Table 9-2B are several studies of the population of residents living near the Techa River in the southern Urals of the Russian Federation. More than 25,000 residents were exposed to external γ-radiation as well as internally from fission products (primarily cesium-137, strontium-90, ruthenium-106, and zirconium-95) released into the Techa River from the nearby Mayak plutonium production facility, predominately in the early 1950s. Studies have been conducted of cancer mortality in residents and their offspring, as well as pregnancy outcomes. Initial dose estimates were based on average doses reconstructed for settlements.
Efforts to estimate individual doses for members of this resident cohort continue. To date, there is no evidence of a decrease in birth rate or fertility in the exposed population, and there is no increased incidence of spontaneous abortions or stillbirths (Kossenko and others 1994). There is some evidence of a statistically significant increase in total cancer mortality (Kossenko 1996). Current estimates of the excess absolute risk (EAR)1 of leukemia in this cohort is 0.85 per 10,000 person-years (PY) per gray (95% CI 0.2, 1.5), and for
TABLE 9-2B Populations Exposed from Atmospheric Testing, Fallout, or Other Environmental Release of Radiation—Cohort Studies
Reference |
Incidence/Mortality |
Cohort Definition |
Comparison Group |
Dates of Accrual |
Type of Exposure |
Type of Dosimetry |
Outcomes Studied |
Number of Cases |
Summary of Results |
Darby and others (1993) |
Incidence and mortality |
Persons who participated in U.K. atmospheric nuclear weapons tests |
Men identified from Ministry of Defense archives who did not participate |
1950s–1991 |
External whole-body dose |
Recorded on film badges obtained from Ministry of Defense |
Broad causes of death; 27 specific cancer sites |
All causes: 2753 (control group—2939) |
No effect on risk of developing cancer or other fatal diseases; some evidence of an increase over expected for leukemia, based on 29 cases |
Kossenko and others (1994) |
Pregnancy outcome and mortality |
Children born to 28,100 residents exposed to discharges of radioactive waste into Techa River |
Unexposed populations living in the same area |
1953–1974 |
External and internal dose: primarily from 137Cs, 90Sr, 106Ru, 95Zr |
Gonadal doses estimated as average for each settlement |
Birth rate, fertility, fetal loss, infant mortality |
56 cancer deaths |
No decrease in birth rate or fertility in exposed population; no increased incidence of spontaneous abortions or stillbirths; no change in cancer mortality |
Kossenko and others (1994) |
Mortality |
28,000 residents exposed to discharges of radioactive waste into Techa River, 1950–1953 |
Unexposed populations living in the same area |
1950–1982 |
External and internal dose: primarily from 137Cs, 90Sr, 106Ru, 95Zr |
Average absorbed dose to bone marrow estimated for each settlement |
All cancer and 13 major site categories |
163 cancers in exposed population |
Increase in total cancer mortality. Leukemia: absolute risk 0.85 per 10,000 PY per gray; relative risk for esophagus, stomach, and lung similar to atomic bomb survivors |
Kossenko (1996) |
Mortality |
28,000 residents exposed to discharges of radioactive waste into Techa River |
Matched control group from unexposed area |
33-year period from 1949 through 1982 |
External and internal dose: primarily from 137Cs, 90Sr, 106Ru, 95Zr |
Average absorbed dose to bone marrow estimated for each settlement |
Leukemia and solid cancer in residents; cancer in offspring |
|
Leukemia: absolute risk 0.85 per 10,000 PY per gray; solid cancer: relative risk 0.65 Gy−1. No increase in offspring of exposed residents |
Davis and others (2001) |
Cumulative incidence |
Persons born to mothers resident in one of 7 counties surrounding Hanford Site from 1940 to 1946 |
Internal control according to estimated individual thyroid radiation dose |
Birth through date of exam in 1992–1997 |
Primarily 131I |
Estimated individual absorbed dose to thyroid |
Thyroid cancer and 12 categories of noncancer thyroid disease |
19 thyroid cancer cases |
No increase in thyroid cancer or any noncancer thyroid disease outcome associated with increasing radiation dose to the thyroid |
Dalager and others (2000) |
Mortality |
Persons who participated in U.S. atmospheric nuclear weapons tests and received highest doses |
Navy veterans who participated in HARDTACK and received minimal radiation dose |
Date of first exposure through 1996 |
External gamma dose |
Film badges |
All deaths; lymphopoietic, leukemia, digestive, respiratory, other cancer |
300 deaths in veterans with 5 rem; 11 cases of lymphopoietic cancer |
All-cause mortality: relative risk (RR) 1.22 (95% CI 1.04–1.44); lymphopoietic cancer 3.72 (95% CI 1.28–10.83) |
Reference |
Incidence/Mortality |
Cohort Definition |
Comparison Group |
Dates of Accrual |
Type of Exposure |
Type of Dosimetry |
Outcomes Studied |
Number of Cases |
Summary of Results |
Kossenko and others (2000) |
Mortality |
10,459 offspring of parents exposed to discharges of radioactive waste into Techa River |
None |
1950–1992 |
External and internal dose: primarily from 137Cs, 90Sr, 106Ru, 95Zr |
None |
Cancer |
25 cancer deaths |
Descriptive analyses only—no estimates of risk |
Koshurnikova and others (2002) |
Mortality and incidence |
Ozyorsk Population |
72,185 persons living in Ozyorsk for at least 1 year under age 15 and born 1948–1988; or born elsewhere 1934–1988 but moved to Ozyorsk before age 15 |
1948–1988 |
Fallout from Mayak facility |
None |
Deaths, cancer deaths, leukemia, thyroid cancer |
4636 deaths; 371 cancer deaths; 53 leukemia deaths; 31 thyroid cancer cases |
Thyroid cancer 3–4 times expected relative to Russia; 1.5–2-fold higher based on Chelyabinsk Oblast rates |
Muirhead and others (2003) |
Mortality and incidence |
21,357 persons who participated in the U.K. atmospheric nuclear weapons tests |
22,333 men who did not participate in tests identified from Ministry of Defense records, matched on a number of characteristics |
1952–1998 |
External gamma |
Film badge readings and potential for exposure based on duties |
All deaths, 27 types of cancer |
2089 deaths; 785 cancer deaths; 16 leukemia deaths; 2641 cases of cancer; 67 cases of leukemia |
Little difference in overall mortality or cancer incidence between exposed and controls; no increase in multiple myeloma; evidence of a small risk of non-CLL leukemia |
Takahashi and others (2003) |
Prevalence |
3709 Marshall Island Residents born before the Castle BRAVO atmospheric nuclear weapons tests on March 1, 1954 |
Internal control according to estimated dose level |
1993–1997 |
Fallout from Castle BRAVO test |
Surrogate estimates of dose based on 137Cs soil deposition levels |
Thyroid cancer |
57 cases |
Prevalence increased with quartile of estimated dose, but was not significant |
solid tumors the relative risk estimate is 0.65 Gy−1 (95% CI −0.3, 1.0). Median dose estimates for soft tissue in this cohort are 7 mSv (maximum 456 mSv) and for bone marrow 253 mSv (maximum 2021 mSv). Estimates of the relative risk for cancer of the esophagus, stomach, and lung are similar to those reported for atomic bomb survivors. There is no evidence of an increase in cancer mortality in the offspring of exposed residents (Kossenko 1996). There has also been one study (Koshurnikova and others 2002) of persons living in the town of Ozyorsk exposed to fallout from the nearby Mayak nuclear facility. This study reported an excess of thyroid cancer three to four times that expected relative to rates for all of Russia and a somewhat lower excess (1.5 to twofold higher) based on a comparison with Chelyabinsk Oblast rates. No estimates of radiation dose were included in this study.
Two other cohort studies of persons exposed to atmospheric releases of radioactive materials are also summarized in Table 9-2B. One is a follow-up study of 3440 persons exposed as young children to atmospheric releases of primarily 131I from the Hanford nuclear facility in eastern Washington State (Davis and others 2001, 2004a). No increased risk of thyroid cancer was found associated with individual radiation dose to the thyroid. The other (Takahashi and others 2003) is a prevalence study of thyroid cancer conducted through screening of 3,709 Marshall Island residents born before the Castle BRAVO atmospheric nuclear weapons test on March 1, 1954. Radiation dose was based on a surrogate constructed from age-specific doses estimated for the Utirik atoll and 137Cs deposition levels on atolls where the participants resided. There was some indication that the prevalence of thyroid cancer increased with quartile of estimated dose, but the increase was not statistically significant.
In summary, some but not all studies of persons exposed to fallout or other environmental releases of radiation have found increased risks of specific disease outcomes. Most notable are findings of a significant increase in death from all causes and for all lymphopoietic cancers combined in a recent study of U.S. veterans who participated in atmospheric nuclear weapons tests, and evidence of an increase in total cancer mortality and thyroid cancer incidence among residents living near the Techa River in the southern Urals of the Russian Federation.
POPULATIONS EXPOSED FROM THE CHERNOBYL ACCIDENT
The explosion at the Chernobyl Power Station Unit 4 in Ukraine on April 26, 1986, released large quantities of radionuclides into the atmosphere, resulting in the contamination of a large geographic area. Initially exposures were due principally to radioisotopes of iodine, primarily iodine-131 (131I), and subsequently to radiocesium, primarily cesium-137 (137Cs), from both external exposure and the consumption of contaminated milk and other foods. Numerous epidemiologic studies have been carried out since the Chernobyl accident to investigate the potential late health consequences of exposure to ionizing radiation from the accident. These studies have focused largely on thyroid cancer in children, but have also included investigations of recovery operation workers and residents of contaminated areas, and have investigated the occurrence of leukemia and solid tumors other than thyroid cancer among exposed individuals.
Overwhelmingly, the published findings are from studies that are ecologic in design and therefore do not provide quantitative estimates of disease risk based on individual exposure circumstances or individual estimates of radiation dose. Most reports are descriptive incidence and prevalence studies that utilize population or aggregate estimates of radiation dose. The principal studies are summarized in Table 9-3A. Only four analytical studies are published that report dose-response results based on individual dose estimates (Table 9-3B). In the sections that follow, current evidence is summarized separately regarding the risk of thyroid cancer, leukemia, and other solid tumors associated with radiation exposure from the Chernobyl accident. Studies of recovery operations workers are considered in Chapter 8 on occupational exposures.
Thyroid Cancer
An increase in the incidence of thyroid cancer first began to appear in Belarus and Ukraine in 1990. After the initial few reports, there was immediate skepticism that such increases were related directly to radiation exposure from Chernobyl. The very early onset of disease after exposure (only 4 years) was unexpected based on existing knowledge of the latent period for radiation-related thyroid cancer; there was doubt about the certainty of the pathologic diagnoses; and there was speculation that the apparent increases were largely the result of widespread population screening.
Numerous reports have continued to describe an increasing number of cases of thyroid cancer, particularly in the most heavily contaminated regions of Ukraine and Belarus, and also in Russia. Collectively, findings reported to date have demonstrated an association between radiation exposure from the Chernobyl accident and an increase in thyroid cancer incidence. Among those under age 18 at the time of the accident, it has been estimated that approximately 2000 thyroid cancers were diagnosed from 1990 to 1998 in Ukraine, Belarus, and Russia. The increase in all three countries for this period was approximately fourfold, with the highest increase observed in the Gomel region in Belarus. More recent data indicate that excess thyroid cancer continues to occur among people in Belarus, Ukraine, and the contaminated regions of Russia. This increase cannot be explained only by the aging of the cohort and the improvement in case detection and reporting. Although there is now little doubt that an excess of thyroid cancer has occurred in highly contaminated areas, there is still very little information re-
TABLE 9-3A Populations Exposed from the Chernobyl Accident—Ecologic Studies
Reference |
Incidence/Mortality |
Population Studied |
Type of Exposure |
Dates of Accrual |
Type of Dosimetry |
Outcomes Studied |
Number of Cases |
Summary of Results |
Prisyazhiuk and others (1991) |
Incidence |
Three contaminated districts in Ukraine: Polesskoye, Naroditchy, Ovrutch |
Fallout from Chernobyl |
1981–1990 |
Calendar year (before and after accident) and district (contaminated areas) |
Leukemia, thyroid cancer, all other cancer |
Leukemia: 105; thyroid: 25; all other: 3804 |
Overall, incidence rates were not different before and after the accident. Leukemia in age 65+ group increased in 1987 and remained 2–3 times higher; three cases of thyroid cancer diagnosed in 1990 in <14 age group (none 1981–1989); all others increased in 1987 by a third |
Ramsay and others (1991) |
Incidence |
Population of Lothian, Scotland |
Fallout from Chernobyl |
1978–1989 |
Calendar year (i.e., from Chernobyl, before and after accident) |
Down’s syndrome |
Ave.: 12.4 cases per year; range 7 (1989)–26 (1987) |
Significant increase in 1986–1987 |
Baverstock and others (1992) |
Incidence |
Belarus |
Fallout from Chernobyl |
1986–1992 |
Calendar year and region |
Thyroid cancer |
104 |
Marked increase beginning in 1990; highest rates in Gomel |
Kazakov and others (1992) |
Incidence |
Six regions of Belarus and Minsk city |
Fallout from Chernobyl |
1986–1992 |
Calendar year and region |
Thyroid cancer |
131 |
Average of 4 cases per year 1986–1989; 55 in 1991; projected 60 in 1992. Most increase in Gomel |
Ivanov and others (1993) |
Incidence |
Belarus: children ages 0–14 |
Fallout from Chernobyl |
1979–1991 |
Two time periods: 1979–1985; 1986–1991. Three levels of contamination by region or city |
Childhood leukemia |
Not given |
No change in incidence after Chernobyl accident, and no increase after accident in areas with higher contamination levels |
Parkin and others (1993) |
Incidence |
20 European countries: children ages 0–14 |
Fallout from Chernobyl |
1980–1988 |
Estimated dose (effective equivalent dose) in 30 countries or regions, obtained from UNSCEAR |
Childhood leukemia |
3679 |
Risk of leukemia 1987–1988 relative to before 1986 was not related to radiation exposure |
Auvinen and others (1994) |
Incidence |
Finland: children 0–14 in 1976–1992 |
Fallout from Chernobyl |
1976–1992 |
Estimated cumulative dose in 2 years after the accident. Based on measurements of dose rate in 455 municipalities. Internal dose estimated from whole-body measurements on sample of 81 children. Municipalities divided into fifths of exposure |
Childhood leukemia |
Not given |
Incidence did not increase in 1976–1992. Relative excess in 1989–1992 was not significantly different from zero |
Hjalmars and others (1994) |
Incidence |
Sweden: children 0–15 |
Fallout from Chernobyl |
1980–1992 |
137Cs contamination by geographic area |
Childhood acute leukemia |
888 |
No significant increase in childhood acute leukemia in contaminated areas |
Petridou and others (1994) |
Incidence |
Greece: children 0–14 |
Fallout from Chernobyl |
1980–1991 |
Three time periods: 1980–June 1986; July 1986–June 1988; July 1988–June 1991. Mean fallout levels (based on 137Cs measurements) grouped into 17 geographic regions |
Childhood leukemia |
968 |
No evidence of an increased incidence of childhood leukemia in periods after Chernobyl accident. No association between childhood leukemia and region by radiation fallout level |
Likhtarev and others (1995) |
Incidence |
Ukraine: children ages 0–14 |
Fallout from Chernobyl |
1986–1993 |
Calendar year; 7 geographic zones defined by estimated average thyroid dose to children |
Thyroid cancer |
418 cases in 0–14 year olds; 248 cases in those 15 and older |
Increase beginning in 1989; rate in 1993 was 5 times higher than 1986; higher incidence in zones with higher contamination levels |
Prisyazhiuk and others (1995) |
Incidence |
Four districts in Ukraine: Naroditchy, Ovrutch, Ivankov, Polesskoye |
Fallout from Chernobyl |
1980–1993 |
Three time periods: 1980–1985 (before accident); 1986–1993 (after accident); 1980–1993 |
All cancer; leukemia and lymphoma; thyroid cancer |
Not given |
Statistically significant increase in thyroid cancer after the accident; no significant increase in all cancer, or leukemia and lymphoma |
Stsjazhko and others (1995) |
Incidence |
Belarus, Russia, Ukraine |
Fallout from Chernobyl |
1981–1994 |
Three time periods: 1981–1985 (before accident); 1986–1990; 1991–1994; 6 geographic regions |
Thyroid cancer |
Since the accident: Belarus, 333; Russia, 23; Ukraine, 209 |
Increase in thyroid cancer incidence after the accident; most pronounced in most heavily contaminated areas |
Sugenoya and others (1995) |
Prevalence |
Two cities in Belarus (Chechelsk and Bobruisk): children ages 10–15 |
Fallout from Chernobyl |
October 1991–August 1992 |
Contamination levels (137Cs): Chechelsk 5–>40 Ci/km2; Bobruisk, control area |
Thyroid abnormalities |
888 screened in Chechelsk; 521 screened in Bobruisk |
Significantly higher prevalence of multiple micronodular lesions in diffuse goiter in contaminated city |
Gunay and others (1996) |
Incidence |
Bursa, Turkey: pediatric cases of malignancy |
Fallout from Chernobyl |
1986–1995 |
Calendar year, 1986–1995 |
Acute leukemia, lymphoma, solid tumors |
Acute leukemia: 101; lymphoma: 44; solid tumor: 31 |
Significant increase in acute leukemia after 1986; no significant increase in lymphoma or solid tumor |
Ivanov and others (1996) |
Incidence |
Seven regions of Belarus: children 0–15 |
Fallout from Chernobyl |
1982–1994 |
Calendar year; 7 geographic regions |
Childhood leukemia |
Not given |
No increase associated with calendar; no difference in rates by geographic region |
Kumpusalo and others (1996) |
Prevalence |
Two villages in Bryansk region of Russia (Mirnyi and Krasnyi): residents ages 3–34 |
Fallout from Chernobyl |
1993 |
Contaminated area (Mirnyi) and control area (Krasnyi) |
Thyroid ultrasound findings |
302 screened in Mirnyi; 200 screened in Krasnyi |
No pathological U.S. findings in either city. Prevalence of thyroid abnormalities higher in contaminated area: ages 0–9, 8.1% in Mirnyi; 1.6% in Krasnyi |
Reference |
Incidence/Mortality |
Population Studied |
Type of Exposure |
Dates of Accrual |
Type of Dosimetry |
Outcomes Studied |
Number of Cases |
Summary of Results |
Parkin and others (1996) |
Incidence |
34 regions of 23 countries in Europe: children 0–15 |
Fallout from Chernobyl |
1980–1991 |
Estimated dose (effective equivalent dose) in 34 countries or regions, obtained from UNSCEAR for first year, 0–4 years and 0–70 years after the accident |
Childhood leukemia |
25,820 |
Small increase in incidence over time, but no association in the period 1987–1991 with estimated dose |
Petridou and others (1996) |
Incidence |
Greece: children ages 0–4 |
Fallout from Chernobyl |
1980–1994 |
Three geographic areas based on 137Cs measurements. Children born during second half of 1986 and all of 1987 considered exposed in utero; remaining births considered unexposed |
Infant leukemia |
Exposed in utero: 55 unexposed in utero: 297 |
Incidence in those exposed in utero 2.6 times higher than unexposed (95% CI 1.4, 5.1) Children born to mothers resident in high-contamination areas were at significantly higher risk of leukemia. No increase associated with preconception exposure of parents |
Remennik and others (1996) |
Incidence |
Bryansk, Kaluga, Tula, Orel, Ryazan, Kursk regions of Russia |
Fallout from Chernobyl |
1981–1994 |
Calendar year, geographic region |
All malignancies; thyroid cancer |
Not given |
Increase in thyroid cancer incidence in children in the Bryansk region after the accident. Incidence of all cancer higher in 6 study regions than all of Russia since 1987 |
Tondel and others (1996) |
Incidence |
Six most contaminated counties of Sweden: persons age 0–19 |
Fallout from Chernobyl |
1978–1992 |
Three levels of contamination based on 137Cs measurements; two time periods: 1978–1986 and 1987–1992 |
Brain cancer, acute lymphatic leukemia, other neoplasms |
746 |
No clear associations between cancer incidence and level of radiation contamination; or increases over time since the accident |
Ashizawa and others (1997) |
Prevalence |
119,178 children examined at 5 centers in Ukraine, Belarus, and Russia |
Fallout from Chernobyl |
Examined May 15, 1991–April 30, 1996 |
Five geographic areas |
Goiter |
42,470 |
Variation in prevalence by region: highest in Kiev (54%); lowest in Gomel (18%). Significant inverse association between goiter and median urine iodine level |
Bleuer and others (1997) |
Incidence |
Belarus: individuals born 1963 or later |
Fallout from Chernobyl |
1986–1995 |
131I contamination levels by raion(geographic unit comparable to a county in the U.S.) |
Thyroid cancer |
528 in persons <15 in 1986 |
Feasibility study to assess the use of existing data to evaluate geographic and time trends—no conclusions risk of thyroid cancer |
Ivanov and others (1997a) |
Incidence and mortality |
Kaluga Oblast, Russia |
Fallout from Chernobyl |
1981–1995 |
Contaminated areas of oblast, based on 137Cs measurements and estimates of whole-body doses by settlement; calendar time (by year and grouped to reflect before and after accident |
All cancer, 8 specific cancer sites; grouped as GI, respiratory, leukemia |
All cancer: 2052; GI cancer: 808; respiratory cancer: 446; leukemia: 35 |
Time trends similar before and after the accident. Significant increase in thyroid cancer incidence in women after the accident |
Kasatkina and others (1997) |
Prevalence |
Two rural areas of the Orel region of Russsia: Uritzky and Kolpnyansky regions. Two samples of children with enlarged thyroids by palpation: (1) 2–3 trimester gestation—1 year at exposure; (2) ages 8–9 at exposure |
Fallout from Chernobyl |
Not given |
Two regions based on 137Cs contamination |
Endemic goiter; thyroid volume; cytology; thyroid autoantibodies |
Goiter: 88 in contaminated area; 20 in control area |
Five times the prevalence of thyroid enlargement in contaminated vs. control area. No evidence of thyroid dysfunction. Higher prevalence of autoantibodies and greater cellular proliferation in contaminated areas |
Lazjuk and others (1997) |
Incidence |
Three regions of Belarus: Gomel and Mogilev (contaminated, based on 137Cs measurements) and Minsk city (control) |
Fallout from Chernobyl |
1982–1994 |
Two contaminated and one control region; 1982–1985 (before), 1987–1994 (after) |
Congenital anomalies: total and 9 specific types |
Total: contaminated areas: before 1201, after 2561; control before 255, after 649 |
Increase in congenital and fetal abnormalities in contaminated regions (1.6-fold increase based on examination of abortuses; 1.8-fold based on examination of tissues). Most increased were multiple congenital malformations, polydactyly, reduction limb defects |
Lomat and others (1997) |
Incidence |
Belarus: children 0–14 |
Fallout from Chernobyl |
1986–1995 |
Three groups: children evacuated from 30 km zone; children residing or moved into areas >15 Ci/km2; children born to parents in 30 km zone or resettled from areas >15 Ci/km2 |
All cancer; thyroid cancer; noncancer: digestive, endocrine, anemia, nervous system, respiratory diseases |
Not given |
Large increase in thyroid cancer from 1987 to 1995 (0.2 = 4.0 × 105). Also increases in the incidence of endocrine and dermatologic diseases and mental disorders. Largest increases in those evacuated from 30 km zone |
Michaelis and others (1997) |
Incidence |
Germany: children 0–15 |
Fallout from Chernobyl |
Children born 1980–1990 |
Three levels of ground deposition of 137Cs; three time periods: born 7/1/86–12/31/87 (exposed); 1/1/80–12/31/85 and 1/1/88–12/31/90 (unexposed) |
Infant leukemia |
Exposed: 35; unexposed: 143 |
Higher incidence in exposed than unexposed cohorts: RR 1.48 (95% CI 1.02, 2.15). Subgroup analyses not consistent with a relationship to exposure levels |
Reference |
Incidence/Mortality |
Population Studied |
Type of Exposure |
Dates of Accrual |
Type of Dosimetry |
Outcomes Studied |
Number of Cases |
Summary of Results |
Pacini and others (1997) |
Incidence |
Belarus: individuals diagnosed with thyroid cancer from May 1986–December 1995 under age 21 |
Fallout from Chernobyl |
May 1986–December 1995 |
Geographic distribution of radioactive contamination |
Thyroid cancer |
472 (372 <14 age; 100 14–21) |
Excess thyroid cancer in both children and adolescents: cases age 5 or under account for majority; youngest ages have greatest risk |
Sobolev and others (1997) |
Incidence |
Ukraine: children born 1968–1986 |
Fallout from Chernobyl |
1986–1995 |
Calendar year; four geographic regions based on estimated thyroid doses |
Thyroid cancer |
2077 |
Increase in incidence of thyroid cancer; especially in youngest age group and highest-dose area |
Vykhuvanets and others (1997) |
Prevalence |
Ukraine: 53 children age 7–14 in 15 contaminated settlements in Chernigov and Kiev regions; 45 children age 6–14 in uncontaminated areas of Poltava region |
Fallout from Chernobyl |
Examined 1993–1994 |
Estimated thyroid doses based on information from the Ministry of Health regarding personal absorption doses, average absorption doses, and average summary doses |
Thyroid ultrasound, autoantibodies, thyroid hormones, lymphocyte subsets |
NA |
Significant association between autoimmune thyroid disorders and radiation dose |
Ivanov and others (1998) |
Incidence |
Belarus: children born 1982–1994 |
Fallout from Chernobyl |
Born 1982–1994 |
Three time periods: 7/1/86–12/31/87 (exposed); 1/1/82–12/31/85 and 1/1/88–12/31/94 (unexposed); all Belarus, Mogilev, and Gomel combined |
Infant leukemia |
Exposed 17; unexposed 89 |
Slight increase in infant leukemia in exposed: Belarus RR 1.26 (95% CI 0.76, 2.1); Mogilev and Gomel RR 1.51 (0.63, 3.6) |
Jacob and others (1998) |
Incidence |
Three regions of Ukraine: Kiev, Zhytomyr, and Chernigov; three regions in Belarus: Gomel/Mogilev, Minsk city, Gomel city; Bryansk region of Russia |
131I from Chernobyl |
1991–1995 |
Estimated average thyroid dose by region |
Thyroid cancer |
Ukraine 175; Belarus 201; Russia 31 |
EAR per 104 PY per gray ranges from 0.9 in Zhytomyr to 3.8 in Kiev city; in Belarus from 2.3 in Minsk city to 3.1 in Mogilev/Gomel; in Bryansk is 2.7 |
Pacini and others (1998) |
Prevalence |
Two areas of Belarus: Hoiniki and Braslav; children 12 at time of accident |
Fallout from Chernobyl |
Examined 1992–1994 |
Contaminated area (Hoiniki) and uncontaminated area (Braslav) |
Thyroid autoantibodies, thyroid hormones |
Hoiniki: 287 examined; Braslav: 208 examined |
Significantly higher prevalence of autoantibodies in girls in Hoiniki; highest in ages 9 years or older. No increase in free T4, free T3, or TSH in contaminated area |
Steiner and others (1998) |
Incidence |
West Germany, children age <15 |
Fallout from Chernobyl |
1980–1990 |
Three time periods: 7/1/86–12/31/87 (exposed); 1/1/80–12/31/85 and 1/1/88–12/31/90 (unexposed) |
Infant leukemia |
Exposed 325; unexposed 1934 |
Slight increase in leukemia in exposed areas: Overall RR 1.48 (95% CI 1.21, 2.78). No clear trend in incidence associated with exposure |
Ivanov and others (1999) |
Incidence |
Bryansk, Tula, Kaluga, Orel regions of Russia, population ages 0–60 |
Fallout from Chernobyl |
1982–1996 |
Three time periods: 1982–1986; 1986–1990; 1991–1996. Thyroid doses estimated for those age 0–17 in Bryansk |
Thyroid cancer |
3082 cases |
Highest risk in children up to age 4 at exposure (14 times risk in adults). EAR in those 0–17 in Bryansk in girls is 2.21 (95% CI 0.74, 3.68) per 104 PY per gray and for boys is 1.62 (−0.04, 3.23) |
Jacob and others (1999) |
Incidence |
Belarus: 2 cities and 2122 settlements; Bryansk, Russia: 1 city and 607 settlements; persons born 1971–1985 |
131I from Chernobyl |
1991–1995 |
Thyroid dose from 131I by settlement or city |
Thyroid cancer |
243 cases |
EAR 2.1 per 104 PY per gray Excess relative risk (ERR) 23 per Gy−1. No differences by countries or cities or rural areas |
Kofler and others (1999) |
Incidence |
Belarus: children 0–15 at the time of the accident |
Fallout from Chernobyl |
1986–1997 |
Calendar year and geographic area (raion) |
Thyroid cancer |
805 cases |
Shorter latency periods (4–5 years) in areas with higher exposure and higher incidence rates |
Tronko and others (1999) |
Incidence |
Ukraine: 27 regions, children 0–18, cases treated at Institute of Endocrinology and Metabolism in Kiev |
Fallout from Chernobyl |
1986–1997 |
Calendar year and geographic area (raion). Estimated thyroid dose based on contamination levels |
Thyroid cancer |
577 cases |
Significant increase in incidence after the accident. Most affected group was 5 years in 1986. Largest increase in those with estimated dose 0.5 Gy |
Vermiglio and others (1999) |
Prevalence |
Tula region of Russia, 143 iodine-deficient children 5–15 years of age from moderately contaminated area and 40 sex-and age-matched children from nearby uncontaminated area |
Fallout from Chernobyl |
Not stated |
137Cs contamination by geographic area |
Thyroid autoantibodies |
183 children and adolescents examined |
High prevalence of autoimmunity in persons from contaminated area, most notably in those born or in utero in 1986 |
Heidenreich and others (2000) |
Incidence |
Ukraine: individuals born 1968–1997 |
Fallout from Chernobyl |
1986–1998 |
Age at exposure and time since exposure |
Thyroid cancer |
Not given |
No increase for 3 years after accident; then linear excess absolute risk through 1998; decrease in excess risk with increasing age at exposure |
Jacob and others (2000) |
Incidence |
Belarus: 2 cities and 2122 settlements; individuals born 1971–1986 |
131I from Chernobyl |
1991–1996 |
Thyroid dose from 131I by settlement or city |
Thyroid cancer |
657 cases |
Number of cases in first decade after accident is a small fraction of what is expected in the following four decades |
Reference |
Incidence/Mortality |
Population Studied |
Type of Exposure |
Dates of Accrual |
Type of Dosimetry |
Outcomes Studied |
Number of Cases |
Summary of Results |
Romanenko and others (2000) |
Prevalence |
236 patients with renal cell carcinoma in Ukraine (Institute of Urology and Nephrology in Kiev) and 112 patients in Spain (University Hospital in Valencia) |
Fallout from Chernobyl |
Kiev: 1993–1999 Valencia: 1975–1989 |
137Cs contamination level by geographic area |
Renal cell carcinoma pathology and cell proliferation activity |
Kiev 236 cases; Valencia 112 cases |
Significant and strong increase of proliferative activity and aggressivity in Ukraine cases |
Noshenko and others (2001) |
Incidence |
Zhitomir and Poltava regions of Ukraine: children born in 1986 |
Fallout from Chernobyl |
1986–1996 |
Two geographic areas: Zhitomir—contaminated and Poltava—uncontaminated |
Acute leukemia |
Zhitomir 21 cases; Poltava 8 |
Incidence rates in contaminated region increased relative to uncontaminated region |
Shibata and others (2001) |
Prevalence |
Four districts of the Gomel region of Belarus: children born 1983–1989 |
Fallout from Chernobyl |
Screened 2/2/98–12/22/00 |
Three time periods: Group III 1/1/83–4/26/86; Group II 4/27/86–12/31/86; Group I 1/1/87–12/31/89 |
Thyroid cancer |
32 cases |
Significant increase relative to Group I in Group III (OR 121; 95% CI 9, 31,000) and Group II (OR 11; 3, 176). Increase was highest in those youngest in 1986 |
Romanenko and others (2002) |
Prevalence |
Two areas in Ukraine (not specified): males with benign prostatic hyperplasia and females with chronic cystitis, living more than 15 years in the geographic areas selected |
Fallout from Chernobyl |
1999–2000 |
Two geographic areas: contaminated and uncontaminated with 137Cs; measurements of 137Cs in urine for some patients |
DNA damage repair indicators in bladder urothelium |
156 males; 48 females |
Significant activation of DNA damage repair in persons from contaminated area compared to those from uncontaminated area |
Tronko and others (2002) |
Incidence |
Ukraine: persons 0–18 in 1986 |
Fallout from Chernobyl |
1986–2000 |
Calendar year; two geographic areas: (1) 6 regions contaminated and (2) 21 remaining regions |
Thyroid cancer |
Ages 0–18:1876 cases; ages 0–15:1318 cases |
Steady rise in incidence after the accident in 6 contaminated regions. Greatest increase in those ages 0–4 in 1986 |
Ivanov and others (2003) |
Incidence |
Bryansk region of Russia: residents age 15–69 at accident |
Fallout from Chernobyl |
1986–1998 |
Mean thyroid doses by raion |
Thyroid cancer |
1051 cases |
Incidence relative to Russia: twofold higher. ERR at 1 Gy, based on external controls: −0.4 (males), −1.3 (females); based on internal controls: 0.7 (males), −0.9 (females) |
TABLE 9-3B Populations Exposed from the Chernobyl Accident—Case-Control Studies
Reference |
Population Studied |
Number of Subjects |
Dates of Accrual |
Type of Exposure |
Type of Dosimetry |
Summary of Results |
||
Cases |
Controls |
Cases |
Controls |
|||||
Astakhova and others (1998) |
Thyroid cancer in children in Belarus <15 at the time of the accident |
Type I: Random sample of children in contaminated raions Type II: Sample of children with same opportunity for diagnosis as cases Both types matched on age, sex, rural or urban residence in 1986 |
107 |
Type I: 107 Type II: 107 |
1987–1992 |
Chernobyl fallout: major contributor to thyroid dose is 131I. Lesser contributions from 132I, and 133I, and external radiation |
Retrospective dose reconstruction. Thyroid dose estimated for individuals based on settlement doses for most cases and controls. For 12 cases (no controls) dose was estimated based on thyroid measurements |
Significant differences between cases and both sets of controls regarding dose. Strong and significant dose-response relationship. Odds ratio (highest- vs. lowest-dose group) in Gomel, Type I controls: rural 10.4 (3.5, 31.2); urban 5.1 (1.3, 20.0) |
Noshchenko and others (2002) |
Leukemia in children age 0–20 at the time of the accident in Zhytomir and Rivno Oblasts in Ukraine |
Two controls per case, randomly selected from the same oblast as the case but not the same raion, matched on age, sex, type of settlement |
98 |
151 |
1987–1997 |
Chernobyl fallout: major contributor to bone marrow dose is external gamma from fallout and ingestion of 134Cs and 137Cs with food |
Retrospective dose reconstruction. Individual accumulated dose to bone marrow estimated, based on settlement measurements and individual dosimetry interviews |
Statistically significant risk (OR for >10 mSv 2.5; CI 1.1, 5.4). Higher risk in males. Risk highest 1993–1997 (OR 4.1; 1.5–11.3), especially for acute lymphoblastic type (OR 13.1; 2.6–65.0) |
Davis and others (2004b) |
Thyroid cancer in children 0–19 at the time of the accident residing in 7 most contaminated raions in the Bryansk Oblast of Russia |
Two controls per case, randomly selected from the same raion as the case, matched on age, sex, type of settlement |
26 |
52 |
April 26, 1986–October 1, 1997 |
Chernobyl fallout: major contributor to thyroid dose is 131I. Lesser contributions from 132I, and 133I, and external radiation |
Retrospective dose reconstruction. Individual accumulated dose to thyroid estimated, based on environmental measurements and individual dosimetry interviews |
Significant dose response (p < .009). OR by dose quartile: 3–60 mGy, 1.0; 66–240 mGy, 1.65 (0.3, 8.5); 290–600 mGy, 3/05 (0.4, 22.1); 610–M 2730 mGy, 44.7 (3.3, 604) |
Cardis and others (2005b) |
Thyroid cancer in children age 0–14 at the time of the accident in Belarus (Gomel and Mogilev) and 0–18 in Russia (Kaluga, Tula, Orel, Bryansk) |
Randomly selected from the same oblast as the case, matched on age and sex |
276 |
1300 |
1992–1998 |
Chernobyl fallout: major contributor to thyroid dose is 131I. Lesser contributions from short-lived isotopes of iodine and tellurium and external radiation from long-lived radionuclides |
Retrospective dose reconstruction. Individual accumulated dose to thyroid estimated, based on environmental measurements and individual dosimetry interviews |
Significant dose-response linear up to 1.5–2 Gy. RR at 1 Gy 5.5 (95% CI 3.1, 9.5). Significant effects of iodine deficiency and iodine supplementation as modifiers of RR per gray |
garding the quantitative relationship between radiation dose to the thyroid from Chernobyl and the risk of thyroid cancer.
There are only three published population-based case-control studies of thyroid cancer in children that utilize individual estimates of radiation dose and provide quantitative information on thyroid cancer risk (Table 9-3B). The first is based on 107 cases diagnosed in Belarus (Astakhova and others 1998). Although a strong relationship between estimated radiation dose and thyroid cancer was found, thyroid doses were inferred for children from estimates for adults who lived in the same villages. The second is based on confirmed cases of thyroid cancer in children and adolescents aged 0–19 years at the time of the accident, residing in the more highly contaminated areas of the Bryansk Oblast of Russia (Davis and others 2004b).
Based on 26 cases and 52 controls and using a log-linear dose-response model treating estimated individual thyroid radiation dose as a continuous variable, the trend of increasing risk with increasing dose was statistically significant (one-sided p = .009). The third is a population-based, case-control study of thyroid cancer carried out in contaminated regions of Belarus and the Russian Federation (Cardis and others, 2005). The study included 276 cases and 1300 matched controls aged less than 15 years at the time of the accident. Individual doses were calculated for each subject. A very strong dose-response relationship was observed in this study (p < .0001). At 1 Gy, the odds ratio (OR) varied from 5.5 (95% CI 3.1, 9.5) to 8.4 (95% CI 4.1, 17.3) depending on the form of the risk model used. A clear linear dose-response relationship was observed up to about 1 Gy, followed by a marked flattening. The risk appeared to be related mainly to exposure to 131I. Collectively, data from these studies suggest that exposure to radiation from Chernobyl is associated with an increased risk of thyroid cancer and that the relationship is dose dependent. These findings are consistent with descriptive reports from contaminated areas of Ukraine and Belarus, and the quantitative estimate of thyroid cancer risk is generally consistent with estimates from other radiation-exposed populations.
A number of the studies have also focused on the potentially modifying influence of a number of host and environmental factors. Results from studies of atomic bomb survivors and persons exposed to external irradiation have shown that exposure at the youngest ages is associated with the greatest risk of thyroid cancer. The available data on exposure from the Chernobyl accident are largely in agreement with this observation. For example, a recent paper (Tronko and others 2002) found the highest incidence of thyroid cancer among those exposed at ages 0–4, who also had the highest doses. There have been few studies in persons exposed at older ages, however. One study of thyroid cancer diagnosed in adolescents and adults in the Bryansk region of Russia reported a small excess of thyroid cancer among adults (Ivanov and others 2003), but the excess was not correlated with the imputed doses, and larger studies with longer follow-up and greater statistical power are needed. It has also been postulated that the risk of thyroid cancer may be especially high among persons exposed in utero, because developing fetal thyroid tissue may be highly susceptible to thyroid cancer induction by 131I exposure. At present there are no data available from Chernobyl regarding the risk of thyroid cancer from in utero exposure.
Fifteen years after the Chernobyl accident, thyroid cancer incidence is still highly elevated. Although based on studies of thyroid cancer in other radiation-exposed populations there is no reason to expect a decrease in the next several years; at the present time the follow-up of Chernobyl-exposed children is too short to determine long-term risks. An increase in thyroid cancer has been observed in both males and females. Most, but not all, of the Chernobyl studies have reported similar relative risks per unit dose for males and females.
Iodine deficiency may also be an important modifier of the risk of radiation-induced thyroid cancer. Some regions contaminated by the Chernobyl accident are areas of mild to moderate iodine deficiency. To date, only two published studies have investigated the relationship between iodine deficiency, radiation dose, and the risk of thyroid cancer in young people. In a study carried out in the Bryansk region of Russia, Shakhtarin and colleagues (2003) report a significantly increased risk of thyroid cancer with increasing radiation dose from Chernobyl that was inversely associated with urinary iodine excretion levels. At 1 Gy, the ERR in territories with severe iodine deficiency was approximately two times that in areas of normal iodine intake, thereby suggesting that iodine deficiency may enhance the risk of thyroid cancer following radiation exposure. The evidence is not conclusive because the study is ecologic and uses approximations for both radiation dose and iodine deficiency. In their case-control study in Belarus and Russia, Cardis and colleagues (2005) also investigated the effects of iodine deficiency and its interaction with radiation exposure in the risk of thyroid cancer. Subjects who resided in the areas of lowest soil iodine content had a 3.1 times (95% CI 1.7, 5.4) higher risk at 1 Gy than subjects residing in areas of higher soil iodine content. It is noted that administration of potassium iodide as a dietary supplement significantly reduced the risk of radiation-induced thyroid cancer.
Finally, relatively little has been published regarding thyroid outcomes other than thyroid cancer, although one study has reported an elevated risk of benign thyroid tumors (Ivanov and others 2003). There have been reports of increases in autoimmune disease and antithyroid antibodies following childhood exposure to Chernobyl (Lomat and others 1997; Vykhovanets and others, 1997; Pacini and others 1998; Vermiglio and others 1999). However, a study by the Sasakawa Foundation, which screened 114,000 children, found no association between a surrogate for thyroid dose
(137Cs) and thyroid antibodies, hypothyroidism, hyperthyroidism, or goitre (Ashizawa and others 1997).
Leukemia
The evidence from epidemiologic studies regarding the risk of leukemia in populations exposed to radiation from Chernobyl comes from studies of recovery operation workers, some of whom were exposed at a high or moderate dose levels and dose rates (depending on when and where they worked), and the general population who have been subject to low-dose-rate exposure (primarily from 137Cs) for a number of years and will continue to be exposed in the future. Worker populations were exposed as adults and are considered in Chapter 8. Resident populations were exposed at all ages, but studies of residents are primarily of persons exposed as children and/or in utero.
Several studies have investigated the risk of leukemia in children exposed to Chernobyl fallout in utero. All are ecologic in design, and results are inconsistent. The initial study compared rates for temporal cohorts born during “exposed” and “unexposed” periods in Greece and found a 2.6-fold increase in leukemia risk and elevated rates for those born in regions with higher levels of radioactive fallout (Petridou and others 1996). However, the numbers of cases in each exposure group were small, and the results could not be duplicated when a similar approach comparing areas with the same categories of contamination (<6 kBq m−2, 6–10 kBq m−2, >10 kBq m−2) was applied to the analysis of data from the German Childhood Cancer Registry (Steiner and others 1998).
In a study in Belarus (Ivanov and others 1998), where levels of contamination are higher by a factor of 10 or more, the results were similar to the Greek study but the trend was weaker. Nevertheless, although the findings are based on small numbers and are not statistically significant, the highest annual incidence rate was in 1987, the year after the accident, and the largest rate ratio (RR = 1.51; 95% CI 0.63, 3.61) was in the two most contaminated regions: Gomel and Mogilev.
A more recent small study published by Noshchenko and colleagues (2001) compared leukemia incidence during 1986 to 1996 among children born in 1986 and thus exposed in utero in Zhitomir, a contaminated region, to children born in Poltava, a relatively uncontaminated region. The reported risk ratios based on cumulative incidence show significant increases for all leukemia (relative risk [RR]2 =2.7; 95% CI 1.9, 3.8) and for the subtype of acute lymphoblastic leukemia (RR = 3.4; 95% CI 1.1, 10.4).
The ongoing European Childhood Leukemia-Lymphoma Incidence Study (ECLIS) has evaluated the risk of leukemia by age using data from population-based cancer registries in Europe (including Belarus and Ukraine). Focusing on the risk of leukemia by age of diagnosis in 6-month intervals in relation to estimated doses from the Chernobyl fallout received in utero, preliminary results suggest a small increase in risk in infant leukemia and leukemia diagnosed between 24 and 29 months.
Thus, at present the available evidence from ecologic studies does not convincingly indicate an increased risk of leukemia among persons exposed in utero to radiation from Chernobyl. However, the statistical power of these studies is low for detecting moderate-sized associations, and the exposure measures are crude. There are no data from analytic epidemiologic studies in which individual dose estimates are available. Consequently, there is neither strong evidence for or against an association between in utero exposure to Chernobyl fallout and an increased risk of leukemia.
Several ecologic studies also have investigated the association between radiation exposure of children from Chernobyl and the occurrence of leukemia. The ECLIS utilized incidence data in children under age 15 from 36 cancer registries in 23 countries. Parkin and colleagues (1996) compared acute leukemia incidence rates before the Chernobyl accident (1980–1985) with those for 1987 and 1988. Although the number of leukemia cases for 1987–1988 significantly exceeded the number of cases expected on the basis of 1980–1985 data, there was no evidence that the excess in leukemia rates was more pronounced in areas that were most affected by Chernobyl-related ionizing radiation exposure. Similar results were observed in the 5-year ECLIS followup report.
Additional reports have focused on changes in childhood leukemia rates before and after the accident in individual European countries and elsewhere. Overall, there was little evidence for an increase in rates of childhood leukemia in Ukraine, Belarus, Russia, Finland, Sweden, Greece, or a number of other countries from Central, Eastern and Southern Europe after the Chernobyl accident. Furthermore, there was no association between the extent of contamination and the increase in risk in these countries. However, one Swedish study (Tondel and others 1996), reported a non-statistically significant increase of acute lymphocytic leukemia (ALL) after the accident in children younger than 5 (OR 1.5; 95% CI 0.8, 2.6). A small study in northern Turkey showed that in one pediatric cancer treatment center, more patients with ALL were seen after the accident than before, but no incidence rates were reported (Gunay and others 1996).
There has been only one analytic (case-control) study of childhood leukemia reported (Noshchenko and others 2002) based on cases identified among residents of the Rivno and Zhytomir Oblasts in Ukraine. Cases were under age 20 at the time of the accident and were diagnosed between 1987 and 1997. Data were collected on 272 cases; however the analysis was based on only 98 cases that were independently verified and interviewed. Controls were selected randomly from
the same oblasts, excluding the raion of residence of the case, and matched according to age at the time of the accident, sex, and type of settlement. The mean estimated dose to the bone marrow among study subjects was 4.5 mSv and the maximum was 101 mSv. The study found a statistically significant increased risk of acute leukemia among males with cumulative doses greater than 10 mSv diagnosed from 1993 to 1997. A similar association was found for acute myeloid leukemia (AML) diagnosed in 1987–1992. These results should be interpreted cautiously, however, because they are based only on approximately one-third of the cases and a lesser proportion of controls, and it is not clear whether cases and controls were selected for dose estimation in an unbiased manner.
On balance, the existing evidence does not support the conclusion that rates of childhood leukemia have increased as a result of radiation exposures from the Chernobyl accident. However, ecologic studies are not particularly sensitive to detecting relatively small changes in the incidence of a disease as uncommon as childhood leukemia over time or by different geographic areas. Further, existing descriptive studies vary in several aspects of study design: methods of case ascertainment (cancer registries versus retrospective record review), methods of classifying radiation exposure, and length of follow-up after the accident (range 2–10 years). The single analytical study is insufficient to draw convincing conclusions regarding leukemia risk after Chernobyl exposure of children.
A few studies have investigated adult resident populations living in highly contaminated areas. Osechinsky and Martirosor (1995) investigated the incidence of leukemia and lymphoma in the general population of the Bryansk region of Russia for 1979–1993 using an ad hoc registry of hematological diseases established after the Chernobyl accident. The incidence rates in the six most contaminated districts (more than 37 kBq m2 of 137Cs deposition density) did not exceed the rates in the rest of the region or in Bryansk city, where the highest rates were observed. Comparisons of crude incidence rates before and after the accident (1979–1985 and 1986–1993) showed a significant increase in the incidence of all leukemia and non-Hodgkin’s lymphoma (NHL), but this was due mainly to increases in the older age groups in rural areas. The incidence of childhood leukemia and NHL was not significantly different in the six most contaminated areas from the incidence in the rest of the region. Similarly, Ivanov and colleagues (1997a, 1997b) found no evidence of an increase in leukemia rates in the most contaminated areas of the Kaluga district of the Russian Federation after the Chernobyl accident.
In Ukraine, Bebeshko and colleagues (1997) examined incidence rates for leukemia and lymphoma in the most highly contaminated areas of the Zhytomir and Kiev districts before and after the Chernobyl accident. Total incidence in adults increased from 5.1 per 100,000 during 1980–1985 to 11 per 100,000 PY during 1992–1996, but there was no excess in contaminated areas of the regions. Similarly, Prisyazhniuk and colleagues (1995) investigated the incidence of leukemia and lymphoma in the three most contaminated regions of Ukraine. There was a steady increase in leukemia and lymphoma rates for both men and women between 1980 and 1993, but there was no evidence of a more pronounced increase after the accident.
Thus, on balance, there is no convincing evidence that the incidence of leukemia has increased in adult residents of the exposed populations that have been studied in Russia and Ukraine. However, few studies of the general adult population have been conducted to date, and they have employed ecologic designs that are relatively insensitive.
Solid Tumors Other Than Thyroid Cancer
There has been relatively little study of the incidence of or mortality from solid cancers other than thyroid cancer in populations exposed to radiation from the Chernobyl accident. Two studies have investigated solid cancer incidence in liquidation workers (Prisyazhnik and others 1996; Ivanov and others 2004a, 2004b) and are considered in Chapter 8. No descriptive or analytical epidemiologic studies of breast cancer risk in populations exposed to radiation from Chernobyl have been published in the peer-reviewed literature. However, one monograph report has cited elevated breast cancer incidence rates based on members of Ukrainian registries (Prysyazhnyuk and others 2002). These included 150,000 residents of contaminated areas close to Chernobyl; 90,000 liquidation workers in 1986 (with mean dose evaluated as 100–200 mSv) and 1987 (mean dose 50–100 mSv); and 50,000 evacuees from Pripyat (mean dose 10–12 mSv) and the 30 km zone (mean dose 20–32 mSv). For breast cancer among the women in these cohorts, the standardized incidence ratio (SIR), based on comparisons to Ukrainian female population rates, was reported as 1.50 (95% CI 1.27, 1.73) for 1993-1997 among residents of contaminated territories. For evacuees from the 30 km zone, the SIR during 1990-1997 was 1.38 (95% CI 1.06, 1.70), and for women who served as liquidation workers during 1986-1987, who comprised only about 5% of the liquidation worker cohort, the SIR for 1990-1997 was 1.51 (95% CI 1.06, 1.96). These registry-derived estimates must be interpreted with considerable caution because they were not subject to diagnostic confirmation and may be influenced by differences in screening intensity.
Similarly, although no descriptive or analytical epidemiologic studies of bladder or kidney cancer risk in relation to Chernobyl radiation have been published in the peer-reviewed literature, there has been a series of papers investigating aspects of possible radiation carcinogenesis in these organs. Romanenko and colleagues (2003) have continued to monitor the incidence of urinary bladder cancer in Ukraine, reporting that it increased from 26.2 to 43.3 per 100,000 PY between 1986 and 2001. In a study of 204
urothelial biopsies of Ukrainian patients, they concluded that activation of DNA damage repair was detected more frequently among residents of contaminated areas, compared to those of putatively uncontaminated areas (Romanenko and others 2002). Morimura and colleagues (2004) observing p53 gene mutations in 54.5% of 11 and 16.7% of 18 Ukrainian bladder cancers collected before and after the Chernobyl accident, respectively, suggesting the possibility of distinct molecular genetic pathways of bladder cancer induction before and after the accident. Romanenko and colleagues (2000) have also reported that renal carcinoma incidence has increased from 4.7 to 7.5 per 100,000 PY.
In summary, there is now little doubt that an excess of thyroid cancer has occurred in areas highly contaminated by radiation from the Chernobyl accident. Analytical studies further indicate that exposure to radiation from Chernobyl is associated with an increased risk of thyroid cancer and that the relationship is dose dependent. Quantitative estimates of risk from these studies are consistent with estimates from other radiation-exposed populations. There is evidence that young age at exposure and iodine deficiency may be important modifiers of the risk of radiation-induced thyroid cancer. There is no convincing evidence that the incidence of leukemia has increased in children or adult residents of the exposed populations; however, few studies of leukemia have been conducted to date and most have employed ecologic designs that are relatively insensitive. There have been very few studies of the incidence of or mortality from solid cancers other than thyroid cancer in populations exposed to radiation from the Chernobyl accident, and there is no evidence of an increase in any solid cancer type to date.
POPULATIONS EXPOSED FROM NATURAL BACKGROUND
Table 9-4 summarizes four studies of populations exposed from natural background radiation. Two were conducted in China, one in Great Britain, and one in India. A number of different cancer outcomes were studied, based on incidence, mortality, and prevalence data. These studies did not find higher disease rates in geographic areas with high background levels of radiation exposure compared to areas with lower background levels. However, these studies were ecologic in design and utilized population-based measures of exposure rather than individual estimates of radiation dose. Thus, they cannot provide any quantitative estimates of disease risk associated with the exposure levels found in the areas studied.
CHILDREN OF ADULTS EXPOSED TO RADIATION
Table 9-5A lists three ecologic studies of children of adults exposed to radiation. The focus is on preconception parental exposure and the risk of leukemia and lymphoma in
TABLE 9-4 Populations Exposed from Natural Background Radiation—Ecologic Studies
Reference |
Incidence/Mortality |
Population Studied |
Type of Exposure |
Dates of Accrual |
Type of Dosimetry |
Outcomes Studied |
Number of Cases |
Summary of Results |
Wang and others (1990) |
Prevalence |
Women ages 50–65 living in Yangjiang, China, vs. nearby control areas |
Natural background (mostly external whole-body gamma) |
1986 (survey) |
Measured external exposure (average annual dose in high-background area: 330 mR; in control area: 114 mR) |
Thryoid nodularity, serum thyroid hormone levels, chromosome aberrations |
Nodules in high areas (95); in control areas (93) |
No difference in prevalence of nodules; no difference in thyroid hormone levels; increased frequency of unstable chromosome aberrations |
Lu-xin (1994) |
Mortality |
Population of Yangjiang, China, vs. control area (not specified) |
Natural background radiation |
1970–1986 |
Measured annual external exposure (mR) |
11 cancer sites |
High-exposure area 914; control 1032 |
No increase in high-background areas except cervix |
Richardson and others (1995) |
Incidence |
Children under age 15 in Great Britain |
Natural background (gamma and radon) |
1969–1983 |
Survey of radon and gamma concentrations in homes; gamma outside; 459 districts |
Leukemia |
6691 |
No association of childhood leukemia with indoor or outdoor gamma levels |
Nair and others (1999) |
Incidence |
Population of Karunagappally tuluk in Kerala, India |
Thorium deposited along coastal areas (gamma) |
1990–1996 |
Gamma measurements made in each house |
All cancers |
Not given |
No evidence of higher incidence of cancer in areas of |
TABLE 9-5A Children of Adults Exposed to Radiation—Ecologic Studies
Reference |
Incidence/Mortality |
Population Studied |
Type of Exposure |
Dates of Accrual |
Type of Dosimetry |
Outcomes Studied |
Number of Cases |
Summary of Results |
Kinlen (1993a) |
Incidence |
Residents of Seascale below age 25 in 1951–1991 |
Paternal preconception whole-body dose |
1951–1991 |
Lifetime preconception dose obtained from employment records (mSv) |
Leukemia and NHL |
Leukemia: 5 in Seascale; NHL: 3 in Seascale |
Significant excess of leukemia and NHL in Seascale among those born in Seascale, and those born elsewhere |
Parker and others (1993) |
NA |
Children born in Cumbria from 1950 to 1989 to fathers employed at Sellafield |
Paternal preconception whole-body dose |
NA |
Total cumulative and 6-month preconception dose, obtained from employment records |
Radiation doses (no diseas |
|
preconception dose is associated with children born in Seascale; mean individual preconception doses consistently lower in Seascale |
Wakeford and Parker (1996) |
Incidence |
Residents of West Cumbria under age 25 |
Paternal preconception whole-body dose |
1968–1985 |
Cumulative preconception dose obtained from worker records |
Leukemia |
41 |
Increased incidence in some groups defined by area and age; no increase associated with paternal preconception dose |
the offspring of exposed parents. These studies followed the findings first published by Gardner and colleagues (Gardner and others 1990a, 1990b) suggesting that an excess incidence of leukemia in children in West Cumbria may be due to parental preconception exposure to ionizing radiation during employment at the nearby Sellafield nuclear fuel processing plant. All three studies were conducted in relation to exposures received by parents working at the Sellafield nuclear facility in Great Britain. One study (Parker and others 1993) is a radioecologic study, examining the distribution of possible doses received by fathers employed at Sellafield of children born in Cumbria from 1950 to 1989; it does not address disease outcome. Although there is some evidence of an increased risk associated with measures of individual dose in the other two studies, the findings are based on very small numbers of cases and the results across studies are not consistent.
A larger number of case-control studies have been conducted to investigate the possible relationship between radiation exposure of adults and subsequent cancer in their offspring. Table 9-5B summarizes the results of seven published case-control studies. Six of the seven studies included in the table are investigations that are related to findings first published by Gardner and colleagues (1990b). The six studies summarized here include investigations in England and Wales, Scotland, and Canada. All but one investigated leukemia and/or childhood cancer. The seventh study by Sever and colleagues (1988) is a study of congenital malformations. All but the study by Sorahan and Roberts (1993) used employment records and recorded doses to estimate individual preconception radiation dose. The study by Sorahan and Roberts (1993) used job histories to estimate paternal exposure to ionizing radiation and the potential for exposure to radionuclides in the 6 months prior to the conception of 14,869 children dying of cancer. For all childhood cancers, the RR was 2.9 (95% CI 1.2, 7.1) for those potentially exposed to radionuclides. There was no evidence of an association between external ionizing radiation and cancer risk. The most recent study by Draper and colleagues (1997) found an increased risk of childhood leukemia and NHL among children whose fathers were radiation workers (RR 1.8; 95% CI 1.1, 3.0). The risk was also elevated for all other childhood cancers among offspring of mothers who were radiation workers (RR 5.0; 95% CI 1.4, 26.9). There was no evidence of a dose-response trend. In summary, none of the studies provides quantitative information from dose-response analyses or quantitative estimates of the risk of disease associated with exposure, and results across studies are inconsistent.
Table 9-5C describes cohort studies published regarding the risk of cancer and adverse reproductive outcomes in children of adults exposed to radiation. Two are studies by Gardner and colleagues (1987) that are not based on individual estimates of radiation dose but rather on proximity to the Sellafield nuclear plant at different ages (at birth and while attending school). A third (Roman and others 1999) is an attempt to confirm Gardner’s findings of an increased risk of leukemia and lymphoma in children born to fathers with preconception radiation exposure. Individual paternal preconception exposure was estimated from employment records. Person-years at risk were accrued from date of birth for 39,557 children of male workers and 8883 children of female workers until age 25, cancer diagnosis, or death. A total of 111 cases of malignant cancer were found, but there was no evidence of increased risk relative to the general population. Rate ratios for all cancers (adjusted for calendar
TABLE 9-5B Children of Adults Exposed to Radiation—Case-Control Studies
Reference |
Population Studied |
Number of Subjects |
Dates of Accrual |
Type of Exposure |
Type of Dosimetry |
Summary of Results |
||
Cases |
Controls |
Cases |
Controls |
|||||
Sever and others (1988) |
Congenital malformations, identified from 3 hospitals in two counties near Hanford |
Selected from hospital delivery room records, next live birth, matched by sex, mother’s age (5 years), race |
672 |
977 |
1957–1980 |
External whole-body radiation |
Recorded doses obtained from Hanford records; estimates in millisieverts |
Overall, no association with employment at Hanford; suggestion of increase with parental preconception dose; some increases evident in subgroups |
Gardner and others (1990b) |
Leukemia and lymphoma in people under 25 born in West Cumbria |
From birth register, matched by date of birth and sex: local group, matched by parish; area group, unmatched |
Leukemia (52); NHL (22); Hodgkin’s disease (23) |
1001 |
Diagnosis: 1950–1985 |
Total and 6-month external whole-body preconception exposure; antenatal X-ray |
Doses from worker radiation records (British Nuclear Fuels) |
Leukemia and NHL higher in children born near Sellafield, and with fathers employed at the plant especially those with high preconception doses |
Kinlen (1993b) |
Leukemia and lymphoma in people born in Scotland since 1958, diagnosed under age 25 |
Randomly selected from births, matched by county and sex |
1024 leukemia; 237 NHL |
3783 |
1958–1990 |
Total, 3-month, and 6-month preconception external whole-body dose |
Doses from worker records (Scottish nuclear industry) |
No significant excess in any subgroup; no association with preconception radiation dose |
McLaughlin and others (1993a) |
Children in Ontario, 0–14, died from leukemia 1950–1963 or diagnosed 1964–1988, born to mothers living near nuclear facility |
Selected from births, matched to case by date of births (3 months) and region of mother’s residence at child’s birth |
112 |
890 |
Deaths: 1950–1963; diagnosis: 1964–1988 |
Whole-body external dose, whole-body external tritium dose, radon dose |
Recorded doses from National Dose Registry |
No increased risk for any exposure period or exposure type |
Roman and others (1993) |
Leukemia or NHL, diagnosed ages 0–4, born in West Berkshire, Basingstoke, and North Hampshire |
Two controls per case selected from birth registers; four per case from delivery registers in study area; matched by sex, date of birth (6 months), area of residence at birth, time of diagnosis |
54 |
324 |
1972–1989 |
Exposure to radiation at work |
Record of employment in nuclear industry; recorded film badge dose if monitored |
Cases were more likely to have a parent employed in the nuclear industry; fathers of cases were more likely to be monitored for radiation; no dose-response evident for fathers monitored |
Sorahan and Roberts (1993) |
Children dying of cancer under age 16 in England, Wales, and Scotland |
Selected from birth register, matched by local authority, sex, date of birth |
15,279 |
15,279 |
1953–1981 |
6-month preconception; external whole-body dose; exposure to radionuclides (unsealed sources) |
Expert assignment, based on job titles |
No association with external exposure; increased risk with potential exposure to radionuclides |
Draper and others (1997) |
Childhood cancer in Great Britain and Scotland |
Selected from birth register for same area, born within 6 months of case, same sex |
35,949 |
38,323 |
Great Britain: 1952–1986; Scotland: 1987–1990 |
Total, 3-month and 6-month preconception external whole-body dose |
Doses recorded by National Registry for Radiation Workers |
Fathers of cases more likely to be radiation workers; no dose-response for any exposure periods for fathers or mothers |
TABLE 9-5C Children of Adults Exposed to Radiation—Cohort Studies
Reference |
Incidence/Mortality |
Cohort Definition |
Comparison Group |
Dates of Accrual |
Type of Exposure |
Type of Dosimetry |
Outcomes Studied |
Number of Cases |
Summary of Results |
Gardner and others (1987) |
Mortality |
Children attending school in Seascale up to 11/84, born since 1950 |
U.K. national rates |
Beginning school–6/30/86 |
Presumed exposures from Sellafield |
Attending school in community near Sellafield plant |
Major categories of causes of death |
Total deaths: 10 |
No increase in relation to national rates for all cancer, all causes, leukemia, or lymphoma |
Gardner and others (1987) |
Mortality |
Children born to mothers resident in Seascale from 1950 to 1983 |
U.K. national rates |
1950–6/30/86 |
Presumed exposures from Sellafield |
Born in community near Sellafield plant |
Major categories of causes of death |
Total deaths 27; leukemia 5 |
Approximately tenfold excess of leukemia deaths vs. national rates; 2.5-fold excess for other cancer; no increase for other causes |
Black and others (1992) |
Incidence |
Children born in Dounreay area 1969–1988; children attending local schools in the same period born elsewhere |
Scottish national rates by tumor site, sex, age, and calendar year |
1969–1988 |
Presumed exposures from Dounreay nuclear reprocessing plant |
Born in or living in Dounreay area of Caithness, Scotland |
Leukemia and NHL, Hodgkin’s disease, other cancers |
Total cancer cases in birth cohort 5; total cases in school cohort 3 |
Increased incidence of leukemia in both birth and school cohorts: birth cohort O/E −2.3 (0.7, 5.4); schools cohort O/E −6.7 (1.4, 19.5) |
Dickinson and others (1996) |
Incidence |
260,060 singleton births to mothers resident in Cumbria, U.K. |
Children of parents who worked at Sellafield anytime between 1947 and 1989 |
1950–1989 |
External dose from ionizing radiation to fathers prior to conception of the child |
Recorded radiation dose obtained from Sellafield facility |
Sex ratio |
Live births to fathers with dose prior to conception: 10,272 |
Significantly higher sex ratio (1.09; CI 1.06, 1.13) for children of fathers exposed at Sellafield than other Cumbria children. Increased sex ratio (1.4; CI 1.13–1.73) for children of fathers with >10 mSv in 90 d prior to conception |
Dummer and others (1998) |
Mortality |
256,066 live and 4034 stillbirths to mothers resident in Cumbria, U.K. |
Observed and expected stillbirth rates by distance (in circles of 5, 10, 15, 20, and 25 km) and direction. Expected estimated from rates in remainder of Cumbria |
1950–1989 |
Presumed exposures from Sellafield |
Proximity to and direction from Sellafield of mother’s residence |
Stillbirths |
Live births to mothers within 25 km of Sellafield: 54,746; stillbirths 888 |
No evidence that proximity to Sellafield increased risk of stillbirth. No significant increase in stillbirths with distance within any of six directional sectors |
Parker and others (1999) |
Mortality |
248,097 live and 3715 stillbirths to mothers resident in Cumbria, U.K. |
Children of father who worked at Sellafield anytime between 1947 and 1989 |
1950–1989 |
External and internal dose from ionizing radiation to fathers prior to conception of the child |
Recorded radiation dose obtained from Sellafield facility |
Stillbirths |
Live births in fathers exposed prior to conception: 9078; stillbirths 130 |
Significant increase in stillbirth with father’s external radiation dose prior to conception: OR per 100 mSV 1.24 (CI 1.04, 1.45). Risk higher for stillbirths with congenital anomaly and highest for neural tube defects |
Reference |
Incidence/Mortality |
Cohort Definition |
Comparison Group |
Dates of Accrual |
Type of Exposure |
Type of Dosimetry |
Outcomes Studied |
Number of Cases |
Summary of Results |
Roman and others (1999) |
Incidence |
Children under age 25 of male employees of three nuclear authorities in Great Britain |
External: national rates from England and Wales. Internal: within the cohort by radiation exposure levels |
For external analyses: born 1965 or later; for internal, born 1985 or later |
External whole-body preconception dose |
Employment in nuclear industry; whether monitored for radiation exposure; dose estimates from records |
All cancer, leukemia, and NHL |
Total cancer 111 leukemia 28 |
No excess incidence over expected; leukemia in children whose fathers received >100 mSv preconception dose was 5.8 times that in children conceived prior to father’s employment, based on 3 cases; no evidence of any dose-response for leukemia |
Doyle and others (2000) |
Incidence and mortality |
Employees of AWE, AEA and BNF, and for AEA and BNF past employees <75 years old who were included in the pension database |
Within the cohort by radiation exposure level |
1993–1996 |
External and internal dose from ionizing radiation to fathers prior to conception of the child |
Whether monitored for radiation exposure; if so, dose estimates from records of the nuclear facility |
Fetal deaths and congenital malformations |
Live births: women 3048; men 20,899 Fetal deaths: women 526; men 2723 |
Risk of fetal death and congenital malformations not related to whether father was monitored for radiation prior to conception or to the dose of radiation received. Risk of early miscarriage (<13 weeks) was higher if mother was monitored before conception (OR 1.3; CI 1.0, 1.6), but no trend with radiation dose. Risk of stillbirth was also higher (OR 2.2; CI 1.0, 4.6). Risk of any major malformation not associated with maternal monitoring or dose prior to conception |
period, age and sex of child, and the number of children born to each parent) were significantly greater than 1.0 among offspring of fathers who received cumulative external doses of 100 mSv or 10 mSv in the 6 months prior to conception (4.1, 95% CI 1.4, 11.8, 5.1, 95% CI 1.6, 16.9), respectively. It should be noted that these results were based on very few cases (four and three, respectively). No trend of increasing risk with cumulative dose was apparent. None of the three studies provide quantitative estimates of risk based on dose-response analyses, and the results across studies are not consistent. Thus, there is little evidence from epidemiologic studies of a link between parental preconception exposure to ionizing radiation and childhood leukemia or other cancers.
Other possible indices of the occurrence of transmissible genetic damage from preconception exposures include spontaneous abortions, congenital malformations, neonatal mortality, stillbirths, and the sex ratio of offspring. Relatively few epidemiologic studies have been conducted to evaluate these outcomes in relation to preconception radiation exposure. Dickinson and colleagues (1996) examined the sex ratio among children born to fathers employed at Sellafield. Exposure was assessed using two methods: total cumulative radiation dose prior to conception and dose received in the 90 days prior to conception. Total cumulative dose did not account for a significant amount of variation in the sex ratio during the period 1950–1988. No significant trend was observed between sex ratio and exposure 90 d prior to conception, although the sex ratio was increased in children of fathers in the highest-dose category (>10 mSv). Chance could not be ruled out as the reason for this result.
A companion study investigated stillbirths in the offspring of men employed at Sellafield (Parker and others 1999). Individual film badge doses were available by record linkage with the British Nuclear Fuels (BNF) dosimetry database. Significant positive associations between both the total cumulative dose (OR per 100 mSv = 1.24; 95% CI 1.04, 1.45) and the dose during the 90 d prior to conception (OR per 100 mSv = 1.86; 95% CI 1.21, 2.76) and risk of stillbirth were observed.3 A nested case-control study was conducted among radiation workers alone using live births matched on sex and date of birth. In contrast with the cohort analysis, the adjusted OR for exposure 90 d preconception was not significantly different from 1.00 (OR per 100 mSv = 1.08; 95% CI 0.68, 1.74). The total cumulative dose, however, did show a significant association with the occurrence of stillbirth (OR per 100 mSv = 1.24; 95% CI 1.04, 1.45). Although based on only a few exposed individuals, neither analysis indicated the presence of an association with internal exposure to radionuclides. Limitations of the study noted by the authors included the possibility of the existence of residual confounding by year of birth, a time-varying uncertainty (30%) in the recorded film badge doses, and the absence of information on concurrent exposures to organic chemicals in the workplace. An earlier study of stillbirth rates around Sellafield (Dummer and others 1998) found no increase in stillbirths in the resident population within 25 km of the facility.
The Nuclear Industry Family Study in the United Kingdom has also investigated possible links between occupational radiation exposures and reproductive health (Maconochie and others 1999). This study population includes all current employees of the Atomic Energy Authority, Atomic Weapons Establishment, and BNF, as well as past employees who were under age 75 and on record at the pension administration office. Information on reproductive health and health of children was obtained through a mailed questionnaire and linked with data from the employers on occupational exposure to ionizing radiation. The database consists of 53,672 pregnancies, 39,557 reported by men and 8,883 by women. Results of the analysis of fetal deaths and congenital malformations were reported by Doyle and colleagues (2000). The risk of neither fetal death nor major congenital malformation was related to paternal preconception radiation dose. Although early miscarriage was more common among mothers who had been monitored prior to conception (OR 1.3; 95% CI 1.0, 1.6), there was no evidence of a dose-response. Risk of fetal death was higher among mothers who had been monitored prior to conception (OR 2.2; 95% CI 1.0–4.6). ORs were adjusted for parental age, birth order, previous fetal loss, calendar year of the end of pregnancy, and manual versus nonmanual job status. No dose response was evident.
In summary, there have been a number of studies of children of adults exposed to radiation. Ecologic studies are based on very small numbers, and none provide quantitative information from dose-response analyses or quantitative estimates of the risk of disease associated with exposure. There is little conclusive evidence from epidemiologic studies of a link between parental preconception exposure to radiation and childhood leukemia or other cancers. Few studies have been conducted to evaluate other possible indices of the occurrence of transmissible genetic damage from preconception radiation exposures, such as spontaneous abortions, congenital malformations, neonatal mortality, stillbirths, and the sex ratio of offspring. Some but not all studies have found a significant positive association between total cumulative dose, as well as dose during the 90 d prior to conception, and the risk of stillbirth. The risk of neither fetal death nor major congenital malformation has been related to paternal preconception radiation dose.
EXPOSURE TO RADIOACTIVE IODINE 131
In evaluating the evidence regarding the risk of cancer associated with exposure to environmental sources of radia-
tion, internal exposure to 131I is of particular concern regarding the risk of thyroid cancer. In contrast to the considerable amount of information that is available from numerous studies of external radiation exposure, there is relatively little information regarding the risk of thyroid cancer in humans exposed to 131I. Existing evidence comes from studies of 131I administered for therapeutic or diagnostic purposes and from various environmental exposure settings, most notably from recent studies of persons exposed to radiation from the Chernobyl accident (reviewed above).
Studies of therapeutic and diagnostic 131I exposures are described in detail in Chapter 7. In brief, early studies of persons receiving therapeutic 131I for hyperthyroidism found no convincing evidence that the risk of thyroid cancer was increased (Dobyns and others 1974; Safa and others 1975; Holm and others 1980a; Holm 1984); most of the participants were adults at the time of exposure, were followed for very short periods, had existing thyroid disease at the time of treatment, and were treated with radiation doses that were quite high (generally 20,000–100,000 mGy). Results from a follow-up (Ron and others 1998a) of one of these studies (Dobyns and others 1974) suggest an increased risk of death from thyroid cancer in patients previously treated with 131I, but the numbers of excess deaths were small and it is likely that underlying thyroid disease might have contributed to these results. Similar results were obtained from a study of 7400 patients who were treated with radioiodine from 1950 to 1991 in England (Franklyn and others 1999). Studies have also evaluated persons exposed to much lower doses (generally 500–1000 mGy) through diagnostic procedures (Holm and others 1980a, 1980b; Hall and others 1996). Although there is some evidence of a small increase in thyroid cancer associated with such exposures, there is a lack of consistency and the small increases in thyroid cancer in some studies are likely due to the underlying thyroid condition. As for the therapeutic studies described above, these too are primarily of persons exposed as adults. The thyroid gland is more radiosensitive in children than adults, most likely because of more rapid growth in infants and children (Williams 2003) and because of differences in metabolism (Mettler and others 1996).
Only a few studies have evaluated the effects of environmental exposure to radioactive iodine. In contrast to the medical exposures summarized above, which were due exclusively to 131I, environmental exposures have generally contained mixtures of 131I, external radiation, and short-lived radioiodines. Initial studies of thyroid disease incidence in Utah schoolchildren exposed to fallout from atmospheric nuclear weapons testing at the Nevada Test Site appeared to show no difference in thyroid disease outcomes compared to children from unexposed areas (Rallison and others 1975). However, a follow-up study reported a slight excess risk of thyroid neoplasms associated with radioiodine exposure (Kerber and others 1993). Although positive dose-response trends were also noted for total nodules and thyroid cancer (when analyzed separately), they were not statistically significant. The study was limited by small numbers of exposed individuals and a low incidence of thyroid neoplasms and by the fact that the examiners were not blinded to exposure. In contrast, a follow-up study of 3440 persons exposed as young children to atmospheric releases of primarily 131I from the Hanford Site found no increased risk of thyroid cancer associated with individual radiation dose to the thyroid (Davis and others 2001, 2004a).
The explanation for the apparent difference in results in the Utah study and the Hanford study is not clear. One possibility is that the exposures were substantially different in terms of the mix of radionuclides and the dose rate. Thyroid dose at Hanford was due almost entirely to 131I, whereas in Utah there was greater contribution from other radioiodines as well as external sources. Exposures in Utah were also more concentrated and episodic than at Hanford, corresponding to specific nuclear tests. This likely resulted in doses being delivered at substantially higher dose rates (although the total dose among 3545 study participants for whom thyroid doses could be estimated [mean 98 mGy] was similar to Hanford doses). A second possibility is that the Utah study’s estimated dose-response could have been biased in the direction of finding an association because the collection of dietary consumption data took place after thyroid disease classification was known for each participant.
Extensive evaluation of the population of the Marshall Islands has shown an increase in benign and malignant thyroid nodules in residents of the northern atolls of Rongelap and Utirik (Conard 1980, 1984). In addition, a retrospective cohort study of more than 7000 Marshall Islanders showed that the prevalence of palpable thyroid nodularity ( 1.0 cm) decreased linearly with increased distance from the Bikini test site (Hamilton and others, 1987). More recently, there has been extensive investigation of populations exposed to radioactive fallout (including 131I as a substantial component) after the Chernobyl accident. Findings from these studies are reviewed and summarized above.
In summary, studies of exposure to 131I from therapeutic and diagnostic uses provide some evidence of a small increase in thyroid cancer associated with such exposures, but there is lack of consistency in the findings. Furthermore, the small increases in thyroid cancer observed in some studies are likely due to the underlying thyroid condition, not to radiation exposure. Results from environmental exposures have been inconsistent. Findings of an increase in thyroid neoplasia in persons exposed to fallout in the Marshall Islands are limited by the lack of individual dosimetry. No excess risk of thyroid cancer was found in residents exposed to radiation from Hanford, and the slight excess risk of thyroid neoplasms associated with radioiodine exposure of Utah residents from the Nevada Test Site was based on small numbers.
In contrast, substantial increases in thyroid cancer have been reported in areas contaminated with radioactive fallout
from Chernobyl, primarily among children. Although much of the thyroid dose from Chernobyl is due to 131I, exposure to a mix of other radionuclides and the lack of individual dose estimates in most of the studies to date have made it difficult to develop quantitative risk estimates for radiation dose from 131I. However, there is now emerging evidence indicating that exposure to radiation from Chernobyl is associated with an increased risk of thyroid cancer and that the relationship is dose dependent. These findings are based on individual estimates of thyroid radiation dose and reveal strong and statistically significant dose-related increased risks that are consistent across studies. Thus, although the precise quantitative relationship between radiation dose from 131I and the development of thyroid neoplasia remains uncertain at this time, recent findings from studies around Chernobyl and Hanford provide important quantitative estimates of risk as a function of dose.
DISCUSSION
A considerable number of papers have been published from studies that have attempted to determine whether persons exposed, or potentially exposed, to ionizing radiation from environmental sources are at an increased risk of developing cancer. The existing published literature consists primarily of reports that are descriptive in nature and ecologic in design. Such studies are limited in their usefulness in defining risk of disease in relation to radiation exposure or dose. They can sometimes be informative in generating new hypotheses or suggesting directions for study, but seldom, if ever, are they of value in testing specific hypotheses or providing quantitative estimates of risk in relation to specific sources of environmental radiation. Fewer attempts have been made to evaluate the effect of environmental radiation exposures using the two most common analytical study designs employed in epidemiology: the case-control study and the cohort study. Such studies are almost always based on individual-level data and thus are not subject to many of the limitations inherent in ecologic studies. They can potentially provide quantitative estimates of risk based on individual radiation dose.
Epidemiologic studies, in general, have limited ability to define the shape of the radiation dose-response curve and to provide quantitative estimates of risk in relation to radiation dose, especially for relatively low doses. To be informative in this regard a study should (1) be based on accurate, individual dose estimates, preferably to the organ of interest; (2) contain substantial numbers of people in the dose range of interest; (3) have long enough follow-up to include adequate numbers of cases of the disease under study; and (4) have complete and unbiased follow-up. Unfortunately, the published literature on environmental radiation exposures is not characterized by studies with such features.
Sixteen ecologic studies of populations living around nuclear facilities are summarized, thirteen of the locations being outside the United States. Most define exposure, or potential for exposure, based on a measure of distance from the facility, and the focus of most of these investigations is leukemia and/or childhood cancer, although a few include all cancers as an outcome. Notably, most of the studies do not specify the nature of the radiation exposure, and none of the 16 contain individual estimates of radiation dose. Although some of these studies report an increased occurrence of cancer that could be related potentially to environmental radiation exposures, none provide a direct quantitative estimate of risk in relation to radiation dose. There have been three case-control studies of persons living around a nuclear facility. One focuses on congenital and perinatal conditions, stillbirths, and infant deaths in relation to exposure from uranium mines. This study does not provide an estimate of radiation risk associated with any of the indicators of exposure. The other two are of leukemia in children and young adults. Neither study found an increased risk associated with parental radiation exposure and X-ray exposure of the child, but both did find an increased risk associated with playing on beaches near the nuclear facility.
Several cohort studies have been reported of persons exposed to environmental radiation under various circumstances: participation in atmospheric nuclear weapons tests conducted by the United Kingdom and the United States; residents and their offspring living near the Techa River in the southern Urals of the Russian Federation and exposed from the nearby Mayak nuclear complex; residents living near the Hanford Site in eastern Washington State; and residents of the Marshall Islands. Overall, studies of persons who participated in U.K. atmospheric nuclear weapons tests found no increased risk of developing cancer or other fatal diseases as a function of estimated dose received, but there was some evidence of an increase in non-CLL leukemia. In contrast, a recent study of U.S. veterans who participated in atmospheric nuclear weapons tests reported a significant increase in death from all causes and for all lymphopoietic cancers combined.
Results from studies of residents living near the Techa River have found no evidence of a decrease in birth rate or fertility in the exposed population and no increased incidence of spontaneous abortions or stillbirths. There is some evidence of a statistically significant increase in total cancer mortality. Estimates of the relative risk for cancer of the esophagus, stomach, and lung are similar to those reported for atomic bomb survivors. There is no evidence of an increase in cancer mortality in the offspring of exposed residents. The one study of persons living in the town of Ozyorsk exposed to fallout from the nearby Mayak nuclear facility reported an excess of thyroid cancer three to four times that expected relative to rates for all of Russia and a somewhat lower excess (1.5 to twofold higher) based on a comparison with Chelyabinsk Oblast rates.
A follow-up study of persons exposed as young children to atmospheric releases primarily of 131I from the Hanford
Site in eastern Washington State found no increased risk of thyroid cancer associated with individual radiation dose to the thyroid. A prevalence study of thyroid cancer conducted through screening of 3709 Marshall Island residents born before the Castle BRAVO atmospheric nuclear weapons test on March 1, 1954, found some indication that the prevalence of thyroid cancer increased with quartile of estimated dose, but the increase was not statistically significant.
Numerous epidemiologic studies have been carried out since the Chernobyl accident to investigate the potential late health consequences of exposure to ionizing radiation from the accident. These studies have focused largely on thyroid cancer in children, but have also included investigations of recovery operation workers and residents of contaminated areas, and have investigated the occurrence of leukemia and solid tumors other than thyroid cancer among exposed individuals. Overwhelmingly, the published findings are from studies that are ecologic in design and therefore do not provide quantitative estimates of disease risk based on individual exposure circumstances or individual estimates of radiation dose. Most reports are descriptive incidence and prevalence studies that utilize population or aggregate estimates of radiation dose. Only three analytical studies are published that report dose-response results based on individual dose estimates.
Numerous reports have continued to describe an increasing number of cases of thyroid cancer, particularly in the most heavily contaminated regions of Ukraine and Belarus, as well as in Russia. Collectively, findings reported to date have demonstrated an association between an increase in thyroid cancer incidence and radiation exposure from the Chernobyl accident. This increase cannot be explained only by the aging of the cohort and the improvement of case detection and reporting. Although there is now little doubt that an excess of thyroid cancer has occurred in highly contaminated areas, there is still very little information regarding the quantitative relationship between radiation dose to the thyroid from Chernobyl and the risk of thyroid cancer. Results from three analytical studies published indicate that exposure to radiation from Chernobyl is associated with an increased risk of thyroid cancer and that the relationship is dose dependent. The findings from these studies are consistent with descriptive reports from contaminated areas of Ukraine and Belarus, and the quantitative estimate of thyroid cancer risk is generally consistent with estimates from other radiation-exposed populations. Available data on exposure from the Chernobyl accident are largely in agreement with observations from other studies showing that exposure at the youngest ages is associated with the greatest risk of thyroid cancer. At present no data are available from Chernobyl regarding the risk of thyroid cancer from in utero exposure. Fifteen years after the Chernobyl accident, thyroid cancer incidence is still highly elevated. An increase in thyroid cancer has been observed in both males and females, and most of the Chernobyl studies have reported similar relative risks per unit dose for males and females. Iodine deficiency also appears to be an important modifier of the risk of radiation-induced thyroid cancer, and there is some evidence that iodine deficiency enhances the risk of thyroid cancer following radiation exposure. Finally, relatively little has been published regarding thyroid outcomes other than thyroid cancer, although one study has reported an elevated risk of benign thyroid tumors and there have been reports of increases in autoimmune disease and antithyroid antibodies following childhood exposure to Chernobyl.
Evidence from epidemiologic studies regarding the risk of leukemia in the general population reflects low-dose-rate exposure (primarily from 137Cs), which has occurred for a number of years and will continue to occur in the future. These resident populations were exposed at all ages, but studies of residents are primarily of persons exposed as children and/or in utero.
At present, the available evidence from ecologic studies does not convincingly indicate an increased risk of leukemia among persons exposed in utero to radiation from Chernobyl. There are no data from analytic epidemiologic studies in which individual dose estimates are available. The existing evidence does not support the conclusion that the rates of childhood leukemia have increased as a result of radiation exposure from the Chernobyl accident. However, ecologic studies of the types conducted to date are not particularly sensitive to detecting relatively small changes in the incidence of a disease as uncommon as childhood leukemia over time or by different geographic areas. The single analytical study is insufficient to draw conclusions regarding leukemia risk after exposure of children to Chernobyl. There is also no convincing evidence that the incidence of leukemia has increased in adult residents of the exposed populations that have been studied in Russia and Ukraine. However, few studies of the general adult population have been conducted, and they have employed ecologic designs that are relatively insensitive.
There has been relatively little study of the incidence or mortality from solid cancers other than thyroid cancer in populations exposed to radiation from the Chernobyl accident. Two studies have investigated solid cancer incidence in liquidation workers. They reported increases of cancer incidence during the periods, but generally the excesses were relatively small and not statistically significant. No descriptive or analytical epidemiologic studies of breast cancer risk in populations exposed to radiation from Chernobyl have been published in the peer-reviewed literature; however, one monograph has cited elevated breast cancer incidence rates based on Ukrainian registries. Similarly, although no descriptive or analytical epidemiologic studies of bladder or kidney cancer risk in relation to Chernobyl have been published in the peer-reviewed literature, there has been a series of papers investigating aspects of possible radiation carcinogenesis in these organs.
Four ecologic studies of populations exposed to natural background radiation have been reported. Two were con-
ducted in China, one in Great Britain, and one in India. These studies did not find any association between disease rates and indicators of high background levels of radiation, and they do not provide any quantitative estimates of disease risk.
Three ecologic studies of children of adults exposed to radiation have been published, with a focus on preconception parental exposure and the risk of leukemia and lymphoma in the offspring of exposed parents. All three studies were conducted in relation to exposures received by parents working at the Sellafield nuclear facility in Great Britain. Although there is some evidence of an increased risk associated with measures of individual dose, the findings are based on very small numbers of cases and the results across studies are not consistent. A larger number of case-control studies have been conducted to investigate the possible relationship between radiation exposure of adults and subsequent cancer in their offspring. In summary, none of the studies provide quantitative information from dose-response analyses or quantitative estimates of the risk of disease associated with exposure, and results across studies are inconsistent. There have been three cohort studies published regarding the risk of cancer in children of adults exposed to radiation. None of the three provide quantitative estimates of risk based on dose-response analyses, and the results across studies are not consistent. Thus, there is little conclusive evidence from epidemiologic studies of a link between parental preconception exposure to ionizing radiation and childhood leukemia or other cancers.
Other possible indices of the occurrence of transmissible genetic damage from preconception exposures include spontaneous abortions, congenital malformations, neonatal mortality, stillbirths, and the sex ratio of offspring. Relatively few epidemiologic studies have been conducted to evaluate these outcomes in relation to preconception radiation exposure, and there is no consistent evidence of an association of any such outcomes with exposure to environmental sources of radiation.
Studies of exposure to 131I from therapeutic and diagnostic uses provide some evidence of a small increase in thyroid cancer, but the small increase observed is likely due to the underlying thyroid condition, not to radiation exposure. Findings of an increase in thyroid neoplasia in persons exposed to fallout in the Marshall Islands are limited by the lack of individual dosimetry. No excess risk of thyroid cancer was found in residents exposed to radiation from Hanford, and only a slight excess risk of thyroid neoplasms was found associated with radioiodine exposure of Utah residents from the Nevada Test Site. In contrast, substantial increases in thyroid cancer have been reported in areas contaminated with radioactive fallout from Chernobyl, primarily among children. Recent evidence from three population-based case-control studies indicates that exposure to radiation from Chernobyl is associated with an increased risk of thyroid cancer and that the relationship is dose dependent. These findings are based on individual estimates of thyroid radiation dose and reveal strong and statistically significant dose-related increased risks that are consistent across studies. They provide important quantitative estimates of risk as a function of dose, primarily from 131I.
SUMMARY
This chapter reviews the evidence from peer-reviewed articles published since BEIR V (NRC 1990) of the relationship between exposure to ionizing radiation from environmental sources and human health.
Ecologic studies of populations living around nuclear facilities neither contain individual estimates of radiation dose nor provide a direct quantitative estimate of risk in relation to radiation dose. Similarly, the one case-control study of congenital and perinatal conditions, stillbirths, and infant deaths in relation to exposures from uranium mines does not provide an estimate of the risk associated with any of the indicators of exposure, and two ecologic studies of populations exposed to fallout from atmospheric nuclear testing or other sources of environmental release of radiation provide no quantitative estimates of the risk associated with presumed exposure.
Several cohort studies have been reported of persons exposed to environmental radiation under various circumstances. No increased risk of developing cancer or other fatal diseases was found in persons who participated in U.K. atmospheric nuclear weapons tests, but there was some evidence of an increase in non-CLL leukemia. U.S. veterans who participated in atmospheric nuclear weapons tests reported a significant increase of death from all causes and for all lymphopoietic cancers combined. There is no evidence of a decrease in birth rate or fertility or an increased incidence of spontaneous abortions or stillbirths in residents living near the Techa River in the Russian Federation. There is some evidence of a statistically significant increase in total cancer mortality, but no evidence of an increase in cancer mortality in the offspring of exposed residents. Persons living in the town of Ozyorsk (Russia) exposed to fallout from the nearby Mayak nuclear facility reported an excess of thyroid cancer (1.5–4 times higher than expected). No increased risk of thyroid cancer was found associated with individual radiation dose to the thyroid in persons exposed as young children to atmospheric releases primarily of 131I from the Hanford Site in eastern Washington State. There is some indication that the prevalence of thyroid cancer among Marshall Island residents born before the Castle BRAVO atmospheric nuclear weapons test increased with quartile of estimated dose, but the increase was not statistically significant.
There continues to be an increasing number of cases of thyroid cancer in populations exposed to radiation from the Chernobyl accident that cannot be explained only by the aging of the cohort and the improvement in case detection and reporting. Results from three analytical studies indicate that exposure to radiation from Chernobyl is strongly associated
with an increased risk of thyroid cancer in a dose-dependent manner, and the quantitative estimate of thyroid cancer risk generally is consistent with estimates from other radiation-exposed populations and is observed in both males and females. At present, no data are available from Chernobyl regarding the risk of thyroid cancer from in utero exposure. Iodine deficiency appears to be an important modifier of risk, enhancing the risk of thyroid cancer following radiation exposure from Chernobyl. Relatively little has been published regarding thyroid outcomes other than thyroid cancer, although one study has reported an elevated risk of benign thyroid tumors and there have been reports of increases in autoimmune disease and antithyroid antibodies following childhood exposure to Chernobyl.
Evidence from ecologic studies does not indicate an increased risk of leukemia among persons exposed in utero to radiation from Chernobyl nor that rates of childhood leukemia have increased. A single analytical study is insufficient to draw conclusions regarding leukemia risk after exposure of children to Chernobyl. There is no convincing evidence that the incidence of leukemia has increased in adult residents of the exposed populations that have been studied in Russia and Ukraine. There has been very little study of the incidence or mortality from solid cancers other than thyroid cancer in populations exposed to radiation from the Chernobyl accident, and there is no evidence of significant excesses of any other solid cancer type.
Four ecologic studies of populations exposed from natural background radiation did not find any association between disease rates and indicators of high background levels of radiation exposure (for a general discussion of the limitations of ecologic studies see the introduction to this chapter and, more specifically in reference to studies of populations exposed from natural background radiation, see Appendix D, “Hormesis and Epidemiology”).
Ecologic studies of children of adults exposed to radiation while working at the Sellafield nuclear facility in Great Britain have suggested some increased risk of leukemia and lymphoma associated with individual dose, but the findings are based on very small numbers of cases and the results across studies are not consistent. A larger number of case-control studies provides no quantitative estimates of the risk of disease in offspring of exposed parents, and results across studies are inconsistent. None of three published cohort studies provide quantitative estimates of risk based on dose-response analyses, and the results across studies are not consistent. Relatively few epidemiologic studies have been conducted to evaluate outcomes such as spontaneous abortions, congenital malformations, neonatal mortality, stillbirths, and the sex ratio in relation to preconception radiation exposure, and there is no consistent evidence of an association of any such outcomes with exposure to environmental sources of radiation.
In contrast to the considerable amount of information that is available from numerous studies of external radiation exposure, there is relatively little information regarding the risk of thyroid cancer in humans exposed internally to 131I. There is some evidence of a small increase in thyroid cancer associated with exposure to 131I from therapeutic and diagnostic uses, but the findings are inconsistent and the small increases in thyroid cancer observed in some studies are likely due to the underlying thyroid condition, not to radiation exposure. Results from environmental exposures have also been inconsistent. An increase in thyroid neoplasia has been observed in persons exposed to fallout in the Marshall Islands, but no excess risk of thyroid cancer was found in residents exposed to radiation from Hanford, and the slight excess risk of thyroid neoplasms associated with radioiodine exposure in Utah residents from the Nevada Test Site was based on very small numbers. In contrast, substantial increases in thyroid cancer have been reported in areas contaminated with radioactive fallout from Chernobyl, primarily among children. Recent evidence indicates that exposure to radiation from Chernobyl is associated with an increased risk of thyroid cancer and that the relationship is dose dependent. These findings are based on individual estimates of thyroid radiation dose and reveal strong and statistically significant dose-related increased risks that are consistent across studies.