Click for next page ( 2


The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 1
Summary More than 200,000 U.S. military personnel participated in atmospheric nu- clear weapons tests between 1945 and the 1963 Limited Nuclear Test Ban Treaty. Questions persist, such as whether that test participation is associated with the timing and causes of death among those individuals. This is the report of a mortality study of the approximately 70,000 soldiers, sailors, and airmen who participated in at least one of five selected U.S. nuclear weapons test series in the 1950s and nearly 65,000 comparable nonparticipants, the referents. The investigation described in this report, based on more than 5 trillion person-years of mortality follow-up, represents one of the largest cohort studies of military veterans ever conducted. We found that, during the follow-up period, overall, participants and referents had similar risks of death; . participants and referents had similar risks of death from cancer; and . specifically, participants had an apparent 14 percent higher risk of leuke- mia death than the referents, although that difference was not statistically sig- nificant and could be a chance finding. Descriptive analyses not specified at the outset of this study showed . statistically significant increased risk of leukemia death among partici- pants at land test series (tests conducted at the Nevada Test Site) compared to land series referents; however, sea series participants (tests conducted at the Pa ~Series selected were Operations GREENHOUSE (1951), UPSHOT-KNOTHOLE (1953), CASTLE (1954), REDWING (1956), and PLUMBBOB (19573. These five series were chosen for an earlier study of atomic veterans to represent tests at both the Nevada Test Site and the Pacific Proving Ground. Fuller discussion is in Chapter 1. 1

OCR for page 1
2 THE FIVE SERIES STUDY cific Proving Ground) have an observed and not significant lower risk than sea series referents; . statistically significant increased risk of all-cause mortality among par- ticipants at sea series (tests conducted at the Pacific Proving Ground) compared to sea series referents; and . statistically significant increased risks among participants of death from external causes (such as motor vehicle accidents), nasal cancer, and prostate cancer. The leukemia findings do not resolve the debate over whether either partici- pation in general or the radiation doses in particular is associated with leukemia mortality. The set of leukemia findings is consistent with the results of other studies of military participants in nuclear tests and is consistent with a hypothe- sis that these are radiation effects. The other findings listed are more likely to be chance occurrences. We discuss the evidence in greater detail in the report. METHODS The participant cohort predominantly white and male was identified from the database maintained by the Nuclear Test Personnel Review Program (NTPR) at the Defense Threat Reduction Agency.2 This study supersedes an earlier National Research Council report (Robinette et al., 1985) that was based on a different NTPR-provided dataset that was subsequently identified as inac- curate.3 The Medical Follow-up Agency (MFUA) and NTPR staff have placed substantial effort into validating the current participation list. We verified that the individuals it includes were indeed participants in these test series and we estimated that the list might have missed approximately one percent of actual participants. We compared the participant cohort's mortality experience with that of a referent cohort of military personnel comparable to the participants with respect to branch of service, time of active military duty, type and general location of assigned unit, age, and paygrade. Department of Veterans Affairs (VA) records and databases provided fact of death for members of both cohorts. For each identified death, cause of death information was requested from VA regional offices and federal archives records centers, where VA claims folders, which hold death certificates, are filed. In those cases for which cause of death was not available through that route, we searched the database of the National Death Index for death certificate-derived cause of death information. 2The organizational locus of the Nuclear Test Personnel Review Program within the Department of Defense was the Defense Nuclear Agency, renamed the Defense Special Weapons Agency in June 1996, and reorganized in October 1998 as the Defense Threat Reduction Agency. 3Chapter 1 of the full report describes the history in greater detail.

OCR for page 1
SUMMARY 3 Using two analytic techniques proportional hazards models and standard- ized mortality ratios we tested for differences between the participant and ref- erent cohorts in all-cause, all-cancer, and leukemia mortality. Analyses based on the proportional hazards model involve direct comparisons of the participant and referent cohorts, whereas standardized mortality ratios involve comparison of each group, separately, with external population rates. Further explorations in- cluded other outcomes (e.g., all major categories of deaths, and specific group- ings of cancers) and possible differences in effect for participants of test series conducted at the Pacific Proving Ground (sea series) and participants at the Ne- vada Test Site (land series). The initial plan for this epidemiologic study included the use of individual- level radiation dose data, compiled and estimated by NTPR, to test for dose- response relationships indicative of radiation-caused adverse health effects. However, the Institute of Medicine committee overseeing the conduct of this study reviewed the dosimetry program and found that the dose data were not appropriate for epidemiologic analysis. Thus, no dose data were used in analyses for this report. The committee's letter report to the Defense Nuclear Agency, this study's sponsor and the Department of Defense entity responsible for the NTPR program, describes the relevant limitations of the data for these purposes and suggests ways to create dose data that could be appropriate for epidemio- logic use (IOM, 1995; reprinted in this report as Appendix A). FINDINGS AND DISCUSSION Veteran concern about radiogenic cancer was a major impetus for this re- search. That leukemia, the cancer most consistently linked with radiation, is fairly rare is fortunate. However, that presents an obstacle to a study of this kind. Only a study cohort four times the size of the one available would have been likely to identify the observed leukemia risk as statistically significant. Although dose data might have increased the study's statistical power to detect an increased risk among participants (if there were an increased risk), these data were judged inappropriate for epidemiologic analysis. In the absence of epidemiologically useful dose data, the focus of this research shifted to an examination of the hazards associated with test participation, irrespective of dose. Overall, no statistically significant differences are evident in all-cause, all- cancer, or leukemia mortality between participants and referents. However, al- though not statistically significant, the risk of leukemia mortality was elevated in participants compared to referents. Among the leukemia subtypes, the highest relative risk of death was for lymphoid leukemia, excluding chronic lymphoid leukemia. While the estimated relative risk of leukemia mortality was higher among participants in land series than among participants in sea series, there was no articulated a priori basis to have predicted such a finding. Also not iden- tif~ed a priori was the association between sea series participation and all-cause mortality that we observed in this study. In addition, significantly elevated rates

OCR for page 1
4 THE FIVE SERIES STUDY of death due to external causes of injury, nasal cancer, and prostate cancer were found among participants, again compared to referents. The set of leukemia findings presented here Is broadly consistent with a ra- diogenic cause, but is not conclusive. An increase in external causes of death has been found in other studies of military personnel who were not exposed to ra- diation. Explanations other than radiation effects might be pursued. Other find- ings-nasal and prostate cancers would not have been expected based on other studies of the health effects of radiation exposure. What, then, is the substantive significance of these findings? We can state that the participant group as a whole did not experience widespread early death. Even for leukemia, for example, there were an estimated 25 excess deaths in the participant cohort. That might be a comfort to those veterans who are not sick and to their families. The report findings to do not rule out, however, possible increased risk among distinct subgroups of test participants that this study did not have the information to identify accurately. Stronger supporting evidence could be acquired from a further study that would make use of data on radiation dose if those data could be developed. A1- though the oversight committee concluded that the dose data in their current form were unsuitable for epidemiologic analysis, it also concluded that carefully carried out custom dose reconstructions done anew for selected participants, using con- sistent methodology, could provide usable dose data. An efficient research design (to minimize the prohibitive cost of custom dose reconstructions) requiring fewer individuals is a nested case-control study, which could focus on specific endpoints of interest, such as leukemia. The pattern of radiation dose among the leukemia deaths (cases) would be contrasted to the pattern among a sampled set of partici- pant controls to assess a hypothesized dose-response association. The size, length of follow-up, and persistence of data collection efforts in- volved in this Five Series Study have helped to assure us that the findings we report are valid. It is unlikely that another cohort study of this type and magni- tude would provide more precise answers than this one, because any atomic vet- eran study of this kind would face the same methodologic problems, namely inadequate exposure (dose) data and imperfect mortality ascertainment, en- countered in this Five Series Study. . ~ . . . . . ..