The Five Series Study.
Mortality of Military Participants in U.S. Nuclear Weapons Tests.

Institute of Medicine
Public Briefing
Oct. 20, 1999

Opening Statement by:
Susan Thaul
Study Director

 More than 200,000 U.S. soldiers, sailors, airmen, and marines participated in atmospheric nuclear tests between World War II and the early 1960s. Now, decades later, some have asked whether participation in those tests can be linked to death or illness among those who took part. To help answer this question, our study looked into the causes and rates of death among nearly 70,000 of these test participants: those at three test series in the South Pacific -- Operations GREENHOUSE, CASTLE, and REDWING; and two test series in Nevada -- Operations UPSHOT-KNOTHOLE and PLUMBBOB.

 We used records compiled or maintained by the Department of Defense, the Department of Veterans Affairs, and the National Archives to identify this group of test participants and a comparable group of about 65,000 service personnel who did not participate in nuclear tests. Then we tracked the death certificates in each group to learn cause of death.

 In analyzing the data, we looked first at overall death rates and total deaths from cancer. We found no difference between participants and nonparticipants in either one.

 We also investigated specific causes of death, focusing primarily on leukemia. Participants had an apparent 14 percent higher risk of death from leukemia than the comparison group, although that difference was not statistically significant and could be a chance finding.

 Because the initial analysis plan had singled out leukemia as a primary target for investigation, we also examined subcategories of participants. For example, land-based participants -- those in the Nevada desert -- had a death rate from leukemia that was 50 percent higher than military personnel in similar units who did not take part in atomic tests. Sea-based test participants in the South Pacific, however, did not differ from their comparison group in leukemia deaths.

 These leukemia findings do not resolve the debate over whether participation is associated with leukemia mortality. However, the set of leukemia findings is consistent with the results of other studies of military participants in nuclear tests and is broadly consistent with a hypothesis that these are radiation effects.

 We also had some unanticipated results regarding two other kinds of cancer -- prostate and nasal. Death rates from prostate cancer were 20 percent higher among test participants than the comparison group, and even higher for nasal cancer. The prostate cancer findings have not been consistently seen in other studies of people exposed to radiation and are therefore difficult to interpret. The nasal cancer finding is even harder to interpret, in part because this is the first study of atomic test participants to look specifically for that cause of death. Nasal cancer has not been among the cancers considered to be caused by radiation.

 How could we better understand the associations we observed? One way would be to examine information on the size of the radiation dose received by each veteran to see whether participants with higher doses were at higher risk of death. But, at the time of the tests, such data were not collected specifically for medical research; dose measurement and records maintenance were neither complete nor consistent. In the ensuing decades, the federal government has used the available information to reconstruct doses -- not for research purposes, but to support veterans' claims for compensation. However, after a review of available data by our expert advisory panel, we chose to not use the dose information, finding it unsuitable for research of this kind.

 Is it possible to reconstruct information on dose that could be useable? Perhaps. Stronger supporting evidence could be acquired from a further study that would make use of such data. An efficient research design would focus on specific endpoints of interest, such as leukemia, and use carefully carried out, custom dose reconstructions done anew for selected participants, using consistent methodology. Then, for example, the pattern of radiation dose among the cases involving leukemia death could be contrasted to the pattern among a sampled set of participants who did not die from leukemia to assess a hypothesized dose-response association.

 Today, though, we meet to discuss not what we might find, but the meaning of what we have found: 1) The participant group as a whole did not experience earlier deaths than those who did not participate in nuclear tests. We cannot rule out, however, the possibility of increased risk among distinct subgroups of test participants, which this study did not have the information to identify accurately. 2) We found an increase in leukemias consistent with a hypothesis that these are radiation effects.

Susan Thaul
Study Director