Conclusion

Chapter 10: The Bomb Tests: Questions of Risk, Records, and Trust

AEC and DOD Risk Analysis for Exposure at Bomb Tests

As we have seen in the case of the cloud flythrough research, by the mid-1950s the AEC and the Defense Department had arrived at a method of operation through which waivers to the basic radiation safety standards for the tests would be granted for particular activities. In the early 1950s, in the context of the Desert Rock exercises, the AEC and the DOD established the precedent for departure from the standards that the AEC relied on for its own bomb-test work force.

At this time the AEC was the main source of expertise on radiation effects. Its guidepost for its own workers (at the Nevada Test Site and elsewhere) was the 3 R per thirteen-week standard established for occupational risk by a private organization (the National Committee on Radiation Protection). This level, it may be recalled from the debates on nuclear airplane experimentation (discussed in chapter 8), was well below that at which the experts assumed acute radiation effects, such as would limit combat effectiveness, could occur.[141]

In 1951, the Los Alamos Laboratory, the AEC's right hand in weapons test management, called on the Division of Biology and Medicine's director, Shields Warren, for "official but unpublicized authority to permit exposures up to 3.9r" for AEC test personnel.[142] Warren granted the request, counseling that "this Division does not look lightly upon radiation excesses."[143]

As we have seen, the DOD shortly thereafter determined to use the tests for troop maneuvers and did so at Desert Rock I, keeping the troops at seven miles distance during the detonation. In early 1952 the DOD asked the AEC to endorse its request to station troops at Desert Rock IV as close as 7,000 yards from ground zero (approximately four miles), far closer than the seven-mile limit the AEC permitted its own test-site personnel. The AEC's Division of Military Applications was willing to concur. Shields Warren, however, dissented on grounds of safety.[144] The dispute was settled when AEC Chairman Gordon Dean advised DOD that "the Commission would enter no objection to stationing troops at not less than 7000 yards from ground zero," provided that proper precautions were taken.[145]

Even so, an internal review of the Desert Rock IV exercise by the Division of Military Applications, generally supportive of DOD's request for troop maneuvers, raised questions about the wisdom of deviation from the AEC standard--and the potential for "delayed" casualties.[146]

Determined to proceed, DOD called for "a study to be made to determine the minimum distance from ground zero that should be permitted in a peacetime maneuver."[147] A December 1952 report recommended that dosages for Army personnel be above the limit set by the AEC for its personnel. The soldiers, by comparison with the AEC personnel, would be exposed "very infrequently." The report summarized the state of knowledge:

There is no known tolerance for nuclear radiation, that is, there is no definite proof that even small doses of nuclear radiations [sic] may not, in some way, be harmful to the human body. On the other hand, there is no evidence to indicate that, within certain limits, nuclear radiation has injured personnel who have been exposed to it.[148]

In response to the DOD's proposal to assume full responsibility for physical and radiological safety of troops and troop observers within the Nevada Test Site, the AEC stated that general safety practice and criteria at the Nevada Proving Grounds was, and must continue to be, the responsibility of the AEC. The AEC did, however, "accept the proposal that the DOD assume full responsibility for physical and radiological safety of troops and all observers accompanying troops within the maneuver areas assigned to Exercise Desert Rock V, including establishment of a suitable safety criteria." The AEC further explained that

The Atomic Energy Commission adopts this position in recognition that doctrine on the tactical use of atomic weapons, as well as the hazards which military personnel are required to undergo during their training, must be evaluated and determined by the Department of Defense.

The Atomic Energy Commission has, however, established safety limits. . . . We consider these limits to be realistic, and further, are of the opinion that when they are exceeded in any Operation, that Operation may become a hazardous one. So that we may know in which particulars and by how much these safety standards are being exceeded, we desire that the Exercise Director transmit to the Test Manager a copy of his Safety Plan. . . .[149]

For the spring 1953 Desert Rock V exercises, the DOD deemed the permissible limit for the troops (for a test series) to be 6 R.[150] In the case of the officer volunteers, a 10 R test limit was agreed to, with the proviso that "it is not intended that these exposures result in any injury to the selected individuals."[151] The Army's limit at Desert Rock was well below the level understood to potentially cause acute effects and far below the recommendation of Brigadier General James Cooney that the military depart from the "infinitesimal" industrial and laboratory limits and accept 100 roentgens for a single-exposure limit.[152] But the level was not only higher than the AEC level but also above the 0.9 R per week being urged by the British and Canadians, partners in U.S. testing.[153] (The AEC itself objected that a 0.9 R-per-week limit would make testing at Nevada impractical.)[154]

Interestingly, in 1952 the Navy, also faced with the need for more-realistic training exercises, considered spraying radioactive materials on ships during training exercises. The Navy's Bureau of Medicine (BuMed) rejected the proposal. BuMed told the chief of naval operations that while it "fully appreciates" the need for more "realistic radiological defense training," it could not approve the use of radioisotopes in a form other than "sealed sources commonly used in basic training . . . since such use might produce an internal radiation hazard serious enough to outweigh the advantages of area contamination for training purposes."[155]

By the mid-1950s, AEC test health and safety staff were continually concerned about radiation safety at the tests and the failure to reduce them to a predictable and assuredly safe routine. "There are," Los Alamos Health Division leader Thomas Shipman wrote to the AEC Division of Biology and Medicine's Gordon Dunning in 1956, "two basic facts . . . which must never be lost sight of. The first of these is that the only good exposure is zero. . . . The second fact is that once the button for a bomb detonation is pushed you have to live with the results no matter what they are. . . ."[156] In fact, while the AEC had set a limit of 50 kilotons (more than twice the power of the Hiroshima and Nagasaki bombs) for Nevada tests, this limit had already been exceeded by 10 kilotons in 1953.[157] "It is all very nice," Shipman wrote in another 1956 memorandum, "to have a well-meaning Task Force commander who by a stroke of the pen can absolve our radiologic sins, but somehow I do not believe that overexposures are washed away by edict."[158] Shipman's comments illustrate an acute awareness among experts at the center of the testing program of the real and continuing element of risk and uncertainty in the attempt to define and control exposures at the bomb tests.

The Aftermath of Crossroads: Confidential Record Keeping to Evaluate Potential Liability Claims

During the summer of 1946, the contamination of ships at the Crossroads tests put officials and medical experts on alert to the radiation risk posed to participants at atomic bomb detonations. "[D]ifficult and expensive medico legal problems," Crossroads medical director Stafford Warren feared, "will probably occur if previously contaminated target ships are 'cleared' for constant occupancy or disposal as scrap."[160] A "Medico-Legal Advisory Board" sought to deal with these questions,[161] and the Navy created a research organization dedicated to the study of decontamination and damage to ships.[162]

Concern for long-term liability stimulated by Crossroads led to more steps to guard against the legal and public relations implications if service personnel exposed to radiation filed disability claims.

In the fall of 1946, General Paul Hawley, administrator of the Veterans Administration, "became deeply concerned about the problems that atomic energy might create for the Veterans Administration due to the fact that the Armed Services were so actively engaged in matters of atomic energy."[163] In August 1947 Hawley met with representatives of the surgeon general's offices of the military services and the Public Health Service.[164] The meeting was also attended by former Manhattan Project chief General Leslie Groves,[165] (Groves reportedly was "very much afraid of claims being instituted by men who participated in the Bikini tests.")[166] An advisory committee was created, which included Stafford Warren and Hymer Friedell, Warren's deputy on the Manhattan Project medical team. The committee was given the name "Central Advisory Committee," as "it was not desired to publicize the fact that the Veterans Administration might have any problems in connection with atomic medicine, especially the fact that there might be problems in connection with alleged service-connected disability claims."[167]

The committee recommended the creation of an "Atomic Medicine Division" of the VA to handle "atomic medicine matters" and a radioisotope section to "implement a Radioisotope Program."[168] The committee further recommended that "for the time being, the existence of the Atomic Medicine Division be classified as 'confidential' and that publicity be given instead to the existence of a Radioisotope Program."[169]

This history is contained in a 1952 report presented by Dr. George Lyon to the National Research Council.[170] The 1952 report records that "General Hawley took affirmative actions on these recommendations and it was in the manner described that the Radioisotope Program was initiated in the Fall of 1947."[171] Lyon, who had worked with Stafford Warren at Crossroads, was appointed special assistant to the VA's chief medical director for atomic medicine, and through 1959 served in a variety of roles relating to the VA's atomic medicine activities. Dr. Lyon's 1952 report recounts that he was present at the August 1947 meeting and involved in the deliberations of the Central Advisory Committee, as well as subsequent developments.[172]

Working with the VA and the Defense Department, we sought to retrieve what information could be located regarding the Atomic Medicine Division and any secret record keeping in anticipation of potential veterans' claims from radiation overexposures. Among the documents found was a Confidential August 1952 letter to the attention of Dr. Lyon, in which the Defense Department called for comment on the Army's proposal to "eliminate the requirement for maintaining detailed statistical records of radiological exposures received by the Army personnel."[173] The requirement, the letter recorded, "was originally conceived as being necessary to protect the government's interest in case any large number of veterans should attempt to bring suit against the government based on a real or imagined exposure to nuclear radiations during an atomic war."[174]

In 1959 Dr. Lyon was recommended for a VA "Exceptional Service Award."[175] In a memo from the VA chief medical director to the VA administrator, Dr. Lyon's work on both the publicized and confidential programs was the first of many items for which Dr. Lyon was commended. Following a recitation of the 1947 developments similar to those stated by Dr. Lyon in his 1952 report, the memo explained:

It was felt unwise to publicize unduly the probable adverse effects of exposure to radioactive materials. The use of nuclear energy at this time was so sensitive that unfavorable reaction might have jeopardized future developments in the field . . . [Dr. Lyon] maintained records of classified nature emanating from the AEC and the Armed Forces Special Weapons Project which were essential to proper evaluation of claims of radiation injury brought against VA by former members of the Armed Forces engaged in the Manhattan project.[176]

The Advisory Committee has been unable to recover or identify the precise records that were referred to in the documents that have now come to light. An investigation by the VA inspector general concluded that the feared claims from Crossroads did not materialize and that the confidential Atomic Medicine Division was not activated.[177] However, the investigation did not shed light on the specific identity of the records that were kept by Dr. Lyon, as cited in the 1959 memo on behalf of his commendation.[178]While mystery still remains, the documentation that has been retrieved indicates that prior to the atomic testing conducted in the 1950's, the government and its radiation experts had strong concern for the possibility that radiation risk borne by servicemen might bear longer-term consequences.

Looking Back: Accounting for the Long-Term Risks

This may have been the case following Crossroads. "Now we are very much interested in long-term effects," a military participant in a 1950 meeting of the DOD Committee on Medical Sciences stated, "but when you start thinking militarily of this, if men are going out on these missions anyway, a high percentage is not coming back, the fact that you may get cancer 20 years later is just of no significance to us."[182]

Decades following the 1946 Crossroads tests, researchers began to study the longer-term effects of the bomb on test participants.

In 1980 the Centers for Disease Control (CDC) reported a cluster of 9 leukemias among the 3,224 (then identified) participants of shot Smoky at the Nevada Test Site in 1957.[183] A later report[184] increased the count of leukemias to 10 compared with 4.0 expected on the basis of U.S. rates, but found no excess cancers at other anatomical sites (the total observed was 112, compared with 117.5 expected). The Smoky test was the highest-yield tower shot ever conducted at the Nevada Test Site; however, the measured doses for the Smoky participants as a group were too low to explain the excess. Whether this cluster represents a random event, an underestimation of the doses for the few participants who got leukemia, or some other explanation remains unclear.

In light of the CDC research, the National Academy of Sciences (NAS) thereafter undertook an enlarged study of five series of nuclear tests totaling 46,186 (then identified) participants.[185] The 1985 NAS report confirmed the excess of leukemia at the Smoky test but found no such excess at any of the test series (as opposed to individual tests) and no consistent pattern of excesses at other cancer sites. Later, however, the NAS study was found to be flawed by the inclusion of 4,500 individuals who had never participated and the exclusion of 15,000 individuals who had participated in one or more of the five series, as well as incompleteness of dosimetry.[186]

The belated discovery that thousands of test participants had been misidentified punctuated the deficiencies in record creation and record keeping faced by those who seek to reconstruct, at many years' remove, the exposures of participants at the tests.

Documents long available, and those newly retrieved by the Committee, provide further basis for concern about the data gathering at test series in which human subject research took place. At the 1953 Upshot Knothole series, which included the Desert Rock V HumRRO research, 1994 DOD data show that only 2,282 of the 17,062 participants are known to have been issued film badges to serve as personal dosimeters.[187] At Desert Rock V, the Army surgeon general's policy that one-time exposure need not be reported led to a determination that maneuver troop units would be issued one film badge per platoon, and observers would be issued one per bus.[188] An AFSWP memo recorded that the Radiological Safety Organization did not have enough pocket dosimeters for efficient operation.[189] A recently declassified DOD memo records that "[a]lthough film badges on the officer volunteers [at Desert Rock V] indicated an average gamma dose of 14 roentgens, best information available suggests that the true dose was probably 24 rem initial gamma plus neutron radiation."[190]

In a 1995 report, the Institute of Medicine found that the dose estimates that were proposed for use in the NAS follow-up study were unsuitable for epidemiologic purposes, but concluded that it would be feasible to develop a dose-reconstruction system that could be used for this purpose. Nonetheless, there are some further studies that are of direct relevance.[191]

Recently, Watanabe et al.[192] studied mortality among 8,554 Navy veterans who had participated in Operation Hardtack 1 at the Pacific Proving Grounds in 1958. This is, to date, the only study of U.S. veterans to include a control group of unexposed military veterans. Overall, the participant group had a 10 percent higher mortality rate, but the cancer excess was significant only for the combined category of digestive organs (66 deaths compared with 44.9 expected, a 47 percent increase). On average, the radiation doses were low (mean 388 mrem), but among the 1,094 men with doses greater than 1 rem, there was a 42 percent excess of all cancers. No categories of cancer sites showed a significant excess or clear dose-response relationship, but the number of deaths in any category was small.

Two sets of foreign atomic veterans have been studied. In a study of 954 Canadian participants,[193] no differences with matched controls were found, but only very large effects would have been detectable in such a small study. In contrast, a large study of British participants of test programs in Australia found higher rates of leukemia and multiple myeloma than in a matched control group (28 vs. 6).[194] However, the cancer rates among the exposed veterans were only slightly higher than expected based on national rates, whereas those in the control group were much lower than expected, and there was no dose-response relationship. No excess was found at any other cancer site. Although the difference between the exposed and unexposed groups was quite significant, the interpretation of this result is unclear. Does it mean that for some unknown reason, soldiers are less likely than the general population to get cancer (the "healthy soldier effect," which is usually not thought to be so large for cancer), or is it an indication of some unexplained methodological bias? This point has never been resolved.

These observed effects need to be put in the context of what might reasonably be expected based on current understanding of low-dose radiation risks and the doses the atomic veterans are thought to have received. Approximately 220,000 military personnel participated in at least one nuclear test. The film badges for those monitored (thought to be roughly representative of all participants) average 600 mrem.[195] As summarized in "The Basics of Radiation Science" section of the Introduction, the consensus among scientific experts is that the lifetime risk of fatal cancer due to radiation is approximately 8 per 10,000 person-rem. On this basis, one might anticipate approximately 106 excess cancer deaths attributable to participation in the nuclear tests. Not only is this a number with considerable uncertainty, it is small in comparison with the total of about 48,000 cancer deaths that are eventually anticipated in this population.

Such a small overall excess would be virtually impossible to detect by epidemiologic methods. In some subgroups, however, the relative increase above normal cancer rates could be large enough to be detectable. Leukemia, for example, is proportionally much more radiosensitive than other cancers and the largest excess occurs fairly soon after exposure, when natural rates are low. Focusing on those with highest exposure would also enhance the relative increase, albeit with many fewer people at risk. The Defense Nuclear Agency estimates that about 1,200 veterans received more than 5 rem (mean 8.1 rem).[196] On this basis, about eight excess cancer deaths would be anticipated. These factors may have contributed to the observed leukemia excess among participants of shot Smoky, for example.

Although these numbers represent the best estimate currently available of the expected cancer excess, there are uncertainties in both the real exposures received by the participants and the magnitude of the low-dose risk. As described in "The Basics of Radiation Science" section, there is roughly a 1.4 uncertainty in the low-dose radiation risk coefficient simply due to random variation in the available epidemiologic data, with additional uncertainties of unknown magnitude about model specification, variation among studies, extrapolation across time and between populations, unmeasured confounders, and so on. These uncertainties are hotly contested, although the majority of radiation scientists believe the figures quoted above are unlikely to be seriously in error. If low-dose radiation risks were indeed substantially higher than this, then there would be a serious discrepancy to explain with the effects actually observed at higher doses. The uncertainties in the doses received by participants are perhaps more substantial, but given the limitations in the dosimetry and record keeping, it may be difficult ever to resolve them.

As is clear from the epidemiologic data available today, there is no consistent pattern in increased cancer risk among atomic veterans, although there are a number of suggestive findings, most notably the excesses of leukemia among shot Smoky and British test participants, the causes of which are still unclear. The low recorded doses, the small size of the expected excesses, and problems in record keeping and dosimetry make it very difficult to resolve whether atomic veterans as a group are at substantially elevated cancer risk and whether any such excess can be attributed to their radiation exposures. The Advisory Committee debated at some length the merits of further epidemiologic studies and concluded that the decisions to conduct such studies should be made by other appropriately constituted bodies of experts.

Looking Back: The Legacy of Distrust

The Veterans Dioxin and Radiation Exposure Compensation Standards Act of 1984 provides for claims for compensation for radiation-related disabilities for veterans exposed at atmospheric tests. The Radiation Exposed Veterans Compensation Act of 1988 provides that a veteran who was exposed to radiation at a designated event and develops a designated disease may be entitled to benefits without having to prove causation.[197]

Notwithstanding the passage of this legislation, the Committee heard from many atomic veterans, and their widows, who complained that the records that were created and maintained by the government--records on which veterans' claims may stand or fall--were inadequate, missing, or wrong.[198] Atomic veterans also stated that the laws and rules do not adequately reflect the kinds of illnesses that may be caused by radiation, that they do not provide for veterans who were exposed to radiation in settings other than atmospheric tests, and that the practical difficulties--in time and resources--of pursuing their rights under the laws are often excessive. The Committee heard from many who told of the time, expense, and difficulty of getting information on the full circumstances of bomb-test exposures. They told of their continued efforts, over the course of the years, to reconcile what they have learned from government sources with that which they have been told by other test participants, that which they recovered from the private letters of test participants to family members, and their own further research.

For numerous atomic veterans, the testimony was not simply that the bomb tests themselves had been large experiments, but that they had been put at risk in the absence of planning to gather the data and perform the follow-up studies needed to ensure that the risks of the unknown, however small, would be measured and adequately accounted for.