Click for next page ( 26


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 25
DISCUSSION In a well-controlled study, subjects are randomly selected for assignment to various test or control groups. In the Edgewood tests, members of short-term control Grouts were later assigned to test groups and thus were lost as long-term controls. There were two main reasons for this procedure: the Edgewood studies were conducted to determine immediate behavioral effects that might be important in a military situation, and the later exposure of controls to experimental chemicals enabled the experimenter to make multiple use of each volun- teer. A need for evaluating long-term health effects was not fore- seen. _ One might therefore say that this arrangement precludes a proper assessment of long-term effects of the Edgewood tests. Strictly speaking, that is true. However, the present evaluation can support many useful inferences. For example, the lack of excess malignancies among a test population that has received a topical carcinogen would be a significant finding. The lack of excess malignancies and other debilitating diseases in the entire test population would be impor- tant. Because of shortcomings in test design, this evaluation is not likely or even intended to reveal minor health deficiencies that might have resulted from the test experience. Only major problems that occur in a large number of men are likely to be uncovered. . ~ . The subjects were not assigned to treatments in a formal ran- _ _ _ ~ domized manner. To be eligible for exposure to the test chemicals, the volunteers had to pass additional physical and mental tests that would have selected the most fit men for chemical testing and the less fit men for testing of equipment and relatively innocuous materials. Nonetheless, two comparison groups could be constructed. The first consisted of the NCT (no-chemical-test) men, and the second, of all men tested with chemicals other than those of interest (the OCT, or other-chemical-test, group). Using the NCT men as a comparison group would tend to under- estimate chemical effects, because the NCT men, having been less fit at the outset, might be expected to have more illnesses than the men tested with chemicals. Using the OCT men as a comparison group - resolves this problem, but suffers from the possibility that, if more than one chemical led to deleterious effects, we would be comparing one potentially affected group with another potentially affected group. However, because the a priori expectation of the kinds of damage that might be anticipated from each class of chemical would most likely be different, it is unlikely that this kind of loss of information would occur. -25-

OCR for page 25
Suggestions to use additional comparison groups from other populations as surrogate controls were quickly turned aside because information on the composition of test groups needed to select com- parable control groups (race, religion, socioeconomic status) was available. Furthermore, the effects of the volunteers' desire to participate in the Army studies could not be controlled for. The Committee has already assessed the possibility of long-term adverse health effects of short-term exposure to the chemical agents tested at Edgewood in two reports (see Appendix A for their executive summaries). The conclusions in those reports were based on a review of literature reporting acute and chronic effects, on the dosages administered to soldiers, and on immediate effects and acute findings reported by clinician observers. The evaluations yielded almost no significant positive findings. The work reported here involved an evaluation, based on a ques- tionnaire, of the current health status of subjects 10-30 years after testing. The questionnaire asked 15 questions, the most important specifically targeted at learning whether test subjects had experi- enced higher prevalences of cancer, mental disorders, necrologic disorders, or reproductive effects than members of comparison groups. The results do not indicate that important effects were seen. Answers to questions 4, 9, 11, 13, and 15 a and b were directly pertinent to the current health status of the subjects. The health status of test subjects does not appear to have been significantly altered, according to responses to the questionnaire. There are several reasons why this study might have low power to detect some long-term effects (see Appendix C). Mail surveys always miss some information and include potential bias; both flaws can result from failure to locate some intended recipients or from failure of recipients to respond. Some questions require the recall of health status or job experience over several years; these kinds of questions often lead to misreporting. Subtle effects on health usually can be assessed only through physical examination and testing. The question- naire used in this study was largely Invalidated, so its sensitivity and specificity for particular health problems and life quality are unknown. Some attempts at validation were made, however. A subset of volunteers who reported having had three to five children were inter- viewed by telephone to validate some questions for larger families. But the questions regarding substance abuse and health effects were not validated. Finally, the study had no true control groups. The baseline groups used in this report for comparison, the NCT group and possibly the OCT group, might be expected to have poorer health than a true control group; if that were the case, the probability of detect- ing some health effects would be decreased. Hence, the objective of this study was to detect major, long-term health effects of the expo- sures to chemicals. Additional information on admissions to Army and VA hospitals was available. The data on admissions during the whole period after -26-

OCR for page 25
exposure were evaluated. The data on Army hospital admissions would give some indication of severe, short-term effects, and the data on VA hospital admissions would provide information on possible severe, long-term effects in men who continued to use the VA system. However, the probability of detecting long-term effects was decreased by the low rate of use of the VA system among men discharged from the service. Again, therefore, the comparisons would detect only major health effects of exposure. Because so many comparisons (more than 756) were included in this study, it is almost certain that at least one chance difference would be declared "significant" at the O.O1 level. To rule out such chance occurrences, attempts were made to develop corroborating evidence, such as evidence of a dose-response relationship or biologic plausi- bility. In addition, comparisons were made after controlling or stan- dardizing for known confounding factors. Another statistical consideration that is relevant to the inter- pretation of the findings of this study is that the sampling errors are large for the outcomes, because of the nature of the response rates and the nature of the data collected. "Sampling error" refers to the notion that accurate estimation of the range of outcomes requires extremely large numbers of responses. Because many of the groups in this study are small, the ability to detect a true effect (i.e., power) is low. Tables 3 and 4 demonstrate the need for very large populations. ANTICHOLINESTERASE CHEMICALS The primary health concern regarding subjects tested with anticholinesterases\was that long-term health effects might occur in the form of subtle changes in EEG, sleep pattern, and behavior--such as increased irritability, inability to concentrate and depression-- that could persist for more than a year (Appendix A). However, if these changes occurred and persisted, they might be difficult to detect. They might have been identified by the subjects as general health problems or, in severe instances, identified by physicians as mental disorders. In fact, answers from subjects who received anticholinesterases compared favorably with answers from NOT subjects. Posttest admissions to Army or VA hospitals for mental disorders did not appear to be significantly increased (Table 18), either during the years immediately after testing or later. The responses to ques- tions about current health status by subjects exposed to anticholines- terases suggest that, as a group, these subjects were no different from the NCT comparison group or from the remainder of the test sub- Jects. If subtle changes occurred, they were not revealed by the subjects' answers about their current health status. -27-

OCR for page 25
There was a borderline significant increase in malignant neoplasms among soldiers who were admitted to VA hospitals (but not Army hospitals) and were exposed to anticholinesterases, compared with those who received no chemical testing. The neoplasms occurred at various sites, and no consistent pattern was seen. Current animal studies show that this pharmacologic class is unlikely to have induced malignancies among the Edgewood subjects; that conclusion is based on a review of NCI-sponsored lifetime studies of animal bioassay for car- cinogenesis at maximal tolerated doses of ten anticholinesterase organophosphate pesticides. \6 - 25 ANTICHOLINERGIC CHEMICALS According to the published literature, the primary health concern for subjects tested with anticholinergics might be short-term cardio- vascular effects. No clear indication of such effects over a long period was found. No evidence of differences between these subjects and others was found, with respect to current health status or first admission to a military hospital. All the chemical-test groups showed a trend toward increased rates of admission to Army hospitals per person-year for the first 5 years after testing (Table 12~. However, the rate was greatest among the volunteers exposed to anticholiner- gics. After the first 5 years, there was no evidence of a higher rate among these volunteers than among the others. An apparent difference in fertility was noted between these subjects and the NCT subjects or OCT subjects (Tables 26 and 27~. However, the exposures to anticholinergics occurred relatively late in the series of tests (Table 2~. Current age and marital status were taken into account in the estimation of expected values in Tables 26 and 27, but other cohort and social differences might account for the smaller family sizes. There has been a trend toward lower birth rates and greater ages at conception during the last decade.15 The men exposed to the anticholinergics have the lowest average current age in this population (Table 24), and their lower fertility might reflect these trends in our society. When the appropriate adjustments were performed to take cohort differences into account, the apparent dif- ference between observed and expected fertility rates disappeared (Table 28~. It was therefore concluded that there was no evidence of an effect of the anticholinergics on fertility among the exposed men. There remained, however, a reduction in numbers of male children among the total number of children born after exposure. This re- duction was of borderline significance (p = 0.04~. It was not seen in comparison of males and females among first children born after expo- sure. Statistics describing proportions of children by sex, such as percentages of male children or sex ratios, are made unreliable by small sample sizes. \2 This small size might have contributed to the finding. No published reports were found of human or animal exposures to anticholinergic chemicals that affected the sex of -28-

OCR for page 25
offspring. Isolated reports of exposures and later distortions of the proportion of male children have been published in connection with uranium,8 2 7 dibromochloropropane (DBCP), 2 ~ ~ anesthetic gases, 7 and air pollution containing metals.4 5 The decrease in the proportion of male children in the present study was judged to be a random finding and was to be expected, because so many comparisons were evaluated. CHOLINESTERASE REACTIVATORS A review of the literature disclosed no long-term effects of cholinesterase deactivators (Appendix A). They are eliminated rapidly and produce a variety of short-term, reversible acute effects. These short-term effects might explain in part the slightly increased (nonsignificant) rates of admission to Army hospitals during the first 5 years after testing (Table 12~. However, there was no evidence of a difference in current health status between these subjects and the other subjects. Nor was there evidence of differences in the current social functioning of these subjects, e.g., in employment, marital status, and family life. PSYCHOCHEMICALS A variety of psychochemicals were tested, including Sernyl (phencyclidine) and dibenzopyrans (dimethylheptylpyran and related compounds). A review of the literature found only sparse evidence of the long-term health effects of these chemicals (Appendix A). The target organs of these substances are the brain and cardiovascular system. However, target mental or cardiovascular effects did not persist beyond a week of exposure to the drugs. It was concluded that, at the dosages used, detectable long-term or delayed effects were unlikely. The data supplied by the soldiers in response to the questionnaire and the patterns of admissions of these soldiers to military hospitals did not contradict these conclusions with regard to specific health effects. Of particular interest were the 86 soldiers who were exposed to some form of Sernyl, a purified form of phencyclidine. (The impure street form is reported to have undesirable properties; see Vol. 2 of this series.) Of these soldiers, six were known to have died since testing. A total of 48 soldiers returned the questionnaire. The proportion of volunteers ever hospitalized was lowest among those exposed to Sernyl (Table 14~. The primary health concern for these subjects was mental disorder. Because few subjects were tested with this compound, the expected number of such admissions was low; in fact, the observed first admissions to Army and VA hospitals for mental disorders were not significantly higher than expected values (Table 18~. Similarly, the expected numbers of admissions for malig- nant neoplasms and diseases of the nervous system were low among this group; no such admissions were observed--an indication that there was -29-

OCR for page 25
no apparent increase in these health effects among these volunteers. In general, most reported few or no problems and little or no need for health care. The primary short-term health effects of the dibenzopyrans were moderate to marked and included prolonged orthostatic hypotension. However, there was no indication from the responses on the question- naire that the current health of exposed subjects was affected. Of 254 soldiers who were exposed to cannabinoids, 161 returned the questionnaire. This group, as a whole, had the lowest rate of admis- sions to Army or VA hospitals (Table 15~. They did not appear to differ from the other groups in any way assessed by the questionnaire. IRRITANTS AND VESICANTS Mustard gas has been shown experimentally to be mutagenic and carcinogenic. Other possible long-term effects, specifically blindness and skin tumors, were expected to be related to local toxicity (Appendix A). However, the soldiers who participated in the Edgewood studies were exposed to mustard gas only at low doses and were wearing gas masks and impregnated clothing. Thirty-eight volunteers had skin damage and erythema after exposure to mustard gas. All these subjects returned the questionnaires; no tumors were associated with skin sites affected as a result of the exposure. In general, there appeared to be no significant differences in current health status, functioning, or previous hospital admissions between subjects exposed to any of the irritants or vesicants and the rest of the subjects. LSD DERIVATIVES Of 571 soldiers exposed to LSD, 317 returned completed question- naires. This group did not differ from the NOT or OCT subjects in total hospital admissions, admissions for malignant neoplasms or mental disorders, or current health. The soldiers exposed to LSD did, however, have an increased number of first admissions for nervous system and sense organ disorders. There was prestudy concern about a possible increase in suicide rate or epilepsy rate that might result from exposure to LSD. There was no evidence of such effects in the data collected. But the soldiers did report more use of controlled substances. In particular, they reported rates of LSD use higher than expected rates, according to age-specific reported use either by the NOT subjects or by the OCT subjects (Table 30~. It is thought that there is underreporting of use of controlled substances, even in self-reporting questionnaires. However, before the testing period, the soldiers were informed as to the substances they might be exposed to; perhaps those who knew that they had been exposed to LSD were more willing to report later use of LSD. l -30-