Since the Persian Gulf War, the VA and DoD have undertaken various activities to address potential health consequences of service in the Persian Gulf theater, including activities involving the cooperation of several federal agencies. At the time of this writing (September 1994), the committee had learned of and evaluated the activities listed in Appendix E.
The committee sought to determine what is currently known about the health status of PGW veterans, and about their opportunities for exposures to various agents present in their environment and through their occupations in the Persian Gulf. We requested, received, and reviewed information from researchers involved in the activities listed in Appendix E. In addition, persons involved in some of these activities presented their work to the IOM committee in January and August 1994 (see Appendix B). The quantity and quality of supporting materials the committee received about the activities varied widely. Initially, the federal government lacked any overall strategy by means of which to evaluate a range of potential health effects. There is a need to develop a strategy that will promote a better capacity for the government to deal with these problems in future situations.
In an attempt to characterize and evaluate the various research activities as a response to the health concerns of PG veterans, the committee offers the comments in this chapter. We derived our findings and recommendations from the background information we present here.
In this chapter we discuss activities completed and underway in four major categories, boards and committees; population-based activities; activities
based on health outcomes, both completed or well underway, and new or just beginning; and exposure assessment activities.
BOARDS AND COMMITTEES
The work of this IOM committee, as mentioned in the introduction, was specified under PL 102-585, with the original charge focusing on possible health effects of the oil well fires. Because concerns about adverse health consequences of the PGW have broadened since the law was passed, the IOM committee has considered the full range of potential health effects. Since PL 102-585 was passed in November 1992, the federal government has conducted and sponsored research activities under the auspices of several agencies, including the Department of Veterans Affairs, Department of Defense, Environmental Protection Agency, and Department of Health and Human Services. The earliest research on potential health consequences included some evaluation of the possible occurrence of post-traumatic stress disorder (PTSD) and the development of models to predict cancer risk from airborne products of the oil well fires. Initially, there were few panels, boards, committees, individuals, or councils to oversee, coordinate, direct, supervise, or advise those conducting the various federal activities.
In May 1993, the VA held the first meeting of an informal "blue ribbon panel" of experts in response to a growing concern that PG veterans were experiencing unexplained illnesses. The group was chartered (October 8, 1993) as the Persian Gulf Expert Scientific Committee to advise the VA Assistant Chief Medical Director for Environmental Medicine and Public Health, and subsequently the VA Undersecretary for Health about medical findings affecting Persian Gulf veterans. The committee is charged to review all aspects of patient care and medical diagnoses, and will provide professional consultation as needed. This VA committee may advise on other areas involving research and development, veteran benefits, and training for patients and staff. The newly chartered committee met in February 1994, and again in April and July 1994. This is an ongoing advisory committee, with no set termination date.
The Department of Defense also assembled a group of experts to examine reports of illnesses that could not be diagnosed. As requested by the Undersecretary of Defense in December 1993, a Task Force of the Defense Science Board (DSB) of the DoD was charged to review scientific and medical evidence relating to long-term health effects of exposure to low levels of neuro-toxic agents. The Task Force on Persian Gulf War Health Effects of the DSB first met in December 1993. Dr. Joshua Lederberg, the Rockefeller University, chaired the proceedings. Following the first meeting, the Task Force requested that the charge be changed to focus deliberations on the cause and effect of the
full range of exposure to low levels of chemicals, and also to environmental pollutants, endemic biologics, and other health hazards that might affect veterans of the Gulf War. An interim report was released March 15, 1994; the final report was released in June 1994 (DSB, 1994). The final report by the DSB task force does not provide evidence for any specific cause-effect relationships between putative exposures and an undefined illness. The medical nature and the cause or causes of GWS remained undefined by the Task Force. However, several hypotheses were considered, including some comparisons between chronic fatigue syndrome (CFS) and GWS. The DSB task force is convinced that GWS is not due to chemical or biological warfare agents. This Task Force has completed its work.
During this same time, as concern about unexplained illness among Persian Gulf veterans continued to grow, the Secretaries of Defense, Veterans Affairs, and Health and Human Services established a Persian Gulf Veterans Coordinating Board on January 21, 1994, to ensure interagency coordination of all efforts, separate and joint, in research, clinical care, and disability determination and compensation for post-Operation Desert Shield/Desert Storm illnesses (PGVCB, 1994). The board released an interim report of activities April 1994, and an updated report in May 1994.
PL 102-585 also requires that the President of the United States designate an appropriate department or agency of the federal government to coordinate all research activities undertaken or funded by the Executive branch on the health consequences of military service in the Persian Gulf theater of operations during the Persian Gulf War (see Appendix C). On August 31, 1993, President William Clinton designated the VA as the lead agency for this effort, which also includes the DoD, HHS, and EPA. This Persian Gulf Interagency Research Coordinating Council serves as the "Research Working Group" of the PG Veterans Coordinating Board mentioned above.
Representatives of the Persian Gulf Interagency Research Coordinating Council sponsored a National Institutes of Health (NIH) Technology Assessment Workshop entitled The Persian Gulf Experience and Health, held April 27–29, 1994 (NIH Technology Assessment Workshop Panel, 1994). The workshop initially was expected to be a Consensus Development Conference on the issue of multiple chemical sensitivities (MCS), which some persons believe is linked to GWS. However, the Director of NIH determined that this issue would not fit the guidelines for Consensus Development, as not enough is known or generally accepted on the issue of MCS. Therefore, he assigned members of his staff to assist the Persian Gulf Interagency Research Coordinating Council in planning an appropriate workshop to evaluate PG-related health issues.
The two-and-a-half-day workshop considered the following four questions: 1) What is the evidence for an increased incidence of unexpected illnesses
attributable to service in the Persian Gulf War? 2) If unexpected illnesses have occurred, what are the components of the most practical working case definition(s) based on existing data? 3) If unexpected illnesses have occurred, what are the plausible etiologies and biological explanations for these unexpected illnesses? 4) What future research is necessary?
The panel was unable to develop a definition of the GWS that could be used to determine whether there is an association between exposures (plausible etiologies) and outcomes (otherwise unexplained illnesses). The panel criticized the lack of research and data available to date, and focused its recommendations on the following: 1) a short health questionnaire aimed at all 700,000 people who served in the PGW; 2) a focused hospital/clinical protocol for DoD and VA to use in their research on CFS; 3) designs for cohort and case-control studies of health effects of the Persian Gulf War; 4) a retrospective cohort study of pulmonary function in veterans; 5) retrospective simulation of exposures of possible health interest; 6) research into potential stressors; 7) development of effective responses to diagnosis and treatment of stress-related conditions; 8) planning for prospective data collection; and 9) further research on leishmaniasis, which has been mentioned as a possible cause of illness in PG veterans (NIH Technology Assessment Workshop Panel, 1994).
Reports of a complex of signs and symptoms among PG veterans that could not be attributed to a known diagnosis have led to concerns about a "mystery illness" or "Gulf War Syndrome." Even though these reports remain controversial, veterans, their families, and veteran service organizations have organized support for persons reporting such signs and symptoms. Although the clinical assessment of PG veterans with undefined illnesses has benefited from some coordination of efforts between the VA, DoD, HHS, and EPA, the responses of various branches of the government have not been fully coordinated, and are not organized around clear clinical and scientific goals.
Through its discussions with the various organizations and groups of investigators charged with investigating the health effects of Persian Gulf War service, the IOM Committee has become aware of the very significant barriers to effective coordination of activities among the various government agencies whose participation in such activities will be essential to their success. It has appeared to the committee that the responsibility of these groups has not been translated into effective action and that the impact on epidemiologic research studies appears to be nil. Large numbers of studies that are too small will not provide answers; if actively coordinated, the joint effort may be more effective in approaching a meaningful study size. The current coordination of many of the activities is on paper only. In fact, when several investigators were contacted regarding their work, they were surprised that other federal agencies had been listed as being involved in the coordination of their research. Addition-
ally, it has appeared to the committee that the term "coordination," means that information is simply shared between federal agencies (VA, 1994), but does not usually mean that efforts are made to reduce duplication of similar research activities. Therefore, the IOM committee believes that the authority of coordinating bodies to implement changes in programs and procedures within each of these agencies must be increased substantially. In particular, the committee believes that the leadership of the Vice President of the United States will be needed to achieve the coordination required for this effort. Given the necessary authority, the Persian Gulf Veterans Coordinating Board and the Persian Gulf Interagency Research Coordinating Council can play an important role in this effort. The Persian Gulf Veterans Coordinating Board, while serving many useful functions, does not appear to be taking the lead in coordinating research activities. This role should be filled by some group, somewhere.
Coordinating bodies also can increase the ability of different agencies to link and to establish compatibility among databases. This will require doing detailed work on eligibility criteria, identifying information for individual records, and defining the variables to be linked. Finally, these coordinating bodies can support efforts to synthesize completed studies and to identify gaps in current knowledge.
GENERAL COMMENTS ON STUDY DESIGN
Although addressing public and individual concerns about long-term health effects of service in the Persian Gulf War were prominent in the initial charge to this committee, this issue has been largely displaced in public discussion by the growing concern that veterans and their families already are experiencing a variety of health effects of Gulf War service, including those known collectively as the Gulf War Syndrome. Groups of veterans and their families were the first to focus public attention on the possible adverse health consequences of service in the Persian Gulf. Their concerns have elicited a variety of institutional responses and investigations, as described in subsequent sections of this chapter. Some of these investigations can be characterized as "hot pursuit" or "cluster" studies of groups of apparently affected veterans. Other responses, especially the creation of the VA Persian Gulf Health Registry, represent efforts to identify and provide care to those self-identified as experiencing health problems consequential to Gulf War service. Though these efforts may be seen as responsive to community concerns and may also provide care to those who are ill, the methods used to select participants and collect data make these undertakings intrinsically unsuitable for systematic study of the health effects of the Gulf War.
Because "hot pursuit" studies and the Persian Gulf Health Registry represent samples of self-identified respondents, the data they provide are not representative of the collective experience of the entire cohort of Gulf War veterans. In the absence of data about the population at risk, these studies provide no information about the frequency of specific illnesses or syndromes in that population. Because no control populations are studied, no attributions of health outcomes to Gulf War service or specific exposures can be made. Moreover, although studies of groups of individuals reporting illness sometimes can yield insight about either the characteristics of the illness or the exposures shared by those affected, these early studies of Gulf War veterans have not achieved either of these objectives. Given the level of public concern about health effects of service in the Persian Gulf War, it is vital that future studies of Gulf War veterans be based on more rigorous epidemiological methods.
Defining the population at risk is the priority task in order to assess the health effects of the Gulf War. Therefore, the Department of Defense must complete its Registry of Unit Locations as soon as possible, and also must request a review of that database by scientists competent to assess its sufficiency for future epidemiological investigations. Developing this population registry will create the potential for several investigations that can define the scope of health effects related to Persian Gulf service. First, it will be possible to perform a population-based survey of the current health status of Gulf War veterans. Second, it will be possible to create a cohort of veterans who can be followed to assess long-term health effects. Third, it will provide a context for the conduct of population-based case-control studies of specific illnesses or syndromes. The merits of each of these approaches to further study the health effects of Gulf War service are discussed in subsequent sections. We emphasize, however, that each of these studies will require the creation of a valid population registry. Finally, the committee believes that every proposal to study Gulf War veterans must be subjected to rigorous scientific review and must be evaluated in the context of other research.
The Department of Defense is assembling a "Registry of Unit Locations" as required originally by PL 102-190 and modified by PL 102-585. The committee concludes that the Registry of Unit Locations has the potential to be an essential data source for epidemiological investigation because it will be the single best source of data about service members discovered to date, and because it will attempt to define the entire population at risk. Designed to include basic demographic and personnel data for every member of the military
who served in the Gulf War between January 1, 1991, and the withdrawal of the last troops, also to be included is information on the day-by-day location of all troops at the unit level. This database is designed to interface with a similar computer-based registry of oil well fire exposure data such that individual exposure matrices can be developed. This registry will include procedures both to verify which veteran units served in the Gulf War, and to standardize quality control of the database. Nevertheless, not enough information was available to the IOM committee to determine whether the data will be of a quality suitable for research purposes. Accordingly, this committee has considered ways to improve completeness and accuracy in data acquisition and extraction. The usefulness of this registry in studies of possible cause-effect relations will depend heavily on the quality of the exposure data.
HEALTH-OUTCOME-BASED ACTIVITIES-COMPLETED OR WELL UNDERWAY
Constructing a Case Definition
During and since the Persian Gulf War, a proportion of returned American service personnel have reported conditions that they attribute to their assignment to the Arabian peninsula and its environs, including their experiences during and after the military campaign continuum called Operation Desert Shield and Desert Storm. Most of these medical problems have been diagnosed and treated in the usual way. However some have been misdiagnosed, have defied diagnosis, or have failed to respond to treatment as expected. Of these, some began in the theater of operations before, during, or after the brief period of combat; others followed the service member's return to the U.S., after a time that ranged from days to many months or even years. Furthermore, some of these men and women subsequently became parents, and some of their offspring have suffered birth defects or other ailments.
Over time the symptoms of these persons (some still on active duty, but most returned to civilian life) became increasingly publicized. Networking of various types, inquiries and reports by the media, and responses by politicians to complaints from constituents have coalesced into a broad public perception that there is or may be some type of "mystery illness" resulting from service in the Persian Gulf. Manifestations have been varied from person to person, but often include arthralgia, weakness, fatigue, headache, memory loss, and other mental impairment. Skin rashes and hair loss have also been mentioned frequently. Problems reported in offspring ranged from miscarriage to Down's syndrome to failure to thrive.
No known etiology could account for this collective picture, although descriptive diagnoses could be used for general symptoms (such as alopecia areata versus toxic alopecia versus male-pattern baldness; or adjustment reaction versus neurotic depression versus post-traumatic stress disorder versus malingering). Several hard-to-verify or obscure diseases have been invoked as causes of these reported symptoms, ranging from leishmaniasis to oxidative phosphorylation disorder. The very legitimacy of some mentioned causes has been questioned, such as multiple chemical sensitivity disorder. A variety of toxins have been suspected, including agents of chemical and biological warfare, fumes of both leaded and unleaded fuels, components of smoke from burning oil wells, illicit substitutes for alcohol, and recreational drugs.
Differences in reported rates of the "mystery illness" among different population segments appear to be both substantial and unexplained. For example, there appears to be no similar condition in the indigenous population, which must have been exposed to local environmental hazards, including products of the oil fires, earlier more intensively, and later than the U.S. military forces (Blanck, 1994; DSB, 1994). Similarly, there seems to be no corresponding problem with the "mystery illness" in members of the Allied Armed Forces. At this time, data are not adequate to determine whether substantial differences exist between men and women, regular and reserve troops, and officers and enlisted personnel. These matters also must be considered in framing research questions.
Mystery illnesses are observed from time to time in every population. Approximately 700,000 U.S. military personnel served in the Persian Gulf. Study of a random sample of that number of persons from the U.S. population, of similar age and sex distribution and similar good health at the start of a period of observation, surely would show that some proportion of them would become ill with some kind of "mystery illness" that causes considerable distress and that has not been fully explained and understood within the U.S. medical system. The question thus is not whether there are any "mystery illnesses" among the returning service personnel, but whether the number or characteristics of such conditions exceeds what one might expect if the same persons had not served in the Gulf War. In short, it is not enough to point to a small number of fully confirmed ''mystery illnesses;" a control population is needed to provide U.S.-based background rates of such conditions.
Unusual psychological stress has been promoted as a possible cause or contributing factor to the "mystery illness," but there appears to have been no direct comparison of levels of stress with stress levels reported in other conflicts, such as those in Vietnam, Korea, and World War II. The committee notes that many of the military personnel in the Gulf were older reservists, not new recruits. On the other hand, the committee has heard from several secondary sources that persons in the regular military services do not want to complain about health problems that are less than intolerable, because of possible adverse repercussions on their military careers.
One of the major obstacles to doing studies of unexplained illnesses is the lack of a case definition. Several attempts have been made to form such a definition for the purposes of identifying people to include in an epidemiological study.
The committee agrees with the report of the NIH panel that ''it is impossible at this time to establish a single case definition" (NIH Technology Assessment Workshop Panel, 1994). The Gulf War illness phenomenon may prove to be a mixture of several illnesses, or may prove not to be associated with a specific exposure or disease. Without understanding the nature and etiology of the phenomenon, investigators must be prepared to adapt their case definition to new information as it emerges.
Still, a provisional case definition is required for certain types of scientific research. For example, a provisional case definition could provide a basis for a case-control study comparing the Gulf War experience of veterans with symptoms meeting the case criterion with veterans without such symptoms. If such a study is undertaken, it should meet usual standards for size, rigor, and clarity of purpose as case-control studies conducted in other settings. In particular, a study should be large enough to identify patterns of exposure or Gulf War experience among those complaining of Gulf War illness.
A provisional case definition also may have a role in clinical management of Gulf War veterans. Such a definition might be used to identify veterans eligible for care for these symptoms at VA hospitals. It should be recognized, however, that any such case definition would not have a firm scientific basis. We concur in the recommendations of the NIH panel that eligibility for medical care not be based on case definition.
The proposals for a case definition developed to date have several elements in common, including service in the Persian Gulf theater (proposed to be the time span of August 8, 1990, to late July 1991 (Sanford, 1994)); a requirement that the person report a combination of several symptoms from a list that includes fatigue; arthralgia; headache; diarrhea; neuropsychiatric complaints (forgetfulness, difficulty concentrating, depression, memory loss, irritability); difficulty sleeping; fever; weight loss; dyspnea; rash; and myalgia; and that the diagnosis be made only when known clinical conditions with similar symptoms have been excluded.
Case definition is, at the moment, handicapped by the lack of any generally recognized pathognomonic physical signs or laboratory findings, and by uncertainty about whether a specific syndrome exists and, if it does exist, its prevalence among Gulf War veterans. The subjectivity of many of the complaints associated with the Gulf War illness creates serious problems for those seeking to investigate the validity and origins of the illness. The symptoms have been widely publicized, and any attempt to estimate the prevalence of the problem from self-reports could be contaminated by well-known problems of
self-reporting. Restrictive policies regarding treatment at VA hospitals could create incentives for veterans to report symptoms now as insurance against a future need to seek medical care.
In response to reports of outbreaks of unexplained signs and symptoms in Gulf War veterans, the DoD conducted two outbreak investigations to determine whether further study of these groups was necessary, or whether these particular "clusters" showed any indicators of common exposures that could be pursued as a connection to the presenting signs and symptoms.
These "hot pursuit" studies, in which investigators focused on reported clusters of symptoms or illnesses among Persian Gulf War veterans, are similar in many respects to the frequent "cluster studies" of illness in the United States with a possible environmental cause (Caldwell, 1990). This analogy is instructive, because many of the investigators who have participated in such cluster studies have become skeptical about their scientific value (Rothman, 1990).
The typical cluster study is characterized by small sample size, an implicit multiple comparisons problem (in that many other groups of people could have been studied but were not, because the individuals did not report a cluster), a poorly identified exposure, and a significant potential for information bias resulting from respondent awareness of the underlying concern. In the environmental domain, these studies rarely have been fruitful (Cutler et al., 1986; Bender et al., 1990). A few of these studies, especially if they are done well, can be useful at the early stages of an investigation by helping to define the problem and by ruling out both some statistical flukes that have been misinterpreted and some possible etiologies. That phase of research on the reported "Gulf War Syndrome" has been completed. The exploratory studies failed to generate useful leads about either the condition or the exposures that might cause it. Proposals for future studies of this type should be scrutinized very carefully. Additional studies of this type are unlikely to be useful, and they may divert attention and resources away from studies that could be useful.
The first "hot pursuit" investigation was reported by Major Robert F. DeFraites, MC, and colleagues (DeFraites et al., 1992) who investigated reports of symptomatic complaints among reservists belonging to the 123d Army Reserve Command (ARCOM), Lafayette, Indiana. Early in 1992, staff of the 123d ARCOM Surgeon's Office became aware of these complaints, which were subsequently reported by members of the 417th Quartermaster Company, Scottsburg, Indiana.
In response to growing concern about these complaints, Major DeFraites visited Fort Benjamin Harrison and neighboring facilities in April 1992. During this visit, 79 reservists were evaluated. All 79 study participants completed a medical questionnaire, and 78 were interviewed using a short symptom inventory. Each reservist interviewed completed a brief psychiatric intake-type interview, and had vital signs measured. All but one of the reservists received a dental examination. All 78 who participated in the interviews also had blood drawn for complete blood count, white blood cell differential, platelet count, erythrocyte sedimentation rate, and liver function studies. All sera were tested for antibody to Leishmania tropica. Sera from selected individuals were tested for antibody for brucellosis. Limited comparative data were available from other groups of veterans.
Fatigue was the most common symptom (70 percent). Other systemic symptoms, including fever, abdominal pain, and diarrhea were much less common. The onset of fatigue and associated symptoms tended to occur after redeployment from SWA, except that the onset of diarrhea was more frequent during deployment. No cases of leishmaniasis, brucellosis, or Lyme disease were detected. The objective findings did not suggest a common pattern of illness among the study group members.
Review of potential exposures during ODS/S provided no evidence that the respondents had been exposed to microwaves, chemicals, radiation, or any other environmental hazard. These reservists did report high levels of stress, which may explain some of the symptoms. The investigators noted that the rapid deployment and subsequent redeployment was stressful for many reservists and their families. The investigators believed, however, that PTSD was present in few, if any, of these reservists. The investigators concluded that this study provided no objective evidence of an outbreak of any disease in this group. They thought that the documentable medical problems and illnesses found were typical of a general population with similar demographic characteristics.
When it is not possible to define a population at risk, as in this case, the investigation does not provide a statistical basis for determining whether rates of symptoms or illnesses depart from expected rates by more than ordinary chance variation. Because of the need to contend with the operations of chance, "hot pursuit" is most effective when it focuses on a cluster of a typically rare disease, or when it uncovers a common pattern of exposure. Thus, as noted above, a "hot pursuit" investigation, in which a small group of apparent cases of disease are studied intensely, can on rare occasions be very useful, but it has important limitations. The problem here is that the group coming to clinical evaluation is a small subset of a larger deployed force, and its representativeness to the larger deployed unit is unclear.
As the investigators noted, this study did not use a consistent definition for a "case" of disease. Also, no evidence for a common exposure was found. Because outbreak investigations are not designed to determine rates of symptoms or disease, or to make comparisons to expected population rates, the investigators were able only to conclude that their findings did not appear to be different for such common problems reported as would be found in the general population.
Although DeFraites et al. included their questionnaires and other research instruments as appendices to their report, it is not clear whether and how these instruments were based on standardized questionnaires, nor is the quality of the laboratory measurements documented. Detection limits, sensitivity, and specificity data of tests used (some of which are quite uncommon in the civilian community hospital setting) are not provided. Thus, the interpretation of these laboratory data is necessarily uncertain. This investigation was aided by the use of trained clinical investigators knowledgeable in dental and in psychiatric issues, in particular. This resulted in more precise clinical descriptions in these two important categories and endpoints.
DeFraites et al. reported that symptoms and objective findings seemed to appear in two peaks, one coincident with return from the Gulf, and another some 6 to 8 months after return. They could find no calendar month associations, clustering by activities, or evidence of "dose-response" relationship with increasing length of time in country, which suggests that at least some of the reported symptoms may be related to the reentry of these reservists. The largest proportion of illnesses that caused time lost from work was attributed to injuries, and thus was recognized and explained. The morbidity assessments reported in this investigation do not include data collected by community (civilian) providers, so that only a portion of the health records are available for an unknown subset of participants.
The absence of demonstrable associations between specific symptoms and reported exposures and activities may reflect a limited ability to detect such associations. The similarity of complaints to those observed in civilian (nondeployed) populations may identify either a similarity of the experiences of deployed groups and their civilian peers or reflect a combination of symptoms consistent with complaints in these same persons prior to their activation. Subsets of the deployed may be "hyper-susceptible," on either a biological or an immunological basis or both if exposed or in adjusting to the stresses of this very disruptive, potentially life-threatening activation to military duty.
This study is useful in documenting the absence of a common underlying malady or environmental exposure among a group of reservists who initially appeared to have similar medical problems. It also shows that the complaints of this group did not represent a common pattern of illness or environmental exposure.
Because of the relatively small number of people studied, the absence of a defined population at risk, and the absence of a control group, this study does not rule out more subtle patterns of illness or exposure among the soldiers participating in ODS/S. Nor does the study address the possibility of increased risk for diseases with long latency periods. Somewhat reassuring, however, is that this one study of aspects of a Gulf War-related illness did not show alarming patterns of disease or exposure, other than the possible impact of stress on the troops.
The second "hot pursuit" investigation reported by the DoD involved the Reserve Naval Mobile Construction Battalion 24 (Seabees) (Berg, 1994). Many members of this unit complained of symptoms that they believed were related to their service in SWA. The Asheville, North Carolina detachment was revisited with follow-up questionnaires, personal interviews, and efforts to verify medical diagnoses by reservist-consented contact with medical providers in the community.
A medical team visited the unit in November 1993 to evaluate those Seabees who believed that they were affected. Eighty questionnaires were administered, and the team interviewed Seabees who had been diagnosed as having a medical problem since returning to the United States. Diagnoses were verified by medical record review and by interviews of medical care providers. A special one-page, anonymous questionnaire was administered to veterans, reservists who were not deployed, and a few spouses and significant others.
Respondents included 61 of the 64 members who had served in ODS/S in SWA, and 19 other veterans. The general characteristics of the veterans were unremarkable.
The report centers on the evaluations of the 61 veterans of ODS/S. These veterans reported a variety of symptoms, including fatigue, joint pains, and sleeping problems, but the symptoms did not point to a common underlying medical diagnosis. This study did not identify patterns of disease or exposure that seem to be useful leads for further research.
The number or percentage with verified diagnoses of psychiatric origin was higher than that reported in other deployed units surveyed to date. Those without verified diagnoses had considerably shorter (presumably milder) illnesses with minimal lost time.
The report included self-evaluation of whether specific symptoms were worse, the same, better, or gone in the past year. These reports are helpful to determine which conditions appear to be self-limiting. The symptoms with the greatest "worsening" (joint pains, fatigue, and confusion) may be useful in defining symptom clusters for additional investigation. Symptoms with the most improvement (diarrhea, rash, sleeping problems, and depression) may reflect infections originating in the Persian Gulf that are self-limited and mild or adjustment reactions that have been resolved. Exposure data in the study
were nonspecific and of questionable value. Sera were not banked in this investigation; however, subsequent recommendations from the Navy indicated that banked sera may be useful after diagnostic categories and etiologies of interest are identified.
This study should be viewed as a simple cluster study of 61 Gulf War veterans in a unit that reported an unusual number of health complaints. Much of the report is devoted to comparisons between the veterans with and without a diagnosis found in the DSM-III (APA, 1987), but this comparison is of little value. Like the study of DeFraites et al., this study does not identify either a pattern of illness or a pattern of exposures that would potentially explain the Gulf War syndrome.
Given the information that these first two rounds of investigation provided, further study of this cluster is not likely to be useful.
VA Health Registry and Clinical Evaluations
PL 102-585 included an instruction to the Secretary of Veterans Affairs to establish and maintain a Persian Gulf Health Registry; the purpose and nature of the registry were not specified. The committee has reviewed the systems that have been put into place to collect data relevant to the follow-up of Gulf War veterans. Major ongoing efforts include the PG Family Support Centers, the Persian Gulf Health Registry Examination ("Health Registry Exam"), and the three PG referral centers.
The committee concludes that the Persian Gulf Health Registry may be a valuable clinical tool, but it does not provide information on etiology or disease frequency, and cannot be used alone for research on these matters.
The Health Registry Exam has been used in attempts to standardize both the clinical approach to the Persian Gulf veteran with health problems, and to standardize clinical data collection. The registry data are computerized; however, there appears to be a considerable lag in time between data collection and entry into the computer database. The Health Registry may help to encourage uniform treatment of some of the ailments related to the Gulf War and make it easier to identify cases of illness that could be regarded as sentinel events. It also will further focus the VA on the directed care of Gulf War veterans. The committee has discussed particulars of the examination and data collection (including data quality), and has found that improvements are feasible and needed.
The committee also is concerned about the desire for the Health Registry to go beyond the provision of clinical data. At present, this Health Registry includes data only on veterans who volunteer for evaluation, and so it cannot be used to estimate the frequency of health problems in Gulf War veterans, or
to derive unbiased information on disease-exposure associations. Hence, this Registry cannot be used in the epidemiological evaluation of possible associations between service in the Gulf War and important disease outcomes.
Some patients seen for Registry examinations in a VAMC have unexplained illnesses. If a PG veteran has unusual symptoms that cannot be diagnosed or managed at the local VAMC, the patient is referred to one of three special referral centers located at VA medical centers in Washington, D.C., Houston, Texas, and West Los Angeles, California. Patients referred from local VAMCs to one of these referral centers are followed with consultation and referral back to the local VAMC. These centers were selected, according to the VA, on the basis of the availability of clinical and academic expertise in such areas as pulmonary and infectious diseases, immunology, neuropsychology, and toxicology (VA, 1992a). The clinical assessments at these three centers were not standardized until recently (Farrar, 1994). This clinical protocol has been shared with the Department of Defense researchers and it is the basis for their clinical assessments in the Comprehensive Clinical Evaluation Program (CCEP). The Assistant Secretary of Defense for Health has initiated a program of standard assessments for Gulf War veterans who are still in active military status. Information from this activity should be combined with that from the VA study, so that for clinical uses each study could take advantage of the other regarding the assessment. However, for reasons already noted including reports that active duty personnel may not report symptoms because of their fear of medical discharge, even a combined data set is unlikely to be useful for research.
Given the wide variation in reporting of numbers of veterans who claim to be ill with undiagnosed signs and symptoms, and the fact that the combined total of referrals to these three centers to date is approximately 140 (August 1994), it seems that large numbers of veterans with complaints have not been seen for clinical examinations in these centers. This committee does not know whether this discrepancy is because veterans are unaware of their right to a registry exam, because information has not been adequately disseminated, or because the number of undiagnosed ill veterans is not as large as publicized.
The committee notes that the open, direct conflict in the Gulf was very brief—approximately 100 hours—though persons far from the front lines were vulnerable to certain kinds of airborne attacks over a much longer time. In-theater casualties everywhere were quite low for an armed conflict, and the war was seen from its opening moments as being highly successful. The committee has no way to directly compare levels of stress with those in other conflicts, but we believe that
the total long-term level of stress among military personnel in the Gulf may well have been substantially lower than the average levels of stress among troops of other wars. However, the stresses of the Persian Gulf may not have been the "usual" combat stresses of war, though they were no less real to the persons affected.
The incidence of PTSD, both acute and chronic, in a given population will depend on a host of variables, including the severity and nature of the stressor(s) experienced by each person; predisposing factors ranging from genetic and developmental to preparatory training and interpersonal support networks, or their absence; comorbidity—both physical and psychiatric—timing and effectiveness of initial treatment; and subsequent experiences. It is not surprising that some Gulf War veterans developed manifestations of PTSD, either immediate or delayed (that is, onset more than six months after the stressful experience). From an understanding of PTSD in other settings, one would expect a higher incidence among reserves than among regular troops, because of factors related to selection criteria, nature of training, likelihood of prior combat experience and command structure differences between these two groups. Physical illness can be a predisposing factor, but it also can coexist with PTSD; so that a diagnosis of PTSD does not eliminate the possibility of an organic illness, obvious or subtle.
Not all psychiatric reactions to environmental stress are classifiable as PTSD. A stressor that generates a classic case of PTSD in one person may produce a monosymptomatic neurosis in one of his or her comrades, a psychotic break in a second, a behavioral aberration in a third, and no apparent pathology in a half-dozen others. The caveat about comorbidity, mentioned above for PTSD, also applies to all of these. Finally, there is an expected incidence of many different psychiatric disorders in any given population; therefore, without appropriate controls we could not even approach a conclusion. Veterans hospitals around the United States have completed several formal and informal psychological studies of veterans returning after the Persian Gulf War. Primarily, these studies were convenience samples not determined scientifically, and had variable rates of participation and completion. Although these studies were essentially descriptive, conducted in many settings, and variable in methods used, the results showed fairly consistent findings, and indications of positive results. The IOM committee received information on some of these individual studies as well as a preliminary report summarizing VA experience with troops experiencing war-related stress (VA, 1992b). The studies were conducted in response to PL 102-25, which required an assessment of the needs for rehabilitative services among PG returnees who experience PTSD, programs and resources available to meet those needs, current plans for providing services, the need for additional resources, and plans to coordinate treatment efforts with the DoD.
These studies indicate that the prevalence of diagnosed PTSD was low among returning veterans, but substantial numbers of veterans appear to have experienced a variety of adjustment problems, and many soldiers may have found their problems highly stressful. It seems likely that post-deployment stress contributed to the health problems some soldiers experienced.
Although the findings strongly indicate the occurrence of PTSD, they make no claim of providing a definitive estimated value. These problems were frequently described as related to work or family. Because most of these studies either recruited volunteers or persons who appeared for medical or mental health evaluations, they do not provide accurate estimates of the prevalence of these disorders among Persian Gulf veterans. In addition, although the use of multiple diverse measures may help to validate observations obtained in several ways, the lack of coordination in the effort (for example, not using a core package of instruments) contributes to the difficulty in estimating the prevalence of problems.
The Fort Devens Reunion Survey, Boston VA Medical Center, was one of the earliest efforts to collect empirical data on returning military personnel following the conclusion of Operation Desert Storm. Phase I was a study of returning military personnel that examined behavioral and psychological impacts of war zone experiences. The study population was drawn from 84 units that returned through Fort Devens, Massachusetts, including National Guard (64 percent), reservists (25 percent), and active duty (11 percent). All participants were surveyed within 5 days of arrival at Fort Devens through a self-reported questionnaire. Included were a number of previously validated measures of psychological distress and PTSD symptomatology. The findings were evaluated with respect to race, sex, service status, and stressor exposure determined by questions related to perceived stressors and 3 other surveys. Only 30 percent of those surveyed had more than minimal level of combat exposure. The survey results are largely descriptive with low frequencies of PTSD scores (9 percent women, 4 percent men) and frequencies of general psychological distress reported (approximately 30 percent of both sexes) (VA, 1992b). The Phase I survey has provided good baseline data against which trends and other changes over time can be measured.
Weaknesses of the study are as follows: 1) some attempt should be made to link the data in Phase I of the Reunion Survey to data that will be generated from the DoD exposure databases in order to assess possible exposure-response relationships; and 2) the 84-unit study population is not a random sample, nor is it necessarily representative of the troops that served in the PGW. Therefore, inferences drawn from this survey cannot be extended with confidence to the total study population of PGW veterans.
To provide data on changes associated with readjustment during the first year of return from the PGW, a Phase II survey was planned for a year after
Phase I. The Phase II survey will provide critical data on the course and possible persistence of the patterns noted in Phase I. Similarly, the follow-up data will elucidate conditions that may have delayed symptom onset. Phase II also will offer data on the longitudinal course of readjustment, and on important subsets of the veteran population. Some findings from Phase II have been completed (Wolfe et al., 1993). Phase III involves an evaluation of cognitive function in a subset (150 to 200) selected from the larger Fort Devens cohort. These people are being evaluated at the Boston VA Medical Center. The study is underway.
In the other studies in which first year findings were reported (VA, 1992b), the prevalence of PTSD is between 4 percent and 16 percent of returnees, depending on demographics and war-zone exposure; women appear to be more prone to PTSD than do men. Studies document the major stressors (including less obvious ones like ''waiting around"). These are good studies of a focused endpoint over a short time period. They document well the possible problems with the data and the analysis (for example, nonresponse and selected samples), apply validated psychometric scales, and make an attempt to include a broad group within each of their "catchment" pools.
The Department of Military Psychiatry, Walter Reed Army Institute of Research (WRAIR), Washington, D.C. provided the IOM committee with extensive information. Concern about the psychological stress that would be created by the high-threat deployment to a harsh environment prompted senior Army leaders to learn more about the stress of the deployment. A research team was sent to SWA to study the deployment from September 20, 1990, to October 6, 1990. More than 500 deployed soldiers ranging from privates to lieutenant generals took part in semistructured interviews either as individuals or in small groups of those of similar rank without their supervisors present. Units that had been in SWA longest, were most forward deployed, lived under the most austere conditions, or had missions judged particularly stressful by higher headquarters were studied. At the time of the interviews, small-unit cohesion was high and morale was generally good. Problems in morale or cohesion generally were traceable to factors existing before the deployment, such as deficiencies in trust or communication up and down the chain of command (Marlowe et al., 1990).
This was a broadly focused program designed to evaluate stress and adaptation in ODS/S and to study it both in the context of community and as factors particular to individuals. Assessments of stressors, adaptation, cohesion, morale, and several indicators of mental health were completed before the air and ground war began in SWA, during combat, and following redeployment. These assessments demonstrate a wide range of military-related and nonmilitary-related (family) stressors, and differences between the wartime experiences of active duty service members and reservists. Peak stresses occurred
prior to the ground war, with acute stresses of combat being different in nature. Compared with statistics from Vietnam-era veterans, the estimated risk of PTSD was low, about 3 to 4 percent. Several factors, military and nonmilitary, are suggested as mediating the effects of combat stress on outcome, including immediate debriefing, good leadership in the field, and family contact.
An exposure of concern to the troops in SWA was the bite of the sandfly, which may have carried the parasite that causes leishmaniasis. Usually, visceral leishmaniasis is caused by L. donovani, L. chagasi or L. infantum, but exposures during Desert Storm were to L. tropica, making the diagnosis more difficult. The committee has considered two aspects of exposure to L. tropica and resulting infection with leishmania: the occurrence of either cutaneous or visceral leishmaniasis; and the possibility that some component of unexplained illness may result from leishmania infection.
One investigator has estimated that 5 to 6 percent of Gulf War veterans are infected with leishmania, which can lead to chronic illness (Magill, 1994). Leishmania have incubation periods from months to years. The organism may persist in the host after treatment that appears to be effective; thus relapse is not uncommon. Illness results from a cellular immune response that produces cytokine mediated symptoms that are compatible with some of the symptoms reported as part of the GWS; these also have been produced by giving interferon gamma to healthy persons. Thus, some observers have concluded that GWS could result from visceral leishmaniasis among soldiers who participated in Operation Desert Storm. Because the presentation of illness associated with L. tropica infection does not fit the classical presentation of visceral leishmaniasis, it is referred to as viscerotropic leishmaniasis. Two issues raised by the eight known, documented cases of viscerotropic leishmaniasis in returning Desert Storm troops are the late presentation due to prolonged incubation and the activation of latent infection in immunosupressed persons. Latent leishmania can be a problem in persons with HIV, after organ transplantation, and during administration of high doses of corticosteroids. If L. tropica can survive in a latent state, it will need to be included in differential diagnoses of otherwise unexplained illnesses in returning veterans from Operation Desert Storm (Cotton, 1992; Kreutzer et al., 1993; Lesho, 1991; Magill et al., 1993; MMWR, 1992; Norton et al., 1992; Ohl et al., 1993; Rashti et al., 1992).
HEALTH-OUTCOME-BASED ACTIVITIES-NEW OR JUST BEGINNING
Issues of Reproductive Health
Illness and toxic agents can affect reproductive health in many ways (Mattison, 1994). One outcome of great concern is birth defects. Two types of approaches could be used to determine whether birth defects are related to parental service in the Persian Gulf. The first is to identify and study all putative exposures likely to produce transmissible genetic defects or adverse health effects in a fetus or newborn. Fathers' exposures could be studied in relation to the genetic defects (sperm-mediated); and mothers' exposures could be related to either genetic or developmental defects (mediated by the ovum or through in utero exposures). The second is to compare all children with some adverse health outcome with a suitable set of controls to establish the location and status of each parent before, at the time of, and after conception, whether or not the defect was genetic or consistent with known toxicity of the potential exposure.
Reproductive health outcomes linked to Persian Gulf War service have been noted prominently in the news media, but have received only limited scientific attention. To date, there are two evaluations that are complete or nearly complete: The Mississippi Cluster Study and a review of Army data by the Army Surgeon General's Office (PGVCB, 1994).
The Mississippi cluster study involved evaluating 54 birth outcomes of the Army Reserve Unit of Waynesboro, Mississippi. The final report should be available by December 1994. The preliminary conclusions of the report are that the birth defect rates were not elevated compared with those in the Metropolitan Atlanta Congenital Defects Program (Mississippi State Department of Health, 1994).
In the Army Surgeon General's evaluation of major and minor birth defects of active duty personnel at all Army hospitals from 1985 to 1993 (n = 346,322), rates of major and minor defects ranged from 5.8 percent to 9.6 percent with 8 percent to 17 percent expected from comparison with the Metropolitan Atlanta Congenital Defects Program. Rates for 1992 and 1993 (following the Gulf War) were similar. Only births in Army hospitals to active duty women or spouses of active duty service members were included. These data reflect birth defects recognized at the time of discharge after birth (after 2 to 3 days of age). Thus, the data would underascertain minor malformations and conditions detected later in infancy.
Neither of these two studies has revealed an incidence of birth defects that differs from expected rates. General Accounting Office (GAO) review (GAO, 1994) of the Mississippi cluster study has questioned whether the Atlanta birth
defects database is an appropriate comparison group, and whether rates of birth defects would be reduced if the Mississippi veterans were selectively healthier than the Atlanta database parents. They also recommended that the predeployment incidence of birth defects be assessed and used as a comparison.
Likewise, reviewers of the DoD study have noted that births to active duty military personnel occur outside DoD hospitals, and that infertility and miscarriages constitute important outcomes of reproductive toxicant exposures that are not included in the DoD database.
A new initiative, the DoD Reproductive Health Study, contains data from hospital records of active duty personnel and spouses for whom the discharge diagnoses include fetal death, premature birth, serious birth defect, or neonatal death (CDR Gregory Gray, 1994, personal communication). Records from the DoD hospitalizations of 350,000 Persian Gulf veterans and 700,000 personnel not deployed in that era will be examined. A recent update (September 7, 1994) from the project director indicates that a survey will be conducted of 20,000 couples to assess miscarriage and infertility frequency. This reproductive health study was not yet begun at the writing of this report (September 1994).
Surveillance of reproductive health is included in the DoD and VA clinical examinations. Information obtained by the VA Health Registry examination includes data about birth defects among children conceived before and after Gulf War deployment. Miscarriage and infertility data were added after the first 20,000 surveys were completed; it is uncertain whether the VA will go back to obtain infertility and miscarriage data from the first 20,000 participants. Information obtained by the DoD includes data similar to that in the VA Health Registry exam. Despite these efforts, the ability to link exposures to reproductive health outcomes will be limited for each of these databases. At this time, high-quality population-based data on reproductive health have not been generated.
VA Mortality Study Plans
The VA has indicated plans to retrospectively compare the causes of mortality of all Persian Gulf War veterans to that of Persian Gulf War-era veterans who did not serve in the Persian Gulf theater. The study population of approximately 690,000 will be defined by service in the Persian Gulf theater of operations at any time between August 1990 and April 1991. The Persian Gulf-era veterans will be a sample of military personnel who served at any time between September 1990 and May 1991, but did not serve in the Persian Gulf theater. The controls will be frequency-matched to the Persian Gulf veterans by branch of service and unit status.
Mortality will be ascertained using the VA Beneficiary Identification and Records Locator Subsystem (BIRLS), Social Security Administration records, and the National Death Index (NDI). The follow-up period will begin for Persian Gulf veterans on the day they left the Persian Gulf area alive, or May 1, 1991, whichever is earlier. Controls will be followed beginning May 1, 1991. Follow-up for mortality ascertainment will end either on the date of death, or September 31, 1993, whichever came first.
This study will be useful in determining whether there are any unusual acute mortality causes in veterans of the Persian Gulf War, as compared with veterans of the same era who did not serve in the Persian Gulf; however, insufficient time has passed to observe any excess mortality from chronic diseases.
VA Health Survey Study Plans
The VA has indicated plans to conduct a health survey to estimate the prevalence of various symptoms and other health outcomes among Persian Gulf veterans and Persian Gulf-era veterans and their spouses and children in relation to certain environmental exposures. The study will include a population-based sample of 15,000 Persian Gulf veterans and an equal number of Persian Gulf-era veterans. This will be a mail survey, supplemented by telephone interviews and physical examination (Health Registry Exam) of a sample of veterans. The study is proposed to be validated through military and VA records.
We hope that the investigators will employ sophisticated statistical sampling techniques to ensure that the study and control populations are properly selected, and that procedures are used to ensure the representativeness and precision of the estimates to be obtained. There are several concerns about this proposed study. Mailed questionnaires are known to have very poor response rates; therefore, the investigators should try to obtain an adequate response rate that is nondifferential in the two groups. Otherwise, there is a high probability that such a study will produce neither reliable estimates of symptoms and other health outcomes, nor estimates that represent the actual experience of either group of veterans. The potential for response bias is great, with PG veterans having problems being more likely to respond than those who do not. Additionally, a tally of symptoms and illnesses will get no closer to understanding whether any of these problems occurred because of some experience or exposure related to PG service.
At the initiation of Senator Tom Harkin of Iowa, the Centers for Disease Control and Prevention (CDC) will conduct a study to assess the prevalence of self-reported adverse health outcomes among Iowa residents deployed to the Persian Gulf during ODS and deployed elsewhere. Of the 20,000 Iowa residents who served in the Gulf War, 3,500 were reservists or National Guard members and 16,800 were on active duty. Current plans call for the study sample to include approximately 2,000 Persian Gulf veterans and PG-era veterans, both men and women, and will include active duty, reserve, and national guard personnel who listed Iowa as their home of residence. The research group plans to use telephone survey methods to assess factors that can be classified as preservice (health status, psychosocial functioning, trauma exposure), service (perceived stress, combat experiences, actual duties, perceived exposures, prior military experience) and post-service (health status, psychosocial functioning, PTSD, depression, chronic fatigue). A more detailed study proposal was not available at this writing (September 1994).
What specific gaps in knowledge will be addressed by this study is unclear, as is whether there is any scientific basis for a study restricted to the veterans of Iowa or any other single state. Whatever other merits, this study seems unlikely to produce useful new medical or scientific information about the health effects of service in the PG. This is because of the relatively small number of veterans included, limited geographical distribution, possible biases in geographic dispersion or in the initial characteristics of veterans from Iowa, difficulty in obtaining information on deceased veterans identified for the study population, and lack of organized expert attention to study design. A single state is unlikely to represent the national experience. There is no reason to believe that Iowa veterans have more problems than others or merit more attention. State-by-state reviews are neither productive nor informative.
DoD/VA Studies of Morbidity among Gulf War Veterans
Researchers at the Naval Health Research Center, San Diego, California have developed a protocol with three primary components: 1) a survey of symptoms among 1,500 Seabee veterans who participated in the PGW and 750 who did not participate; 2) a comparison of hospitalizations among 350,000 PG veterans (Navy, Marine, and Army), with those of 700,000 persons not deployed; and 3) a review of adverse pregnancy outcomes among both PG veterans and persons not in the war, as reported at discharge from Army hospitals (Gray, 1994).
This is a well-funded enterprise of nearly three million dollars over a five-year project period. The investigators have described their general methods,
but the details of the protocols are still being revised. As of this writing, the IOM committee has not received or reviewed a final protocol for these studies.
The studies address important questions and proceed logically. There will be an analysis of self-reported symptoms, hospitalizations, and pregnancies, which will make good use of available data. Incomplete follow-up is a threat to data quality, especially when it is related to an outcome of interest. Thus, there should be a tracking system to follow participants who leave military service.
Key concerns are the baseline comparability of persons who did and did not serve in the PGW. The investigators propose to look at this, and careful and detailed assessment is central to the enterprise. Any notable discrepancies will need to be addressed by stratification and adjustment, and all of these will need a sensitivity analysis. This concern is no different from that in all observational studies, but data on some relevant variables might not be available to the researchers, and some variables might not even be known to be important.
Data on an individual's exposures related to PGW service are an important aspect of the study, and it remains to be seen how well exposures can be estimated from available data. In fact, the major strength of these studies may lie in uncovering a dose-response (exposure-response) relationship among PG veterans in relation to specific exposures, rather than in comparing those who did serve against those who did not. Also, persons who did not go to the Persian Gulf may not be a valid control group, even after adjustment for a wide variety of preservice imbalances.
The studies are complex scientifically and organizationally. Follow-up of selected samples can be very informative, provided the follow-up is accurate, detailed, and sufficiently complete. It is not clear that the project has the infrastructure or staff to ensure success. Additionally, attention needs to be given to the issue of career military service versus reserve status, particularly because the current indications concerning GWS focus primarily among the reservists who are not included in this set of studies. The proposed five-year duration of this study seems too short to detect cancers, though it may be longer than is necessary for observing more acute effects. In all of the studies, primary attention should be paid to collecting reliable data, and documenting problems with the data sets, which should be communicated to other researchers in a timely manner. Some details of the individually proposed studies follow:
This study will assess health conditions, environmental exposures, and potential confounders among 2,250 active Seabees (two battalions of a total of 1,500 PG veterans and one battalion of 750 who were not deployed to the
Gulf). Data are to come primarily from an interview that is designed to identify clinical, occupational, geographical, temporal, and demographic risk factors. Data from computerized hospital records, unit diaries, and PG veterans registries will be available. Limited physical examination will be done. Whole blood and serum samples will be stored for future analysis; urine samples will be collected and stored; and a lung function test will be done.
Missing information is likely to cause an assessment problem in this portion of the study. If only active duty Seabees are included in the study, bias is likely. If not, there must be prompt and detailed decisions about how discharged persons are to be found. An added concern is whether the Seabees will provide accurate information. With the current military reduction in forces, there may be a tendency to under-report problems to protect their jobs. This should not, however, affect the assessment of physical signs or laboratory findings unless persons with problems have a low participation rate as well as biased reporting of symptoms.
A computerized hospitalization discharge database will be used to study 350,000 PGW personnel who are still in active service. With respect to hospitalizations after July 1991. these persons will be compared with a randomly selected cohort of 700,000 soldiers not deployed to the Persian Gulf The investigators will assess pre-PGW health status, and will try to track persons who leave the military, and capture information on their hospitalizations.
Again, missing information is likely to cause an assessment problem in this portion of the study. Hospitalizations of only persons still on active duty are likely to be a biased identification of the sick population. Also there is the concern that methods have not been determined to track persons discharged from the service. The possible differential tendency of PGW veterans and those not in the war to seek care outside the military system is a further threat to the study. Some of these concerns are addressed by the addition of the California nonfederal hospitalization component that will obtain information from nonmilitary hospitals.
The hospital database should be quite accurate and complete regarding the fact of hospitalization, but data on specific diagnoses, symptoms, and signs may be much less accurate and complete. There is a good data system for evaluating the hospitalization of active duty service personnel; however, these data systems do not track non-DoD hospitalizations and outpatient visits.
Pregnancy Outcome Study
The women studied for pregnancy outcomes (fetal death, premature birth, serious birth defects, and neonatal deaths) will be a subset of the hospitalization study. This study has been mentioned above in the section entitled ''Reproductive Health."
There are likely to be problems in the assessment of pregnancy outcomes, particularly specific birth defects, based on routinely available hospital records. Similarly, gestational age estimates will be subject to error. Nevertheless, given the scarcity of informative data on reproductive health outcomes among PG veterans, either completed or in progress, analyses of available data are justified.
ASSESSMENTS OF EXPOSURE
Only limited information is available to study possible links between environmental exposures and the unusual medical conditions that have been reported to be associated with service in the Persian Gulf War. While environmental exposures may have been important, the data needed for sound epidemiological study are very limited. Of data available, the most reliable findings seem to refer to oil well fires and lead. However, the information is skimpy for several reasons. First, there was minimal systematic sampling and measuring of possible environmental hazards; even expert descriptions of environments rarely are available. Also, the wide range of health outcomes that some persons have attributed to service in the Persian Gulf have yet to be grouped into useful and biologically plausible categories. Therefore, it is premature to narrow the consideration of exposures to specific agents or different types of environments. In addition, the vast majority of the health investigations that the committee reviewed were neither designed around one or several precise hypotheses nor based on well-characterized "populations at risk." Therefore, the consideration of types of exposures potentially important with respect to the reported problems can be presented only in a general way.
The environment that veterans encountered in the Persian Gulf is usefully considered in two categories: general and occupational. General environmental exposures of interest include temperature and humidity, sand and dirt, sanitary conditions, fauna, oil well spills and fires, pesticides, petroleum products (especially jet fuel and diesel fuel containing lead additives), mycotoxins, and decontamination solution. Exposures to consider under occupational environment are related mainly to general maintenance operations and include
battery repair, cleaning/degreasing, electronic/radio repair, generator repair, grinding/sanding, sand blasting, lathing/milling, painting (especially with isocyanate-based paints), refrigeration servicing, vehicle repair, weapons repair, and welding and cutting.
An adequate description of the general environmental conditions is still being developed. The following is a summary of the information available at the time of writing (September 1994).
Temperature and Humidity.
Mean daily low and high temperatures were 80°F and 108°F in July, and 45°F and 65°F in the winter. Except in coastal regions, the relative humidity was less than 40 percent during the summer, but over 60 percent during the rainy season (December through March). Solar heat was intense in the summer. Rain was minimal (3 to 8 inches) but sometimes caused the flooding of tents (DSB, 1994).
Sand and Dirt.
Most troops were located in desert settings where sand was ubiquitous. The sand often was powdery, and persons with respiratory conditions sometimes reported respiratory symptoms. Whether these symptoms were properly attributed to sand rather than to the type of living structure (tent versus air-conditioned building), or other problem has not been determined (DSB, 1994).
Sanitary conditions in the Gulf theater need a systematic review that is not yet possible from available records. However, the following has been reported as a summary of the sanitary conditions.
Staging areas near ports of entry were characterized by crowded tent living with strains on latrine facilities, showers, and feeding. The prototype four-scat latrines were mass produced by contract in country. Latrines were designed for suction removal of waste by contractors or bum-out. Early designs permitted ingress of flies. In the desert environment, daily burning out of waste cans employed mainly diesel fuel. Smoke from such fires was common, though rapidly dispersed by prevailing winds. Solid waste disposal was handled by contract in the staging areas and by burning in pits in the desert. Locations for burn pits and latrines were usually chosen carefully to minimize nuisance from smoke, smells, and flies. When shower/bath units were not available, many field expedients were improvised. Likewise for laundry. As shower setups became available or were improvised, an unforeseen problem was heating the water once the cooler weather set in. Time in-
tervals between showers and uniform laundering were sometimes lengthy (O'Donnell, 1994).
Food and Water.
The final report of the United States Army Environmental Hygiene Agency (USAEHA) does not contain a discussion of the possibility that troops were exposed to contaminated water or food. This is because local drinking water was not used. The troops and civilians were provided with sealed containers of drinking water; food was provided in the form of meals ready-to-eat (MREs) sealed in plastic, or provided at the mess hall. Personnel generally did not eat food produced from the local areas. This helped prevent diarrheal illness. In all previous military engagements, diarrheal disease was a major problem; therefore, incidence was surprisingly lower in PGW. In addition, ''there was no reason to believe that the local food was contaminated with oil fire residue" (USAEHA, 1994a).
Exposures to fauna also need a systemic review that is not possible at this time. The following has been reported as a summary of certain exposures.
Filth flies were a universal problem in the warm months. Latrines and food sources were attractants. The use of screening, self-closing doors, fly traps, fly bait, and pesticides were moderately successful suppressants. Various types of scorpions and snakes such as the horned viper were common in the desert, and envenomation of personnel occurred occasionally. Although biting spiders were present, they were not a problem. Mosquitoes were a factor only in the Euphrates Valley. Sandflies were present, as evidenced by the cases of leishmaniasis, but were difficult to find even when searched for. Sheep and camels were commonly observed in the desert. Dead sheep were often reported, but veterinary inquiries disclosed no signs of unexpected causes of death. Unit pets or mascots were officially banned but some small units adopted stray dogs and obtained veterinary care through Saudi sources (O'Donnell, 1994).
The report that no signs of unexpected causes of death among sheep or camels were noted is not consistent with a report the committee received in the public meeting, where one veteran described his investigation of an episode of widespread unexplained sheep deaths (IOM public meeting, March 1, 1994).
DoD apparently has or could produce a listing of all pesticides shipped to the Persian Gulf, but the committee knows of no central record of their distribution and use. The following are the pesticides reported as likely
to have been used at some time in some locations: allethrin/resmethrin, azamethiphos, bendiocarb, chlorphyrifos, diazinon, malathion, d-phenothrin, permethrin, and pyrethrin. There are no reports of the use of herbicides and no reports of acute pesticide poisonings. Summary information indicates that pesticides were generally applied either using one-or two-gallon sprayers or directly by hand. There was also an indication that marines did some "ultra low volume fogging" (PGVCB, 1994). The virtual absence of reports of sandfly fever suggests that vector control was in general good (Richards et al., 1991; Richards et al., 1993).
Many different fuels were used in the Persian Gulf to power vehicles and for heaters, cooking stoves, and portable generators. The exhaust produced in the course of using these fuels (particularly the use of diesel heaters in living quarters) could have caused a variety of exposures to combustion products including lead. Even though there are several accounts of exposure to these sources, the committee has not received information either to directly evaluate the likelihood that such sources resulted in important exposure to the troops or to determine levels of exposures (either substance-specific or through measurements of surrogates or indicators for the exhaust gases). It appears that approval of the use of diesel fuel for firing stoves was based on work undertaken at the Natick Research and Development Center before the Gulf War (Riley, 1992).
The committee found only one reference to the decontamination solution for the treatment of equipment exposed to chemical warfare agents, which stated the following:
Decontamination Solution 2 (DS-2) is a liquid mixture to decontaminate equipment that has been exposed to chemical warfare agents. The constituents are propylene glycol monomethyl ether or ethylene glycol, sodium hydroxide, and diethylenetriamine. DS-2 does not seem to have produced widespread problems in the Persian Gulf region and chronic effects after a single exposure are still conjectural (PGVCB, 1994).
Oil Well Spills and Fires
Several efforts have been made to determine whether the oil well spills and fires created by the retreating Iraqi forces caused any health effects among U.S. troops. These effects have been summarized by USAEHA (USAEHA, 1992; USAEHA, 1994a). They include Kuwait and Saudi environmental assessments; reports from EPA (EPA, 1991), France, and Great Britain that were
beginning or completed in March and April of 1991; reports from Japan and Harvard University beginning or completed in April; reports from the National Institute of Standards and Technology, the National Center for Atmospheric Research, Norway, and Germany beginning or completed in May; reports beginning or completed in June from the National Oceanographic and Atmospheric Administration (NOAA)/Air Resources Lab; reports beginning or completed in July from EPA/National Aeronautics and Space Administration (NASA) and Lawrence Livermore Labs; and reports from Germany in October. The Defense Nuclear Agency carried out some ground-based measurements during an unknown period. The interim USAEHA report provides very limited information about these studies and none about the deoxyribonucleic acid (DNA) study. The final USAEHA report suggests that further examination of these reports, along with the published literature on the fires, will be done in conjunction with the Geographical Information System (GIS) project, which is still in process.
The major effort at environmental assessment was undertaken by USAEHA to evaluate troop exposures to the oil well fires and oil spills. While exposures began when the first oil well fires were ignited in February 1991 by the Iraq armed forces during their retreat and lasted until November 1, 1991, USAEHA's air sampling effort could not be successfully launched until the beginning of May. Therefore, the more stagnant air conditions of the winter months were missed. Those who undertook the sampling efforts did so with this knowledge and intended to address the problem as well as possible by the use of meteorological modeling. A GIS is being developed to integrate information on airborne and soil-based exposures, on meteorological conditions throughout the study interval, and on individual troop movements during Operation Desert Storm. Results from a pilot project in developing the GIS indicate that the approach should prove useful (USAEHA, 1994a). Once this model is available, individual exposure estimates can be made for troops throughout the region, though further work will still be needed to validate the model and estimate its precision.
At the maximum there were ten fixed air-sampling stations, but two of them operated for less than two weeks and two more operated for two months. Only three were in operation through the end of December.
These fixed sites were located where troops were concentrated, and soil was sampled from the same areas. The results are to be used with NOAA-assisted modeling to estimate reasonable maximum individual exposures (RMEs) to the chemical substances sampled according to records of troop movements using the GIS. These included air and soil pathway analysis and industrial hygiene sampling. Air and soil quality was estimated not to have deteriorated during the sampling interval, and a reference to earlier sampling suggests that quality at some sites was even higher than before the war. Soil
metals did not increase during sampling except for metals unrelated to Kuwaiti crude oil.
The air pollutants expected were classified into four categories: reactants (crude oil components not combusted); combustion products (such as carbon dioxide and water); incomplete combustion products (for example, carbon monoxide), and products of secondary reactions (from photolysis). The substances included short-chain and low-and medium-molecular weight aliphatics such as butane and heptane (both straight and branch chain in the range C2 to C40); single and polycyclic aromatic hydrocarbons (PAHs); heterocyclic compounds including benzene, naphthalene, and xylene; and substituted compounds such as methylated and halogenated compounds. Samples were assayed for suspended particulates, both total and particulate matter less than 10 microns, a series of volatile organics, PAHs, and metals. A subsample was examined for sulfur dioxide, nitrogen dioxide, coal-tar pitch volatiles, and acid aerosols. These agents were chosen as likely to provide a reasonable estimate of the impact of the oil well fires and spills. The sampling results were designed for use in estimates of cancer risk and subchronic, noncancer risk.
As yet, none of the agents sampled or detected are obvious sources of problems that would cause persistent symptoms for months or years after return from the Persian Gulf. The modeling may, however, offer some improved understanding of the differences in the general environment of troops located in different parts of the war zone.
The most striking outcome of the USAEHA sampling and the report of the EPA sampling is that there was very little documented deterioration of general air quality during the period of the burning oil wells and oil spills. The only substantial increase noted was in particulates. The concentration of particulates, while high, was considered "normal" for this area of the Middle East. Exposures to organic compounds were similar to levels observed in Houston and Philadelphia, cities with major petrochemical industries. There were relatively high concentrations of naturally occurring metals, apparently resulting from wind-blown surface soils. The levels of air pollution were high enough to raise concerns about ingestion and dermal absorption, and these routes of exposure have been considered (USAEHA, 1994a). However, no measurements were taken, so that the possible effects of exposure through these additional routes can be estimated only by mathematical models.
When individual exposures can be estimated, the effects of different levels of exposures to pollution from the oil fires can be estimated by using troop location during the largest part of the time in the Persian Gulf region. For example, the prevailing winds (from northwest in Kuwait and from north-northwest in Saudi Arabia) should have caused substantial differences in potential exposure of troops, because some troops were both close and downwind; others were either distant or to the side of the prevailing windflow.
Further work by USAEHA is expected to provide a model of the exposure distributions and incorporate information from earlier, more limited sampling that might improve the estimates of exposures to troop encampments. The work by USAEHA also will examine the frequency and duration of exposures. When this information is available, the time course of symptoms in different troops could usefully be studied.
In summary, these assessments of exposures to products of the torched and destroyed oil wells will apparently include information on relatively low-level exposures to a variety of petroleum-related chemical substances. It will then be possible—and important—to consider more specific questions about symptoms and health outcomes, including any effects of troop location and troop movement, and the time-courses of each of these. Because accurate information on exposure of individual persons is not available, epidemiological assessment of differences in illness or symptoms reported by different troops in various locations holds the greatest promise for indicating health problems related to chemical exposures.
"Small groups" from the 11th Armored Cavalry Regiment were surveyed in June, August, and October of 1991 before, during, and after 90-day deployment from Germany to Kuwait (USAEHA, 1994b). Tests included questionnaires on symptoms, moods, exposures, and work; a pulmonary function test; blood samples for metals, volatile compounds, and sister chromatid exchanges (SCEs); and urinary measures for metals and benzo-a-pyrene (BaP). Approximately 2,800 of 4,700 troops volunteered to complete a questionnaire in predeployment, 350 during service in the Persian Gulf, and 1,700 post-deployment. There is no evidence of a change in pulmonary function, although the predeployment values did not meet technical standards for usability. Serum and urine tests for metals were carried out in 150 to 175 persons in the three periods; for volatile organic chemicals (VOCs) in 28 to 32 persons; for SCEs in 3,661 and for PAH-DNA in 34 to 42. Some increase in symptoms was reported in the Persian Gulf. Metal analysis was remarkable only for a slight and nonsignificant increase in the average level of blood lead. Elevations were noted in a minority of the troops, suggesting that a local rather than a general explanation should be sought. No ready explanation for this increase has been provided, though lead exposures from the use of diesel fuels for tent heating were apparently not considered. Nickel (Ni) and vanadium (V) (present in Kuwaiti crude oil) were not elevated. VOCs have a very short half-life in the body, and therefore their measurements relate to very recent exposure. The VOCs were, in general, lower during the time in the Persian Gulf than before or after, with the exception of a small, unexplained elevation in
perchlorethylene. Since perchlorethylene is a chlorinated hydrocarbon it was thought that blood level increases were unlikely to be caused by exposure to oil well fires. Acetone levels seemed to have been slightly elevated, but still need to be examined for normalcy. SCE levels were elevated during time in the Persian Gulf; PAH-DNA adducts were lower during service there than either pre-or post-deployment. Even though the elevated SCE levels were evidence of some genetic stress, no source of such stress has been identified. The reduced levels of PAH-DNA were attributed to unusually low exposures during service in the Persian Gulf Analysis of urinary tetrols and overall data analysis is continuing.
Some battlefield casualties were further examined at autopsy for possible connections between exposures and mortality or morbidity. The Armed Forces Institute of Pathology examined 85 corpses of troops who died in the Gulf War theater. Some were autopsied, some had only a study of body fluids, and some had both studies. These included deaths both before and after smoke exposure.
Blood was examined for contaminants, mostly heavy metals: V, Ni, chromium (Cr), lead (Pb), copper (Cu), and zinc (Zn). Autopsy findings did not suggest acute pulmonary, liver, or renal failure. Except for expected pathology nothing unusual was seen. Because the smoke from the oil fires contained V and Ni, a finding of excess V or Ni would suggest exposure. No such excess was found. However, in seven cases, a high blood Pb level was seen:
These tests (*) were thought to be contaminated because of the high value. The author indicated that the normal range was 7 to 19, with an average of 12.5 (AFIP, 1993). Although isotope studies of the samples might help, it is impossible to determine whether there was contamination. However, these reported high measurements are compatible with acute lead poisoning, such as might occur from burning leaded diesel fuel in heaters in unvented sleeping
tents. These high values might indicate substantial exposures in the staging area. Also, the elevated blood lead levels might indicate higher exposures to a variety of petroleum combustion and evaporation contaminates. If surviving troops suffered from acute lead poisoning, blood lead levels today would not be abnormal, but X-ray florescence might show that blood and bone levels are still increased. Troops might still be asked about transient gastrointestinal pain immediately after exposure. Despite the purported negative findings, the paucity of the above data, and its non-scientific format would still not permit any conclusions about toxicity.
Little summary information has been identified that characterizes the range of occupational exposures that may have occurred in the Persian Gulf beyond those that might be associated with the occupation of "soldier." It appears that the majority of possible occupational exposures would be related mainly to repair and maintenance activities to keep equipment in functional order. No information is currently available on the numbers of troops who were assigned regularly or intermittently to the different components of support work in the Persian Gulf.
Because the operating conditions were far from ideal, and the environments generally were not sufficiently fixed for adherence to recommended occupational hygiene controls, some of these exposures could have been substantial. Trained industrial hygienists in the Gulf reported that exposures in fixed work environments generally could have been kept within current standards there, but that operations in the field may not have allowed sufficiently stringent controls (Riley, 1992). This was particularly the case for painting operations that called for laminar flow hoods. Personal protective equipment was a key feature of adequate exposure control in painting operations. The industrial hygiene staff used the American Conference of Governmental Industrial Hygienists' (ACGIH) Threshold Limit Values (TLVs) as their primary guide. Commentary indicates that these limits were considered sufficient to protect against any health effects (USAEHA, 1992). However, Rappaport (1993) has shown that many of these limits are based on feasibility rather than on protecting against all health effects.
The following listing includes most of the potential chemical hazards associated with common maintenance or repair operations. With the exception of vehicle painting, information on actual exposure evaluations for these work settings is not available. Therefore, the operations that are likely to have occurred are characterized primarily by noting the potential chemical exposure risks.
Battery Repair: Corrosive liquids, particularly sulfuric acid, lead.
Cleaning/Degreasing: Degreasing solvents generally including a range of chlorinated hydrocarbons such as trichloroethylene.
Electronic/Radio Repair: Soldering fumes and cleaning solvents.
Generator Repair: Carbon monoxide.
Grinding/Sanding: Abrasive particulate.
Sand blasting: Abrasive (possibly crystalline silica) particulates in the respirable range.
Lathing/Milling: Metal working fluids.
Vehicle Painting: Paint solvent vapors and mists. During primary and repair painting operations there was a potential for overexposure to isocyanate paints, which can cause primary sensitization and asthma as well as exacerbation of existing asthma. The chemical agent resistant coating (CARC) paint system was directly used by an estimated 1,000 troops (PGVCB, 1994). The number of vehicles painted each day varied between 10 and 100 at the major work site in Al Jabayl. At least one episode of overexposure was reported (December, 1990) (Riley, 1992).
Refrigeration servicing: Lead fumes and exposure to refrigerant such as fluorocarbons.
Vehicle Repair: Asbestos from brake repair, carbon monoxide, organic solvents.
Weapons Repair: Lead particulate.
Welding/Cutting: Chromates, ozone, nitrogen dioxides, heated metal fumes.
USAEHA industrial hygiene sampling showed increases in personal or general air monitoring results of outdoor occupational environment from selected locations, but these increased levels still did not exceed recommended standards.
Threat of Chemical and Biological Warfare Agents
Direct and indirect statements to this committee by several Gulf War veterans and groups speaking on their behalf have ranged from worried concern to firm statements that troops were exposed to unspecified war agents, chemical, biological, or both. The strong suspicions seem to be that exposures were to nerve agents stockpiled by the Iraqis for use against Iran and the Kurdish population. Such stockpiling and use would have preceded Operation Desert Storm. Some claims have also been made that U.S. CDC Type-Cultures (anthrax and botulism) supplied by the United States to Iraq in the 1980's for research and vaccine production could have been fashioned into weapons and
used, overtly or covertly, against U.S. forces. Testimony before several Congressional committees has carried into the press strong assertions that members of Congress or their staffs actually have concluded that self-reported signs and symptoms said to be associated with soviet-designed surface-to-surface missiles (SCUD) attacks and concomitant chemical and biological warfare (CBW) alarms were sufficiently convincing to make a prime facie case for CBW use. These assertions have been amply reported elsewhere in the public domain and will not be repeated here.
The committee could find absolutely no reliable intelligence, and no medical or biological justification for any of these purported claims. Furthermore, analysis of the attacks indicated that the alarms were false positives generated by dust particulates. When analysis of the alarms was followed by more sophisticated tests, the results were confirmed to be negative. This information speaks convincingly against the claims of CBW. Official analyses by the Central Intelligence Agency (CIA) Directorate of Intelligence, the CIA Non-Proliferation Center, the Defense Intelligence Agency, and a high-level all-source DoD committee, the Task Force of the Defense Science Board have concluded, independently and based on all-source/codeword access, that no use by intention or by accident occurred, and that bombing of storage areas both in theater and to the rear also did not cause CBW exposure. These reports are also in the public domain and will not be further referenced here.
One member of the IOM committee obtained classified briefings on known Iraqi capabilities, believed intentions, and prior use of weapons of mass destruction. The committee as a whole obtained briefings on theater operations and CBW suspicions. This material was obtained and reviewed prior to and after disclosure and declassification of most of the same material and the committee found no inconsistencies. This material is now in the public domain and will not be further referenced here.
The committee is not aware of any other actual data, complete or incomplete, that are inconsistent with the above conclusions and is not aware of any reports or studies, classified or unclassified that are inconsistent with the above. Therefore, the committee believes that the illnesses that have been reported are not the result of chemical, biological, or toxin warfare, or accidental exposures to stored weapons or research material. The committee believes further that, in light of this negative evidence from highly placed sources, claims of exposure to chemical or biological warfare agents should not be made or given credence in the absence of reliable data to the contrary.
Vaccines and Prophylactic Treatment
Anthrax and Botulinum Toxoid
The threat of CBW was quite real, as mentioned above, when troops were deployed. Threats of chemical and biological warfare were anticipated. DoD vaccinated some troops against anthrax and botulism, which were considered the biological warfare agents most likely to be encountered. These vaccines were not administered together. Individuals receiving vaccines signed their consent to a list. Prophylactic treatment against possible chemical warfare agents was provided in the form of self-administered pills.
Respiratory-acquired anthrax is a fatal disease unless the diagnosis is made and appropriate antibiotic treatment is begun within hours of exposure. Botulism is a fatal disease caused by potent neurotoxins released by Clostridium botulinum that must be promptly neutralized by antitoxin therapy. There was concern that the Iraqi forces could use one or both of these agents against U.S. forces during ODS. Vaccines against them were made available to selected groups. Several factors have created questions about these vaccines: they were unique in the military immunization program; the botulism toxoid was distributed to the Army as an investigational new drug (IND) by the Food and Drug Administration (FDA) and was to be given on a voluntary basis with informed consent; the anthrax program was involuntary; many persons believed that these vaccines were not properly tested; the program was classified as secret; and each vaccine was intended to prevent serious disease induced by these potential biological weapons. These concerns have raised further questions about a possible role of these vaccines in the long-term health effects reported by some recipients.
Anthrax vaccine contains the known protective antigen of a killed avirulent strain of Bacillus anthracis. In order to enhance its immunogenicity, it is absorbed onto an adjuvant, aluminum hydroxide. The vaccine has been produced by the Michigan State Department of Public Health and licensed since 1972. The primary nonmilitary use has been for persons with occupational risks—those workers who process wool, goat hair, animal hides, and bone and bone products. It has also been administered to laboratory workers. Over 7,900 doses have been given in the past 22 years to employees at the U.S. Army Medical Research Institute of Infectious Diseases (USAMRIID).
The recommended dosage schedule of anthrax vaccine is 6 injections over an 18-month period. Few service members in ODS received more than the first 2 doses, given 2 weeks apart. The program was discontinued when the war was over. It is estimated that about 150,000 service members received at least one dose. The known side-effects of the vaccine include tenderness, erythema, and swelling at the injection site. These have been reported in about
6 percent of recipients. Less that 1 percent have more severe local reactions, which may limit the use of the arm for 1 to 2 days. Systemic reactions rarely have been reported, and no chronic sequelae have been noted. The reaction rates among those receiving the vaccine in ODS were unmeasured. One person was hospitalized because of an infection at the injection site.
Botulinum toxoid vaccine consists of five of the most common types of toxins (A, B, C, D, E) that have been toxoided by use of formalin. It is also manufactured by the Michigan State Department of Public Health. Alum is used as an adjuvant. Botulinum toxoid vaccine has status as an IND with the FDA, as mentioned, and has been used as an investigational vaccine for over 20 years. More than 3,600 doses have been administered at USAMRIID. This experience has permitted an estimate of the reaction rate. There have been more mild local reactions (up to 10 percent) than noted with anthrax vaccine, and about 3 percent experienced mild systemic reactions, such as headache, myalgia, fever, and malaise for 48 to 72 hours. No chronic sequelae have been reported.
The vaccine is to be given in 3 injections: an initial dose, followed by other injections at 2 and 12 weeks, with a booster at 1 year. About 8,000 service members received at least 1 dose. The recipients were primarily members of the U.S. Marine Corps, First Marine Division, and the U.S. Army VII Corps. None of those troops received the anthrax vaccine. All members of these units were to have had the opportunity to volunteer and to give informed consent before receiving the vaccine. No reaction data were collected. A retrospective postcard survey was conducted at Camp Pendleton for the Marines who received the vaccine. Of the 123 cards mailed, 121 responded. Of respondents, 96 percent had received 2 doses; 12 percent reported mild local reactions; 14 percent reported pain that limited use of the arm temporarily; and 2.5 percent reported systemic symptoms that did not limit activity. There was no evidence of long-term effects.
The military has had long experience with vaccine administration. There are 7 vaccines administered during basic training to Army recruits. All active duty Army personnel receive one additional vaccine, as well as boosters to two of the basic training immunizations. Seven other vaccines, excluding anthrax and botulinum toxoids, are given for deployment to high-risk areas, to alert forces, as required by host country, or as directed by the surgeon general. Each of these 15 vaccines can produce local or systemic reactions in recipients. The reactions noted with the anthrax and botulinum vaccines are not unique. The tetanus-diphtheria vaccine is a toxoid vaccine, and is employed worldwide without evidence of inducing chronic medical conditions. The ''antigenic load'' injected into service members is considerable; an even larger volume of antigens was given (up to 75 cc) to employees at Fort Detrick, Maryland. These people were studied over a period of 15 years and no evidence of im-
mune-associated illness was identified. No other chronic disease has been attributed to vaccine use.
The swine flu episode of the mid-seventies caused large numbers of Americans to receive a new antigenic strain of influenza virus. Over 40 million people received the vaccine. Many of these recipients claimed that the vaccine initiated rheumatoid arthritis or exacerbated their arthritis. Others claimed to have developed neurological complications from the vaccine. However, several prospective studies demonstrated that there was no association between the vaccine administration and the claimants' arthritic conditions (Herron et al., 1979; Kurland et al., 1984). Conflicting data have been presented regarding neurological complications, but the best epidemiological data, from military records, show no increase in the incidence of Guillain-Barré syndrome.
Most or all of the claims for both rheumatic diseases and Guillain-Barré syndrome seem to have been a result of coincidences that were wrongly interpreted to be causal. The U.S. Government did establish a fund to settle claims, but this was done in order to avoid legal expenses rather than because of a confirmed association between exposure and complications.
The introduction of other vaccines in the last 30 years has been associated with similar examples of coincidences, such as a concern about neurological disease after use of a measles vaccine. One striking exception to this was the polio vaccine. The attenuated living strains have caused paralytic disease in a few recipients or persons infected from a vaccine recipient. Also, a "killed" polio vaccine that had been incompletely treated caused a form of polio. These are examples of disease that clearly can be attributed to the vaccine. We have no evidence that vaccines in general cause the non-specific complaints associated with service during Operation Desert Storm.
Pyridostigmine bromide (PB) is a drug that belongs to the group of agents classified as anticholinesterases (AChE) that bind reversibly with acetylcholinesterase. This allows the temporary and partial buildup of acetylcholine, which may allow continuous stimulation of cholinergic receptors throughout the central and peripheral nervous system. This pharmacological action has been taken advantage of in the protection of military personnel from the effects of organophosphate toxic gases; these agents bind irreversibly with AChE causing life-threatening complications. PB can compete for binding sites and allow escape of AChE to permit more controlled transmission of nerve impulses. Anti-AChE agents are not new drugs. The first drug in this group, Physostigmine, was isolated in 1864. PB has been used for decades to treat patients with myasthenia gravis. (The dose employed is 360 mg to 6,000 mg
daily.) There is a great deal of pharmacological data, therefore, about the use of this drug.
Some persons have claimed, however, that the dose recommended for use in military personnel in the war against Iraq (30 mg three times a day) had not been adequately tested in persons who have no diseases such as myasthenia gravis. Certain medical conditions can be made worse by anti-AChE agents. Thus, persons with asthma, coronary artery disease, or cardiac dysrhythmias (especially bradycardia) may have a worsening of their condition, and those who are sensitive to the drug may develop anaphylactic shock. These conditions occur promptly after an overdose of the drug; discontinuation of the drug and the administration of atropine lead to rapid recovery in most people. PB is poorly absorbed from the gastrointestinal tract. The bioavailability is 8 percent to 29 percent of ingested dose. It is 70 percent to 90 percent eliminated unchanged in the urine with the half-life of 3.7 hours. The maximum plasma concentration is reached in 1.7 hours. There is minimal penetration of the blood-brain barrier by this drug.
The side effects of PB include those expected from stimulation of the peripheral parasympathetic nervous system, including nausea, vomiting, diarrhea, abdominal cramps, increased peristalsis, increased salivation, bronchial secretions, miosis, and diaphoresis. Skeletal neuromuscular junctions also are stimulated; the effects involve striated muscle groups—muscle cramps, fasciculation, and weakness. The bromide radical can cause skin rashes which subside when the drug is stopped. There have been no documented long-term side effects of the drug. No effects on pregnancy are known, but there is no evidence to suggest any adverse effects. All of the side effects noted above have been reported in PG military personnel taking PB for varying periods, including up to 50 percent in the U.S. Army Medical Research Institute of Chemical Defense (USAMRICD) study, 38 percent in flight crews, and 48 percent of those seen in health care facilities in the theater of operation. Gastrointestinal symptoms were the most common complaint.
The frequency of these complaints was surprising in light of the few complaints among volunteers who took the same dose, but under more controlled experimental conditions. A possible explanation of these effects was the Stress of being in a hot desert environment and under the threats common to battle-grounds. A study of a small number of volunteers in a desert environment failed to show any difference in symptoms while taking PB versus a placebo (Cook et al., 1992). PB does raise body temperature slightly because it decreases blood flow to the skin, thus limiting heat loss by convection. Offsetting this effect is the increased sweating that occurs as a consequence of PB. In the volunteer study, the slight increase in temperature did not affect exercise tolerance in the heated chamber.
Other explanations for a possible increase in symptoms following PB ingestion have been raised by staff of the Senate Committee on Veterans Affairs. First, the reactions of women may differ from those of men, especially if they are on birth control pills; the effects of PB on women have not been studied. Second, the troops were not screened for those conditions that are contraindications to use of PB, including asthma, peptic ulcer disease, liver disease, kidney disease, or hypersensitivity to PB, though predeployment medical examinations should assume that these conditions, other than sensitivity, were present in few of the troops on active duty. A third hypothesis is a possible synergistic reaction between PB, the pesticides, the insect repellent-impregnated clothing used by the troops, or all of these. It has been known for many years that the simultaneous or sequential administration of two anti-AChE drugs would have an additive or even a synergistic effect. The staff of the Senate Committee cited recent studies in cockroaches, in which a combination of the "pesticide" DEET (actually an insect repellent) and PB was reported to be 10 times as toxic as DEET alone. Several pesticides in low doses are not toxic to most humans, but are effective against insects. Robbins and Cherniack (1986) have reported that DEET is only partially absorbed through the skin of humans and is rapidly but not completely excreted. DEET impairs mammalian biochemical pathways (reversibly inhibiting the urea cycle) and can block lactate-dependent synthesis of glucose (Heick et al., 1988; Brini and Tremblay, 1991).
Pesticides such as permethrin (i.e., synthetic pyrethoids) modify the ionic permeability of nerve membranes and produce a neuroexcitatory toxic response (Casida et al., 1983; Vijverberg and van den Berchen, 1990). They have been shown to act in a stereospecific fashion, that is, in a fashion directly dependent upon the structural conformation of the pesticides, on sodium channels in nerve membranes (Eells et al., 1992). Permethrin has also been found to inhibit calcium-dependent ATP-ase enzyme activity in cells from the central nervous system (Kodavanti et al., 1993).
Inhaled pesticides plus PB conceivably could have synergistic effects, but such effects would be immediately obvious, as was true for the combination that killed the cockroaches. The drugs do not persist in the body, and the IOM committee knows of no pharmacological reason why they should have any long-term effect.
Other medications could possibly increase the risk of side effects of PB. Those drugs would include the beta-blockers used for treating hypertension. This combination could cause a further reduction in cardiac output and blood pressure. In rare persons, there could be bronchial constriction. A combination of PB with medications that cause vasodilation, for example calcium channel blockers, or direct-acting vasodilators in circumstances of poor hydration, could lead to lightheaded feelings or syncope. Antimalarial medications
in combination with PB could lead to diarrhea. Quinidine and PB could induce heart block, but the former is not used routinely for malaria prophylaxis or treatment. All these possible drug interactions (and others not mentioned) cause acute and short-lived problems. There is no evidence of any chronic effect.
In summary, PB is a well-studied medication belonging to a class of drugs about which extensive knowledge exists. The DSB Task Force recommended doing a study of the effects of the drug and its metabolites to determine whether they cause long-term effects. The DSB Task Force believed that this was necessary because PB can prevent or ameliorate serious consequences of chemical warfare and must therefore be kept in the military's pharmacopoeia. It seems unlikely that these studies will be revealing, because so much information is available already, and the drug is excreted essentially unchanged.
Daxon (1994) has reported that:
Natural uranium is composed of three radioactive isotopes: U-238 (99.3 percent by weight), U-235 (0.7 percent), and U-234 (0.006 percent). The depleted uranium (DU) used by DoD is a by-product of the uranium enrichment process that increases the concentration of U-235 to levels suitable for nuclear power and nuclear weapons. The uranium that remains is "depleted" in both U-235 and U-234 and typically contains 0.2 percent of U-235 and 0.001 percent of U-234. Because the radioactive decay rate of U-235 and especially of U-234 is so much higher than U-238, this decrease reduces the radioactivity of DU by approximately 50 percent when compared to natural uranium. The specific activity of natural uranium is 0.7 µCi/g while DU is 0.4 µCi/g. DU is used by DoD in antitank munitions and in tank armor because of its high density and pyrophoric properties.
Intact DU munitions and armor radiation exposure is minimal and within accepted standards. A series of studies conducted by the DoD quantified the radiation exposure received by personnel during the transportation, storage, and use of DU-containing systems. Studies have shown that, with the exception of warehouses where large quantities of DU munitions were stored, the estimated annual exposure did not exceed the current standard of 100 mrem/year for the general population. Where this limit might be exceeded, standard radiation protection programs are in place (Daxon, 1994).
DU can enter the body on the battle field as a result of fires involving DU munitions, use of DU munitions, or working in a vehicle or other space con-
taminated with DU. Daxon (1994) continues: "In general, only a small fraction (less than 1 percent) of DU particulates from storage or tanks fires were respirable, and more than 90 percent of the respirable particulates were insoluble. When a DU munition strikes an armored target, the fraction of particulates generated that are respirable ranges from 50 percent to 90 percent. Approximately 17 percent to 48 percent of the respirable particles are soluble in lung fluid."
The potential for inhaled DU particles to exceed current safety standards exists only inside vehicles when they are penetrated by DU munitions. This was shown from estimates of simulated bulk storage fires, test fires involving vehicles with DU munitions, and tests in which DU munitions struck armored vehicles. The Persian Gulf experience showed that DU can reach internal body sites through wound contamination and the injection of DU fragments.
DU inside the body presents both radiological and toxicological risks. Irreparable kidney damage has been shown in studies of injected or inhaled uranium salts. The toxic level for uranium in the kidney for single exposures is 1–3 µg of uranium per gram of kidney (Kearsley and Daxon, 1993). However, there is considerable uncertainty regarding the toxicity of long-term exposure to uranium.
The toxicological hazards of DU inhalation and ingestion are probably more significant than the radiological hazards if the uranium compounds internalized are soluble. For insoluble compounds, the ingestion hazards (both radiological and toxicological) are minimal because the majority of the ingested compounds are rapidly eliminated from the body. The risks associated with the inhalation of insoluble compounds are primarily radiological and are determined by the total radiation exposure to the body. The radiological and toxicological hazards associated with long-term exposure to imbedded fragments are uncertain. There are no known studies of the long-term effects of uranium metal implanted in tissues.
During the Persian Gulf War, few personnel were exposed to DU. A friendly fire incident wounded 35 U.S. soldiers, of whom 22 were suspected to have retained DU fragments (Daxon, 1993). Soldiers (27) from the 144th Supply and Service Company involved in fighting fires, vehicle recovery, and cleaning of the 29 tanks damaged by DU munitions were potentially exposed. There were 32 others that had a chance of exposure. DU rounds were left on the battlefield.
A five-year follow-up policy has been developed for the 22 soldiers suspected of having retained DU particles and for all others suspected of being exposed to DU (Keogh, 1993). The Armed Forces Radiobiology Research Institute (AFRRI) reviewed this policy and found that there were sufficient uncertainties with the potential chronic effects to warrant long-term follow-up of veterans with fragments and to conduct research. The Department of De-
fense and the Department of Veterans Affairs are conducting the recommended patient review and the Department of Defense is initiating AFRRI-recommended research.
Environmental Research Centers
In January 1994 the VA announced a program to establish centers for basic and clinical science studies of environmental hazards. Up to $500,000 per year would be provided for up to five years of support for centers to engage in basic research on environmental health effects, with special emphasis on the diagnosis and treatment of medical problems currently being reported by the PGW veterans. An additional $100,000 will be available for equipment in the first year. The VA is especially interested in new initiatives that complement current activities, and has suggested that interorganizational agreements and scientific affiliations are encouraged if they are justified and set up properly. Links with non-VA researchers are allowed.
As of May 13, 1994, the VA Central Office reported that a group of six "distinguished and internationally renowned scientists was selected to be the ad hoc Scientific Review Committee for the proposals." They were selected on their reputation as "outstanding in the field of environmental science." The committee was to operate much like an NIH study section group; it was to "exclusively make judgment on the scientific quality of each proposal."
These environmental research centers were established in July 1994, in the Veterans Affairs Medical Centers in Boston, Massachusetts; East Orange, New Jersey; and Portland, Oregon. The environmental centers are expected to be fully funded by October 1, 1994. The grants are for five years and there will be a mid-point review conducted by a panel of experts other than the ad hoc committee.
Constituting the ad hoc committee to review proposals and having a mid-term review are excellent ideas; however, the halfway point for a review may come too late. This committee recommends a review by an advisory committee at six months and then every year. The VA should examine how the entire problem can be investigated, and should provide a sufficient degree of central evaluation and advice. Possibly, an external advisory committee could be established, and collaboration with qualified investigators and other federal and non-federal agencies more strongly encouraged.
Review should ensure that appropriate infrastructures will be in place and that funding will support new initiatives and not simply fund current projects. Limiting the non-recurring instrumentation expenditures to $100,000 will surely lead to core laboratories with limited or out-moded testing capabilities.