This chapter provides an overview of the cohorts that were assembled to study the symptoms of the 1990–1991 Gulf War veterans. It discusses the limitations of the studies and summarizes the finding of each. Some of the cohorts were brought together in the first few years after the Gulf War; others were assembled later. Most of the studies compare sizable groups of deployed veterans with groups of veterans who were not deployed during the period of the Gulf War (often referred to as era veterans) or who were deployed to locations other than the Persian Gulf (for example, Bosnia) during the period (referred to as deployed elsewhere). In addition, a number of volunteer registries were assembled by the US Department of Defense (DOD) and Department of Veterans Affairs (VA) and by other governments whose forces were included in the Gulf War coalition. All those groups constitute the major cohorts that are described in this chapter.
The chapter’s focus on symptom studies, rather than on studies of well-characterized diagnosed medical and psychiatric conditions, provides the foundation for the committee’s understanding of the symptomatology experienced by the Gulf War veterans. Some of the authors have conducted factor analyses or proposed case definitions, and those studies are discussed in Chapters 4 and 5.
The chapter begins with a discussion of the limitations seen throughout the studies, which is followed by a description of the cohorts assembled after the Gulf War. The studies detail the symptoms found in the Gulf War veterans. They are organized by a key feature of the studies’ design: how the study populations were defined. Three categories of studies were identified—population-based, military-unit–based, and registry-based. The discussion of each study includes a summary of its methods and its major findings pertaining to reported symptoms. Table 3.1, summarizing all the studies reviewed, is at the end of the chapter.
The cohort studies of Gulf War veterans and their derivatives have contributed greatly to our understanding of veterans’ health but are subject to limitations that are commonly encountered in observational epidemiologic studies. They include selection biases that limit the studies’ representativeness and generalizability
to the larger veteran population; self-reporting of health outcomes and exposures, which is affected by recall bias; outcome misclassification; and reporting bias.1Box 3.1 briefly summarizes some of the general biases that affect this body of literature.
Bias, refers to the systematic, or nonrandom, tendency for an observed value to deviate from a true value because of flaws in study design and methods. A biased design may result in weakening an association, in strengthening an association, or in generating a spurious association. Because all studies are susceptible to bias, a primary goal of research design is to minimize bias or to adjust the observed value of an association by correcting for bias if its sources are known.
There are two major types of bias: selection bias and information bias. Selection bias involves a systematic error—in how subjects are identified, recruited, included, or excluded or in how they participate in a study—that leads to a distortion of a true association. Information bias results from how data are collected and can result in measurement errors, imprecise measurement, and misclassification. Those biases may be uniform in an entire study population or may affect some subgroups of the population more than others.
Common Medical-Research Biases That Affect Studies of Gulf War and Health
Selection bias: Bias can result from selection of participants in such a way that they do not represent the target population or the probability of selection is related to exposure or disease status. This may be due to a poor definition of the eligible population or failure to obtain a random sample. Includes
Nonresponse bias: Participants have a different exposure or disease status from nonparticipants.
Volunteer bias: Participants who volunteer are more likely to have the exposure or disease of interest; this is a particular problem for registry studies that collect information on participants who enroll voluntarily.
Healthy-warrior effect: Veterans or personnel who were deployed may be healthier than those who were not deployed or than civilians; selection of healthier people occurs at enlistment and separation (ill and injured personnel are more likely to leave the military).
Information or measurement bias: Misclassification of participants’ exposure or disease status may be based on the information collected by various methods (such as a mailed questionnaire, a telephone interview, record review, or a medical examination). Includes
Recall bias: The presence of disease influences participants’ reflection and perception of possible causes and can make them likely to report more exposures than or different exposures from nondiseased participants.
Reporting bias: Participants are more likely to report responses that they perceive as favorable and to underreport undesirable responses.
Temporal ambiguity: This occurs when it cannot be established that an exposure occurred before the onset of disease; it is common in cross-sectional assessments.
Confounding: This occurs when a risk factor for the disease that is also related to the exposure creates a spurious exposure–disease association; in other words, a risk factor may cause the exposed and nonexposed participants to have different background disease risks.
SOURCE: Delgado-Rodríguez and Llorca, 2004; Levenson et al., 1990; Pearce et al., 2006.
1Biases previously described by the IOM (IOM, 2006, 2010)
An important limitation is selection bias, which results in a lack of representativeness and limits one’s ability to generalize results to the entire population of interest; this is related to what is known as external validity. For example, six of the cohorts are made up of veterans that were selected according to where they served in the military (a military-unit–based study) (Fukuda et al., 1998; Gray et al., 1999; Haley et al., 1997; Pierce, 1997; Proctor et al., 1998; Stretch et al., 1995). Military-unit studies are not representative of all Gulf War veterans with respect to their duties and locations during deployment, possible exposures, military status during the war (active duty, reserves, or National Guard), military status after the war (active duty, reserves, or discharged), branch of service (Army, Navy, Air Force, or Marine Corps), or ease of ascertainment (IOM, 1999).
In population-based cohort studies, a sample or the entire defined population is selected for longitudinal study. Ideally, a population-based study starts prospectively with a cohort that is convened before the exposure or onset of symptoms. The study of a cohort, that is representative of a defined population offers several advantages. For example, it allows the estimation of distributions and prevalences of relevant variables in the reference population; risk-factor distributions measured at baseline in a study involving periodic examinations of the cohort can be compared with distributions in future cross-sectional samples to assess risk-factor trends over time; and a representative sample is the ideal setting in which to carry out unbiased evaluations of relationships not only of confounders to exposures and outcomes but also among any other variables of interest (Szklo, 1998).
Some population-based studies of Gulf War veterans sample a cohort of veterans by contacting them where they live as opposed to where they seek treatment or where they serve in the military (for example, a particular base or a particular branch, such as the Air Force). Studies of military units or other military groups are less representative of the broader Gulf War veteran population than are population-based studies. Military-unit–based studies are generalizable only to members of that unit and not to the broader veteran or military population Large population-based studies of Gulf War veterans have been conducted in each of the three major countries that participated in the Gulf War coalition: the United States, Canada, and the United Kingdom (Cherry et al., 2001; Goss Gilroy Inc., 1998; Kang et al., 2000; Unwin et al., 1999).
Representativeness is also compromised when some demographic groups are underrepresented in the study sample, such as women. Some studies used methods to increase a sample’s representativeness by oversampling specific groups. For example, Kang et al. (2000) oversampled women and those serving in the National Guard and reserves, and this resulted in a study sample that was about 20% women, 25% National Guard, and 33% reservists. The controls were stratified by sex, unit, and branch of service to mirror the population of deployed veterans.
A study’s representativeness, even if it is population-based, can be compromised by low participation or response rates, which may result in nonresponse bias. For example, Gulf War veterans who are symptomatic may choose to participate more frequently than those who are not symptomatic. Response rates in the studies discussed in this chapter are highly variable; they range from 92% (Steele, 2000; Wolfe et al., 1998) to 28% (Salamon et al., 2006). In some studies, researchers not only try to measure nonresponse bias by comparing participants with nonparticipants from both deployed and nondeployed populations but make adjustments to overcome it, for example, by oversampling nondeployed populations.
Quite different from population-based studies are ones that rely on voluntary participants to identify themselves, such as those who volunteer to participate in a registry. Registry studies may be subject to volunteer bias. They should be interpreted with caution inasmuch as registry participants are self-selected (sicker people are more likely to join) and not representative of the entire Gulf War veteran population. In addition, they often do not include a control group for comparison.
Selection bias might also occur through the so-called healthy-warrior effect. That bias has the potential to occur in most of the major cohorts that compare deployed veterans with nondeployed personnel. The healthy-warrior effect is a form of selection bias in that chronically ill or less fit members of the armed forces might be less likely to have been deployed than more fit members. That is, there might have been nonrandom assignment of those selected and not selected for deployment.
Some studies attempt to measure the potential for selection bias and adjust for it in the analysis. Other studies compare Gulf War deployed veterans with two or more groups, such as veterans deployed to other locations or missions (Hotopf and Wessely, 2005).
Many issues can contribute to information bias or measurement bias and result in the misclassification of people as sick or healthy when they are not. Symptom self-reporting might sometimes introduce outcome misclassification, in which there are errors in how symptoms are classified into outcomes and analyzed. One Gulf War study sought to document outcome misclassification by comparing veterans’ symptom reporting on questionnaires with results of clinical examination about 3 months later (McCauley et al., 1999). The authors found that the extent of misclassification depended on the type of symptom being reported; agreement between questionnaire and clinical examination ranged from 4% to 79%. The overall problem led the investigators to caution that questionnaire data, in the absence of clinical evaluation or adjustment, might lead to outcome misclassification. Another study also found poor reliability and validity of self-reported diagnoses compared with medical records (Gray et al., 1999). In contrast, a study by VA (Kang et al., 2000), which verified a random subset of self-reported conditions against medical records, found a strong correlation between the two (above 93%). Those data, however, were available only for the 45.2% who signed consent forms that allowed researchers to verify records.
Another important limitation is that most cohort studies rely on self-reporting of symptoms on questionnaires. Most of the larger epidemiologic studies described here were conducted through mail or telephone surveys, and this precluded clinical examination and diagnosis. Studies based on self-reporting have inherent limitations because of potential inaccuracies in recalling past events and difficulty in verifying the reports. Symptom self-reporting potentially introduces reporting bias, which occurs when the group being studied (such as deployed veterans) overreports particular symptoms (such as symptoms that are more intense or more recently experienced), that is, reports the symptoms more frequently than a comparison group (such as nondeployed veterans). Reporting bias, in this example, would lead to an overestimation of the prevalence of symptoms or diagnoses in the deployed population. Similarly, self-reporting of exposures is problematic and subject to recall bias in that sick soldiers may be more likely to report that they were exposed. Issues arising as a result of symptom self-reporting are best addressed through clinical evaluations, as has been done by some researchers (Ishøy et al., 1999a; Kelsall et al., 2005; McCauley et al., 1999). Many
limitations of reporting and recall bias are present in Gulf War research (e.g., Murphy et al., 2006, 2008).
Virtually all the studies cited in this chapter are cross-sectional—surveys, questionnaires, interviews, and the like were conducted at a single time—and can thus be subject to temporal ambiguity. Even though some studies conduct serial cross-sectional assessments over time (Kang et al., 2000, 2009), the timing of retrospective exposures and symptoms is difficult to ascertain, and this limits the opportunity to learn about symptom duration and chronicity, latency of onset, prognosis, and potential causal factors.
Many other factors can affect the association between an exposure and an outcome, including lifestyle, hereditary factors, and additional exposures, which are known as confounding factors. Confounding occurs when a variable or characteristic otherwise known to be predictive of an outcome and associated with an exposure (and not on the causal pathway under consideration) can account for all or part of an apparent association. A confounding variable is an uncontrolled variable that influences the outcome of a study to an unknown extent and whose effects cannot be precisely evaluated. Carefully applied statistical adjustments can often control for or reduce the influence of a confounder (IOM, 2010).
Other limitations of the body of evidence are that studies might be too narrow in their assessment of health status, the measurement instruments might have been too insensitive to detect abnormalities that affect deployed veterans, and the period of investigation might have been too brief to detect health outcomes that have a long latency or require many years to progress to disability, hospitalization, or death. Finally, research into the health effects of Gulf War deployment is limited by the interval between the war and when the studies were conducted. Many studies were conducted years after the war, and this limits the ability to determine when symptoms developed and the ability to detect causal associations —for example, the earliest assessments were conducted in 1993 by Pierce (1997) and Wolfe et al. (1998), in 1994 by Gray et al. (1999), and in 1995 by Fukuda et al. (1998). Followup of the cohorts is also limited because some active military separate each year. And the delay between the war and the studies allows the dissemination of speculation by the media and others that may have affected veterans’ recall (Hotopf and Wessely, 2005).
The following are population-based studies of samples of veterans or military personnel. VA conducted a nationally representative study of Gulf War veterans (Kang et al., 2000, 2002, 2009). Several studies of selected population-based samples of veterans defined by state of residence were conducted (Bourdette et al., 2001; Iowa Persian Gulf Study Group, 1997; McCauley et al., 1999; Steele, 2000). Finally, several additional studies of populations of allied military personnel are described (Cherry et al., 2001; Goss Gilroy Inc., 1998; Hotopf and Wessely, 2005; Ishøy et al., 1999a; Kelsall et al., 2004; Simmons et al., 2004; Unwin et al., 1999).
Department of Veterans Affairs Study
VA conducted a study that used the National Health Survey of Gulf War Veterans and Their Families to estimate the prevalence of symptoms and other health outcomes (including
reproductive outcomes in spouses and birth defects in children) in Gulf War veterans vs nondeployed Gulf War–era veterans. The three-phase retrospective study was designed to be representative of nearly 700,000 US veterans who were deployed to the Persian Gulf and 800,680 veterans who were not deployed but were in the military during September 1990–May 1991. In the first phase, questionnaires were mailed to 30,000 veterans (15,000 Gulf War deployed and 15,000 era veterans) identified by the DOD Data Manpower Center as representing the various branches and units of the military. The questionnaire contained a list of 48 symptoms and questions about chronic medical conditions, functional limitations, and other items from the National Health Interview Survey and included questions about exposures. The overall response rate was about 70%. The second phase used telephone interview software in an attempt to capture those who did not respond to the mailed questionnaires. In addition, medical records were obtained for a random sample of 4,200 respondents to validate self-reports of clinic visits or hospitalizations within the preceding year. The third phase was a comprehensive medical examination, including laboratory testing, of a random sample of 2,000 veterans drawn from the Gulf War population and a comparison group (Kang et al., 2000).
The investigation found that Gulf War veterans reported statistically significantly greater functional impairment in the preceding 2 weeks (27.8% vs 14.2%), limitation of employment (17.2% vs 11.6%), and health care use in the preceding year as assessed on the basis of clinic visits (50.8% vs 40.5%) and hospitalizations (7.8% vs 6.4%) than era veterans. Gulf War veterans reported higher prevalences of all 48 symptoms on the health inventory. The most frequently reported were runny nose, headache, unrefreshing sleep, anxiety, joint pain, back pain, fatigue, ringing in ears, heartburn, difficulty in sleeping, depression, and difficulty in concentrating (see Table 3.2) (Kang et al. 2000). Those 12 symptoms are similar in prevalence to the same symptoms in a UK cohort (Unwin et al., 1999). In a randomly selected subset of veterans, medical-record reviews verified more than 90% of self-reported reasons for clinic visits or hospitalizations (Kang et al., 2000).
TABLE 3.2 Results of Department of Veterans Affairs Studya
|Most Common Self-Reported Symptoms||Prevalence in Gulf War Veterans (%)||Prevalence in Non–Gulf War Veterans (%)|
|Ringing in ears||37||23|
|Difficulty in sleeping||37||21|
|Difficulty in concentrating||35||13|
|Most Common Self-Reported Symptoms||Prevalence in Gulf War Veterans (%)||Prevalence in Non–Gulf War Veterans (%)|
|5 Most Common Self-Reported Chronic Medical Conditions|
aSubjects were asked whether symptoms were recurring or persistent during previous year. Differences in prevalence are all statistically significant (p < 0.05).
SOURCE: Kang et al., 2000. Adapted with permission from Lippincott Williams and Wilkins/Wolters Kluwer Health: Journal of Occupational and Environmental Medicine (2000).
A followup study of the same population (Kang et al., 2009) was conducted in 2005 to obtain survey-based health information from the 15,000 Gulf War–deployed and 15,000 Gulf War–era veterans originally surveyed in 1995. In phase I of the followup, VA and Social Security records through December 2002 were used to identify the 29,607 living participants and mail health questionnaires to them. Phase II consisted of telephone interviews with 2,000 participants who did not respond to the initial mailed questionnaire (nonrespondents) and a sample of 1,000 participants (respondents) who had indicated a clinic visit or hospitalization within the previous 12 months. In all, 6,111 (40%) deployed and 3,859 (27%) nondeployed veterans participated in both phases I and II, but the overall response rate was low (only 34%). However, there were no differences in deployment status between respondents and nonrespondents. The administered questionnaire was a modified version of the 1995 questionnaire and included the Patient Health Questionnaire, the 12-Item Short Form Health Survey (SF-12), and other items used to assess general health status. Unexplained multisymptom illness, in this study, was defined as having several symptoms that persisted for 6 months or longer and were not adequately explained by other diagnoses. Those symptoms included fatigue; muscle or joint pain; headache; memory, digestive, respiratory, or skin problems; or any other unexplained symptoms. Unexplained multisymptom illness was identified in 36.5% of the deployed and 11.7% of the era veterans, for a risk ratio of 3.05 (95% confidence interval [CI] 2.77–3.36), after adjustment for age, sex, race, body-mass index, current cigarette-smoking, rank, branch of service, and unit component (active duty, National Guard, or reserve). Multisymptom illness was the most widely reported medical condition in Gulf War veterans except for arthritis. Gulf War veterans also had higher rates of functional impairment, of limitations of activities, of at least one clinic or doctor visit, and of hospitalization.
The Iowa Study
The “Iowa study” was a cross-sectional survey of a sample of 4,886 military personnel who listed Iowa as their home of record at the time of enlistment (Iowa Persian Gulf Study Group, 1997). The study examined the health of military personnel in all branches of service who were still serving or had left service. The sample was randomly selected from the 28,968
military personnel who listed Iowa as their home of record. Of the study subjects who were contacted, 3,695 (90.7%) completed a telephone interview in 1995–1996. Study subjects were divided into four groups: Gulf War–deployed regular military, Gulf War–deployed National Guard or reserve, non–Gulf War–deployed regular military, and non–Gulf War–deployed National Guard or reserve. Trained examiners used standardized questions, instruments, and scales in interviewing the subjects.2 The study found that Gulf War veterans had significantly higher prevalences of symptoms of depression (17.0% vs 10.9%, p < 0.001), posttraumatic stress disorder (PTSD; 1.9% vs 0.8%, p = 0.007), chronic fatigue (1.3% vs 0.3%, p < 0.001), cognitive dysfunction (18.7% vs 7.6%, p < 0.001), bronchitis (3.7% vs 2.7%, p < 0.001), asthma (7.2% vs 4.1%, p = 0.004), fibromyalgia (19.2% vs 9.6%, p < 0.001), alcohol abuse (17.2% vs 4.1%, p = 0.02), and anxiety (4.0% vs 1.8%, p < 0.001). Gulf War veterans scored significantly lower on all eight subscales for physical and mental health on the 36-item Short Form Health Survey (SF-36); this indicated lower quality of life than that of nondeployed personnel. The subscales for bodily pain, general health, and vitality showed the greatest differences between deployed and nondeployed veterans (Iowa Persian Gulf Study Group, 1997). In short, this large, well-controlled study demonstrated that some sets of symptoms were more frequent in Gulf War veterans than in nondeployed military controls.
Oregon and Washington Veteran Studies
Veterans from Oregon and Washington were studied in a series of analyses by investigators of the Portland Environmental Hazards Research Center (McCauley et al., 1999). A mailed questionnaire, to assess general health through symptom self-reports, was sent to a random sample of 2,343 of the total of 8,603 Gulf War veterans who listed Oregon or Washington as their home state of record at the time of deployment. The study did not include a nondeployed comparison group. The response rate was 48.4%. The study found high rates (21–60%) of self-reported symptoms, including cognitive–psychologic symptoms, unexplained fatigue, musculoskeletal pain, gastrointestinal complaints, and rashes. The 225 veterans who participated in the clinical examinations displayed differences between the symptoms that they reported on questionnaires and the symptoms that they reported at clinical examination. The difference might suggest high rates of outcome misclassification based on either the questionnaire or the examination.
Kansas Veteran Study
Kansas established the Kansas Persian Gulf War Veterans Health Initiative to determine the patterns of veterans’ health problems. Using lists of eligible veterans from DOD, Steele (2000) conducted a population-based survey of veterans who listed Kansas as their home state of record. A stratified random sample of 3,138 was selected, of whom 2,396 were located with in-state contact information. The survey, mailed out in 1998, asked about 16 specific medical or psychiatric conditions, 37 symptoms, service branch, locations during the Gulf War (including whether the veterans were notified about the Khamisiyah demolitions; see Chapter 2), and
2Sources of questions included the National Health Interview Survey, the Behavioral Risk Factor Surveillance Survey, the National Medical Expenditures Survey, the Primary Care Evaluation of Mental Disorders, the Brief Symptom Inventory, the CAGE questionnaire (for alcoholism), the PTSD (Posttraumatic Stress Disorder) Checklist—Military, the Centers for Disease Control and Prevention Chronic Fatigue Syndrome Questionnaire, the Chalder Fatigue Scale, the American Thoracic Society questionnaire, and the Sickness Impact Profile.
vaccinations. Kansas Gulf War veterans reported greater prevalences of 10 physician-diagnosed conditions than Kansas nondeployed veterans: skin conditions, stomach or intestinal conditions, depression, arthritis, migraine headaches, chronic fatigue syndrome (CFS), bronchitis, PTSD, asthma, and thyroid conditions. The investigators used their own definition of Gulf War illness, which was similar to that used by the Centers for Disease Control and Prevention (CDC) (Fukuda et al., 1998) and which required having at least one moderately severe or two or more chronic symptoms in at least three of six domains: fatigue and sleep problems; pain symptoms; neurologic, cognitive, and mood symptoms; gastrointestinal symptoms; respiratory symptoms; and skin symptoms. The symptoms had to persist or recur in the year before the study interview and had to have been a problem for the study participants in 1990 or later. Using their case definition, the researchers found that 34.2% of Gulf War veterans and 8.3% of nondeployed veterans met criteria for Gulf War illness (odds ratio [OR] = 4.68, 95% CI 3.25–6.75). On the basis of the CDC case criteria, the study found that 47.2% of Gulf War veterans and 19.8% of nondeployed veterans had chonic multisymptom illness (OR = 3.26, 95% CI 2.48–4.28). The prevalence of Kansas-defined Gulf War illness was lowest in Gulf War veterans who served on ships and highest in those who served in Iraq or Kuwait.
Canadian Veteran Study
A 1997 survey mailed to the entire cohort of Canadian Gulf War veterans found high prevalences of several chronic conditions (Goss Gilroy Inc., 1998).3 Some 3,113 respondents from Canada who had been deployed to the Gulf War were compared with 3,439 respondents who had been deployed elsewhere during the same period. The Gulf War veterans who responded reported symptoms of cognitive dysfunction, multiple chemical sensitivity (MCS), major depression, PTSD, chronic dysphoria, anxiety, and respiratory diseases at higher rates than the controls. The greatest differences between Gulf War deployed forces and those deployed elsewhwere were in symptoms of cognitive dysfunction, MCS, and major depression. Symptoms of cognitive dysfunction had the highest overall prevalence: in 34–40% of Gulf War veterans and 10–15% of veterans deployed elsewhere. Gulf War veterans also reported significantly more visits to health-care practitioners, greater dissatisfaction with their health status, and greater health-related reductions in recent activity.
United Kingdom Veteran Studies
The United Kingdom sent 53,000 personnel to the Gulf War. Two teams of researchers each studied a separate, nonoverlapping, stratified random sample of those Gulf War veterans. The first team was from the University of London (Guy’s, King’s, and St. Thomas’s Medical Schools), the second team from the University of Manchester. A third team of researchers from the London School of Hygiene and Tropical Medicine surveyed the entire cohort of 53,000 veterans for a focused study of birth defects and other reproductive outcomes.
3In January 1997, Goss Gilroy Inc. was contracted by the Canadian Department of National Defence to carry out an epidemiologic survey of Canadians who served in the Gulf War to establish the overall health status of Gulf War personnel.
University of London Veteran Studies
Unwin et al. (1999) at the University of London investigated the health of UK servicemen in a population-based study. The study used a random sample of the entire UK contingent deployed to the Gulf War and two comparison groups. One of the comparison groups (n = 2,620) was deployed to the conflict in Bosnia; this is the only study that used a comparison population that had combat experience during the time of the Gulf War. The second comparison group of era veterans (n = 2,614) was deployed to noncombat locations outside the United Kingdom in the same period. As opposed to what was done in some studies, the nondeployed control group was recruited from among the subset of nondeployed service members who were fit for combat duty, and this avoided selection bias related to the healthy-warrior effect. A mailed questionnaire queried symptoms (50 items), medical disorders (39 items), exposure history (29 items), and functional capacity. The authors compared ORs for each symptom after controlling for potential confounding factors (including sociodemographic and lifestyle factors), using logistic regression analysis. Only results on male veterans were analyzed, because female veterans’ roles and symptoms were distinct enough to warrant separate consideration. Responses to the questionnaire were received from 70% of the 53,462 Gulf War–deployed, 61.9% of the 39,217 in the Bosnia cohort, and 62.9% of the 250,000 era cohort members. Bosnia-deployed veterans were more likely to be in service, unmarried, younger, and drank more alcohol than Gulf War–deployed veterans. The era veterans were similar to Gulf War–deployed veterans but included more non-smokers.
The Gulf War–deployed veterans (n = 2,961) reported higher prevalences of symptoms and diminished functioning than did either comparison group. Gulf War veterans were 2–3 times more likely than comparison subjects to have met symptom-based criteria for chronic fatigue, posttraumatic stress reaction, and CDC-defined chronic multisymptom illness (CMI; Fukuda et al., 1998). More specifically, 25.3% of Gulf War veterans met CDC case criteria for CMI compared with 11.8% of Bosnia and 12.2% of Gulf War–era veterans. The most frequently reported symptoms were feeling unrefreshed after sleep, irritability or outbursts of anger, headache, fatigue, sleeping difficulties, forgetfulness, joint stiffness, loss of concentration, flatulence or burping, pain without swelling, and redness in several joints. Gulf War veterans were 2–3 times more likely to report those symptoms, but results were not statistically significant. On the SF-36, Gulf War veterans reported significantly worse health perception but not worse physical functioning. It should be noted that the members of the Bosnia cohort, who also had been deployed to a combat setting, reported fewer symptoms than the Gulf War cohort, and this suggests that combat deployment itself does not necessarily account for higher symptom reporting (Unwin et al., 1999).
In a followup study, a postal survey was sent 11 years after the war to a random sample of 3,305 participants (1,472 Gulf War–deployed, 909 Bosnia-deployed, and 924 era veterans) from the total who completed the first study described above. The response rates were as follows: 74.0% Gulf War–deployed, 70.2% Bosnia-deployed, and 69.7% era veterans. Respondents completed the Chalder fatigue scale, the General Health Questionnaire (GHQ), SF-36, and the count of physical symptoms (Hotopf et al., 2003). Compared with the first survey (time 1), respondents reported a modest reduction in fatigue, modest reduction in psychologic distress on the GHQ, and slight worsening on SF-36. Compared with the two groups of non–Gulf War–deployed veterans surveyed at time 2, deployed veterans performed worse on all measures.
Deployed veterans reported a mean of 10.7 symptoms vs 7.9 and 6.4 in the two non–Gulf War–deployed veteran groups. They had no higher incidence of new illnesses.
University of Manchester Veteran Study
The University of Manchester study used a random sample of UK veterans 7 years after the Gulf War (Cherry et al., 2001). The cohort was deliberately separate from that studied by Unwin et al. (1999). The 9,505 eligible deployed veterans were divided into two groups—4,755 in the main cohort and 4,750 in a validation cohort to permit replication of analysis and to assess consistency. The control population of 4,749 consisted of nondeployed veterans who were in good general health. Veterans were sent a questionnaire about the extent to which they were burdened by 95 symptoms in the previous month. By asking them to mark their answers on a visual analogue scale, investigators sought to determine the degree of symptom severity. Investigators also sought to determine areas of peripheral neuropathy by asking veterans to shade on two pictures of mannequins the body areas where they were experiencing pain or numbness and tingling. Deployed veterans reported greater severity of almost all 95 symptoms. The overall mean severity scores of the two Gulf War cohorts were similar and significantly greater than the score of the non–Gulf War cohort. Deployed veterans’ severity scores for 14 symptoms—including memory, concentration, and mood problems—were at least twice those of the nondeployed veterans. Numbness and tingling were reported by about 13% of deployed and about 7% of nondeployed. Widespread pain was also reported more frequently (12.2% vs 6.5%).
London School of Hygiene and Tropical Medicine Veteran Study
The third British study was a large mail survey conducted by researchers at the London School of Hygiene and Tropical Medicine (Maconochie et al., 2003; Simmons et al., 2004). It was designed largely to assess reproductive outcomes in Gulf War veterans, but it contained open-ended questions about their general health. The exposed cohort consisted of all UK Gulf War veterans, and the unexposed cohort consisted of a random sample of nondeployed UK military personnel from the same period. Although the number of surveys returned in the study was large (25,084 by Gulf War veterans and 19,003 by non–Gulf War veterans), the participation rates were low (47.3% and 37.5% of male and female Gulf War veterans, respectively, and 57.3% and 45.6% of male and female nondeployed veterans). Simmons et al. (2004) reported that 61% of responding Gulf War veterans and 37% of responding nondeployed veterans reported at least one new medical symptom or disease since 1990. Some 85% of symptoms and diseases were reported more frequently by Gulf War veterans. The strongest associations were for mood swings (OR = 20.9, 95% CI 16.2–27.0), memory loss or lack of concentration (OR = 19.6, 95% CI 15.5–24.8), night sweats (OR = 9.9, 95% CI 6.5–15.2), general fatigue (OR = 9.6, 95% CI 8.3–11.1), and sexual dysfunction (OR = 4.6, 95% CI 3.2–6.6). Adjustments were made for age, service, rank, serving status, alcohol consumption, and smoking. Veterans’ belief that they had “Gulf War syndrome” was associated with the greater reporting of symptoms or disease, but only 6% of Gulf War veterans believed that they had the syndrome.
Danish Peacekeeper Studies
Military personnel from Denmark were involved in peacekeeping or humanitarian missions that occurred predominantly after the Gulf War ceasefire but were in the same areas as other coalition forces that served in Gulf War combat (Ishøy et al., 1999). A total of 821 Danes,
deployed from August 1990 to December 1997, were eligible for inclusion in this population-based cohort, and 686 (83.6%) agreed to participate in the study. The deployed veterans were matched by age, sex, and profession to 400 members of the Danish armed forces who were not deployed to the Gulf War; 231 (57.8%) agreed to participate. Participants completed a detailed questionnaire, including 22 neuropsychologic symptoms, and then received detailed clinical health and laboratory examinations (height, weight, blood pressure, a battery of urinary and blood work, and a battery of neuropsychologic tests) and participated in physician interviews about their medical history and symptoms. The examinations were conducted in 1997–1998.
The results showed that Danish peacekeepers were significantly more likely to have a wide variety of symptoms (with onset during or after August 2, 1990), including headache, blurry vision, numbness or tingling of hands or feet, balance difficulties, depression and concentration problems, fatigue, sleep difficulty, nightmares, nervousness and agitation, and difficulty in pronouncing words. Analyses did not adjust for potential confounders. All together, Gulf War veterans reported higher prevalences of 17 of 22 neuropsychologic symptoms, eight of 14 gastrointestinal symptoms, and eight of 19 skin symptoms. Rates of symptoms that appeared before August 2, 1990, were no different among groups. Only minor differences were found among the groups in hematologic measures (Ishøy et al., 1999). The authors concluded that because Danish peacekeepers' symptoms were consistent with American Gulf War veterans’ symptoms, results indicate the existence of some common risk factors that are independent of combat.
Australian Veteran Studies
Investigators at Monash University conducted a cohort study of Australian service personnel who had (1,456) or had not (1,588) been deployed to the gulf as part of the multinational force (Kelsall et al., 2004). Participation rates were 80.5% and 56.8%, respectively. In the Australian contingent sent to the Gulf War, members of the Navy were heavily overrepresented (86.5%). Participants completed mailed questionnaires: a physical and mental health screening questionnaire (SF-12), a test for nonpsychotic psychologic illness (GHQ-12), a PTSD checklist (PCL-S), and a questionnaire about military service and exposures.
Kelsall et al. (2004) stated that participants in the deployed cohort reported higher prevalences of all 63 symptoms (all but seven were statistically significant) and reported more severe symptoms. The symptoms that had the highest prevalences were feeling unrefreshed after sleep (adjusted OR = 1.6, 95% CI 1.3–1.8), fatigue (adjusted OR = 1.6, 95% CI 1.3–1.8), headache (adjusted OR = 1.3, 95% CI 1.1–1.6), sleeping difficulties (adjusted OR = 1.6, 95% CI 1.4–1.9), and irritability and outbursts of anger (adjusted OR = 1.6, 95% CI 1.4–1.8).
French Military Study
Salamon et al. (2006) surveyed all French troops who were deployed to the Gulf War. The study was able to survey 5,666 of 20,261 French veterans (28% response rate) and perform clinical examinations on 1,008. On the basis of a health questionnaire administered from 2002 to 2004, the study found that signs, symptoms, and ill-defined (SSID) conditions were self-reported by 10.9% of Gulf War veterans. Among all participants, the symptoms reported most often were headache (82.9%), sleeping difficulties (70.9%), irritability (68.8%), backache (62.9%), and memory difficulties (56.0%).
Hawaii and Pennsylvania Active-Duty and Reserve Study
One of the first epidemiologic studies of US Gulf War veterans examined the psychologic and physical health of active-duty and reserve Army, Navy, Air Force, and Marine Corps personnel from bases in Pennsylvania and Hawaii (Stretch et al., 1995). Questionnaires were mailed to 14,167 potential study participants with questions regarding demographics; physical, psychologic, and psychosocial symptoms; deployment type; and perceived sources of stress before, during, and after combat or deployment. A total of 4,334 veterans returned the questionnaires, for a response rate of 31%. Of those, 715 active-duty personnel and 766 reserves were deployed to the Gulf War, and 1,576 active-duty personnel and 948 reserves were not deployed; the remainder deployed to other locations. Significantly more deployed personnel reported 20 of 23 queried symptoms than nondeployed personnel. For 12 symptoms, deployed personnel were more than twice as likely as nondeployed personnel to report head colds, sinus trouble, sore throat, difficulty swallowing, headaches, back problems, stomach upset, muscle aches, aching joints, cough, chills or fever, and “other problems.” Adjusted ORs for those symptoms ranged from 2.14 to 3.76 (all p < 0.001)—adjustments were made for age, rank, education, marital status, and branch of military.
Ft. Devens and New Orleans Cohort Studies
The symptom experience of two deployed cohorts of Gulf War veterans was studied by Boston-based researchers (Proctor et al., 1998; Wolfe et al., 1998). The first, an Army cohort based in Ft. Devens, Massachusetts, was surveyed longitudinally at three times (1991, 1993–1994, and 1997). The second Gulf War deployed cohort was from New Orleans. The study’s 252 subjects were the result of a stratified random sample of 2,949 troops from Ft. Devens and 928 from New Orleans; both groups consisted of active-duty, reserve, and National Guard troops. A third unit consisted of 48 members of an air-ambulance company of National Guard from Maine that had been deployed to Germany for handling wounded personnel evacuated from the gulf.
In comparison with veterans deployed to Germany during the Gulf War era, random samples of both Gulf War cohorts had higher prevalences of 51 of 52 items on a health-symptom checklist (Proctor et al., 1998). The greatest differences in prevalence of reported symptoms were of dermatologic symptoms (such as rash, eczema, and skin allergies), neuropsychologic symptoms (such as difficulty in concentrating and difficulty in learning new material), and gastrointestinal symptoms (such as stomach cramps and excessive gas). With a separate checklist, researchers found a higher prevalence of PTSD, according to the Clinician-Administered PTSD Scale (5% Ft. Devens, 7% New Orleans, and 0% Germany). The Ft. Devens group reported significantly higher rates of 35 of the 52 symptoms than the unit in Germany. The New Orleans group reported significantly higher prevalences of 24 of the 52. Among the musculoskeletal symptoms reported more frequently by the Ft. Devens deployed veterans were joint pains (OR = 2.6) and neck aches or stiffness (OR = 2.7), and among the neurologic symptoms with greater prevalences in both cohorts of deployed veterans was headache (OR = 4.2); all were statistically significant. About 30% of the Gulf War veterans and 11% of the comparison group reported an inability to fall asleep (OR 3.4–3.6, p ≤ 0.05).
In a subanalysis of 2,119 veterans in the Ft. Devens cohort conducted in 1993, Wolfe et al. (1998) reported on symptoms on a 20-item symptom checklist. They found that the most frequent symptoms were general aches and pains, being overly tired/lack of energy, headaches, trouble sleeping, nervous or tense, depressed mood, and difficulty concentrating. Some 30% of the sample reported that their physical health had become either “worse” or “much worse” since their return.
Seabee Reserve Battalion Studies
Numerous studies of the Seabees called to active duty for the Gulf War have been conducted. Haley et al. (1997) studied members of the Twenty-Fourth Reserve Naval Construction Battalion who lived in five southern states and were called to active duty. The unit was a mobile construction battalion for other branches of the military. More than half the battalion’s members had left the military by the time of the study and so were not included in the study cohort. Participants were recruited among those whose addresses were available and from veterans’ meetings. Of the 249 participants, 175 (70%) reported having had serious health problems since returning from the Gulf War. A telephone survey of a random sample of nonparticipants found that although they were demographically similar to participants, fewer (43%) reported having had serious health problems since the war. Some 11% of participants and 3% of nonparticipants were unemployed. Of the 606 men in the battalion, 41.1% participated; there was no nondeployed group for comparison. The study was the first to cluster symptoms into new syndromes by applying factor-analysis techniques (discussed in Chapter 4).
In the first of a series of studies by Gray et al. (1999), investigators surveyed active-duty Seabees who remained on active duty for at least 3 years after the Gulf War. The Seabees were in 14 Seabee commands at two locations (Port Hueneme, California, and Gulfport, Mississippi). Those who were deployed to the Gulf War were in mobile construction battalions serving in the same tasks and at the same sites as the reserve Seabee battalion studied by Haley and collaborators. During the Gulf War, Seabees built airports, supply points, and roads. Gray et al. excluded Gulf War veterans who were no longer active at the time of the study in 1994–1995.
Gray et al. (1999) enrolled 1,497 study subjects: 527 Gulf War–deployed veterans and 970 nondeployed veterans. Study subjects filled out symptom and exposure questionnaires and answered additional questions that screened for PTSD, CFS, and various psychologic symptom domains; blood and handgrip strength were also tested. The study had a 53% participation rate. Findings indicated that 55.8% of Gulf War–deployed and 31.7% of nondeployed era veterans reported prolonged symptoms (lasting for 1 month or longer) that occurred after the war; the prevalences of 35 of 41 symptoms were significantly higher in the deployed than in the nondeployed. The groups had similar pulmonary function and reactant assays (C-reactive protein, transferrin, and haptoglobin). Gulf War veterans had higher adjusted serum ferritin measurements, but results were within the normal range. Handgrip strength was lower in Gulf War veterans on the average, and they were more likely to have PTSD (15% vs 9%).
Beginning in May 1997, Gray et al. (2002) mailed a questionnaire to all 18,945 regular and reserve naval personnel who served on active-duty Seabee command during the Gulf War period. The questionnaire collected information regarding medical history, current health status, symptoms and medical problems, and environmental exposures. Of the 17,559 participants located, 11,868 completed and returned the questionnaire: 3,831 Gulf War–deployed, 4,933
deployed elsewhere, and 3,104 nondeployed. Compared with the two control groups, the deployed were more likely to report having more symptoms; the greatest differences were evident in MCS, nightmares or flashbacks, rash or skin ulcer, general muscle weakness, and unusual irritability. Gulf War–deployed Seabees were significantly more likely to report having more of all 33 self-reported medical problems than personnel in the other two groups on the basis of logistic regression analyses that controlled for age, sex, race or ethnicity, and duty status. Gulf War–deployed Seabees were also significantly more likely to report suffering from a wide variety of physician-diagnosed disorders than those nondeployed or deployed elsewhere, including CFS, PTSD, MCS, and irritable bowel syndrome (IBS). Gulf War–deployed Seabees also reported more depression, cognitive failure, digestive diseases, lost work days, and were more likely to report being in fair or poor health than the other two groups. Of the Gulf War–deployed Seabees, 22% met criteria for Gulf War illness, which the authors defined as having any of five conditions: a self-reported physician diagnosis of CFS, PTSD, MCS, or IBS or self-reporting of 12 or more other medical problems or symptoms. The percentage of members of the control groups meeting the case definition was not reported.
Pennsylvania Air National Guard Study
In response to requests from DOD, VA, and Pennsylvania, a team of investigators from CDC conducted a study to assess health status and prevalence and causes of unexplained illness in Gulf War–deployed personnel (Fukuda et al., 1998). The index unit consisted of 667 members of active-duty Air National Guard members. Three demographically similar Air Force units were used as comparison groups: Unit A consisted of 538 personnel from the Pennsylvania Air National Guard who had a mission different from that of the index unit, Unit B consisted of 838 members of a US Air Force Reserve unit, and Unit C consisted of 1,680 active-duty Air Force personnel from Florida who had a mission similar to that of the index unit. Questionnaires regarding military characteristics, demographics, health status, and 35 specific symptoms previously identified to be of concern were distributed and completed by 3,675 participants (taken together these units included 1,155 Gulf War veterans and 2,520 nondeployed veterans). Response rates were as follows: index unit, 62%; Unit A, 35%; Unit B, 73%; and Unit C, 70%. To assess symptom prevalence, the investigators combined the four units and compared questionnaire responses of deployed and nondeployed. Of 3,723 participants surveyed, those deployed to the Gulf War experienced higher prevalences of chronic symptoms than nondeployed veterans (33 of 35 symptoms of more than 6-month duration were reported to be more prevalent).
Air Force Women’s Study
Only one study was devoted to the effects of deployment on symptoms in women. Pierce (1997) surveyed by mail questionnaire 525 US Air Force women who had and had not deployed to the Gulf War. The survey was conducted in 1993 (time 1) and again in 1995 (time 2). Response rates were 82% at time 1 and 92% at time 2. The sampling was random with oversampling of those deployed to the theater of operations and of reserve and National Guard components to achieve a representative study sample. Analyses were adjusted for age. At time 1, the study found that deployed veterans reported rash, cough, depression, unintentional weight loss, insomnia, and memory problems more frequently than nondeployed veterans; differences were not statistically significant, but differences were apparent when data was stratified by
duration of deployment to the Gulf. At time 2, the most commonly reported symptoms were rash, cough, memory problems, and sex-specific problems, such as breast lumps or cysts and abnormal pap smears. The pattern of symptom reporting was similar to that by men and women in other Gulf War studies (Unwin et al., 2002).
Several registries have been formed to track and collect information to assist in the investigation of health concerns related to service in the Gulf War. The committee reviewed the registry studies with caution. Registry participants cannot be considered representative of all Gulf War veterans in that they are self-selected subjects, many of whom have joined the registries because they believe that they have symptoms related to Gulf War illness; they were not randomly selected from all Gulf War military personnel, and there is not a nondeployed control group.
In 1992, VA developed and implemented the Persian Gulf Registry. Its original purposes were to ease returning veterans into the VA health care system, to create a registry containing medical and other data on Persian Gulf veterans that would assist in addressing questions about possible future effects of exposure to air pollutants and other environmental agents, and to serve as a basis of future medical surveillance. Exposures, particularly those associated with the oil-well fires, were included as part of the registrants’ history. As time passed, it became apparent that a number of exposures and a host of symptoms being reported needed further investigation.
DOD also developed and implemented a Persian Gulf clinical program to diagnose and treat conditions in active-duty military personnel who had medical complaints that they attributed to service in the Gulf War. DOD and VA collaborated and used experts to develop clinical protocols; by 1994, they had implemented similar and parallel clinical evaluation programs. In light of continuing concern about the potential health consequences of service in the Persian Gulf, DOD and VA revised their clinical programs to improve diagnosis of veterans' health complaints. DOD instituted the Comprehensive Clinical Evaluation Program (CCEP), and VA instituted the Persian Gulf Registry and Uniform Case Assessment Protocol (UCAP). Both programs included a medical history, physical examinations, laboratory tests, and specialty consultation as needed. By early 1994, over 20,000 veterans had been examined as part of VA’s Persian Gulf Registry program (IOM, 1998, 1997).
Department of Defense Registry Studies
Four investigations have used the CCEP to identify cohorts for study. A study of the first 20,000 cases seen in the first phase of the CCEP was conducted by Joseph et al. (1997). Findings indicate that 17.8% of Gulf War veterans in the registry had SSID, the most common of which were fatigue, headache, memory problems, and sleep disturbances. In the Gulf War veterans who indicated a date of onset, symptoms were reported to have begun more than 6 months after return from the Gulf War. Gulf War veterans who had SSID did not have any characteristic signs of disease or consistent laboratory abnormalities.
Kroenke et al. (1998) reported on findings from a provider-administered symptom questionnaire on 18,495 Gulf War veterans from the CCEP. The most common symptoms found were joint pain, fatigue, headache, memory and concentration difficulties, sleep disturbance, and
rash. The study tracked timing of onset of symptoms relative to the war. Symptom onset was found to be delayed: 66% of symptoms did not appear until after the war, and 40% more than 1 year after the war. According to the authors, there was no association between individual symptoms, types of combat experience, self-reported exposures, or patient demographics. Increased symptom counts were associated with loss of work, the number of self-reported exposures, the number of types of combat experience, and particular International Classification of Diseases, Ninth Revision (ICD-9) diagnostic categories (such as psychologic disorders).
Roy et al. (1998) reported on 21,579 Walter Reed Army Medical Center patients who had participated in the CCEP and were referred for additional evaluation. Physicians at Walter Reed conducted a series of evaluations, including a patient health questionnaire, medical history, laboratory studies, and physical examination. The investigators reported that 17.2% of the CCEP participants had a primary diagnosis of SSID, whereas 41.8% had SSID as a primary or secondary diagnosis. The most common symptoms were fatigue, headache, and memory loss. The authors concluded that an analysis of the SSID diagnoses in the large series of Gulf War veterans did not identify a new or unusual syndrome.
Erickson et al. (1998) described musculoskeletal complaints in participants in the CCEP. Of the 1,250 evaluated, 18% were referred to a rheumatologist at the Fitzsimmons Army Medical Center for evaluation of musculoskeletal complaints from March 1994 to March 1995. The most common symptoms reported to rheumatologists were polyarthralgia (pain in more than three joints with or without swelling or widespread pain), knee pain, back pain, myalgia, ankle pain, and hand or wrist pain. Extensive laboratory testing was not specific enough for any diagnosis to explain symptoms.
Department of Veterans Affairs Registry Studies
One of the VA sites, in south Texas, referred 145 potential rheumatologic cases to a nearby clinic (Escalante and Fischbach, 1998). Rheumatologists at the clinic administered a health questionnaire, elicited pain symptoms, and administered the SF-36 for health-related quality of life. Almost all the patients had pain, which was widely distributed and spared no body part. Widespread pain was reported in 65.1% of Gulf War veterans. The most frequent painful areas were knees (in 65%), low back (over 60%), shoulders (50%), and hands and wrists (35%). The average values on the SF-36 were below the 25th percentile of published national norms. Pain and nonarticular rheumatic symptoms explained most of the diminished health-related quality of life.
Hallman et al. (2003) conducted a health survey in 1995 of 1,161 participating Gulf War veterans who represented a random sample of a VA registry that covered seven states. Of 48 reported symptoms, participants endorsed an average of 9.9 mild symptoms, 9.5 moderate symptoms, and 6.1 severe symptoms. The average total number of symptoms was 25.5.
A 5-year followup of the VA registry members originally surveyed by Hallman et al. (2003) was conducted by Ozakinci et al. (2006). A mail survey was sent to 390 Gulf War veterans who were later interviewed by telephone in 2000 (time 2). Compared with time 1 (1995), there was no significant change in number of symptoms reported or their severity. Subjects who were more symptomatic at time 1 showed some improvement at time 2 but remained much more highly symptomatic than those who had less severe initial symptoms. Adjustments were made for sex, rank, race, marital status, education, branch, and duty.
United Kingdom Registry Study
The health status of 3,000 consecutive registrants in the Gulf Veterans Medical Assessment Programme (GVMAP) was reported by several researchers (Coker et al., 1999; Lee et al., 2001, 2002). The GVMAP provides British Gulf War veterans, who are referred to the program by their regular clinicians, with free specialized health assessments. Some 75% of the registrants were considered well (without organic or psychiatric conditions or able to function normally both physically and mentally): 10% (303) had no conditions or symptoms, 21% (619) complained of symptoms, and 44% had diagnoses of incidental organic or psychiatric illnesses. Among all registrants, the most commonly reported symptoms were affective symptoms (mood, emotions, or feelings), joint and muscle aches and pains, and fatigue.
Many Gulf War veterans suffer from an array of health problems and symptoms (for example, fatigue, muscle and joint pain, memory loss, gastrointestinal disorders, and rashes) that are not specific to any disease and are not easily classified with standard diagnostic coding systems. Studies since the middle 1990s have found a higher prevalence of self-reported and clinically verified symptoms in Gulf War veterans than in nondeployed Gulf War–era veterans or other control groups. United States, United Kingdom, Canadian, Australian, and Danish Gulf War veterans report higher rates and greater severity of nearly all symptoms or sets of symptoms than their nondeployed counterparts; that finding was reported consistently in every study reviewed by this committee. However, Gulf War veterans do not all experience the same array of symptoms, and the symptoms reported are also found in the nondeployed. Furthermore, the studies are beset with limitations; there is the likelihood that bias distorts the findings and that the representativeness of many, if not most of the studies, is uncertain.
TABLE 3.1 Studies of Symptoms in Gulf War Veterans
|Department of Veterans Affairs Study|
|Kang et al., 2000||Cross-sectional survey, conducted in 1995||Population-based sample of 15,000 troops deployed, 15,000 troops not deployed; National Health Survey of Gulf War–era veterans and their families. Includes any person who served in the US military on active duty, in reserves, or in National Guard irrespective of whether they were still in the service or separated. Enrolled 11,441 GWVs and 9,476 NDVs.||Health questionnaire (48-item symptom inventory), functional impairment, limitations of employment, healthcare use.||GWVs vs NDVs reported higher rates of functional impairment (27.8% vs 14.2%), limitations of employment (17.2% vs 11.6%), and health-care use as assessed by clinic visits (50.8% vs 40.5%) or hospitalizations (7.8% vs 6.4%). GWVs reported more frequently all 48 symptoms on health inventory; most frequently severe were back pain, runny nose, joint pain, headaches being anxious, difficulty in getting to sleep, feeling tired, skin rash, excessive fatigue, and heartburn or indigestion.||Sex, branch of service, unit status||Response rate: 70% total, 75% GWVs, 64% NGWs. Compared survey participants with VA health registry participants (n = 15,891). From Kang et al., 2002: 225 NGVs were actually GWVs. Some misclassification in original sample. Assessed selection bias|
|Kang et al., 2009||Followup cross-sectional medical evaluation and survey, conducted in 2005||Derivative of Kang et al. (2000). Eligible population selected randomly from Kang population. Eligible 6,111 GWVs and 3859 NDVs; enrolled 5,469 GWVs and 3,353 NDVs.||Slightly modified version of questionnaire used by Kang et al. (2000) plus Patient Health Questionnaire, SF-12, Multisystem Illness, CFS-like illness.||GWVs vs NDVs have higher rates of multisymptom illness (36.5% vs 11.7%, adjusted RR = 3.05, 95% CI 2.77–3.36), higher rates of CFS-like illness (9.4 vs 3.4, adjusted RR = 2.38, 95% CI 1.97–2.87), more functional impairment (31.6% vs 16.5%), more limitations on activities (29.0% vs 19.2%), higher rates of at least one clinic or doctor visit (56.2% vs 45.9%, p < 0.001), and more hospitalizations (10.5% vs 8.0%, p < 0.001).||Age, sex, race, body mass, rank, branch of service, unit component (active duty, National Guard, or reserve)||Response rate 34%; differences between respondents and nonrespondents were nondifferential by deployment status.|
|Iowa Persian Gulf Study|
|Iowa Persian Gulf Study Group, 1997||Cross-sectional study; Interviews conducted 1995–1996||Iowa listed as home of record on initial military record, and service in regular military or activated National Guard or reserve some time from August 2, 1990 to July 31, 1991. Eligible 4,886; Enrolled 1,896 GWVs and 1,799 NDVs.||Telephone interview using standardized questionnaires, instruments, and scales.||GWVs reported significantly higher prevalence of symptoms of depression (17.0% vs 10.9%, p < 0.001), PTSD (1.9% vs 0.8%, p = 0.007), chronic fatigue (1.3% vs 0.3%, p < 0.001), cognitive dysfunction (18.7% vs 7.6%, p < 0.001), and fibromyalgia (19.2% vs 9.6%, p < 0.001).||Age, sex, race, branch of military, rank||Response rate 76%. Limited assessment of selection bias. Stratum random sample with proportional allocation—64 strata (GW, type of military, age, sex, race, rank, branch).|
|Oregon and Washington Veteran Studies|
|McCauley et al., 1999||Case-control study, survey and clinical examination, conducted in 1995–1998||GWV residents of Oregon and Washington in Operation Desert Shield or Desert Storm. Eligible 1,396 GWVs. Enrolled 158 GWVs with unexplained illness and 67 healthy GWV controls.||Unexplained health symptoms reported on mailed questionnaire and symptoms reported at clinical examination. Symptoms were not attributable to organic condition. Emphasis on disease misclassification in cross-sectional surveys.||Significant differences between symptoms reported on survey questionnaire and those confirmed at clinical examination. Self-reported fatigue was confirmed in 79% of participants, self-reported GI symptoms in 20%, and musculoskeletal pain in 35%.||Randomly selected population, reservists oversampled, all women selected. No nondeployed control group. Response rate 32.5% for clinical examination, 53.2% of cases, 51.5% of controls. Blinded clinicians reviewed case-control status.|
|Kansas Veteran Study|
|Steele, 2000||Cross-sectional survey, conducted by telephone in 1998||Kansas residents on active duty August 1990–July 1991, separated or retired from military, or currently serving in reserve component. Eligible 3,138; enrolled 1,548 GWVs and 482 NDVs. 37-item symptom questionnaire.||34.2% GWVs and 8.3% NDVs met Kansas criteria for Gulf War illness (OR = 4.68, 95% CI 3.25–6.75).
47.2% GWVs and 19.8% NDVs met CDC case criteria (OR = 3.26, 95% CI 2.48–4.28) Prevalence of Kansas-defined Gulf War illness lowest in GWVs serving on ships and highest in those in Iraq or Kuwait.
|Sex, age, income, education level, rank, service branch, component||92% response rate. No assessment of selection bias.|
|Canadian Veteran Study|
|Goss Gilroy Inc., 1998||Cross-sectional postal survey, conducted in 1997||All Canadian Gulf War veterans. Enrolled 3,113 GWVs and 3,439 veterans deployed elsewhere.||Survey of chronic conditions and symptoms.||GWVs reported higher prevalence of symptoms of chronic fatigue (OR = 5.27, 95% CI 3.95-7.03), cognitive dysfunction (OR = 4.36, 95% CI 3.80-5.01), MCS (OR = 4.01, 95% CI 2.43-6.62), major depression (OR = 3.67, 95% CI 3.04-4.44), anxiety (OR = 2.20, 95% CI 1.55-3.12), PTSD (OR = 2.69, 95% CI 1.69-4.26), chronic dysphoria (OR = 2.68, 95% CI 2.13-3.35), fibromyalgia (OR = 1.81, 95% CI 1.55-2.13), and respiratory diseases (OR = 1.35, 95% CI 1.16-1.57) than veterans deployed elsewhere (all p < 0.05).
47% of GWV and 74% of controls did not report any health outcome.
|Income and rank were important confounders; sex did not affect associations||Response rate 73% GWVs, 60% controls. 189 respondants were female. Controls were age-and sex-matched. Also compared with general population using 1990 Ontario Health Survey.|
|United Kingdom Veteran Studies|
|Unwin et al., 1999||Cross-sectional survey, conducted in 1997-1998||Stratified random sample of 53,462 UK military who served in gulf region, excluding special forces, compared with 39,217 personnel who served in||50-item symptom questionnaire, SF-36, GHQ.||25.3% of GWVs, 11.8% Bosnia, and 12.2% era cohort met CDC case criteria. GWVs reported higher rates of all 50 symptoms. Most frequent symptoms were feeling unrefreshed after sleep, irritability or outbursts of anger, headache, fatigue, sleeping difficulties, forgetfulness, joint||Age, smoking, alcohol consumption, marital status, education, rank, employment, civilian or military status||Response rates 70% GWVs, 63% Bosnia cohort, 62% era cohort. Authors note that 800 Bosnian veterans later moved to GWV group.|
|Bosnia in 19921997 and 250,000 era cohort in military but not gulf in 1991. Enrolled 2,961 GWVs, 2,620 Bosnia, and 2,614 NDVs.||stiffness, loss of concentration, flatulence or burping, and pain without swelling or redness in several joints. SF-36 significantly worse in GWVs' health perception but not physical functioning.||Extensive assessment of selection bias. Stratified by service, sex, age, service status, rank, fitness. Oversampled women.|
|Hotopf et al., 2003||Cross-sectional survey, conducted in 2000-2001||Followup of Unwin et al. (1999). Selected all Women; all male veterans with fatigue score >8; 50% sample of GWVs, with scores 4-8, all 4-8 in Bosnia and era; 1 in 8 sample of veterans with scores <4. Enrolled 1,089 GWVs, 638 Bosnia, and 634 NDVs.||Chalder Fatigue Scale, GHQ, SF-36, count of physical symptoms.||Modest reduction in fatigue since time 1, modest reduction in psychologic distress on GHQ since time 1, slight worsening on SF-36 since time 1. When compared with two groups of nondeployed veterans, SF-36 was worse, GHQ was worse, and fatigue was worse in GWVs, and they had 10.7 total symptoms in vs 7.9 and 6.4 in NDVs. No higher incidence of new illnesses.||Age, sex, rank, marital status||8,196 reported to first survey (Unwin et al., 1999).
Response rate 71.6%.
|Cherry et al., 2001||Cross-sectional survey by mailed questionnaire. Conducted 19971999||UK military who served in gulf region September 1990-June 1991, excluding special forces, compared||Health questionnaire surveying 95 symptoms (experienced in preceding month) and two mannequins on which to shade areas||On every symptom, score was higher for GWVs. Symptoms on which score was at least twice as high included difficulty in concentrating, poor memory, sudden changes in mood, feeling too weak to complete what you||Age, rank, still serving, marital status, sex||Sample stratified by sex, age, service, rank. Matched with randomly selected sample of NGVs. Three stratified|
|with nondeployed military as January 1, 1991. Total eligible 14,254; enrolled 4,076 GWVs, 4,135 GWVs validation sample, 4,749 NGVs.||of pain or numbness and tingling. Personal followup visits. Factor analysis and cluster analysis.||are doing, and feeling incapable of making decisions. In shading on mannequins, GWVs more likely to report symptoms consistent with peripheral neuropathy: 6.0% limited, 8.5% extended in GWVs vs 4.5% limited, 2.3% extended in NGVs. Widespread pain reported in 12.2% GWVs and 6.5% NGVs.||random samples: main GWVs, GWV validation sample, non-GWVs. Assessed selection bias. Relatively high response rate: 85.5%. No overlap with Unwin et al. (1999).|
|Maconochie et al., 2003; Simmons et al., 2004||Retrospective cohort study of veterans who believed they had GWS; assessed by mailed questionnaire in 1998–2001||All UK armed forces personnel. Enrolled 24,379 GWVs and 18,439 NDVs.||Incidence of self-reported adult ill health.||61% GWVs and 37% NDVs reported at least one new symptom since 1990. 85% of symptoms or diseases queried were higher in GWVs; strongest associations were in mood swings (OR = 20.9, 95% CI 16.2-27.0), memory loss or lack of concentration (OR = 19.6, 95% CI 15.5-24.8), night sweats (OR = 9.9, 95% CI 6.515.2), general fatigue (OR = 9.6, 95% CI 8.3-11.1), and sexual dysfunction (OR = 4.6, 95% CI 3.2-6.6).||Age, service, rank, serving status, alcohol consumption, smoking||47% response rate for men.
Analyses restricted to men only for statistical power.
|Danish Peacekeeper Studies|
|Ishoy et al., 1999||Cross-sectional study with questionnaire and medical examinations; health||Danish Gulf War veterans stationed in gulf August 2, 1990-December 31, 1997, UN peacekeeping force,||Questionnaire on neuropsychologic, gastrointestinal, and skin symptoms.||Deployed veterans reported higher prevalence (p < 0.05) of 17 of 22 neuropsychologic symptoms, 8 of 14 gastrointestinal symptoms, and 8 of 19 skin symptoms.||None||Response rates 83.6% GWVs 57.8% NGVs. Limited assessment of selection bias|
|examinations conducted in 1997-1998||officers, noncommissioned officers, and enlisted privates compared with Danish armed forces who could have been but had not been deployed in gulf.
Eligible 821 GWVs and 400 NDVs; enrolled 686 GWVs and 231 NDVs.
|81% deployed veterans and 71% controls had one or more ICD-10 diagnoses at examination (p = 0.002).||(most frequent reason for not participating was lack of time). Comparison group matched on sex, age, profession; randomly selected at end of 1996.|
|Australian Veteran Studies|
|Kelsall et al., 2004||Cross-sectional mailed questionnaire and comprehensive health assessment, conducted in 2000-2002||All Australian veterans who served in gulf August 2, 1990-September 4, 1991, compared with 26,411 Australian Defense Force personnel in operational units at the time but not deployed. Eligible 4,795; enrolled 1,456 GWVs and 1,588 NGVs.||63-item symptom questionnaire, medical-condition questionnaire, exposure questionnaire, functional impairment, 44-item military service experience regarding war stressors.||GWVs reported higher rates of all 63 symptoms (all but seven were statistically significant) and reported more severe symptoms. Top 5 symptoms were feeling unrefreshed after sleep (adjusted OR = 1.6, 95% CI 1.3-1.8), fatigue (adjusted OR = 1.6, 95% CI 1.3-1.8), headache (adjusted OR = 1.3, 95% CI 1.1-1.6), sleeping difficulties (adjusted OR = 1.6, 95% CI 1.4-1.9), and irritability or outbursts of anger (adjusted OR = 1.6, 95% CI 1.41.8).||Service type, rank, age, education, marital status||Assessed selection bias.
Used telephone survey-only results.
Response rates: 80.5% GWVs, 56.8% NDVs. Comparison group randomly selected and frequency-matched by service type, age, rank. Limited to men only.
|French Military Study|
|Salamon et al., 2006||Cross-sectional survey and clinical exam, conducted in 2002-2004||5,666 French GWVs.
Eligible 20,261; enrolled 5,666.
|Health symptoms and medical conditions, medical evaluation.||SSID self-reported by 10.9% of GWVs. Top five symptoms were headache (82.9%), sleeping difficulties (70.9%), irritability (68.8%), backache (62.9%), and memory difficulties (56.0%).||Response rate: 28%.
No nondeployed control group. 1,008 GWVs completed the clinical exam.
|Hawaii and Pennsylvania Active-Duty and Reserve Study|
|Stretch et al. 1995||Cross-sectional survey of service members from Hawaii and Pennsylvania||16,167 active-duty and reserve personnel assigned to all Army, Navy, Air Force and Marine Corps units in Hawaii and Pennsylvania. Enrolled 1,481 GWVs and 2,524 NDVs.||GWVs were significantly more likely than NDVs to report 12 symptoms: head cold, sinus trouble, sore throat, difficulty swallowing, headache, back problems, stomach upset, muscle aches, aching joints, cough, chills/fever, other problems (all p<0.001).||Age, rank, education, marital status, branch of military||Response rate: 31%.
No formal assessment of selection bias (speculated on reasons for nonresponses).
|Ft. Devens and New Orleans Cohort Studies|
|Proctor et al., 1998||Cross-sectional study of 3 cohorts followed longitudinally; dates of enrollment: Ft. Devens 1991, 1992-1993, 1994-1996; New Orleans 1991, 1994-1995;||Stratified random samples of two Gulf War-deployed groups: Ft. Devens and New Orleans compared with nondeployed group (Germany). Ft. Devens: US Army active, reserve, and National Guard||52-item symptom questionnaire.||GWVs report higher prevalence of all but one of 52 symptoms. Ft. Devens group reported significantly higher prevalences than Germany group. New Orleans group reported significantly higher prevalences of 24 of 52 symptoms. Among musculoskeletal symptoms reported more frequently by Ft. Devens-deployed veterans were joint pains (OR = 2.6) and neck||Age, sex, education||Germany group only studied at time 2.
300 completed Ft. Devens questionnaires were analyzed. Response rates: 53% Ft. Devens, 34% New Orleans, 49% Germany. Assessed selection
|Germany 1995||veterans; New Orleans: active, reserve, and National Guard US Army, Navy, Marine Corps, and Air Force troops; Germany: Maine National Guard air ambulance unit. Final participation 252 GWVs (Ft. Devens and New Orleans) and 48 NDVs (Germany).||ache or stiffness (OR = 2.7); among neurologic symptoms with greater prevalences in both cohorts of deployed veterans was headache (OR = 4.2); all confidence intervals excluded 1.0. About 30% of Gulf War veterans and 11% of Germany group reported inability to fall asleep (OR = 3.4-3.6, 95% CI excludes 1.0).||bias.
|Wolfe et al., 1998||Longitudinal study, cross-sectional survey conducted in 1993||Population included Proctor et al. (1998) sample. GWVs from Ft. Devens. Eligible 2,949; enrolled 2,119.||20-item health-symptom checklist||Most frequent symptoms were general aches and pains, overltiredness or lack of energy, headache, trouble in sleeping, nervous or tense, depressed mood, difficulty in concentrating;
30% of sample indicated that their physical health had become either “worse” or “much worse” since their return.
|Age, educational level, marital status, race||Response rate: 92%.
|Haley et al., 1997||Cross-sectional survey, factor-analysis survey conducted individually in supervised, in-person, group||Active and retired members of 24th Reserve Naval Mobile Construction Battalion, called to active duty in||Self-reported exposure to neurotoxic chemical combinations and association with factor-analysis–defined syndrome.||Survey results indicated 6 symptom factors (called syndromes by the authors): impaired cognition, confusion–ataxia, arthromyoneuropathy, phobia–apraxia, fever–adenopathy, and weakness–||No assessment of selection bias. Small cohort, no control group. Response rate: 41%.
|sessions in 1995||GW, residents of Alabama, Georgia, Tennessee, South Carolina, and North Carolina in November 1994. Eligible 606; enrolled 249.||incontinence. They accounted for 71% of observed variance. 63 (25%) veterans had one of the 6 “syndromes,” authors noted. 70% had serious health problems since returning from the Gulf War.||factor weights to assign veterans to “syndromes” with 1.5 cutoff.|
|Gray et al., 1999||Cross-sectional survey conducted in 1994||Active-duty Seabees in Navy In 1994 and serving at Port Hueneme, California, or Gulfport, Mississippi. Eligible 1,497; enrolled 527 GWVs and 970 NDVs.||Questionnaire on postwar symptoms; screening for chronic fatigue and PTSD; Hopkins Symptom Checklist for psychologic symptoms. Clinical evaluation: serum collection, handgrip strength, pulmonary-function testing.||55.8% of GWVs and 31.7% of NGVs reported postwar symptoms lasting ≥1 month. GWVs reported significantly (p < 0.05) higher prevalences of 35 out of 41 symptoms than NDVs. Both groups had similar clinical evaluations except GWVs had reduced hand grip strength and were more likely to have PTSD (15% vs 9%) than NDVs.||Assessed selection bias.
Response rate: varied by unit, 26–71%.
|Gray et al., 2002||Cross-sectional survey conducted in 1997-1999||All regular US Navy Seabees. Eligible 18,945; enrolled 3,831 GWVs, 4,933 veterans deployed elsewhere, and 3,104 NDVs.||Health questionnaire, working case definition.||GWVs reported poor general health, higher prevalences of 33 medical problems, and higher prevalences of CFS, PTSD, MCS, and IBS. 22% met criteria for Gulf War illness.||Age, sex, active-duty or reserve status, race or ethnicity, current smoking, current alcohol drinking||Study limited by recall bias. Response rate: 68.6%.
Large sample. Assessed selection bias with telephone survey.
|Pennsylvania Air National Guard Study|
|Fukuda et al., 1998||Cross-sectional survey conducted in 199.||Everyone on base when survey was conducted was||Survey of 35 symptoms, in-person interview. Clinical||GWV vs NDV: mild to moderate cases 39% vs 14%, severe cases 6% vs 0.7%. Veterans who met||Rank, sex, age, smoking status||Response rates: 61.6% index unit, 35.4% Unit A,|
|eligible. Index population 667 in ANG unit in Lebanon, PA. 3 Comparison populations: Unit A, 538 in ANG unit from PA with different mission; Unit B, 838 in US Air Force Reserve; Unit C, 1,680 active-duty Air Force from FL with missions similar to those of index.||evaluations of index unit only.||case definition had significantly diminished functioning and well-being.||73.4% Unit B, 69.8% Unit C. Deployed to Gulf: 47% index unit, 22% Unit A, 32% Unit B, 28% Unit C. Started as cluster investigation in Lebanon, PA. No assessment of selection bias.|
|Women in the Air Force|
|Pierce, 1997||Cross-sectional survey conducted in 1993 (Time 1) and 1995 (Time 2)||Stratified random sample of women in Air Force. Eligible 638; enrolled 525.||Health survey sex-specific health concerns.||At time 1, symptoms more frequent in GWVs vs NDVs were rash, cough, depression, unintentional weight loss, insomnia, and memory problems. At time 2 symptoms more frequent in GWVs vs NDVs were rash, cough, and memory problems. At time 1, no major difference between GWVs and NDVs in sex-specific symptoms. At time 2, most common symptoms in GWVs were lumps or cysts in breasts, abnormal PAP results, headache, and genital herpes.||Age||Response rates: 82% at Time 1, 92% at Time 2. 88,415 women in Air Force at the time.
Sample composition: 47% active duty, 25.5% reserve, 27.4% Guard.
|Joseph et al., 1997||Case series conducted in 1994-1996||Veterans seen by DOD for CCEP. Enrolled 20,000 GWVs.||Physical examination, medical and family history.||17.8% of GWVs had diagnoses of SSID, including primarily fatigue, headache, memory problems, and sleep disturbances||Registry study. Self-selected sample, no control group.|
|Kroenke et al., 1998||Case series conducted in 1994-1996||Veterans seen by DOD for CCEP Registry. Enrolled 18,495 GWVs.||Provider-administered symptom questionnaire.||Most common symptoms were joint pain (50%), fatigue (46.9%), headache (39.7%), memory or concentration difficulties (34%), sleep disturbance (33%), and rash (30.2%).
66% of symptoms did not appear until after Gulf War, and 40% of symptoms had a latency >1 year. Increased symptom counts were associated with loss of work.
|Registry study. Self-selected sample, no control group.|
|Roy et al., 1998||Case series conducted in 1994-1997||Veterans seen by DOD for Comprehensive Clinical Evaluation Program Registry. Enrolled 21,579 GWVs.||Health questionnaire, physical examination, medical history, laboratory studies.||17.2% of veterans had primary diagnosis of SSID. 41.8% had primary or secondary SSID diagnosis. Most common symptoms were fatigue, headache, sleep disturbance, and memory loss.||Registry study. Self-selected sample, no control group.|
|Escalante and Fischbach, 1998||Case series of rheumatologic referrals||GWVs enrolled in Persian Gulf Registry from South Texas Veterans Health Care System. Enrolled 145.||Symptoms, self-reported pain, and SF-36 for health-related quality of life.||Pain was present in almost all patients and was widely distributed. Widespread pain was reported in 65.1% of GWVs. Most frequent painful areas were knees (65%), low back (>60%), shoulders (50%), and hands and wrists (35%). Average values on SF-36 were||None||Registry study. Self-selected sample, no control group.
Dates of examinations not reported. Registry enrollment began in 1993.
|below 25th percentile of published national norms, with pain and nonarticular rheumatic symptoms explaining most of decrease in health-related quality of life.|
|Hallman et al., 2003||Cross-sectional assessment, conducted in 1995||GWVs residents of Delaware, Illinois, New Jersey, New York, North Carolina, Ohio, and Pennsylvania randomly sampled from VA Gulf War Health Registry. Enrolled 1,161.||Mailed survey||84.5% attributed their medical problems to service in gulf. Participants endorsed average of 25.5 symptoms: 9.9 mild, 9.5 moderate, and 6.1 severe.||Registry study. Self-selected sample, no control group.
Response rate: 60%.
Registry contained more than 70,000 veterans at time of study.
|Ozakinci et al., 2006||Cross-sectional assessment, 5-year followup of Hallman et al. (2003) conducted in 2000||Derivative of Hallman et al. (2003). Enrolled 390.||Mailed survey and telephone interview.||Compared with time 1, there was no significant change in number of symptoms reported or their severity. Subjects who were more symptomatic in 1995 showed some improvement but remained much more highly symptomatic than those who were less symptomatic.||Sex, rank, race, marital status, education, branch of service, duty||Response rate: 62%.|
|Coker et al., 1999; Lee et al., 2002, 2001||Cross-sectional, case series conducted in 1993-2001||First 3,000 British veterans attending GVMAP.||Medical and psychiatric diagnoses (ICD-10) and nonspecific health symptoms.||75% of first 3,000 GVMAP registrants assessed were well and symptom-free; 21% were well with symptoms but no disease.
Most common symptom groups reported were affective (45%), joint and muscle aches and pains (39%), and fatigue (38%).
|None||Registry study. Self-selected sample, no control group.|
|Of registrants assessed as unwell, 11% had psychiatric conditions, 5% organic medical conditions, and 9% both.|
NOTE: ANG = Air National Guard; CCEP = Comprehensive Clinical Evaluation Program; CDC = Centers for Disease Control and Prevention; CFS = chronic fatigue syndrome; CI = confidence interval; CMI = chronic multisymptom illness; DOD = Department of Defense; GHQ = Global Health Questionnaire; GI = gastrointestinal; GVMAP = Gulf Veterans Medical Assessment Programme; GWS = Gulf War syndrome; GWV = Gulf War veteran; IBS = irritable bowel syndrome; ICD-10 = International Statistical Classification of Diseases and Related Health Problems, 10th Revision; MCS = multiple chemical sensitivity; NDV = nondeployed veteran; OR = odds ratio; PTSD = posttraumatic stress disorder; RR = relative risk; SSID = signs, symptoms, and ill-defined conditions; VA = Department of Veterans’ Affairs
Bourdette, D. N., L. A. McCauley, A. Barkhuizen, W. Johnston, M. Wynn, S. K. Joos, D. Storzbach, T. Shuell, and D. Sticker. 2001. Symptom factor analysis, clinical findings, and functional status in a population-based case control study of Gulf War unexplained illness. Journal of Occupational and Environmental Medicine 43(12):1026-1040.
Cherry, N., F. Creed, A. Silman, G. Dunn, D. Baxter, J. Smedley, S. Taylor, and G. J. Macfarlane. 2001. Health and exposures of United Kingdom Gulf War veterans. Part II: The relation of health to exposure. Occupational & Environmental Medicine 58(5):299-306.
Coker, W. J., B. M. Bhatt, N. F. Blatchley, and J. T. Graham. 1999. Clinical findings for the first 1000 Gulf War veterans in the Ministry of Defence's Medical Assessment Programme. British Medical Journal 318(7179):290-294.
Delgado-Rodríguez, M., and J. Llorca. 2004. Bias. Journal of Epidemiology and Community Health 58:635-641.
Erickson, A. R., R. J. Enzenauer, V. J. Bray, and S. G. West. 1998. Musculoskeletal complaints in Persian Gulf War veterans. Journal of Clinical Rheumatology 4(4):181-185.
Escalante, A., and M. Fischbach. 1998. Musculoskeletal manifestations, pain, and quality of life in Persian Gulf War veterans referred for rheumatologic evaluation. Journal of Rheumatology 25(11):2228-2235.
Fukuda, K., R. Nisenbaum, G. Stewart, W. W. Thompson, L. Robin, R. M. Washko, D. L. Noah, D. H. Barrett, B. Randall, B. L. Herwaldt, A. C. Mawle, and W. C. Reeves. 1998. Chronic multisymptom illness affecting Air Force veterans of the Gulf War. Journal of the American Medical Association 280(11):981-988.
Goss Gilroy Inc. 1998. Health study of Canadian Forces Personnel Involved in the 1991 Conflict in the Persian Gulf. Ottawa, Canada: Goss Gilroy Inc. Department of National Defence. http://www.dnd.ca/site/reports/Health/vol1_TOC_e.htm Gray, G. C., K. S. Kaiser, A. W. Hawksworth, F. W. Hall, and E. Barrett-Connor. 1999. Increased postwar symptoms and psychological morbidity among U.S. Navy Gulf War veterans. American Journal of Tropical Medicine and Hygiene 60(5):758-766.
Gray, G. C., R. J. Reed, K. S. Kaiser, T. C. Smith, and V. M. Gastanaga. 2002. Self-reported symptoms and medical conditions among 11,868 Gulf War-era veterans. American Journal of Epidemiology 155(11):1033-1044.
Haley, R. W., T. L. Kurt, and J. Hom. 1997. Is there a Gulf War syndrome? Searching for syndromes by factor analysis of symptoms. Journal of the American Medical Association 277(3):215-222. [Erratum appears in Journal of the American Medical Association 1997;278(5):388.]
Hallman, W. K., H. M. Kipen, M. Diefenbach, K. Boyd, H. Kang, H. Leventhal, and D. Wartenberg. 2003. Symptom patterns among Gulf War registry veterans. American Journal of Public Health 93(4):624-630.
Hotopf, M., A. S. David, L. Hull, V. Nikalaou, C. Unwin, and S. Wessely. 2003. Gulf War illness - better, worse, or just the same? A cohort study. British Medical Journal 327(7428):1370-1372.
Hotopf, M., and S. Wessely. 2005. Can epidemiology clear the fog of war? Lessons from the 1990–91 Gulf War. International Journal of Epidemiology 34(4):791-800.
IOM (Institute of Medicine). 1997. Adequacy of the Comprehensive Clinical Evaluation Program: A Focused Assessment. Washington, DC: National Academy Press.
IOM. 1998. Adequacy of the VA Persian Gulf Registry and Uniform Case Assessment Protocol. Washington, DC: National Academy Press.
IOM. 1999. Strategies to Protect the Health of Deployed US Forces: Medical Surveillance, Record Keeping, and Risk Reduction. Washington, DC: National Academy Press.
IOM. 2006. Gulf War and Health, Volume 4: Health Effects of Serving in the Gulf War. Washington, DC: The National Academies Press.
IOM. 2010. Gulf War and Health, Volume 8: Update of Health Effects of Serving in the Gulf War. Washington, DC: The National Academies Press.
Iowa Persian Gulf Study Group. 1997. Self-reported illness and health status among Gulf War veterans: A population-based study. Journal of the American Medical Association 277(3):238-245
Ishøy, T., P. Suadicani, B. Guldager, M. Appleyard, and F. Gyntelberg. 1999a. Risk factors for gastrointestinal symptoms. The Danish Gulf War study. Danish Medical Bulletin 46(5):420-423.
Ishøy, T., P. Suadicani, B. Guldager, M. Appleyard, H. O. Hein, and F. Gyntelberg. 1999b. State of health after deployment in the Persian Gulf. The Danish Gulf War study. Danish Medical Bulletin 46(5):416-419.
Joseph, S. C., R. R. Blanck, G. Gackstetter, R. Glaser, K. C. Hyams, S. Kinty, C. Magruder, J. Mazzuchi, F. L. O'Donnell, M. D. Parkinson, R. E. Patterson, and D. H. Trump. 1997. A comprehensive clinical evaluation of 20,000 Persian Gulf War veterans. Military Medicine 162(3):149-155.
Kang, H. K., B. Li, C. M. Mahan, S. A. Eisen, and C. C. Engel. 2009. Health of US Veterans of 1991 Gulf War: A follow-up survey in 10 years. Journal of Occupational & Environmental Medicine 51(4):401-410.
Kang, H. K., C. M. Mahan, L. Y. Lee, C. A. Magee, and F. M. Murphy. 2000. Illnesses among United States veterans of the Gulf War: A population-based survey of 30,000 veterans. Journal of Occupational and Environmental Medicine 42(5):491-501.
Kang, H. K., C. M. Mahan, K. Y. Lee, F. M. Murphy, S. J. Simmens, H. A. Young, and P. H. Levine. 2002. Evidence for a deployment-related Gulf War syndrome by factor analysis. Archives of Environmental Health 57(1):61-68.
Kelsall, H., R. Macdonell, M. Sim, A. Forbes, D. McKenzie, D. Glass, J. Ikin, and P. Ittak. 2005. Neurological status of Australian veterans of the 1991 Gulf War and the effect of medical and chemical exposures. International Journal of Epidemiology 34(4):810-819.
Kelsall, H. L., M. R. Sim, A. B. Forbes, D. C. Glass, D. P. McKenzie, J. F. Ikin, M. J. Abramson, L. Blizzard, and P. Ittak. 2004. Symptoms and medical conditions in Australian veterans of the 1991 Gulf War: Relation to immunisations and other Gulf War exposures. Occupational and Environmental Medicine 61(12):1006-1013.
Kroenke, K., P. Koslowe, and M. Roy. 1998. Symptoms in 18,495 Persian Gulf War veterans: Latency of onset and lack of association with self-reported exposures. Journal of Occupational and Environmental Medicine 40(6):520-528.
Lee, H. A., R. Gabriel, A. J. Bale, P. Bolton, and N. F. Blatchley. 2001. Clinical findings of the second 1000 UK Gulf War veterans who attended the Ministry of Defence's Medical Assessment programme. Journal of the Royal Army Medical Corps 147(2):153-160. [Erratum appears in Journal of the Royal Army Medical Corps 2001;147(3):260]
Lee, H. A., R. Gabriel, J. P. G. Bolton, A. J. Bale, and M. Jackson. 2002. Health status and clinical diagnoses of 3000 UK Gulf War veterans. Royal Society of Medicine (Great Britain). Journal of the Royal Society of Medicine 95(10):491-497.
Levenson, L. J., C. Colenda, D. B. Larson, and J. C. Bareta. 1990. Methodology in consultation-liaison research: A classification of biases. Psychosomatics 31(4):367-375.
Maconochie, N., P. Doyle, G. Davies, S. Lewis, M. Pelerin, S. Prior, and P. Sampson. 2003. The study of reproductive outcome and the health of offspring of UK veterans of the Gulf War: Methods and description of the study population. BMC Public Health 3:4.
McCauley, L. A., S. K. Joos, P. S. Spencer, M. Lasarev, and T. Shuell. 1999. Strategies to assess validity of self-reported exposures during the Persian Gulf War. Environmental Research 81(3):195-205.
Murphy, D., R. Hooper, C. French, M. Jones, R. Rona, and S. Wessely. 2006. Is the increased reporting of symptomatic ill health in Gulf War veterans related to how one asks the question? Journal of Psychosomatic Research 61(2):181-186.
Murphy, D., M. Hotopf, and S. Wessely. 2008. Multiple vaccinations, health, and recall bias within UK armed forces deployed to Iraq: cohort study. BMJ 337:a220.
Ozakinci, G., W. K. Hallman, and H. M. Kipen. 2006. Persistence of symptoms in veterans of the first Gulf War: 5-year follow-up. Environmental Health Perspectives 114(10):1553-1557.
Pearce, N., H. Checkoway, and Kriebel D. 2007. Bias in occupational epidemiology studies. Occupational Environmental Medicine 64:562-568.
Pierce, P. F. 1997. Physical and emotional health of Gulf War veteran women. Aviation Space and Environmental Medicine 68(4):317-321.
Proctor, S. P., T. Heeren, R. F. White, J. Wolfe, M. S. Borgos, J. D. Davis, L. Pepper, R. Clapp, P. B. Sutker, J. J. Vasterling, and D. Ozonoff. 1998. Health status of Persian Gulf War veterans: Self-reported symptoms, environmental exposures and the effect of stress. International Journal of Epidemiology 27(6):1000-1010.
Roy, M. J., P. A. Koslowe, K. Kroenke, and C. Magruder. 1998. Signs, symptoms, and ill-defined conditions in Persian Gulf War veterans: Findings from the comprehensive clinical evaluation program. Psychosomatic Medicine 60(6):663-668.
Salamon, R., C. Verret, M. A. Jutand, M. Begassat, F. Laoudj, F. Conso, and P. Brochard. 2006. Health consequences of the first Persian Gulf War on French troops. International Journal of Epidemiology 35(2):479-487.
Simmons, R., N. Maconochie, and P. Doyle. 2004. Self-reported ill health in male UK Gulf War veterans: A retrospective cohort study. BMC Public Health 4:1-10.
Steele, L. 2000. Prevalence and patterns of Gulf War illness in Kansas veterans: Association of symptoms with characteristics of person, place, and time of military service. American Journal of Epidemiology 152(10):992-1002.
Stretch, R. H., P. D. Bliese, D. H. Marlowe, K. M. Wright, K. H. Knudson, and C. H. Hoover. 1995. Physical health symptomatology of Gulf War-era service personnel from the states of Pennsylvania and Hawaii. Military Medicine 160(3):131-136.
Szklo, M. 1998. Population-based cohort studies. Epidemiologic Reviews 20(1):81-90.
Unwin, C., N. Blatchley, W. Coker, S. Ferry, M. Hotopf, L. Hull, K. Ismail, I. Palmer, A. David, and S. Wessely. 1999. Health of UK servicemen who served in Persian Gulf War. Lancet 353(9148):169-178.
Unwin, C., M. Hotopf, L. Hull, K. Ismail, A. David, and S. Wessely. 2002. Women in the Persian Gulf: Lack of gender differences in long-term health effects of service in United Kingdom Armed Forces in the 1991 Persian Gulf War. Military Medicine 167(5):406-413.
Wolfe, J., S. P. Proctor, J. D. Davis, M. S. Borgos, and M. J. Friedman. 1998. Health symptoms reported by Persian Gulf War veterans two years after return. American Journal of Industrial Medicine 33(2):104-113.