Toxicological assessments of Gulf War veterans

(a) Early environmental health concerns

The relatively short 1991 Gulf War was seen as highly successful. Iraqi forces were quickly routed from Kuwait during a 4-day ground war with low casualties among Coalition troops. However, as reports emerged in 1992 and 1993 of increasing health problems, many Gulf War veterans and their supporters expressed concerns that their illnesses were the result of exposure to one or more hazardous agents during that conflict. Specific concerns included the first war-time use of special depleted uranium (DU) armour-piercing munitions, vaccines and other drugs used to protect troops, deliberate or accidental exposure to chemical warfare agents and smoke from oil-well fires ignited by retreating Iraqi troops.

In 1993, VA created a special voluntary Gulf War clinical registry, which by July 2005 had evaluated about 96 000 (about 14%) Gulf War veterans in this clinical surveillance programme. Principal findings were that these veterans are suffering from a wide variety of recognized illnesses, but no new or unique syndrome was identified (Combined Analysis of the VA and DoD Gulf War Clinical Evaluation Programs 2002). As the lead federal agency on Gulf War-related research, VA was responsible also for coordinating federally sponsored epidemiological and other relevant scientific studies in the USA. As of today, this coordinated approach has spent more than $250 million for more than 200 Gulf War health-related research projects. However, as noted in the 2003 MRC report, ‘research studies everywhere had been hampered by a lack of baseline exposure data on which to base trends, and by the difficulty of retrospectively verifying self-reported exposure data, which are sensitive to bias, and vary and become less reliable over time.’

Because most VA benefits depend in part upon establishing a connection to military service, growing environmental health concerns put pressure on VA to determine which Gulf War hazards had caused specific illness and disease among these new veterans. Establishing service connection usually requires a simple medical review in the case of a veteran with an injury clearly related to military service, e.g. from a bullet or shrapnel wound. Establishing service connection can be much more contentious when the illness or injury involves a chronic disease possibly caused by environmental or occupational exposure that may have occurred years or decades in the past. For example, it can be difficult for a veteran to provide evidence that their leukaemia was specifically caused by exposure to benzene during their military service, in comparison to all other possible causes that they may have been exposed to over their lifetimes.

(a) Formal veteran's health advisory role

The NAS was established in 1863 with the signature of President Abraham Lincoln, to ‘investigate, examine, experiment and report upon any subject of science or art’ for agencies in the federal government. In 1970, the NAS created the IOM to provide independent, objective, authoritative, credible and timely scientific analyses on medical and health issues. The US Congress through US government agencies regularly seeks the IOM's unique scientific advice on a broad range of health-policy issues. The IOM does not compete for funding in response to federal requests for proposals.

The IOM has developed a highly formalized process designed to provide authoritative, objective and scientifically balanced answers to difficult public health questions of national importance. Their studies are conducted by independent committees of volunteer scientists composed of leading nationally and internationally recognized experts, selected by the IOM based on their expertise, good judgment and freedom from conflict of interests. Committee members serve as individuals, not as representatives of organizations or interest groups. Although some committee meetings may be opened to the public, committee findings and conclusions are based upon their review of the relevant scientific and medical literature, with a primarily focus on peer-reviewed materials. The IOM acknowledges that although peer review ensures high standards of quality, it does not guarantee the validity of a study or the ability to generalize its results. Therefore, committee members are expected to critically read each publication for relevance and quality before reaching conclusions.

Recognizing that even scientific subject matter experts inevitably bring their own biases and experiences, the IOM has specific procedures to ensure appropriate committee membership balance and to avoid real or perceived conflicts of interest. For example, a committee member would virtually never be put in the position of reviewing his or her own research—a safeguard that has been absent from the work of other ‘expert’ committees. Committee members are required to itemize in writing all professional, consulting and financial connections, as well as pertinent intellectual positions and any relevant public statements. Because potential committee members with professed bias, prejudice or significant preconceptions on the subject at hand are carefully excluded, IOM committee findings are usually considered to be independent and highly credible.

The IOM requires that a committee's formal findings and recommendations are evidence-based whenever possible and noted as only expert opinion when that is not possible, making them similar to ‘Cochrane Reviews’ in depending on a predetermined methodology and standards of evidence, albeit with a special focus on military and veteran health issues. Each IOM report undergoes extensive formal internal and peer review by a group of external experts who are anonymous to the committee, and whose names are revealed only once the study is published.

(b) IOM veterans' health reviews

The IOM's highly developed formal review process has proven invaluable to VA for establishing fair, scientifically based disability policies for veterans. Their unsurpassed reputation for objectivity, scientific integrity and independence means that their reports stand as authoritative even when their findings fail to please all stakeholders. By law and historical precedent, VA depends upon formal IOM reviews to help establish compensation policies for veterans exposed to Agent Orange, other herbicides and dioxin used in the Vietnam War, for veterans experimentally exposed to chemical warfare agents prior to 1975, and for veterans exposed to a wide variety of environmental hazards during the 1991 Gulf War. The IOM is also often requested to examine a wide range of general clinical issues, and since 1991, IOM has completed 17 independent reviews of Gulf War health issues (table 1).

Since IOM's Gulf War environmental health studies are designed to inform veteran disability compensation policies and to identify specific human diseases that may be associated with military environmental hazards, they primarily focus upon human studies, supplemented by a thorough review of all relevant animal studies. They concentrate on scientific literature documenting long-term chronic and delayed health effects that are most relevant to potential disabilities. Their conclusions are applicable to specific exposures and diseases that may occur among military veterans. For example, the IOM's finding that dioxin exposure is not only a carcinogen, but is specifically associated with Hodgkin's disease led VA to establish a compensation policy specifically for veterans exposed to dioxin and who develop Hodgkin's disease.

Since the end of the 1991 Gulf War, at least 14 different committees have been established, both in the United States and the United Kingdom, to help evaluate Gulf War veteran health issues. However, the IOM's reputation for scientific rigour, independence from the political process, and freedom from bias has made it one of the most influential sources.

The UK MRC and Royal Society also achieve such an international stature and credibility, and together these sources have almost certainly had the greatest impact on our understanding of health issues among veterans of the 1991 Gulf War. For example, the 2003 MRC report included an insightful comparison of UK research findings with related studies conducted in other nations involved in the 1991 Gulf War: they noted, ‘There is a notably similar pattern of findings irrespective of where Gulf War veterans came from, or what their Gulf service experiences were. Gulf War veterans from several coalition countries consistently report suffering from more symptoms than non-Gulf War veterans, and these symptoms are similar despite markedly different exposures to vaccination, nerve agent pretreatments, oil fire smoke and other potential hazards.’ Their bottom line conclusion, consistent with findings from the IOM, is that ‘Increasingly detailed medical investigations have detected very few actual abnormalities and no consistent associations. In short, there is no evidence from UK or international research for a single syndrome related specifically to service in the Gulf.’

Other committees (and date of publications) include:

1. Armed Forces Epidemiological Board (1996, 1999, 2000a,b), US Department of Defense (1999).

2. Goss Gilroy Inc. Canadian epidemiological study of Gulf War veterans (1998a).

3. The Rt Hon The Lord Lloyd of Berwick. Independent public inquiry on gulf war illnesses (2004).

4. US Department of Veterans Affairs, Research Advisory Committee on Gulf War Veterans Illnesses, James Binns, Chair. Scientific progress in understanding Gulf War veterans' illnesses: report and recommendations (2004a,b).

5. US Department of Defense Special Oversight Board for Department of Defense Investigations of Gulf War Chemical and Biological Incidents. Final Report (2000).

6. US Department of Defense. Report of the Defense Science Board Task Force on Persian Gulf War health effects (1994).

7. US Department of Health & Human Services, National Institutes of Health Technology Assessment Workshop Panel. The Persian Gulf experience and health (1994)

8. US Government Accountability Office (GAO). Gulf War illnesses: DOD's conclusions about US troops' exposure cannot be adequately supported (2004).

9. US Presidential Advisory Committee on Gulf War veterans' illnesses: interim report (1996a).

10. US Presidential Advisory Committee on Gulf War veterans' illnesses: final report (1996b).

11. US Presidential Advisory Committee on Gulf War veterans' illnesses: special report (1997).

12. United Kingdom Parliamentary Office of Science and Technology. Gulf war illnesses: dealing with the uncertainties (1997).

13. United States Senate, Committee on Veterans' Affairs, Report of the special investigation unit on Gulf War illnesses (1998).

(c) Congressionally mandated studies on Gulf War veterans' health

Recognizing the IOM's potential contribution, in 1998, Congress directed in two statutes (Public Laws 105-277 and 105-368) that VA request the IOM to conduct reviews of the scientific and medical literature on the long-term health effects from exposure to a wide range of environmental hazards related to the 1991 Gulf War. Congress modelled these statutes after the earlier Agent Orange Act of 1991 described earlier, which established a similar review process to evaluate possible health effects among Vietnam veterans from exposure to Agent Orange, other herbicides and their impurities. The two statutes also listed nearly three dozen different specific environmental exposures or categories of exposures of concern (table 2). Congress directed the NAS-IOM to ‘identify the biological, chemical, or other toxic agents, environmental or wartime hazards, or preventive medicines or vaccines to which members of the Armed Forces who served in the Southwest Asia theatre of operations during the Persian Gulf War may have been exposed by reason of such service.’

Based on these statutes, the IOM has produced three major reports and one update for VA on health effects from a variety of hazardous exposures associated with the 1991 Gulf War (table 3).

(d) Evidence categories

The IOM has established a standardized method for rating the quality of published scientific and medical evidence associating a specific environmental hazard to a specific illness or disease among veterans. These are essentially identical to those used to evaluate Agent Orange health effects.

Their default category of association is ‘inadequate/insufficient evidence to determine whether an association exists.’ That is, an IOM committee begins with a neutral assumption that the ‘evidence is of insufficient quantity, quality, or consistency to permit a conclusion regarding the existence of an association between exposure to a specific agent and a specific health outcome in humans.’ Of course, the purpose of the IOM review is to search for evidence that can change this default. They have three categories for a positive association between an environmental hazard and health effect: (1) ‘sufficient evidence of an association,’ (2) ‘sufficient evidence of a causal association’ and the weakest positive category (3) ‘limited/suggestive evidence of an association.’ There is also one category for a negative association as ‘limited/suggestive evidence of no association.’

It is important to note that a possible association between a specific hazard and health outcome may remain in the default category of ‘insufficient/inadequate evidence of an association’ either because (i) there is simply too little evidence to evaluate, or (ii) because a substantial body of evidence does not allow any more definitive conclusion.

(a) Depleted uranium

DU, a by-product of the uranium enrichment process, is a low-level radioactive element as well as a heavy metal. About 340 tons of DU were used in munitions during the 1991 Gulf War. Used for both vehicle armour and armour-piercing munitions, DU primarily emits alpha radiation, which has a very short range and little penetrating power, making it hazardous only when in close proximity to tissue, e.g. via ingestion or inhalation. Uranium is also considered to have toxicity as a heavy metal, with the kidney as the vulnerable organ. DU munitions may release significant amounts of DU particles in various forms upon impact, which may be inhaled. DU munitions also produce fragments that leave DU shrapnel in wounds. Exposure to DU during the 1991 Gulf War could have occurred from ‘friendly fire’ incidents, cleanup operations and accidents, via inhalation and ingestion of particles and retention of DU shrapnel. Long-term health effects from such exposures have led to significant public controversy in the United Kingdom and the United States. The topic of DU exposure and health is considered in more detail elsewhere in Squibb & McDiarmid (2006)

The IOM committee focused primarily on 12 epidemiological studies of uranium processing workers (milling and processing uranium primarily for nuclear weapons and nuclear reactors; table 5). Most were mortality studies of uranium millers, and workers at uranium enrichment facilities, nuclear fuel fabricators and workers in a phosphate fertilizer plant (processing phosphate ore can result in uranium exposure). Exposure of uranium millers came primarily from dust inhalation, the most relevant route for exposed Gulf War veterans. Many of the workers evaluated in these occupational studies would have experienced nearly daily exposures over much of their working lifetime, and adjustment for cigarette smoking was a common problem. Studies of miners were also examined but were considered less relevant as they focused primarily on radon health effects.

Based upon their analysis of these studies the IOM committee reached three major conclusions regarding the association between exposure to DU and illnesses.

1. Lung cancer. The committee concluded that there is limited/suggestive evidence of no association between exposure to uranium at cumulative internal dose levels lower than 200 mSv or 25 cGy. However, they reported inadequate/insufficient evidence to determine whether an association does or does not exist between exposure to uranium and lung cancer at higher levels of cumulative exposure.

2. Renal function. They concluded that there is limited/suggestive evidence of no association between exposure to uranium and clinically significant renal dysfunction. That is, the kidney does not appear to be a significant target in humans exposed to uranium.

3. Other health outcomes. The committee concluded that there is inadequate/insufficient evidence to determine whether an association does or does not exist between exposure to uranium and the following health outcomes: lymphatic cancer; bone cancer; nervous system disease; non-malignant respiratory disease; or other health outcomes (gastrointestinal disease, immune-mediated disease, effects on haematological parameters, reproductive or developmental dysfunction, genotoxic effects, cardiovascular effects, hepatic disease, dermal effects, ocular effects and musculoskeletal effects).

At about the same time, in the United Kingdom, the Royal Society responded by convening an independent expert working group, under the chairmanship of Sir Brian Heap, to review scientific literature on long-term health effects from DU exposure. That expert committee also was charged with considering whether exposure to DU could be a cause of the ill health seen in UK service personnel.

The two Royal Society ‘working group’ reports and one summary on DU reviewed the very extensive relevant published epidemiologic and animal research (The Royal Society 2001, 2002a,b). Like the IOM committee, they concentrated upon peer-reviewed materials. Their reports were reviewed and endorsed by the Council of the Royal Society. The working group also heard from a broad range of stakeholders with an interest in DU and its potential consequences.

Using broadly similar approaches, the Royal Society working group reached similar conclusions to those of the IOM committee on this issue. However, the Royal Society took the additional step of attempting to estimate actual DU exposure and the corresponding radiological and toxicological health risks experienced by veterans. As with essentially all risk factors associated with the 1991 Gulf War, little or no actual measurements are available of DU exposure, and exposure therefore must be estimated. The Royal Society estimated battlefield exposure levels based upon a range of scenarios, including a worst-case scenario considered as ‘as not likely to have been exceeded.’ They concluded that regardless of uncertainties, they could establish reasonable upper and lower limits on battlefield DU exposure and the subsequent health effects from that exposure.

Based on this approach, the expert working group concluded that the greatest risk from DU inhaled on the battlefield related to radiation was an increase in lung cancer risk (The Royal Society 2001, 2002b). Estimated average risk of other cancers, including leukaemia, were lower than those for lung cancer for all DU exposure scenarios. Under worst-case exposure estimates, excess lung cancer risk ‘could be about one in 15.’ For a large exposure, such as surviving in a tank struck with a DU munition, the working group estimated an average excess lung cancer risk of about one in a thousand. For smaller exposure, the excess lung cancer risk was about one in 40 000, or less. These figures compare to lifetime general population lung cancer risk of about one in 250 for non-smokers and one in six for cigarette smokers.

For toxicological health effects, the working group noted that the few human studies involving large uranium exposure result in limited information about serious adverse effects in kidney, the likely target organ in humans. ‘Very few humans have had sufficiently large acute intakes of uranium compounds to lead to severe kidney dysfunction or kidney failure. Studies of these few cases indicate that kidney failure is likely to occur within a few days at concentrations above about 50 μg uranium per gram kidney.’

Based on these observations and estimated battlefield DU exposures, the working group concluded that ‘it is not expected that adverse effects on the kidney would occur. Levels of uranium in the kidney of soldiers surviving in tanks struck by DU rounds, or of soldiers working for protracted periods in heavily contaminated vehicles, could lead to some short-term kidney dysfunction, but whether this would lead to any long-term adverse effects is unclear. According to worst-case assumptions, kidney uranium levels in some soldiers could be very high, and would probably lead to kidney failure within a few days of exposure, although we are unaware of any such cases of kidney failure' (The Royal Society 2002a).

Using their battlefield DU exposure estimates and review of relevant scientific and medical literature, the working group reached the following main conclusions about the increased risk to veterans of the 1991 Gulf War:

1. Except in extreme circumstances, any extra risks of developing fatal cancers as a result of radiation from internal exposure to DU arising from battlefield conditions are likely to be undetectable above the general risk of dying from cancer over a normal lifetime. This remains true even if the estimates of risk resulting from likely exposures are one hundred times too low.

2. The extreme circumstances will apply only to a very small fraction of the soldiers in a theatre of war, for example those who survive in a vehicle struck by a DU penetrator, or those involved in cleaning up struck vehicles. In such circumstances, and assuming the most unfavourable conditions, the lifetime risk of death from lung cancer could be about twice that in the general population.

3. Any extra risks of death from leukaemia, or other cancers, as a result of exposure to DU are estimated to be substantially lower than the risks of death from lung cancer. Under all likely exposure scenarios, the extra lifetime risks of fatal leukaemia are predicted to be too small to be observable.

4. Many soldiers on a battlefield may be exposed to small amounts of DU and the risks of cancer from such exposures are predicted to be very low. Even if the estimates of risk for these conditions are one hundred times too low, it is unlikely that any excess of fatal cancer would be detected within a cohort of 10 000 soldiers followed over 50 years.

5. The radiological risks from the use of DU in munitions are for the most part low, but there are uncertainties in the levels of exposure that could occur under unfavourable conditions, and for small numbers of soldiers there could be circumstances in which the excess risks of lung cancer are substantial. Therefore, they recommended that further work be undertaken to clarify the extent of intakes on the battlefield.

6. The estimated DU intakes for most soldiers on the battlefield are not expected to result in concentrations of DU in the kidney that exceed 0.1 μg per gram of kidney, even transiently. Consequently, in these cases, it is not expected that adverse effects on the kidney or any other organ would occur.

7. Levels of uranium in the kidney of soldiers surviving in tanks struck by DU rounds, or of soldiers working for protracted periods in heavily contaminated vehicles, could reach concentrations that lead to some short-term kidney dysfunction, but whether this would lead to any long-term adverse effects is unclear as adequate studies of the long-term effects on the kidney of acute or protracted exposures to elevated levels of uranium are not available. According to worst-case assumptions, kidney uranium levels in some soldiers could be very high, and would probably lead to kidney failure within a few days of exposure. However, they were not aware of any cases of kidney failure, occurring within a few days of exposure, in US soldiers who would have received the highest DU intakes during the Gulf War, but they could not rule out some kidney damage for such soldiers under worst-case assumptions.

8. For those returning to live in areas where DU munitions were deployed, including peace-keepers, the inhalation intakes from re-suspended DU are considered to be unlikely to cause any substantial increase in lung cancer or any other cancers. The estimated excess lifetime risk of fatal lung cancer is about one in a million, although there could be higher risks for some individuals with worst-case intakes of DU due to higher levels of local contamination. Estimated risks of other cancers are at least 100-fold lower. There are, however, large uncertainties in the estimates of inhalation intakes in the years following a conflict.

9. No effects on kidney function from inhalation of re-suspended DU are expected for most individuals who return after a conflict. Small effects on kidney function are possible using worst-case assumptions, but would at most only apply to a small number of individuals.

10. Although there is no clear evidence that occupational exposures to uranium have consequences for reproductive health, effects on reproductive health have been observed in mice after high intakes of uranium. Accordingly, epidemiological studies of the reproductive health of Gulf War veterans and of the Iraqi population are underway. If effects are seen, then further investigation would be required to determine the relative contributions from DU and from other possible causes.

Similarly, citing the Royal Society report, the 2003 MRC reported concluded that ‘although animal studies indicate that sufficiently large short-term doses of uranium can cause kidney damage, the very limited human data available from uranium miners and accidental exposures, suggest that the risk of acute poisoning in humans is not great.’

(b) Sarin

Health concerns from exposure to low levels of sarin became a significant issue for veterans of the 1991 Gulf War following revelations that some Iraqi munitions destroyed by US forces at Khamisiyah, Iraq, after the March 1991 cease-fire contained this chemical warfare agent. To evaluate short- and long-term health effects from organophosphorus (OP) military and insecticidal nerve agents, reviewers including IOM committees and the MRC have an abundant scientific literature to consider. This literature, published over more than five decades, includes data from actual human experimentation, from occupational and accidental exposures, laboratory animals and from terrorist attacks including the 1994 and 1995 sarin attacks in Japan.

Four distinct health effects have been described, including (i) acute cholinergic toxicity, (ii) organophosphate-induced delayed neuropathy (OPIDN), (iii) subtle long-term neuropsychological and neurophysiological effects and (iv) a reversible muscular weakness known as ‘intermediate syndrome.’ Some effects described in this literature are subtle, often difficult to differentiate from health effects caused by other diseases or occupational exposures.

Consistent with general toxicological principles, each described health effect has data suggesting threshold exposure levels below which it is unlikely to be clinically detectable. Therefore, meaningful interpretation of human and animal studies requires strict characterization of the exposure.

Because precise exposure levels are often difficult to reconstruct, a system for characterizing exposure has been proposed based upon observed initial acute signs and symptoms, as high-level (definitive cholinergic poisoning); intermediate-level (threshold cholinergic effects including miosis, rhinorrhoea or clinically measurable depression of cholinesterase); and low-level (no immediate clinical signs or symptoms) exposure (Brown & Brix 1998).

A basic conclusion of reviews by the IOM and others is that there exist threshold or minimum exposure levels for all known long-term effects from OP nerve agent, and that those thresholds are at or above intermediate level exposure. Long-term health effects seen at intermediate-level exposures or in many survivors of high-level exposure are subtle, detectable in exposed populations but not individuals, and not reported in individuals experiencing low-level exposure alone. From a practical perspective, this means that long-term health effects from these agents are seen only in the aftermath of acute clinical poisoning, or to put it another way, such long-term effects are unlikely in the absence of evidence of acute signs and symptoms of poisoning.

Two IOM committees (2000 and 2004) reached several essentially identical conclusions about long-term health effects from sarin for different exposure levels, as defined by the magnitude of initial acute poisoning signs and symptoms. As the IOM committee pointed out, medical reports by the US Army Medical Corps at the time of the Khamisiyah incident were consistent with the absence of any acute sarin poisoning: ‘US troops did not report acute cholinergic symptoms at the time, but the possibility of low-level, asymptomatic exposures cannot be discounted.’ Therefore, the key health issue for veterans of the 1991 Gulf War was the possibility of long-term health effects following sub-clinical exposure to sarin, that is, following an exposure causing no acute poisoning signs and symptoms.

A 2003 UK MRC report also noted that subtle effects following acute OP poisoning have been reported, but that no evidence for these effects has been reported following low-level non-acute exposures. They concluded that, ‘The evidence linking UK Gulf War veterans' illnesses to actual measures of nerve gas or OP pesticide exposure is extremely poor. There is no confirmed evidence that nerve gases were used in the 1991 Gulf conflict. OP pesticides used by troops stationed in the Gulf were the only likely OP hazard. However, there is little information on the quantity of OP pesticides people handled, or how these were used. No cases of acute OP poisoning were reported at the time, making it unlikely that exposure levels were ever high enough to account for the kinds of symptoms experienced later.’

The MRC report also considered whether some individuals might be uniquely sensitive to OP poisoning as a result of variations reported in the enzyme paraoxonase among some ill Gulf War veterans, which break down these chemicals. However, the MRC concluded that the significance of these findings was not clear and would require further study.

The 2003 MRC report also concluded that although ill veterans of the 1991 Gulf War consistently report an unusually high level of neurological symptoms, clinical investigations had not uncovered significant evidence of malfunction in the peripheral nervous system or nerve/muscle junctions.

The IOM committee reviewed a number of human studies on sarin health effects, as well as a wide range of laboratory animal studies. They focused upon studies of four human populations exposed to sarin, including military volunteers in experiments conducted in the US and UK several decades ago involving exposure to sarin and other chemical warfare agents, industrial workers involved in the manufacture of sarin and victims of terrorist attacks in Matusumoto City and Tokyo, Japan, in 1994 and 1995 (table 6).

The 2000 IOM committee also reviewed the hypothesis that exposure to sub-clinical traces of sarin might produce a new, previously undescribed disease—a ‘Gulf War Syndrome’ (Haley & Kurt 1997). Supporters of this hypothesis presented evidence to the committee during an open two-day meeting 16 and 17 September 1999, in Washington, DC (appendix B, ‘Gulf War and health: volume 1. Depleted uranium, pyridostigmine bromide, sarin, and vaccines.’ National Academy Press, 2000). Nevertheless, the committee did not endorse this hypothesis, and, in fact, their most important conclusion relative to Gulf War health effects was that there was ‘inadequate/insufficient evidence of an association’ between exposure to sub-clinical levels of sarin and any subsequent long-term health effects.

Not surprisingly, considering the fact that these chemical agents are designed to kill or incapacitate, the committee also found ‘sufficient evidence of a causal relationship’ between exposure to sarin and a dose-dependent acute cholinergic syndrome that is evident seconds to hours subsequent to sarin exposure and resolves in days to months.

The IOM committee also reported ‘limited/suggestive evidence of an association’ between exposure to sarin at doses sufficient to cause acute cholinergic signs and symptoms and subsequent long-term health effects, based primarily on studies of three groups of people exposed to sarin: (i) workers occupationally exposed to sarin in the 1950s and 1960s; (ii) a terrorist attack on civilians in Matsumoto, Japan in 1994; and (iii) a terrorist attack on civilians in Tokyo, Japan in 1995. As previously described, no Gulf War veteran was known to have experienced an exposure this large, which caused acute symptomatic illness.

The committee pointed out that the Japanese survivors of terrorist attacks obviously experienced a wide range of exposures, as well as the stress of the event itself. Some terrorist victims showed severe cholinergic poisoning that required hospitalization or even resulted in death, some showed milder signs and symptoms, and some were exposed at levels leading to no acute effects. Commonly reported long-term health consequences including increased risk of PTSD and reports of ‘fear of subways’ are likely to have derived from the psychological stress of the terrorist attack rather than directly from cholinergic poisoning.

Perhaps host variation may explain differences in individual susceptibility. Haley has claimed that polymorphisms in the enzyme detoxification pathways for organophosphate compounds are related to symptoms (Haley et al. 1999), which is theoretically plausible (Furlong 2000), but has not been confirmed in two studies from the UK (Mackness et al. 2000; Hotopf 2003). Finally, animal experiments have failed to confirm any adverse delayed neurobehavioural effects or other significant effects from low-dose pyridostigmine alone or in combination with low-dose sarin, as well as confirming that pyridiostigmine did convey some protection against sarin, the purpose for which it is administered (Scremin et al. 2003).

The IOM committee also addressed the issue of a possible unusual genetic susceptibility to sarin toxicity in humans, and reported that the data was equivocal. The enzyme paraoxonase (PON1) found in humans in the brain and the blood is an esterase involved in sarin detoxification. The IOM committee pointed out that the human PON1 gene has polymorphisms at positions 192 (Arg/Gln) and 55 (Leu/Met) that affect serum PON1 activity (Furlong et al. 1993).

Because human serum PON1 catalyses the hydrolysis of OP insecticides and nerve gases such as sarin, those polymorphisms might substantially alter a person's susceptibility to the toxicity of the chemicals. The polymorphism at position 192 accounts for three genotypes (QQ, RR and QR) related to the catalytic properties of two forms of the PON1 enzyme (types R and Q allozymes), which hydrolyse some organophosphates at different rates.

The R allozyme (Arg₁₉₂) hydrolyses the organophosphate paraoxon at a high rate; however, it has a low activity against OP nerve agents, such as sarin and soman (Davies et al. 1996). Lower activity means that more sarin would be bioavailable to exert its anticholinesterase effects. The Q allozyme has high activity against OP nerve agents and low activity against paraoxon. Thus, people with the Q allozyme (genotype QQ or QR) are expected to have greater hydrolysis of sarin than people homozygous for the R allele (genotype RR). Animal studies support the role of PON1 in protection against the toxicity of some OP compounds (Costa et al. 2003). The prevalence of the R allele is about 0.3 in Caucasian populations but 0.66 in the Japanese population (Yamasaki et al. 1997). Because that form is associated with low hydrolysis of sarin, the authors hypothesized that it could make the Japanese population more sensitive to the toxicity of sarin, which might contribute to their morbidity and mortality after the terrorist attacks in Japan. Yamada et al. (2001), however, reported that of 10 of the victims of the Tokyo attack, 7 expressed the PON1 Q allele (6 QR, 1 QQ). The genotype that confers high hydrolysing activity toward sarin, therefore, did not appear to play a role in protecting those exposed against the toxicity of sarin.

The relationship between illness in Gulf War veterans and the PON1 genotype and serum acetylcholinesterase (AChE) activity has been investigated by Haley et al. (1999). The enzyme activity, or ability to metabolize acetylcholine, can be quantified in serum samples from the veterans. That activity is, in part, a function of the genotype of the veteran. Ill veterans (n=25) were more likely than controls (n=20) to possess the R allele (genotype RR or QR; OR 3.50; CI 0.26–2.80) and to exhibit lower PON1 type Q arylesterase activity. That study raises the possibility that the R allele represents a risk factor for illness in Gulf War veterans, but in a nested case–control study, Hotopf et al. (2003) did not find any differences in PON1 activity between symptomatic and asymptomatic Gulf War veterans. Those researchers studied symptomatic Gulf War veterans, healthy Gulf War veterans, symptomatic Bosnia peacekeeping veterans and symptomatic non-deployed military controls. The main outcome measures were PON1 activity and genotype for PON1-55 and -192. The authors observed statistically significant differences in PON1 activity among the four groups, but the two gulf groups did not differ in PON1 activity. However, those deployed to the gulf had significantly lower PON1 activity than the non-Gulf War groups (median difference 70.9; 95% CI 20.2–121.5; p=0.012); the differences were not explained by PON1 polymorphisms. PON1 activity was lower in Gulf War veterans than in military control groups. The effect is independent of ill health in Gulf War veterans.

Those studies do not entirely clarify the role of PON1 in Gulf War veterans. A study by Mackness et al. (2000) suggests that symptomatic Gulf War veterans have lower PON1 activity, but this is not explained by the various genotypes in Hotopf et al. (2003). Nonetheless, the decreased activity of PON1 would result in an increased susceptibility to OP insecticides and gases, such as sarin.

(c) 2004 sarin update

Following completion of the 2000 review, several new studies on sarin effects in laboratory animals were published. In particular, the VA Gulf War Research Advisory Committee, chaired by Mr James Binns, concluded that the new animal studies required a revision of earlier IOM committee conclusions on the lack of evidence supporting human health effects from trace level exposure to chemical warfare agents including sarin. The two laws that established this formal review process also anticipated the need for periodic updates as new scientific studies became available. Therefore, in February 2003, VA requested the IOM to update their 2000 investigation on the possibility of sarin causing long-term health effects, which they completed in August 2004.

For their update, the IOM committee focused on about 250 peer-reviewed articles published after their earlier 2000 report, including 19 epidemiological studies of sarin health effects (tables 7 and 8), as well as a wide range of animal studies. These included three studies of non-Gulf War veterans, four studies of Gulf War veterans potentially exposed at Khamisiyah, six population-based studies of US and UK Gulf War veterans using self-reported exposures and six studies of specific military units of Gulf War veterans also based on self-reported exposures. All of the studies reviewed in the initial IOM investigation were also considered in the update.

As with their earlier review, non-Gulf War veteran studies were based upon: (i) US military volunteers who had been experimentally exposed decades ago to non-lethal doses of sarin and other chemical warfare agents; (ii) industrial workers with documented acute exposure to sarin; and (iii) victims of the sarin terrorist attacks in Matsumoto City in 1994 and Tokyo in 1995. The committee also considered other relevant issues, including exposure to multiple chemicals at the same time and genetic susceptibilities. As with previous IOM committees, the new committee included toxicologists as well as epidemiologists and other public health scientists, to ensure that both animal and human data were considered.

The committee noted that virtually all human sarin studies lack good exposure data. For example, following their 2000 report, DoD exposure modelling for the Khamisiyah sarin incident was criticized as inaccurate by the US Congressional GAO. The IOM committee also noted the significant uncertainties surrounding DoD Khamisiyah exposure assessments as, ‘none of the studies using exposure information showed persistent neurological effects in Khamisiyah-exposed troops compared to non-Khamisiyah exposed troops. Because of the uncertainty in the [Khamisiyah] exposure assessment models… those studies do not provide strong evidence for or against the presence of neurologic effects.’ That is, studies based upon the DoD Khamisiyah sarin exposure modelling added little to our understanding of long-term sarin health effects. Rose & Brix (2006) consider the data on long-term damage to the peripheral nervous system in Gulf war veterans, also concluding that the evidence for any such effects is not compelling, suggesting that even if there was sarin exposure, it has not resulted in damage to the peripheral nervous system.

The 2004 IOM update confirmed the findings of the earlier IOM analysis, that: (i) there is sufficient evidence of a causal relationship between exposure to sarin and a dose-dependent acute cholinergic syndrome that is evident seconds to hours subsequent to sarin exposure and resolves in days to months; (ii) there is limited/suggestive evidence of an association between exposure to sarin at doses sufficient to cause acute cholinergic signs and symptoms and a variety of subsequent long-term neurological effects; and (iii) there is inadequate/insufficient evidence to determine whether an association does or does not exist between exposure to sarin and subsequent long-term cardiovascular effects (this last finding was not contained in the 2000 report).

The committee also reported that the new published data from experimental animals that had precipitated the interest in updating the earlier IOM review, which were designed to mimic the potential exposures in the Gulf War, were an important step in ‘determining whether a biologically plausible mechanism could underlie any long-term effects of low exposure to chemical nerve agents, but more work needs to be conducted to elucidate potential mechanisms and clarify how the cellular effects are related to any clinical effects that might be seen.’

In conclusion, exposure to sarin nerve agent remains an implausible cause of the illnesses experienced by Coalition Armed Forces. No military or intelligence authority accepts the proposition advanced in some quarters that there was deliberate use of nerve agents by Iraqi forces at the start of the ground war, which remained undetected and caused no immediate acute poisoning signs and symptoms at the time (in contrast to the Japan sarin terrorist attacks). Turning to the accidental destruction of sarin containing munitions at Khamisayah at the end of hostilities, the absence of any acute effects on personnel at the site also speaks against a serious contamination. Likewise, United Kingdom, Canadian, Australian and Danish military personnel, all of whom have experienced changes in subjective ill health similar to those of US personnel, were all too far distant, or even not in theatre, to have experienced exposure likely to cause any health effects.

On the other hand, the belief in exposure to sarin is the single strongest risk factor seen in studies of Gulf veterans. The prevalence of those who believe the cause of their ill health varies—being greater in the US than in the UK—but the strength of the association is very similar (Stuart et al. 2003).

(d) Pyridostigmine bromide

The drug pyridostigmine (PB), as either its bromide or iodide salt, is a reversible cholinesterase inhibitor used as a chemical warfare agent prophylaxis by some Coalition troops during the 1991 Gulf War. The IOM committee reviewed the large number of clinical and the fewer epidemiological studies conducted for this drug, as well as an extensive toxicological literature based upon laboratory animals. They noted that PB has been widely used for more than 40 years in the conventional treatment of myasthenia gravis in much larger doses.

Based upon their review of this abundant literature, the committee reached two conclusions: (i) that there is ‘sufficient evidence of an association’ between this medication and transient acute cholinergic effects in doses normally used in treatment and for diagnostic purposes; and (ii) that there is ‘inadequate/insufficient evidence’ to determine whether an association does or does not exist between PB and any long-term adverse health effects.

The committee reached several conclusions about the acute effects of PB: ‘a large number of clinical studies report acute, transient cholinergic effects in normal volunteers and patients with a wide variety of clinical disorders given PB as a diagnostic test of hypothalamic pituitary function and patients with myasthenia gravis treated with the drug for extended periods…’ PB is routinely used to treat patients diagnosed with myasthenia gravis with daily doses from 120 to 600 mg given for the lifetime of the patient. The committee reported that studies show about 34% of patients receiving PB have one or more, mostly mild, side effects, usually gastrointestinal, although a few experience other cholinergic symptoms such as hypersalivation, increased perspiration, urinary urgency, increased bronchial secretion and blurred vision. Patients rarely stop taking the drug because of abdominal complaints. By comparison, during the 1991 Gulf War, acute accidental poisoning with PB in doses ranging from 390 to 900 mg resulted in mild to moderate cholinergic symptoms within several minutes of ingestion, which lasted up to 24 h.

The committee explained their key finding of ‘inadequate/insufficient evidence to determine whether an association does or does not exist between PB and long-term adverse health effects:’ ‘The epidemiologic data do not provide evidence of a link between PB and chronic illness in Gulf War veterans. Most epidemiologic studies of Gulf War veterans focused on whether a unique Gulf War syndrome exists and on defining its characteristics. Only two epidemiologic studies specifically investigated the possible association of PB use and chronic symptoms among Gulf War veterans (Haley & Kurt 1997; Unwin et al. 1999).’

The IOM committee noted: ‘the suggestion of a unique manifestation of organophosphate-induced delayed polyneuropathy associated with PB exposure alone or in combination with other wartime exposures, in the absence of acute symptoms of organophosphate toxicity, requires further investigation.’ ‘Although Haley and colleagues provide evidence that chronic neurological changes are present in a small number of ill Gulf War veterans compared to a small number of well veterans from the same unit, the validity and causal nature of this association are uncertain due to the large potential for selection and information biases in this study population and the lack of a non-deployed comparison group.’ This echoes concerns expressed in other peer-reviewed literature and expert reports on either side of the Atlantic, such as the UK Medical Research Council review.

In addition, the evidence that some types of chronic neuropsychological changes may be linked to acute responses to administration of PB, also suggested by Haley & Kurt (1997), is limited by the lack of consistency with results from toxicological and clinical studies; uncertainty about the selection, administration, and interpretation of the neuropsychological tests employed; the highly select nature of the small number of Gulf War veterans studied; and the lack of comparable studies in a non-deployed comparison group.

Noting that in general studies have reported a general association between ill health and recall of hazardous exposures, rather than specific links, the IOM committee concluded: ‘the other epidemiologic study was of UK servicemen (Unwin et al. 1999), and all exposures studied (PB, diesel or petrochemical fumes, oil fire smoke, viewing dismembered bodies, etc.) showed an association of similar magnitude with adverse symptoms. Recall and reporting bias may also explain this finding. Thus, neither of these two studies provides a basis for determining that a specific association between PB and chronic adverse health effects exists.’

Similarly, the 2003 MRC report noted that most UK armed service personnel deployed to the Gulf were given PB nerve agent pre-treatment tablets. They also noted that the acute effects of PB are well characterized in human and animal studies, and are typical of other agents that inhibit AChE. Further, PB is routinely used to treat myasthenia gravis, a disease of nerve/muscle junctions, at doses generally much higher compared to that used to protect against nerve agents, and over much longer periods. They also noted that the effects of PB in combination with other agents such as insecticides and insect repellents, vaccines, genetics or stress are not known and are difficult to study.

(e) Natural experiments

Canada's experience as a Gulf War Coalition member provided a unique ‘natural experiment’ in deployment-related veterans' health effects. Three Canadian ships left for the Gulf War in August of 1990, and all three remained in the Gulf for the duration. One of the three ships (the HMCS Protecteur) rotated its entire crew between Christmas and New Year of 1990. Because they served early in the deployment prior to any hostilities, the first crew did not have exposure to most of the common Gulf War hazards of concern, including PB, anthrax or plague vaccine, oil-well smoke, chemical or biological warfare agents (including from the Khamisiyah incident), pesticides including OP pesticides, etc. The HMCS Protecteur did have DU weapons, and apparently although practice firing of DU rounds occurred, no rounds were fired in a conflict situation.

A study of the health of the sailors involved on this ship found that the initial crew who served only in 1990 and prior to the outbreak of actual hostilities later reported increased rates of a wide range of symptoms compared to controls, and at a rate similar to Canadian sailors who had served on ship actually during the conflict (Goss Gilroy 1998a,b). In other words, these early sailors manifested a ‘Gulf War deployment’ effect, even though their prime exposure was through seeing the conflict second hand while back in Canada.

Similarly, other Coalition troops who served in the Gulf War but had little or no exposure to combat or other Gulf War-related environmental hazards are reported to have war-related health problems, compared to matched controls. Danish Gulf War veterans were nearly all involved only in peacekeeping or humanitarian roles during the 1991 Gulf War (Ishoy et al. 1999a,b). Danish Gulf War Veterans have been reported to have significantly higher prevalence of self-reported neuropsychological symptoms up to six years after their return associated with various exposures (Suadicani et al. 1999). Reported symptoms include concentration, memory problems, repeated fits of headache, balance disturbances or fits of dizziness, abnormal fatigue not caused by physical activity, and problems sleeping all night. A wide range of ‘exposures’ were strongly associated with neuropsychological symptoms. Each of 17 self-reported symptoms was significantly more prevalent among deployed veterans compared to controls, and many of the symptoms were correlated with each other. Moreover, Danish veterans reported higher prevalence of unspecific symptoms such as repeated fits of headache, fatigue, memory and concentration difficulties, sleep disturbances, agitation, dyspnoea, diseases of the skin, and intermittent fever. Researchers noted that this symptom pattern was consistent with the findings among US Gulf War veterans, even though Danish Gulf War veterans were predominantly deployed after the war in peace-keeping missions (Ishoy et al. 1999a,b).

An obvious concern about PB health effects is that the drug is prescribed as a means of treating a neuromuscular disease (myasthenia gravis)—yet a common symptom among Gulf War veterans is muscular weakness or fatigue—is there a connection? The IOM committee reviewed clinical studies of PB, fatigue and neuromuscular effects, and noted one report of a group of 17 patients with post poliomyelitis syndrome, who along with 10 controls were studied for response to PB and similar drugs. Given 180 mg PB daily for one month, nine of the 17 patients reported improvement in fatigue, and continued using the drug for an average of 1.2 years, despite mild gastrointestinal effects (Trojan et al. 1993). Thus, ‘it appears that PB seems to be well tolerated and without significant neuromuscular side effects at the prescribed dose.’

(f) Interactions between Gulf War hazardous exposures

The 2000 IOM committee also examined the possibility that various Gulf War exposures in combination might interact to cause illness. They noted there are limited studies on pharmacological interactions between PB and other Gulf War risk factors, including the OP pesticide chlorpyrifos, and the insect repellent N-diethyl-m-toluamide (DEET). Co-exposure to these three agents at sub-lethal doses to hens over 2 months resulted in increased toxicity (Abou-Donia et al. 1996a). Although the IOM committee noted that the pathology and neurological impairment reported in this study were ‘hallmarks’ of OPIDN, nevertheless ‘symptoms of this neurotoxic disorder are not consistent with those reported in Gulf War veterans' illnesses. Similarly, none of the agents involved inhibits neurotoxic esterases (a target marker for OPIDN potency) at levels consistent with being capable of causing this disease in humans. Similarly, co-exposure to hens of combinations of PB and DEET along with permethrin also showed increased toxicity, possibly because two of these agents are esterase inhibitors and the third is an ester (Abou-Donia et al. 1996b). Finally, co-exposure to large amounts of these three agents to rats increases their lethality (McCain et al. 1997), although the relevance of such a dosing regime to humans is not clear. Moreover, the IOM committee concluded, ‘there is no a priori reason to suspect that PB is capable of causing OPIDN,’ and ‘there are insufficient data to determine whether exposure to other chemicals, either before or after PB, enhances its potential to produce delayed neurotoxicity.’ Relative to human health effects, the committee similarly concluded that the lack of data supporting an association between PB exposure and long-term adverse health effects was true ‘particularly when PB exposure occurs in combination with other combat exposures.’

(g) Vaccines

All military service personnel are routinely vaccinated against a wide variety of common infectious diseases, and against certain military-specific diseases, as indicated. Some veterans have expressed concerns that these vaccinations, or some combination of vaccines, could result in long-term health consequences. The IOM committee noted the difficulty of studying this issue among actual Gulf War veterans, again because of poor data on who actually received the vaccines. DoD recorded that about 311 000 doses of the anthrax vaccine were sent to the Gulf War theatre, and about 150 000 US service members received at least one vaccination. However, little information is available on who actually received them. Similarly, according to DoD, about 138 000 doses of botulinum toxoid were sent to the Gulf, and about 8000 US service members were estimated to have received the vaccine, but medical records provide little information about who actually were the recipients.

The peer-reviewed scientific literature available to the committee on adverse effects among humans from anthrax and botulinum toxoid vaccines included results from human volunteer studies, US post-market surveillance via the voluntary Vaccine Adverse Event Reporting System, and active surveillance from clinical studies. The committee noted that studies on these two vaccines primarily involved passive surveillance with only short follow-up periods relative to the periods that concern Gulf War veterans. Similarly, available studies were often limited by small sample size, involvement of multiple different vaccinations, large numbers of potential health outcomes evaluated, and lack of unique symptoms specific to vaccination.

The committee found only one randomized peer-reviewed study of the type of anthrax vaccine used in the United States (Brachman et al. 1962), and it involved no long-term monitoring for adverse outcomes. The committee also had data from animal studies, including anthrax vaccination of agricultural live-stock.

Not surprisingly, the committee concluded, ‘there is a paucity of published peer-reviewed literature on the safety of the anthrax vaccine.’ Nevertheless, they also concluded that published studies have not reported any significant adverse effects. Similarly, the committee found few published peer-reviewed studies on potential adverse health effects from the botulinum toxoid vaccine administered to humans. The small number of limited studies probably contributed to their default conclusion of ‘inadequate/insufficient evidence of an association’ for long-term health effects.

(h) Simultaneous vaccinations

The committee also noted the paucity of peer-reviewed studies on potential effects from multiple simultaneous vaccinations. They focused on studies of three relevant populations, including a group of Finnish military recruits who received numerous vaccinations during their first weeks of service; a group of laboratory workers at DoD's Fort Detrick, Maryland, who received an intensive vaccination regime as part of their occupational safety; and several limited studies of US and UK veterans of the 1991 Gulf War. The committee reported that these studies provided little evidence of long-term adverse effects (beyond the transient localized and systemic effects seen with any vaccination). However, they made it clear that the available evidence was of insufficient quality, consistency or statistical power to allow developing any firm conclusions about the presence or absence of any association between multiple vaccinations and any adverse health outcomes.

(i) Pesticides and solvents

The next ‘Gulf War & Health’ IOM committee reviewed health effects from pesticides and solvents used in the 1991 Gulf War. In their 2003 report, all of the 21 health effects identified as positively associated with these risk factors were previously well-documented associations between common occupational and environmental hazards and various cancers, neurological, and other specific diseases (table 4). These were primarily various cancers and serious haematological disorders (e.g. leukaemias, non-Hodgkin's lymphoma, multiple myeloma and aplastic anaemia), subtle general neurological effects detected via neurobehavioural tests, and other health effects (i.e. reactive airway dysfunction syndrome, and allergic contact dermatitis).

Nearly all the data the IOM reviewed came from occupational studies involving large or prolonged exposures to industrial workers. Consequently, most positive findings were based upon occupational exposures the committee characterized as ‘chronic,’ ‘high-level,’ or at doses ‘sufficient to cause poisoning’ (the one exception was an association between propylene glycol and insecticides exposure and allergic contact dermatitis).

Occupational studies are clearly relevant, but results from studies of civilian workers may be difficult to extrapolate to military personnel, since civilian workplace exposures are typically much larger and longer compared to the typical Gulf War veteran, or indeed veterans from any era. The committee commented on the absence of exposure data for the 1991 Gulf War, and it is not clear that Gulf War veterans were on average exposed to significantly larger amounts of the common pesticides and solvents used in that conflict compared to any other service members, or US civilians. None of the committee's findings came from studies of Gulf War veterans themselves.

Some of the positive associations noted in the 2003 IOM report represent a clear public and occupational health issue for humans exposed to high levels of cholinergic poisons such as the OP pesticides and military nerve agents. However, they are probably less relevant to concerns about long-term neurological health effects among veterans of the 1991 Gulf War (table 4). For example, the committee found ‘limited/suggestive evidence’ of an association between acute OP pesticide poisoning or solvent exposure, and long-term neurobehavioural effects including abnormal results on neurobehavioural test batteries and symptom finds. In fact, a very large body of scientific literature exists showing such long-term effects in the aftermath of acute, generally severe, cholinergic poisoning (see for example Brown & Brix 1998). As a result, a range of occupational safety programmes have been established to protect workers (primarily in agriculture) from such effects. However, the clinical significance to veterans of the 1991 Gulf War is unclear because of the lack of any reported instances of acute cholinergic poisoning during that deployment. Gulf War veterans were apparently exposed at most to sub-clinical levels of OP cholinergic toxins such as pesticides and military nerve agents. For example, in the Khamisiyah sarin exposure modelling conducted by the US Department of Defense, the exposure levels used were developed to be completely protective of civilians, by setting them orders of magnitude smaller than that leading to acute toxicity.

Interestingly, the IOM committee also placed in the category of ‘inadequate/insufficient evidence to determine whether an association exists’ insecticides and solvents and peripheral neuropathy, Parkinson's disease, amyotrophic lateral sclerosis (ALS) and Alzheimer's disease (AD). They noted the difficulty in studying the relationship between environmental exposures and neurological disease due to uncertainty in diagnoses, and long latencies associated with these illnesses, and problems with self reported exposure data. For example, although there are many epidemiological studies that attempted to evaluate associations between agricultural chemicals factors and Parkinson's disease, subjects often have difficulty recollecting which specific pesticides they may have been exposed to. The committee identified 6 case-controlled studies of Parkinson's disease, who were primarily agricultural or other pesticide workers. They reported conflicting results among the studies, and concluded that there is only ‘inadequate/insufficient’ evidence of an association between insecticide exposure and Parkinson's disease.

Similarly, with ALS and pesticide exposure, the IOM committee identified only a single peer-reviewed study (McGuire et al. 1997). They concluded that also the study provided ‘evidence of a relationship,’ but that possible selection bias resulting in an overestimate of an effect made it ‘inadequate/insufficient’ evidence of an association.

Finally, the committee reviewed the literature associating AD with pesticide exposure. They found several studies on the relationship between insecticides and AD, which were either negative or not relevant to exposure among Gulf War veterans. Therefore, the committee concluded that there was again only ‘inadequate/insufficient’ evidence of an association between this disease and insecticide exposure.

(j) Fuels and combustion products

The third ‘Gulf War & Health’ IOM committee reviewed chronic health effects associated with exposure to oil-well fire combustion products, diesel-heater fumes, hydrogen sulphide (a specific oil-well fire product), hydrazines and red fuming nitric acid (as rocket propellants), and gasoline and jet fuel. Their 2004 report included nine positive findings linking certain long-term health effects with exposure to combustion products and rocket propellants (table 4). Most of studies supporting the committee's findings involved people living in urban areas who were routinely exposed to ordinary ‘smog’ from motor vehicles, stoves, heaters and other common sources of ambient and indoor air pollution. Other studies involved occupational exposure to engine exhaust, for example, among professional motor vehicle drivers. A few studies were of actual veterans of the 1991 Gulf War, based upon their exposure to burning oil-well fires in the Gulf War theatre.

As with earlier IOM reports, all positive findings were previously very well-documented associations between the generally commonplace environmental exposures and various specific diseases, with an abundant scientific health literature available for review. These include an association between combustion products (i.e. exhaust from engines ‘smog,’ oil-well-fire smoke, and fumes from tent heaters) and lung cancer, cancers of nasal cavity and nasopharynx, cancers of the oral cavity and oropharynx, laryngeal cancer, bladder cancer, low birth weight/intrauterine growth retardation and exposure during pregnancy, preterm birth and exposure during pregnancy, and asthma. They also reported an association between exposure to hydrazine rocket propellants and lung cancer, although according to reports from the US Department of Defense, these fuels were not used during the 1991 Gulf War.

Interestingly, the IOM committee reported no long-term health consequences from exposure to fuels. That is not to say the IOM did not identify certain acute and immediate health effects from exposure to such fuels. However, their review did not reveal any long-term effects that manifested only months or years after an acute poisoning from exposure to fuels.

(k) Problems extrapolating to deployed military personnel

There are always uncertainties extrapolating health effects observed among civilian workers and urban dwellers to health effects expected among deployed service members. The best epidemiologic studies on long-term health effects from combustion products and fuels are those of urban dwellers and workers exposed over years to decades. However, their relevance to health effects anticipated among military personnel serving in the relatively short 1991 Gulf War is not clear, since the exposures involved are typically much longer (e.g. years to entire lifetimes for urban dwellers and workers) or much larger (e.g. based upon highly exposed industrial workers). Similarly, urban dwellers and workers with larger or longer exposures are more likely to develop a particular disease than people who experienced smaller or shorter exposures.

Thus, the IOM committee noted that reviewed studies primarily ‘included people whose exposure was over a lifetime (as in community air pollution studies), or included workers employed in a particular industry over many years.’ ‘In contrast, the exposures experienced by veterans in the Persian Gulf were relatively short although the intensity might vary from occupational exposures,’ and ‘[t]herefore, the exposures experienced in the gulf might only approximate exposures described in the occupational literature used in this report.’

(l) Gulf War exposure to combustion products

The 2004 IOM committee noted that poor exposure data for Gulf War veterans made it difficult to predict health effects from exposure to combustion products among this population. However, some relevant data on combustion product exposure is available from the US Department of Defense (DoD September 2000) investigations on exposures to oil-well-fire smoke and related combustion products during the 1991 Gulf War. From January through late February 1991, retreating Iraqi forces set fire to more than 600 Kuwaiti oil wells that lofted huge columns of smoke up into the atmosphere, and these oil-well fires were not completely extinguished until about 9 months later. Their report concludes that although the oil-well fires produced spectacular smoke plumes, the actual exposure to combustion products of US service members in that region was generally unremarkable. Furthermore, unlike many Gulf War environmental hazards of concern, an extensive monitoring effort by various agencies for air pollutants and combustion products from the 1991 Gulf oil-well fires is available to support firm conclusions about such exposure. Nevertheless, the DoD report also concludes that some individual veterans who were near the oil-well fires could have been exposed to high levels of large particulates, primarily as material deposited directly to skin or clothing rather than through inhalation.

The DoD report concludes, ‘[a]t the time of the destruction, the medical and environmental community feared exposure to the fires would result in catastrophic acute and chronic health effects. However, the fires' high combustion efficiency, the nature and amount of the smoke's contaminants, the lofting effect created by solar heating, and the local wind and weather conditions combined to reduce the fires' impact on military and civilian populations.’ ‘Results of air monitoring studies indicated, except for particulate matter, air contaminants were below levels established to protect the health of the general population. However, there were self-reports by a number of veterans who complained of acute symptoms they allege were a result of their proximity to the burning oil wells.’

The report also notes that exposures to US service members were quite short compared to civilians dwelling in US cities exposed to urban ‘smog’ and indoor air pollution, or workers exposed to engine exhaust. ‘Fortunately, the time period during which military and civilian populations were subjected to the fires' pollution was relatively short.’ Although, ‘[w]hile smoke plumes occasionally touched the ground, enveloping nearby personnel, few were in those areas for extended periods of time.’

In an interesting comparison of deployment-related combustion product exposure, blood and urine samples from a group of US troops were collected before, during and after deployment to Kuwait during the 1991 Gulf War, the centre of the burning oil-well fires (US Army Environmental Hygiene Agency 1991). The concern was that troops stationed in Kuwait near the burning oil-well fires would be exposed to high levels of polyaromatic hydrocarbons (PAHs) and other pollutants. Samples were measured for volatile organic carbons (VOCs), PAH–DNA adducts, metals, and sister chromatid exchange frequency in lymphocytes. Surprisingly, this troop biomonitoring study found low PAH concentrations in Kuwait (Poirier et al. 1998). Biomonitoring with DNA adduct assays showed the lowest exposures to troops located in Kuwait, with significant increases among soldiers after they returned to Germany. Levels of metals, VOCs, and PAH–DNA adducts were essentially identical or showed decreases among troops while they were in Kuwait, compared to troops living in Germany. In particular, lead levels in blood were not statistically altered during deployment to the Gulf War theatre. ‘Overall, the data suggest that this group of soldiers was not exposed to elevated levels of PAHs while deployed in Kuwait.’

(m) Gulf War exposure to hydrazine rocket propellants

According to another DoD investigation, rocket fuels used by Iraqi forces in Scuds and several smaller missiles during the 1991 Gulf War contained a type of kerosene and red fuming nitric acid (also known as IRFNA). They state that apparently Iraq had experimented with hydrazine rocket fuels including UDMH, however they concluded that these fuels were not used during that conflict. ‘In addition to IRFNA, missile fuels were also potential battlefield hazards to coalition forces. The missile fuel that Iraq used in its older Soviet systems was a specially refined kerosene-like substance (called kerosene in the literature). Some improved missiles used UDMH in combination with IRFNA. The Soviet Union used UDMH in their Scuds, but we have no evidence that Iraq used UDMH’ (emphasis added). Therefore, it appears unlikely that any Coalition service members were exposed to hydrazine rocket fuels during the 1991 Gulf War.

(n) Reproductive health effects among Gulf War veterans

Compared to previous IOM committees, the 2004 committee decided to focus greater attention on potential reproductive health effects (table 4). They noted that previous IOM reports on Gulf War veteran health focused primarily on health effects to adults, because pregnancies during the 1991 Gulf War were thought to be rare and pregnant women were thought to have been immediately evacuated from the Gulf War theatre. However, the committee described recent published data indicating that there were more pregnancies than previously reported. Furthermore, pregnancies were typically not identified until 2 to 6 weeks after conception, and most pregnant service members would have been evacuated during the first 6 to 12 weeks of pregnancy. Therefore, the committee expanded its search of epidemiological literature to include reproductive health outcomes from exposures during the first trimester of pregnancy. The committee reported finding ‘limited/suggestive’ evidence of an association between exposure during pregnancy to hydrazine rocket fuel and low birth-weight, intrauterine growth retardation, and preterm birth. However, since Coalition forces were apparently unlikely to have been exposed to these agents during the 1991 Gulf War, the clinical significance of this finding is not clear. IOM committees have not found any other reproductive health effects to be associated with any other Gulf War risk factor they reviewed. This general topic is covered in greater detail in the contribution from Professor Doyle and colleagues (Doyle et al. 2006).