All the POPC active ingredients were registered for use by the EPA at the time of deployment, with the exception of azamethiphos. Most of the POPC active ingredients are still registered for at least some uses, although EPA has made a number of changes in specific uses since the Gulf War, and plans future changes. Azamethiphos has never been registered. The following are currently registered: DEET,[565] d-phenothrin, methomyl,[566] and propoxur.[567] The following are currently registered, and are scheduled for reregistration eligibility decisions by the end of FY2001: permethrin, dichlorvos, diazinon, malathion, and lindane.[568] In 1995, EPA proposed to cancel most uses of dichlorvos, including all uses in or on residences, but so far, no decision has been made.[569] All bendiocarb products will be canceled as of December 31, 2001.[570]
EPA has reached agreement with the registrants to eliminate various uses of chlorpyrifos. For example, the agreement will cancel and phase out nearly all indoor and outdoor residential uses. Notably, under the agreement, ULV fog applications for mosquito control will be allowed to continue.[571]
I. Uncertainty and Variability Analysis
"Uncertainty" refers to those aspects of the data and exposure factors which are not presently known, and/or are unknowable. "Variability" refers to the variation within a population which may or may not be presently known and/or well characterized. For example, the actual levels of pesticide active ingredients to which veterans were exposed can never be known because no real-time monitoring was conducted at the time, and it is impossible to collect the data today. Thus, there may be a high level of uncertainty associated with some of the estimates of exposure levels calculated. In contrast, the data on body weights among veterans who served in the Gulf War exists and either has been or can be quantitatively described, if desired, and such a description would be a description of variability. What follows, however, is predominantly a qualitative description of uncertainty and variability.
1. Uncertainty in the Selection of POPCs
The group of all pesticide products identified as definitely or possibly having been used during the Gulf War was assembled from many sources with varying shortcomings. There is solid documentation regarding the pesticide products ordered by deploying units through the military supply system. However, no documentation was identified detailing what was actually shipped, received, used, disposed, and/or returned. Military guidance consulted from the period indicate only which pesticide formulations were recommended under various circumstances. Discussions with the Armed Forces Pest Management Board helped fill in many gaps about what was probably used. The survey and interviews of the veterans provided much additional information, but were also based on years-old recollections and had many inaccuracies identified.
The criteria used to limit the group of all pesticides and related products to a manageable subset, known as the POPCs, while logical and defensible, are imperfect. For one thing, both the survey and interviews have varying degrees of bias. In the case of the survey, recall bias is an issue. Among other things, veterans have had much potential exposure to media reports on Gulf War illness issues. In the case of the PM interviews, while these individuals are presumed by investigators to be the most knowledgeable about pesticide application, they bore some of the responsibility for application, and some may have avoided conveying information with the potential to reflect negatively on them.
2. Uncertainty in the Exposure Assessment
The most significant uncertainties in the exposure assessment, and in this risk assessment as a whole, are associated with the estimation of exposure concentrations and doses of POPCs active ingredients. No documentation from the period has been identified which records, at the unit level, what pesticide products were applied, how they were applied, and in what quantities they were applied. There are also no real-time monitoring data for pesticides. Thus, it was necessary to employ many location-specific and generic assumptions in order to estimate concentrations. While many of these assumptions originated with EPA, are widely used, and are based on credible field and laboratory studies, they have varying degrees of applicability in this HRA exposure assessment. All attempts were made to develop realistic exposure scenarios and assumptions, based for example on veteran survey and interviews, yet varying degrees of uncertainty remain.
In many cases, pesticide product exposures during deployment would have been very similar to exposures normally occurring in the US at the time. On the other hand, there were certainly conditions existing some of the time that would have contributed to higher-than-normal, or otherwise unusual exposures for some servicemembers.
3. Uncertainty in the Toxicity Assessment
Some uncertainty is inherent in the toxicity values used to quantify potential systemic and carcinogenic effects. EPA toxicity values for systemic effects are derived using available toxicity information and standard uncertainty factors. Uncertainty factors are assigned based on the quality of the available data. Consequently, the less information that is known about a given constituent, the more conservative the systemic toxicity value. Carcinogenic toxicity values (slope factors) are based on the 95% upper confidence limit on lifetime risk and, therefore, will tend to overestimate cancer risks.
A full set of ideal toxicity values for the HRA do not exist. The ideal reference dose is a verified duration-specific and route-specific value. For example, the ideal subchronic dermal reference dose is based on a reliable subchronic dermal no-observed-adverse-effect level (NOAEL). Since ideal values do not exist in many cases, provisional values were identified or developed, and used. Published toxicity values from various sources were compiled and modified to provide the most comprehensive and appropriate values available. Route-to-route extrapolation was carried out; for example, oral toxicity values were modified to be used for dermal exposure risk assessment.
Toxicity values were identified for POPC active ingredients only. The active ingredients were formulated and/or applied along with other chemicals, generally referred to as "inerts." While inerts are typically of low acute toxicity, some of them may be associated with adverse health effects, at least upon repeated high-dose exposures. For example, lindane dust was 99% talc. Talc has many applications in pharmaceutical and pesticide formulation, as well as in cosmetics. However, some forms of talc, when inhaled repeatedly and in high concentrations over long periods of time, have been associated with adverse respiratory system effects. One source states that there are many case histories available on talc workers who developed pneumoconiosis after employment periods ranging from 1 to 37 years.[572] Pneumoconiosis is an inflammation of the lungs, commonly leading to another pathological condition known as fibrosis, caused by the long-term inhalation of some types of dust. The potential health effects related to talc alone were not evaluated as part of the HRA.
4. Uncertainty in the Risk Characterization
Most of the hazard quotients (HQs) calculated for individual pesticide active ingredients should be considered reasonably reliable, depending on the reliability of the specific components. As HQs are added together across exposure routes and different chemicals, to yield a hazard index (HI), certainty declines. In the HRA, organophosphates and carbamates were grouped together because investigators presumed a common mechanism of action. This particular grouping is widely practiced and accepted. The effects of additional pesticide active ingredients (e.g., permethrin, d-phenothrin, lindane) are unknown. It seems more likely that those pesticide active ingredients that potentially affect the central nervous system are likely to have equal to or greater than additive effects (synergism), rather than less than additive effects (antagonism). Additional uncertainty is associated with the assumption that all toxicity values used in the evaluation have an equal degree of reliability, which in reality is not the case.
There are many factors that would have significantly reduced actual exposure, while there are many conservative assumptions used in the HRA that significantly increase apparent (estimated) exposure. Pesticide applicators who followed the required procedures would have reduced their hazards and risks substantially. For example, proper use of PPE would have reduced associated hazards and risks by factors ranging from 10 to 100. However, a key assumption in the HRA was that, "the most exposed were the least protected," which probably tends to exaggerate hazards and risks.
The hazard quotients, hazard indices, and risks calculated are based on the specific endpoints of animal and human testing, such as the suppression of cholinesterase within a specified time frame. Normally, the most sensitive endpoints are used to establish the toxicity values; however, there are other potential endpoints and effects which the HRA does not take into account, such as:
Investigators do not believe that OPIDN due to pesticide exposure plays any significant role in veteran health. First, investigators did not uncover any reports of OPIDN symptoms. Second, chlorpyrifos is the only pesticide active ingredient identified which has been associated at all with OPIDN, and it is only weakly associated. Additionally, RAND concluded that the relevance of OPIDN symptoms to the issue of undiagnosed illnesses in veterans is limited.[573] OPIDN is a very well-characterized syndrome, with very specific sequelae. A major problem with studying unexplained illnesses in veterans has been the lack of objective measures. OPIDN begins with acute symptoms, and later presents clearly recognizable clinical and histopathological manifestations (e.g., peripheral neuropathy). Investigators uncovered no reports of the occurrence of organophosphate intermediate syndrome.
Delayed cognitive toxicity (DCT) is a condition of delayed signs and symptoms that is not akin to OPIDN. In two studies, industrial workers exposed to OPs during sarin manufacture reportedly showed electroencephalogram (EEG), sleep, memory, and personality changes sometimes 2 years after exposure.[574] Given that literature reports describing DCT are rare, and it has only been associated with sarin exposure, it too is of questionable relevance to pesticide exposure during deployment.
Perhaps most importantly, the HRA does not account for the potential combined effects of some pesticide active ingredients and other chemicals. The HRA addresses to some extent the combined effects of the specific pesticide formulations identified as having had a fairly high level of use during the Gulf War, but does not fully account for any additional effects due to the following:
One unanswered, and probably unanswerable, question then is, "What were the individual body burdens of all like-acting chemicals, such as cholinesterase inhibitors, and what is the full spectrum of health effects that may have resulted?" Possible interaction of organophosphate and carbamate pesticide active ingredients with PB is of particular interest, since PB is a carbamate, and all act by a similar toxic mechanism, namely, inhibition of various forms of cholinesterase.
Discussion of uncertainties related to the other human benchmarks is presented in Tab C-8.
Some of the exposure factors and assumptions used in the HRA are associated with no significant variability. For example, one can say with a high level of certainty that the floor area of a GP large tent used during the Gulf War was 900 ft2, and that there is little variability associated with this datum. On the other hand, each of many of the other exposure factors and assumptions used in the HRA is associated with an underlying variability which may be well known and well characterized, or may be unknown but knowable, or unknowable. In this risk assessment investigators assumed that all receptors weigh 70 kg. This is a commonly used assumption for adult body weight, but in reality the weights of the veterans vary substantially, and the data exist to characterize the data distribution with a high degree of certainty. Investigators selected the value of 70 kg for BW because it is close to the mean adult body weight for males and females combined, and because of the precedent for its use.
The HRA took the traditional "deterministic" risk assessment approach of using point estimates for each input rather than the "probabilistic" approach using data distributions for each, and then applying a technique such as Monte Carlo Analysis. While we acknowledge that there are more rigorous quantitative methods for evaluating and describing data distributions, and that it would be possible to do this for some exposure factors, we decided not to employ the more rigorous approach. The reasons for our decision to use the deterministic approach are that 1) we can meet the objectives of the HRA by doing so; 2) it would be a much larger effort to use the probabilistic approach; 3) we lack sufficient data to truly characterize the distributions for many exposure factors; 4) the results used to draw conclusions would likely not change dramatically. During the course of an in-depth review of the HRA, neither EPA nor peer reviewers recommended that we should use the probabilistic approach.
| First Page | Prev Page | Next Page |