In many cases, pesticide formulation 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. Mass delousing does not occur domestically. Azamethiphos is not registered for use in the US. Most people in the US do not live under relatively primitive conditions in sweltering regions of extreme pest infestation, which require intensive and sustained control measures. Symptomatic pesticide formulation exposures (i.e., overexposures), however, appear to have occurred rarely by all accounts.

Investigators evaluated the low-, medium-, and high-exposure levels for each POPC, although doses were not always calculated for low and medium levels. The exposure levels are defined by the exposure factors selected, such as the volume of the formulation used each day. Some exposure factors, such as assumed body weight, are the same for all three exposure levels. The degree of overall conservatism increases as one moves from low exposure to high exposure. We applied, as much as possible, the following guidelines for selection of exposure factors:

• Low-exposure group. The nominal 5^th percentile exposure factors derived from the survey. The nominal 10^th percentile exposure factors derived from the interviews. The nominal 10^th percentile exposure factors or low-end values available from published sources.
  
  
• Medium-exposure group. The average exposure factors derived from the interviews. The nominal 50^th percentile exposure factors or mid-range values derived from the survey and published sources.
  
  
• High-exposure group. The nominal 95^th percentile exposure factors derived from the survey. Nominal 90^th percentile exposure factors from interviews. The nominal 50^th percentile exposure factors or midrange values available from published sources. The nominal 50^th percentile published values or other midrange values are used in the high exposure group in order to minimize unreasonably high bias per EP^[181] recommendation. The EPA believes that this approach avoids compounding many conservative assumptions to produce a highly unlikely or impossible exposure estimate.

One exception to the selection of percentile values are the assumptions regarding the wearing of PPE by pesticide applicators. In general, investigators assumed that the low-exposure group wore the most PPE, and that the high-exposure group wore the least. So, for example, the 90^th percentile for percent of servicemembers wearing PPE is associated with the highest level of protection, but serves as the basis for the low exposure group PPE assumption. The PPE percentile values are not used directly in the calculation of doses, as the algorithms do not accommodate this factor, and, due to the shortcomings in the relevant data from the PM exposure interviews, they can only be used qualitatively to indicate the level of PPE use. In short, investigators assumed that all individuals within an exposure group (e.g., low exposure) wore the same components of PPE, such as gloves and respirators.

The results for the medium-exposure group are the most reliable in this HRA. The "high-exposure group" is equivalent to what is commonly referred to as the "reasonable maximum exposure" (RME) receptor group, which likely applies to a small number of individuals. The EPA has typically used the RME level for decision-making purposes in the regulatory setting. The results for the low and high-exposure groups are less reliable than those for the medium-exposure group, due to generally greater uncertainty surrounding the associated assumptions in the former groups. While the EPA is interested in what might happen in the future, we are particularly interested in what most likely happened in the past.

As used here, a "nominal" 10^th percentile value is the point below which approximately 10% of the values from the associated sample distribution fall. A nominal 50^th percentile value is the point below which approximately 50% of the values from the associated sample distribution fall. A nominal 90^th percentile value is the point below which approximately 90% of the values in the associated distribution fall. However, the sample data arise from the years-old recollections of veterans, and the exposure factors thus derived may not be accurate values for the relevant veteran subpopulations.

One important factor that impacted pesticide formulation use was the prevailing pest pressure or the perceived pest threat (and any associated threat of disease). "Pest pressure" is directly related to the active pest population in the vicinity. If the active pest population was very large, then the pest pressure was high, and would most likely have been associated with a high level of pesticide product use, if pesticide products were available. Usually, the pest threat was tied to time of year and location. For example, fly populations were worse in August than they were in January, and servicemembers were more susceptible to flies if they were quartered in temporary shelters in rural locations versus permanent shelters in urban locations. "Urban" denotes a moderately to highly developed area, normally having a high human population density. "Rural" denotes a relatively undeveloped area, normally having a low human population density. In the scenarios described later in the exposure assessment, each pesticide exposure level is associated with an assumed level of pest pressure. A low exposure level would equate to minimal pest populations and low levels of pesticide use due to time of year and/or location, while medium and high levels would have corresponding moderate and high levels of pesticide use.

Another important distinction between exposures is whether they occurred during application or whether they occurred following application ("post-application exposures"). Application here includes mixing, loading, and applying pesticide products. People may experience exposure during application and/or post application.

Other factors impacting exposure level include the level of PPE used by applicators, and for post-application exposure, whether the applications were made outdoors or indoors. In cases where PPE should have been used, investigators assumed that the low-exposure applicators donned appropriate PPE, and high-exposure applicators did not. Investigators assumed that medium-exposure applicators either did or did not don appropriate PPE, depending on the information provided by the PM interviews. Investigators also assumed that post-application exposures occurred outdoors for the low- and/or medium-exposure group, and indoors for the medium- and/or high-exposure group (depending on our information). While investigators did not evaluate applicators who wore PPE at the high-exposure level, it is safe to conclude that, in a given case, the estimated dose for them would lie between the dose at the medium-exposure level (wearing PPE) and the dose at the high-exposure level (without PPE). In a few selected cases which we examined, a person not protected by PPE would typically receive 10 times the inhalation dose and 10 to 100 times the dermal dose versus a person protected by appropriate PPE. In some cases, however, BDUs may have afforded significant dermal protection, perhaps more so than is reflected in some HRA assumptions.

2. Quantifying Exposure in Personnel

Investigators grouped the 15 formulations evaluated in the exposure assessment based on similarities in designated uses, application, and/or exposure scenarios, as shown in Table 20.

For the pesticide formulations considered in the HRA, the most common routes of exposure are dermal contact and inhalation. For this exposure assessment, investigators calculated exposure point concentrations (EPCs) only when they evaluated the oral route of exposure, and when they conducted air modeling to estimate the air concentrations of pesticide active ingredients. Otherwise, investigators calculated doses by other means, as described in the following subsection. EPCs are the concentrations of POPCs in a given medium to which a receptor may be exposed at a specific location known as the "exposure point."

Per EPA recommendation, investigators calculated one type of dose for the evaluation of potential noncarcinogenic effects, and another for the evaluation of potential carcinogenic effects. ^[182] For evaluation of noncarcinogenic effects, we calculated doses for the day of application; that is, we did not reduce the doses by averaging over time. In contrast, for evaluation of carcinogenic effects we calculated lifetime average daily doses (LADDs). Thus, the LADD is always substantially lower than the dose calculated to evaluate noncarcinogenic effects. All doses in the HRA, whether for noncarcinogenic effects or carcinogenic effects, are expressed in units of mg active ingredient per kg body weight per day (mg/kg/d). Intakes for carcinogenic effects are averaged over a lifetime since risk of cancer is accrued over a lifetime. In any case where investigators calculated a dose for the evaluation of noncarcinogenic effects, they calculated at least the high dose. Investigators calculated low and medium doses in most cases; however, if investigators deemed the low and/or medium dose inconsequential, they only calculated the high dose. Investigators calculated a LADD based on the medium exposure level only, where possible; they used the high exposure level in cases where they did not evaluate the medium exposure level.^[183]

As used in the HRA, the term "consequential" means of sufficient magnitude to potentially affect the outcome of the health risk assessment. A "consequential exposure" is known or likely to be greater than or equal to one-tenth of any relevant risk-based threshold such as a reference dose (Tab D, Section D, "Toxicity Assessment"). The term "inconsequential" means of insufficient magnitude to potentially affect the outcome of the health risk assessment. An "inconsequential exposure" is known or likely to be less than one-tenth of any relevant risk-based threshold such as a reference dose.

Table 20. Formulations evaluated in the exposure assessment

Investigators initially identified all plausible scenario/route combinations of consequential exposure (Tables 21 and 22), and calculated doses for them as described throughout the exposure assessment. In some cases, investigators evaluated both application and post-application together as the contribution of one so far exceeded the other. This is explained in detail later for each formulation. Evaluation of potential carcinogenic effects is indicated only for those pesticide active ingredients having an appropriate carcinogenic slope factor (Tab D, Section D, "Toxicity Assessment"). The most common routes of exposure are the dermal and inhalation routes; the oral route is relevant in only a few cases.

Table 21. Summary of application scenarios and exposure routes evaluated^a

Table 22. Summary of post-application scenarios and exposure routes evaluated^a

The types of doses calculated in the exposure assessment are as follows:

Doses for noncarcinogenic effects:

• PDR_D = potential dose rate for dermal contact
• AD_D = absorbed dermal dose
• PDR_I = potential dose rate for inhalation exposure
• PDR_O = potential dose rate for oral exposure

Doses for carcinogenic effects:

• LADD_D = lifetime average daily dose via dermal contact (absorbed dose)
• LADD_I = lifetime average daily dose via inhalation exposure (potential dose)
• LADD_O = lifetime average daily dose via oral exposure (potential dose)

The HRA expresses dermal doses for the evaluation of noncarcinogenic effects as both potential dose and absorbed dose. In general, these two types of doses, for any route of exposure, are defined as follows:

• Potential dose. The amount of pesticide active ingredient contained in material ingested, air breathed, or bulk material applied to skin.^[184]
  
  
• Absorbed dose. The amount of pesticide active ingredient penetrating across the absorption barriers (exchange boundaries) of a test organism or human (same as "internal dose"). ^[185]

Either one of these may be used to characterize risk, as appropriate, depending on the type of toxicity value identified (Tab D, Section D, "Toxicity Assessment" and Tab D, Section E, "Risk Characterization"). The HRA presents the inhalation and oral doses for evaluation of noncarcinogenic effects only as potential doses, as this is appropriate for the available toxicity values. The HRA presents the LADD for dermal exposure only as an absorbed dose, because this is appropriate for the available slope factors (Tab D, Section D, "Toxicity Assessment"). The HRA presents the LADDs for inhalation and oral exposure as potential doses, as appropriate for the available slope factors.

To convert a potential dose to an absorbed dose for dermal exposure, it is first necessary to find an appropriate chemical-specific dermal absorption factor (ABS). The ABS represents the fraction or percentage of the pesticide active ingredient which can be expected to pass through the skin into the body. Then the HRA calculates the absorbed dose as follows: potential dose times ABS = absorbed dose.

All equations used to calculate doses are described in the following subsections. All the exposure factors used in the equations are described as well. There are three exposure factors used throughout the HRA, and investigators assume the following:

• Body weight (BW) is 70 kg ^[186] for all receptors. The value of 70 kg is the value most commonly used for adults (males and females combined) in EPA risk assessments over the past 15 years or so.
  
  
• A human lifetime is 70 years.^[187] This is the value most commonly used in EPA risk assessments over the past 15 years or so. The value for lifetime used for averaging time (AT) in the equations is 25,550 days (i.e., 70 years x 365 days).
  
  
• The inhalation rate (IRA) is 1.6 m³/h, which is the EPA-recommended value for adult short-term exposures during moderate activities.^[188] This inhalation rate takes into account that troops were engaged in a variety of activities over the course of the average day ranging from light activities to heavy activities.

• acute/subacute: exposures lasting <2 weeks,
  
  
• subchronic: exposures lasting 2 weeks to 180 days, and
  
  
• chronic: exposures lasting >180 days.

One source defines acute exposure as lasting less than 1 day, and subacute as lasting one month or less.^[191] The subchronic and chronic exposure durations are adapted from EPA guidance.^{[192,] [193]}

Most pesticide product exposures during the Gulf War were acute/subacute and subchronic; however, there were probably a small number of chronic exposures as well. In some cases there are varying acute/subacute, subchronic, and chronic toxicity values, and investigators combined the assumed exposure duration with the appropriate toxicity value.

In the following subsections, there are six possible exposure scenarios for each formulation: one scenario for each of three exposure levels for application exposure, and one scenario for each of three exposure levels for post-application exposure. Investigators evaluated the high-exposure scenarios in detail in all cases for application and/or post-application exposure; however, they did not evaluate lower-level exposure scenarios if they concluded that the lower-level scenarios are inconsequential. For example, if investigators concluded that lower-level exposure scenarios provided little or no opportunity for contact with the pesticide active ingredient, such exposures were not quantified.

As part of risk characterization (Tab D, Section E, "Risk Characterization"), investigators added together the risks due to multiple scenarios, exposure routes, and formulations in some cases, as justified by the evidence, to estimate cumulative risks.

The Pesticide Handlers Exposure Database (PHED) is used to estimate exposure to pesticide workers (applicators).^[194] PHED is a huge database of actual field monitoring data that provides dermal and inhalation exposure estimates based on type of pesticide product and application method, which are not chemical specific. Thus, a PHED user can identify appropriate scenarios to estimate exposure for many different pesticide active ingredients. The EPA has reviewed the exposure data in PHED for scientific integrity and has graded the data based on analytical quality assurance.^[195] Besides the PHED computer program software, there are a number of related PHED hard-copy documents. Of the latter, investigators used the PHED Surrogate Exposure Guide in conducting the exposure assessment. The data listed in the PHED Guide are associated with varying levels of confidence based on numbers of replicates and data quality as described in the PHED Guide. The levels of confidence listed are: high, medium, and low, and are included in the footnotes to the present exposure assessment.

(1) DEET

Investigators quantified both application and post-application exposure to DEET under Post-Application Scenarios below. Investigators did not quantify application exposure alone because they judged it to be inconsequential.

(2) Permethrin

Up to about 44% of servicemembers in the RAND survey (Table 8) indicated that they used permethrin or witnessed its use. However, the value of 44% is probably biased somewhat high, as the product was identified as "personal-use spray," and probably included a minor proportion of other products. Forty percent of the PM interviews cited use of permethrin, 0.5% aerosol. The 0.5% aerosol formulation, in a 6-oz spray can, was the most common form, and is the one evaluated in this exposure assessment. Permethrin was applied mainly to battle-dress uniforms (BDUs), and to a lesser extent to tents and bed nets. The spray was to be applied to clothing while it was hanging up then allowed to dry for 2 to 4 hours before being worn. The formulation in the can is a mixture of the active ingredient, and various inert ingredients such as solvents and/or propellants.

An important determinant of the level of absorbed pesticide active ingredient dose is the level of PPE worn by applicators. The PM interviews indicate that some servicemembers wore PPE; however, only 14% of servicemembers who identified permethrin addressed PPE. It is probable that the majority of servicemembers did not wear PPE when applying permethrin, as it would not normally have been required, unless someone was assigned the task of spraying many sets of BDUs. Thus, investigators evaluated the following clothing scenarios for servicemembers applying permethrin:

• Low exposure: long pants, long-sleeve shirt, gloves, half-face respirator with organic vapor cartridges and pesticide prefilters (respiratory protection factor = 90%)^[196]
  
  
• Medium and high exposure: long pants, long-sleeve shirt ("single layer, no gloves" from the PHED Guide).

Table 23 presents the assumptions for application of permethrin, 0.5% aerosol. As shown in Table 13 from the PM interviews, 26% of the servicemembers who identified permethrin aerosol indicated that it was used mainly outdoors, while 9% indicated that it was used mainly indoors; 65% did not provide information. The label directed the user to make all applications outdoors; thus, indoor application was a misuse.^[197]

Table 23. Permethrin assumptions for application

(3) Personal-Use Products Doses - Application

Investigators did not evaluate application exposure separately for DEET. Table 24 presents doses potentially resulting from application exposure to permethrin, 0.5% aerosol. There are three types of doses presented for the evaluation of noncarcinogenic effects: PDR_D, AD_D, and PDR_I. Table 25 presents the doses for the evaluation of potential carcinogenic effects for permethrin due to exposure during application. The two types of doses shown are LADD_D and LADD_I.

(1) DEET

The RAND survey determined 50% of servicemembers used DEET or witnessed others using DEET. Thirty-one percent of the PM interviews cited DEET, 33% stick/cream, while 18% cited DEET, 75% liquid. The DEET preparations used were similar to those widely available to the American public for many years. Servicemembers applied DEET to the skin as a repellent for numerous arthropods, including sand flies, mosquitoes, ticks, fleas, and mites. For purposes of this assessment, investigators assumed the conditions of use for DEET, 33% stick/cream are identical to those for DEET, 75% liquid. The only difference in the scenarios, exposure factors, etc. is the active ingredient concentration in the formulation.

Table 26 presents the assumptions for DEET post-application exposure. In this analysis, investigators considered application and post-application together. In reality, "application" is completed within a few minutes, and then post-application exposure continues for hours, until either all active ingredient evaporates, is absorbed, or is removed from the skin surface (e.g., while bathing). The only relevant exposure route is the dermal route. In a recent exposure assessment, the EPA quantified exposure via the dermal route, but not via oral or inhalation routes.^[204] Little if any DEET is likely to be ingested by adults. A small amount of DEET may be inhaled under some circumstances; however, the EPA does not provide an inhalation toxicity value with which to evaluate inhalation exposure (Tab D, Section D, "Toxicity Assessment")

Table 26. DEET assumptions for post application

Table 26 presents the source and rationale for each exposure factor. The application rate (AR) in units of mg formulation per cm² listed is based on input from the EPA. The EPA described a study by the DEET registrant where adults were asked to apply DEET with varying percentages of active ingredient.^[209] The registrant is the person or company attempting to register a pesticide product with the EPA. The mean amount of active ingredient per application was 925.25 mg active ingredient and 649.31 mg active ingredient for males and females, respectively. If one assumes the 925.25 mg active ingredient is applied over 5,000 cm² of skin surface area, then the AR for the study described is about 0.18 mg a.i./cm². Based on the latter result, one can estimate ARs of 0.55 mg and 0.24 mg formulation per cm² for the 33% and 75% formulations, respectively. Thus, the AR listed in Table 26 of 1 mg formulation per cm² is certainly conservative, being about 2-4 times that based on the study submitted by the registrant.

(2) Permethrin

Table 27 presents the assumptions for post-application exposure to permethrin, 0.5% aerosol. The label for a 6-oz can provides the following directions for use: "Treat the clothing for a minimum of 30 seconds on each side ... Use approximately � of this container to treat one complete field uniform. Use remainder on mosquito netting."^[210] Investigators assumed that 90% of the sprayed active ingredient adheres to the clothing, while the remaining 10% becomes suspended in the indoor air. Post-application dermal exposure occurred daily by migration of permethrin from the treated BDU to the skin. Inhalation exposure occurred for 8 hours in tents where servicemembers treated BDUs. Investigators assumed indoor application, a misuse of the product, occurred fairly infrequently for low-exposure and medium-exposure receptors, but more frequently for high-exposure receptors.

Table 27. Permethrin assumptions for post application

(3) Air Modeling for Permethrin

Investigators conducted air modeling to estimate the permethrin emissions and indoor air vapor concentrations resulting from treatment of one BDU inside a GP medium tent. A soldier sprays his/her uniform as directed, and other servicemembers in the tent receive a secondary exposure to the spray during this treatment and after the spraying is completed. The secondary exposure continues until the suspended permethrin is removed by ventilation or until servicemembers leave the tent. This secondary exposure to permethrin is considered in the air modeling analysis. For air modeling purposes, investigators assumed that an entire can is used to treat one uniform and that the treatment takes approximately 10 minutes to complete. Investigators also assumed that 10% of the contents of the can becomes suspended in the indoor air. Investigators assumed the airborne permethrin to be either vapor or fine mist (very small aerosol droplets that behave and disperse similarly to vapor).

(a) Permethrin Emission Calculation

Investigators calculated the emission rate of permethrin to the indoor air according to the following mass balance equation:^[216]

E = S x p x f x 1,000 / t

where,

E = emission rate, mg/min
S = amount of pesticide formulation applied, grams
p = fraction by weight of active ingredient in pesticide formulation, unitless
f = fraction of pesticide formulation which becomes airborne, unitless
t = duration of application, minutes

Based on the manufacturer’s label information and the assumption that the entire application takes 10 minutes to complete, investigators calculated the emission rate from treatment of one uniform as follows:

E = 8.505 mg/min
S = 6 oz = 170.1 g
p = 0.005 (0.5%)
f = 0.10 (10%)
t = 10 min

(b) Calculating Indoor Air Concentrations

Investigators calculated indoor air concentrations resulting from a single application event using a standard box model approach assuming complete mixing. The derivation of the box model is detailed below.

Investigators calculated indoor air concentrations in two phases. The first phase considers concentrations during the duration of the application (10 min). The second phase considers concentrations during the remainder of the 8-hour period following the onset of the application. Investigators calculated the average concentration during this period based on the average concentration during each phase (weighted by the duration of each phase).

One can develop the box model equation from mass balance considerations. The rate of change in the mass of active ingredient in the air inside the room is equal to:

the rate of emission of active ingredient to the air
plus the rate of transport of active ingredient from the outside
minus the rate of transport of active ingredient from the room to the outside
minus the rate of decay of active ingredient inside the room

One can write this in the form of a differential equation, which describes the change in concentration over time:

This equation has the following general solution:

C = [1/(I+K)][(E/V) + (C_a)(I)][1 - exp{-(I+K)(t)}] + C_o exp{-(I+K)(t)}

where:

C₀ = initial concentration in room (mg/m³)

Consistent with EPA draft guidance for conducting residential exposure assessments, the investigators assumed: 1) the active ingredients to be nonreactive (K = 0), and 2) contributions from outdoors to be negligible (C_a = 0)^.[217] Outdoor application of permethrin near tents almost certainly occurred. However, since conditions were usually very windy, the small volumes discharged would have been rapidly and extensively diluted, and the assumption of negligible contribution from outdoor to indoor air is justified.

With these assumptions, the equation for concentration within the room simplifies to:

C = [E/(I)(V)][1 - exp{-(I)(t)}] + C₀ exp{-(I)(t)}

For an initial concentration of zero (C₀ = 0), this equation for concentration at time t simplifies to the following expression:

C = [E/(I)(V)][1 - exp{-(I)(t)}]

Investigators applied this equation to calculate concentrations in the room during the assumed 10-minute period when spraying of the uniform occurred (i.e., during Phase I). The concentration in the room at the end of the application period is given by:

C₁₀ = [E/(I)(V)][1-exp{-(10)(I)}]

One obtains the average concentration over the time interval from t₁ to t₂ by integrating the concentration equation over the interval and dividing by the duration of the interval:

C_avg = [1/(t₂-t₁)][E/(I)(V)][(t₂-t₁) + (1/I)(exp{-(I)(t₂)} - exp{-(I)(t₁)})]

The average concentration over the first 10 minutes (t₂ = 10 and t₁ = 0) is then given by:

C_I = (1/10)[E/(I)(V)][10 + (1/I)(exp{-(10)(I)} - 1)]

For Phase II exposure (following the end of application), the initial concentration (C₀) at the beginning of Phase II is the same as the concentration at the end of Phase I. Returning to the general box model equation presented earlier and again setting C_a = 0 and K = 0, the concentration in the room during Phase II is given by the expression:

C_t* = C₀ exp{-(I)(t*)}

where,

t* = time (in minutes) following the end of application
C₀ = [E/(I)(V)][1-exp{-(10)(I)}]

One can obtain the average concentration during Phase II over the interval from t₁* to t₂* by integrating this equation over the interval and by dividing by the duration of the interval:

C_avg = [C₀/(t₂*-t₁*)][-1/I][exp{-(I)(t₂*)} - exp{-(I)(t₁*)}]

Setting t₁* = 0 and t₂* = 470, the average concentration in the 470 minutes following the end of the application (i.e., during Phase II) is given by:

C_II = [C₀/(470)][-1/I][exp{-(470)(I)} - 1]

The average concentration over the entire 8-hour period encompassing the 10 minutes of application (Phase I) and the remainder of the period following application (Phase II) is then given by:

C_avg = (1/480)[(10)(C_I) + (470)(C_II)]

Investigators calculated low, medium, and high permethrin exposure levels. The three cases differ only in the fresh air infiltration rates investigators assumed for the tent. Investigators assumed ventilation rates of 6, 4, and 2 air changes per hour for the low, medium, and high indoor exposure scenarios. Investigators believe these rates are reasonable given reports that strong winds readily penetrated the tents. The air volume of the tent was set to 123 m³ to correspond to that for a GP medium tent.

Although the EPA recommends the use of the Multi-Chamber Concentration and Exposure Model (MCCEM)^[218] in conducting high-end exposure assessments, investigators used the simple box model equations instead for several reasons. The tent is treated as a single chamber structure, so investigators did not need the multi-chamber capability of MCCEM. In addition, the default ventilation rates included in MCCEM were not realistic for the structures and scenarios under consideration. Nonetheless, the basic physics embodied in the box model equation and in MCCEM are the same.