Post-Vietnam Dioxin Exposure in Agent Orange-Contaminated C-123 Aircraft (2015)

4

Evaluation of Assessments of Possible Exposure of Air Force Reservists from Service in Operation Ranch Hand C-123s

Before the US Department of Veteran Affairs (VA) asked the Institute of Medicine (IOM) to convene this committee, a number of parties had expressed opinions about the sampling results¹ from C-123 aircraft that were formerly used to spray herbicides in Vietnam and were subsequently used by Air Force (AF) Reserve personnel in the United States between 1972 and 1982 (detailed sample results are presented in Chapter 3 of this report). The positions developed were based on qualitative, or at most semi-quantitative, treatment of the sampling data and their sources fall into two general categories:

• interpretations by individuals and entities associated with the military (USAF, 1994, 1997b, 2009a,b, 2012a,b; Young and Young, 2012, 2013a), and
• statements provided to the C-123 Veterans Association (ATSDR, 2012, 2013a,b; Berman, 2011; NIEHS, 2011, 2013; Schecter, 2013; Stellman, 2012, 2013).

_______________

¹ The one exception is that the results of a second dioxin sampling of “Patches” in 1995 were not entered into the data set until spring of 2014, when their existence was pointed out by Nieman (2014) in a letter to the editor responding to Lurker et al. (2014).

In this chapter, the committee reviews the approaches adopted in these documents.

Just as this committee started its work, Lurker et al. (2014) published a paper that used the available C-123 sampling data in a much more quantitative fashion in three models of exposure. The strengths and weakness of these estimation models are also addressed in this chapter.

The committee then compares the existing indoor contamination guidelines for 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) (as presented in Table 2-1) to the range of exposures supported by the existing sampling data. This serves to put in context the degree to which the AF Reservists may be at increased risk of adverse health consequences in association with exposure to components of the herbicides sprayed in Vietnam from working in Operation Ranch Hand (ORH) planes after they were returned to the United States.

This chapter concludes with the committee’s integration of these various pieces of information in a qualitative fashion. From this, the committee makes its judgment about whether the existing information related to sampling of the ORH C-123s support there having been a meaningful elevation in the risk of adverse health effect among the AF Reservists.

INTERPRETATIONS OF SAMPLING RESULTS BY ENTITIES ASSOCIATED WITH THE MILITARY

Between 1979 and 2009, air and wipe samples were taken from several C-123 aircraft that had formerly been used to spray herbicides in Vietnam and were then used by the AF Reservists in the United States from 1972–1982. These sampling efforts are detailed in Chapter 3 of this report. Internal memos, evaluations, and laboratory reports provided by the VA regarding the sampling included some comments regarding elevated levels, but their bottom lines were largely dismissive about the possibility of there being potential health hazards related to exposures to these aircraft. Specific instances are noted below:

• After noting that the results for 2,4-D and 2,4,5-T in air were below the 10 mg/m³ threshold limit value for each of these herbicides, the conclusion was “Sample results show contamination levels to be below amounts considered to be possible health hazards” (USAF, 1979).
• Another AF report concluded “the interior of the C-123 aircraft under discussion is heavily contaminated with PCDDs.” When referring to the 25 ng/m² standard for office workers established by the National Research Council (NRC) (1988) as applied to 70-kg restoration workers for 375 days, however, the report also noted that “a higher surface concentration would be acceptable” and recommended that, during restoration, exposure “be maintained at the lowest possible level” by use of protective gear and procedures (USAF, 1994).

• “[T]he C-123 exterior and the majority of the interior are not contaminated with PCDDs or PCDFs above detectable levels . . . contamination is confined to a very small area of the plane’s interior and to the inside of the rear inspection ports” (USAF, 1995a).
• A consultative letter (USAF, 1997b) addressing the 1996 sampling at Davis-Monthan Air Force Base (AFB) of two ORH C-123s for TCDD and 12 for phenoxy herbicides, concluded that “there is potential for individual exposure” and that, before sale or transfer, at least 10 TCDD samples should be gathered from each plane and any with detectable dioxin needed to be fully decontaminated.

Reports in Preparation for Smelting of C-123s at Monthan-Davis AFB

In 2009, air and surface sampling was conducted for TCDD and phenoxy herbicides in 2 ORH C-123s and another 10 ORH C-123s were tested only for herbicide residues on surfaces (USAF, 2009a,b). The final report concluded that “concentrations of 2,4-D and 2,4,5-T detected inside the aircraft were very low with respect to risk-based screening levels of concern and do not pose a significant risk” and that “low levels of dioxin/furans, near the risk-based screening level, on all interior surfaces that were sampled” represent “low level contamination that does not pose a significant risk to personnel involved in short term recycling activities” (USAF, 2009b). No comments were made about the experience of the AF Reservists.

The risk-based screening levels presented in the 2009 USAF report (USAF, 2009b, Appendix F, and reproduced here in Table 4-1) were said to incorporate consideration of both oral and dermal exposure resulting from contaminated surfaces. The committee has determined that the aggregate screening levels derived from the pathway-specific surface contamination screening levels for the dermal and oral routes shown in Table 4-1 were calculated incorrectly. The USAF report calculated the screening levels by simply adding the individual screening levels for the two routes, rather than by using the inverse-of-the-sum-of-the-inverses formula, which is the US Environmental Protection Agency’s (EPA’s) recommended method for calculating aggregate screening levels (EPA, 2001).

As a result, the screening levels used were established on the basis of the less sensitive rather than the more sensitive of the two pathways; hence the derived guideline levels are not necessarily protective. Had the aggregate screening levels been calculated correctly from the stated dermal and oral screening levels, the TCDD result would be 1 ng/m² and the results for 2,4-D and 2,4,5-T would be 200 µg/m². For the two C-123s sampled in 2009 (see Tables 3-2 and 3-3), the average surface contamination levels reported for TCDD (14 and 16 ng/m²), 2,4-D (590 and 590 µg/m²), and 2,4,5-T (520 and 500 µg/m²) exceeded these values.

TABLE 4-1 AF Screening Levels the Assessment of Surface Contamination with Committee’s Correction

SOURCES: Appendix F: Risk Screening Level Assessment, Table 1 (USAF, 2009b,d).

US Air Force School of Aerospace Medicine Report

In April 2012, the US Air Force School of Aerospace Medicine released a report commissioned by the Headquarters Air Force Medical Support Agency detailing an exposure assessment related to Agent Orange (AO) for UC-123 aircraft previously used in support of ORH in Vietnam (USAF, 2012a). The cover memo to the report (USAF, 2012b) summarized that

• There was no relevant personal exposure or laboratory data found.
• [E]xisting information is inadequate to accurately determine individual exposure.
• [I]t is unlikely that the exposures experienced between 1972 and 1982 would have been sufficient to cause harm.
• Given the absence of a clear finding of potential harm, we believe it unnecessary to relay such individual findings to persons who had entered or worked on C-123s between 1972 and 1982, and who may be unaware of this assessment.

The USAF (2012a) review described the USAF sampling study conducted in 2009 as being the “most comprehensive” and it adopted the screening levels and the associated conclusions from the 2009 report as valid. The committee, however, has determined that these screening levels were incorrectly calculated (as described in the previous section) and therefore cannot be assumed as protective as asserted by the USAF review.

The USAF review (2012a) also concluded that the air sampling data for the phenoxy herbicides were “within acceptable exposure limits” although no air exposure limits were presented in this document, and it is not clear what exposure limits are being referred to by the authors. The committee has concluded that the air sampling data were minimal and of unknown quality. Two of three air samples from a former ORH C-123 were collected in 1979 and tested positive for herbicide exposure. In addition, air samples from two other ORH planes were collected in 2009, decades after the relevant period of exposure, using a low-volume screening protocol. Those samples were acquired at ambient temperature and pressure unlikely to be generally representative of in-flight work conditions of the AF Reservists. All the 2009 TCDD air samples (including blanks and controls) were reported as non-detect without corresponding quality assurance/quality control (QA/QC) materials, bringing into question whether the laboratory conducting the analysis could in fact capture and extract relevant amounts of TCDD with the sampling media used.

The Executive Summary of the USAF report concluded that, since there is a lack of TCDD “regulatory standard or consensus standard of practice . . . in the occupational health profession, application of wipe sampling data to estimate personal occupational exposures is not warranted” (USAF, 2012a, p. 4). The

committee concurs that there is no agreed upon method for estimation of dose from wipe sampling data, but notes (as reviewed below) that multiple agencies and individuals have attempted to make such estimates. Based on those efforts, it is unreasonable in this case to assume that uncertainty regarding the threat presented by observed surface loads can be interpreted as evidence of negligible risk. Dismissal of the results from wipe sampling was not justified.

The last of the four points in the conclusions section of the report (USAF, 2012a, p. 5) reads:

The committee finds this paragraph problematic in three respects:

• The report’s “assessment of risk” could not have been “dependent upon the findings of the National Academy of Sciences’ Institute of Medicine RANCH HAND studies” (USAF, 2012a); that work is officially known as the Air Force Health Study because it was conducted by the Air Force, and is in no way a product of the IOM or any other part of the National Academies.
• It asserts that “It is reasonable to assume that any exposures associated with HO [Herbicide Orange] in post-Operation RANCH HAND utilization of the UC-123 would likely be less than exposures associated with HO during Operation RANCH HAND” (USAF, 2012a). The committee agrees that workday exposures during active use of AO were likely greater than post-war exposures on a comparable task basis. However, ORH personnel were in Vietnam for a median of about 320 days (CDC, 1988), and undoubtedly did not access spray aircraft on every day of in-country service. Post-war exposures to AF Reservists would have been generally less frequent on an annual basis, but may have extended for up to 12 years. So some fraction of the AF Reservist cohort could have conceivably spent more time in contaminated C-123s than did some fraction of the ORH cohort (especially if the AF Reservists worked full-time or were scheduled for additional hours throughout the year). In addition, some post-war tasks

• could have resulted in workday exposures that exceeded the workday exposures of some less-exposed ORH personnel (for example, flight crew officers) in terms of overall duration, more work inside the planes on the ground under conditions of reduced ventilation, etc.

• In asserting “Consistent with the findings of the National Academy of Sciences’ Institute of Medicine biennial report ([Update] 2010), it is reasonable to conclude that it is not possible to derive quantitative estimates of any increased health risks [emphasis added] for” the AF Reserve personnel, the AF report (2012a) misrepresents the charge and determinations of the Veterans and Agent Orange (VAO) committee. The second task assigned to VAO committees by the 1991 Agent Orange Act was to estimate risk to Vietnam veterans for each specific adverse health outcome. These outcome-specific risks are what are being referred to in the statement “estimation of risks experienced by veterans exposed to the chemicals of interest during the Vietnam War is not feasible” (IOM, 2012, p. 11). Not only is determination of potency factors from epidemiologic results for each health problem far from being feasible, there are no exposure data to be drawn upon for non–Ranch Hand Vietnam veterans. In contrast, the issue concerning the AF Reservists and ORH C-123s is whether the risk of adverse health outcomes of any sort was elevated among those with potential exposure. This committee considers the sampling data for the AF Reserve C-123 population, albeit limited, to be an adequate basis for a qualitative exposure assessment and for some judgment about the overall health risk from TCDD exposure that the AF Reservists may have experienced, and so finds this justification for not doing so from the authors of the USAF report (2012a) to be invalid.

Investigations into the Allegations of Agent Orange/Dioxin Exposure from Former Ranch Hand Aircraft

In a report with the above title (Young and Young, 2012), a similarly named one on 2,4,5-T (Young and Young, 2013b), and a formal briefing document (Young and Young, 2013a), the VA’s Compensation Service received an interpretation of the “dry” Agent Orange residues found on surfaces inside C-123s used by AF Reservists after their use in Vietnam, which has been incorporated onto VA’s webpage concerning this issue (http://www.publichealth.va.gov/exposures/agentorange/locations/residue-c123-aircraft/scientific-review.asp; accessed August 21, 2015). The committee found this description of the chemical properties and behavior of TCDD and its propensity for dermal absorption to be inaccurate. Notable issues include the following:

• Young and Young often refer to TCDD residues as “dry” and immobile. Semi-volatile organic compounds (SVOCs) that are nominally solid at

• room temperature are generally not found in pure, crystalline form in the environment. Even below their melting point, dilute SVOCs do not crystallize and are found sorbed or dissolved in various matrices (Mackay, 2001). Because TCDD is not deliberately manufactured, but only exists as a very dilute contaminant, it would be expected to behave as a sub-cooled liquid rather than a solid. Therefore, it would not be immobile as Young and Young asserted.

• No attempts to investigate the removal efficiency of the selected method or of alternative methods for surface sampling for TCDD specifically have been reported. Unlike the irreversible binding of organics by activated carbon, binding to metals would be expected to be negligibly small. For instance, when investigating surface wiping techniques, Deziel et al. (2011) used stainless steel as a substrate to avoid issues associated with wiping porous surfaces, and Slayton et al. (1998) stated “wipe-sampling procedures provide semi-quantitative data for non-porous surfaces (i.e., metal) but are considered poor for porous surfaces (i.e., concrete).” The claim that hexane is required to remove TCDD from surfaces in the C-123s appears to be conjecture and not evidence-based.
• The assertion that “studies of dermal contact with TCDD have found that any exposures that occurred were ‘negligible’ because the skin is a major barrier to TCDD uptake, contributing less than 1% over the long term to the body burden” (Young and Young, 2012, p. 3) is based on interpretation of a Dow-funded study of the link between soil contamination and body burden in nonoccupationally exposed adults (Kimbrough et al., 2010). The passage conflates dermal exposure with exposure to contaminated soil. Kimbrough et al. do not discuss dermal absorption, dismissing it a priori as “a minor contributor to body burdens of the general population.” However, dermal exposure can be an important source of exposure in occupational settings. For instance, the Dioxin Registry Report for the Dow manufacturing facility in Midland, Michigan (NIOSH, 1991) notes that “air sampling for chloracnegens were [sic] discontinued by 1966 because skin contact was recognized as the primary route of exposure.” Similarly, Kerger et al. (1995) state “The available literature suggests that dermal uptake of dioxin in the workplace may be the primary source of occupational exposure.” It should be further noted that, at least initially, the TCDD in the C-123s would have been dissolved in herbicides, which are themselves well-absorbed (Harris and Solomon, 1992; Moody et al., 1990) and that at no point would the surface matrix in the planes have been soil.
• The claim that TCDD is not volatile below its melting point is also incorrect. As discussed in Chapter 2, TCDD behaves like other SVOCs with similar physical-chemical properties. At room temperature, it is generally dissolved or sorbed into a film and is in constant flux around equilibrium with surrounding media.

STATEMENTS PROVIDED TO THE C-123 VETERANS ASSOCIATION

The C-123 Veterans Association (CVA) circulated the available sampling data to several dioxin researchers requesting replies regarding their interpretations of possible exposures (ATSDR, 2012, 2013a,b; Berman, 2011; NIEHS, 2011, 2013; Schecter, 2013; Stellman, 2012, 2013). The statements that were provided agreed with the CVA position that the sampling data support the possibility that the AF Reservists’ risks of adverse health effects were increased by their service in ORH C-123s.

Interpretation of the Agency for Toxic Substances and Disease Registry

Prior to the publication of Lurker et al. (2014), which is discussed in the next section of this chapter, the most detailed response to the CVA came from the Agency for Toxic Substances and Disease Registry (ATSDR), which provided three documents with the same approach and conclusion (ATSDR, 2012, 2013a,b). The first from Sinks compared the average TCDD concentration from surface wipe samples collected from the interior of “Patches” (USAF, 1994) to TCDD screening guidelines corresponding to a 10⁻⁶ cancer risk, as recommended by the US Army Center for Health Promotion and Preventive Medicine (CHPPM) in Technical Guide 312 (TG 312) (CHPPM, 2009). TG 312 derived surface wipe screening levels (SWSLs) for 70-kg office workers over a 10-year period assuming exposure to contaminated surfaces through dermal contact and absorption, incidental ingestion by hand-to-mouth behaviors, and inhalation through breathing resuspended particulates. Environmental samples with concentrations above the SWSL for TCDD indicate the need for a more thorough health risk assessment.

ATSDR calculated an average TCDD surface concentration of 636 ng/m² for the three C-123 interior wipes collected in 1994. ATSDR concluded that this value exceeded the TG 312 screening guideline of 3.5 ng/m² by a factor of 182, which corresponds to a 200-fold greater cancer risk than the screening value. The opinions expressed in the initial report were subsequently upheld twice, first by Ikeda and by Portier (ATSDR 2013a,b), and again in a presentation to this committee on June 16, 2014, by Sinks (ATSDR, 2014).

The initial ATSDR opinion letter acknowledged the limitations of the available data, most notably the questionable representativeness of TCDD surface concentrations from the sampled C-123s to surface levels in the other ORH C-123s and the 20- to 40-year lag time between when the AF Reservists worked on them and collection of the comparison surface wipe samples. ATSDR further acknowledged a lack of information on flight crew activities, work histories and duration of work in C-123 aircraft, and minimal information regarding their interior environment at the time when the C-123s were used by the AF Reservists. Additionally, ATSDR noted that the TG 312 SWSL was derived from an office worker scenario, and thus, was likely to be under-protective for the TCDD expo-

sures of the AF Reservist personnel when they worked inside the confined space of an aircraft. Representative exposure levels of the AF Reservists would depend on skin surface area, duration of exposure, hand washing, and food intake. In their June 2014 presentation to the committee, ATSDR additionally pointed out that the newly available sampling data had been collected from wipe samples from different aircraft and sequential sampling of the same surfaces in a given aircraft had not been conducted. Thus, it is not possible to infer degradation rates for TCDD on the sampled surfaces. Further, the purpose of the additional surface wipe sampling was for estimating exposure risk for personnel preparing the aircraft for destruction or recycling, not for retrospective evaluation of exposures to C-123 Reservists.

ATSDR’s summary interpretation of the available information was as follows:

• inhalation exposures to TCDD in C-123 aircraft could not be excluded;
• aircrew operating in “this and similar environments were exposed to TCDD”;
• it was not possible to accurately establish the degree of exposure (high or low), or the risk of adverse effects among C-123 AF Reserve flight crew;
• the contamination levels in 1994 in at least one plane greatly exceeded current Department of Defense screening guidelines; and
• the observed levels would likely have required the use of personal protective equipment or the grounding of these aircraft.

These observations from ATSDR are largely in accord with the committee’s assessment of the available data. A comprehensive comparison of the wipe samples reviewed by ATSDR to the TG 312 screening guidelines is provided in Table 4-2.

CRITIQUES OF THREE MODELS PRESENTED IN LURKER ET AL. (2014)

The CVA had provided Jeanne Stellman with the then available data from sampling of the C-123s (that is, without the sampling conducted on “Patches” in 1995). Dr. Stellman and colleagues, Fred Berman and Richard Clapp, entered into collaboration with Peter Lurker, who had been involved in the sampling conducted on the C-123 aircraft at Monthan-Davis AFB. A report of this group’s efforts to perform quantitative exposure estimation using these data (Lurker et al., 2014) was published just before this committee was convened.

Lurker et al. (2014) presented three modeling approaches to address potential exposures to C-123 flight crews and maintenance personnel in post-Vietnam operations. The first approach involves estimation of exposure due to hand-to-mouth contact using a protocol based generally on prior exposure assessment guidance from EPA (1989) developed for management of hazardous waste sites and more specifically on May et al. (2002) and US Army guidance in TG 312 (CHPPM,

TABLE 4-2 ATSDR’s Comparison of Interior Surface Wipe Samples Collected from C-123 to Screening Guideline of 3.5 ng/m² TCDD from TG 312 (CHPPM, 2009)

^a This committee found these two sample readings without specification of the area sampled could not be compared with other results in units of ng/m².

^b The range of surface wipe concentrations from each sampling period has been added by the committee.

SOURCES: ATSDR, 2012, 2013a,b, 2014.

2009) developed for evaluation of risks to office workers from exposure to chemicals on indoor work surfaces. The committee calls this the “Dermal-oral Direct Contact Model.” Lurker et al.’s (2014) second approach evaluated the potential degree of vapor supersaturation of herbicides in air samples collected in a C-123 in 1979 and then predicted potential inhalation exposures to vapor and particle-bound TCDD. The committee calls this the “Maximum Saturation Vapor Pressure Model.” Lurker et al.’s (2014) third model adapted the methodology of Little et al. (2012), which uses an approach based on thermodynamic arguments to predict indoor air concentrations from surface residues of SVOCs. The committee calls this the “Thermodynamic Emissions Model.” Exposures that were predicted using this third model were again attributable solely to inhalation. Lurker et al. (2014) interpreted the results of their modeling as supporting the premise that the AF Reservists had experienced exposure to residual herbicide components that could have exceeded guidelines of EPA, The Netherlands, and the World Health Organization.

The committee endeavored to reproduce the estimates Lurker et al. reported from these three models as a means of validating the models and of gaining full understanding of the assumptions that were made. With the validated models, the committee was able to explore the sensitivity of their results to various scenarios and changes in assumptions. The committee found none of the estimation models

assessed to be without weakness and considered no particular exposure scenarios for the AF Reservists to be well documented. Consequently, the committee members did not find any specific quantitative estimates likely to be representative of the range of exposures experienced and so refrained from presenting any quantitative exposure estimates that might be construed as representing predictions it favored for the AF Reservists’ actual exposures. This decision is also in accord with the committee’s charge to conduct a qualitative assessment.

Dermal-oral Direct Contact Model

Although abbreviated as “dioxin dermal-oral exposure,” the first model in Lurker et al. (2014) can more appropriately be described as a dose estimate of non-dietary oral ingestion due to hand-to-mouth contacts. The model has roots in EPA’s guidelines for risk assessments at Superfund sites (EPA, 1989). EPA’s original equation for intake (I) in mg/kg body weight-day is

where CR is the contact rate, EFD is exposure frequency and duration, C is the contaminant concentration, BW is body weight, and AT, the averaging time.

Lurker et al. (2014) added terms to correct for the sampling efficiency of the wipe technique for surface concentration estimation (FT_we), and to convert units representing surface area (CF_a) and expanded the terms for contact rate and for exposure frequency and duration in the original equation as follows:

• CR = (SA)(FT_ss)(FT_sm)(FT_re)(CF_wt)(FT_ga)
• Here, SA is the exposed skin surface area, FT_ss is the fraction of the mass of the contaminant that is transferred from the surface to the skin, FT_sm is the fraction of skin area that touches the mouth, FT_re is the fraction of contaminant transferred from the skin to the mouth, CF_wt is a conversion factor, and FT_ga is the fraction of contaminant absorbed from the gastrointestinal tract.
• EFD = (RH)(EF)(WD)(ED)
• Terms used include RH, the probability of being on a Ranch Hand aircraft; EF, the count of hand-to-mouth events per day; WD, the number of work days per year; and ED, the duration of exposure.

The equation in Lurker et al. (2014)² after these substitutions is

May et al. (2002) and TG 312 (CHPPM, 2009) were both mentioned in Lurker et al. (2014) as sources that also had adapted EPA’s intake model. The committee found that the three versions differ slightly in how they compute the ingestion due to hand-to-mouth events. Most notable are terms explicitly accounting for the frequency of contacts between the skin and surfaces (in the May model) and the fractional surface area involved in the contact between skin and surfaces (in the CHPPM model). Lurker’s model does not seem to account for either of these factors. Aside from some confusion on the interpretation of this transfer input, the Lurker et al. (2014) model is specifically for non-dietary ingestion dose due to hand-to-mouth contacts. The model does not aim to include non-dietary ingestion due to object-to-mouth contacts, let alone inhalation, dermal, or dietary ingestion exposure. This exclusivity could be a problem when the contaminant of interest is assumed to reside in multiple media and where other routes of exposure may be important.

Additionally, this model embodies a simple linear estimate. The direction in which an increase or decrease in the value of any input variable will modify the resulting exposure estimate is entirely predictable. The model does not account for chemical-dependent properties (for example, transport, decay, partitioning) or uncertainty in input parameters. It can be used for individuals of different body weight, but aside from this, there are some restrictions in the ability to include varying activities (such as contact with surfaces) from person to person or over time.

This model is quite representative of commonly used models for exposure estimation, but almost any result could be obtained depending on what values are assumed for the large number of input variables. A broad spectrum of values is usually feasible for any exposure situation under investigation, but the extensive uncertainties about the actual work histories of the AF Reservists in this case make this model even more flexible. The committee agrees with the point made by Driver (Driver and Solomon, 2014) at the committee’s workshop that with its multiplicity of component variables, each of which can be assigned a wide range of values, the model is too plastic to provide any real insight into what levels of exposure might actually have occurred.

____________________

² Personal communication with the authors confirmed that the term FT_sm was missing from Equation 2.

Maximum Saturation Vapor Pressure Model

The saturation vapor pressure model presented in Lurker et al. (2014) is a simple thermodynamic model premised on the assumption that the maximum vapor air concentration in an enclosed environment will be a function of the saturated vapor pressure. However, applying it to semi-volatile compounds in a mixture can present a number of challenges. In addition, the saturated vapor pressure is temperature- and pressure-dependent. To apply to an air concentration, a closed system is assumed with no air exchange. This is not the case when an aircraft is in use, however, so this pathway would represent a maximum concentration that would result through volatilization under static conditions. The herbicide 2,4-D or 2,4,5-T concentrations in the Lurker paper were derived by assuming each was a pure substance at 760 mm Hg air pressure. These substances are solids at room temperature. If they are dissolved in an oil-based film then the mole fraction of other compounds in the film with similar vapor pressure should be considered. Driver and Solomon (2014) noted that the air concentrations to which the calculated vapor pressures were being compared were not the acid forms of 2,4-D and 2,4,5-T, but rather the esters which have a considerably higher vapor pressure.

Based on the premise that the 2,4-D and 2,4,5-T acid vapor concentration calculated was lower than the measured concentration, the measured air concentration would contain both vapor phase and particulate phase of these compounds. The next step was to estimate the TCDD air concentration by assuming that, because the 2,4-D and 2,4,5-T were in both the vapor and particulate phase and TCDD is less volatile than either of the herbicides, it also would be partitioned from the vapor phase to a particulate phase. The committee was unable to reproduce the TCDD air concentrations reported in the Lurker paper by following the indicated calculations with the specified inputs. There are a number of questionable assumptions involved in these calculations. Driver and Solomon (2014) objected that the value used in the Lurker paper for the concentration of TCDD in AO represents the upper range of what has been found to have been present in AO. Lurker et al. (2014) also assumed that the degradation of TCDD within the aircraft was similar to that of the herbicides within the aircraft, which is questionable.

Independent of whether the model predicts appropriate saturated vapor pressure for TCDD, the applicability of this model for an inhalation route is questionable because, once the doors and hatch of the aircraft were opened, the air exchange would increase within the aircraft and while the aircraft was flying the increased air exchange would drastically reduce the air concentration associated with volatilization. The basic premise is that there would be redistribution onto particles of SVOCs with resulting exposure. However, the TCDD concentration on the particles cannot be determined by this method. The committee found that this model is unsuitable for estimation of exposures experienced while a contaminated plane was in flight and that its use is further

hindered by the difficulty of establishing appropriate values for the saturation vapor pressure.

Thermodynamic Emissions Model

The third model in Lurker et al. (2014) was a general thermodynamic model developed for emission of SVOCs, a class of compounds that includes TCDD. The model is based on a screening level model described in Little et al. (2012), which was proposed for use in health-based evaluations of chemicals prior to entry into commerce. Lurker et al. (2014) applied this model by using an expected gas phase concentration immediately above the surface of the source material, the exchange between the air and organic film expected to be throughout the aircraft on surfaces and dust based on an octanol-water coefficient, the surface area of the interior of the aircraft, and the ventilation rate. This approach can provide a steady state air concentration in a specified microenvironment with a constant air exchange rate and a source of SVOCs that is not exhausted over the time period of interest.

This is an appropriate model to use for TCDD in the post-Vietnam ORH C-123 aircraft if the input parameters can be properly established. However, the committee identified problems for several of the input parameters presented in the Lurker paper:

• The value given in Table 6 of Lurker et al. (2014) for the convective mass-transfer coefficient (h) appears to be an order of magnitude low. Based on the air concentrations given in the paper, it appears that a value of 3.68 m/h (rather than the tabled value of 0.368 m/h) was actually used.
• The ventilation rate of 170 m³/h, which is said to be adapted from Meek (1981), is much smaller than the ventilation rate of 8,464 m³/h for a flying C-123 aircraft reported in that document. It should be noted that the crew were also in the aircraft before and after flights and during training session while on the ground when the ventilation rate would be only a fraction of the 170 m³/h.
• The interior surface area was underestimated by assuming a cylindrical shape because portions of the interior were exposed substructure of the aircraft which present a much larger surface area than a smooth surface.
• The largest uncertainty is associated with y₀, the gas-phase concentration in contact with the emission surface, which was calculated using maximum saturation vapor.

As indicated above there are a number of problems with the assumptions in the Lurker paper calculations. Even if those values were accepted as accurate, Lurker et al. (2014) corrected the gas phase TCDD concentration for the portion that would be on the particle phase, which should not be included in y₀. Estimat

ing y₀ is recognized as the largest shortcoming of this model approach (Little et al., 2012). Thus, the committee does not have confidence in the TCDD air concentrations calculated from this model based on the input parameters used.

The approach taken by Little et al. (2012) could, however, reasonably be applied to a scoping analysis of the inhalation exposures to the C-123 flight crews. Based on that approach, vapor phase air concentrations would be expected to be highest while planes were stationary and lowest during flight. Time spent in the planes while they were on the ground is, therefore, a key determinant of exposure due to inhalation and dermal absorption of vapor. Based on projected on-ground air concentrations, and given that aggregate exposures would be expected to be a multiple of vapor inhalation exposures, average workday exposures to flight crews could be problematic, but maintenance personnel whose on-ground time in the planes exceeded that of flight crews, or flight crews who participated in static training missions, probably had even higher exposures.

UNDERESTIMATES RESULTING FROM ALL POSSIBLE PATHWAYS NOT BEING FACTORED INTO SOME GUIDELINES

The committee used the existing indoor contamination guidelines for TCDD that were presented in Chapter 2 as a means of assessing the degree to which possible exposures of the AF Reservists may indicate possible adverse health consequences. The committee considered the guidelines for surface loading the most applicable to the occupational situation of the AF Reservists. These guidelines for surface loading range from 1 to 25 ng/m², with 3.5 ng/m² being the level derived by CHPPM and 22 ng/m² being the guideline derived in the 2009 AF report. (More correctly, the AF value should have been 1.1 ng/m², if dermal exposure actually had been taken into account; see Table 4-1.) Although such guidelines are derived with the intention of amply protecting the health of exposed individuals, some guidelines may not be as health-protective as the risk level nominally stated. The committee finds that the existing indoor contamination guidelines for TCDD (see Table 2-1) expressed in terms of surface loading or air levels are incomplete with respect to consideration of all possible exposure pathways.

Table 4-3 indicates which of all possible exposure pathways were factored into the guidelines derived for building re-entry, and which the committee found to be most applicable for the situation of the AF Reservists. Examination of the bases for the surface loading standards shows that they are largely driven by estimated ingestion exposures related to hand-to-mouth contact. Hand contamination is presumed to occur via contact with contaminated surfaces. Object-to-mouth contact is discussed in some of the relevant documents, but not actually included in dose estimation. Inhalation exposure is more often formally addressed, but typically not directly tied to surface contamination (even though, in a physical-chemical sense, air and surface residues are contiguous and exchanging phases). Hence, separate standards are derived for ingestion and inhalation pathways,

TABLE 4-3 Exposure Pathways Considered in Prior Surface Standard Relevant Literature

^a Adopted by NRC, 1988.

^b Included but not explicitly linked to surface contamination.

with the proviso that measured contamination in either phase should lower the acceptable level of contamination in the other. Dermal exposure is often considered, but only as a consequence of direct contact with surfaces. As discussed below, dermal protocols use rates of contact with surfaces that are low compared to traditional values used in occupational and environmental health, assume that normal clothing is chemical-protective, and address dermal absorption in a crude manner. Dermal absorption of vapor-phase TCDD is not considered in any of the proposed protocols. Given that some potentially important pathways are ignored, and that others are treated in a perfunctory manner, the degree of protectiveness of the existing guidelines expressed in terms of surface loading or air concentration should not be assumed.

Transfer Coefficients

Transfer coefficients (TCs) are routinely used to characterize occupational exposure to contaminated surfaces. A TC represents the equivalent surface area from which 100% of the dislodgeable chemical residue is removed and transferred to the skin or clothing of a worker per unit of time. Application of this approach to assess the use of pesticides in indoor spaces is proceeding out of necessity, but is less firmly established than for agricultural applications due to greater heterogeneity in indoor surfaces and behavioral patterns. Nevertheless, methods for translating surface contamination to human dose are needed, and the TC approach is a likely candidate. Derivations of the guidelines for contamination of building interiors discussed here typically do not explicitly use an estimated TC (the 2003 World Trade Center [WTC] document is the exception). However, implicit transfer coefficients are inevitably found in all of the guidelines expressed in terms of surface loading. In agricultural occupational health practice, TCs routinely exceed 1,000 cm²/hr. The WTC protocol utilized

a TC of 1,200 cm²/hr, but described it incorrectly as a skin contact rate and then inappropriately applied a surface-to-skin transfer efficiency.

Dermal Absorption of Residues on Skin

TCDD’s rate of absorption into skin can be estimated from surface loads, transfer coefficients, and resulting predicted dermal doses. Fluxes are lower than available experimental results. For instance, the 3% availability assumed in the WTC protocol is based on absorption experiments conducted by Poiger and Schlatter (1980), who applied TCDD in soil to rats in vivo. Observed flux in the experiment producing the 3% estimate was nearly 500 pg/cm²-hr or 10⁶ times the flux implicit in the WTC protocol. Fraction absorbed is not independent of chemical load on skin (Kissel, 2011). Sampling skin surfaces with low loading is unlikely to provide an appropriate measure of potential dermal dose due to depletion, whereas testing at a high loading is unlikely to show dose dependence due to saturation. Application of fractional absorption data from high load experiments to low load conditions can lead to gross underestimation of absorption efficiency. Dermal absorption of the phenoxy herbicides has been shown to be substantial in laboratory animals and in human volunteers (Harris and Solomon, 1992; Moody et al., 1990); 2,4-D and 2,4,5-T penetrate the skin relatively rapidly and could potentially take TCDD with them. Dermal absorption of PCBs, which are chemically similar to TCDD, has been found to exceed bioaccessibilty via either the ingestion or respiratory route in humans (Ertl and Butte, 2012; Lees et al., 1987).

Dermal Absorption of Vapor

For compounds that are sparingly soluble in air, pulmonary absorption may or may not exceed absorption through the skin. Weschler and Nazaroff (2012) have presented methods for estimating the contribution of vapor absorption via skin to total exposure to vapor phase chemicals. Compounds for which absorption of vapor via skin might plausibly exceed absorption via inhalation include both TCDD and the n-butyl esters of the herbicides 2,4-D and 2,4,5-T.

PLAUSIBILITY OF AO-RELATED IMPAIRMENT OF HEALTH AMONG AF RESERVE PERSONNEL

With increasing awareness of the toxic potential of various agents in the environment, substantial resources have been expended to develop methods to evaluate health risks in a rational and consistent fashion. Over the past several decades a general framework has been accepted for developing quantitative guidelines (like those discussed in Chapter 2) for use in assessing exposures to toxic substances, but objectives vary and, by its very nature, risk assessment is fraught with uncertainties. Specific efforts in risk assessment invariably come to

points of uncertainty where assumptions must be made and numerical inputs must be selected in order to move forward; however, opinions may differ on exactly what the appropriate decisions are in a given circumstance.

Given a range of options for addressing what could be concluded about health risks from the sparse set of results from exposure sampling on some of the C-123 aircraft that had been used in ORH and subsequently by AF Reservists, the committee undertook a number of approaches in an attempt to be inclusive and thorough. In the end, the committee reached the following consensus about the central question in this task: It is plausible that working on these ORH C-123s could have contributed to adverse health effects for some AF Reservists.

Historic reconstruction of occupational exposures relies on combining the two components of exposure, exposure concentration and contact time. Knowledge of the exposure levels during the time period of interest are combined with the job histories of the workers that delineate the time spent in the different activities (job classifications) that bring workers into contact with those exposure concentrations. Depending on the quality of the data available, the results can vary from a quantitative exposure evaluation usable for a quantitative risk assessment to a qualitative evaluation of whether a problematic exposure may have occurred. The data available to this committee fall into the latter category.

There were only limited numbers of measurements of TCDD and the herbicides in AO on the surfaces and in the air of aircraft flown by the AF Reservists (the exposure concentration), and those measurements were made years to decades after the exposures occurred. Although the paucity of measurements and delay in obtaining them increase the uncertainty about exposures, it is clear that decay and loss processes would result in overall lower levels at the time of measurement than had been the case when the AF Reservists actually were exposed. The committee, in the absence of knowledge of the losses of TCDD and herbicides during the time period of operation and storage of the aircraft, did not have a basis for predicting the decay in concentration and adjusting the measurements accordingly.

A second constraint on the committee’s deliberations was an incomplete knowledge of which aircraft were employed by the AF Reservists, the specific jobs and activities they were involved in, and the duration of time spent in the aircraft during their service. These factors would have been needed to estimate the contact rate, frequency, and time. Thus, the committee based its conclusions on a qualitative evaluation of the AF Reservists’ potential exposures, although quantitative estimates based on various models and assumptions were explored. The committee considers the AF Reservists to have been exposed to TCDD and herbicide through multiple routes when on aircraft that had previously been used in ORH.

A further consideration in interpreting the evidence is that background exposures were higher during the period that AF Reservists worked on the ORH

C-123s. In reaffirming its oral tolerable daily intake (TDI) of 2 pg/kg-d for TCDD and other dioxin-like chemicals, the Environment Agency of the United Kingdom (UKEA, 2009) noted that the average adult is estimated to consume about 49 pg TEQ from food and drinking water and to inhale approximately 0.2 pg TEQ on a daily basis, corresponding to a mean daily intake (MDI) of 0.7 pg/kg-d for a 70-kg adult. Twenty years earlier in 1988, the average person consumed on the order of 1.2 pg/kg-d TEQ (WHO estimate in NRC, 1988). Any AF Reservists ever assigned work on one of the ORH C-123s during their years of service most certainly received at least some increment in exposure to these substances. Superimposed upon the yet higher background levels of 1972–1982, such increments in exposure would have posed a more substantial threat to health than they would if experienced today. As discussed in Chapter 2, very small increments in exposure are not necessarily innocuous. They can increase the risk of adverse health effects either through a linear relationship or by crossing a threshold to a level at which adverse effects are plausible.

As another way of addressing its charge of putting in context the levels of exposure plausibly experienced by the AF Reservists, the committee attempted to compare them to existing guidelines for TCDD exposure within enclosed settings. When making such comparisons, the committee considered differences in the activities that influence exposures of the population for which the guidelines were developed, along with differences in duration of exposure. Such modulating activities include contact rates with surfaces, breathing rates, hand-to-mouth frequency, access to hygienic facilities before eating, surface area of the skin exposed, etc. Greater activity and movement of the AF Reservists would have resulted in contact with multiple surfaces since Reservists frequently sat on the floor of the aircraft with limited or no access to lavatory facilities prior to eating. This would suggest that they would have higher exposure from inadvertent ingestion and dermal pathways than office workers. On the other hand, the number of hours and years of work that the AF Reservists were in the aircraft were less that that used for establishing the guidelines for the office workers. Based on these varying conditions, it is the committee’s judgment that it is plausible, in some cases, the AF Reservists’ exposures exceeded TCDD guidelines for workers in enclosed settings.

All projected exposures will necessarily be uncertain due to limitations in the data available. The C-123 crews were potentially exposed to TCDD from air by inhalation and dermal absorption of vapors. Because of the methods used to collect the air samples, the committee did not consider the resulting measurements useful. Surface contamination with an SVOC can lead to exposure by inhalation of the chemical released into air, dermal absorption resulting from contacting a contaminated surface, and ingestion arising from hand-to-mouth transfer. There were 24 usable interior wipe samples of TCDD from ORH C-123s (see Table 3-3). Sixteen of them were collected in 2009, 27 years after the C-123s had been retired. The remaining eight were collected in 1994 and

1995, 12 years or more after retirement. These samples came from only three C-123s, and there is only a little anecdotal information on how representative these three C-123s are of all the ORH planes used back in the United States. There is no record of where in the interior of a C-123 some of the samples were collected. The three highest wipe samples (200, 250, and 1,400 ng/m² collected in 1994) seem rather inconsistent with five samples of the same plane conducted in 1995 for which the largest concentration measured was 30 ng/m², and the remaining four were non-detects (< 20 ng/m²). Moreover, these samples were from “Patches,” a plane selected for sampling because of chemical odors (due to the herbicides themselves or, perhaps more likely, the insecticides that had also been sprayed by this aircraft), so consequently these samples may not be representative of the ORH C-123s in general. The values of samples gathered from the other two C-123s in 2009, however, were quite closely clustered, perhaps an indication of redistribution toward homogeneity on their interior surfaces as they sat sealed on the desert. The only two TCDD air samples collected were collected in 2009 using a screening sampling method (USAF, 2009b), apparently under conditions of artificial mechanical ventilation (Nieman, 2014), and they both were non-detects. It is clear that any estimate of TCDD exposure to C-123 crews based on these meager data will be very uncertain. Nevertheless, the committee worked with these data in an effort to fulfill its charge by gaining at least some perspective on potential exposures.

To assess whether industrial sites require clean-up, it has become common practice to do a “worst-case scenario” evaluation as an initial screening procedure. Assumptions are made in selecting variables for estimation models to be protective of all exposed individuals; that is, they are “conservative.” The committee considered this approach as a potential starting point. In light of its charge to provide the VA with a sense of the plausibility of an increase in health problems actually occurring among the AF Reservists who had worked with ORH C-123s, the committee thought it more constructive to limit the inputs used in exploring various exposure models to ranges more likely to reflect the residues and the work experience of the AF Reserve personnel.

Exposure assessors frequently perform computer simulations, or Monte Carlo modeling, in which the exposure model is run repeatedly with values sampled from the theoretical distributions of each input variable. The results of all the iterations produce a distribution of exposure estimates that describes central tendency and variability for the situation in question. Because of the sparse nature of the available sampling results and the committee’s instruction to perform a qualitative assessment of exposure, such an intricate approach was not deemed necessary or scientifically credible.

The limitations of the available sampling data (as described previously) make them inadequate for deriving definitive estimates of exposure, but the committee did explore different approaches to quantitative assessment to gain a sense of magnitude and variability for answering its second charge. The committee does

not, however, endorse any particular estimates generated in the course of its quantitative explorations, so specific estimation results are not presented in this report.

The committee had defined the sampling results that would be best for any numerical considerations to be interior TCDD wipe samples. Unfortunately, this set of measurements was limited to eight samples from “Patches” (three in 1994 and five in 1995), and seven and nine samples collected in 2009 from two ORH C-123s that had been stored in the desert for many years (see Table 3-3).

The set of guidelines adopted as being most appropriate for comparison with the sampling results were those for surface loadings on indoor surfaces, which ranged from 1 to 25 TEQ ng/m² (see Table 2-1). Derived between 1983 and 2009, they seem to have tended toward greater stringency as data and understanding of the toxicity of dioxin-like chemicals increased. They include the 3.5 ng/m² screening level derived by CHPPM and the 22 ng/m² guideline found in the 2009 AF report (which should have been 1.1 ng/m², if dermal exposure had been taken into account as claimed; see Table 4-1). The committee considered this range of surface standards as roughly defining zone in which observation of sampling measurements suggests transition into a level of exposures plausibly associated with consequential increases the risk of health problems.

It was the committee’s understanding that the VA’s interest is in determining whether the AF Reservists had plausibly been at risk of experiencing exposures hazardous to their heath. With this objective, strict adherence to the public health practice, or “precautionary principle” that adopts very protective assumptions to ensure that no health threat to a population might be overlooked seemed inappropriate to the committee. Therefore, in its efforts to establish the plausible magnitude of the AF Reservists’ exposures, the committee considered values reflective of their actual work experience, rather than the extreme or “worst-case” values that are often used in risk assessments. When putting its perceptions of the available data in context, however, the committee accepted international screening guidelines, generally derived in accord with the precautionary principle, as defining a range of values at which taking further action in the interest of health would be merited.

Most of the guidelines that the committee used to put the possible exposures of C-123 personnel in context were developed for hypothetical long-term office workers (an application of the precautionary principle), but information on the work profiles of the AF Reservists was far too unclear to permit adjustment of the guidelines to match their work experience. Although the cumulative time the AF Reservists spent in ORH C-123s was less than 30- to 40-year working lifetimes, the nature of the C-123 personnel’s work activities may have increased the exposure experienced per unit of time, thereby reducing the “protectiveness” of the guideline. Also, several of the guidelines expressed in terms of surface TCDD levels do not account for the extent of dermal absorption. While such guidelines are generally derived with the intention of being protective, if the guidelines do not factor in all routes of exposure or activities of the workers,

then their tendency toward protectiveness from “worst-case” estimation of risk would be diminished.

The extent to which the levels of TCDD in the aircraft would degrade or be depleted over time was also uncertain. Several committee members thought that the levels of sampled residues would reflect only a small fraction of the concentrations present during the decade of use after the ORH C-123s returned from Vietnam. There was no consensus on the value for the fraction degraded or depleted over time, but all agreed that the contamination at the time of sampling would have been less than it had been at the time of exposure at least 10 years earlier. Therefore, the long delay between the time of exposure and the time of sampling would at worst contribute to underestimation of the AF Reservists’ exposure.

In other respects, comparisons of exposure estimates for the AF Reservists to established guidelines are associated with a great deal of uncertainty, but there is no reason to anticipate an overall trend of systematic over- or underestimation of health risks. For example, it was not at all clear that the three sampled planes and the 24 interior samples gathered were representative of the ORH C-123s used by the AF Reservists or the surfaces with which they had contact. Although this is a substantial source of uncertainty, there is no evidence suggesting their selection was biased toward over- or underestimation.

These arguments recognize considerable uncertainty, but they do provide support for concluding that the reported TCDD surface levels are very unlikely to be systematic overestimates in comparison to the existing surface loading guidelines. Table 4-4 is a compilation of sources of uncertainty demonstrating that there are a substantial number that might be expected to tend toward underestimation of the exposures and associated health risks experienced by the AF Reservists, which together may go far toward neutralizing the protective bias normally built into guidelines. Screening guidelines are intended to provide protection for workers with a vast range of far less extreme values than those of the hypothetical office workers used in deriving the guideline.

The committee became convinced that simply comparing the unadjusted surface measurements from the ORH C-123 to the existing guidelines for surface loading provides a valid qualitative means based on international regulatory standards for envisioning the degree of health risk associated with these results (see Figure 4-1).

Given the variety of approaches pursued in its efforts to interpret the available data and the intrinsic weaknesses of those data, the committee was unable to determine which, if any, of the various models and exposure scenarios investigated should be regarded as the most reliable representation of the experiences of the AF Reservists. The committee observed, however, that under at least some reasonable scenario, all the models that were considered generated exposure estimates for the C-123 personnel that were larger than what screening guidelines deemed to be “acceptably” safe. Several factors contributing to uncertainty

TABLE 4-4 Likely Directionality of Various Sources of Uncertainty Involved in Assessing Exposure of AF Reservists Who Used ORH C-123s on the Basis of Available Sampling Results

NOTES: AF, US Air Force; AO, Agent Orange; ORH, Operation Ranch Hand; TCDD, 2,3,7,8-tetra-chlorodibenzo-p-dioxin; TEQ, toxicity equivalency quotient.

discussed in this report (such as the long delay between when the activities leading to possible exposure occurred and sampling, the failure to adequately account for the extent of dermal absorption, considering only TCDD measurements in comparison to the screening levels developed in terms of TEQs for all dioxin-like chemicals contributing to exposure overall, etc.) would be expected to bias exposure estimates toward underestimation and hence an understatement of projected risk for the AF Reservists. Such tendencies toward underestimation would countervail to a certain extent against the built-in “worst-case” nature of

FIGURE 4-1 Existing guidelines for TCDD Surface concentrations in relation to results of interior wipe samples. NOTE: TCDD surface concentrations obtained from the total of 24 wipe samples from the interiors of post-Vietnam C-123 aircraft by year. The horizontal band represents the 1–25 ng/m² range of existing surface guidelines for TCDD. Clear points represent non-detect samples plotted at their detection limit. The concentration of TCDD in four samples gathered from “Patches” in 1994 were below their limits of detection, which in turn were not sensitive enough to evaluate whether or not the samples’ concentrations were less than 1 ng/m² (below the range of concern).

the available guidelines considered. When even these quite probably understated values fall in the region delineated by the screening levels for interior surfaces derived by several expert regulatory groups, the committee’s reasons for attributing plausibility to the occurrence of nontrivial increases in the risk of adverse health outcomes have a firm basis.