National Center for Environmental Health: Chemical Weapons Elimination | Reports - VX Report - Attachment #2

Attachment #2

Report for Congress: Review of the U.S. Army Proposal for Off-Site Treatment and Disposal of Caustic VX Hydrolysate From the Newport Chemical Agent Disposal Facility

Review of the Toxicology and Health Hazard Considerations for Safe Management of Newport (Indiana) Caustic VX Hydrolysate

By

Agency for Toxic Substances and Disease Registry in collaboration with the Centers for Disease Control and Prevention Atlanta, Georgia
November 3, 2004

SUMMARY
The Centers for Disease Control and Prevention (CDC) requested that the Agency for Toxic Substances and Disease Registry (ATSDR) assess DuPont Report 14523, Toxicology Assessment of Health Hazard Considerations for Safe Management of Newport Caustic Hydrolysate, dated March 3, 2004, and its supporting documentation as part of a larger evaluation of the proposed transportation and disposal of caustic VX hydrolysate (CVXH), waste material produced by the reaction of the nerve agent VX with sodium hydroxide. In response to this request, ATSDR conducted the following assessment in collaboration with CDC. Please note that in this report, the more technically accurate term CVXH generally is used in place of Newport caustic hydrolysate or NCH.

It should be noted that the CVXH toxicity testing discussed in ATSDR’s assessment was conducted on 33 weight percent loading material. The current treatment plan by the Army is to process at an 8 weight percent loading. Because of the lower loading in the current plan, the toxicity testing that was conducted at the higher loading percentages should be considered “worst case” in terms of the potential toxicity of the CVXH.

The major findings and conclusions of the ATSDR assessment are as follows:

The untreated CVXH is highly corrosive. The major human exposure pathway for the material is dermal contact, which could result in severe, possibly irreversible, burns to the skin or eyes. Overall, the health risk from exposure resulting from an accidental spill appears comparable with that expected for any highly corrosive material with high pH.
Although the individual toxicity studies are limited in scope and applicability, the studies—considered in their totality—do not preclude the handling and transportation of untreated CVXH if appropriate engineering and administrative controls and personal protective equipment are used.
The supporting studies do not provide adequate data on the nature of the toxicity of ethyl methylphosphonic acid (EMPA) and methyl phosphonic acid (MPA) (constituents of CVXH). EMPA and MPA are highly water soluble; therefore, if an accident occurs during handling and transportation, groundwater or surface water contamination and subsequent human ingestion are unlikely, but possible, outcomes. Limited data are available to determine the risks from exposure to nonlethal ingestion of EMPA and MPA. However, oral lethality studies indicate the two substances have a Hodge and Sterner toxicity rating of 4 (slightly toxic).
While the effects in animals following administration of CVXH are not likely due to residual VX or EA 2192 (a degradation product of VX with nerve agent properties potentially present in CVXH), the data in one of the cited studies are not conclusive due to lack of appropriate controls.
Clearance criteria for VX and EA 2192 are suitable for the risk management approaches presented.

INTRODUCTION
ATSDR was provided copies of the toxicity studies examined by DuPont, as well as other studies commissioned by the Army or its contractors. The studies examined major components of the CVXH. Because neither the studies cited by DuPont nor the other toxicity studies provided were peer-reviewed, ATSDR first had the studies peer-reviewed. An ATSDR contractor identified nongovernmental independent professionals for the peer review. After receiving the peer-reviewer comments, ATSDR reviewed DuPont’s report and referenced studies to generate the following comments.

DuPont stated that its assessment of potential health risks of CVXH was conducted to support decisions related to the transportation and treatment of CVXH at the DuPont Secure Environmental Treatment (SET) facility. DuPont and the Army proposed that the CVXH be transported from the Newport Chemical Agent Disposal Facility in Newport, Indiana, to the DuPont SET Facility in Deepwater, New Jersey, for final treatment and discharged into the Delaware River.

The DuPont assessment states that the composition of the CVXH is 80% water with minor amounts of sodium hydroxide (Chemical Abstract Services [CAS]# 1310-73-2), diisopropylamino ethylthiolate (thiolamine, CAS# 5842-07-9), ethyl methylphosphonic acid (EMPA, CAS# 1832-53-7), and methylphosphonic acid (MPA, CAS# 993-13-5). Approximately 1% is composed of “other compounds,” including ethanol (CAS# 64-17-5), diisopropylamino ethyl disulfide (CAS# 65332-44-7), and diisopropylamine (CAS# 108-18-9).

ANALYSIS AND DISCUSSION
DuPont’s assessment concludes CVXH is not a Department of Transportation (DOT) poison or toxic material and has no nerve agent characteristics. DuPont indicates that CVXH is corrosive and capable of damaging the eye and skin after contact exposure. Gastrointestinal injury can result from ingestion. In support of these conclusions, the DuPont assessment of CVXH cited the following studies:

Finlay, C. Ethyl Methylphosphonate: Oral Approximate Lethal Dose (ALD) in Rats. Haskell Laboratories, February 26, 2004.
Finlay, C. Methylphosphonic Acid: Oral Approximate Lethal Dose (ALD) in Rats. Haskell Laboratories, February 26, 2004.
Manthei J, Way R, Gaviola B, Burnett D, Bona D, Durst H, Thompson S. Toxicological Evaluation of VX Decontamination Wastestreams According to DOT Test Procedures, February 1999.
Kemper, R. Ethyl Methylphosphonate: Computational Toxicology Analysis. Haskell Laboratories, March 1, 2004.
Kemper, R. Methylphosphonic Acid: Computational Toxicology Analysis. Haskell Laboratories, March 1, 2004.

The Army subsequently provided additional studies:

Manthei J, Way R, Gaviola B, Bona D, Burnett D. Alternative Technology Program: Intravenous Toxicological Evaluation of Four VX Wastestreams in Mice.” U.S. Army ERDEC, ECBC-TR-173, August 2001.
Janus, E.R. Analysis of EA2192 Monitoring and Sampling Issues at Newport Chemical Agent Disposal Facility. Environmental Health Risk Assessment Program. U.S. Army Center for Health Promotion and Preventive Medicine, November 2001.
McDonald, J., and Campen M., Revised Final Report, Acute Inhalation Toxicity Testing of 2-(diisopropylamino)Ethyl Mercaptan. Lovelace Respiratory Research Institute, April 2, 2004.

Analysis of the Finlay (2004) Studies
The studies conducted by Finlay (2004) determined a lethal dose of 2300 milligram per kilogram (mg/kg) and 3400 mg/kg for MPA and EMPA, respectively. The chemicals were administered as a single oral (intragastric intubation) dose to one rat per dose level; body weights and clinical signs of toxicity were observed for 14 days postexposure. These studies provide useful information about lethality. The Finlay (2004) studies were “approximate lethal dose” studies that use fewer animals but have been shown to closely predict the results of classical lethal dose in 50% of animal population (LD₅₀) studies. However, the studies presented no information to assess the nature of the acute toxicity— that is, this study generated no information about the type of toxic effects (i.e., organ system affected). Therefore, DuPont’s statement in its toxicology assessment—“…MPA and EMPA have relatively low acute oral toxicity…”—provides limited perspective on the toxicity of these components of CVXH. In reality, the Findlay studies were lethality studies, not acute exposure studies; the “acutely toxic effects” observed at 2300 mg/kg MPA and 3400 mg/kg EMPA were death. With respect to handling and transportation of CVXH, however, the likelihood of ingestion of CVXH (including MPA and EMPA) is low. The Hodge and Sterner toxicity rating for MPA and EMPA is 4 (slightly toxic). Therefore, although cited studies were limited in scope, when considered in conjunction with the toxicity rating and potential exposure scenarios, MPA and EMPA components do not introduce excess risk in handling and transportation activities.

Analysis of the Manthei et al. (1999) Study
The Manthei et al. (1999) study, performed by the Army, provided toxicity data to establish shipping and packaging criteria (for CVXH) according to 49 Code of Federal Regulations (CFR). In this study, severe dermal injuries occurred when the CVXH homogenate was applied to rabbit skin at 1000 mg/kg; and gastrointestinal injury and death (two of 12 rats) occurred in rats dosed orally at 500 mg/kg. The study concluded that this compound was less than a Level III toxic according to 49 CFR. If, as is our understanding, the Level III requirement is for an LD₅₀ of <500 mg/kg, then the CVXH would appear to meet this requirement. For caustic compounds, 40 CFR outlines corrosivity characterization needs. Under some circumstances, DOT recommends further toxicity tests for more complete characterization (49 CFR 173.137 and 1992 Organization for Economic Cooperation and Development Guideline No. 404).

Additionally, toxicity testing of the top organic layer of test material killed 12 of 12 dermally treated rabbits (500 mg/kg) and 12 of 12 orally treated rats (1000 mg/kg). The animals died from agent (VX)-associated effects. Subsequent testing revealed that the organic layer contained 2000 ppm VX. The Manthei et al. (1999) abstract states that a follow-up study would be conducted, but as of this writing, no follow-up study has been provided. However, it is clear that the samples were contaminated with VX as a result of laboratory error, rendering the results of this study questionable. Furthermore, this high-level VX contamination was not consistent with other work by the same laboratory. In summary, the results of this particular part of the Manthei et al. (1999) study must be discounted as not representative of the toxicity of CVXH.

DuPont’s assessment states that the CVXH contains no VX (later clarified to “no detectable VX) with a MDL (method detection limit) of twenty parts per billion (ppb) or less” (DuPont Position on the Question of VX in Hydrolysate, July 24, 2004). The ATSDR review assumes this to be the case because the CVXH will be analyzed for VX and must meet the 20 ppb criteria before shipment.

Analysis of the Manthei (2001) Study
In another study by Manthei (2001), adult, male ICR mice were dosed intravenously with CVXH. LD₅₀ values were calculated to be 349.5 mg/kg, 39.0 mg/kg, and 279.3 mg/kg for the bottom, top, and homogenate samples, respectively. Chemical analysis indicated no VX at or above the detection limit of 20 ppb in the bottom layer or the homogenate. The top layer was not analyzed for VX. Effects observed included convulsions, exophthalmus, straub tail, collapse, and prostration. Although the toxic signs in the mice probably resulted from by-product salts, the investigators did not use controls needed to determine whether the effects were due strictly to the by-product salts and not to residual VX or EA 2192. The conclusion was based on the absence of observed tremors and salivation. The use of controls or acetylcholinesterase activity would have provided more definitive results. ATSDR concludes that the upper organic layer material on CVXH is more toxic than the aqueous lower layer, and the effects in the animals probably resulted from by-product salts and high pH (caustic nature).

Analysis of the McDonald and Campen (2004) Study
The McDonald and Campen (2004) study was designed as an acute toxicity screen for diisopropylamino ethylthiolate (thiolamine), which typically is used as a basis for establishing a dose regimen in subchronic and other studies. Decreased body weight gain and nasal porphyrin accumulation was observed in the high dose groups (316 mg/m3). Because no sham or age-matched control animals were used in this study, it is not possible to draw definitive conclusions about these effects. McDonald and Campen (2004) noted the pathology analysis was a crude indicator of a lack of toxicity of this component of CVXH. The usefulness of this study in assessing inhalation toxicity of thiolamine for use in the CVXH assessment is limited.

Analysis of the Kemper (2004) Studies
As stated in the DuPont assessment, the computational toxicology analyses of MPA and EMPA (Kemper 2004) did not provide useful predictions of the acute toxicity of these chemicals. The positive predictions of toxicity for developmental effects for both MPA and EMPA (by the Toxicity Prediction by Computer-Assisted Technology [TOPKAT] model), and bacterial mutagenicity for EMPA (by the Deductive Estimation of Risk from Existing Knowledge [DEREK] model), and the negative prediction for skin sensitization (by TOPKAT) are not reliable because the query structures are poorly represented in the TOPKAT or DEREK models’ datasets. The report also provides a nonuseful large predictive oral LD₅₀ range (which appears to be the predicted 95% confidence limits), instead of the single predicted LD₅₀ value it should have provided. Thus, ATSDR agrees with DuPont that the Structure Activity Relationships analyses performed did not provide useful predictions of the toxicity of these chemicals.

The results of the DEREK analysis (by Kemper 2004) suggested that EMPA could cause mutagenic effects in bacteria. The DuPont document states that mutagenicity is unlikely on the basis of negative test results for isopropyl methylphosphonate (IMPA), a close structural analogue of EMPA. However, because of its chemical structure, IMPA would not be expected to react similarly in the body as EMPA. Thus, whether IMPA should be used as a surrogate to make conclusions about the mutagenicity of EMPA is not clear.

Analysis of the Janus (2001) Study
The purpose of the Janus (2001) paper was to calculate a Performance Indicator (PI) value for EA 2192. The document states that PIs are “developed to monitor and evaluate discrete subsystem requirements that must be demonstrated to achieve the design and technical performance goals of the Newport Pilot Plant.” The document briefly discusses the relative potency of VX and EA 2192, stating that EA 2192 toxicity is generally within the same order of magnitude as VX, therefore, it is appropriate to use the interim VX reference dose (RfD) to calculate the PI for EA 2192. The document uses an algorithm to calculate the PI that is based on U.S. Environmental Protection Agency (EPA) Region IX’s Preliminary Remediation Goal (PRG) approach. In this algorithm, the interim oral RfD for VX (of 6E-07 mg/kg/day) is used to develop a dermal PI value of 1.128 ppm for EA 2192. The PI methodology appears appropriate; however, the EPA PRG User’s Guide/Technical Background Document states, “For many chemicals, a scientifically defensible data base does not exist for making an adjustment to the oral slope factor/RfD to estimate a dermal toxicity value.” Whether the permeability coefficient, as used in the PI algorithm, is appropriate is unclear because the caustic nature of the CVXH will compromise the ability of the stratum corneum to serve as a protective barrier, thereby allowing more direct entry. Nonetheless, Manthei et al. (1999) did not observe VX or EA 2192 effects after dermal application of caustic VX hydrolysate to rabbits (1000 mg/kg for 24 hours). Therefore, ATSDR believes that the PI appears to be suitable for worker protection when appropriate personal protective equipment is used to handle CVXH.

FINDINGS

Although the individual toxicity studies were limited in scope and applicability, the studies considered in their totality do not preclude the handling and transportation of CVXH, assuming appropriate engineering, administrative, and personal protection policies are in place.
Although the studies on MPA and EMPA do not provide data on the nature of the toxicity, the oral lethality studies indicate that the two compounds have a Hodge and Sterner toxicity rating of 4 (slightly toxic). Furthermore, oral ingestion of MPA and EMPA during handling and transportation of CVXH is unlikely.
MPA and EMPA are highly water-soluble; therefore, if an accident occurs during handling and transportation, groundwater or surface water contamination and subsequent human ingestion is an unlikely, but possible outcome. Data are insufficient to determine the risks from exposure to nonlethal ingestion of MPA and EMPA.
Information about thiolamine is limited. Mercaptans in general are well-known noxious volatile odorants and skin irritants.
Although the effects noted in the intravenous studies (Manthei et al. 2001) probably do not result from residual VX or EA 2192 in the CVXH, the data are not conclusive because of a lack of appropriate controls to distinguish between agent effects and by-product salts or high pH (caustic) at the 33% VX loading. In another study (Manthei et al. 1999), lack of nerve agent effects were observed after CVXH exposure in dermally exposed rabbits and orally exposed rats.
The PI of 1 ppm for EA 2192 appears to be adequate given the Manthei et al. (1999) data, which did not note any VX or EA 2192 effects in rabbits after dermal exposure to CVXH. Although no chemical analysis for EA 2192 was conducted, this CVXH fraction obtained from a 33% VX loading is assumed to have contained at least representative quantities of EA 2192. For the 8 weight percent loading CVXH planned for disposal, the concentration of EA 2192 probably would be lower than that found in these experiments.
As the DuPont assessment indicates, CVXH is highly corrosive. This is supported by the Manthei et al. (1999) study and the chemical property information. The major human exposure pathway is dermal contact, which will result in severe, possibly irreversible damage. Eye injury is also possible, and inhalation of aerosolized CVXH potentially could damage the respiratory tract.

CONCLUSION
ATSDR believes that, in the event of an exposure after an acute release, the greatest concern would be the caustic nature of the CVXH, which potentially could cause severe burns upon contact. Overall, the risk from an accidental spill appears to be comparable with what would be expected for any highly corrosive material with a high pH.

Previous | Table of Contents | Next