m. \ UNITED STATES ENVIRONMENTAL PROTECTION AGENCY a •--•' WASHINGTON, D.C, 20460 March 9, 1988 SAB-EHC-88-013 Hbn, tee M. Ihcmas TH Administrator U.S. Environmental Protection Agency 401 M Street SW Washington, D.C, 20460 Dear Mr. Thomas; On August 13-14 1987 the Halogenated Grganies Subcommittee of the Science Advisory Board's Environmental Health Oonwdttee met in Washington, D.C. to review two documents prepared by EPA's Office of Research and Development that assess health effects associated with dichlorcroethane (methylene chloride). These documents included; o a June 1987 Craft Technical Analysis of New Methods and Data Regarding Dichlorcmethane Hazard Assessments, and o a July 1987 Draft Addendum to the Health Assessment Document for DLChlorcme thane: Pharmaeokinetics, Mechanism of Action and Epi- demiology. The Subectanittee's evaluation of these documents is presented in two parts: 1) a discussion of scientific issues related to phannaco- kinetics and metabolism, and 2) review of specific issues pertinent to the addendum. Ihe Subcommittee focused less attention on the former document because much of its scientific content overlapped with the Addendum. The Subcownittee concludes that the Addendum was one of the best docu- ments it has reviewed in terms of its clarity, coverage of the data and analysis of scientific issues. This document clearly demonstrates the potential utility of phamacokinetic data in risk assessment. EPA should continue to use this approach in future risk assessments, whenever scien- tifically possible. For reasons discussed in the attached report, the Subcommittee concludes that the level of uncertainty is greater and the hazard posed by dichloro- methane may be less than that expressed ty the categories of EPA's cancer risk assessment guidelines. ------- —2— The Subcommittee appreciates the opportunity to conduct this scientific review. In igbalf of the Subcommittee we request that the Agency formally respond to the scientific advice provided in its attached report, Sincerely, Por'tbn Nelson, Chair Executive Ccnwiittee Richard A. Griesemer, Environment* u. air ,ee y Qrganics Subcoratrattee ------- Halogenated Grganics Subcommittee Review of the June 198? draft Technical Analysis of New Methods and EBta Regarding Dichloranethane Hazard Assessments: and July 1987 Draft Addendum to the Health Assessment Document for ttchloromethane: pharntaco- kinetics, Mechanism of Action and Epidemiology Pharntacojci'netics and Metabolism The application of a physiologically based pharmacokineties (PBPK) model presented in the document is generally well conceptualised and organized. This application represents a novel approach that can sharpen EPA's ability to refine human risk estimates in the future. Ihe Subcommittee commends SPA for incorporating such information into the weight of evidence deter- mination of the carcinogenic potential of dichlorcroethane. .Adoption of the Reits-Anderson model, with certain modifications, is a positive step for- ward for the Agency's risk assessment process. Ihe critical analysis of the constraints Of the model are thoroughly discussed and scientifically balanced. EPA, for example, is justified in adjusting the estimates of Reitz-Andejrson for breathing rates traditionally used in EPA models. Ihe rationale for using surface area factor adjustment, and contrary arguments, are clearly described, The PBPK model has appeared in the published peer reviewed literature. The novel applications of the latest data concerning the model that were presented at the meeting are new and received enthusiastic support from the. Subcommittee, which also recommends publication of this approach. Ihe Sub- committee recognizes that validation will be required for this and other compounds before using this approach generally for human risk calculations. One possible problem is that the metabolic conversion of dichloromethane by different animal species by either the cytochrorae P-450 oxidase system (mixed function oxidase or MFO) or the glutathione-S-transferase system (GST) was not supported by data indicating that measurements in each species were conducted at conditions optimal for pH» ionic strength or temperature for that substrate in that tissue for each specific species, unless such conditions are utilized, interspecies quantitative data may be meaningless, and the basis for the use of this approach in human risk estimation could be flawed. Comparative _in vitro studies with rat, mouse, hamster and human cytosol showed that the dichlororoethane-GST conjugation rates in humans were at least 50 times lower than in mice. The Subcommittee points out that: 1) mice have the highest activity and liver tumor induction that correlates with GSH-r&etabolite production above saturation of the MFO system; 2) hamsters have much lower activity and no liver tumors? and 3) humans have even lower activity indicating very low, if any, liver tumor inducing potential for dichloromethane. Ihere was a good correlation between the relative rates of dichloromethane-GSH conjugation and susceptibility of liver tumors. The conclusion that, at low exposure levels, the carcinogenic hazard to humans ------- frcm dichlorcitiethane appears very low needs to be clearly stated in the document. Ihe document could be simplified by eliminating the Chapter 7 discussion of a "rationale" for surface area scaling and replacing it with the statement on page 107 that, "Ihe fact that there is no clear basis for choosing the use of surface area correction or not..*is a weakness of the current state-of-the- art of quantitative risk assessment." jDLscusslon of Specific Scientific Issues Related to the Addendum 1. In considering an overall weight of evidence approach to risk assessment, other factors, such as the nature of the animal tumor response, mechanistic data (such as binding of the chemical to ENA), genotoxic activ- ity and epidemiological data should also be discussed. 2. In evaluating the tumor data* the Subcommittee urges caution in ex- trapolating the existing animal bioassay to humans. Although diehlorcraethane induced both lung and liver tumors in the mouse models, these observations occurred only at high doses which likely influenced the compound's overall metabolism. Other bioassays in other species, or at lower doses in mice, induced negative results. Ihe fact that the Reitz-Anderson model is able to predict these responses suggests that an interspecies correction factor based on surface area may not be necessary for extrapolating the tumor data to humans* This is particularly true when hamster and rat data (GSH transferase) are considered using the PBPK analysis. Ihe observation of benign mammary tumors and salivary gland tumors in rats should not be used as strong evidence for human carcinogenic potential given the uncertain significance of these lesions. The benign mammary tumors have very low potential for predicting malignancy even in the rat, and salivary gland tumors were reported in only one of the studies. 3. EPA should discuss the findings of several investigators (Shumann et. al., Dow Chemical; Green et. al., ICI, U.K.) that indicate that dichlOTO- methane or its metabolites do not exhibit any potential to alkylate liver or lung ENA following j.n vivo exposure. Such findings raise the clear possibil- ity that dichloronethane may have produced its carcinogenic responses in mice by non-genotoxic mechanisms, and may include an important contribution of cytotoxieity in the overall tumorigenic process. Such data become par- ticularly relevant as carcinogenicity was observed only at extremely high exposures and was absent at lower, potentially noncytotoxic doses. 4. Critical uncertainties remain regarding the relationship between dose to target tissues and tumor incidence, since little information on the mechanism of action is available for dichloronethane. The Subcommittee accepts EPA's use of a surface area scaling factor for delivered dose as appropriate for calculating an upper bound estimate, but it views this usage as more conservative than the usual "default" assumption frcm the Agency cancer guidelines, scaling administered dose by surface area frcm animals to humans. Further research may indicate that, at least for some substances, scaling delivered dose on the basis of body weight is more appropriate than scaling by surface area. ------- -3- 5. the degree of nonlinearity in the dose reponse relationship for deliver- ed dose is an important source of uncertainty. As noted on page 110 of the Addendum, EPA uses the linearized multi-stage model to calculate an upper bound estimate. The true dose response curve may fall off more rapidly than a linear relationship at low doses. Biological information supporting a non- linear or threshold type of dose response relationship is potentially ijmpor- tant for risk management decision making because it becomes less likely that the default plausible upper bound linear estimate will be an accurate estimate of human risk, especially at low exposure levels in the ambient environment, 6. The Subcommittee was presented with a brief report on the current status of the Kodak epideraiologieal study of dichlorcmethane. A slight excess of pancreatic cancer deaths has been separately reported. However, the study is based only on death certificate data and has not included a histopatho- logic review of biopsies or surgical specimens from such patients. Ihe incidence of pancreatic cancers tended to cluster, and only with further surveillance of the population can a more definitive statement be made on human health risk. The clinical diagnosis of pancreatic cancer is difficult and may be easily confused with other abdominal malignancies. Thus, without pathologic confirm- ation, the Subcommittee cannot necessarily conclude that an excess of pancreat- ic cancer deaths has occurred. Kwever, neither can it be concluded that dichloromethane is safe for humans at the occupational exposure levels seen in the study. The Agency should determine the criteria of the Kodak epidendo- logical study necessary to substituting the animal derived risk estimate with a human derived risk estimate. Finally, the Subcommittee recommends the continua- tion of this important study. 7. Although there is an impressive weight of evidence implicating metab- olites of dichlorcmethane in tumors, the possibility should not be discounted that the actual tumor inducing agent is the parent ccmpound(s). In order to present a more balanced document, this possibility should be discussed at greater length, perhaps in Chapter 8. 8. Both the scaling factor and the shape of the dose response relationship are important areas for further work in order to aid development of risk assess- ment methods that incorporate available scientific data and judgement on bio- logical mechanisms. As better information is developed on pharmacokinetlcs, phamaccdynamies and mechanisms for chemical carcinogenesis, it should be possible to further reduce uncertainties in human risk estimates. 9. For all of the above reasons, therefore, the Subcommittee concludes that the level of uncertainty is greater and that the hazard for dichlorc- methane may be less than that expressed by the Agency's classification system in its cancer risk assessment guidelines. More detailed discussion of these and other issues by individual Subcom- mittee members has been forwarded to the office of Research and Development. ------- U.S. Environmental Protection Agency Science Advisory Board Environmental Health Ccnmittee Halogenated Qrganics Subooranittee Roster for August 13-14, 1987 Review of the Draft Assessment Documents for DLchloreraethane Dr. John Doull, Chairrnan, Professor of Pharmacology and Toxicology, Univer- sity of Kansas Medical Center, Kansas City, Kansas 66103 Dr, Seymour Abrahamson, Vice-chairman, Professor of Zoology and Genetics, Department of Zoology, University of Wisconsin* Madison, Wisconsin 53706 Suboonroittee Membersand Consultants Dr, Linda Birnbatjm, National Institute of Environmental Health Sciences, P.O. Box 12233, Research Triangle Park, North Carolina 27709 Dr. George T. Bryan, Department of Human Oncology, University of Wisconsin, K-4, Roan 528, 608 Clinical Science Center, 600 Highland Ave., Madison, Wisconsin 53792 Dr. Janes Bus, Pathology and Toxicology Research, Upjohn Company, Kalarnazoo, Michigan 49001 Dr, Robert Dedrick, Chief, Chemical Engineering Section, National Institutes Health, Building 13, Rocra 3W13, Bethesda, Maryland 20892 Dr. David Gaylor, National Center for Toxicological Research, Jefferson, Arkansas 72079 Or* Ronald D* Hood, Professor and Coordinator, Cell and Developmental Biology Section, Department of Biology, University of Alabama, and Principal Associate, R»D. Bood and Associates, Consulting lexicologists, P.O. 1927, University, Alabama, 35486 Dr. K. Roger Hornbrook, Department of Pharmacology, P.O. Box 26901, University of Oklahoma, Oklahoma City, Oklahoma 73190 Dr. Curtis Klaassen, Professor of Pharmacology and lexicology, University of Kansas Medical Center, 39th and Rainbow Blvd., Kansas City, Kansas 66103 Dr. Karl K. Rozman, Department of Pharmacology, Toxicology and Therapeutics, University of Kansas, Kansas City, Kansas 66103 Dr. Stephen Safe, Department of veterinary Medicine, Physiology and Pharmacology, College of Veterinary Medicine, lexas ASM University, College Station, Texas 77843-4 DC. Robert Squire, 1515 Labelle Ave., Ruxton, Maryland 21204 Dr. Thomas Starr, CUT, P.O. Box 12137, Research Triangle Park, North Carolina 27709 ------- ' Participating Members of the Environmental Health Ccnniittee DC. Richard A, Griesaner, Biology Division, Oak Ridge National Laboratory, Martin Marietta Energy Systems, Inc., P.O* Box Y, Oak Ridge, Tennessee 37831 Cr. D. Warner North, Principal, Decision Focus Inc., IDS Altos Office Center, Suite 200, 4984 El Canuno Real, IDS Altos, California 94022 Executive Secretary Cr. C* Richard Cothern, Executive Secretary, Environmental Health Committee, Science Advisory Board {A-10IP), U.S. Environmental Protection 401 M Street, SW, Washington, D.C, 20460 ------- |