Document rpq7RjVRRLKYdvnkXQoJyvgpE

CMiza CHEMICAL MANUFACTURERS ASSOCIATION July 5, 1994 James Cogliano Chairman, Carcinogen Review Assessment Verification Endeavor U.S. EPA (8602) 401 M Street, S.E. Washington, D.C. 20460 Dear Mr. Cogliano: It has come to my attention that EPA's Carcinogen Review Assessment Verification Endeavor currently is reassessing human health risk for vinyl chloride. The Chemical Manufacturers Association Vinyl Chloride Panel applauds EPA's efforts to review the latest health effects information to determine risks posed by specific chemicals. Dr. Richard Reitz, formerly of Dow Chemical, has developed a physiologically-based pharmacokinetic model for vinyl chloride risk assessment. Dr. Reitz recently presented his model to the Vinyl Chloride Panel and has agreed to present his model to EPA prior to its publication, if the Agency provides an opportunity for the presentation. Therefore, the Vinyl Chloride Panel requests a meeting with appropriate EPA and/or contractor staff to discuss the new vinyl chloride risk assessment model. Dr. Reitz's abstract of the model is enclosed with this letter. Also, the Panel conducted an epidemiologic study of vinyl chloride workers covering a period of time from 1942-1982. The final report on the study was submitted to the EPA Administrator in 1986. A copy of the report is enclosed with this letter just in case the report did not reach CRAVE or CAG personnel. The following articles related to the CMA-sponsored vinyl chloride epidemiology study also are enclosed: 1991 An Industry-Wide Epidemiologic Study of Vinyl Chloride Workers, 1942-1982; 1993 Letter to the Editor, Diagnostic Bias in Occupational Epidemiologic Studies (H. Shah); and, 1993 Response to Letter to the Editor, Diagnostic Bias in Occupational Epidemiologic Studies: An Example Based on the Vinyl Chloride Literature (0. Wong). 2501 M Street, NW, Washington, DC 20037 Telephone 202-887-1100 Fax 202-887-1237 BFG 02020 James Cogliano July 5, 1994 Page 2 I will call you in the next few weeks to discuss the possibility of meeting to examine scientific issues related to the vinyl chloride risk assessment. Meanwhile, if you have any questions or need additional information, please contact me at (202) 887-1192. Sincerely, STvjjJ-_____ Hasmukh C. Shah, Ph.D. Manager, Vinyl Chloride Panel Enclosures cc: William Farland, Director, Office of Health and Environmental Assessment, EPA Hugh McKinnon, M.D., Director, Human Health Assessment Group, EPA BFG 02021 I Estimating Human Risk from Exposure to VC Quantification with PB-PK Modeling Richard H. Reitz Michael L. Gargas McLaren/Hart, ChemRisk Division for CMA Vinyl Chloride Panel May 19,1994 i Collaborators | McLaren/Hart: R. H. Reitz M. L. Gargas IC1 Toxicology Lab (Zeneca) T. L. Green W. M. Provan U. S. E. P. A. (Res Tri Park) M. E. Andersen S/2(W* 2 VC History j Low Acute Toxicity Occup. Expos. Limits - 500 ppm Viola (1970,1971) Rats, Increased Tumor Incidence Maltoni (1974) Confirmed Viola's Results Identified Rare Liver Angiosarcoma Dose Response Flat > 1,000 ppm Creech & Johnson (1974) Found Same Cancer Type (Liver Angtosacroma) In Homans Human Tumor Registry (to Present) 14,000 Subjects, I VC Plants S20/94 3 Objectives / Opportunity | Develop A Process for Quantitatively Estimating Risk in Humans + Low, Non-Occiipatiooal Exposures - Superfund Sites - Fugitive Emission - Drinking Water Test the Utility of our Cancer Risk Assessment Procedures + Rkh Animal Data Set In Rats and Mice -t- Unique Opportunity to Compare Risk Assessment with Actual Results In Humans s/2om 4 BFG 02023 I Expectations for Pharmacokinetic Modeling Modeling Cannot Eliminate ALL Uncertainty from Risk Assessments Modeling Can Quantitatively Describe: Metabolic Saturation Changes In Dose Route Physiological Differences In Species PREMISE: Risk Assessments based on Estimates of "Delivered Dose*1 will be More Reliable than Risk Assessments based Only on Administered Dose &T20/M S Classical Pharmacokinetics: "Stripping the Curve"_________ Curve Stripping fEraonentlalsl C(t) - A, * e'ai 5120m d&O/dt > -ka+Doao dJU/4t - ta*DoM - KJ3-C1 + X21*C2 - ke*Cl 4*2/dt - +K21*C1 - K21+C2 dElia/dt - ~ka*Cl sawn 7 Advantages of PB-PK Models: Compound Specific Information Vapor Pressure Solubilities (Partitioning) In Tissues Species Specific Information Physiology Metabolism Route Specific Information Oral Route, 1st Pass Through Liver Allow Extrapolations Between Dose Routes Between High Dose / Low Dose Between Species sawn 8 BFG 02025 I A PB-PK Model for VC s/2<m Metabolites Bawd ob Ramsey 4t Aodena, 1464 Capabilities of PB-PK Models Will use examples from studies of Reitz et al., at Dow Chemical Co., with 1,1,1-trichloroethane (Methylchloroform, MC) Data used to Illustrate potential applications for VC PB-PK Model. worn ifl BFG 02026 BFG 02027 BFG 02028 BFG 02029 I VC Metabolic Rate Constants Gas Uptake Apparatus snom 5/20/!M BFG 02030 BFG 02031 I Validation of Rat Model (Watanabeetal., 1976) Siltm 21 Estimating Mouse Metabolic Rate Constants Small, Halogenated Hydrocarbons Metabolized by CyP450 2E1 In Vivo VMax's from Experiments Methylene Chloride (MeCI*) (Rats, Mice, Humans) Chloroform (CHCI3) (Rats, Mice) Calculate VMax / gram Liver Normalize to Rat In Vivo MeCI2 CHC13 Average Mouse 2.57 2.71 2.64 Human 0.21 0.21 s/jom 22 3 t S BFG 02033 I Validation of Human Model (Barettaetal., 1969) V20m 25 Deriving Rat Potency! Based on Maltoni's Experiments 12 Months Exposure 0,1,5,10,25,50,100,150, 200, 250,500, 2500,0000,10000,30000 ppm tested Poor Survival 10000 and 30000; Use Remaining 13 Dose Groups Use PB-PK Model to Calculate Pose Average Amount VC Metabolites per day per Liter of liver Tissue Howe & Crump's GLOBAL83 Multistage Model Dose Response (Maximum Likelihood Estimate) Comparison: Linear Model Pitted to Top Two Doses (MTD, MTD/2) 26 BFG 02034 I 5/20/94 PKOom PPM Vinyl Chloride MmnDosa B Mdma 27 Extrapolating Rat -> Mouse (Maitoni, Swiss Albino Mice) Cone 0 50 250 500 2,500 Males 0/80 1/30 9/30 6/30 6/29 Females 0/70 0/30 9/30 8/30 10/30 LADD 0.0 38.4 173.1 265.2 331-0 Equivalent Amounts of Metabolite produce Equivalent Tumor Yields No Surface Area Correction Factor Used. The 1(H RSD = 0.80 x lO1 S/20<94 28 BFG 02035 I Comparing RSD'sfr Maltoni et al., (Rat) Maltoni et al., (Mouse) Lee et al., (Mouse) 0.177 0.080 0.120 Drew et al., (B6 Mouse)^^r0.0032 Drew reported angiosacromas in )Is. Lee and Maltoni saw none. B6C3F1 Ultrasensitive? Reported to have partial oncogene activation in absence of any chemical treatment. s/2<m 29 Extrapolating Rat -> Humans (Maltoni. Rat Potency) Equivalent Amounts of Metabolite produce Equivalent Tumor Yields No Surface Area Correction Factor Used. Calculated "Unit Risk" Lifetime Exposure to l/^g/m3,24 hr/day. PBPKMLE = 4xia7 PBPKUCL = 6xl0-7 IRIS Number = 840 x 107 3/20/94 30 I VC Tumor Registry (Simonatoetal., 1991) 12,706 Individuals from Population of 14,351 Completeness of Followup = 97.7% Cohort has> 25 Years since 1st Exposure to VC Exposure Groupings: + 0 - 2,000 ppm years + 2^09 * 6,000 ppm years + 6,600 * 10,000 ppm years + > 10,006 ppm years Absolute Risks Estimated to Range from 6.2/100,000 to 280/100,000 31 PBPK Risk Assessment | Versus Simonato et ai (1991) | PPM PPM Veu Ten Ytan Exposure SO 500 100 1,000 200 2,000 -- 4,000 500 5,000 -- 8,000 1000 10,000 -- >10000 2000 20,000 Twenty Yon Exporar 50 1,000 100 2000 200 4,000 -- sno 500 lOjOOO -- 15,000 1000 2000 20,000 40,000 PB-PK LADD PB-PK Pitdirtmi perlWmO Observed Cm* per 100,** 3.33 6.63 13.06 -- 26.68 -- 3128 -- 36.03 188 374 736 -- 1,497 -- 1,753 ~ 2,532 -- (62)* -- 422 -- 152.3 -- <280.O>b -- 6^6 1326 26k.ll -- 53.35 -- 62.57 72.07 376 747 1,465 -- 2971 -- 3/476 3,993 (62) -- 422 152.3 cmafi s/2tm >2 BFG 02037 I Summary) Straight-Forward Modification of Existing PBPK Model Based on Rat In Vivo Studies. Validated with Mouse and Human Data Described Tumor Data 1-6,000 ppm in Rats and Predicted Tumor Data in Mouse Studies Unit Risk Based on PBPK Principles 150 Fold Lower than Current IRIS Value. Tumor Predictions Most Accurate WITHOUT Surface Area Correction Factor When Mechanism is Known, PBPK Procedures Should Be Capable of Giving Much More Accurate Estimates of Risk. snon* a BFG 02038