Document 3JXQJO8rbN7QyR69MvvLJyGox

August 27, 1991 Ms. Meredith Scheck Assistant Director The Vinyl Institute Wayne Interchange Plaza 155 Route 46 West Wayne, NJ 07470 Dear Ms. Scheck: Joe Ledvina informed me of the potentia interest that the Vinyl Institute has in a new quantitative cancer risk asses ment for vinyl chloride. He also stated that you need more definitive information on what I am proposing to do, its importance, and how it could be use by the Vinyl Institute. In August last year I was contacted by Vista Chemicals to evaluate why the State of Mississippi was using a more s tringent ambient air emission standard for vinyl chloride than had been used pi eviously. The former number was based on an EPA report dated 1984. The more decent stringent number which resulted in a 12 fold higher risk was taken from a more recent 1985 EPA publication, Both documents were prepared by contrac tfors for EPA and neither really underwent a good scientific peer review. Apparently EPA's regional office in Atlanta is recommending states use the uantitative risk assessment methodology in the 1985 document. I foil:nd several errors in the 1985 risk assessment procedure and discussed thes errors with EPA's Hazard Assessment Group (formerly Csr.c-r Assessment Group) They agreed that the cuartitati-o risk assessment was in error and thought until new data become available, the quantitative risk assessment procedure i n the 1984 document should be used. EPA has not conducted a recent compreher sive review on vinyl chloride but is aware of new data and risk modeling tool s that may be useful in conducting a new quantitative cancer assessment. Alt]hough conducting a new risk assessment for vinyl chloride is not currently on t heir high priority work agenda, at least two staff members in the Hazard As sessment Group are continuing to conduct work on vinyl chloride. Now tha t EPA is devoting so little time to this work, it is an opportune time to do an independent, thorough examination of the recent data and risk assessment methods that would be applicable to a new quantitative assessment. Such an ass essment would most likely improve on the current simplistic and conservative vinyl chloride model currently in use. Because of EPA's current workload, they do not have the time or expertise to improve on the existing methodology and vould be receptive to having an independent scientific assessment perfor ned. SPI-00971 BACKGROUND Vinyl chloride is a category A a rcinogen (a known human carcinogen), however, the epidemiology data have not beien of sufficient quality to use to quantitate risk. In lieu of this, EPA used rat studies where the exposures were adequately known, in order to es tablish a relationship between cancer incidence and exposure in the hi gh exposure range. The linearized multistage model was employed to extrapolate to exposures and cancer incidences that were about 100,000 less then observed in the study. It is assumed that the same relationship between exposure and cancer at very high levels of vinyl chloride holds for very low levels of vinyl chloride and cancer incidence. This assumption has always been crit:iiclized by the scientific community. The linearized multistage model fits the observable data to a quadratic relationship and essentially is a curve fitting exercise (no biological information concerning cancer mecfi anisms or pharmacokinetic data on vinyl chloride was/can be used in this aodel). A statistical 95% upper bound of the linear term of the quadratic is e aployed (ql*), which at low concentrations results in a simple linear relati nship cancer incidence - ql*(expo ;ure) Usually a lifetime cancer inciden: e of 1 in 1,000,000 is considered acceptable and the exposure that is determin sd from the above equation results in this risk. For vinyl chloride this exi osure is 0.00015 ppm using the current EPA methodology. Because of criticisns from the scientific community and recommendations from EPA's science advisory board, the agency is becoming more receptive to alternative cancer mo dels for the linearized multistage (personal communication with Dorthy Felton, EPA Risk Astess.-r.sr.t Fcrum, Chairperson), Therefore, the development of new more scientifically justifiable quantitative risk assessment methodology using biological data is both timely and needed and can result in lower risk esti:ihates than the conservative model now in use. Chemicals can initiate and/or pron>ote cancer. Initiators are agents that interact with DNA in some manner to cause a change in the sequencing that results in a mutation. Recent st ijdies have shown that for many cancers to develop more than one mutation mu. t occur. This is in contrast to the "one hit" theory of carcinogenesis on vfhich regulatory agencies base their current model. The cancer incidence woulc be nonlinear with respect to exposure concentration and would suggest tl at the currently used linear model overestimates the risk at low con dentrations. Pure promoters such as dioxin are not mutagenic, rather they in iluence the proliferation of cells such as preneoplastic cells or clones that have already undergone one mutation (transformation). The relationship between promoter concentration and cancer is also likely to be nonlinear Vinyl chloride is an initiator but has to be-metabolized to an active form in order to cause a mutation. This ac tive form is also deactivated in the body SPI-00972 and excreted. Therefore, there is a balance between formation of the active carcinogen and its deactivation th at is dependent on exposure concentrations, Establishing a relationship betwe en external exposure and active carcinogen in the body, and then between active arcinogen and cancer incidence would result in a more accurate cancer model th .in the current model, where only the relationship between external cone antration and cancer incidence is used to estimate risk. At higher concentrations used to dc tect cancer in experimental animals, vinyl chloride may also act as a promote:- by causing proliferation of liver cells, Current models used by regulatory agencies do not account for promotional activity, but treat all carcinogen;: as initiators. By not addressing the promotional component would likely lead to overestimating the risks at lower exposure levels. If adequate data are available to demonstrate that vinyl chloride has promotional activity t could be used in the development of a new quantitative model. Epidemiology studies on vinyl chlo:- ide have been updated (personal communication with Charles Risk of EPA's Hazard Assessment Group). These studies should be evaluated to detc rmine if a dose related response can now be attained. If so, it may be possib e to use the human data for a quantitative assessment. Also, a determination can be made to see if the current linearized multistage model (rat s :udies) discussed above accurately predicts the observed cancer incidence usint the updated epidemiology studies. For benzene, the animal studies overpr dieted the leukemogenic risk in the human epidemiology studies (personal comnunication with Todd Thorslund). I have learned from EFA that they :ave data on vinyl chic tide that indicate young rats are more susceptible to vinyl chloride toxicity than adult animals. These data have not been publish'- ' -: should he evaluated. EPA is concerned that their current risk assessment may underestimate the cancer risk, because the risk assessment to set ambient levels was performed on adult animals to set ambient levels and may not suf iciently protect sensitive young individuals. They verbally agreed to give me the data to review. PROPOSAL There are essentially two parts to any quantitative risk assessment. The first part is to assess and organi :e the data available to determine if dose response relationships can be deve . oped. The second part involves taking the data and modeling it using statist Leal techniques to best describe the biological results. Depending upofi the data available, certain conservative assumptions have to be made in ord sr to err on the side of protecting public health. Usually, quantitative ass ssments performed on limited data result in a more conservative risk assessmen : (e.g., the 1984 risk assessment for vinyl chloride). This proposal deals only with the irst part of the quantitative risk assessment as first the data avail ible for vinyl chloride must be assessed and SPI-00973 organized to determine if the new data are of sufficient quality to develop a new quantitative model. This ini tial assessment should provide indications as to whether the development of a n sw quantitative model will result in a more or less conservative prediction of risks at ambient levels compared to the existing EPA model. The four part s of the initial assessment are: a) a complete review of the pharmacokinetics (metabolism) of vinyl chloride b)conduct a complete review of in ormation pertaining to the multistage process of carcinogenesis as it p rtains to vinyl chloride and an assessment of the potential for vinyl chlorijie to act as a promoter c) obtain and review the updated pidemiology data on vinyl chloride and compare it with the animal studie Determine if the human data are sufficient to quantitate risk d) obtain and review data regarding the greater sensitivity of young animals to vinyl chloride toxicity A decision can then be made if the development of a new model is worthwhile, Each of a these points is describe d in more detail below. A. Pharmacokinetic review Below is the current concept behind EPA's 1984 risk assessment. It essentially curve fits the experiinjentally high external vinyl chloride concentrations with cancer incider ce and extrapolates orders of magnitude downward to a risk of 1 in a 1,000 ,000. vinyl chloride(external exposure concentration) + normal cell -> cancer The actual cancer risk can be more accurately determined from the relationship between activated carcinogen and c ancer incidence. vinyl chloride---------(external exposure) ->activated carcinogen + (internal concentration) normal cell.............> cancer inactivati on of carcinogen \y excretion of metabolites To evaluate this more accurate mod j! a critical review of the pharmacokinetic SPI-00974 data on vinyl chloride will be mad co determine if che current data base will allow establishing a relationship tetween external vinyl chloride concentration and internal activatd carcinogen levels, particularly at low levels where the formation of actiA e carcinogens is not saturated and deactivation may be more prominent If possible, a qualitative assess will be made if the use of the relationshij is likely to result in an over- or underpredictior. of the current risk use<. by EPA. If the data base is inadequate the research needs will be provided to establish a relationship that can be used in a quantitative assessment. B. Multistage process review current model normal cell + vinyl chloride ------ > cancer cell two stage model ONE STAGE LINEAR RELATIONSHIP promotion (increase in ce 111 proliferation)' (vinyl chloride or me tapolite(s) ) normal + vinyl ------ > transformed + vinyl ------ > cancer cell cell chloride cell chloride (active (active metabolite) metabolite) TWO STAGE NON LINEAR RELATIONSHIP The two stage carcinogenic model i s a function of the square of the concentration of the internal acti ve carcinogen (metabolite). The promotional effect is likely to be nonlinear i nd may have a threshold. A review of the literature vill be performed to determine if data are available which indicate that the cancers induced by vinyl chloride (e\g., angiosarcomas) are multistage involving the mutation of more than one oncogene and/or regulatory gene and whether vinyl chloride has promotional activity. If the process is multistage then the current EPA method of estimating the vinyl chloride cancer risk is likely to overestimate the risk. C. Updated Epidemiology data rev ew SPI-00975 This part involves: 1. Acquire the updated data from EPA and determine its adequacy to establish a dose-response relation ship. 2. If the data are good eno igh, the current EPA quantitative risk assessment procedure (rat data) wi .1 be used to determine if it accurately predicts the risks in the updated luman epidemiology study. D. Review the vinyl chloride toxi :ity data for young animals This part involves: 1. Acquire the unpublished iata from EPA scientists. 2. Determine whether the da :a can be used to show that young rats are more susceptible to vinyl chloride toxicity than adult animals. 3. If young animals are mor s susceptible, assess whether this data can be used in a quantitative risk ass issment and can be extrapolated to humans. A report will be written that will contain an intrc luc.___ _____ ,^-ibing the purpose of the qualitative assessm snt, four sections (A,B,C,D) that will be organized into an assessment revie of the data with references, and conclusions based on the assessmen t. A discussion covering all sections will follow that describes how the asse ssments in A, B, C, and D are likely to influence a new quantitative cance : assessment for vinyl chloride, along with recommendations whether new studie > and/or a new quantitative assessment is justifiable. This constitutes the first part of the proposal. A proposal for the second part (modeling and stat istical analysis) depends on the findings in part 1 and can not be written at t lis time. It is possible, depending on the n ieds of C'.vj Vinyl Institute and the findings, to publish the first par : of the proposal in a peer reviewed journal. Regulatory agencies find published results more convincing and useable (to be decided after compl stion of Part I). Costs of PART 1 $29,000 Time to complete project The project can be completed withi 6 months, and could be initiated this November. A monthly progress repo t will be provided to the Vinyl Institute until the project is completed. T avel (e.g.,'meetings with the Vinyl Institute) is not included in this cost. SPI-00976 I hope this provides you with sufficient information to initiate the first part of the proposal in order to determine if a new quantitative cancer risk assessment for vinyl chloride is warranted and beneficial to the Vinyl Institute. It has been my experience that regulatory agencies like EPA find it difficult to change from established procedures. However, their established procedures are being criticized by the scientific community, and because of this EPA is likely to be more receptive to alternative scientifically justifiable models One cannot predict if a new model will be more or less conservative than th e existing model, but since the existing models are so conservative, it s ems unlikely chat any new model will result in lowering existing standards. If you need any additional infornation or wish to address any points that I overlooked, please call me at 70]-934-3126. I have enclosed my resume and Dr. Suresh Moolgavkar's resume. Suresh is a member of the Fred Hutchinson Cancer Research Center and professor of Epidemiology at the University of Washington. He would be involved in the epidemiology review in the first part of the proposal and would have a major i ole in part 2. I have also enclosed a recent workshop review paper on benzene that will be published this Fall, as well as a previous publication concerning different mechanisms of asbestos carcinogenicity. Sincerely, Peter Voytek Enclosures SPI-00977 PETER E. 70YTEX EDUCATION 1953 1969 1980 3.A., Chemi|scry, University of Vermont PhD., 3ioch|emistry, University of Vermont Program in Environmental Policy and Management:, Harvard University POSITIONS AND AFFILIATIONS Senior Science Advisor and Vi ce President, Clement International Consultant to the Center for Environmental Health and Human Toxicology, affiliated with Georgetown ard George Washington Universities Guest research scientist, Laboratory of Molecular Microbiology, NIAID.NIH Member of the Senior Executive Service, Acting Deputy Director of EPA's Office of Health and Environnental Assessment; Director of EPA's Reproductive Effects Assessment Group Senior Toxicologist/Pharmacologist and Science Advisor to the Deputy Assistant Administrator for the Office of Toxic Substances, EPA Senior Toxicologist in the Criteria and Standards Division, Water office, EPA Senior Research Fellow at NldHS, NCI, NIH Research Associate/Instructoi , Department of Pharmacology, Yale University AWARDS Selected into the Federal gov ernment senior executive service 1980, EPA. Bonus for Outstanding Achievement, 1980. NSF undergraduate summer grar t for research in Physical Chemistry 1962. SPI-00978