Microsoft word - 127th meeting sab minutes nov 15 2007.doc

ONE HUNDRED-AND-TWENTY SEVENTH MEETING OF THE SCIENCE ADVISORY BOARD (NCSAB) ON TOXIC AIR POLLUTANTS Proceedings of the November 15, 2007 Meeting Dr. Starr called the meeting to order at 2:05PM. NCSAB Members Drs. Thomas Starr, James Gibson (by telephone), Wayne Spoo, Woodhall Stopford, and Elaina Kenyon were in attendance. Dr. Yim had a schedule conflict and could not attend. The minutes of the 125th and 126th meetings were approved as amended. Acrylonitrile The introductory material for the acrylonitrile (AN) risk assessment document has been completed. The drafting of the remaining sections of that document should be completed and transmitted to NCSAB members by the end of the first week of December for review. Dr. Starr reported that in a recent conversation with Bob Fensterheim, he was told that: (1) Jack Mandel, Bill Cole, and Jim Collins were writing a review of the AN epidemiology literature. A draft of that review should be completed about the first week in December. The NCSAB may wish to have Dr. Mandel (Emory University) present the results of this review, perhaps at the January NCSAB meeting, if he is available. (2) the Symons paper, an update of the DuPont epidemiological study of two cohorts, has been accepted for publication. This is important because the exposure levels were about 10 times higher that those in the NCI epidemiological study. (3) Richard Albertini is writing a review of the genotoxicity of AN. The NCSAB Weight of Evidence assessment could be affected by an addition of a genotoxic component to the mode of action. We have been working under the assumption that there is a genotoxic mode of action, attributing all of the excess lung cancer mortality in the NCI study to that mode of action. The Albertini review may lend additional support to this assumption. Dr. Spoo reported that he transmitted the data on AN exposure from cigarette smoking to NCSAB members. Perfluorooctanoic Acid (PFOA) Dr. Stopford led the ensuing discussion. He commented that he thought it would be useful to include PFOS in the NCSAB discussions on PFOA because there are similarities in the toxicological endpoints. In addition, Dr. Stopford wished to recap the work of other state and federal agencies on PFOA in groundwater. A review of pertinent, published, peer-reviewed literature on PFOA was completed and summarized for the NCSAB; this summary is appended to these minutes. In addition, Dr. Stopford reviewed the similarities between PFOA and PFOS, and then led a discussion of how this information might inform the NCSAB in making any recommendations to the Division of Water Quality concerning PFOA. Concerning the regulatory agency actions, Dr. Stopford summarized the basis for their approaches and commented that: (1) The New Jersey approach (see the Appendix for more detail) suffers from several weaknesses: a. Dr. Stopford did not find where differences between species were addressed. There were great differences in half-life between the various species studied. The human half-life is 3-4 years, where animal half-life is on the order of days-to-weeks. b. The New Jersey standard is based on a study by Sibinski (Sibiniski, L.J. (1987). Final report of a two-year oral (diet) toxicity and carcinogenicity study of fluorochemical FC-143 (perfluorooctanate ammonium carboxylate) in rats. Vol 1-4, 3M Company/RIKER. October 16, 1987 (cited in USEPA, 2005), an unpublished study that was apparently not reviewed by New Jersey. It appears that they pulled a table of endpoints out of the Sibinski report to use. c. For cancer effects a 10% outcome level was used and a straight-line extrapolation was made to determine the concentration corresponding to10-6 risk based on animal data. This approach was used even though New Jersey did not know what was causing cancer in the animals. Dr. Stopford disagrees with this and explained that while PPAR-α cannot explain all animal cancers, in rodents one would expect an increase in liver cancer and possibly other organs. There are two sites where one would expect increased cancers: in Leydig cells and in mammary cells (due to decrease in testosterone levels and increased estradiol levels). (2) The German Ministry of Health approach does not extrapolate acceptable exposure concentration from blood levels and does not use pharmacokinetic modeling to address marked differences in half-lives. These give rise to additional uncertainty and are weaknesses in the study. (3) The Minnesota Department of Health approach uses a ratio of half-life in humans to half-life in monkeys of 70 to adjust their assessment: 3.8 yr in humans vs. 20 days in Cynomolgus monkeys. Minnesota referred to a computer model for determining water consumption by age, but it was not referenced in their report. This report did not consider serum PFOA levels or pharmacokinetic modeling. Dr. Stopford commented that he likes the monkey data: (1) There is consistency between PFOS and PFOA for every endpoint examined (2) Primate data more closely approximates human response than does rodent data (3) Except for the rodent cancer study, the monkey study is a relatively long-term study (4) The toxicological endpoints are non-PPARα effects and that explains more completely the cancer risks, even those in the rodent studies. Dr. Spoo agreed with Dr. Stopford that the monkey data were the primary data that should be used for a PFOA recommendation. Dr. Kenyon also agreed, adding that very little weight should be given to the rodent cancer study. Dr. Spoo commented that the same conclusions were reached in Klaunig, J.E., M.A. Babich, et.al. 2003. PPARα Agonist-Induced Rodent Tumors: Modes of Action and Human Relevance. Critical Reviews in Toxicology, 33(6): 655-780. Dr. Stopford added that the rat tumors observed in the two-year study could have resulted from hormonal disruption and not from peroxisome proliferation and thus should not be discounted altogether. Dr. Starr commented that another issue to be considered was if a PBPK model for humans was going to be used to get to a human intake rate what is needed in the way of an interspecies factor? One of the ways New Jersey adjusted for the interspecies half-life issue was to use an additional uncertainty factor of 10. It was generally agreed that the NCSAB would not take this approach. Dr. Kenyon stated that it would be a good idea to include in the narrative a tabulated summary of studies that estimated serum PFOA concentrations at which no effects were observed in humans. Dr. Starr added that the Clewell PBPK model could be used to back out the intake rates needed to produce those serum level concentrations. Dr. Gibson stated that he liked the Minnesota approach and generally agreed with their uncertainty factors, but was not settled in what he would recommend as an approach for development of a recommended PFOA concentration in groundwater. Dr. Starr commented that these issues should be further discussed. If the monkey is used in a pharmacokinetic model to derive an equivalent effect level in humans, then what is the need for an additional interspecies factor? At a minimum, the pharmacokinetic part of the interspecies factor will have been already taken into account. If there are no observable effects in humans (even at relatively high (occupational) exposure levels), should the pharmacodynamics part of the interspecies factor be considered at all? Dr. Kenyon stated that since there are no available data on interspecies variability, relative to PFOA, there is no justification for not using an interspecies factor of 3. Dr. Starr asked that insofar as PFOA is concerned, whether there is any reason to believe that humans are more sensitive than monkeys? Dr. Kenyon responded that there were no data to suggest anything about sensitivity, one way or the other, so there would be no reason to depart from the traditional use of the pharmacodynamics part of the interspecies factor. Both Dr. Spoo and Dr. Gibson concurred. Dr. Kenyon continued by stating that there are also no available data on human variability, either. In addition, there are issues with the LOEL-to-NOEL factor, and this factor, and the value used for it, provides perhaps the most room for discussion. There was general agreement within the NCSAB that the monkey data were the data that should be used in the risk assessment. Action Items for next NCSAB Meeting (1) send the EPA draft risk assessment on PFOA to NCSAB members (this assessment has apparently been taken down from the EPA website). (2) find out the source for the Minnesota water intake rate used in their Health Risk Limit (3) post Klaunig PPARα (reference above) article on ftp site for member distribution. (4) complete draft of AN risk assessment. Other Business A proposed 2008 meeting schedule was transmitted to members for comment. The proposed schedule includes the following dates (all Wednesdays except as noted, 2:00PM): Jan. 31 (Thursday), Feb. 20, Mar. 26 (moved because of conflict with SOT), Apr. 23, May 21, June 26, NO JULY MEETING, Aug. 27, Sep. 24, Oct. 29, Nov. 19, and Dec. 19. Agenda for the Next NCSAB Meeting The next meeting of the NCSAB will be held at 2:00PM on THURSDAY, December 13, 2007 in Room AQ526 of the 2728 Capital Blvd. building. A call-in number has been reserved for those not able to attend in person. The call-in number is (919) 733-2441. Agenda items will include: (1) a progress report on the first draft of the acrylonitrile risk assessment and (2) a discussion on how the NCSAB will proceed on PFOA. Dr. Stopford will be the lead discussant. The meeting was adjourned at 4:15PM. Respectfully submitted, Reginald C. Jordan, Ph.D., CIH Liaison, Science Advisory Board Approved by the NCSAB, 128th Meeting, January 31, 2008. Regulatory Actions Concerning PFOA Woodhall Stopford, MD, MSPH 15 November 2007 New Jersey DEP Risk Assessment for PFOA (ca. 2006) Because of the disparate biological half lives of PFOA in experimental animals and man, New Jersey decided to use blood levels in experimental animals for a given end point for comparison instead of external dose. When blood levels were not available, they used data from EPA’s draft risk assessment for PFOA (2005) for pharmacokinetic modeling of blood levels in experimental animals given certain doses of PFOA. To extrapolate to humans, they used data from Emmett et al (2006) that compared blood levels in humans exposed to drinking water contaminated with PFOA. This latter study found that for a given drinking water concentration, blood levels in exposed individuals averaged 100 fold higher with median serum levels of 354 µg/L and water levels of 3.5 µg/L. At these exposure levels there were no evidence of liver, renal or thyroid abnormalities among 371 randomly chosen residents drinking the contaminated water. . New Jersey used the following uncertainty factors in their analysis: Interspecies: Less than chronic (monkeys but not rodents): Default 20% contribution of exposure from water For cancer effects, they used a linear model starting at the dose associated with 10% tumor response (for 3 different tumors) and extrapolated to a 10-6 risk level. EPA discounted both risk of liver effects and cancer to humans from exposure to PFOA based upon there finding that these effects were mediated through a mechanism involving the activation of peroxisome proliferator activator receptor-alpha (PPAR-α) and that mechanism is not relevant to humans. Humans are unresponsive to peroxisomal proliferators (such as hypolipidemic drugs) that are active in rodents. New Jersey, however, accepts the EPA SAB assessment (2006) that there could be alternate mechanisms for liver effects, including cancer. The SAB felt that PFOA should be considered likely to be carcinogenic to humans based on two positive animal studies, the finding of a mammary cancer effect when the comparison group is concurrent controls, disagreement that Leydig and pancreatic acinar cell tumors are related to PPAR-α agonist activity and the finding of liver effects in knock out mice lacking PPAR-α. PPAR-α agonists inhibit testosterone biosynthesis with associated compensatory increases in LH, Leydig and cell proliferation, a mechanism not involving peroxisomal proliferation (Kennedy et al, 2004). For a cancer end point, New Jersey found than an acceptable water level would be 0.06 ppb. For anemia as an end point, using an uncertain factor of 100 and a relative source contribution factor of 20%, they found an acceptable water level would be 0.04 ppb. They recommended this latter level. The end points considered by New Jersey are summarized as follows: Table: Measured and estimated blood PFOA levels at various end points Species German Ministry of Health Guidance for PFOA and PFOS in drinking water, 2006 Germany relied on data from EPA’s 2005 draft risk assessment to determine an acceptable water level for PFOA. EPA specifically found that PFOA was not genotoxic. Germany used reproductive toxicity in rats as the end point of concern with an LOEL of 1 mg/kg/d. They used the following uncertainty factors: LOEL to NOEL: Prolonged ½ life in humans: 10 Their tolerable daily intake for PFOA was then calculated at 0.1 µg/kg/day. For PFOS they reviewed a 2-year rat study (OECD 2002) with an NOEL of 0.025 mg/kg/d and the NOEL of 0.15 mg/kg/d seen in monkeys in a 6-month study (Seacat et al, 2002). They accepted the risk assessment of Thayer and Houlihan (2002) that derived a tolerable daily intake for PFOS of 0.083 µg/kg/d (rounded to 0.1 µg/kg/d). For each study they assumed a 10% exposure allocation to water, a drinking water consumption of 2 L/day and a 70 kg weight to derive a lifelong guidance level for PFOA and PFOS of 0.3 ppb. Minnesota Department of Health risk assessment for PFOA and PFOS, 2007 Minnesota noted that EPA has derived an action level for PFOA of 0.5 µg/L as part of a Consent Order for Dupont. They note that the North Carolina health-based value for PFOA is based upon a monkey study and modeling to estimate a dose of concern and determine an acceptable water level of 0.63 µg/L (Williams, et al, 6/20/07 memo) . North Carolina accepted a reference dose calculation made by CIIT researchers based on serum levels of PFOA, a 30 fold uncertainty factor, a relative source contribution factor of 0.2, a water intake of 2 L/day and a body weight of 70 kg. Monkey serum levels were scaled to humans using a pharmacokinetic model. The resulting reference dose was 0.09 µg/kg/d. Minnesota used this same end point for their assessment but a different approach to extrapolate from monkeys to man. The LOEL in monkeys in a 6-month was 3 mg/kg/d with a finding of hepatomegaly (Butenoff, et al, 2002). Based on the 70-fold difference in the biological half-life for PFOA between the two species (3.8 years in man and 20 days in Cynomolgus monkeys), an equivalent human dose would be 0.043 mg/kg/d. They factored in the following uncertainty factors: Intraspecies: to determine a reference dose of 0.14 µg/kg/d. They used a relative source contribution factor for water of 0.2 and a water intake of 0.053L/kg/d (intake from birth to 19 years) to determine an acceptable water level of 0.5 µg/L. Minnesota also used the 6-month monkey study with PFOS, where an LOEL of 0.15 mg/kg/d was found for hepatomegaly for their risk assessment. The difference in biological half lives between humans and monkeys for PFOS is 20 fold (5.4 years in humans and 110-132 days in Cynomolgus monkeys). They therefore calculated that a human daily dose of 7.5 µg/kg/d would be equivalent to the LOEL dose in monkeys. For this risk assessment they used an uncertainty factor of 100 (an uncertainty factor of 3 was used to go from LOEL to NOEL because the effect was of minimal severity) to determine a reference dose of 0.075 µg/kg/d. They used a relative source contribution factor for water of 0.2 and a water intake of 0.048 L/kg/d (intake from birth to 27 years) to determine an acceptable water level of 0.3 µg/L. Similarities in toxicity of PFOA and PFOS Endpoint References: Scialli et al (2007), rodents; Butenoff et la (2002) PFOA monkeys; Seacat et al (2002) PFOS monkeys At 0.75 mg/kg/d for 6 months in monkeys, PFOS caused increased mortality, decreased triiodothyrodine, lowered estradiol levels, and hepatomegaly without peroxisomal proliferation. At 3 mg/kg/d for 6 months in monkeys, PFOA caused hepatomegaly secondary to mitochondrial proliferation with little evidence of peroxisomal proliferation based on small increases in peroxisomal CN-insensitive palmitoyl CoA oxidation. Deaths occurred at higher doses as well as decreases in thyroxin, triiodothyronine an estradiol.

Source: http://daq.state.nc.us/toxics/risk/sab/minutes/127.pdf