Document Ra3Xx4qnByKDyD7226eDYrkJz

. m-- .s April 25, 2001 S2C Pmal,hMNo55144-1000 AR 226 - Joy 2 = - EHR - 060l _ 0373. a k Mr. Charles Auer Officeof Pollution Prevention and Toxics U.S. Environmental Protection Agency 4W0a1shMingSttorene,t DS.CW. 20460 a = epa-oTs Re: Serum Half-Life of Perfluorooctane Sulfonate Dear Mr. Auer: (WURHETEa 000811820K "wTihtihs wyiolul,sOusmcmaarrHiezrenathnededzi,scMuiscshioanelduSrainntgoroourancdonmfeerernecgearcdailnlgonneMwarincfhor1m9a,t2i0o0n1 obtained as partofthe ongoing serum halflife study on PFOS. As you know, 3M has been attempting to calculate a half-life for perfluorooctane sulfonate (PFOS) in serumofretired workers. We have previously reported on two diinftfeirciumltdtaotainsteetrs.pret3. MThhaesycdoontniontuecdortroosbtourdaytethoiusripssruieo.r cCaulrcruelnattiloynsavailable data are In January 1999, 3M provided EPA with limited data from 3 workers indicating a hinalpf-rloigfreesfosraPlFaOrgSerinstruedtyi,readnwdoornkeJruslys1e5r.u2m0o0f0a,pspurbomxiitmtaetdetlhye 1fi,r4st00indtgyesr.im 3reMporhtad on serumhalflife of PFOS among 27 retired employees. These employees had worked at 3M's Decatur. Alabama and Cottage Grove, Minnesota manufacturing plans. `This initial analysisofthe 27 retirees used data points collected in November 1998. May 1999 and November 1999, designated t, t; and tz. Using a one-compartment `model. the mean serumhalflife determined for 18ofthe 27 employees was 303 days (range 139 -- 640 days). The remaining 9 employees either had less than three data points or did not show a trend. The 303 days was near the serum halflife that was initially reported for cynomolgus primates (274 days). Although we were aware of several limitations in this study. 3M considered the interim data to be an improvement over our previous estimate. based on only 3 employees, and using varying analytic techniques for determining PFOS. However. the Interim Report advised caution in interpreting the limited data set. Serum was collected in May 2000 for a fourth data point (t;). Initial analysis was returned several months later and showed results inconsistent with those expected. This caused us to reexamine ofsourcesof error in the laboratory measurement. The: analytic technique exhibits variability of 20 to 30% when the same sample is. Contain NO "Ri Mr. Charles Auer April 25, 2001 Page 2 `measured at different time points. The best way to reduce this variability is to run all samples from one subject during one analytic run. Random errors represent another typeoferror, and multiple analysis of the same sample can be used to reduce errors from this source. Nineofthe 27 subjects, representing arangeof PFOS values, were selected for reanalysis. Triplicate analysis was done for all time points (t through 13). All samples for each subject were analyzed during one analytic run. In this way, controllable sources of variability were reduced. Preliminary analysisofthese data does not corroborate the range for serumhalflife reported in the first Interim Report. The data from these nine subjects for t through t; suggest a. minimum half life near 1500 days, but we are still exploring various issues that confound interpretation of the data. Other factors complicating assessment of the serum half-life in retired workers may include the possible presence of other fluorochemicals from occupational exposure that may be metabolized to PFOS during the study; ongoing non-occupational exposure during the study; or difficulty deriving average valuesif the elimination Kinetics are concentration-dependent, since the workers began the study with varying levels. At this point, the caution admonished in both 3M's Interim Report and the October 2000 draft OECD Initial Assessment is appropriate. Neither the current calculation nor prior values should be considered conclusiveygstimates. `We plan to submit a second interim report in the summerof 2001.Ourplan after that, however, is to collect data for a periodof at least two years and possibly longer before doing any more analysis, which will then be done in such a way as to reduce laboratory variability as much as possible. Although this type of study continues to carry many limitations, it remains a good opportunity to measure human serumhalflife. albeit at concentrations many times higher than those observed in non-occupationaily exposed individuals. Please do not hesitate to contact meif you have any questions. Hog Rp , MPA |rc LarryRY Zobel. M.D.. M.P.H. Medical Director : Mr. Oscar Hernandez Mr. Michael Santoro