Document zbbMgDrpNgk1EqBdEvvqp3ddz

; '., ~lroili'l~liil ~RV@B9l\!11Ml UrnsltD!mlill I 1220 l Street, Northwest- Washingion, D.C. 20005 - ' ~o~ril "ir. ~~. ~llaJO. OiiOCIGr-(:lll2") 682"-830"8 - - Soioncoo D>poltm:>ftl (2D2) EIM!i70 (FAX) August 29, 1995 Dr. lfolm Douli!Department of l?'harmacology, Toxicology & Therapeutics University of Kansas Medical Center 3901 Rainbow Blvd.. OCansas ~r 66103-7417 JOear P1(DOull':- The American J?etroleum Instirute (API) wishes to bring to your attention several aspects of the presentation by Dr. Robert Spirtas entitled, "New 1LV foe Benzene" -at the Benzene '95 Conference on June 20, 1995. n..nus_ talk was presented as a description of the rationale for the current proposed V of 0.3 ppm foe benzene. DY. Spirtas-was listed on the program as representing ACGIEL API is extremely cont:emed because the rationale in the peesentation by Dr. Spirtas bore little resemblance to the rationale presented in the written documentation for the intended change dated March 20, 1995, copies of which were made available to the audience by ll>r. Spirtas. 'I'htee major points that Dr. Spirtas made Wring his presentation in support of the current proposed change 3I"e not found in the official, written -documentation. o li>r. Spirtas indicated that the 11.V Committee decided to use all leukemias as the endpoint for determining a 1LV rather. than AMLs because changes in the KCJD codjng for leukemia over time affected the reliability of cell~type-specific analyses. 'This explanation is incorrect and is not found in the written "documentation. o One explanation Dr. Spirtas gave for rejecting of the full likelihood analysis of Dr. Crump wa.s-the possibility that the data were not normally distributed. 'This objection is not in the written documentation either. De. Spirtas subsequently agreed with Dr. Crump (in a telephone conversation after the conferencetthat there is no requirement that the--data themselves be normally distributed. JH!e agreed with li>r. Crump that the distribution of the residuals between observed and expected leukemia rates would be expected to be normal. o Dr. Spirtas further rejected the entire Crump analysis in favor of an SMR approach because theJifetime occupational risk calculated by ~ of the nonlinear lfoxms of the Crump model emibited a large variation in predicted lemremia incidence in responseto different exposure estimates suggesting that the model is unstable. While the appropriateness of using a riskmodel that is sensitive to the exposure input variable can be debated, the APl's concern is again that this criticism does not appear in the written documentation. An equal opportunity employer BP-00017259 Dr. .Dolm IDouU Augusn: 29. !995 ll"age 'JI'wo lHiearlng mch a presentation and .contrasting it to the rationale published by the ACG:ffiLfor the proposed bemene n.V is disturbing in that we do not know what arguments we are attemptlng to address in our oommments to the 1I'1LV Committee. lit appears from IDr. Spims' talk that illl responding 1t0 ~ms raised in ~ doommentation, we may rot be addressing all of the concerns of the 'IlLV Committee. lF".Dlll2l!ly, ooe other comment by E>r. Spirtas was ~ssibiy the most distwbing of SIJ!l. lllll cfiscussing fue Crump full] lli!re1iliood analysis of the !Pliofilm ~ Rle stated an additional reason for rejecting that mnalysis in favor rDf tlhe SMJR-based estimates was that the atter were simpler anci could be explained to til worker. lDlr. Spims Wed that he did not mow Rtow the Crump model couid be explained to a worker population sod nmt be wasn't sure that he understood it-himself. You have stated many times in referring to the approach 1l!Se4ll by the 1LV Committeet!o.at, "...we follow the data.". APll agrees-that scientific data should be the basis lforr any TlLV and hopes that ACGm is not abandoning this approach in favor of setting a benzene n V based upon the "data that anyone can follow". APl! wouid like to request that the manuscript submitted by Dr. Spirtas for publication in the proceedings of the lBemene '95 Conference be revised to be consistent with the written documentation for the benzene 11LV as .issued by the TLV Committee. lif the final manuscript contains inaccuracies and inconsistencies such IDS the ones listed, then the manuscript should either be withdrawn or contain a disclaimer clearly indicating it to be the opinion of one member of the n..V Committee rather-than that of the entire body. As was evident at the Benzene 95 Conference there- is growing consensus that Al\m.. is the only leukemia type associated with benzene exposure, and that the dose response is clearly nonlinear. We are unhappy ofwith the widening disconnect between the growing base scientific knowledge about benzene-induced lewemia and the proposed change for the benzene 111..V. Dr. Spirtas' ora1 presentation at Benzene '95 further widens this disconnect. API. nonetheless, hopes that we can continue our open dialogue with the TiL'V Committee to arrive at a 1fLV for benzene that has a firm foundation in science and protects the health olf exposed workers. Sincerely, ~ Robert T. Drew, lPhD. cc: Mr. William Wagner BP-00017260