Document NmqneY1N9p11vdQp5QxMDjDE

DownloadSearchable PDFRandom document
EDITORIAL RESPONSES TO THE RINSKY ARTICLE IN THE NEW ENGLAND JOURNAL OF MEDICINE INCLUDING THE EDITORIAL BY WILLIAM F. O'KEEFE (October 15, 1987 Vol. 317, No. 16) SAL 000001667 PROJECT TIM ING PROGRAM FUNCTIO N COST BUDGET SOURCE b C a E a c a 9 a a E a a c c- 3 3U ite - a V r k. b a c aa --g c ue <ak a3 4# o-- a ** 4a c a *- oa a e* bc a to a b 3 a k. 0 I3 g-- oa a a s -- Wo ao a 9 a b 9 a a a --g e as-- o k a u 3 a c e >m a 9 ac s *- a n w a >- c < + a ak. Ea eV --a ac lb a > a X am Q e -w Ca a e a e BA* r 9 S . me M3 k a r 9a 9 > -U VOI ak ga 0L a aca a oi b axe 9 - a w x L e l. a aa3 -c a a3a a v a- a 3aaa aec oao a a < a aoca a k. a a - ax w oo e e m --a a 3 9 9 e a o o o u ow Z a c- s> 9k :a aa & uo Ik z a c 9 c a a aa a a *4 a 4a4 ca X4* 9a k --a ea a9---- 3c > bO' a --a aw -- --b *c4 <ab ak uaba E& 9 Ca a a a a xk 4> a a--a > a a --g Ea -- Ea > a X0 a^ > C o k -- k a c a 444" o 44 r - a E ga x oa b0 a <b u oo > 4" a M X o -- a C a --a U -- Sec a- a 0 u k 9 --9 *u aa c* a *4 e a a a ** > a 9 3 a a 4 44 aa a b a c k a 44 O bc -- --a 44 a 3 b> 44 ka a ak a a a aa 0 > 9 0 # 9 k be M 4a4 a Eb 44 a a x a k 9 C 3 k M bb c o 44 u wa> k a 9 C 3 a 9 k a 9 c c * 3 O 4O4 ba 44 44 a a a Ck u a --c X a E bC < a u bo a bba 44 e a 3 9 a a 3 a a C b0 ba u a 9 ba k c bo ba u a 9 m> b9 a a E a a ba* < aa u 3 a< ak > a ba b a aU # aa a --a k0 a k a a a3 c a Ml ba u C 44 b b a a- aX u w bc a> a 4 k b> ba bk a 9 ba e> e 0 k< ac C 9 rb a < a a c 9 - a 9 a * a 3 b b a a-- 44 (J C a E 9 k a e tv ca k 3 9 a 4a4 a 9 a aa U U 44 a 9 o- k a c 4a4 --b k a a a3 aa a a ao a be a bc o ao-- > aac 9>o fr 141 E O k k a c a N C a axm ab JL a a b> a 44 b> b c k ba b a a 9k a 9 c a a< C b a -- u a 9 9 30ft a 9 k 4e. * V a vt 9 a bc a a a b k o b -- -- B U a w a b a bk> b a b9 k b m W w0 w xa a9 oc -o 3 b 3 a 4cW a bc 9 fM b b C a h- a ba a bk o ab > ac a9 a - a a e b 44 k. x k C - > a - b a> b u *> o W k. - a 1- a E a - <w k a am a a a ab aa b9 ^ a C aa Lb >aa a 3 4k k 44 c aa aO o a oo c a a -- k ao a a a9 a 9 a >9 am 44 C a9 a a bC --* 9 a o ** Ca -- E a --s a 44 c - Ik e 3 a b kaJ a *- a a> o n-- a wa . 3 b9 a a a a >4 b e > U k C 9 a 9 - k -- a a kka aa k 44 C -ab b-. bc bO ae c e-o k aa a Dc b *4 b c 9c b a 3 0 a a O O O ob ca 9 3 9 44 . b b k. bM o k E v* a a a a a a -- a9 a a al u a al 44 u a 3 a 0 a -- --aw -- 44 t k b >. 0344 -- ac-aa- a a a u-- a b a 3 a o a 4* a k3 a a* a- bE -- a k b o w 44 # -- a ac a a u k aH C a a^ a Eac k -t a a e -- a M E 4a> a ek E3 ke 9e 9 C C a x a 9 a 9 E a a k am bg k* c x aa 3c aw > at oc Ji -- a- m & aa a K --u 9e g a --ai E 3 a C ek 9c 0 ft afi a a a 44 b ca aa w Q a. s -- 9 ac a b a aw aa ct E w a-- a aa o a eac k k-- > a 99 a-- ba 3> a k aa a o n JC a a a Ca a a 3 a u a k a a *4 9a 0 C U c -- ax a -- KB C a u9 9 C o u aa o aa c cc 9 9a i. eu ea ? w a O c3 a e -- x s-- U <k --a o lo k 9ma e ta< fc e *> e -- ae -- a b id oac k 9Caa o Cn --e aw o s. icaa a JC u k 9a a k 3 S a *kC wot ie ca !h -a aE *> a a e t. aa a* w l ra a kc 3k e ~ a~ wa aa c o ak U Development E ffo rt s u a itr 1980. m p o iu ri iis s tm in t p ro ja ct by providing te c h n ic a l and a n a ly tic a l s u p p o rt fo r ,4C approval SAL 00000166* VoJ. 317 No 16 CORRESPONDENCE 1027 lom-ierm survival among patients uh Stage III disease is 25 per* cent (two <rf eighti. Among patirnu mih Stage IV ditr^ir. 6 of 34 (18 percent) have survived overall. of the 23 patterns with bunr nwiuiatu and bunc marrow trtvulvcmeni at durttukii -- a group generally considered to have a particularly poor prognosis 5 (22 percent) have survived- The survival ol these five patients indicates that narrow "purging" mav not always be needed. .* Then has been a price for survival -- considerable treatment- related shromc toxicity Six of the eight children hav^ auditory injury affecting their hearing range9 and wear hearing aids. Five patients h\ve reduced glomerular filtration ram.4 twa> of whom have chronk renal failure: one of them requires amihvpertensivc therapy. Apiri from thesr handicaps, the eight children mow 6 3 lo 12.4 vears oid\median, 8 21 are leading full and acuvg lives, includ ing normal school attendance. This senes is (datively small, but a comparable rdte (71 percent) of response to thi\chemothrrapeutic regimen plus surgery has been observed in a larger study bv the European Neuroblastoma Studv Croup.9 Previous reports state or imply that no progress has been made over the past <10 years in treating advanced neuroblastoma and that the number of survivors who are mre man one sear old at diagnosis is negligible.V We challenge these assertions. Our experi ence suggests that the median survival of patients in Stage III and IN* has been extended nWh-dose melphalan k probablv an impor tant component of this improved prognosis/, therefore, the long term results at centers using more intensive "consolidation* regi- mens*'6 will be awaueo wufi great interest. Rather than adopt a negative attitude, those treating children with neuroblastoma should energetically search foV better induction and consolidation regimens so that reponse rate*can be improved. London. WC IN 3JH. Vnued Kingdom J PxrlCHAaix E. Kutv, DW Roccxs. L. SrrrL Li. Smaffoxo. R Bhixctov Ml i-LLB. AND M U|r,wfT Hosp.:al for Sick Children Evans A. D'Aagio GJ. Randolph J prepowd staging for efctMnn with ocureblastoms Cancer 19*1. 2? 37a Shafford EA. Rogers PnuJurd f Advanced neuroblastoma improved response rate using a suiLaprn: reuneclOPEC/ tncludisg soqueeuai cuptaun and VM-26 J Clin Oncol lit ! W-` McHxacy VA. Thibsdovu C. Hagv FA. drucr AA Hearing lots in ctui- dren receiving cirplauo eftemotfverap' J Psliatr 1913 102 314-7 Weaver RB. Pniehart J Bkrrrr 7^*1 Rersai tqucity of ctsplatifl in Children J Pediatr 1915: 106 639-63 S. Pmchari i. CemooC S. Jones <k Knku i vt S Is high dme melptu- laa (HDM) of value in lunmary results of a m of sdvsn neurobUsiaou iANi* Premo bv tbe pern Neuroblastoma Stud) Croup lENSCl foe Am n Oncol I 5 203 abstract 6 Grotfcld J. Neuroblastoma /isys DM ed Ip surgical oncology Orlando. Fla.. Cruae A Sopnon 1986 63-83 7. Voute FA. de Knka J. BuJger* JMV Tumours sympathetic nervous system: neuroblastoma. ganglioneuroma and hranocytoaa la: Voute PA. Barred A. Bl^Sm HJG. Lctaerw J. Ne MK. cds Cancer children: cbaieal taaipfemest 2nd od Bertui ponger-Vertag 1916. 231-51. t. AugustCS.Scrota FT.JCacbPA.ctal Tremnntof cod ncwoHsuo- a w-ith suprakthal edetnosbempy. mdsaaoo. and al k or ataotogous marrow iccorsmuikm J Cha Oncol 19VI 2 60 9. Haimaan O. JUiif/C, Bcwjcaa F. Bayk C. E. Lenerte J. Treatment of adva v* hifh-doae chemotherapy regimens and aiuolofous boot marrow A. D`Ang OilSeegCT RC. ads Adv mo is *aos New York: AlaiUW. Uu. 1915.563-1 10. fohp T. Btron P. Rullip I. et al Auiolegowt bone marral niplawiinun for very bad nn neurobli'tuma 1a Zvara a D'Angm Secft? RC. di. Advances us neuroblastoma research New York- in R 1915.369-86 AIDS .AND RENAL FAILURE 7* tht Editor. Rao et al. observed accelerated wasting anB early death in paiiems with the acquired immunodeficiency syndrome (AIDS) who required maintenance hentudulv su for therapy of endstage renal disease (April 23 issue).' The course appears to br asure rapid than has been observed in patients with AlDb who do not have renal failure Hus observation may indicate that patients w nh renal involvement have more extensive human immuntidefirietin viruwmlstrd diseaar and iherriorr die sooner. Huwpyer. h it also ptmihirVihai hemudiulvsis acrrlerates the procrestfun nf AIDS Normal immune stimulation of T helper eelis muy/truirr repluatton nf thc\AlL)S virus ' Furthermore, patirniyeeceiv ine mainte nance hrmudtaJvsis have an increased proportion of activated T cells in peripheral bloud. as defined bv oppression of interieuLin-2 reccpton\VV hether such acttvaimnyof T ceils reflects long term exposure to qluud transfusions, infectious agents, or dialvsis membranes or tubing, or is a mantfestaubn of the uremic state itself, it unknown. Intravenous drug use. ponent in a high proportion of the patients studied bv\Rao et ai .iiwuld also be associated uuh chronic immune stimulatwi. It mu^nt be interesting to see whether patients on long-term ambuVjtorypenioncat duly sis have increases in immune activation of T Ivirlgnocv to comparable to the increases observed in patients on maintraance hemodialvsis If increased Tcell acuvation is more marked in maintenance hemodtalvus than in other forms of therapy of Vnd-siage renal disease, suivival of pa tients with this disorder apa AIDS co^ld br enhanced bv ar ahernauve therapy (such as lo/g-ierm ambuUtorv dialysis, if Icasible. Philadelphia. P.- Xla.s NVasscxstco. M D N2aolvn Ksliv. M.D UrmVrsitv of Prnruvlvama \School of Medicine 1. Rao TICS. Fndma* EA, Ntcssm AD TNs rypes oKrena) disease m the Bquired imafuMdefieietiry ivodrorw N Lag] j Med raST 316 1062 8 2. Nabei G. kiuinore D A* tftdueible nasenpoao (sew amvia cspmsioo of bumaa'immunodeficiency virus in T <eUs Nanut 1917. 326 *11 3 3. ChuerwCd L.. Dugas B. Beaurasn A. ei si Presence of pnacuvaicd T ceils in bemo/alyied pauenu tbeir possible rote is sltemd unmusin Free Nail Acs/Sc. L'SA 1986 83 7457-61 BENZEN'E AND LEUKEMIA Te the Editor- In the studv on benzene and leukemia b. Rmik. ct al (April 23 issue).* there were but nine cases of leukemia anc four cases of myeloma. These are vrn small numbers on which to base conclusions, especially when the overall mortality has been as expected and the mortality due to malignant neoplasms has also been as expected. Of special interest it the authors' Table -I. which lists the cases Some had minuscule exposure, and one wonders what other malig nant disease--causing agents they may have been exposed io during their lives Of special interest is the feci that no employee first exposed after 1954 had leukemia or mveioma. even though hiring continued for II years and even theugh 7 of the 13 cases had the onset of disease within 20 years of first exposure. (Cases were fol lowed through 1981, so there should have been some cases ) In the study, two subjects died within 3'/i years of exposure. Looking at the year of initial exposure of the 13 cases, one can see that 9 had first exposures before the end of World War II. One wonders about the measuring instruments in that situation. Having spent some time studving childhood malignant disease more than 25 yean ago, ! am very sensitive to the tendency to observe dusten of cases sometimes living within short distances of each other. Statisticians have generally attributed these to chance, and there is no reason why that may not be true in the cases in this studv. - Seattle. YVA 98115 Max Radix, M.D. M.P.H 6536 29th Ave NE Rifwky RA. Smith A8. Hunsung R. el al. Bourne and leukemia aaepuk-miokjgw n^ aucumcnt. N Engl J Med 1987. J16 l(kU-30 Tv ikt Editor: The paper by Rinskv et al. on workers in rubber hydrochloride plants exposed to benzene reports a cohort study, a case-control study, and a risk assessment for leukemia in'rrLnon to bcuarnc. Table l. which uses dau from that paper, indicates a discrepancy between the cohort and case-control studies. The eac- SAL 000001669 1030 THE NEW ENGLAND JOURNAL OF MEDICINE Ori 15. IW7 Tdfcfo i. Ddith* from Ltuhoma. 9* StafldardttM Mortality Ratio (SMR). and trw PraU>ciad Odd* Ratio. Aocording to Cumufotiv* Expoaura to Bamana. Cuuwianv* bmui UFfl 0-0 40-200 200-400 >400 'From m nU'MMim Mil I In <4*0 ppmr*n t* Dum no* LevUNI* Omiaio SMt 2 l.l 2 2.2 2 11.9 3 66 Xuwt> *J.. M| Ub um^mt hunta OOoe llano* 13 4.5 43 1 423 N Ml lignum control risk assessment predicts, in the higher-expmurr categories, odds ratios much larger than the standardised mortality ratios ob served in the cohort. A possible explanation for ihe discrepancy is that the exposure of the controls was much lower than that of all cohort members without disease. In support of this, it can be in ferred from the paper that the mean cumulative levels of benxene exposure of the 1156 subjects without disease and ihe 90 controls were about 6" and 50 parts per million (ppmi-vcars. respectively It is reasonable to ask whv the case--control studs was done A case-control studv within a cohort u usually done to reduce the cost of determining the exposure of all cohort members without disease However, this reason is not applicable here, since the exposure to benzene of all cohort members was known. Rmsky et al oiler sever al reasons for doing the case-control studv. bui all the objectives could have been met effective! v bv a cohort analysis of ail the data. The disparity between the mean exposure level of all cohort members without disease <67 ppm-\earsi and that of the 90 controls (50 ppm-vears: requires explanation. It is probablv not the result of selection due to matching, sine- this should have raised the level in the companion senes The investigators mav have reviewed the information on the benxene exposures of cases and controls, with the result that the average level of the controls but not that of the cases, was lowered- If such a rev iew explains why the control expo sure levels were lowered bv 26 percent, then the validity of the cohort studv is questionable Finalh. if chance explains the differ ence between the cohort data and the case-control ml assessment, a nsk assessment based on the entire cohort would be superior Irrespective of the explanation, it would be appmpnace to evaluate the data further. Althoueh this might mvolve reevaluatmn of the benzene exposures of all cohort members, the results should clarify the relation between benxene and leukemia. The Occupational Safetv and Health Administration (OSHA) proposes to reduce the benxene standard of a 10 ppm time-weighted average to 1 ppm. A principal basis for this is the results of the rubber hydrochloride study. However, the magnitude of the doseresponse relation in this cohort is uncertain. We. in a review aupponed in pan by Texaco. Inc., and others have done risk assess ments based on this study and estimate that exposure to 300 ppmyears of benxene would cause 50 to B0 excess leukemia deaths per 1000 workers so apoaed.* The nsk assessment of Rinsky ct al. predicts 250. Birmingham, AL 35294 Hasland Austin, DSc. Euxaxxtii Dclxcll. D Sc Phiup Colc. M D.. Dx.P.H University of .Alabama at Birmingham School of Public Health Amu H. Detxell E. Cok P Benxene and leukemia: a rev** of tbe btervurr and a nsk assessment. Am J Epadesuot tie press) 7i itu Editor: In his editorial on new scientific evidence and public health imperatives (April 23 issue).1 Dr. .Ashford stated that 'an other example is the reanalysis of the National Canrrr Institute data on human exposure to formaldehyde.1 which now indicates a dear risk of cancer to humans." The authors of the study of humans expraed to formaldehyde* concluded that "these data pro vide little evidence that mortality from cancer is associated with formaldehyde exposure at the levels experienced by workrn in this uudv " Since substantial numbers nf health workrn air rxposrd to formaldehyde, it would hr of considerable mifrni if Dr Ashford could provide the evidence that indic-ain "a clear nsk uf cancer to humans.'' St.Johns. NF AIB 3Y6. Canada M Himv Gallt. M D Memorial Lmvmiu of Newfoundland I Ashford NA New wicniiAh evidence and public health im^nuvcs N En|l ) Med 1917. 316 I0SA-' J. Blair A. Sttwan P. OBcry M <J Mortality among industrial workers exposed io formaldehyde JNCI >9X6 76 10714a Tt >A/ E4\ter The editorial bv Ashford sounds a false alarm about the state of public health regulation today. Far from "waiting fur ihe bodies to foil." federal agencies long ago adopted the "preventive public health policies" that Professor .Ashford now pleads tor The Supreme Court's decision in the benxene case accepts this approach by authorising OSHA to "n*k error on the side of overproirctiun rather than underproiection The essence of the court s ruling a ihai this principle mu>i be applied within realistic limns, so lhai instead of seeking "to eliminate completely and with abso lute certamiv any risk of serious harm." regulators must fairu evaluate all the available scientific data and "make a rational judgment about the relative significance of the risks' involved Federal agencies hav e recognized that prudence, rather than an illconceived rush to judgment, is crucial io sound public health regu lation and have had no difficult implementing ihe vital mu assessment and pobcv-makifig disciplines contemplated m ihe bentene decision * 4 In fact. ObHA has alreadv regulated asbestos* and ethviene oxide" as carcinogens and will shortly presenbe stricter limns for formaldehvde The report bv Rjnskv et al on a much-studied cohort of benzene workers, which Professor Ashford cues as "bene.* science" requ.rmg an immediate "governmental response." actualK illustrates thr p::falls of reacting uncnncalU to a single publication Indeed, u fails io meet the standard of scientific excellence to wrhieh he subscribes Dunng OSHA's public hearings on benzene last summer.* major shortcomings in ihe paper by Rrnskv et al were uncovered, includ ing selective use of ihe available data on industrial hvgiene sub stantial underestimation of exposures, partieularlv during the J940s, at the work stations where thr excess cases of leukemia oc curred and where workers were known io have died from the acute blood-poisoning efTecis of benxene*. and failure to account for the prolonged, heavy exposures io benxene that manv of the work er! in this cohort experienced at work stations excluded from ihe analysis.9 10 There is nothing to be gained and much to be lost by fomenting a crisis atmosphere based on a flawed analysis that greativ exagger ates the level of risk, especiallv since average benxene exposures in the petroleum industry and elsewhere have for some time been re duced to levels comfortablv below the l-ppm limit OSHA has pro posed." Instead. Professor .Ashford should recognize, as ihe Ameri can Petroleum Institute doei. that thr best way to improve health protection against thr hazards we face today -- occupational and otherwise -- is to insist on a thorough scientific review of all the available data, focusing on the question of whai exposures pose significant risks. Washington, DC 20005 Wiluah F. O kurt American Petroleum Institute 1. AFICIO v. American Petroleum Ituiinne Industrial Union Dcpanmem Ul US. 19*0 607. 656 2. Russel M. Gruber M Bui assessment in environmental pol*)-rTMk>nf Science 19*7.236 2*6-90 3. Lave LB Health sad vaf.Ty nsk analvsev infurmanun for feuer ik\Uiorw Scenes 1917; 236 291-3 4 Okfciti D. The safety goals of the U S Nuclear Regulator} Commission Scsancc 19*7; 236 29b-3UU 3. PcpMUncat of Labor Ovxwpnnurul exposure to asbeuoa. irrn*au`. aathophyllite. and actinoiitc Fed Regisi iunc 20. 1916. 31 22612 <0 6. Dvpanmeniof Labor. Qv-i^ntiu*uJe\fvw*iie tocthykneosMk Fedftcgisi hue 22. 19*4. 49 25714-aw SAL 000001670 Voi 317 No. 16 CORRESPONDENCE 1029 7. DxiwAtw* of LNn. Tim mm* regnlimry Fm **p* April r. 199?. 32.14M1 I. Puu-lwannj cuwiiwwa of tht Amcncan Nootxwn Inwitow. OSHA Dorset No H459C. Ei 260. Jww 10. 1996 9-14. 3U-***- 97a-9a 9. D* PA Tone wbuancw a fct rubber wdusvy bnurt ftubb* Age 1929 367-8 10 Wilson RH Benzene poisonm* in utftiur? ! LabOia M 1942. 27 1317- 21. II. Rumen HE. Seal LM hum exposure Mihe United Stam 1978-19*3 M o**rvw AnJIMIM 1913; 7.313-93 Thr above letters wore referred to the authon of the articles in question. who offer the following replies 7* Mr Edit* Dr. Bader i concerned that our analysis rests on only nine deaths from leukemia and four from multiple mveloma and mav therefore represent merely a description of a random 'can cer cluster." That conclusion it highly unlikely, because of the close association of th* cases with an experimental!* proved carcino gen1'* and (he demonstrated strong, internally consistent doseresponse relation. Similar features are characteristic of senes of cases of other chemically induced cancers, such as mesothelioma in asbestos workers4 and hepatic angiosarcoma in vtnvl chloride work ers.5 Dr. Bader a correct in noting that no cases of leukemia devel oped m our senes among workers first exposed to beiuene after 1954. The most likelv explanation for this finding is a reduction in cumulative exposure; relatively 4c* workers were hired after 1954. and among those workers, iota) lifetime exposure was kept low b* decreases m production and progressive diminution in exposure standards. Austin et al. are concerned bv what the* perceive as 'discrepan cies" between i)jf ns* tsun-rM denved 1mm our case-com-r! and standardis'd monal't* ratm analyses Given the different nature of these two analyses, surn cnr.crm is misplaced The standardized mortality ratio study is intended simpl* to delineate overall patterns of mortality. It provides a poor basts for ascertainment of an expo sure-response relation, because each dose category within the co hort is a distinct subset with a different duration of follow-up, a different proportion of deaths, and diffrrrni age distributions. In uddi'-or.. the standardized mortality ratio uses an external popula tion for comparison -- namely. the entirt population of ihe U nited States. B\ contrast, the case--control study provide* an elegant means for modeling cumulative dose w hile controlling for age. dura tion of follow-up. and other potential eonfounders and effect modifi ers. Furthermore, it employs an internal comparison group. .All that said, however, the apparent discrepancy noted by Austin et al- is largely an artifact of their manner of representing our data. They arbitrarily chose to derive risk estimates by using the midpoint of each of our exposure categories. A more appropriate procedure, given the stated highly skewed distribution of (he cumulative expo sures. would ha'C been io use the person-vran weighted average exposure for nth category. With tha* approach more highly con cordant risk estimates wuukd have been ootuinrd. Austin et al. are also concerned bv a "disparity** between the mean level of exposure to benzene of ait cohort members with out disease (67 ppm-years) and that of the 90 controls (50 ppmyean). This apparent discrepancy is easily explained. The appro priate procedure for tabulating dose* in a case-control incidence study requires that cumulation of the controls' exposure be trun cated at the time of death of each corresponding case* For all other members of the cohort, dose continues to accumulate until cessation of exposure. Austin et al. hate overlooked this methodologic point, and thus the two average doses cited by them are not comparable. \\r are reassured that a separate risk assessment undertaken recently by Austin ei al produced overall findings quite similar to ours.T Both studies demonstrate an elevated risk of leukemia with cumulative exposure to benzene, both derive similar dose-response curves, and both indicate that there ii t substantial risk of leukemia at exposure levels below 10 ppm. Mr. O'Keefe naira that during public hearings on benzene held bv the U-S. Department of Labor m 1966. "major shonromings" were uncovered in our research. O'Keefe fails to note that with our cooperation, he and In* staff scrutinized rarh datum in our analysis. They found unly minor discrepancies among thousands of coded work hbionex and apoaurr measurements. We subsequent!* cortoned each of these errors, and the corrections are included in our published report Correction of those enun actually increased our OUmates of leukemia nsk slightly Furthermore, the environmental measurrmerm were reviewed bv the company in whose plants the stud* was undertaken. Thr company took no exception to our char acterization of exposures, and they rejected the notion that we had substantially underestimated raposure.' We are plraaed to note that mi September 1. 1987. OSHA an nounced a reduction in the standard for permissible occupational exposure to benzene from 10 to I ppm. Cincinnati. OH 45326 Rotuit A. Rimsky. Alijlandi* B Suit*, Rjcmvxq Horvcnc. Andrla H. Oxts. amp Puttie J Lanoiicav National Institute for Occupational Safety and Health 1. SayOrr CA. Gottfitetn ID ScUzkwmar aR. Brombcry I. Lnim S. Albert PE The lahaiaaon soSKstof* of benzene incidence of hemstoomeoc oro- pixsms and betnaneucm AXR/} aad C37BL6J owce Toxicol Appl niamucoi 1990. 34 323-31. 2. MaitomC. Conu B. Cora G Bcazeoe a muhtpatemiaJ earatnofM te*ulu uf kMf-iemi tnoaxuvs perforated m Ok Botepas littuiuie of Oncology Aa J lad Mod 1913. a 5*9-630 3 NT? tactvncal report on me totzeofctfy aad evciaofeaesii saidies of brue-e in F?as N rat* sod B6C3F1 okt Kudin) Research Tnsnjjt Para. N C National Tessoiop 1996 (Publicaaon no *6-2345 ) 4 Seltkoff U. Hammond EC- Cherr J. Cascinufciuctfy of ammtie ashesiot Arch Esnne Hcaltt 1972. 23 1*3-6. 3. Creech fL h. iohnaoB hO Angiosarcoma of liver is the naaufacrure of pel>vay| chloride J Occur Med 197a, I6 15CM 6 Bm(o NE. Uibm JH. Mares P. LanghoU B Multiplicative aodels and ennort analmi 1 Am Sui Asm 1913. 78.J-I2 7. Ausun H. OcLzel* E. w'e.e P Brume and leukemia, a mie* of dw Lieranirc and a nik xivcnrwr- Am J L>i.Tuoi id 7. mi I Letter (nore Goodyear Tire me Rubber 10 MOSH. October 5. 1981 Ir Occupational Safer* and Hears Admuuscratiae Betuenc Docket H-059 Watfunfun. D C . Oervpnooca' Safety and Hesiut Adreuuioxuoe. Tecftr.. cal Dsu Center. 1986 To tht di/er Despite asseroorti to the contrary bv Mr. O'Keefe federal (regulatory! agencies have not adopted preventive hea;;h policies OSHA waited 15 'ears (unul 1986; to regulate aibesros as a carcinogen -- and then suspended the new standard for non asbestiform fibers This coJi hardlv be called a pmenme ap proach. since thousands of asbestos workers died because of the delay. OSHA also failed to establish a short-term exposure level. Elhvlenc oxide was only ftceniiv regulated because OSHA waj sued. OSH.A has still refused to define a short-term exposure level for ethylene oxide that would be required for adequate protection. What M-. O'Keefe woold probably call an ill-eoecetved rush to judgment is what thuse concerned with public health - ou! ! recc . mend as a prudent measure Sufficient evidence tojusufi a siandarj is. like beauty, in the eye of the beholder. Decisions about whether to regulate can be in error because of unerruimie* about the nanarr and extent of the risk or about the economic and technologiol feasibility of regulatory controls One type of error ia committed when, because of insufficient evidence, aocietv fails to regulate an activity that turns out to be harmful. Another type of error is committed when society regulate* an acm it> that turns out not 10 be harmful and resource* are neediest)* expended. Aversion to making these two types of error* reflects differing values regarding the nature of the mistakes made and the extent, prevalence, or magnitude of the mistakes. In my opinion, reliance on scientific conventions of causalir* for public health purposes is simply too cosily in terms of human life. To dignify a deciaion not to regulate or# decision to w-xji for more evidence, under the mantle rd its being scientific, is io do a duserv ice to both science and public policy. Furthermore, u obcurrs the vatur judgments made With regard to the regulation of chemicals over the past 17 years, I cannoi think of a single regulated substance for which the evi dence has not grown mure convincing with time or for which a reversal in regulatory policy occurred. Sooner or later, the tip of the SAL 000001671 1030 THE NEW ENGLAND JOURNAL OF MEDICINE Oct. IS, 1967 iceberg mail thr iceberg. We vr very for from B conceived rushes in judgment. On the contrary, the regulatory agencies have tomefimn hern dragged kicking and screaming into action necesarv for fulfilling ihnr statotory mandats. In the meanwhile, work* en and citizen* are harmed. An example it the solution of science regarding the carcinoge nicity of formaldehyde. The National Cancer Institute (NCI) studv in 1986 asserted that there waa 'little evidence that mortality from cancer h aaaociaicd with formaldehyde exposure at levels experi enced In workers in this study.* In a later publication1 the same authors concluded that Despite small numbers, the dose-dependent association of naso pharyngeal oncer with ctpmurt to formaldehyde and particu lates deserves further unTsttganon through ease-control studies, where the influence of formaldehyde and particulates mav be evaluated with more statistical power than by standard cohort Studies- Further analysis of the NCI data by Sterling and Weinkam7 reveals a drartv roerraaed risk of lung cancer. At congressional heannes m 1986. the Director of the NCI. in commenting on the earlier work, tatified that "in association with particulates, there is in association hrrwren formaldehyde and nasopharyngai cancer, which dnrh has to be followed up."1 The irons concerning the eriiihsms of the adequacy of the sci ence is that the scientific evidence is euenualiy there. OSHA has just decided to reduce the permissible exposure level for benzene 10 I ppm. It waned Car too long. Let there be a better gmemmental response to the orw realities. Cambridge. MA 03139 Nicholas A. AsHronn Massachusetts institute of Technology 1 Blair A. Slew Pa. Hoover RN. Pramxi JF Jr. ei al Caacm of the wopnarvru mt orapOaryw and formaimtivde ciposwe. JNCl 198?. 71 191-2. 2. Sterling ID, Wemkaa U. Submitted for dke tecort si (he OSKa Hnnfift on FwmaJdehvde. Submission lt2B. Nmember 1986 3 Formaldehyde Sody Hcanog before the Subeommittr oa Ovmight and mupnonsof thr Committee on Energy and Commerce. House of RepitseauDva. )uh 3. 1986 106 THE EC-1C BYPASS STUDY T Ou Editor: The recent discussion (March 26 issue)1'4 *o*f the extracrznial-iniraeraRial (EC--1C) arsenal bvpass nudv is very timely. I would like to sprai io one statement in the editorial: "The conclusions of the EC-IC trial are valid for the population of pabents studied.' * The population that underwent surgery included patients who had only one eptsede of the transient ischemic attack. Moat neuro surgeons w ho are experienced with the bypass procedure would not accept for surgical consideration any patient with a single transient ischemic attack. Three is no discunion in the original report of the randomized study3 about bow many of the parents who underwent surgery were in this ordinarily nonopevadve group. Subsequent re ports by the authors of the randomized study have never clarified this issue. Regardless of wrheiber the randomization of the patients was proper or not. the mere fact that patients with one transient ische mic attack were included in thr studv legates the concept that the conclusions of the EC-IC trial are valid for the population of the patients studied. Lot Angeles. CA 90048 Mu.tow D. Htimr. M D. Cedarx-Sinai Medical Center 1. Reiman AS The esscmwtil j--rsmiusl wscnal hypes* study; SOua bare wt Iramed* N Eegl J Med 198?. 316 109-10 2. SunJi TM Jr. the lainnsiiona) randomued oul of csmmeial-iMra' cntual wwnal Mpms it( siire of thr population mi? N EafI i Mctf 1987. 3t6 IM-6. ). Gottis^ S. Zcrras N. Lanffat T. The Eiswntiil-lmnwl Bypass Studv. a tepret i4 tt* committee gpusS H the Amrnrm Asaenum of Wcutotopcal Srepcrew eistmnr studv. N EnglJ Med 1917. 316117- 20 BemmHfM. SKfccttO.TaylBr9W.aal Are dw remits of ttzexnoasek-mowramai bypass mal %------ `irsTV N Engl J Mad 1917. 316. ISM 3. The EC/1C Bypms Study Ctm* Pasture at eioaciaiuaJ-imcmnial ancnaJ bypass w reduce die nak of sOtme sookc resutts of ei unemaitonaJ raadonuud anal N Engl J Med 1913 313.1191-200 7e tin Editor: The Special Rrpoeu in the Jmmal of March 26 on the validity of the intevnsitonal randomized trial of eairacramalintracranial artenal bypass, first published in theorem*/ m 1965.* induce some thoughts about the findings Specifically, the implica tion of thr findings for a better understanding of the pathogenesis of ischemic stroke is of interest. The results are fully consistent with the hypothesis that ischemic strokes are largely embolic in origin. The relations among transient ischemic attacks, completed strokes, and atherosclerotic plaques of thr internal carotid antry were well established before the study was done. Emboli from ulcerated plaques or occlusive lesions of the ioiemat carotid artery or middle cerebral artery would presumably lodge in a beam vessel distal to thr anastomosis, whose only effect could then be to improve marginal or collateral circulation to the periphery of the temtorv where the infarct occurred. Is it too simplistic io assume that in the case of the carotid lesions, a natural bypass alreadv exists -- the arcle of Willis? One of the postulates of the studv was that lesions of the middle cerebral arterv would be espeoaily suitable for bypass surgerv. In fort, the existence of such lesions genetallv indicates more severe, widespread atherosclerotic disease, with the attendant nsk of stroke from embolism originating in the internal carotid artery or aorta or death from coronary hran disease. Montreal. PQ H3A 2B4. Canada Sean Moo*e. M D. McGill University 'The EG1C Bypass Studv Croup Failure of nncfinial-iimmmil aneniJ bypuss io reduce the nsk of ischemic moke results of an unemaaonai randomiced trial N Engi i Med 1983. 313 1191-200 To ttu Editor- The Journal is to be congratulated for its com plete handling of thr controversy surrounding the EC-IC oper ation.1'4 However, we strongh disagree with the conclusions of both the neurosurgeons1 and the Editor.1 who seems to have been impressed bv the neurosurgeons' arguments and to cast a shadow on the proper interpretation of such vials. We agree that generelizability is strengthened when most eligible patients are ran domized. but we insist that the results should stand unless or until an even better study repudiates them. In the meantime, those who insisi on using this oowndiscredited operation should not expect third-party reimbursement until ihev provide credible (i.e.. well-controlled experimental) evidence that contradicts the EC--1C study. The called-for retrospective review of the outcome in the nonran domized patients wrould be as worthless as any uncontrolled histori cal survrv. It would be impossible to sort out sclecbon factors from treatment effects. Some patients selected for operation rather than randomization might do worse than those randomized, because the more threatened patients were created surgically, and some might live longer because their gond condition made them attractive surgi cal cand'dates. Whichever predominated, the mult would be uninterprrtable, as demonstrated by srvml published comparisons of survival of randomized and nonrandomizrd patterns who teemed to have thr same prognosis.*'* If surgeons have reservations about the validity of randomized controlled trials because their patients were not entered, thr problem is easily solved by including al) eligible patients in future studies that have the prior approval of those surgeons. In our experience, patients readily accept randomization when their physicians are convinced of its validity. When told of the mulls of the trial and the lark of anv evidence favoring surgery, they would prefer randomization tn an operation baaed on surgical judgment alone. As suggested hv one of us.* the study could be paid fiir with the money saved bv operating on only half as many pa tients. In the mrantime. there ia no acceptable evidence that the SAL 000001672 J Cancer Res Clin Oncol (WS'UI J.3J9-55S Cancer Zesearch Clinical C neology 2 Spr-ngcr-Wrlag 198? Possible implications from results of animal studies in human risk estimations for benzene: nonlinear dose-response relationship due to saturation of metabolism . - --------'--l.OGY- Sandro Grilli 'Werner K. Lutzand Silvio ParodiJ* Centro Imemmuerotine per la Ricerca sul Cancro and Ittitmo di Canoerotogu deirUmvcmta di Bologna. *A CCC v >iilt Filopanli 22, I-*0126 Bologna. Italy 2 Institute of Toxicology. ETH and University of Zurich. CH-K603 Schwereenbaeh, Switzerland * Cceiro Intentoisenitario per la Ricerca sul Cancro and Isutuio Naaonaic per la Ricerca tul Cancro. vialc Benedetto XV. ]0. 1-16122 Genova. Italy ;$cS Summary. To date, all risk assessment studies on ben* 2ene have been based almost exclusively on epidemiological data. We have attempted a more integrated and quantitative evaluation of carcinogenic risk for hu mans, trying to utilize, in addition to the epiaemiological data, all data available, specifically data on me tabolism, genotoxicity, and carcinogenicity in small rodents. An integrated evaluation of the giobaiity of the a'aiLble deta seems to suggest a progressive satu ration of metabolic capacity both for man and rodents between 10 and 100 ppm. The most susceptible target cells seem to be different in humans (predominant in duction of myelogenous leukemia) and small rodents (induction of a wide variety of tumors). Nevertheless, both epidemiological and experimental carcinogenic ity data tend to indicate a flattening of the response for the highest dosages, again suggesting a general satura tion of mechanisms of metabolic activation, extended to different target tissues. From a quantitative point of view, the data suggest a carcinogenic potency at 10 ppm two to three times higher than that computable by a linear extrapolation from data in the 100 ppm range. These observations are in accord with the re cent proposal of the European Economic Community of reducing benzene time-weighted average occupa tional levels from 10 to 5 ppm. Key words: Benzene - Risk estimation - Carcinogenic ity - Genotoxicity - Metabolism saturation - Dose-re sponse relationship Introduction Benzene is probably the chemical with the largest pro duction and utilization in the world (IARC 1982s). As a consequence, both occupational exposure and en * Contributions to this work by Dr. Annamana C>>lacuiMano Manullo *, and Maunno Tanmghcr2 are gratefully acknowledged Offprint requests i<> S. Grilli vironmental contamination occur to a rather signifi cant extent. Benzene has been recognized for many years as a carcinogen in humans wuh the .r.duction of acute myelogenous leukemia (IARC 1982a). The most common regulatory standard for occupational exposure in industrialized countries is a limit of 10 ppm (time-weighted average JWA) (IARC 1932a), For the corresponding risk assessment often only epi dermic logical data have been considered (IARC 1932 a). Correlations with carc.nogcnjcit> data in ro dents have less frequentlv been established (Lee et al. 1983). In this paper, we have tried to enrich current in formation by compiling the data related to genotoxicirv versus noneenotoxiciry in the mechanism of action of benzene, by comparing metabolism in humans and rodents, and by pharmacokinetic considerations of a possible saturation of benzene metabolism and geno toxicity at higher exposure levels. Epidemiological data In the annex to the LARC Monographs on the Evaluation of the Car cinogenic Ruk of Chemicals tc Humans. vol29 (IaRC :b). a working group attempted a qualitative mk estimation for benzene, based entirely on human epidemiological data. Minimise ascites of 1^-1 *9 excess leukemia deaths per 1000 workers exposed ever a working lifetime were eaJcuiatad for as exposure level of 100 ppm. Their computations were based oo three epidemiologic*] studies (Vigliam 1976: Ott et al. 1971; Rinsky et al. 1931) for which ranges of excess cases and exposure levels are shown m Fif. i. Usisg the words of the authors (IARC 1982 b): "The working group found that quan titative estimation was more feasible than teemed irutiallv to oe the case. The group was impressed that rather large amounts of quanti tative information could be extracted from ihe published epidemio logical data and that rcasooabie mk estimations couid then be based on that information.'* The IARC estimate is in rather good agreement with (he esti mate of Goldstar! tl983). The lifetime risk per parts per million com puted by this author was apparently about 10 times higher than ihe lifetime nsk computed by (he (ARC Group. The difference is only apparent and due to the fact that in the Goldstein computations a continuous life exposure was considered, whereas m the compuia- SAL 000001673 ptt 100U hy<nnk ! M 350 S. Gnili ct al.: Risk estimations for Prnzs.ie Fig. I. Lsukrarja incidence in workers exposed to benzene iccmpuu- tiooj o> IARC Working Group. 1ARC I982a>. Range* of bet- zcnr coacssuatioa include varuuons m exposure leveis end uncer tainty in u-*\tie*i determinations performed by "old" methods which did a precision of - 50% (Rituxy et ai. I9S1) lions of :ie - ARC Group the exposure *u referred to JO h week for 45 years. the computations of Goldstein are corrected accord ing to tie standard of exposure uuitzsd by the [ARC Croup, the te ens cun of leukemia deaths expeetee for 100 ppre become 332 1000 workers, m good agreement with the \alue deduced from rig. 1. In the Goldstein computations, two out of three eptdtraiolopcaJ stud ies were the same as those utilized b> tie lARC Group. results that we have considered relevant is presented in Table ). Differences in strain susceptibility and in route of administration can be observed: Sprague* Daw-ley rats are less susceptible than Fischer 3J4 rats to benzene administered by gavage. A quantitative comparison of the carcinogenicity results from oral with inhalation exposure could not be made becauser; we could not find enough data on serum levels of hen- * zene at different concentrations and times after beg:n- ning of treatment. If we confine ourselves to the high dose inhalation experiments of Maltoni et al. (1985). we can compute a tumor incidence over controls (animals with a: least one tumor) of 0.27 for an average continuous concen tration (on the basis of 8 h/day, 5 days a week) of 243 ppm for 2 years (a difTerem value (TWA - 2S4 ppm) is reported in the legend of Table 1, but this value refers to a daily exposure of 4h for 7 weeks plus 7h for 97 weeks), equivalent approximately to an incidence over controls of 111,1000 animals for 100 ppm for rat lifespan. If we consider the inhala tion experiment with mice of Snyder et al. (1980). we can compute a tumor incidence over controls of 0.15 for an average continuous concentration of 225 ppm for 2 years, equivalent approximately to an incidence over controls of 67/1000 animals for 100 ppm for mouse lifespan. Consideringcarcinogerv.c po tencies can span a range of more than 107 times (Parodi et al. 1982; Peto et al. 1984), a difference by a factor Carcinogenicity in animals To our knowledge, the first suggesiion that benzene could be carcinogenic in small rodents came from work published in 1979 by Maltoni and Scamato. The induction of Zymbal gland carcinoma in female rats was statistically significant (P -- 0.0031) at 250 mg/kg (administered by gavage once daily. 4-5 days each week, for 52 weeks). A suggestion that benzene could be carcinogenic in small rodents upon inhalation came from a work pub lished in 1980 by Snyder et al. Benzene was given to C57BL/6J mice at a concentration of 300 ppm, 6 h/ day, 5 days/week, for life. Eight hematopoietic tumors were found in 40 treated animals, whereas only 2 were found in 40 controls. The difference was suggestive but only borderline from a statistical point of view (P-0.044, one-tailed, according to the Fisher exact test). After the publication in 1982 of the work of the IARC group on quantitative cancer risk estimation based only on epidemiological data (IARC 1982b). the carcinogenicity of benzene in small rodents was better defined und confirmed (NTP 1984; Cronkite et al. 1985; Maltoni ct al. 1985). An overall picture of the Fig.2.Oncogenic potency index (OPIJ is x fuoctioo of dosage of ben zene in mice end rats. OPl- -ln(l /)'/. where l is the excess inci dence over controls (proportion of animals with at lease one malig nant tumor for all tumors). / is tne length ofob*erviiou (;:me unu 2 yean) and D is the dosage in millimoles per kilogram per day equivalent to the total dose divided by a 2-year exposure (Parcdi e: al. 1912). Only dosages which produced sutisucall) significant ef fects (P<0.05) were considered. These data are not directly ccmparabie with those of Table t where malignant tumor incidences of dif ferent target organs are reported. A F>44 rats. gavage (NTP !$$Ji. B-B6C3FI mice, gavage (NTP I9S4); CS-D rats. fiige |Maltona et al. 1985) SAL 000001674 S. Gnlli ei al.: Risk estimations for benzene 35! Table I. Malignant tumor* obsencd m long-term assays of benzene carcinogenicity in rodents Species Route of administration Inhalation Gavage Mam target organs Tumor incidence over controls TWA chronic dose (ppm) Response Strain Main target organs* Tumor incidence over controls Chronic Redose sponse (mg.kg per day) Strain Rat Zymbal glands 0 093 :4 (+*-) S-O* Oral cavity 0 079* :s4 -) S-D* Liver (hepatomas) 01084 :s4 S-D* Breast 0084* 284 (4-) S-D* Nasal -avuy oo:i 284 ( + ) S-D* Zymbal glands Zymbal giandi Oral cavity Oral cavity Liver (angiosarcoma) Skin Skin Nasal cavity Forestomach (in situ ca.) 022 0.42 0.19 0.52 0 063 0.23* 0.16* oos 0 15* too (4-4-) F344* 500 (4--) S-D* 50 (4-4-) FJ44* 500 (4*-) S-D* 500 <-) S-D* 500 <-> 200 <-> 500 (*) 500 (-> S-D* F`J4 S-D* S-D* Mouse Hematopoietic 0 15' 300 (i) system' Hematopoietic 0 145* 300 (~) system' Oiry C 09* 300 (4-) Zymbal g!ods or* 300 (*> C57BL.6* Zymbal glands 0.24 100 (4-*) B6C3F1* C57BL.6* C57BL.6* C573L t>` Liver (adenomas * carcinomas) Hematopoietic stttem' Breast Lung (adenomas) 0.14* 0.16 o :ts 022 100 25 5'X) 500 (4-*) <+) (-) (-) B6C3FJ ` 86C3FI* S* u* Swiss* ' Tumor incidence m treated group* was higher than that of controls from a statistical point of view i?<0 Ofi. Oni> the most effective dosage was reported * Malloni el al. <1985). In tnhaiation studies. S-D rats were exposed to benzene pre- and postnaully {i~" q day. 5 da>s week for '/* weeks (TW42SJ pprri or 15 weeks TWA *200 ppm)). Experiments with Wutar rats and Swiss mice are nagesig ('OO-week nud> i Ttr'scr inctcences from this work have been calculated bv referring the aumber of vumcr-peancg animals to the correct r.-mbe? ft.e.. tie rubber of survivors at the time that the Hrst tumor appeared, regardless of the organ considered! * NT? study (1984). Malignant tumors of lung, breast, and preputial glasd* were also induced 4 Only in females * Only in males r Malignant lymphomas * Snyder ei al. (1980). Mice wrere exposed for 6h day. 5 day;*eek for iife. A single cose was tested * Cronkite et al. rj9Si), Subchron;c exposure (16 weeks) and 99-week observance. ^ single dose was tested TWA time-weighted average. F)44~Fischer 3*u rats. S-D*Sprague-Daw;-y -au (-- -- ) dose*related rumor inductioo. t -- .doseunrelated tumor inducuon. (-) borderline evidence of tumor induction '0 I > ?>0.05) of 2 to 5 between steal! rodents and humans is remark* ably small. Obviously, in making this statement we imply that the ratio of tumor latency to lifespan is es sentially the same for both humans and small rodents. Most experimental evidence seems to favor this broad assumption (behind this apparently simple hypothesis very complex problems are hidden when considering carcinogenesis as a multistep process. Mutation-like events seem to have a similar frequency in human and rodent cells, but perhaps the number of steps is differ ent in the two species, in order to adjust latency to the respective lifespan). The dosage extrapolations at low levels of risk, performed by the USA Environmental Protection Agency (Lee et al. 1983; Goldstein 1985), showed a similar picture of concordance between hu man and animal data. From the point of view of overall tumor frequency small rodents and humans showed very similar sensi tivity. Considering target organ, only few leukemias (usually not of the myeloid type) were induced in ro dents. Zymbal glands and oral cavity were the primary targets in rodents, but ned^lastic effects were also ob served in liver and breast, and occasionally in skin, nasal cavity, lung, forestomach, adrenals, ovary. Harderian and preputial glands (Table 1). We are aware that the application of animal data s as proposed here) to humans would be safer if j*.ot only a quantitative but also a qualitative similarity of tumor incidences had been shown in the two systems. Some indication of a possible dose-response rela tionship is offered by the experiments when benzene was given by gavage. In Fig. 2 the oncogenic potency SAL 000001675 35: S Grilli cl al Rk estimation* fnr hen?c Table 2. Short-term of potential carcinogenicity of benzene: <' a.'uuiion of most gvnoirmuity duiu from lit.raiure. Figures mdiojie :he number ji' horwerm assay* which were cither pontile ( +) or negative ( - ) Unless ocherw ise indicated, data are taken from (ARC (1912a. ci. ECETOC 11914) and Dean 11935j reviews. Data lacking in experimental details, equivocal for the .mthors or for IARC working group, or successively changed b> the ame authors after having modified experimental design, have not been considered Er.dpoist System Prokarsotes Fungi greet plants Insect* Araphibtas* Maaur.a-iaa rails (m w.ro DNA damage <-) 1 -) 0 poi A .. 1` MtMBl 0 1* MMBI 0 1* a A* : uds 4* Mittaxs i in vivo) j* 1' 0 Hucanj :a m*o) Class pOSilivtty 8,17.47.1% Overall activity (positive tests total tests): Mutation Chromosome anomalies (-) (-) (+> ( -) 0 ST 1* 0 ST. BS 7 0 ST 5* 0 SC 1 3* GM.GC.CO 16* % TRD 0 3* A 1* 0 RM 1 1* RM 2* 1* CA 0 0 L5l?8V 1 S" L5l*8Y.VT9. 9* CHO.KL 0 1* 1 A0 t* CA 0 7 CA 1 3* CA 4* 2* A/P 3* 2 SCE 1 0 SCE 6* 0 MN 1" 13 CA 1 r CA 0 2 SCE 0 It MN 0 1' MN 0 19 CA 2 4* CA 0 0 SCE 2 14 58 -24.1% 71,93 76.3* 96180- 53.3% Other 0 MC 3* 3* TR 5* 0 DL ] 3 1 2 - 23 0*'. * F;cm Let e; al. (19J2> * From Table 1 of 1CPS collaborative study of Ashby et al. (1985) * From Kalf et al. (1985) - From Arfellmi et al. (1985) * From Parodi et al. (1983) alkaline elution 1 From Parodi et al. (1983) viscostmetnc unwinding 1 From Lutz sod Schlatter (1977b) and Arfellmi et al. (1985) DNA adducts * Point mutation in rat chromosomes 1 From Cad-cl'Kanm et al. (1984) 1 From Cboy et al. (1985) , CHO- Chinese hamster ovary. ST -Salmonella lyphimarumiimiemomt. BS-Bacillus subtilis. SC- Saccharomyxes ctmuia*. CM gene mutation. GC-ftae convmion. CO - crossing-over. A-tneuploidy RM recombination and mutation. CA- chromosome aberration. MtMBl - raitoehondria-mediated binding to mitochondrial DNA. MMBI microsome-mediated binding to calf thymus DNA. AE- alkaline elution. UDS * unscheduled DNA synthesis. HL-humaa lymphocytes. P polyploidy. SCE - sister chromatid exchange. MN miBonucieus mi. MC - metabolic cooperation. TR - eel! transformation. DL - dominant lethal test (rat), pal A* DNA polymcrasedeficient sod proficient Escherichia coii. TR.D - Trudescantii assay index (OPl)computed according to Parodi et al. (1982) for increasing benzene dosages is reported, and sug gests a higher relative sensitivity to the carcinogenic effect of benzene at lower dosage. Carcinogenicity data in rodents for vinyl chloride given by inhalation also showed typical flattening of the response at the highest dosages (Malioni 1975). However, in the case of benzene, dose-response experiments with in halation exposure are not available. A quantitame correlation between human and animal data was reported by Rall(1977)forchlomapha2ine. benzidine, diethylstiibestrol. and tobacco smoke. 000001676 SAL S. Gnlli ct al.: Risk estimations for benzene Short-term tests The qualitative results obtained with short-term as says are illustrated in Tabic 2. As is typical of a com pound examined in many different short-term tests and in many laboratories, there are a mixture of positi\e and negative results. In order to express a global opinion on the data, we have followed the approach suggested by Ashby et al. (1982). These authors have examined the results obtained with short-term tests, performed in 66 different laboratories, on 10 different chemicals. Clearly genotoxic and carcinogenic com pounds like acrylonitrile and tcluidine were positive in 40%-6G% of the tests. Presumed noncarcinogens and/or nongenotoxic compounds like benzoin and ca prolactam were positive in about 20% of the tests. The authors suggested that potent carcinogens with a genotoxic mechanism would be positive in --80% of the tests. Noncarrinogens could be positive in -20% of the cases. Weak carcinogens should produce an in termediate number of positive responses. We examined all the data that could be collected on benzene without any selection or exclusion of data. Bcnzcre appeared positive in 53% of the cases 353 (Table 2). In the perspective of the above consider ations, the results are definitely more typical of a geno toxic carcinogen than of a noncarcinogen or nongeno toxic agent. However, looking at the results in more detail, the percent of positivity of DMA damage data was 47%, whereas for point mutations and chromo somal anomalies the values were 24% and 76%. re spectively (Table 2). It seems reasonable that there might be something peculiar in the DNA damage in duced by benzene, in the sense that the damage seemed clearly mere dastogenic in vivo and in vitro than mu tagenic. We have examined the publications available on chromosomal damage induced by benzene in ro dents in terms of dose-response relationship after treatment in vivo (Erexson et al. 1984; Gad-el-Karim et a!. 1984; Styles and Richardson 1984; Choy et al. 1985). The results obtained are shown in Table 3. The potency of the response (effect,dosage) clearly de creased from low to medium to high dosage. In the ex periments where benzene was given by inhalation (Erexson et al. 1984, Styles and Richardson 1984) the net effect was only doubled when the exposure in- Table 3. Dose-response relationship Tor the clastotemc activity of benzene io rodents Species Route of admiiustralion Endpoint and dose unit Effect* dosage ratio at different dosage -- Lo* Medium Hied Male Swiss mice Male B6C3F1 mice Male Wistar rat* Male DBA/2 mice Oral* Oral* Oral* Inhalation' Inhalation* Inhalation* MicronucJei. 1000 PCE mg,kg Chromosome aberrations per metaphasa. mg/kg Micrcnuclet 1000 NCE mg kg % cells with chromosome ncmazrs. ppm SCE metaphase ppm Micronuciei 1000 PCE ppm 006-003 ---- ------------------ a 0 '>/- f 3* ----------- 9*s: 25' 3.03-1 5* 10* 7,6-39 10' o.9-o.:i 00*4 10' 6-2.* 38 0 16-003 88 0 iU 0 001 s 5.3-1: too 4*3-1,54 100 9 3-3 9 100 3.03-0.31 100 1221 0036 0018 ;s 880 1.1-003 !?0 03 o > i: 9 3 -:: 600 7,81-1.54 100U 13.S-5 9 1000 3.81 -0.3! 1000 Q-0C~9 0 0036 * Treated values minus control values * Two doss ware admmmercd 24 h apart and animals were sacrificed 30 h after the first dose (Gad^l-Kanm et al. 1984) * Corresponding to 3.76ppr according to EPA computations (Lee et al. 1983), assuming a respiratory absorption coefficient of 0.5 and a ventilation capacity of 0.03 l/tnm for mouse (Gold et at. 198*) 4 Subchrooic treatment within NTP study (198*). Micronuclet were measured after exposure for S days week for 4 months <Cho> et al. 19SJ) * Corresponding to 10.7ppm according to EPA computations (Lee et ai. 1983). assuming a respiratory absorption coefficient of 0 5 and a ventilation capacity of 0.03 l/min for mouse (Gold et ai. 1984) 1 Exposure of 6 h (Styles and Richardson 1984) * Corresponding to 7.1 mgkg according to EPA computations (Lee et al. I9K3). assuming a respirator) absorption coefficient of 0 5 and a ventiiatioa capacity of 011,'mm for rat (Gold et al. 1984) * Exposure of 6 b (Erexson et al. 1984) 1 Corresponding to 23.4mg/kg according to EPA computations (Lee et al. 1983). assuming 4 respiratory absorpiton coefficient of 0 5 and a ventiiatioa capacity of 0.031/min for mouse (Gold et jI. 1984) PCE polychromatic erythrocytes. NCE w nomtochronuiiu erythrocytes SCE - sister chromatid exchanges SAL 000001677 354 S. Cnlliei at.. Risk eumations Tor benr.e Table 4. Rjr.k daksiftuaurtt uf potency of oncogenicity (OPI) and of DNA damaging activity (CBI) for 36 carcinogen* tested in rat and mouse' and for benzene Rank Loglft OPI Log(0 CBI Rank Log,e OPI Lo;,0 CBI 1 -142 Ami --0.35 Eth 19 2.66 CP -0.4e AA8 3 --0 ~a CCt* 4 -0 io MMS -0.30 DES 0.21 2-NA 0.38 AAB 20 2.72 MDAB 21 173 ENU 22 2.96 BP J 0.00 CHC1, 0.45 MCA 6 0 ?4 ;.nA ` 0.57 Anil 7 \.Zb SPY 0.83 DAB 8 i.40 Eta 0.S6 CHClf 9 1.5* EDB to i.6i Dab 11 1.63 DMH t: 1.79 NPt 1.18 BP 1.33 DM9A 1.41 uC'l, 1.61 EC 13 : i: nm 14 2.14 AAT l.W NM 1.79 CP If 2.23 MNNG 1 "9 EMS ;6 2.33 MCA * 2.35 afb: :s 2 50 EMS 1.82 MDAB 1.92 NMua: 2.03 .AAF 23 2.97 DEN 24 3.9! AAF 23 3.05 HAAF 26 3.06 EC 27 3.07 Benzid 28 3.13 DMN 29 3.43 MNU 30 3.52 DES 31 3 60 NMusi 32 3.73 NHMI 33 3.73 MAMA 34 3.95 AFGI 35 4 36 DM BA 36 4.41 AFBI B&zcze OPI values Q.S2V 0.13*. 0.3SV 0.036*. 0.0J6f. Ugw OPI - -0086*. -0.638*. -0.056*. --1.09*. --1.25* cat values 7. i."\ if; Leg,* cat - 0.845*. 0.23*. 1.176* 2.06 NHMI 2.0? NPI 2.1* DEN 2.16 AAT 2.25 NPY 2.26 Benod 2.26 EDB 2.29 NU 2.59 HAAF 2.60 MMS 2.7] MNU 2.75 afb: 2.33 AFGI 3.00 MNNG 3.40 DMH 3 47 DMN 3.64 MAMA 4.09 AFBI * Data from Parodi et al. (1951) * F-3--4 ri>. gavjge (NTP study :9S-t) * S-D m. gavage (Maltor.i s: al :4aft * B6C'r I rues, savage (NT? study i9jj) 4 Swim sics, gavage (Malton et al. 19*5) ' Wiiur rati, gavage (Maltou ei il 1985) w sin rats. i.p. (ArfcUim et a!. :aB5) * SlV-50 rats, inhalation (L.:* ar.c Scr-ar.er !9*"b) Exposure in a closed inhalation chamber: the amount of benzene administered *is kr.ou r. ' BALB.c mice. i.p. (Arfc'.Um et al. 1985] Arrows indicate the rank positions of log, OPI and of log,,, CBI of benzene Anil-aniline. AAB-4-amjnoazobenzeae. CC1*--carbon tetrachloride. MMS-methyl methaoe sulfonate. CHCt* -- cUoroform. 2-NA:-eaphihylamice NPY - N-nitrosopyrrolidine. Eth-ethiomne. EDB-ethylene dibromkJe. DABsN.N-dimethyl^aisinoazobenzene DMH - l^-dimethylhydr3Zir.e NP( - N-rjtro*cpperidine. NM - N-nitrosomorpholine. AAT- o-aminocrotolucnc. MNNG - N-methy!-N nitro-S*nitrosoguamdme. MCA 3-meihyIchotanthrene. AFB2-afIatoxin 8*. EMS-eihyl methane sulfonate. CP- cycJophosphamiCe. MDAB-3'-meihyi-t-dimethvUmtnoazobenzeac. ENU-N-nitrosoeihylurea. BP benzo<a)pyrene. DEN-diethylnitrosamine. AAF2acetylaminofluorme. HAAF-N-hydroxy*2-acetylammofluorene. EC--ethyl carbamate. Benzid -- bensdioe. DMN--dimethylnurosomine. MNU - N-muosomethylurea DES -- diethylstUbcstrol. NMust - nitrogen mustard. NHMI - nitrosohexamethyleneiminc. MAMA - meihvlazoaymethanol acetate. AFCl -aflatoxtn O, DMBA --7.12-dimethylbcn: >a)anthrac*oe. AFBt -- aHaiosin B, creased 10-fold from 10 to 100 ppm. In their work on terms of potency coefficient, OPI, which is, as a first micronuclei induction in polychromatic erythrocytes approximation, the reciprocal of the TDJ0 index of of CD-I mice Hite et al. (1980) reported very large Peto et al. (1984). From a database ofcarcinogeoic po fluctuations in each time group after oral treatment, tency (OPI) and liver DNA adducts in vivo (CBI) that but the average trend was flattening of the response we had published previously (Parodi et al. 1982) we above 110 mg/kg, corresponding to approximately 47 were able to find a group of 26 chemicals for which ppm according to EPA computations (Lee et al. 1983) bothTOPI and CBI data were available. These data, in (see also legend of Table 3). ascending rank of potency are presented in Table 4 Carcinogenic potency of mutagens in rodents Potency data concerning benzene were inserted at the seems to correlate semiquantitatively with adduct for appropriate places, between the I and 5* position for mation in liver DNA. In 1979 Lutz proposed an index OPI values and between the 3* and 9** position for of potency, the covalent binding index (CBI) for DNA CBI values. For both OPI and CBI values benzene adducts. In 1982, Parodi et al. discussed a procedure ranked amongst the weakest chemicals on the scales of for the transformation of carcinogenicity data in potency. The data seem to suggest not only a quahtaI SAL 000001678 S Cnlli al.. Ri*k cnmations for benzene tive correlation between carcinogenicity and adduct formation but also some degree of quantitative corre lation. Kinetics and metabolism In humans, following benzene exposure at relatively low concentrations (amount of absorbed benzene -4-lOmg/kg) (Tngret al. 1952; Nomiyama and Nomiyama 1974a, b), about Vs is expired unchanged in the air and 2/s is excreted as metabolites in the urine within 24-48 h. Approximately 87% is excreted as conjugates of phenol, 9% as conjugates of catechol, and 3% as conjugates of hydroqumone. In Swiss mice injected s.c. with 440 or 880 mg/kg benzene, approxi mately 3/j or J/4, respectively, is excreted unchanged in the air, and l/j or V*. respectively, is excreted as metabolites in the urine within 24 h (Andrews et al. 1977). Other studies on the urinary metabolites in mice (Longacre et a!. 1981), have indicated that 81% is represented by phenol, 15% by catechol, and 4% by hydroquinone. Trace amounts of trans-trans muconic acid and of 1,2,4-mhydroxvbenzene have also been Fif. 3. Benzene eoDcntnuon in ihe doted inhalation chamber at a funciioa of time after injection of benzene mio (he chamber, In phase I, the major phenomenon uptake and distribution of ben zene m the rac In phase II. benzene concentration decrease*, es sentially by zero ordo kinetics In phase MI. a shift to I* order kin etics occur* 355 detected. Benzoquinones are, possibly, the electro philic reacting intermediates; involvement of free radi cal mechanisms in phenol production has also been postulated (Johansson and Ingelman-Sundbere 1983). The same metabolites and similar ratios have been found in the urine of many other mammals (IARC 1982a), which suggests that the major pathways of benzene metabolism are similar in all mammalian spe cies tested so far. In view of these considerations it seems acceptable to utilize the information obtained from experimental data (both short-term tests, metabolism, and pharma cokinetics) for an investigation of the relationship be tween dose and effects. In order to investigate the kin etics of uptake and metabolism of benzene by inhala tion. adult male SIV-50 rats (270-340 g) were exposed to benzene evaporated in a closed inhalation system (a 2 I desiccator) which has been described in detail else where (Lut2 and Schlatter 1977 a). In this system, ex pired carbon dioxide was adsorbed on soda lime and replaced by pure oxygen to keep the total pressure at a constant level. A typical time-course of benzene con centration in the desiccator is shown in Fig 3: similar patterns were obtained for the experiments reoorteo in Fig 4. During the first 2 h. .he benzene concentration in the air rapidly decreased until an equiiibriisn was reached between absorption and exhalation. Within 2 h, about 90% of the total amount of benzene had been taken up by the rat, about 10% was in the air. In a second phase, the level of benzene in the chamber decreased slowly, always in equilibrium with the inter nal level of benzene so that the rate of the zaetabolic turnover could be estimated from the decreased ben zene concentration in the air. Using different doses of benzene, the phase II data were analyzed for the period of time required to halve the benzene concentration determined at 2 h. The re sults are shown in Fig. 4. In the range 200-1200 ppm (0 f-V* rr.g l). the "halflife" of V -ter.e no; a con stant value but seemed to increase Linearly with ben zene concentration is the air. This suggests a zero order reaction, as if the overall metabolic system was saturated and working at a maximum constant speed. A regular feature of the experiments reported in Fig. 4 was the presence,of a phase III. as reported m the ex ample of Fig. 3: starting from about 50-30 ppra and below, the system was no longer sarurated and first order decay of benzene levels was regularly observed. Other findings seemed to confirm the observation that the raie of benzene metabolism is progressively saturated and flattened at concentrations above 30 ppm. The results of Andrews t al. (1977) seemed to indicate the same, when mice (50 g) were given 3Hbenzene s.c., one group receiving 440 mg kg and an other group receiving 880 mg kg. The metabolites in 000001679 SAL 356 5. Grillj er J.: Risk estimation' :'ur ben/.-* Fi|.4. Reciprocal rate of benzene consumption m rats exposed to increasing concentrations of beruene [phase II of Fig. 3) m a closed inhalation chamDer (Lutz and Schlatter I977a>. On the abscissa, the loui amount of benzene applied to (he t chamber and expressed per kilogram bou> weight is given along with the corresponding concentration in the chamber after 3 h. This sccood scale was obtained from a linear regression (/-0.W9 + 0.014.*, ,m0 9U. n m S) between benzene concentration observed in single experiments ana totai amount of benzene applied urine were measured over a 2--h period. The micro mole equivalents of benzene found as metabolites in the unne were similar (60.1 and 72.6j. irrespective of doubling the dose. Again, metabolic processes were saturated. Furthermore, rats inhaling benzene con centrations of 125. 250. 625. or 1250 ppm for 6 h ex creted similar amounts of phenol in urine: hence ben zene metabolism was already capacity-limited at 125 ppm (Gut and Framik 19S0). Although no appropriate metabolism experiments are available in humans, the epidemiological data on the rumor incidence seem to suggest a flattening of the dose-response curve, for concentrations above 50 ppm (Fig. 1). Conclusions The quantitative concordance between experimental data and epidemiological data seems to allow some degree of utilization of experimental results in human risk estimation. A few years ago. in response to the publication of the IARC report on benzene (IARC 1982 b) it was hotly debated (Sun 1982; Tomatis 1982) whether a linear extrapolation from 170 excess leukemia deaths (for a working lifetime exposure to 100 ppm) to 17 excess deaths (3 times background risk) for a 10 ppm exposure could be defensible. We have presented evidence that nonlinear doseresponse relationships suggested by the data on tumor induction in animals and man and on chromosomal damage in rodents might be explained by the metabo lism of benzene which seems to become saturated in the dose range analyzed. A similar situation has been described previously for vinyl chloride where it was shown that metabolism (Gehring et al. 19*78), macromolecuiar binding (Watanabe et al. 1978), and tumor induction (Maltoni and Lefemine 1975) al! become saturated at inhalation exposure levels between 500 and 2500 ppm. The threshold with benzene appears to be at a lower concentration in the rat (30-50 ppm) ar.d it remains to be elucidated whether saturation of ben zene metabolism in humans occurs in the same expo sure range. Since this cannot be investigated in vivo, alternative methods might be required. One possibility would be the use of human tissue samples fer the study of the maximum metabolic capacity in liver and bone marrow. Because of the flattening of the curves at high ex posure levels, probably due to saturation of the activa tion of benzene to reactive metabolites, the data ob tained at the lowest exposure should be considered the most relevant as a basis for an extrapolation. Dat3 cbtained from dose levels above saturation wiil be of less value for defining the shape of the dose-respcr.se curves in the lower range. It is inevitable that a 10 ppm exposure will result in a higher risk than if the high dose data are given the same weight, regardless of the model chosen. Our utilization of the available in formation tends to suggest that going from 100 ppm to 10 ppm for a 45-year exposure the risk ecu id be two to three times higher than .Jui expected fr^m < Iir-_: extrapolation. Usually benzene workers are exposed for an average of about 5-10 years and not for 45 years; therefore, we are dealing with a small increment of tumor incidence in comparison with the natural background of */* to Vs cancer-related deaths. It is in teresting to note that the European Economic Com munity (COM(85)669 def., Bruxelles. December 3. 1985) is considering the possibility of putting the limit of occupational exposure at 5 ppm instead of 10 ppm. as generally accepted by different countries. Aekmowtedgemenis. This work was supported by grunts from Progetto Firulioato "Oncologia" C.N.R. (contract nos 86 0CM-uii and 86.00697.44). Progetio Finalizzato "Medicine Prcventi'u iR:- S4L 0oO0i6ao S. Gnlli ct !.. Risk estimations for benzene 357 bilitati'a" CN R. (contract no. 36 0206: 56). Mmistcro della Pubblica Istmncne, Rome and Associazione Itatiana per la Ricerca sul Cancro. Milan. Italy. References Andrews IS. Lee EW. Wiuncr CM. Kocsi* JJ. Snyder R (1977) Ef fects of toluene on the metabolism, disposition and hemitopoieuc toxicity of'H-benzcnc. Sioehem Pharmacol 26:293-300 Arfellini G. Gnlli S. Colacri A. Mazauilo M. Prodi G < 1985) Ip vjvo and ia vitro bwding of benseae to nucleic acids and proteins of various mt and mouse organs. Cancer Leu 28:159-168 Ashby J. deSerre* FJ. Omper M. Ishidaie M. Jr. Margolin BM. Mat ter B. Shelby MD (1985) Overview and conclusions of the IPCS collaborative ciudy on is vitro assay systems. In: Ashby FJ. de Senes FJ et al. (edi) Progress in mutation research 5:117-174. Elsevier Science Publishers, Amsterdam (WHO Ger.cve) Choy WX. MacGsvgorJT. Shelby MD, Marnopot RR, 1985) fndjenon of micronudei by benzene ut B6C3FI mice: retrospective analysis of paphcral blood smears from the NTP carcinogen esis bioassay. Mutat Res 143.55-59 Cronkite P. Drew RT. Inone T. Builis JE (1985) Benzene hematotoxicity and keukemogeresis. Am J Ind Med 7 147-436 Dt-n BJ (1985) Recent findings on the geneuc`oncology ofbenzene. toluene, tylters and phenols. N!uu' Res 154.153-181 ECETOC Techomi Report No. 16 (1984) A review of recent litera ture on the toxicology of benzene. European Chemical Industry Ecology and Toxicolofy Center. Brussels ErezS'.'D JL, Wilmer JL. Auman CD. Kligrrman AD (1984) Induc tion of sister chromatid exchanges and micronucivM in mzie DBA/2 mice afle: irhi'.at.cn of centre. Environ Mut-gcn 6:408 Gad-el-Karim MW. Harper 3L Legator MS (1984) Modifications in tbe myctodastogcnic effect of benzene in mice uh toluene. phcnobarbitaL 3-meihylchoianthrme. Aroclor 1254 and 5KF 525-A. Mutal Res 135.225-243 Gehriog PJ. Watanabe PG. Park CN <.19'3) Resolution of dose-re sponse toxiciiy data for chcmxals requirr.g meubchc actua tion: example - vinyl chlonde. Toxicol Appl Pharmacol 44 581591 Gold LS. Sawyo' CB. Magaw R. Backman GM. de Veciana M. Levinson R. Hooper SK. Havcnder VVR. Bernstein L. Pete R. Pike MC. Anars BN (19S4) a carcinogenic potency database of the standardized results of animal bioassays. Environ Health Pexpect 53.9-319 Goldstein BD (1985) Risk assessment and risk management of ben zene by the Environmental Protection Agency. In: Hoel DG. Meml RA. Perm FP feds) Risk quantitation and regulatory policy. Cold Spring Harbor Laboratory. Banbury Rep 19:293304 Gut I. Franuk (1980) Kineiio of benzene metabolism in rats in in halation exposure. Arch Toxicol (Suppli aji J-J|7 Kite M, Pecharo M. Smith 1. Thornton S (1980) The cfTect of ben zene m the osacronucleus tor. Mutat Res 77:149-155 LARC Monographs on the Evaluation of Carcinogenic Risk of Chemicals lo Humans (1982a) `Benzene'*. International Agency for Research on Cancer. Lyon 29:95--14* IARC Monographs on the Evalutauoa of Carcinogenic Risk of Chemicals to Humans (1982b) `Annex'*. International Agency for Research on Cancer. Lyon 29:391-398 IARC Monographs on the Evaluation of Carcinogenic Risk of Chemicals to Humans t'.982c). International Agency for Re search on Cancer. Lyon (Suppl) 4:56-57 Johansson I. Ingclman-Sundbcr(1983) Hydroxyl rjdicul medi ated. cytochrome P-450-depcrdcnt metabolic activation of ben zene in microsomesand reconstituted enzyme systems from rab bit liver. J BkXChem 253 731 (-7316 Kjlf GF. Snyder R. Rushmore TH (1985) Inhibnton of RNA syn thesis by benzene metabolites and their covalent binding 10 DNA in rabbit bone marrow mitochondria m vitro. Am J Ind Med 7:485-492 `Lee SD. Dourson M. Mukerjee D. Stara JF. Kawccki J (19831 As sessment oi beuene hetHh effects m ambient w*ter. In: Mchlman MA (ed) Advances in modem environmental toxicol ogy. Princeton Scientific Publishers. Pnnceton 4 91-125 Longacre SL. Kocsi* JJ. Snyder R (1981) Influence of strain differ ences in mice on the metabolism and toxicity of benzene. Toxicol Appl Pharmacol 60 398-t09 Lutz WK (1979) In vivo covalent binding of organic chemicals to DNA as a quantitative indicator m the process of chemical car cinogenesis. Mutat Res 65:259-356 Lutz wk. Schlatter C (1977 a) A closed inhalation chamber for quantitative metabolism studies of volatile compounds with smut! laboratory imnalt. Toxicol Lett 1:83 -87 Luu WK, Schlatter C (1977 b) Mechanism ofthe cjrcmcgtnic action of benzene: irreversible binding to rat liver DNA. Chem Biol In teract 9:253-237 Maltom C (197$) The value of predictive experimental bioassays m occupational and environmental carcinogenesis. Ambio 4 10 Malioni C. Lefcmme G (1975) Carcinogenicity assays of vinyl chlo ride. current results. Ann NY Acad Sci 21$ 195-224 Maltom C. Scamato C il J`9j First expenmenul demonstration of the carcinogenic effects of benzene. Long-term bioassay on Sprjgue-Dawley rats bv oral administration. Med Lav 70:352357 Malioni C. Conti B. Com G. Beiocegi F (1985) Experimental studies c* brnzere carcinogemetty n '.he Bo'egna .r.stituie of anca'cgy uur-e-; results ano ongoing rrsez.vb. Am j ind Med ' 4;5--i46 Ncrr.:v jma K. Nomiyama H 119"-a< Respiratory retention. uptake jrJ excretion of organic solvent* m man. Benzene, toluene, nheur.e. tnchloroeth)ler.e. jcetor.c. etfaylaceiaie and ethyl alcohai. Ir.t Arch Arbeitsmee 32."5--A3 Nonvyamj K. Nomiyama H 't9"4b) Respiratory ehejiunor of or gan.c solvents in mart. Benzene, toiuene. n-kexane. tr.eLrroethy.ene. jcetctie. ethyiacetate and ethyl alco.no:. In: .Area Arhef.smed 3255-91 " NT? 'N.rcr.a! Toxicology Programme) (19S4i NT? Technical Re po*:: Toxicology and carcinogenesis studies of benzene x FJ44, N rats ar.d B6C3FI m:cr igivage studies). NT?v-0'2. N'lH r-5.icjf.on 34-2545 US Dept of H H.S. Ott *.`.G. Townsend JC. rishbe.x Wa. Langner R.A (19'5 Mortality omvr.g individuals ocrapauor.all) exposed to benzene. Arcs Env::-n Health 33 3-9 Pared- b Tamr.jher M. Santt L <1982) Alkaline elution :n vivo flu.rtmrtnc analysis m rati. Quantitative predicuvttv of caretn.'gsr.trtty. as compared with ether short-term tests la- Bridges B*\. Bzttsrwonh 3E. Weinstein IB teds) Indicators of gesoioxic evrosure. Coid Spnng Harbor Laboratory. Sl.-?i_-v Rep 13 !3`-l55 Paredi S. Baibi C. Abefmoschi ML. Pa!a M. Russo P. Saati L 11953) Studies on DNA damage: discordant responses of rate of DNA disentanglement (vtsconmetnally evaluated) asd alkaline elu tion rate, obtained for several compounds. Cell Siophvj 5:235300 Pet R. Pike MC. Bernstein L. GoM LS. Ames BN i1934i TheTD.y a proposed general convention for the numerical description of the carcinogenic potency of chemicals in chromc^xposure ani mal experiments. Environ Health Perspect 58:1-8 Rail DP (1977) Species differences :n curctnngtr.ic testing. In. Wiau IIH. Watson JD. Winsten JA led*) Ongms of human cancer. Cold Spring Harbor Laboratory. Book C. pp 1333-1390 Rmsky RA. Young RJ. Smith A8 (I98t( Leukemia in benzene workers. Am J tnd Med 2 217-245 (this study completes the ful- SAL 000001681 I358 S. Cnlli et L: Risk boos for benzene low-up by Infante PF. Rmsky RA. Wagoner JK. Young RJ (1977) Lancet ll:7$-78) Snyder CA. Goldstein BD. Sellakumar A. Bromberg!. Laskm S. Alben RE (1980) The inhalation toxicology of benzene: modenee of hematopoietic neoplasms and hematoioxicity in AKR/J and C57BL 6J mice. Toxicol Appl Pharmacol 54:323-331 Styles JA. Richardson CR (19$4) Cytogenetic effect of benzene: dosimetnc studies on rats exposed to betuene vapour. Muiat Res 155 203-209 Sun M (1932) Risk estimate vanishes from benzene repon. Science 2J7:9t4-9l5 Teismger J, Bergerova-Fiserova V. Kudme J (1952) The metabolism of benzene in man. Pracov Lek (Pol) 4:175-10 Tomatis L (1982) IARC benzene repon. Sclera 218 214 Vigiiani EC (1976) Leukemia associated with fceaane exposure. Ann NY Acad Sci 271 143-151 Waianabc PG. 2cmpel JA. Pegg DO. Gehnag PJ (1976) Hepatic macromolccular binding following caporae to vinyl chloride. Toxicol Appl Pharmacol 44:571-579 Received January 23, 1987/Accepted February 25, 1987 Oo 00 '*6