Tobacco Documents Online

1~7 Env'Fonmental Science & Te 3ogy, Jan., 1979 Pages - 8 Workplace cancers: politics vs. science I Industry~faul~t~s the logic and the~~meth~odology~~of a recent ~ government ~.rtudy~ which attempts~to estimate future cancer~ deaths from present and past worksite exposures Watch a tennis match to its con- clusiom and you lcavc satisfied: the resolution is dccisive, the winner krto~%n. Follow the IOOCA cluinis ana countercltri'ms on the estimates of cancers of occupational origin, and youl arc left in a disconcerting tizzy: the resolution is unccrtain~ the consc- qucnces monumental. Except for the matter of rcsolution, the tennis analogy holds up well in t'his latest exercise in mulhcmatic manip- ulation. The players are known: I I EW scicntists vs. industrial scientists. The _ basic equipment used by both sides is idcnticali epiderniological studies and bioslatistical methods. The court-- OSHA's cancer policy hearinE~ and the prcss-is sharcd'. The rallic.,trc fierce and cxciting, but whcn the match of tltiti "estimates gtnnc" , is played out, the spectators suill don't know thc score. Players ready? I'la~! The opening serve wa,, ntadc on September 11, when I II?«' Secretary Joseph Califano, Jr., speaking beforc, the AFL-CIO National Conference on Occupational Safcty and Ilcalth„ quoted from the thcniunrcleaticd I t'd_'W study "Estimates of thc Fraction of Cancer in the United States Related to Occupational Factors," writteni byy scientists from the National Cancer Institute (NCl); the National Institute for Em•ironmcntal I Icalth,Scicnccs, (NIEfIS) and the National Institute .. for Occupational Safcty and Health (N105FI). Califano noted industry's frequent reference to the csliniatcs of 1' -5% ass theoecupaiionai'contribution tothc totol cancer burdan in the U.S. Then he qltutcdl frrnn thc I1I.W study: "1f Chc full conncquenccs of occuPational' exposures in the present and the recent past arc taken into,tccounU, estimatcs uf at least 30'.tt-,appcstr much more reasonablc; and tnay even be conser- vative." Califl;tno charactcriecd the report's findings as "alarming." As cxpected, his stalemcnt received wide media play; industry was not slow in returning the servc. That samc day. Perry Gchring, di- rector of health and environmental research for pow Chcmical USA, was quoted in the tf'n.chirlgtorr Post as .. saying: "Whaf it boilkduwnito is that Califuno is all %Nct. Ille doesn'it h;vc the data to support those findings.'. Two wccks later, on October ?5, the Antcric;in Industrial I Ic;tlth Council (AIF{C), amassoci:rtion of 119 mcrn- bcr companics and, 66 trade groups formed sulcly to onr al"coordinatcd, constructive" response to OSHA's proposed cancer policy, held a press conference to present its rebuttal to the governmenU study, which it ternied "exaggerated speculation." . At the press confcrcncc, Fred Flocrgcr, chairman of tihe All IC sci- entific committcc, said that hc could ' charitabl~," tcrm the HEW study a "poor paper based on infl,ttcd esti- mat'cs of the population at risk, and erroneous atiswnptionx in s4rtistics and cpicicmiology."ln its introduction, the AIIl(_ rcbuttul rcinfurces IlocrRCr's statements with phrases that fault the govcrnmcn0 study for using "qucs- tionublc logic," "selective ... datti that is ouUdmcd"and of "yuestionalblc sci- cntiific v,ilidity," and "unsoun ' d mcth- odi;'' which lead to "unscientific and highly speculative conclusions." Sonic of industry's criticisms, discusscd latar, are valid, and are supported bz inde- pendent scientists. Despite these criticisms', the two reports share sonicvery tcllih~ char- actcrititics. Qothiwcre prepared hastily for inclusion in the posthearing record on OSI I'A's cancer policy. F3othihavc

HEW's Califano NCI'.s Schrtnidermart NIL'NS' h'rrll been widely publicized. And neither posure to the carcinogen comparcd to has been peer rcviewed'4 although,the tile number expected in tile gcneral H'EW paper has since been widely popultiUion--which they dcriN.cd from distributed for comment. In fact, it is studies publishcdlbctwccn 1947- 1978, now being revised based on the cont- mcnts received (which have been sur- prisinglymeagcr to dittc), and will bc- submittcd' to a scientific journal for publication, perhaps later this year. } I E1V: rankings change Ten government scientists from~ NCI. N'IEI-IS and NIOSH preparcd' thc government report. Although this paper was written quickly, discussions on the concepts and mct~odologics containcdl in it had supposedly been going on for months. . According to scientists contacted by ES& T. tile 10 contributors had long felU that the l-5%cstimatcs of cancers of occupational origin putl forth earlier by John Higginson, director of the International Agency for Research on Cancer (Lyon. France), Gio Gori of NCI, Ernst Wynder of the Americani 1lcalth Foundation (Valhalla, N.Y.) and Richard Doll of Oxford University understated the true contribution of occupational factors. This is cspeciirlly true, they, fclt, once Irving Sclikoff s (Mt. Sinai Schoollof Medicine, New York, N.Y.) data on asbestos-rclated! discascsarc factored into tile calcula- tions. They thcn began to develop wh,it they considered to be more realistic cstimates. To predict the number of futurc cancer dcaths which nrty be attribut- ablc to past atibestos exposure, the I IFyV scientists rclicd he;rvily on Si:l- ikoff and coworkers generally well- received work,, plus some Ilss than well-defined cstimatcs of thc numbcr of wurkcrs exposed, to asbcstos in the U.S. since World War 111. . To project annual excess deaths as- sociatcd with occupational exposure too five other carcinogcns-arsenic, ben- zene, chromium, nickel and petrollum fractions, including the aromatic hy- drocarbons--for which ample datal cxist, the scientists used a mathcmat- ical formula. This calculation is a se- ries of multiplications: risk ratio - the number pf cancers expected from ex- multiplicd by the numbcrof workcrs cxposcd-a numbcr usually obtai'ncd from the NIOSH-commii;sioncd Na, tional Occupational Hazard Survey (NQf 1S) conducted between 1972-74, multiplied again byt'~he incidence for tile tumor type induced by the cherni- calL The AIHC rcbut'tal argues that! "serious crrors" were made in dcter- mining all three of these multipliers. For cxamplc, AIHC claims that the risk ratiosuscd arcnot applicable to recent workplace conditions, that! tile estimates of thccxpo}ed population arc "inflated," and that incidence rates arce uselS intcrchangcably with mortality without adjusting "for surviv,tl or conipc.ting c wties of death." A Canadian canccr epidemiologist, A. B. M illcr, who rcccivcd a copy of the IIE'V1' papcr from NCI's Marvin Schneidcrman for review, agrecs, He toldi L-.S& T thatl, in gcncral,"hc fccls that the governmcnt! paper overesti- mat'cs the number of people exposed and f or the degree of risk. kliller cites the cases for nickel and chromium, with which he is familiar, as incorrect cxtrapolations of risks from a subgroup to the whole group. Another example ol?ainiinapproprintc cxtrapeflation, AII IC s,iys; is th,it (lrom coke ovcn and gas workers exposed to pok,nuclear aromatic hydrocarbonti (I'NQs) to,all workcns potentiiil9y cx- posed to I'NAs in olhcr indu~trics, Nlillir quickly addsl how'evcr,,thai tile I I I:W papcr is ai"uscfullcontribu- tion bccause it will [stimulittc]'people to cah:ulatc the numbers more ,iccu- rately." Crucial points in first set The bulk of the difference in dis- puocd cstimate rangcs;, I-5`%o vs. 20. 40%, may be cxplaincd! by asbestos- related deaths, if you accept the gov- crnmcnt calculatiun++: Inithc next three decades, the government scicntists predict thatlcxcesti canccr deaths front past asbestos exposure may reach . NIOSN's Bridbord 2-2.3 million, or about 58' 000-75 000, deaths per year. If this conics to pass,, ` atibcstos-relatcd deaths will, then ` comprise 1'3-1 x%of all annual.canccr deaths, assuming that total canccr mortality reaches 400 000-450 000/y in the next few decades. Annuai excesscanccr deaths for the five other carcinogens (box) may ac-, count for another 10-2011% of thc an- nual cancer deaths from; all causes. The contribution to total annual can- cer mortality made by asbestos plus the five other carcinogcns may range , from 23-38%, accordin to HEW. A[HC's bll~stin 03~~2528. AIFIC bclicves that'scienlific evi- dence does not support tile niugnitudee of tile projt ctions made by the H EW scientists. The Council takes HEW's. pivotal argument-that for a~;bcs- to!;--and! asks tile question: Arc we today in the midst of tile cancer epi- dcmic pi~edictcd by the governmcnt study? To answer thi's, AI l IC analyzes the casc for mcsothclioma, a very rare type of cancer which serves as a marker for asbcstos-induced~canccrs, . Using very conservative assump- tions and HEW-gcncraicd numbers, AIIIC calcultttcs that we should be seeing from 5000-10 000 deaths per year from mcsatlhelioma. Instead, AI I IC says, NCI's ll-ci'ty/statc Sur- vcillltncc l:pidi mioli)ry and l:nd' Rc sults Program (SEER ) tabulatCs only 950 deaths per year fi-om masotliclio- noa, with no upwurd surge eviiJcnt.. Ni>llsci, says Schncidcrmun, a NC1 contriliutor tu the IIII;tiL' sludy,. Ile poilnts out thaC deaths from mcsothc- lionia have increased 1i25ON', during the 1970-76 period, and' that there has been a highcr death ratc in men lhnn in wanen, which is consistcnt with previous industrial exposure. Further, he counters, seven out of the 1I SEER's cities/states werc not ship- build4ngarcas during World War II, and still we are finding incrcascs in the disease. Nevertheless, there is still an order of magnitude difference between 950 deaths and. the cpidemie-like 5000-10 000 deathscalculatcdl byAI'IIC. '

l C ~ N~YVs .S'i•!rAoJJ~ ~' ALM"s ll nergf•r Both Scaineidcrnutn and NIEIIS' director David RaiL another cont'rib- utor tu tlic HI:W study, feel that the ,.inpondcrables of diagno,is.' have probably resultcd in an undcrreporting of mcsothclioma, AIIIC counters by saying thnt the publicity given to World War 11 shipyard exposures to asbestos probably results in an ovcr reporting uf this rare cancer. Whatever the true rate of rcporti'ng may bc, ntcsothcliumu is not thc only furrn of cancer induced by atibcstus. According to III:W"s catcultttions, somc 4:S -5.6 million people wcrc ex- posed toasbestos in shipsards during W'orld' War II, and another 3.5 G.5 /)un'`.c (ic'1irf7t,t; While acknowledging tlutU work- place exposures to asbestos have de- clined, Schncidcriman recounts Chat the first asbcstos standard was set au I2 p,trticlcs/niL in 1968, unly 10 years ago; ,tnd that thii; was furthen lotticncd to 2 rarticlcs/nol- only in 1976. Hc points out'thau t'hcrc is sonic evidcncc th:tt even 2 particli*,./'n1L is too high: O51I A, in fact, is considering loN~cring the st;tndard to U particlcs/mL, Schneidcrman says. To the charrc that I'il:W used outd,ttcd dat,r. Sithncid'erman turns to the rcgulatory, history of thc five other carcinogens considered. W'itlhi the cx- c:eption,of nickel„hcsays, the dccrcascs I)uu•'s Blhir many, uother chemicals also, have, hoNNcvcr, impruvcd Lrnn;iderably since tficn." Althuugh the guvcrnment' sci- cntists citt; a range of risk ratios fur cach of thc fivc carcinogens, and never usc the hio,licsti value, they do apply p risk facttrrs as.oci,ttcd t+ith hcitt y cx- posurc to workcrs thut m,ty not.ltov, be .' "subject to [thut] heavy cxposurc," the AI I IC rebuttal cnophusizcs. Thc Council also rebukes the gov- ernment scientists for their careless interchange of cancer incidence with . cancer mortalit'y. IloN~cvcr, at le,tst, for asbcstos-induccd lung cancer, in- cidcncc and mortality are virtually identical, unless, ot cour.e;' a lung cancer p;tticnt dies in an automubilc accident, in u hich case lhis dcath cen- million have been exposed to the min- iniczposurcr httvc been rcuthcr rcccnt, eraf in other occupations. About 4 and "old dnt;rouglti to bc uscd." million of thcse are believed to have In response to the chargc that the been heavily exposed. And uf the 1972 -74 \•OIiS data infft,tc thc c,ti- heavily exposed \corkcrs who havcal- niat'es of the number of «orkers cur rcady dicd, cpidemiolugic evidence rcntl% and aca'uufly cxposed'. Rall indicates thnt'?0-?i'!9)died'from lung concedes that "thc NOl IS data crre not! c<tnccr.7-10'!irdicd! fromi metiothelio- the `tror.ge;t'. but %kc have no, other ma and K-9+'t,diud from g:tstrointcsti- data. Industry could'huvc +korkcd with nal cnnccrs. us in the past and supplied better AIf(C claimt; that HI;SS"s projac- data." tion of futurcdeatlth from past asbes- Turning to the five carcinogcns, toti exposure ovcrctitimates the number AI I IC chides the I'1G1V scientists for of heavily cxposcd, World War I I using risk ratio,, obtaincd "I'romistud- sltipyard %%orkcrs who arc still alive. ics of workrrs exposed in the . today; "disregards the fact that as- 1920's I95Q7s. Wurkhlacc conditiuns bestos exposure has bccn declining by for knu%% n carcinogens and for ,r unat i %%or ers an f k d both number o by degree ' of exposure in rcccnt, ycars: and ignures any changes in smokinfg hubits;" whicli anwng asbestos wurkcrs, would reducc lung cancer mortality r,utcs signifi- cantly. (l:vcn if the numbur of rrcvi- ously exposed shipyard workcrs alive today wcrc accuralcly known, precise calculations oG future mortality would - not be possible because the age distriF bution of these workers, and the du~ ration an&cxtcnt of their exposures arc not known.). ttl•:1V: hack on the offentiite Sclincidcrnt,tn, studying the age dititributioniof shipy,ard'"orkcrs dur- ing World War Ii, found this working population "very young"; therefore, he bclicvcs, many more arc alive today than "were guessed at (without age- distribution data) by AIH!C." tific:atv would not notc thc cause of dc:tthiati lung cancer. AIl IC says to I IE1S': double fault Ini,uannrarizing thc dcficicncies ul":' the government paper, AIIiC con- cludc,: "By selecting high risk ratios .rppl!icablc to a smalllgroup of highly exposed worker.,,by inflating cstiinutcs of workers actually exposed, by con- fusing incidence with mortality, by ncgli?ct'inug age and sex distribution of worker cohorts and by confusing at- tributabl'c risk witit statistical associ- ation, the [1'IL`V] Paper :trrivcs at a hemical' Number of workers potentially exposed isk ratios Arsenic About 1.5 3-8 (for respiratory tract million cancers) 1 Benzene About 2.0 About 5(for leukemia) million Chromium (trioxide & other About 1.5 3-40 (for nasal cavity & hexavalent chromium million. sinus;,lung 8 larynx) compounds) Nickel (oxides) About 1.4 5-10 (for respiratory million tract cancers) Petroleum products„ About 3.9 2-23 (for lung cancer) including aromatic million hydrocarbons . Projected excess annual cancers 7300 14000 CJ 7900W 9100 • Source: FEW's"Estirnates of the Fraction of Cancer in the United States'RetSted to Qccupational. Factors" ..n. .6 1 i --

t speculative concfusion conccrIllTntg the probabilities of an i'mmincnt! cancer epidemic due to occupational causes. Kenneth Bridbord'; a NIOS'II scir cntist whose unpubGshcd paper "New Ilorizons in Occupati.unal Medicine" supplied the philosophic~,l under- pinnings for the IIEW rcporu, feels "that even with its faults, whiclr arc listed in the report itsclf, it still reprc- scnt:s as good a referenced estirnate as anything done before." If anything 13ridbord points oul', the I Istudy: may underestimate the risk of cancers from occupational exposure bccausce the paper only considers a small num- bcr of compounds: agricultural occu- pations arc not included, which climi- nates the vast number of workers cx= posed to pesticides: and only a Iimitedl number of cancer sitcs are notcd'1 (bladder cancers, for cxample, are not estimated because thc data arc too inadequate, E3ridbor,: says, to make projections). - The confines of tlie I I[:1V studv are further restricted by not including skin cancers and radiation exposures, Bridbord notes. Moreover, thc epide- miologic studies thntan uscd'doinot follow workers to the end of their lifc- tihies, so the cstimatcs of true risks among 'thc populations at risk arc probably low, Qridbordl contends. Schneidcrnian an& Rall, of course, concur. Play suspended The 11 f:W paper is now being rc- vised: howevcr, the estimates will probably not change substantially., Schneidern,an told f:S&'T that hc is. "conrfortabllr with [,tliej estimatcs:"• Rall says that "20% is probably a goo& figure." And 13ridbord feels thati "20% %tiill hold pretty firm, although there may be sonic adjustment on the higher side." Inithc process ol' revision, Schneid- crnrcui sent the I i E«`' paper out to a Icgioniof scientists. When asked if he had heard from I IiprinsonI he rcplirrdl "No, but then I haven't hcard from a lot of peoplc.;" . I ligginson, the author of the widely quotud estimate that 60 90"s; of all human cancers mayy bc associated with environmental factors (including dict', smoking and drinking habits, and %%orkplace carcinogens),,has publicly stated that the data in the I I E%V study .'do nou make a good case," and by cmphasizing occupational cmccrs, thce report may be "detrimcntalto the best iritcrests of canccr control." John Wcisburgcr of the Amcrican~ Ilealth Foundation agrees with Higginson: He told E,S&T that he bcjicves 'that the HEW' paper m1f places the emphasis on occupational cancers whilc ignoring, or not properly stressing, the major contribution that lifestyle habits-smoking, drinking and dirrt-make to thc total canccr burden. Selikoff, on the othcr hand,s;tys that the "I-IIEW report doesn't deflect fromi the more important contributions to cancer [madc by such factors] as smoking." On the contrary; hcstresses, it rcinforccs "Ihc importance of con- trolling those factors such as.moking that are recognized as causes itndi which can be controlled." The largest U.S. manufacturcr of asbestos•prod- ucts, the Johns-Manvillc Corp. (Den- ver, Colo,)', has an antismoking pro- gram. Sclikoff says, that "should be cited as evidence against Flil;ginson's fears." Adiantage of replay Califano's motives for announcing the results of the study at a union meeting have been tluestioned by in- dustry,. Etcyl [3llrir, Dow ChemicalPs vice presidcnti, dircctor for ccaltih and environmental sciences, is disturbcdibyy the timing and the manner of its rc- leasc, stating that Califano made cancer "a political not a scientific issue." Furthcrmore, Blair feels that the study itself "discredits the Amcr- ican chemical ind'ustry." Rail counters by asking what better forumdid Caliifano~havc than a labor mccting, if, in fact, IIE1V's estimates of occupational canccrs prove to be realistic. "To havc withheld such in- forniation from the public would be unconscionable," Rulllstresscs.. Certainly the government study N~as, in part, writtcn to bolstcr OSf1A'ss proposed cancer policy, but it's pri~ mary purpose; the HEW scientists clttim, was tojug complacent thinking. Whilc not underrating the majorcon- tribution made by tobacco smoking to cancer mortality, the government of- ficials were reminding othcr public- health scientists to look for the occu- pational componcnts-thc prcvcntablc fractiion--of thc total cancer burden. They were also reminding public- hcalkh scientii;ts, ofcanecrs'basic charact'erii;tacs: multiple, interacting causes; arid' multiple stages of dcvcl- opment and; its corollary„longlatcncy, periods. As Sclikoff told LS& T, "to be auarc of the multifactorial infcrenccs, and to use them arc two different things." Miller agrees: "A lot! of us tend to overlook the industrial com- poncnts of the multifactorial causes of canccr," The question nevertheless remains: should the government scientists have. made these points by cstiiriating future cancers ofoccupational origin based on a limited, if not poor, data basc'? Sclikoff answers that the task the government scicntistshadtoperforriv was "a difficultione, an urgent one, but an impossible one to complete with any rigor or accuracy." I Ic feels that they had the "guts" to address issues that neodui reevaluation, and that they "correctly enumerate and detail the reasons why previous judgmcnts were inadequate," especially why.thc one- effect, effcct, one-cause approach, or as Scl- ikoff terms it "the singll-slot ap- proach" to thc classification of cancers . "is wrong." Recognizing that the government scientists used data bases (NOt-IS data) that were sometimes "vcry thin," and made It:aps in assumptions that research scientists would be faultc6 for taking, Selikoff queries: "Should they as public-hca{th scientists be frozen into saying nothing" because the bodyy counts aren't in'? Sclikoff answers in the negative. "Thrauthors of this report performed a vcry valuable function in calling at- tention to the difficulties in thc,past evaluations, the critical factors that must be taken into- account, thc inr adcquacics of past data, and'thc im; portance of proceeding to acquire more accurate data." In short, he says, they acted as responsible public-health of- ficials. Match to res03732530 . ume The government report has var- iously been called a"uscful cxcrcise," a "numbers game," and a "political witch hunt." Whatever its faults-and . thcrc. arc many=it has stimulatcd' fresh debrte, principally, licrhaps, bccause industry fears the estimates will be uscdto set public policy. Pauli Kotin„senior vice president of Johns-Manvillc„ in his posthcaring submission to thc'OSHA cancer policy record said, "I think there's a IcgiUi- niatc right and need•for suchia docu- ment," but hc implied that he would like to see one of higher quality: Sol would'the government scientists. I lowcvcr, bettcr estimates will re- quire better data: the number of workers exposed't the duration and extent of cxposurc; the age, sex and health history of thc exposed workers; andl thc precise formulations of the compounds to which they are exposed. The repository for this information is industry. And industry will have to be. more forthcoming with this informa, tion or its v;ilid criticisms of the current I I E1V study will'e be rendered mean- ingless, and its right to future criti- cisms negated. • LRE