Document R2aQRn9g43Qq4dJ6YDn9ky7GX

MEMORANDUM n;uuc Mr.a i.'iii snuviciCI'-MI-H 1 OK mil Ai!): fl INI HOI. n a t io n al in -s i run a i o r -o c c u p aiio n al sAv^nr a n d utA-urii TO TROUGH TROUGH TROUGH Deputy .Director, HTOSH DATE: August 24, 10,76 -N sf CKhiieef, Industry-wide SSitudies Branch, DSHEFS, NI03i{R^tAJ f3Chhii;af, Bioraetp'- Section, IWSB, DSHEFS, NIQSH i\L- M Dii;rector, Division of Surveillance, Hazard Evaluations and Field Studies, NIQSH FROM Epidemiologist, Biometry Section, IWSB, DSHEFS, NIQSij SUBJECT: -Review of DCMA .Lead Chromate Report The "Epidemiological study of Lead Chromate Plants'' prepared by Equitable Environmental Health, Inc. examines the mortality experience of 577 lead chromate workers. The study population is defined as any male worker who had been exposed to lead chromate for at least six months prior to December, 1S74. Ninety-five percent.of 548 workers' vital status was ascertained and 50 deaths (with certificates) had occurred in this group. The study population was of necessity divided into three groups since the workers were employed in three separate plants as fbllows.: Total Hired before I960 Deaths/Certificates Plant 1 Plant 2 Plant 3 . 238 workers 152 workers 15:8 workers 548 workers 131 45 158 : 334 12 deaths 5 deaths .33.deaths *- 50 deaths The analytical section of this study, despite the small numbers of deaths involved, show excesses of both respiratory and digestive (primary stomach) cancer- Plants 1 and 2 had exposures solely to lead chromate:. Plant .3 had mixed exposures to lead chromate and to zinc chromate, (refer to attached table) with only one worker having sole exposure to lead chromate, (He died .. of lung cancer). - The most, apparent serious deficiency of this study is the short latent period of this worker cohort. Table 5 of the study shows it to be 17 years at most. A further complication is that most of the workers with the longest latency periods fall into the mixed exposure group at plant 3.' The relatively purer .exposure of workers at plants 1 and 2 is diluted by recent hires, and short term employees resulting in a cohort of only 176 workers (total from both plants combined) who- were hii'cd prior to 1960. Plant 3!s total is 158 workers hired before 1960, Of the workers who were hired before I960 and worked 10 years or more - plants 1 and 2 total 59, whereas plant 3 has 138. (Refer to attached table), N 27680 DUP040010992 * Page 2 - Meaorandum, -Review of DCMA Lead Chromate Report, August 24, 1976 It is a generally accepted theory that the latency period (time from first exposure to the onset of disease or death) for occupationally induced cancer is usually in the range of 20-35 years. These lead chromate workers have most likely not had sufficient latency for any significant excesses to appear. L Other noticeable limitations to this data set are the extremely small sample . sizes, as stated above.' Essentially there are three cohorts which must be analysed individually - refer to above for numbers of workers by plant. The-possibility of Type II Error, or failing to detect a significant difference when one does in fact exist, should be considered in the evalua tion of this data set, due to the very small numbers of deaths involved. h *1 No age distribution for this cohort is presented .anywhere in the study. Racial differences are presented as follows: Plant 1 - "overwhelmingly white". Plant. 2 - "conservative estimate is that half of the plant population is ncn-vhfte", Plant 3 - "is 50% non-white". Nevertheless, the .authors chose to use -the crude death rates (for the state in which each plant was located)to generate expected deaths. The reliability-of the SMR's generated using this-type of control population, which is unadjusted for confounding variables such as age and face is questionable. The contributory role associated with both age and race, in both respiratory and digestive tract cancer is well documented. u \ When the authors used age-sex-rac.e-adj-ust.ed rates for the S.MjS.A. for each plant location, as a control; the SMR's changed considerably, a drastic illustration or the effect of the choice of an appropriate control popula'tlon on statistical analysis. Considerable bias may well be present in the control, group. Another limitation on respiratory cancer inferences is the confirmed smoking history of nearly all those wno died of lung cancer. This problem could probably be resolved If latency were longer and if cell types of lung cancers were given. In spite of the above methodological problems, an increasing'gradient of the SMR's was observed with Increasing time from first exposure to lead chromate at plants 1 and 2. The mixed exposure to zinc Chromate and lead chromate at plant 3 resulted in consistantly elevated SMR's fpr respiratory and digestive cancer - consistently elevated due to cohort composition, a statistical artifact. If the records available at plant 3 (prior to 1957, plant began In 19-11) had been included in this study, more of a gradient in cancer deaths would probably have shown up in the analysis, similar to that observed in plants 1 and 2. J The results of using S.U.S.A.'s for' a control population and thereby ' controlling for variables such as age, race, sex are interesting. The .SMR calculated for respiratory cancer dropped somewhat, but still shows an excess. For plant 3, the SM!i for digestive cancer elevated to an S-fold excess (from 209.9 to 77 S,3). { DUP040010993 Page 3 - M^ncrandum, Royiew of DC5-IA 'Lead Chromate Report, August 24, 1976 The authors stats that sinew no mention of stomach cancer as a consequence of chromate exposure is made in the scientific literature on zinc chromate and since no stomach cancer was found in plant 1 and 2, it is unlikely that these stomach cancer deaths arc attributable to cither of these two '! exposures'. One purpose of epidemiologic research in the field of occupa tional health is to uncover associations between certain chemicals or compounds and the resultant health status of workers occupationally exposed to these substances'. Therefore, it does not seen justifiable to rule out nr.y association between stomach cancer end the exposures to zinc chromate in plant 3 just because there has been no previous mention in the x scientific, literature. An 8-fold excess of stomach cancer deaths in any plant can not be lightly passed over. A more thorough detailed investiga tion of work history and other factors should bo dons, as the authors mention. The authors contend that no definitive conclusions can be made from this Study, due to the small numbers involved. But, the respiratory cancer excess holds up for both the 12 deaths in plant 1 and the 33 deaths in plant 5, shewing a gradient with-increased exposure and latency, which should be investigated further. A re-analysis of this data is in order. '^ `i This reviewer contends that the latency was not long enough, sample size was too small and that the control population was not 'well enough defined for a clear conclusion to be made concerning the exact magnitude of cancer risk being experienced by workers exposed to lead chromate or zinc-lead- chromate. A tors rigorous data presentation and re-analysis of mortality data should probably be done in order to clarify the excesses of respirator;/' and digestive cancer already demonstrated by way of this insensitive study. ' *& Cynthia Robinson < DUP040010994 SIIMHAUY OF STUDY RHSULTS rtf p < jj vciu Un cx. X vt JS> o ou- ** .50C "IOS r_, CM Wl (M V3 o> u. Mrt fCjl VT C o .n n. .0 P S V) O CM P *N* >0 O' 0 <0 so c. o tj "NVI. =>l -0 O CM = 1 vi 3--;f O'** .VPi iz: s gi >: ^1 ^ J w O -CSMs v <r* K to '<w>HU uOcs & sV.I 'S wo W Su3 y.0 -- c vSi.ScJs PO == SO Pi K *ft. 0 O Uo4 -o z0 MvVo so* -O O 0 >OSr cn to CM i-i CM 80 CM 1 #0 M CM s* 0 -w s*-o* ' -r *p M 'S*' n N * >9 . CO rt H " * Cl :<--0fl w to <*t<. r HI f* M OCM. ' OV CM <3 0 0 O Cl , ' 0 W 90 o N ^ O > SO O .O 0 O - 10 M 0 -O S* 00 - to CM *r M CM CM 0O 0O w o <* 0O >*. p O CM i** Cl N N to *--1 fy .Hi O CM >CM* O d I 0 N, O >9 d 0 CM *-i --H CM rH r> to to s~1 N to o>HI Vo. *V-S wC o ucj rtf so cu til VI X5 0 O O O N O .*er4 V| VI y .0 1.-9 6 rt ii >VvI 0 > V *u1c, *r>t 0 O MO jt O Cl l< tfi LJ y U O t. O tv. W 'I O d Cl 0 - >*. 0 "O a s- Vt oC c: 0 #< fl V> cl. V.I "J 1. b c 0 O AJ ,v,` i>i rl i. --i ^ O v a Pu t* 1. ,u -.1 U Wi a Jj e > n, ^3 u .0 0 *0 0 s 0 * 0 "tJ *0 9?3j 0X s:i ,*.4. O y ^ n. e-J i ny .*n # U% vs i< .0 W| <J si V> O .** i. ;> m - ,i a 0 0u Jt tc. >.* 0 >00 ./ U -* .0 0 1>1 r. X a S v C V**I CffC ft zi 0 0 ^ C.N 0 'V}<t]. n % C-*3l OO ^3 O 10 U >% rs . I 4i .VI H0- t: - x t "3 <# O A { VJ.t *.Oi >*1 V, *T! VI :.s 0 0 ij ft-"S ; 8c. t s 0 - P" ijt t- 0 -rft -f * t^8j -- 0 ir:._v.-: A-: C4 r- Ji 0 .10 a 0 DUP040010995 t EMORANDUM DEPARTMENT OF HEALTH, EDUCATION, AND WELFARE .Pl'inUC HI-.ALTM m-itvicK Cf.MJiK fOH mbt-ASi: (ON1OC. NATIONAL INSmUML fCfH OCCUPATIONAL SAFEST,AMl HEALTH : Deputy Director, NIGSH m7*: August 24, 1976 ZH: Chief, Industry-wide Studies Branch, DSHEFS, NrTOSIf^/uJ 1H: Chief, Biometry Section, liVSB, DSHEFS, HIOSH < \Vv-/ w JH: Director., Division of Surveillance, Hazard Evaluations and Field Studies, NIQ5H s Epidemiologist, Biometry- Section, IHSB, DSHEFS, MXGSII CT. Review .of BOIA Lead Chromate Report . The "Epidemiological study of Lead Chromate Plants" prepared by Equitable Environmental Health, Inc. examines the mortality experience of 577 lead . chromate. writers,, The study population is defined a;s .any male worker who had been exposed to lead chromate for at least six months prior to December, 1574. Ninety-five percent of 548 workers' vital status was ascertained and 50 deaths (with certificates] had occurred in this group,. The study population was of necessity divided into three'groups since the workers were employed in three separate plants as follows: Total Plant .1 238 workers ' Plant 2 152 workers Plant 3 , . 158 workers 543 workers Hired before I960 GID/s/ 45 158 ' ` , O 38i Deaths/Certificates .12 deaths .. 5 deaths 33 .deaths 50 deaths The analytical section of this study, despite the small numbers of deaths' involved, show excesses of both respiratory and digestive (primary stomach) cancer. Plants 1 and .2 had exposures solely to lead chromate. Plant 3 had mixed exposures to lead ci-.vmate and to zinc chromate, (refer to attached' table) with .only one worker having sole exposure.to lead .chromate; (He died of lung cancer)"- " _ u?e> iAjarKet-5 of Ima The most apparent, serious deficiency of this study is the short latent periodof this worker cohort. Table .5 of the study shows .it to be 17 years at most. Afurther.. complication ip : that most of the -.workers with the longest latency periods fall into the mixed' exposure group at plant 3, The relatively purer exposure of workers at plants i and .2 is diluted by recent hires, .and short . term employees, resulting in-.a cohort of only 176 workers (total from both plants combined) who were hired prior to 1260. Plant o's total is 158 workers hired before 1960- Of the workers .who were hired before 196.0 .and worked 10 years or more - plants 1 and. 2 total 59, whereas plant 3 has 138, (Refer to attached table.). ,> * **'! be <5-5 m&icsX'b j ffyst t/?e DUPQ40Q10996 Page 2 * fJ-iporaiulMn, Review of. DCMA Lead Chromate Report, August 24, 1075 It is a generally accepted theory that the latency period (time from first exposure ta ths onset: of diso-i.se or death'} for`occupationally induced cancer is usually in the range of 20-55 years. These lead chromate workers have most likely, not had sufficient latency for any significant excesses to appear;-. .... .Other noticeable limitation? * " this dots set arc the "r*remelv .-roll sample sines, as stated above. P-sentlcjiv i n l c a-c trres C'no.t: wr'i n nust be analyzed indi/idually - xerer'to above for numbers e workers by plant. The possibility of Type II Error, or failing to detect a significant difference when one does 2" fart rx.st, siiotiii be conn _rM it ;h" evalua tion of this data set, due to the very small numbers of deaths involved. No age distributio-' for this cohort is presented anywhere in the study. Pacts! dicier;1rice;, are presented as follows: Plant 1 - "overwhelmingly white", Plant 2 - "conservative estimate is that, half of the plant population is non-white". Plant 3 - "is 50% non-white". Nevertheless, the .authors chose to use .the crude death rates (for the state in which each plant was located)to generate expected deaths. The reliability of the SMS's generateo using this type of control population, which is unadjusted :||j^p:s!^&gjii5||l3ni! . -;i|p . and digestive tract cancer is well documented. . .?**#< occwipa-tt+m. / -A-rf ' /Pfey AJft'Afe.rf in '/f"Aisc/*(&**A *fr-yj r ' *" /s~ Afa. isy /? <?rt- excess **f c S-h c ^s s - '/^Z4~fr sbfpetZ* *^,S> . S.em- //J^ ' " y` ' When, the authors used age-sex-race-adjusted rates for the S.M.S.A. for each-` li|j J jtsffifewted ' cbhgiiigBkbliy a':.dfattic. i|||^||ffStglp;: jcp [ t|i^i spil'pect-: Sppij[ifiteipiCpatrqi: popbla- tier! an statistical analysis. Considerable bias may well be present in the costas!. group;; .. ` 1 Ttc/e/far&'jiya *rs s-'&?<er<3dr/i" Another limitation on respiratory cancer inferences is the confirmed smoking ^iltigsy dil'ffelafiiiy.'.all. those; who':iipd!-;4; Jjuhg caftpeif:. iSiiSrprobleni-could probably be resolved if latency were longer and if cell types of lung cancers wsruitgiveni;-- ' In spits of the jabove. methodological problems, an; increasing gradient of the ^P'|>hs|ii(fe^vea' i-tith i-incfeasing. fii|e.j frpti? .irst:":pxpassid# to, lead chromate at plants 1 and 2. The liaised exposure to zinc chromate and lead chromate at pjLjsjbjl debited, in. .cidnsis.cpnt-ly: lelevated.; -for*: TOspiriaRbty and digestive cahter - .eShsjtsicntiV elevated. !pfte to idb-hprt eomsdsitxbfe, h statistical artifact.' If "the records available at plants (prior to. 1957, plant began in. I94i|{. haH)!,-tlfs-sffiij. included in this, study! more: .ojf 'a.grddiebf. in .cancer deaths would probably have shown up in the analysis, similar to that observed in plants 1 and 2. " rf/f h* S'efTT p, CSC. yr/e?/1cS/~c MS*~ fyfr/S- Ht ` *&/ The results of using S.M.S.A.'s for a control population and thereby eprittpllite; fpx Variables such as. age, face,, sex are interestihg. the SMR ehicmabeS ifor Respiratory-dancer-.dropped somewhat, but still shows an .excess. For plant 3, the SMR for digestive cancer elevated to an S-fold excess (fror.r 209.9 to 773,3j).. , rfte SsfX *#?.? /y yos- o'// ta?/y <P*S~s tS'S" C.aes7<=-*" ^ . f$ zn SMR Zfof^aerA Crrr<c/- s>sr/y* ^ > * 9 DUP040010997 Page 3 - Memorandum, Review .of DCMA Lead Chromate Report, August 24, 197-6 The authors stats that since no mention of stomach cancer as a consequence of chromate exposure is made in the scientific literature on zinc chromate and sines no stomach cancer was found in plant 1 and 2, it is unlikely that these stomach cancer deaths are attributable to either of these two exposures. Cr.e purpose of epidemiologic research in the field of occupa tional health is to uncover associations between certain chemicals or compounds and the resultant health status of workers occupationally exposed to these substances'. Therefore, it does not seen justifiable to rule out any association between stomach cancer and the exposures to zinc chromate in riant 3 just because there has been no previous mention in the scientific literature. An 8-fold excess of stomach cancer deaths in any plant can nor be lightly passed over. A more thorough detailed investiga tion of work history and other factors should be dons, as the authors mention* The authors contend that no definitive conclusions can be made from this study, due to the small numbers involved. But, the respiratory cancer excess holds up for both the 12 deaths in plant 1 and the 33 deaths in plant 3, showing a gradient with-increased exposure and latency, which should be invesnigated further. A re-analysis of this data is in order. This reviewer contends that the latency was not long enough, sample size, was- too small end that the control population was not well enough defined for a clear conclusion to be made concerning the exact magnitude of cancer risk being experienced by workers exposed to lead chromate or zinc-leadchromate. A more rigorous data presentation and re-analysis of mortality data should orobably be done in order to clarify the excesses of respiratory and digestive cancer already demonstrated by way of this Cynthia Robinson DUP040010998 SUMMARY OF STUDY miMUM'S t I ! DUP040010959