Document pB5nZp2kGVwKVBNXQrqL3yoD7

FILE NAME: Shipbuilders Council of America (SHIP) DATE: 1967 Apr DOC#: SHIP046 DOCUMENT DESCRIPTION: Medical Journal Article - Mortality Pattern in a Cohort of Asbestos Workers Jo u rn a l of ^ f ^ fccupational MD-462 Mortality Pattern in a Cohort of Asbestos Workers A Stud}? Based on Employment Experience THOMAS F MANCUSO, M D and A.VAS A EL-ATTAR, M B , Cn B .Sc D , Pittsburgh, Pa. IN previou s pa pers, the basic epidemiological workers exposed to asbestos and specific causes approaches to the study of industrial popula- of death i tions, demographic and cohort, were discussed, For epidemiological purposes, die question and their potential contributions and limitations is what combination of resources and methods cited.' *-5 Recently, we have considered several can be brought to bear, in the absence of com facets of long term studies, the dynamic pany data, to establish long-term epidemiolog changes of a cohort and their influence on the ical studies of industrial populations, and to ap- .statistical evaluation of a cohort's progress and pioxmiute e\ ii on mental exposure and the de observation over time, and the concept of the gree of usk for the population under study. In successive cohort approach.4 5 In these studies, the epidemiological evaluation of environmental our objectives have been to improve on the exposures, duration of employment, length and methods of studying industrial populations over continuity of exposure, and piovision of some long periods of time and to cast some light on consistent measurement of employment expo .'particular aspects of the occupational hazard tor sure over the years constitute the initial broad the population at risk. separation of the population under study and a The purpose of the present paper is to carry lout these objectives pertaining to methodology .by', introducing an essential refinement--the 5characteristic of "duration of employment"--to the cohort studies established through the same ,national resource (Social Security records), and .to determine what relationship exists between V-- ^ ----- - From the Deportment of Occupational Health, Graduate method of defining the population more likely to be at risk. This measure of risk--duration of employment--was not included in our earlier study1 because such data were not available at that time. Since then, following exploratory ef forts which adhered to the policy of the Social Security Administration to safeguard the confi dential nature of the data, arrangements xvere completed to obtain information on duration of School-of Public Health, University of Pittsburgh. Pitts- TWrgV ra. V.Supported by Research Grant CA 0T588 from the Na` n^Jnatitutes of Health, V S Public Health Service employment on a stati^tival basis for each mem ber of the cohort. The principal hypothesis being tested is 147 N '- 148 MORTALITY IN ASRESTOS WORKERS OccmmS) whethei the duration of cumulative employ ment experience among workers in the asbestos products industry is related to an increased mortality rate generally for all causes of death and particularly for specific underlying causesof death. W agnei, in his pioneeiing observations of the relationship between mesothelioma and asbestos exposure, focused attention on the biological effects of asbestos, which study has served as a basis for international research and effort per taining to experimental pathology and epidemi ological studies.'"9 The importance of duration of emploMiient in such epidemiological studies (for evaluating the strength of association be tween asbestos and malignancies in longitudinal cohort studies) has been well recognized by pre vious investigators Unfortunately, this type of information has been quite limited and not gen et ally available, consequently, there are only a few cohoit icpoits on asbestos in the literature which include some tvpe of leference to dura tion of employment. Further, in each study, official vital-statistics data of the general popu lation of the country, state, or county were uti lized fm control purposes. The limitations of the use of the genera! population for contiol purposes lias been tecogmzcd Doli1' defined and o b s m e d a cohoit of men emploved m scheduled jobs m an asbestos tex tile plant who had been emploved for 20 veais oi inoir m the s.mie plant He concluded that the avei.igc lung-cancer risk win m the oulei of 10 tunes that experienced by the geneial popu lation and that the iisle had become piogiessiveh less as the dm at ion of employment e\poxme under old dusty conditions decieascd (following the institution oi ceitam legiilatums and contiol measmes at a given point m time) These ohst-iv ations weie tiuthci substantiated by Knox et ul m a continuation of this basic stmlx,` " which levealed a clcueased lung can cel usk subsequent to that point in tune ( 191") In this in.uuiei. the observation ot the effective ness oi the env noninental contiol measmes. \ e the deciease in asbestos cvposme insulting in the deciease ot lung cancel, augmented data on the pievumslx established direct association b e tween lung cancel and asbestos exposme (t- was also appaient from Kuos and Doll's ohscivations that the effectiveness of the em iluumcntal contiol measmes at vaiimis pm ods could have an influence m testing mu piuposed Itvpothi sis It is evident that a 10-vcar pemxl nf nuplov incut exposme in the asbestos indus try during 1920-1939 would be complete! ferent from that of 10 years of employm the same plant during 1940-1959. This was' of the essential reasons for using the gen cohort method for testing this hypothesis, Braun et al ,15 in their study of asbestos in Canada, established a cohort at a given ^ in time (1950), the members of which had more years of previous employment; the in gators observed this cohort for a period of 6 years. The cohort composition was chi . terized by varying lengths of employmentas they related to other factors such as smoking and weighted-exposure categories. An analyst of this study has been m ade,18 citing the infier! ent limitations and the lack of justification foi the negative conclusion derived. ' Mancuso,3 through the use of 1938-39 Social^ Security records, established a company cohorts; of asbestos workers who were observed to mid-4 1960 In this study, the need for data by dura tion and the development of more appropriate' controls for industrial cohort populations w expressed and served as the basis for the present' xtudv . Sc'hkoff et f t l 17 utilized union records to es tablish a cohort of insulation workers as of 1943, m which group the elapsed time from initial exposure was the essential criterion, characterizmg the cohort as those with a history of 20\ xears or longer since the first exposure to asbes tos dust The consistency of employment for tins occupational trade group has been histori cally ieffected in the expression "once a pipe coxcrer. always a pipe coverer." Consequently, the duration of employment may have been extensive Enterline,1'* utilizing the Social Security ap- pioach, established an industry cohort, consist ing of a number of the asbestos-products plants, ot workers who were employed at some time duung 1948-1951 and observed to nud-1963. Cohoits \xeie established m the same manner for cotton textile plants, xxhieh together xvith the specific death lates for the U S xvere uti lized for comparable purposes Enterlme fur ther cited the difficulty of interpreting findings m the absence of data on duration of employ ment. and the limitations of the Social Security data for this purpose. We have been able to m oi come these problems in the present studv The piesent study constitutes an application nf the obset vations denved from our explora tions in methodologv of studv mg industrial pop ulations It differs from piexious prospective April 1*67 MANCLSO & E L -V m R 149 asbestos studies in the basic utilization of a generation cohort and internal contiols, with gradient levels of employment as assesssed through Social Secuntv teeords The pioblem of "adequate contiols has been a major obstacle m industnal health studies and is further accentuated in piospectne longi tudinal studies covering a period of m ain yeais This problem, too. has been consideied in om present study. The primary emphasis and dncction of the cohort study was to 1 Demonstrate the method of utilizing an internal control with a geneiiition eohoit, accoidmg to age and specific lesels of employ ment expenence foi a company under studs (as derived from Social Secuntv data) 2 Esaluate the effects of duration of cumu lative employment expenence on the moitaht) pattern of workers exposed to asbestos, specifi cally in relation to malignant neoplasms, eaidiovasculur diseases, and asbestosis To our .knowledge, the present study is the fust repott to provide such essential data pet taming to all gradients of levels of duration of cumulative employment experience m an asbestos industry, and the relationship to the moitality pattern m a yvell-defined "total" population at risk, fol lowed for a long penod of time (24.5 years) m "which a generation cohort method and internal control is used. Method of Study The study was conducted m the same manner as the previous asbestos study1 m the establish ment of the cohort for 1938 and 1939 through Social Security records, the method of follow up, and the determination of the deceased and .the causes of death. `Data pertaining to sex, color, year of birth, 'date of death, and a complete employment hisory in the asbestos plant (by quarters of each y ear) were m ade available to us by the Social _urity Administration for the living and the deceased from 1938 and 1939 to mid-1964. The Additional data made it possible to classify the [population by sex, age, and cumulative employ m ent expenence during the period of 1938 to inid-1964. The dynamic changes pertaining to ^ga, cumulative duration of employment, and Years of observation in a long-term cohort study ^discussed in a previous paper1) were considered *.the computation of person years and analysis the data. Age The eohoit was classified in age giuups of 15 24 25--44, 45-64. and 65 and over. The date foi which the age-specific death rates weit* cal culated \y.is the date at death, for the age at enhance generation method, the age was calcu lated as of 1940 The limitations and m euts of the \ at ions methods of compat isons. age at deatli and age at enhance,4 seised as a basis for om selection and classification of the piesent cohort In age at 1940 Cumulative Employment Experience Each sex-age gtoup was classified actoulm g to the cumulative employment expenence then acquiK'd dining the penod 1938 timmgh imd1964 mto tlie following 5 levels I HH Hththtf * u i f i l ' t i u m li t t j }n 11? JKP ni in IV X /, tit fih l l f lilt lit r J t H It* r Cnr ) 0 l- 2 0 2 1-70 7 1-12 0 12 1 -1 7 0 17 1-27 V Death l.ites for the selected underlying causes of death weie calculated for each sex-age lex el of cumulative employment expenence and foi Level 11 through V combined Development of an Internal Control In the present study, employees of the 1938 and 1939 cohort who had a cumulative employ ment experience of 2 years or less during the penod 1938 to mid-1964 were selected as an in ternal control. Average annual death rates for employees who had a cumulative employment experience of 2 years or less (Level I) were computed by sex and age group. For each sex and age group of employees having cumulative employment experience of more than 2 years (Levels II, III, etc ), the expected number of deaths was de termined for the selected specific underlying cause of death by applying the average annual death rates of the employees having cumulative employment exjoerience of 2 years or less for that age group. The standard mortality ratios (SM R) were calculated based on the observed and expected number of deaths due to under lying selected causes of death for age at death and age at 1940. " yiux.u.M.i. ' vment experience for an em- pinvee who uhs a h \e by mnl-l9G+ and has a continuous em ploym ent in the asbc&tos p la n t since 1938 is 27 5 y ears lU'.tt 0 | JflJ) %[ *wv ^ u u u ir r>i|j u jjj Ml Ul p J|/Tlj >\ > t|V ( l ) ->r><i|-.K|-r [jiii- ( i ) m i j i ( n f| j u i n n |ilui III ]Ul Il2fif 111 1 m p J 1V M[|l l| Ip III 1l MX (1 ) ^ p p jlts p u u ` (L ) ( D hIMJ|>.Hj>U ( J ) a'lU ^III |i i u q 0) si* |t yi j MI'm |H11 H>pU i *>l >ji > lapoa jjiiuk q ju jp |U j ^ b J 11 SjU J v \ t j - U, jjw p w jii bi HpllJUJ l>| t>Ul * f t * I i dUMJII >>u M||,- >p qiiiM j r in i" jl III *| Vpii| IS tfi 9 -- -- -- I -- I p c 1 15 05 t S5 66 fc 9 9 -- 9 1 u a 61 01 6 IL 69 8 68' 98 8 855 i(>l 18 11V l --- [ -- -- ---- -- -- -- -- -- 1 l '''< V 8 9 5 -- z 1 i -- ;---- -- -- ---- *--- -- 1- 1 1 1 <)~88 T i -- , s. t. z 6 z -- 8 L 1 81 II 8 I 8-91 8 5 1 11 9 9 61 PI c h l z 19 Sh 81 1 1-81 9 s I il 91 l oy iE 8 81 8l -- till 81, S' 11 - 8 5 8 8 S , ---- 0 1 91 51 51 -- 51 51 -- 15 -8 1 U 1V I\ J i 181 QL 1 --- I 51 t b tu i8 zz H(> 89 ot 9S1 eh PI 19 19 -- .1 7 </ tu y / .7) A ti 61 9ii 6U UL 19 ( p / -- L $)Z I IS 01 -- 01 111 8 Ul (It 9 91 N o 51 u 91 i! 8 f 1 it- 91 II 1- p -- U 15 1 i 7 (o 'i- l C l) Al </ / 7 <1 ( t u ,/ / :) if/ /anttfutftfin Jt` Mink 688 195 88 9 -- 9 51 8 6 SiO 8 zz 99 95 M 981 O il il Hi on I SIIVJV 1 7 <) (" : i . ) it / 151 c i r (>1 8 9 5 1 8 1 0 O H v.TU- 1 i i 1 l 1 15 M U i ^ ( if 1 -- t ! 9 () 1 8 1 9 -8 8 ' 5l 91 91 (il 1 S8 [0 18-81 51 S5 h Si 5 O il SO ( 1-8! 151 1 1 1 1 1 151 ill 18 151 11 1 t rn o i r 51 11 - 8 5 15-81 / 7 (. ' l i 7 tf W / U>' * >n > tU.i t*> >i*\ *{ ( I ) <IV 1(1 >11) ( '1 ) JN lA l'l VMJfcJ ( 1\ (1, M i l m ill IS l( [ M II II l l s s \ | ) ) SlIJNIlOyW S. AMVJINOf) SIMMUSy I N( ) 4 0 m a i M I ^ ( I ) I V I O X 4 0 . I l l - w i l l I O. | | <X)I -<HI\ I III' I. T able 2. C ause of D eaiii im 1 9 4 0 - mid-1964 Follow-up of A sulaios-I'j.a m W ohki hs ( i s C i assihi i v T i m ) t * r \ t f trf r lilll tthtftX t <m/dur/riff Mf 4 4 ftt m lit < M utt s f't lim it a ( 'arfx ttf tftitfh unit code IH l ix t h / 11 in n 1 Ut l.i.K 1 II ill n I Toi \L Tuberculosis, all forms (001-020) All neoplasms (140-219) All malignant neoplasms (140-205) Malignant neoplasms of digestive oi gaits and peritoneum (150-159) Malignant neoplasms of peutimeiun a n d ! or mesothelioma! Malignant neoplasms of lung, brum litis, and ti.iclic.i (162-16')) Unspecified in.itignaiil tieoplusms ( 199) All athei neoplasms (140-149, 160 101 164-198 & 200-239) Caidiovaseulai diseases (330-334 & KKI-168) Diseases nl lespiiatois system (470 527) Disejses o f digestive system (530-557) All oilier causes Total Population L and D lit) 10 J>5 84 26 ,8 55 .5 10 1 12 4 5 1 1 39 1265 49 h 61 56 76 n 5 (> -- 1-- i 13 22 H 11 22 11 1 5 21 11 14 22 11 1- 0 37 t 1 T 1 l i 5 9 f 3 It) ft -- 2 -- i - i i >t * 25 to 20 11 ,S > 9 to 11 11 s -- 0 i 2 3 i 0 15 9 i 7 3 121 149 159 7*3 151 22.5 !s 9 5 () -- 1 ------ ------- -- 1 ) 1 1 ! 1 1 1 1 ! 1 1 1 1 1 1 t ____ 1 -- . - ----- ------ ____ 1-- 1 > 1 -- 1 -- 1 ___ ____ 77 1 2 1 t 1 3 ___ _ _ 1 I_ 19 16 =>7 ' I n d i u U s - t n d t \ i d u . i U u t v i l n t t i i , i s h i s l o ' . i s , m i s r c | K > r h > d i h d . i t l i 1 4 r\ i i m i h S St 1 M i l l f M l .. I U M I I I i h i l l ) t 'l l i i s j r r * m p a f n u i i ^ n a n t n< u p l a s i u s is i m l u i h it . in | . i r t n l Utiliormi of jtliu ra i s l i l s t t:I d i p s t n > n r ^ . i i i s i i d 1" III"' I) U l i t J'il tiifitlis. in Ilk na.iii s mu l all tin ili.itlis m tin* ft m a l t s t\ n d m hi .isIm\ I hm % in.! 1 .1 . . I l l . | u . h III) ---------- i r >iinT i-a m M ! m ui i ,Sj,\sinni) Mum u h x K \ j i i > ( S M R ) Ham i>on O ivsehvid ( O ) and E u 'l c ie o ( E ) N u M a tn oe D ea th s D ue to S l l K 111! I ' M i l H I M M ( ' \ l . S I VVli I III, M o t U A I I IT H AT ES ( I t ) P i;|l 1 0 0 ,0 0 0 I 'o i'U LAI ION HJI I A ( l i A1 1 9 4 0 AJUD \i i \i Ui \ iii \ \ i o v , Asmmos WonkuisSrunui> (srs Tk\ t) J i f f o f /*' /f# * `f t * l -/ / 1.11 M \U Ml i *|>l. ii n o j to) Nt.ilijlii.uit m{ <(iu<is { i \ r n i u w** tltl JM'I in o r/)) \l.k f im i.in i IK u p [ (MIIV o| i ilmuMitu ,id ' hi in# miflie lidnid * Vf.thi m\ >l Iuhu 1IfOIKIni'. & u.u lu ( [02 l(v M (. - hi <1 iK'upl.tuns ( MImt m'upl.isMis (1 10-1 to lOU-llil. |f.|-19S 200-21')) 1>|V.|* i s t <* fif r \ , t \ t nl.u s\ stem i h o t o * lo o ir ) 1J im .lu s ill IrspitAllll V \ \ 'll! Ill ( tTO- 727) \ 0 m`Mi >m >. t " > 2 ) 2 } Diseases n f (Iik <'s O%c s\ stem (5 7 1 )-5H 7 ) O iH n tansos A ,, It sx it: \n It SM I! \l> H S \|]t \n H S \tli Nn It NMU Nil H SM I! Nn 11 SMIt 11 SM It \.i II SMH Xu 11 s \m NSi H SM I) No n s \in > > a i' 4' <f i> t: N sT l!7 2 is 5.H -- 25 -- OS (-- ) -- 1 2 -- il i- 1 12 17 1 I.SSO 7 2 s 170 12 0 .) 70S 1 )> 17! 10 1 'J M2 20 .7 10 >,} o t: tij A over c> /; 12(i to S *1102 2.MU 1J7 2fi 12 0 l `)2 o SSti 217 12 4 0 S.S0 2 0 5 .)()() IS 7 1 1109 1 5525 272 4 00 3252 -- (+ ) r:> -9 o li -- OS -- u .s -) -- 0 4 -- 84 (-- ) -- 1 2 -- M (-- ) 1 27 27 OS 10 1 00 y~ __ ( !) 1 00 27 -- (4 ) 2 37) 75 102 71 1 00 TT -- 11-) 10 1 s 270 2f> 0*33 fi (HI M -- (i) 7 5 55 100 70 I 2fi If) JO 7 770 277 210 27 7 7 )27 7! 017 10 0 0 270 -- (+ ) 4 0 0 57 -- (4 ) 27 7.41 728 100 310 ! 20 2 2 1 i IS 170 0 SO Itfil 79 0 U ! 1 00 74 -- (4 ) 1 (10 297 -- (4 ) fifi 22 0 1 9 Iff 10 2 4 700 IT 0 0 1275 -- (+ ) 12 0 0 S80 -- (+ 5 5 00 369 -- (+ ) 12 4 0 886 295 300 2 00 in n -- (+) 1 00 816 -- ( *-) I 00 810 -- (+ ) JO 7 I S i f t '! 5 8 2 5 140 1 00 816 -- (+ ) 3 0.0 2449 -- (+ ) -- 0 4 -- 84 (-- ) " 04 -- 84 -) I u i\A j iv. nu -u i S w lu r v in v .v t\ tin ilitln Tt\ n 'll IIIlie f r o m S S J S o n ly . ` M l Ml nftltfcMit* n r i i.iiiH |inH il 1 pauelne U relh eod#d \ Rroin* o f nnwmlrl|.i." 4n l t . i n io p ! iiti u r^in.-i l'n..ottai iaMi Apt: at tlcath i!J over O E O E O E 3 2 21 3 Jilt 200 451 IJ 6 5 JOT (i 3 168 4 J 0 39 16 245 16S 35 7 2403 940 256 48 17 9 687 256 268 15 9 0 215 128 167 81 4 0 5 14056 8007 174 13 7.7 2256 1345 168 3 00 521 -- (+) 4 00 39 -- (+) 1 00 to -- (+) 2 4 9 19 48 41 4 49 39 48 82 4 33 39 32 J22 2 00 J9 -- (4 -) 4 00 39 -- (+ ) 9 6 .5 87 63 J38 ___... wf- 6 3.0 89 43 201 21 9 0 300 128 234 6 00 86 -- (+) 6 00 86 -- (+ ) 63 38 8 901 556 162 31 0 0 443 -- (+ > 2 6 0.0 372 -- (+ ) 2 00 29 -- (+ ) 24 9.0 3 4 3 ,...128 1 00 174 (+) 5 77 868 1345 65 1 00 174 -- (+> 4 00 694 -- (+ ) 50 34 9 8677 6050 143 6 00 1041 -- (+ ) .3 0 0 521 -- (+ ) 5 0.0 521 -- <+) 7 39 1215 672 A N /ifk i. . -f3i*l: y $ X '1 i Yol. S No. 4 - April 194? M \.\C L 5 0 8 EL-A ITMi i33 Characteristics of the Cohort The asbestos plant had unK white employees in the yeais the colioit was w lock'd. 19 38 oi 1939 (Table 1). About 855 ot them w eir male The age coinpdiisnn shown! a considuiublc minibei of voting employ et s Specifically. mon than two-thirds oi the employers in the colioit m 193S oi 1939 wcie m the a g e gioup 15-34 teats in 1940 During the pound 193S to mid1964, the cumulative employment eypeil e u c i classified by sex and .ige at 1940 toi the cnhoit population is illustrated m Table 1 It wax noted, for example, that about 255 oi the male employees weie m Level 1 2S'. weie in Lewel ii, and 477 weie in the combined Ley els 111 IV, and V of cumuhitiye employ incut eypctience in the asbestos plant During the penod 1940 to mill-1964, theie were 330 males and 31 females m the plant co hort yvho died, excluding 14 yyai casualties death at sea, and 6 deaths dming 1938-1939. The distribution of the numhei of deaths in level of cunmlatire employment expend ice V. among the cohort population is illnstiated in Table 1 For all le\els of cumulative employ ment expeuence, there weie 330 deaths (26 6'c) .in a total male population ot 1265 and theie ,s were only 31 deaths ( 1.3.67) among 228 total female employees IT ^Findings In our study the colioit population icpie'sented the combined cohoits of those employed 'in 1938 and those whose first employment yvas ^recorded by Social Seemitv as being in 1939. ^Information was not available to characterize the members of the cohort bv fiist employment frithe plant prior to 1938 and this, consequent ly, deferred any analysis by the successive co urt m ethod5 Howevei, in our preliminary .analysis of the individual cohorts, employees of 1938 and those of 1939, yve found an apparent " Terence. The 1939 cohort employees, repre- ting mostly new employees, sustained a low , mortality for all causes of death and for the lected specific causes of death under study. VOf the total 1265 male employees in the 1938- cohort, there were 328 male employees ,'at belonged to the 1939 cohort, of rvhich there were 47 observed deaths due to all causes in clu d in g 2 deaths c a u se d by- m a lig n a n t neolasms of lung, trachea, and bionchus, and none ue to asbestosis. Of the lemammg 937 male loyees in the 1938 cohoit, there were 283 deaths including 31 deaths due to asbestosis and 33 deaths due to malignant neoplasms ul (lie Iiimu tiuchea and bionclios In esseuci the )u tti l wnvivallni (Ihisc ol the 19 39 colioit and the nhst i s at i nn t hat none ol die cases ol asbestosis o c t m t e d anion" thosi whose u colds appealed m 1939 as n- mided by Social Secmify when c o m p a n d to those wIiom k colds a ppeal ed m 1938 could ht din to flu iniloeiice ol pitwiniix employ tot lit puoi to 193S in the same plant among the I93S loboit and also because ol the Hlechvetioss oi die e m nonmi iital changes ami lllipi in u n i uts in subse quent seals Tlieielote mu obsi i s .dions on u - diiction oi iisk although limitid .in m agiee- iiieiit with the ubs e n atlons mdicali <1 bv l)o|| a nd Is"m o v r t n l ' ' 11 The subsequent findings {winch ioilow ) t i late to the combined 1938-39 colioit of employ ees ol tin* asbestos pindncts plant Table 2 cll.U actei i/es the numhei of deaths h\ the 11]idei lying selected causes of death as noted bv se\ and level oi cumulative cmplr>vni nt eypeiieucc> Among the males tbeie w tie 35 malig nant tommy ot the lung (425 oi all malignan cies). of winch ~ occulted in Level I (0 1-2 0 cumulative ve.us oi employment) Of the 4 3 deaths due to icspiiutoiv disease, 31 deaths weie due to asbestosis Fm the lemales theie y\eic II deaths due to malignant neoplasms mid of this gioup 6 (547) were due to malignancies of the* lung (one was m Level I), and all of the 8 deaths clue to lespiratoty disease were due to asbestosis Theie weie S malignant neoplasms of the1 peritoneum and/or mesothelioma, iden tified in the 330 male deaths of all causes, 1 of which occuried in Lcs'cl I. Table 2 provides the initial distribution and indication of the influ ence of dmation of employment As was seen in Table 1, the age gioup in each corresponding level varies and consequently influences the ob servations lelative to duration of employment and specific causes of death Table 3 applies the average annual mortality rate by age and undo lying causes of death of the internal control (Level 1) to those with Cumulative years of employment experience of more than 2 years (Levels JI-V inclusive), and as such provides an "ovei view" of the moitalitv patterns of the cohort, without relating to spe cific levels of employment It must be borne in m ind th a t th e in tern al co n tio l (L evel I) is a su b cohort of the same cohort population and differs only in being charactei lzed by the shortest duration of cumulative years of employment ex- i* V iin T ilift iiIT |ii 154 MORTALITY IN ASBESTOS WORKERS Otfcupti perience { ? ........ !"*) in the asbestos indus try during 1938 to mid-1964, In Table 3 the mortality rates are calculated by 2 methods(1) age at entrance, the "generation cohort ap proach," and (2) age at death, die "usual age specific death rates." In Table 3, the SMR calculated hy age at death and age at entrance illustrated higher values as compared to the control. Most of the obseived additional deaths were due to malig nant neoplasms (specifically those of the lung, tiachea, and bronchus, the digestive organs and peritoneum, and also "unspecified malignant neoplasms"), diseases of the cardiovascular sys tem. diseases of the respiratory system (specifi cally asbestosis), and diseases of digestive sys tem The SMR and .iveiage annual moitahlv lutes tor ages 25-44, classified In age at enhance, uas much higher than the age 4.5-64 foi malig nant neoplasms ot the digestive oigans and of the lung This dilleience was moie inaiked foi malignant neoplasms of the lung for age 25-44 and vsas 10 times that of the control group. For the age at death, the highest late tor lung cancer occurred in the age gioup 45-64 The .SMR and m oitjhty Kites foi lung cancer ni tlie oldei age gioup 45-64 (as classified by both age at eiiti.uice and age at death ) lllnstiated tlie effect ot the `intein.il control" on the dilleience between the obseived and the ex pected When (lie expenence ot this age gioup with Ivvel I implov incut expel icute (0 1-2 ve,ns) is applied to the cmlesponding age group to di live the expected i.ite, the expected i.ite would be iugliei Iw vutue nl (lie influence' ni emplov (IK nt pi mi to 10 JS ill this inspect tlieieime tin dilleience obseived constitutes a veiv conservative estimate Effects ot Cumulative Employment Experience on the Mortality Pattern I i>i tin' p u rp o se ot assessing th e iel.itionsliip betvvien the moitalitv pattern and tlie dotation 0 I i m ploi riu nt tlie rates In age at 1910 wen u se d tin th e i ni ions reasons to be discussed To 1\atitim this n latioiislup. moitalitv rates u o r us< (I as Inflow > all causes all neoplasms malig nant ni oplasms of digestive oigans and peutojieum malign.ml neoplasms ot lung bjoncluis and tiachca c.udiov ascul.u diseases (ex cept vawnlai lesions alb'cting the cential luivmis ,vsh in i and .isbestoMs Table f illustiates the obsiiw d the tv pic ted and the aviingt iimn.il mot talitv Kites jit i KM) Ot ill fm sefei terl underlying cause of death and SMR a$ c*` by age at 1940 and level of cumulative e ment experience. The mam description of the result seated in Table 4 is directed at the age 25--44 age at entrance of the generation according to the corresponding cause of i in order to lessen the possible influence of. vious employment. ' r ` *_% All Causes of Death It is noted that, with the exception of l?j and more years of level of cumulative empTmerit experience, there is a consistent inc m the average annual mortality rates as cumulative employment experience increa For example, the average annual mortality rates, due to all causes for age group 25-44 are for 0 1-2.0 years (control group); 1091 for 2.1 7 0 years, 2280 foi 7.1-12 0 yeais, 6788 for 12.1 17 0 years, and .3398 for 17 1 and more years of cumulative years of employment experien The same holds true with SMR values wt in the same age groups, there are values of 207, far 1 1-7 0. 432 for 7.1-12 0. 1287 for 12.1-17.0, and 644 for 17 l and more years of employment expel lence All Neoplasms Table 4 indicates an apparent increase in the uveiago annual moitalitv Kites foi all neoplasms as the cumulative employment experience mc.ieased Fm example, the uveuge annual mort.ilitv Kites foi all neoplasms fm tlie age group 25-41 aie 132 fm 0 1-2 0 vears (control gioup) 348 fm 2 l-i 0 v(Mis 5i0 foi . 1--120 ve.its 3133 foi 12 1-170 ye.tis. and 1081 for [7 1 and moie veais ot cumulative employment expeneute The same holds tivie with die SMR values whete m the same age gioup. SMH values aie 264 (m 2 1-7 0 veais. 432 foi 7 1-12 0 vc.ns 2 376 fm 12.1-170. and 819 fm 17 I and mme veais of cumulative emplov ment Malignant Neoplasms--Digestive Organ* and Peritoneum llieie uas a total of 25 deaths m the age gioup 2-5-61 as of 1910 due to malignant neo plasms of digestive oigans and peritoneum mile 3 deaths ucclined m the control gioup ,'0 1-20 veais of cumulative emplov ment rvpeiicm r) and 22 deaths oceuned m the highet I. vi Is Because ot tlie veiv small number of c.isis Im tach level ol cimmlativo emplov meat ami agi gioup it is difficult to demnustiote a wu -Yet. I * 1* < 'Apnl IW NHMASO ii LL-MTAH I'- relationship between the- motl-difx i.dcx bom Ineopiasms of digestive nig.uix and pm tonnm i and the cumulative em plm m nit cxpriionic The aieiage annual rooit.ihts ink's. however viett: higher for aiJ the cumulative emplox ment experience of any age gionp than loi the contiol group. Foi example die avci.ige humt.ditx i.itex foi employees with 7 1-12 0 u m i\ ut <.minilativ < employment experience wax 12 times that ot the coiitiol foi the same ag ginup 25-44 vi.ijx Malignant Neoplasm*--Lung. Bronchus, anti Trachea As is noted in Table 4 the mueased mnifalffx rates in association with the iiu.ie.isc of cmmil.,tive emploxment expeneiue is app.uent lm the age group 25-44 yo.nx e g i.ttes ot 26 lor <> 1 2.0 yeais, lates of 156 hn 2 1-7 0 yeuix, !6> lot 7 1 12 0 years, and 1567 hu 12 1-170 xeais There was, howexei, a decline in the moitahtx Tate for 17 1 and mine xeais of eiimulatixe em ployment expeitend* It is evident that the late and risk for lung cancel does nutcase with ^duration of employment The clegtee ol nsl: iiappears highest after emplox ment ol 12 xi.ux more At this stage of on mvestigahon, wc mot attach undue significance to the rate foi Hung cancel for Level 1, since the data and spe cifically the Level-I employees' emplox ment aie lited to that which occurred after 1938, and ye do not know what specific earlier yeat vould represent the actual first yeai of employ ment in the company--in contrast to the year as ^utilized in this study foi enhance into the coJ'hort. However, Level l {0 1-2 years) for age group 25--44 can he considered a possible over estim ate of the lung-cancer kite of individuals this age group and level of employment e\ience. The consistency of the increase in ate, recognizing the built-in factors for sticcesye levels of increasing employment, piovides additional support to the relationship of lung ancer and asbestos exposure. Cardiovascular Diseases l^Jt is noted from Table 4 that there is a con- ent increase in the ax'eiage annual mortality iltes as the cumulative employment experience picreas ...4.aple, the average annual ttortah'ty rates due to eatdiovascular diseases' jtcept vascular lesions affecting cential nerv- --- ' r-- - - 35-44 are 211 for -A ' ' - 651 for 7 1 - ... md S49 for 11 I .11ill l l l e l C X(Ml x o l t ll II III 1.1fix < i UijliriX Il i e u I xpciniuc Tin viiiii1 p u tm c holds Him xxiili 5MR x.dues xxheir in the same .igc 'jump then .in' xi|i11"i nl J >2 loi 2 1 7 <i \ i .n s >04 )i 7 1 12 0 x i.us {45 loi 12 1-17 0 xc.iix and 40} lm IT I .md IIIOIC xc.ils ol (.iiiiiul.it I' i ciiiplox IllCtU expel ll'licc Asbestos is In T.d>l( 1 excluding die "limp xxit 11 IIn !<\ <I of uiiniil.ttix c emplnv mciit expe?inni ol 17 1 .md nioic xi.ii- die avri.igt .imm.il nioitalilx i.lies and S \|R xalm x nil 11 ,ixe xxlib fIn .lx( of the i imil ll.lIIX( imploxmilit (x p iitiiiit I>m deaths due to nxbcxinxk I in i x.unplc lot agt ginnp 25--14 latex dm to ,ixh( stoxix xxiu 0 i/c m ) loi 0 1-2 0 xc.ux 116 loi 2 1-7 0 xc.iix 24 4 foi 7 I -! 2 0 xcai x 2084 lm !2 !--I, 0 xi.ux and 540 loi [7 I and oioh xi.ux ul cinnul.ilix c emplox ment expet iciu c ll lx of intcicst to noli dial niiitiiii' tin total tl obxcni d dcathx due to dixeuxe of k xpii.ifoix xwtem {Ioi .ill age gioupx ol cmplmicx xxito have woiktd moie than 2 xc.iix m tin axbcxtox nidiixtix I (itcic u ric >1 dcathx dot to axlii x(o- six, 5 deaths due to chtonte mteistittul pn< u- tiuima. and ! death tine tu mIuoxix--a total of 47 deaths Excluding these 57 deaths limit the 4l ohsi'tvcd deaths the lrruamdei will he 4 observed versus 3 7 expected fax wax denxcd by applying the axmage annual moitahtx tate of internal contiol) This is an illustiation ot the fact that the excess in moitahtx among em ployees with piolongctl cumulative seats ol emplox ment experience tx a tine significant ob- seixation and not due to anv Inns of selection m the internal contiol Table 4 also pi ox ides iate and SMR by age gionp, in which the cohoit is diaractei i/ed as of 2.1 or more xeais of employment expeiience m comparison to Lose! f ( 1-2 veins) The cohoit is also characterized as "all levels,'' i e , it includes Level I, which also compares the total with Level I as the internal contiol. These two additional characterizations of the cohoit {2 I and more, and all levels) serve to illustrate the averaging effect m the former and fhe dilu tion effect tn the latter, which occurs in contiast to the rates observed for specific levels of em ployment experience foi the corresponding age gi oups. Discussion Although we have oveicome some of the limitations of our previous asbestos study'1 and 3=- *VMrlfetaa vttl t I S i W i i i l i l i M o t u u II S i U l t t > U \ M ) J i > \ OlSM H \t U ( O ) -\\1> L \ P i C l LI) ( E ) N U M H L Ii Of- D j . V l t l i s O u t . 1 0 Sk.Uh.CIM> I 'mii hi . i\( ( u.M (mi \W.ittun-, \ i \ i i s ( N) 5*111 lUMXW l'ui-oi t. 1 ion Auonc. Couom Pts.ciiiUi in Tfcvr nt't>ihthr> t iHplwiiH Hi t f i n n i M C (yr ) "/ U / ," / l tl t; i no t 7 $ aver J` f ^ MOrB A ll levels 1'/ fi~ f f t'>i*l*`i U *t 1 u i: t $ i: <> n t> H a K 17-2 1 r>-11 i5-i> i \n It SMU N.i U su it v. ' it SMU No U SMU : 2 is --- 'll 528 ____ 1>J 2S<>! -- t 5S25 -- 1 U2 17 llfill U 77u2 l) IfiMiT 6 7 2 is Ml 22 7 >2.s 2\ >7 11 1 2SU-1 27'. 2 1 *>S2"j 2-S0 \\ 1 < STM<1 t 2 is 2 8 22S1) 2 j SS65 7 Hi IT I I n 2 is Kill b => >2S 1)2 7 `J 2s<> 1 <36 2 r> 5825 2s i (* U 1 -- 21S -- 11 J 0 (>78S 52S 1287 )9 8 1 11518 2804 482 1 2 2 Klttfi? >825 18) 0 0 8 -- 2)8 -- 44 6 8 3398 528 644 31 10 7 8126 2804 290 1 04 15185 58-25 264 5 1 17 132 1880 126 9302 18 14694 a 7 238 58 37.1 528 356 .39 8 2804 317 7 1 5825 252 12 182 152 1406 145 7136 21 12069 157 238 77 57.1 528 266 570 2804 254 10 J7 5825 207 a i i . s m >I`I.a s m s (140-239) 25-14 15-04 No H SMR No 1) SMU 5 1> 5 7 7 1 6 6 0 3 14 17 42 93 47 14 3 nz US 132 570 132 3133 132 1081 132 598 132 435 132 -- 21.4 4 12 2376 819 453 329 . 6 (. 36 a 25 6 26 8 34 20 120 32 .. 18.0 886 1191 886 2128 886 2080 886 2097 888 1920 886 -- 168 240 235 237 2J.7 , Ni. It SMII N' r. MM U \,VM A'.Mv-mCUTIVU OltLAN'S AND r E M X W f ^ V U ( 2 ^ - 1 5 9 ^ I 1 l I o.i 0 0.1 , _!U utt aa ~ - - ?TJ*4 T T Wl ->> IV1l 295 ~>Mh 2.15 - 1\)5 a&e ' k * MjUUGNAHT N EOPL&lfoS---- LL'NC, BBONF H I'S. 1 BAt 1IFA ( 1 6 2 - 1 0 ) ) 2 5 -4 4 ' '.No:1' ` -1 R 26 SMR -- 45-64 No.' 4 R 590 SMR -- 25-44 45-04 No. 8 R 211 SMR -- No 9 R 1328 SMR -- 25-44 No _ R -- 45-64 No -- R -- 8 11 186 20 715 1 2.4 248 590 42 12 91 279 211 132 11 53 2734 1328 206 5 ------ - 116 -- 1 -- 248 -- 2 03 163 26 017 3 17 1064 590 180 .3 0 1 1567 26 5939 I 17 317 590 59 6 03 461 20 1781 4 23 1048 590 17S CARDIOVASCULAR DISEASES (400-468)* 8 26 651 211 309 11 37 3901 1328 294 O 04 1044 211 49 > 20 38 7279 112S 548 11 27 819 211 40) 17 5 1 1456 1 128 3)5 Asncsrosis (523 2) 3 -- 4 -- 7 -- 244 -- 2089 -- 510 -- 5 -- 5-- 1 -- 1773 -- 1713 -- 262 -- 19 1 S 270 26 1038 9 82 664 590 109 11 N 8 170 211 221 59 ISO 4 r,(> 112S )2S 19 -- 270 -- 12 SV> -- *Kx< )l vft^cnlnr lesion* afleet mg C S S 20 2 8 185 26 713 l) 120 040 590 108 41 22 8 379 211 ISO 68 27 0 3 )4I5 | )2S 252 19 -- 176 -- 12 -- 591 -- I P WftVw 13W lo o MOi. '> WOlXKtts some of the problems inherent to prc., studies m general, there still remains the ba.,.limitation of establishing more precisely the population at risk, as defined by department and occupation m the utilization of the Social Se-. curtly approach in establishing company co horts The total population of a company may not be at risk. It is more likely that a smaller group of employees in specific departments and/or operations rather than those throughout the plant would be more directly exposed Con sequently, the asbestos-company cohoit utiliz ing the Social Security apptoach constitutes a conservative estimate of iisk because of the dilu tion factor of including all employees. This `dilution factor' was present also and probablv was augmented m Enterlinc's asbestos industry study, which combined a number of asbestos companies 1' It must be icahzed that employment data prior to 19.3b are not as tillable from Social Seciiutv records and tins unknown length of em ployment prior to 1933 could have some heating on the interpretation of observations. It is well kilimn that poor employment m the same oi m a different plant depends on the age of the em ployee m 1949. Therefore, tin example an emploxee whose age is 1-8 by 1910 is assumed to luxe no thdiistii.ii einplovinent pum to 193S and an eniphnie whose age is 4b In 1940 is assumed lo have about 30 xeais ol emploxinent pnoi to 19 >S since age 1*S is cuiisuleied to he the fust age of actixe employment The classi fication ul (he sohoit In lex el of cumulative emplm mm n f expel n ute u Inch is based on the emploMueiit lei old dm mg 1933 to mid-1964 docs not th<i<!oie lepnsent the actual level ot ummlativi <mploy meiit expel lence since tlie Inst xeai ol ein|)lox meiit m the asbestos plant could not lie detu innieil Tins is specific.ills line loi the oldei age gionps, as ot 1940 I'm . .ic.1l Ii m I ol euiplox meiit the classification foi (lie \ oiing age gionp ( ie 17-24! is considcied lo lie the actual level ol t mplov meiit in the aslnstos plant since tins age gionp has no emplox mint pi mi lo !<)3S Couseipieiitlv the apjihcatiou <4 l.evel I as an mtiinnl contiol loi the age group 15-24 pictvides the actual compal ison between the vaiiuns levels of cumula tive eniplovinuit expenence In the olclei age gi mips lumrvi'i, the i ompai attv<pic tin e ts chi li u ut 1oi example in flu- age giimp 45-04. lit. ie is a jxntttl of wuknovwi vuiptoxiweiit pilot io I93S winch could axmage to 27 xeais (4.3 -- IS - 27 I f i \ i I 1 loi tfie age gionp 17-6 { ' " ' a c u u> G.i-2.0 years of cum'' 1.. TnAni Bxmerience during the period mid-1964, could actually represent a hi level of employment when the unknown*' ployment nnnr tn 1938 [S also considered. The age group 25-44 lies between the'* age groups, 15-24 and 45-64, relative to 7, employment experience prior to 1938, . effects on the employment classification are f sented. Consequently, any comparison be' Level 1 and other levels for the older age gjr' * represent a conservative comparison. In i when the employment prior to 1938 is acH then the Level 1 for age 15-24 could be sidered to be the '`actual" Level I, and for agp* 25--14, when this prior employment is addedj Level I could be a mixture of the equivalent oh Levels I. II, and III, finally, for age 45-64,,' Level I could be a mixture of Levels III, IV; and V. The mortality rates for Level I used in', the study, therefore, would for the older age groups constitute a rate which represents that: of a level of employment that exceeds the limits of the designated employment classification Foi age group 45-64, this high rate of Level ft ts earned through as the control comparison in the development of the expected rates and SMR m each succeeding level of employment experi ence and for each ol the various categories of cause of death For these reasons. we have focused on the geueiation cohoit utilizing the age group 25-44 at cufiance as a homogeneous cohort, which pun ulex a mote representative indication of the [iiogiessiou and changes of iate according to v,nmux levels of employment and years of expei lenec Since there still remains the limitation of not knowing flit* einplovinent prior to 1938, (he Kites denied will, in tins respect, constitute a umseixatixe estimate lit this method of es tablishing moituhtx late's, however, the use of in internal control ot the same age gionp (age at enhance) equalizes the previous einplovinent lac tin In the age-at-entiunce method, all of the population oi the specified age gunip will grow uliiet be a constant nuntbei of years during the peime! of obseixation {1940 to nud-1964) Tins aging process, however, would not influence the iesuits since the SMR and the expected nurnbit ot deaths are cleiixed by applying the a\cr ane .iiiini.il niortalitx Kites fiom the mtcm.il uuiUot ot the same age, which is treated am! obseixed m the same maimer Thi' mean dmutiftn of employment and the V0I. 9 No. 9 Apnl i W MANCl'SO 4- EL-ATTAR 159 degree of en\ ironmentnl expoviie prior to 19% would be the same fot any smnlai age gioup and different for tanous age gioups as clarified by age at 1940. Similar!}, previous medical caw* prior to 1940 would then be the same as well as for subsequent sears, but will van with each age gioup. In this method, cfassthmg the u>hoit by age as of 1940 and followed for 24 5 sears provides a complete mortality pattern of a generation of each corresponding age gmup during the period of observation The age-at- entiance (generation cohoit) method ptm ides a more piecisc measutement hv icstrictmg all the deaths to membeis of the coi responding geneiation at usk who have eonnmm chaiactei- istics pertaining to the veai of hut!), period of death, changes m eunnmmeot.il exposure, and previous employment exposure as well as changes m medical eaic and subsequent levels of employment espenence. ft is e\ tdent that au internal control, such as Level 1, applied to the generation cohoit method, does not have the limitations of age at death method and has a f.number of inherent advantages1 ` In this respect, too, follow-up studies on eon- -firmatien of primary causes of death and sup plem entary histological examinations when ap plied to the cohoit as a whole, include the sub cohort designated as the internal control This ^overcomes some of the problems previously ev- iperienced of the inability to apply the same fort of confirmation of diagnosis to the study /and control groups because of size, composition, r;or location of the control group It was noted fiom the data presented that, with the exception of Level V of cumulative ^employment experience, there is a consistent ^increase in the average annual mortality rates la s the cumulative employment experience in- ccreases. This relationship was more apparent ffor age 25-44 than in other age groups for the ans explained above. Level V was an ex- tfption to the rule. In fact, the mortality rates r. Level V were higher than those for Level 1 j^bevel II, but were not higher than that of ayel' IV. This drop of the mortality rate for V as compared to that m Level IV could i due'tb one or more of the following reasons, employment experience for Level V differs ithat of Level IV by the additional employ ; history during 1955-1964, which repre- `a period with much better environmc"*'-' t?ol; those who are classified as Leve1 * ipcohort represent those with belt- office employees or sab----- ' ' mortality picture lot Level V was observed for the penod 1955 to und-1964 .is compared to the penod of 1950 to miil-1964 for Level iV ,ts a consequence of the dynamic changes of the cohort * The present studs ptovules Imtbei confirm.!' tion of the high iisk of cancel of selected sites in asbestos ptoduct workvis and also tuiiuslies the basis of an association of uuieased nsk fiom cancer with the gradient of emplmnu nt expen- ence (Table 4). As was indicated, tim e vs.is a consistent metcast* in the aveiage mortality iate foi all causes of death m geneia! and specificallv lot malignant neoplasms tauliovasiulai ihs- eases, and asln'stosis m association with mote ptolonged emplovmeut m asbestos mdustrx Tins met eased mottabtv iate was evident fot all c.mcei m genet a! and spec ificalh foi cancel of lung, bioncluis. and tiachea and digestive or gans, Howevei, m evaluating malignant neo plasms In (filiation of employment, the problem of competing usk still holds a pet son who was exposed to asbestos dust could die Irom asbestosis and/or cauliov.iscular diseases and/or ma lignant neoplasms, Theiefoie, an important factor which inHu- eiued the extent of the obsenations lelative to the strength of the association m this study was the question of competitixe risks in the analysis of the causes of death '* The significance of the frequency of asbestosis ns a cause of death, and its occurrence as a secondary cause of death, was noted m mir earlier study. Deaths at eai her ages lessen the number of sumxors, who may live long enough to meet the requirements of the latent penod, to develop lung cancer or mesothelioma Env irormiental exposure and dust concentrations were undoubtedly higher in the earlier years of manufacture and prior to the establishment of the cohort and chrvsotde was used almost exclusively.18 18 Consequently, more workers developed asbestosis more readily and died at an earlier age and this has lessened the extent of the findings relative to malignancy. Undoubtedly, continual long-term observation of the present cohort, with a further extension of the latent period, would be most appropriate, and in particular for the age group 15-24 and 25-44 at entrance to the generation cohort. ~ ' ,.,esent study, there was an unusually high number of malignancies specified on the ''-rth certificate as "primary site unknown" 8 i > m mates employed 2 or more years. 'x'~ ' ` 'irute queries, to obtain fur- histological re- 160 MORTALITY 3N ASBESTOS WORKERS view of these cases This situation was unfortu nate because the resolution of these "unknown" cases might have had some bearing on the obseivations pertaining to cancer of the diges tive system, stomach, colon, and rectum or some other site. In our first study, such a histological review added to the number of lung cancers and detected several mesotheliomas. There were 9 cases of mesothelioma of the peritoneum and pleura, which occurred among 3S1 deaths of the cohort (2.5'?). Hammond et ai. observed 10 mesotheliomas (peritoneum and pleura) among 307 deaths. For malignant neoplasms of the central nervous system, there was 1 subject among the females who died m 1939 and 3 among the males who died between 1940 and mid-1964 There was 1 death due to lymphn- saiconu and 4 deaths due to leukemia observed in the male gioup. Further interpretation of ihese malignant sites was deferred. Of interest, too, was the higher mortality rate than that expected for suicides and accidents and homicides combined There were S deaths due to suicides and 15 deaths due to accidents ami homicides in the total white male eohoit. Ail of the suicides except 1 were employed m the asliestos inchtxtiy tor more than 2 yeais. Useful siipplcinentan epidemiological mfoimation relative to disease patterns can be ob tained In consulting the multiple causes of death and then sequence Foi example, m oiu present studs, included m the 142 deaths of the cnidmv ascul.u svstem illO -554 and 400-463) were II subjects in whom asheslosis was indi cated on the death certificate as a secondare cause of death and 1 with cancel of the large intestine ,is a secondare cause ot death There* were* 6 cus< s coded pi miai v as toi pulmonale, in 4 at v huh asbestosis was indicated as secondare c.tuse of the subjects' death We obsened m the fiist studv tin* contubution denied fiom fmther inquire and how the supplementalv infoimation would have altered the sequence of coding of the cause of death as lecoided mi the death certificates It was not possible m this extension of urn onginal shk to obtain fuitliei inhumation to amplify adequatek the relationship:* between secondare and pi unary causes of death To a ceitam extent, the problem of multiple causes of death masked the piopei mteipietation of the caidiovascidar problem observed m oiu asbestos studv. Enleiline1' indicated a significantly high moitalitv fiom lupci tensive heatt f!iseure among asliestos precincts woikeis as compared to that of United States and ers m other industries. Our findings did ini cardiovascular diseases (excluding vi siOfis affecting the central nervous system), | associated with employment, as was evidepfc by the consistent increased mortality rate^re more prolonged duration of employment in asbestos industry. The risk of dying fronts diovascular disease was 2.5 times that of'' internal control for age 45-64 (as of 1940). posure to asbestos dust, therefore, might s the cardiovascular system either directly. ( to an unknown mechanism) and/or in (due to cor pulmonale and/or right heart' ure) In the present study, there was a total of deaths in which the primary cause of death was due to asbestosis-31 m males and 8 in females!, These usbestosis deaths, coded as a primary cause of death, provided a higher mortality rate in association with the increased level ot eui-,-.i ployment experience m the asbestos industry. 5 In addition, there were 21 deaths in whichgf asbestosis appeared as a secondary cause of death. As indicated before, there was no at1*? tempt to utilize the secondary cause of death in the analysis of the findings. O'Donnell,''' a pathologist who had initially observed a high association between asbestosis and lung cancer among the employees of this plant, has now iepotted--based upon the admis A sions to only 1 of 3 mam hospitals m the comnmmtv area--55 cases of pioved asbestosis with 28 malignant neoplasms 23 bronchogenic and 5 mesotheliomas of the jientoneum and pteurad" Tins is of particular interest to out study be cause of the 23 lung-cancer cases Only 12 em ployees could be identified as being m our tc)3S-lcJ3c) cohort, a significant factor when considered in light of the finding that of the 12 matched deaths of our cohort as compared with O'Doi moll's lung-cancer cases, only 10 were so classified m. om cohort and 2 were classified as deaths from other causes These additional 2 lung-cancer cases had accumulative employ ment evpeueuce of Levels II and IV, when these are added to the cm responding levels of employment experience of the cohort, the mort-ditv late shown in Table 4 would be raised corresponclmgk and the tot.Vi W .g cancers for the white males increased to .*? ~ total cohort. This lack of identification i nnc-iii'th of the matched cas valence ot the nuclei estimate 160 MORTALITY JN ASBESTOS WORKERS t -M* view of these cases This situation was unfortu nate because the resolution of these "unknown" cases might have had some bearing on the obseixattons pertaining to cancer of the diges tive system, stomach, colon, and rectum or some other site. In our first study, such a histological review added to the number of lung cancers and detected several mesotheliomas. There were 9 cases of mesothelioma of the peritoneum and pleura, which occurred among 3S1 deaths of the cohort (2.5?). Hammond et ah observed 10 mesotheliomas (peritoneum and pleura) among 307 deaths. For malignant neoplasms of the central nervous system, there was 1 subject among the females who died in 1939 and 3 among the males who died between 1940 and nud-1964 There was 1 death due to lymphnsaicoma and 4 deaths due to leukemia observed m the male gioitp. Further mterpietation of these malignant sites was deferred. Of mteicst. too, was the higher mortality rate than that expected for suicides and accidents and homicides combined There were S deaths due to suicides and 15 deaths due to accidents and homicides m the total white male eolioit. Ail of the suicides except 1 were employed m the asbestos industiy tor mine than 2 ycais. Useful supplemental \ epidemiological mfoimation ud.dixe to disease patterns can lie ob tained In consuleimg the multiple causes of death and then setjiieuce Foi example, m out pieseot stmlx, included m the 142 deaths of the c.udioxasciil.ii sxstem i 110-534 and 400-463) were 11 subjects in whom asbestos's was indi cated on the death ceitificate as a stcoudaix cause of death ami 1 with cancel of the large intestine as a secomlaix cause of death These weie 6 cus< s coded pi tinai \ as coi pulmonale, in 4 ot whuh asbe.stosis was indicated as secondaix cause of tile su b le ts' death U'e obseixed m the fiist stndx tin1cootubutioo dom ed fiom Outlier mquuv and how the supplementalx mfoiuiation would have altered the sequence of coding of the cause of death as leeouled on the death certificates it was not possible in this extension uf out oiiginal stndx to obtain tmthei information to amplify adequutdx the lelationsliips Isetween sccondaix and pi unary causes of death To a ceitam extent, the problem of multiple causes of death masked the ptopei mteipietation of the caidioxnscular problem obserxed m oui asbestos stuck, Eutciline1' indicated a s i g m f i t antlx high moitalitx fiom Inpei tensee heait dtsc-u-e among asbestos pmciuits wotkeis as compared to that of United States and ers m other industries. Our findings did in cardiovascular diseases (excluding va sions affecting the central nervous system), n associated with employment, as was evid by the consistent increased mortality rate y more pioionged duration of employment in1 asbestos industry. The risk of dying ronv~ tiiovascular disease was 2.5 times that oF internal control for age 45-64 (as of 1940);* posure to asbestos dust, therefore, might a the cardiovascular system either directly. ( to an unknown mechanism) and/or inc (due to cor pulmonale and/or right heart-f ure) In the present study, there was a total of deaths in which the primary cause of death was. due to asbestosis-31 m males and S in females.',. These asbestosis deaths, coded as a primary,cause of death, provided a higher mortality ram' m association with the increased level ot eui-. ployment experience in the asbestos industry. In addition, there were 21 deaths in whichd, asbestosis appeared as a secondary cause ofA death. As indicated before, there was no a t tempt to utilize the secondary cause of death in the analysis of the findings. O'Donnell,t!l a pathologist who had initially obserxed a high association between asbestosis and lung cancer among the employees of this plant, has now icpoi ted--based upon the admis sions to only I of 3 mam hospitals in the com munity area--55 cases of piox'ed asbestosis with 26 malignant neoplasms 23 bronchogenic and 5 mesotheliomas of the pentoneum and pleura-*0 Th is is of particular interest to om study be muse of the 23 lung-cancer cases Only 12 em ployees could be identified as being in our [938-19 39 cohott. a significant factor when unisideied in light of the finding that of the 12 matched deaths of our cohort as compared xvith O'Donnell's lung-cancer cases, only 10 xvere so classified m om cohoit and 2 were classified as deaths from other causes These additional 2 lung-cancer cases had .iccumulatixe employ ment expellence of Levels II and IV, when these aie added to the cot responding levels of employment experience of the cohort, the mor. t ibtx late shown in Table 4 would be raised Lorrespondingk and the tot.:' ejneers for the white males increased to total cohort. This lack of irlentificatior inic-sis th of the matched c t \ idencc of the nndciestm . Vol 9 No 4 April 1967 \ 1 A \ C I S 0 & L 'L -V rT A H 161 in our cohort, particularlv when it lealim l that the observations made on the total matched cases (related to one-fourth ot the lung cancels m our asbestos cohoit) constitute admissions to only one of the hospitals. O'Donnell, moreover, reported only cases ot respiratoiv cancer that were associated with asbestosis Although we o b sm ed 9 cases ot mesotheli oma, 1 of the pleura and 8 of the pentomuim, O'Donnell reported 3 additional cases of meso thelioma m em ployes of the same plant (which workers were not part of our 1938-1939 cohoit) Another death due to mesothelioma of the peritoneum occurred in a fonuei employee who was not m our cohoit and uas not in O'Donnells group of cases of mesothelioma This total of 13 cases of mesothelioma (all of employees from a single plant m a small com munity engaged in the manufactm e of the same asbestos products) provides a very strong asso ciation for such a raie disease, in which all ex cept 1 of the 10 mesothelioma of our study weic employed for more than 2 sears In general, the pathological confiimatum by O'Donnell of lung cancers and mesothelioma among former employees of the same plant (from which plant our cohort was established, over the same period of time--1940-1965) fur ther substantiates the lelationship of malignan c ie s and asbestos exposure, observed in our co hort study of employees classified by duration of employment. Summary and Conclusions A cohort study of 1265 white males and 228 `.white females who were employed m a plant manufacturing asbestos products m 1938-1939 Lwas established through Social Security records ad observed to mid-1984. The cohort study represents a series of refine- its in methodology pertaining to prospective Epidemiological studies of industrial popula- aris. The Social Security cohort approach has pen developed to include the classification of ^members of a cohort by duration of emplov- it from 1938 m the plant under study. A thod also is demonstrated of establishing a bmogeneous generation cohort and an internal atrol according to age and specific levels of upmyment experience for a company under iidy, as derived from Social Security. -was observed in this study that when the 1533 considered as individual s, that the xwu one had a lower mortality exprnence than the 1938 one which was prinninh iiillnenced lv the years of employment prior to 1918 Fuither. the 1939 cohort observa tion pim ided an indication of the effectiveness of environmental changes and control measures m subsequent \ears. For the purposes ot our studs, the combined cohort of 193S-1939 was utih/.ed and the levels of accunuilatise employ ment experience from 1938 to nml-}964 and were classified as follows Lev el 1. 0,1-2 0 vears, Lev el 11,2 I-T 0 years, Level IIi 7 1-12 0 ve.us, Level IV. 12 0-170 yeais: anil Le\ el V, 17.1 27.5 yeais Ol employees in the 1938 and 1939 cohort that had a cumulative employment experience of 2 yeais oi less dining (he pet tod oi 1938 to mid-1964. Level 1 was utilized as an internal um tiol classified by age Tlieie was a total of 330 deaths among the white males and 31 lor the females Among the males theie were 35 malignant tumrns of the lung and 43 deaths due to respiratory disease, 31 of which weie due to asbestosis. Among the females, ot the 11 deaths due to malignancies, 6 were due to malignancv of the lung and all 8 deaths due to respuatois diseases were due to asbestosis. There were 9 malignant neoplasms of the peritoneum and/or mesothelioma, 8 in the males and 1 m the females Among the 71 malignancies for the white males employed 2 or more years, 8 were specified on the death certificate as "primary site unknown." The relationship of cumulative employment evpei tente on the mortality pattern was as sessed foi all causes, all neoplasms, malignant neoplasms ot the digestive organs and peutoiieum, malignant neoplasms ot lung, diseases of cardiovascular system (except vascular le sions affecting central nervous system), and asbestosis There was a consistent met ease in the average mortality rates as the cumulative employment experience increased This was true for each of these categories with the exception of the digestive system, which involved a small er number of cases and analysis could not be made The problem of '`competing'' nsks of deaths due to asbestosis, cardiovascular disease and/or malignant neoplasms, and the relationship of primary and secondary causes of death are also discussed. It is concluded that the duration of cumulative employment experience among as bestos workers in the asbestos-products com pany was directly related to a general increased mortality rate for all causes of death and spe- iM iflfai I*rw1 cifically (or malignant acuyunu^, lar diseases, and asbestosis. . -- . jiij intemaHoad'i C' i!J Health, Wien, Sept ^ 6, Wacneh, J, C Epidemiology of diffare _ 11 Addendum hal tumors evidence of an association in South Africa and the United fCing A few months after the close of the cohort follow-up in mid-1964, I t deaths among the ,Vy Acad Sci 132 575, 1985. 7 Report of the U I C C (Intenwti Against Cancer), Working Croup on . 1938-1939 cohort were reported. The causes of Cancer Arch Environ Health 11:221, these t l deaths follow l, lymphosarcoma of the lung, 6. cardiovascular diseases (including 1 case of cerebral embolism), 1, pulmonary fibro sis; 2, asbestosis; and 1, suicide. All of the de 8 CtLsow, J C. Health hazards of asbent Soc Occupat Med 16 62, 1968. 9, Chios', J D. Problems and changing hazards of exposure to asbestw?) Acad Sa 132 896, 1965. ,-x.' ^ ceased were employed in the same asbestos plant for more than 7 sears since 1938 10 Meters, R J Mortality of workers insured-) 4 OASDt for the year 1955. Social, tin, July 1981 H.E W., Washington, D. Ct ^ A cknow ledgm ent 11 Ciocco. A,, Mancuso, T F., and Thompson, Four years mortality experience of a-s We wish to thank Dr Antonio Ciocco. Pro fessor and Head of the Department of Biosta the United States working population! -/ Public Health 55 587. 1965. Entehline, P Estimation of expected rate* in , tistics. G r a d u a t e School of Public Health, Um- veisity of Pittsburgh, for his aduce and encour agement, ('iiiurufij of Pitidjurph Graduate School of Public Health Department of Occupational Health Pittsburgh, Pa 15213 cupotional disease epidemiology Public Hei Rep 79.073. 1964 __ n Doll, R Mortality from /ung cancer in asbeMntr? workers Brit } Irulustr Med 22 81, 1955. 14 Knot, J. F . Doll, R. S , and Hill, I. D. analysis of changes in incidence of bronchial < noma in textile asbestos factory. .Ann NY Acad! 132 526, 1965 References 15 Braun-. D C , and Traun, T. D Epidemiological:, study of lung cancer in asbestos miners. Arch 1 Msictso. T F nm) Cm i ter E j Methods of Indi'iur Health 17 634, 1958 sltidimn (lie i(.`latum "I eroplmtnenl ami Imm- 16 Mascuso T F Discussion of studies. Ann .VP tvun tSIlK'ss solum .mats sis \ mi, r / Fiibhi Ifcnlth Actul Sct 132 589, 1965 to pm 17 Silikoff. I J . Cuuno, | , and Hammond, C. As 2 W ssriM i I J" .mil I i i i n m b I Meflimlnlnm bestos exposure and neoplasia JAMA 18H 22, m kicIunIi i.tl lie.itlli studies-- (lie dctnoei.iplm. up- 19 6 1 pin.itii \ if /i I in nmi litnlth 6 515 196? 15 Enthuise P E Mortality among asbestos prod 1 M s m i n i i 1 F .mil f i n i ii it E f Methmliilniiv ucts \sorkers m the l tilled States .Ann A'7 Acad in imliistiial health siuflics;-- (lu> uiluirl ap pu uU i Su 1 >2 156. 1965 mill spest-d retc-iente lii -ill asbestos (iini|i.iiiv 19 O D osntll W H . and M sss. K H Asbestos; \n h Lm ouii Health h 2 In 196] .in cstimstc factor m the pathogenesis of broncho- 4 \ K \ n in T I mil F i - A r i s n . A A Dynamic t'ciiic. carcinoma Amcr f Path 73 610, 1957 lti.mlies in imliiNlit.il cufinit studies fni/tntr \htl 20 O'iViwtLL, W M, Mann, R H , and Crosh, Sure >5 1059 1966 f L Asbestos, an extrinsic factor in the pjlho- " M w i i mi T F unit E i - Ai tsh A A Evaluation cteitesis of bronchogenic carcinoma and mesothe of i ilium vimhr* oI (mItiNiii.il popni.ilihik Pre- lioma Cancer 19 i 141. 1966 OH \C r F, OF \F.SCUL\PIUS Serpents ucre sacred lo Aesculapius the healer probable because of a super stition tint those animals b.uc a faculty of renewing their south bv a change of skin. Tlw- worship of Aesculapius "as introduced into Rome in a time of great sickness, and an embassy sent to the temple of Epidaurus to entreat the aid of the god Aesculapius si,is propitious and on the return of thc.ship accompanied it in the form of a serpent Arriving m the river Tiber the serpent glided from the vessel and took possession of an island in the riser, and a temple is as there erected to ins honor B u t f m i h ' \ Mi/tliolosij, bv Thomas Bulfineh Thomas Y Crowell Comp.uis. New York. 1947. p 29S