Document wr95bYO8QwQQBaRwyV12obkED

62U5 SCT21 P 35 5*1? J XHO H60 British Journal of Industrial Medians 19H4;4^:35-3iJ Mortality experience of glass fibre workers /I 2-2- 7^ . H S SHANNON. M HAYES. J A JULIAN. AND DCF MUIR From ihe'Octvpadonol Health Fforrom. McMastar University Health Sciences Centre, Hamilton, Ontario LBN JZJ. * ****** abstract A historical prospective mortality study was conducted at an insulating wool plant in Ontario, Canada, on 2576 men who had worked for at least 90 days and were employed between 1955 and 1977. Eighty eight deaths were found in the 97*2% of men traced. Mortality was compared by the person-years method with that of the Ontario population. Measurements taken since 1977 show very low fibre concentrations. The overall standardised mortality ratio (SMR) was 78%. significantly below 100. Among plant only employees, seven deaths were attributed to lung cancer compared with 4-22 expected, a non-significant excess (SMR - 166; 95% confidence limits 67 to 342). No confirmed cases of mesothelioma were observed and no other disease was significantly increased in plant workers. Glass fibre has many uses but is also under suspicion had been employed for a total of 30 days at any time as a caresnogen since animal implantation work between the opening of the plant in 1948 and the strongly implies that the dimensions of fibres are end of 1977. It was discovered, however, that the more important than their composition with respect main source of records of employees who left befese to potential carcinogenicity.1 Indeed animal studies 1969 was incomplete. By using income tax records have found that implantation of glass fibres can pro of all employees who worked between 1960 and duct tumours in rats.11 but studies of exposed indi 1969 together with union records, group insurance viduals have failed to detect such effects,'* although rolls, and seniority lists from 1955 to 1962 H was Bayiiss found an excess of noo-malignant respirat possible to generate a complete list of all the men ory disease.' who had worked at least 90 days and were employed The present paper repons a historical prospective at any time from 1955 onwards. Although records of mortality study of workers engaged in the manufac other employees who left before then were located, ture of glass fibre. Of especial interest were lung and it was impossible to be sure that the list was com gastrointestinal cancers, particularly mesotheliomas. plete and would not omit a group among whom the Attention was also to be given to non-malignant number of deaths might be disproportionately high. respiratory disease. In addition, it was noted that few, if any. women had genuine exposure to glass fibre. Methods aad material It was thus decided that the study group would comprise only those men who had worked at dm SITE OF STUDY ....... plant between 1955 and 1977 and who had had at An insulating wool plant in Sarnia, Ontario, least 90 days' total employment; 2576 such men Canada, was the location of the study. It opened in were studied. 1944 with about 200 hourly rated employees aad currently has about 500. latdaly, a steam Mown PERSONAL OATA ANO WORK HISTORIES fibre was produced but a change wan made to cen- Wherever possible detailed work histories were trifugally formed ttaas in 1961. A small Bams compiled for the study subjects. Not all of these attenuation proeam occupied a separate area from were available, however, although individuals could the mid-1950s uadi 1974. be divided into three groups: thoes who worked in the plant only, those who worked in the office ama STUDY POPULATION only, and those whose job entailed a combination of It was originally planned in include all workers who plant and office work C*mixed exposure"). In addition, other persons! information snrii as XeewwC IS NevtaOv INI Amtona to imm ties hill name, date of birth, and last known sddrsm were extracted from the records. These data were to 35 SSrriSHHItOlCAL A|S L , niuOuri wCIH 9 tn 08 M mist in tracing thoac who had left the company. The data wen IranKribed into machine readable form by study team personnel using mainly a microfiche copy of the original documena Shannon, Hayrs. * over 0-2 flbrea'ml and the mean levels in were below 01 flbrewml MORTALITY ASCERTAINMENT Vita) status was unknown for meet of the population and various methods were used to trace these men. Two students were hired to make contact with sub* jeers who might have remained in the area or with family or friends who could provide information. Driver licence bureaus in Ontario and other Cana* dian provinces consented to a search of their data bases. Statistics Canada's Vital Statistics and Dia* ease Registry found that some subjects had died and provided dates and official causes of death for these men as well as for those found to be dead through other means. Other follow up methods included the use of street directories and telephone books of the local area, current plant employees, driver licensing records, and the United States Department of Immigration and Natursliation. Sources of tracing were recorded and documentation retained. ANALYSIS The primary analysis adopted the penon-years approach described by Case and Lea.* using a com puter program written by one of us (JJ). Age and calendar specific death rates were computed and compared with those of the Ontario male popula tion. Indirect adjustment was made and SMRs calcu lated. Statistical tests of significance were applied by treating the observed number of deaths as a Poisson variate with a mean equal to the expected number of deaths. When the expected number was greater than three, the approximation to the normal distribution was adopted. The statistical tests were two-sided; an a of 0-05 was used as the primary level of significance. Because the numbers are small 93% confidence intervals have been calculated for tome SMRs to indicate the range of possible values. ENVIRONMENTAL DATA Historical exposure data were not available for the period of study. Samples taken in 1978 and 1979 suggest that optical microscopy counts are rarely Table 1 shows the distribution of person-years by age and calendar period. It should be noted that tome men were employed before 1933 but were only under observation from 1933 onwards. Only 302 men contributed person-years beyond 20 yctrs from their lint exposure (13 of these were office workers who would have hid very little exposure to fibrous glass dust). The number of men exposed IS or mote yean was 303. tight of whom were Office employees. AH but 74 men were traced, giving a follow up rate of 97-2%. Seventeen of those not found were believed to have emigrated to the United States. Of the remainder, a search of the National Mortality Data Base at Statistics Canada did not yield a death record. Since it is very probable that if any had died in Canada a death certificate would have been located, the effective follow up can be considered greater than 97-2%. Eighty eight men were found to have died and death certificates were located for all but one of them. Additional data suggested that this death was due to ischaemic heart discesc and this cause has bean used in the analyse. The expected number of deaths was 113-3 giving a significantly low SMR of 78% (x,* * 5-63. p 002). While at least a part of the deficit might have been expected in an occupational population, when the analysis was restricted to obscrvttion at least 10 yean after first employment the SMR was still low--36 deaths observed v 73-34 expected. SMR 74. One would have expected the healthy worker effect to have largely attenuated. Mortality was examined by dividing the popula tion into the three gtoupe described ftovc (plant only, office only, and mixed exposure). Table 2 shews the observed and expected deaths by ICO chapter headings for thorn categories with at tenet three deaths. Circulatory dhesses wen fewer than expected in plant only employee* but slightly more numerous in men with mixed exposure. Neither dif ference was statistically significant. Mortality hem Ttbtel Pentm nmrttkkyafinltndmpitied CiMvrNU 41 41- 41- Aft- JJ- 41- * nml |iIIf9ff*?)U*MM-M# Tirsoui-f ME MM Ufl MimM mSUaIm4i . MM t1a444n*? IioSStaI MM 4n2S4o4? iMi?4i 4414 M Ml 237 IJ4420M9? IIMI?4f M10O1 a M2s 4IS4 2SS4?M42t*I 1202 2*444t?4 r.'r.v.wnv,- A-.V V V.- V.-.\V.Y.`.Y % `UiitfTffTf .\r.v.-.\\v.-.\er.*.v 7Y*tmnr* J.v.vrJr.lTfrrf.-iim wKYT.-rr.-.- ...-TTr.'T.rrrr--.\\r 'renrmtft rvtiittt:* ".innunntKnnnSW'.WV.tf.v.-.v.v.-Rv.tTrtTtmY Kv.tv.nv.'. YttnMMimr.rtirrmttr.Tf.m: iwrr.wn i..m\..........y-inei *S*mw^v.wv.w.-r rmwv.-.v.---*.-.. cT tL-: vv-- (IBSi txptrimn ofgtaa fibrt workm Mormiirjr by mo/or com* ffwie br P^om. offico. v4 miir* rxpomn prompt tr^m .......1,`.v,v.vw T QraUuof* i Plomomt* o8 M 4074 17 IOJ4 1) nsi 4 4-10 1 440 1 071 71 4447 SMI 74 101 SO 41 74 74 74 tiffin oott 00 J >70 2 . IS] 0 2-S0 0 042 0 040 1 ' 070 4 02S sum ii in 0 0 0 14] 45 37 0 0 KM 7 >1} IS] 1 I-S4 44 2 134 .144 0 OM 0 0 014 0 0 0S7 0 10 010 in all cancan was almoat identical with that expected (20 observed v 19*63 expected) with little variation by exposure. Deaths due to accidents and other vio lent causes were low. significantly so for plant only worken (x* " 6*02. p 04)4). Noo-malignant respiratory diseases accounted for four deaths. AO these occurred in plant worken among whom the SMR was 91. The remaining causes of death showed no excess. Mortality from more specific causes is shown in table 3. Ischaemic heart disease was low in plant worken and high in mixed exposure employees, although neither difference from expected was statistically significant (*,* 3*29 and 3-44. p 0-07 and 0*06 respectively). Seven deaths from lung cancer were observed in plant only employees against 4*22 expected. The SMR of 166. though, was not statistically significant Or,1 * 1*83. p 0*18). Among office worken the SMR was 360 but with only two deaths the excess was not statistically significant (p D10). The 93% confidence interval on the lung cancer SMR for plant only worken was 67 to 342. Because of the width of this interval the data were considered by duration of exposure and by time since first exposure. An increase in the SMR with increasing levels of these variables would suggest a genuine relationship. Table 4 shows these data. While the numben of deaths become even fewer, the data do not support the hypothesis of a link between occupa tional exposure to glass fibre and increased lung cancer mortality. Permission was granted by the Ontario Registrar General's Office to contact the hospitals in which the deaths from cancer had occurred. All but two such deaths occurred in Ontario. The records showed one man with stomach cancer in whom dm dinical history suggested the possibility of mesothelioma. He started working with the compsny in 1948 at age 38. left in 1966. and tied in 1976. Tissue samples were sent to the r%--u-- Tumour Reference Centre to confirm or refute this diagnosis. Of three pathologists who reviewed the * material, two believed it to show a sarcoma, while the third thought that mesothelioma was a dating possibility. On balance, therefore, it was concluded that a diagnosis of mesothelioma could not be confirmed. Cerebrovascular deaths and those due to cancers of the digestive organs and peritoneum were roughly equal to their expected numbers. Three deaths were ascribed to "bronchitis, emphysema, and asthma.** although since these three causes aceojinted for one death each, furher comment teems unwarranted. Among plant only workers the three cancers of the genitourinary organs gave a SMR* of 161. Three different sites were affected, however genital organs, kidney, and bladder. Lymphatic and haematopoietic malignancies had a SMR (in plant workers) of 92 (2*0 observed v 2*18 expected). Table 3 Montihy from sHtcmd coma jbrplum. offict. 0*4 miu4 apotmn {roupt fr 10 2042 4T s7 4s*-1ia2 ISrCt i MS as 1 1*41 IS4 S 1*44 141 2 MS 41 Offinmtp tIM 1 247 111 2 014 SCO 044 0 047 0 0o 0o1ia4 o0 on SMq SMM 7 i-44 ne 0 044 4 oat e 0t ooOnii4l seo4t a oia e ( tv-*' `tTTfttm* I...... t...... Err U::: oou 7647 Shannon. Hayrt. Julian. Usd Muir hzZZ TaM* 4 Ln*t ctnctr rntm/iff tj iurmon oftxpomrt and by turn anti fnt txpomtt --rkm otty wottm nmumnfm mfmmttytmti <10 to waa ftmUtt tymttl <1 >9 O 1 SMR o C 1 >SI SSI 1 oaj m 9 2S i US SMR 0 1ST A major focus of the study was oo lung cancer. As noted, the 95% confidence interval oo the SMR for that cause for plant workers was very wide ((7 to 342). Indeed, the power ofIhe study to detect (at an a of 04)5) a true twofold risk of lung cancer mortal* icy was only 42%. The more detailed breakdown (table 4), however, did not nicest a genuine rela tionship. Examination of hospital records found a possible case of mesothelioma, but further patholog ical examineboo could not confirm this. Among other diseases of interest, none showed a significant excess. Indeed, both cancers of the diges tive organs and peritoneum and non-malignant respiratory disease were slightly fewer than expected. No other disease produced a statistically significant SMR, except far accidents and violence. The reason for the shortfall is not dear. The age standardised male death rate from these causes in Lambton county (in which the plant is located) from 1966 to 1976 was 103-6 compared with 1009 for Csneda as a whole, so local facton srere apparently not responsible.' Because the plant opened in 1941 and the study period ended in 1977 no fallow up beyond 30 yean from lint exposure could be achieved. (It expected when the study was planned that collated mortality data after 1977 for Canada would not he available.) Nevertheless, h is hoped that this popula tion will continue to be followed up and the analysis repeated in several yean. We ate grateful to ad who helped with this research. In particular we thank Fiberglass. Canada, for finan cial support and for their coodnued involvement in this research endeavour; the Energy and Chemical Worken Union, Local 914, for their unremitting cooperation; members of the joint managementlabour steering committee; and pathologists. Ota Elliot McCaughcy. Rena Lefabrc, and Andrew Oturg for reviewing tissue samples. We also dunk Mn Pat Yakt for typing the many drafts of this report and for organising materials sad meetings throughout the project and Ms Sandy Martin for her valuable help in the early stages of this project. IMf;e:ttfr44. DwaJMO. Ta aw pwwai m*WT cm nww tenJ3: no-Mi. KM IW. .. ................ CWM JO. Hiasii* OA. Qwa Stow: Wwm af taiaoaary Wove W traeacaaa awtm. Sr tuM Mat ierajonra-e. ' Ofwt t. Tawa I. DaTmUa RTF. Uap of wrtm ipan a sen 0MK wMyM tkaa awMbpc ckaapa a* UwW aa aaai. AtWt Mm MmU l*l JS;*T-7A > Bayw OI_ OwaaM M. Wafoaar IK. Ikw MP Moraarr m- wa iaon| Straw (law stoSawWa awtats. Am MY Amt ia insjrvJM-ss. -- Cw RAM. La AJ. tat Walim. Mwa*? at* a/ Ottawa: NMW. ItaOUIS. VMS.