Document ykEYNZ95Y5w0Q966pMGe634r4

Scand J Work Environ Health 1989:15:43--46 Mortality of auto mechanics A ten-year follow-up by Eva S Hansen, MO, PhD1 HANSEN ES. Mortality of auto mechanics: a ten-year follow-up. Scand J Work Environ Health 1989; 15: 43--46. This study was set up to investigate whether work in car repair workshops is associated with an increased risk of ischemic heart disease and specific malignant neoplasms. For this purpose, a cohort of auto mechanics has been followed through 10 years with regard to cause-specific mortality. Comparisons were made with another cohort of skilled male workers who were not exposed to asbestos or petrochemi cal substances. The auto mechanics' mortality was increased for ischemic heart disease (standardized mor tality ratio (SMR) 121,95 7 confidence interval 195 Cl) 102--145j. other cardiovascular diseases (SMR 112.95 9k Cl 82--150), cancer (SMR 115.95 % Cl 97--136). other diseases (SMR 119.95 7. Cl 94--149). and external causes (SMR 131.95 Cl II3--153). For specific cancer sites, increases were seen for pan creatic cancer, urinary cancer outside the bladder, and pleural mesothelioma. Key terms: asbestos, car repair, carbon monoxide, ischemic heart disease, mesothelioma, occupation, pan creatic cancer, petrochemical substances, polycyclic aromatic hydrocarbons, urinary cancer. The work environment of auto mechanics involves potential exposure to asbestos, mineral oils, solvents, paint pigments, anticorrosive substances, and automo bile exhaust. The literature on work-related health haz ards is sparse regarding auto mechanics. Increased risks of both bladder cancer and lung cancer have been reported (I--3). In addition a few cases of malignant mesothelioma have been reported for auto mechan ics, especially among persons employed in brake service centers (3--5). With regard to other occupational groups whose exposure may resemble that of auto mechanics, increased cardiovascular mortality has been reported for motor vehicle examiners (6), whereas in creased mortality from digestive cancer has been found among automobile manufacturing workers and other metal workers (7. 8). The'actual study was set up to investigate further the potential health hazards associated with the com plex exposure situation typical for the work environ ment in Danish car repair workshops. Malignant ne oplasms of the respiratory system, the urinary system, and the upper digestive organs were anticipated to be of particular interest, together with cardiovascular dis eases. Subjects and methods The research design was that of a historical cohort study in which the exposed part of the cohort was com pared with the unexposed pan in terms of cause-spe cific mortality through a ten-year period of follow-up. 1 Institute of Community Medicine. University of Odense. Odense. Denmark. Reprint requests to: Dr ES Hansen. Institute of Community Medicine. J B Winslowsvei 17. DK-5000Odense C. Denmark. Auto mechanics constituted the exposed part of the study cohort. The comparison population, ie. the un exposed part of the study cohort, was made up of other specific groups of skilled workers. The members of the study cohort were identified from (he files of a nationwide census carried out in Denmark on 9 November 1970. Self-reported occupa tion. trade, industry, and employment on the day of the census was recorded for all the Danish inhabitants older than 14 years of age. These data were used to select the persons included in the study. The study comprised only skilled workers and ap prentices. only men between IS and 74 years of age. and only persons who were occupationally active (ie. employed) on the day of the census. Everybody with the title "auto mechanic" or "mechanic, employed at a car-repair workshop" was included as an exposed person. As unexposed persons all men with one of the following job titles were included (proportion of to tal comparison group in parentheses): carpenter (45 7a), electrician (32 <ra), instrument maker (13 7a), dairyman (S 7e), upholsterer (3 7o), or glazier (2 7b). Table I shows the number of exposed and unexposed persons in ten-year age groups. The Danish National Bureau of Statistics has linked the records of the 1970 census with the current cen tral register of persons and the national register of deaths (publications numbers 37 (1979) and 41 (1985) from the Danish National Bureau of Statistics, both in Danish |. This procedure has enabled the total cen sus population to be followed for 10 years. The ac tual study cohort formed pan of the census popula tion and was traced by the use of the existing register linkage. The study persons were traced until 8 Novem ber 1980 or till death or emigration prior to this date. The follow-up data arc shown in table 2. The auto % Table 1. Number of exposed persons (auto mechanics) and the comparison population ol unexposed skilled workers by age. Age (years) on 9 November 1970 Auto mechanics N% Comparison population N V, 15--24 25--34 35--44 45-54 55--64 65--74 15--74 13 682 4 661 1 545 949 662 101 21800 64 21 7 4 3 0.5 100 25159 12 633 6118 3 976 3146 713 51 747 49 24 12 8 6 1 100 Table 2. Follow-up data lor the auto mechanics and the com parison population of skilled workers from S November 1970 through 8 November 1980. Auto Comparison mechanics population Number of deaths Number of emigrants* Number of persons alive and living in Denmark at the end of the follow-up 586 581 20 633 1 903 1 665 46 179 Includes persons lost to follow-up. The register linkage did not allow for separating emigrants and persons lost to fol low-up. comparison population and the number of personyears at risk accumulated by the auto mechanics. For the statistical evaluation, the observed numbers of deaths beyond 100 were assumed to follow a nor mal distribution (Yates' correction was employed), whereas, for smaller numbers, a Poisson distribution was assumed. The death rates of the comparison popu lation were assumed not to be influenced by random variation. Confidence intervals (Cl) for the stan dardized mortality ratio (SMR) were calculated by the method proposed by Miettinen (9) whenever the ob served number of deaths exceeded 100. For smaller numbers of deaths, exact Poisson limits were calcu lated (10). Results The number of deaths observed among the auto mechanics exceeded the expected number by 21 Vs. The increased mortality was not confined to any single group of causes of death (table 3). As regards specific cancer sites, increases were seen for pancreatic cancer, pleural mesothelioma, and uri nary cancer outside the bladder (table 4). Table 3. Mortality of Danish auto mechanics in 1970--1960. (SMR = standardized mortality ratio, 95 % Cl =95 % confi dence interval) Cause of death* Ob Ex served pected SMR number number 95 % a Cancer (140--209) 148 128.8 115 97--136 Ischemic heart disease (410--414) 132 108.7 121 102--145 Other cardio vascular diseases (390--404, 420--458) 46 41.0 112 82--150 Other diseases 76 64.0 119 94--149 External causes (E000-- E999) 175 133.2 131 113--153 Unknown 6 5.9 101 37--220 Total 583 481.5 121 112--131 * The code of the international Classification of Diseases (eighth revision) in parentheses. mechanics and the comparison population of unex posed skilled workers contributed 192 300 and 481 642 person-years at risk, respectively, to the study. Deceased persons, emigrants, and persons lost to foliow-up contributed person-years at risk up until the day of death, the day of emigration, or the last day of notification in the central register of persons. For deceased persons, information on the underly ing cause of death was derived from the death certifi cate. The numbers of deaths expected for the auto mechanics were calculated from the death rates of the Discussion The auto mechanics' mortality has been evaluated against that of skilled men in other trades, free from occupational exposure to petrochemical products, as bestos, and paint pigments. However, some of the per sons included in the comparison group may have been occupationally exposed to other agents associated with increased mortality from, eg, cancer. In this case, a potentially increased risk for auto mechanics may have been masked. According to the concept introduced by Hernberg (II), the actual cohort study was rather insensitive. Only occupationally active and therefore healthy work ers were enrolled in the cohort, and the period of fol low-up was rather short. "Exposure" was defined by a person's occupation and industry and determined at one point in time, namely, the day of the census. In Denmark skilled workers tend to stay in one industry throughout their worklife. Thus, for the actual study, the census data may represent a reasonably good proxy for lifetime occupation and industry. However, the crudeness of the employed definitions dilutes the ex posure contrast under study, and some effects of the auto mechanics' work environment may have gone un detected. Selection of the occupational groups included in the comparison population enabled the group to resem ble auto mechanics with respect to work-related de mands on physical strength and fitness, social class membership, and geographic distribution. In spite of this group control for social class, life-style habits may have differed between the compared populations. The 44 Table 4. Danish auto mechanics' mortality from specific malignant neoplasms in 1970--1980. (SMR* standardized mortality ratio, 96 % Cl =confidence interval) Cancer site* Observed number Expected number SMR 95% Cl Esophagus (15(8 Stomach (151) Pancreas (157) Other digestive organs (152--150, 158--159) Bronchus and lung (162.1) Pleura (163.0) Other respiratory organs (160--162.0, 163.1--163.9) Bladder (168) Other urinary organs (169) Other sites (140--149. 170--167, 190--209) Total (140--208) 4 3.4 7 6.6 17 7M 16 17.0 41 40.7 1 0.0 1 0.3 4 4.1 8 4.6 49 44.2 148 128.8 The code of the International Classification of Diseases (eighth revision) in parentheses. 116 104 219 94 101 400 98 174 111 115 32--298 42--214 12B--351 54--153 72--137 10--2229 27--250 75--343 2--147 97--1% auto mechanics' high mortality from external causes points to differences regarding risk-taking behavior. On the other hand, the fact that no differences were found for lung cancer indicates that the results have not been positively biased by smoking habit differ ences. Actually, smoking is prohibited in many places in car repair workshops, and the auto mechanics may in fact have smoked less than the persons in the com parison group. The number of deaths ascribed to ischemic heart dis ease exceeded the expected number by 21 an ef fect which is of the same order of magnitude as that of smoking 1--14 g of tobacco per day (12). The ob served excess of deaths from ischemic heart disease is remarkable, particularly in view of the rather short period of follow-up. Ischemic heart disease usually has a protracted course rich in symptoms and increasing disablement. Persons with symptoms of cardiac ische mia arc rare among active blue-collar workers, and with short-term follow-up one usually finds a very low mortality due to ischemic heart disease in such groups. The difference between the auto mechanics and the comparison group with respect to ischemic heart dis ease may either be due to specific occupational ex posures or to confounding by life-style factors. With regard to the latter, smoking can be ruled out because of the aforementioned lack of difference in lung cancer mortality between the groups. Hypothetically, the com pared populations may have differed in dietary habits or leisure-time activities. However, such differences would have had to be unusually extreme to bring about an effect equaling the actual order of magnitude. With respect to work-related hazards, the auto mechanics may be exposed to suspected atherogens like carbon monoxide and polycyclic aromatic hydrocarbons through the inhalation of automobile exhaust and dur ing (he handling of petrochemicals such as, eg. anticor rosive substances, solvents, and mineral oils. Previous ly, a slight increase in cardiovascular mortality (SMR = 105) has been reported for motor vehicle examiners. whose exposure may resemble that of auto mechanics (6). Pancreatic cancer occurred in excess among the auto mechanics. However, for such a rare disease, the com parison population may have been too small to justi fy the assumption that random variation has not in fluenced the death rates. Therefore, reservations should be made as to the statistical evaluation of the data on mortality from pancreatic cancer. Little is known about the etiology of pancreatic cancer. Tobac co smoking seems to be of importance, alcohol not (12--15). Studies on occupational mortality have in dicated that this cancer site is associated with work in the metal industry (7. 8). the chemical industry (16, 17), especially in the production of beta-naphthylamine and benzidine (18), and with exposure to petroleum products (19, 20). Animal assays have pointed out nitrosamino compounds, acetaminofluorene, p-dimethylaminoazobenzene, and mcthylnitrosurea as car cinogens with affinity to the exocrine pancreas (21--24). The chemical substances in the auto mechan ics' work environment may include nitrosamino com pounds formed from, eg, sodium nitrite and triethanol amine. both of which are in current use as mineral oil additives. Asbestos exposure is known to occur during the replacement of brake linings, and the single case of pleural mesothelioma is an indication that this exposure has not been negligible. On the basis of this back ground. it is remarkable that the auto mechanics' lung cancer mortality was not increased. The possibility that smoking has been less prevalent among the auto mechanics than in the comparison group has already been mentioned. If present, such a difference would result in negative confounding as regards lung cancer and other tobacco-related diseases. For specific cancer sites other than lung cancer and pancreatic cancer, the observed numbers of death were too small to state or rule out a potentially increased risk. Acknowledgments The author thanks Mr O Andersen of The National Bureau or Statistics for providing the possibilities to make use of census records. This work was supported by a grant from The Dan ish National Anti-Cancer League. References 1. Smith EM, Miller ER, WooUon RF. Brown CK. Blad der cancer risk among auto and truck mechanics and chemically related occupations. Am J Public Health I985;73:MI--3. 2. Schwartz E. Proportionate mortality analysis of automo bile mechanics and gasoline service station workers in New Hampshire. Am J Ind Med 1987; 12:91--9. 3. Jirvholm B, Brisman J. Asbestos associated tumours in car mechanics. Br J Ind Med (in press). 4. Elliehausen HJ, PaurR, ROdetsperger K. WottowitzHJ. Zum Risiko von Asbestinhalationsfolgen bci Kraftfahrzeugmechanikern in Brcmsendienstcn. Arbeitsmed Sozialmed Praeventivmed 1985; 11:256--61. 5. Paur R, Woitowitz HJ, ROdelspcrgcr K. Jahn H. Pleuramcsotheliom nach Asbeststaubgefahrdung bd Bremsreparaturen in Kfz-Handwerk: Kasuistische Beobachiungen. Prax Klin Pneumol 1985;39:362--66. 6. Stern FB. Lemen RA, Curtis RA. Exposure of motor vehicle examiners to carbon monoxide: a historical prospective study. Arch Environ Health I98l;2:59--65. 7. Vena JE, Sultz HA. Fiedler RC, Barnes RE. Mortality of workers in an automobile engine and pans manufac turing complex. Br J Ind Med !985;42:85--93. 8. Milium S Jr. Cancer monality patterns associated with exposure to metals. Ann NY Acad Sci 1976:271:243--9. 9. Miettinen OS. Simple interval-estimation of risk ratio. Am J Epidemiol 1974;IOO:5I3--6. 10. Documenta Ccigy. mathematics and statistics. 5th ed. Copenhagen: Ciba-Gcigy, 1975:107. 11. Hernbcrg S. "Negative" results in cohort studies -- how to recognize fallacies. Scand J Work Environ Health I98!;7 (suppl 4): 121--6. 12. Doll R, Peto R. Monality in relation to smoking: 20 years' observation on male British doctors. Br Med J 1976:2:1525--36. 13. Mack TM, Yu MC, Hanisch R. Henderson BE. Pan creas cancer and smoking, beverage consumption, and past medical history. J Natl Cancer Inst 1986;76:49-60. 14. VeUma JP, Walker AM, Cold EB. Alcohol and pan creatic cancer. Epidemiol Rev 1986;8:28-41. 15. Williams RR. Horm JW. Association of cancer sites with tobacco and alcohol consumption and socioeconomic status of patients: interview study from the third national cancer survey. J Natl Cancer lost 1977;58:525-47. 16. Li FP, Fraumeni JF. Mantel N, Miller RW. Cancer mor tality among chemists. J Natl Cancer Inst I969;43:l 159-- 64. 17. Olin GR, Ahlbom A. Cancer mortality among three Swedish male academic cohorts: chemists, architects and mining engineers/meiallurgisis. Ann NY Acad Sci 1982:381:197--201. 18. Mancuso TF, El-Attar AA. Cohort study of workers ex posed to beta-naphthylamine and benzidine. J Occup Med 1967;9:77-85. 19. Lin RS, Kessler JJ. A muhifactorial model for pancreatic cancer in man. JAMA 1981:245:147--52. 20. Norell S, Ahlbom A. Olin R, et al. Occupational fac tors and pancreatic cancer. Br J Ind Med !986;43: 775--8. 21. Hoch-Ligeti C. Primary pancreatic tumors in rats fed p-dimcthylaminoazobenzene. Br J Cancer I949;3: 285--9. 22. Pour P. Salmasi SZ, Runge RG. Selective induction of pancreatic ductular tumors by single doses of N-nitrosobis(2-oxopropyl)aminc in Syrian golden hamsters. Cancer Lett I97g;4:317--23. 23. Reddy JK. Rao MS. Pancreatic adenocarcinoma in in bread guinea pigs induced by N-methyi-N-nitrosurea. Cancer Res 1975;35:2269-77. 24. Wilson RH. DeEds F, Cox AJ. The toxicity and car cinogenic activity of 2-aceioaminofhiorcne. Cancer Res 1941;1:595-607. Received for publication: 1 July 1988