Document DDL4nJGp0ybpnVQzBEgvzMQ8O

NOTICE: Thk? rllatwhl may be protected by copyright law (Title I?, Unitea Stares Code). Mortality Experience of a Cohort of Rubber Workers, I964-1973 PLAlNTIFF'S Dragana Andjelkovic, M.D., M.P.H.; john Taulbee, M.S.P.H.; and Michael Symons, Ph.D. I n 1970, the United Rubber Workers Union and six major rubber products manufacturing companies in the United States agreed to undertake joint occupational health programs. As a result of this pioneering effort, comprehensive studies of work related health problems have been initiated. This article presents the results of a cohort mortality analysis for a single rubber manufacturing plant located in Akron, Ohio. The objective of this study was to explore whether the general mortality experience of rubber workers is greater than in other industries or in the general population. Design of the Study The basic approach used in the present analysis was a historic prospeclive study design. The population at risk (PAR), conveniently called "the 1964 Cohort", was defined as follows: any Person who, as of January 1, 1964,was Jo or more years of age and was an active or living retired hourly worker from the plant under study. The mortality ex- Perience of the population at risk was observed for the period of ten years, from January 1, 1964 to December 31, 1973. Procedures used to collect the necesQrY data for the reconstruction of the PAR required a thorough check of the Occupational Health Studies Group, f F h l of Public Health, Departments of Epidemi- nW 0. Andjelkovic) and Biostatistics (Mr. T''(bee and Dr. Symons), University of North GrdiW chapel Hill, N.C requests to Dr. Dagana Andielkovic, De- PInment of Epidemiology, School of Public Health, Univeeihlof North Clrolina, Chapel Hill, NC 27514. various company and union records. A large number of death certificates were collected from the company and, in the process of follow-up of the cohort, the remainder from bureaus of vital statistics of several states. The causes of death were coded by a trained nosologist according to the Eighth Revision of the International Classification of Diseases. Follow-up of the cohort is 96.7% complete. The remaining 3.3% of workers in the cohort will be followed-up in a subsequent study. The study was designed so that mortality experience among active and retired workers could be examined separately. Since many workers retire prior to age 65,we obtained information on age and date of retirement for all retirees in the cohort. Workers were treated as active until the time of retirement and as retirees thereafter, regardless of age. It was felt that this design would allow for better study of active versus retired mortality experience than would an attempt to differentiate between the two groups by age for normal retirement. Methods Overall and age specific death rates were calculated for all deaths and for deaths due to selected causes. These rates were expressed -as standardized mortality ratios (SMRs). The e x w e d number of deaths for each five-year age group was calculated by multiplying the number of person years at risk in that age group by the age-race-specific death rate in the standard population for the burnal of Occupational MedicineRol. 18, No. 6/June 1976 same age group for the cause or group of causes of interest. The SMR, then, is seen to be the ratio of the death rate in the population at risk (PAR) to the death rate that would have occurred in the PAR if deaths occurred at the same agespecific rates as in the standard. To obtain the expected number of deaths for an age range longer than five years, the constituent five-year figures are added together. To examine whether the difference between the observed and expected number of deaths was due to chance alone, a chi-square test of signifi- cance was calculated for each of the two age-ranges 40-64and 65-84 and for the entire range 40-84. Unless othelwise noted, for all SMRs presented in this paper, the standard population used is the 1968 US. population of comparable age, race, and sex. The use of the general U.S. population as a standard in studies of the morbidity or mortality of an industrial population has been recognized as one of the major shortcomings of a study such as this one. This is because this standard population consists of all persons, working, not working, retired, chronically ill, institutionalized, physically handicapped, among other conditions, while workers are, by virtue of the fact that they are working, healthier than the remainder of the U.S. general population. This favorable selection operates not only at the time of a preemployment physical examination which is routinely performed in most industries, but also in selecting particular occupations in 387 L_ - which the worker is capable of performing. Because of this selection, the mortality experience of an industrial population is expected to be more favorable than that of the general population, giving rise to the phenomenon called the "healthy worker effect". For comparative purposes, a general industrial population consisting of all industrial workers in the U.S. would be a more preferable standard population. In a study1 of mortality by occupation and industry conducted in 1950 by the U.S. Department of Health, Education, and Welfare (HEW), standardized mortality ratios for various industries and occupations were calculated using workers employed in all manufacturing industries as the standard. Although several problems, mainly related to incompleteness of data, were encountered in the conduct of this study, it i s the only one known to us in which the standard population consisted of all U.S. industrial workers. If such a standard population was available as of a more recent year, it would be preferable to the U.S. general population as a standard. However, since recent data for such an ideal standard population are not available, the use of the U.S. population has been common practice in studies of this type. Results of Related Studies During the last several decades there has been a growing awareness of occupationally related health hazards, and numerous mortality studies of industrial populations have been conducted. Industries which have been investigated in some depth include the following: asbestos, cotton textile, steel, chemicaI, coal and uranium mining, and some others, including the rubber industry. Cohort mortality studies of industrial populations are limited in number, and in the U.S. and Great Britain, most of these studies calculated mortality indices based on the general population as the standard. The overall mortality experience of rubber workers has been examined in several studies. S M R s for all causes which have been reported for the rubber industry range from 86 to 99. The aforementioned HEW study,' using the U.S. industrial population as the standard, reported an S M R of 86 for the rubber industry. From a study of a cohort of over 13,000 rubber workers in Akron followed from 1940 to 1971,zthe S M R was 388 reported to be 86,using the US. general workers employed between 1948 and populatioi is the standard. In a group 1951 and followed through 1963,a an life insurance study for the period 1955 S M R of 112 was found for asbestos to 1964, the US. Society of Actuaries products workers employed in 10 in- found an S M R of 89 for rubber workers., dustries, contrasted with an S M R of @ A recently reported study4 of mortality in among cotton textile workers. An SMR the rubber and cablemaking industries in of 82 was reported for a cohort of over Great Britain for the period between 59,000 steelworkers followed for nine 1968 and 1971 found an overall S M R of years.9 The general population of the 97.This study was based on over 40,000 local county was used as the standard in workers employed on lanuary 1,1967,in this study. 381 firms throughout the country. The The above discussion refers to overall standard population used was that for mortality. Several of the aforementioned England and Wales. For a cohort of 6678 studies of rubber workers have produced male rubber workers in Akron followed results for specific causes of death which from 1964 through 1972,5 the S M R was are of interest. Mancuso's studylo 11 of 99 for the age range 40 to 84, based on mortality in the state of Ohio suggested U.S. general population data. The defini- that there was an elevated proportion of tion of that cohort of 6678 workers is deaths due to cancers of the respiratov. very similar to that for the cohort of rub- genitourinary, and central nervous ber workers presented in this paper. systems among workers of the rubber For informative purposes we present and plastics industry as compared to some all-causes SMRs which have been workers in other industries. Subsequent- reported for other industries. An S M R of ly, he reported an excess mortality from 123 was reported for chromate workers malignant neoplasms of the gall bladder. for the period of 1937 to 1960.6For 5370 bile ducts, and salivary glands in a uranium workers, all races combined,' cohort of 1877 rubber workers followed an S M R of 119 was found for the period from 1940 through 1968.13 1950 to 1960. Among white male In Great Britain, Case reported In Mortality Experience of a Cohort of Rubber WorkerslAndjelkoviC et a' `International Classification of Diseases, Eighth Revision. + -+Malignant Neoplasms. $ = P < 0.05, p < 0.01. # = p < 0.001 X'Statistical test used to determine probability, p. was Chi Square: I (obnrvedsipected)* upKted `~ntwrutionat Cluriiation d Disara, E i t h Revision. ----+Malignant Neoplasm. t p < 0.05 Sbtistical test used to determine probability, p, was chi %Jar@: x' (observedupected)' =, apeacd burnal Of Occupational MedicineBol. 18, No. 6/June 1976 ~ ~ ~~~~~ ~ 195414 that rubber workers were at increased risk of developing cancer of the bladder in connection with the use of a - and fl -naphthylamine as antioxi- dants. These findings were confirmed by other investigators15 16 who identified particular work areas in the industry at greatest risk. Little evidence has been ob- tained of an excess of bladder tumors among rubber workers in the U.S. However, some studies" I* indicate that there may be an increased risk of bladder cancer among rubber workers in Akron and Massachusetts. A recently reported study in Great Britain4 suggests an excess of cancer of bronchus and lung in tire manufacture, pointing out two problem areas which require further study. These two areas are: (1) moulding, press, autoclave, and pan curemen, and (2) finished goods, packaging, and dispatch. In a study of another group of rubber workers in Akron,S an excess mortality was found for malignant neoplasms of the stomach and prostate, and for neoplasms of the lymphatic and hematopoietic tissue. Unpublished data* on the mortality of rubber workers identifies some specific work areas with an. excess of deaths due to certain causes. For instance, an excess of cancer of bronchus and lung was found for workers who cure tires, while for the excess of leukemia deaths, no particular operations within the industry could be singled out. In most of the above reviewed cohort studies in the rubber and other industries, the overall and cause-specific SMRs were based on data for the general population of the country or a local political subdivision. In the published reports on these studies some discussion was devoted to the phenomenon of the healthy worker effect as an inevitable shortcoming when studying employed populations. However, none of these studies distinguished active worker mortality experience from retired worker mortality experience, particularly in the case of workers who retired before age 65. The study presented here distinguishes the mortality experience of all workers cornbined from that of active workers and that of retired workers. Particular attention is given to the mortality of workers who retired prior to age 65. Results Study Population. - Only the male 389 members of the 1%4 Cohort were included in the present analysis. The distribution of this all male PAR by age group and color is given in Table 1. Of the total 8938 males, 8418 or 94% are white. As of January1, 1964, about onethird of all males belonged to each of the following age groups: 40-54, 55-64, and 65-84. Although this holds for white males, it can be seen that non-white males are present in higher proportion in the younger age groups. The distribution of deaths by age at death for the cohort is found in Table 2. Of the 8418 white males, 2373 or 28% are known to have died during the ten year period. For black males, 135 of 520, or 26% are known to have died during the period of observation. Also, we note that 98.4% of the cohort had 70 or more years of service. Deaths Among Male Hourly Workers. - We first analyzed mortality among all males, white and non-white combined. The main purpose for this analysis was to allow a general comparison of results from this study with those reported in 1974 for another population of rubber workers in Akron, Ohio.5 Definitions for the two PARs are very similar, and the plants under study are in the same city. The two PARs consist of workers from two different companies. Members of one PAR may have been employed, at some previous point in time, with the other company. Also, there are slight differences between the two PARs in average age at hire, raw materials used in the plants, and other factors which may have an effect on mortality. For the above reasons the comparison of mortality experience between the two PARs is not complete, and must be regarded as preliminary. We anticipate doing a more detailed comparison of mortality experience among workers employed in different rubber plants at some future date. SMRs were calculated using the all males combined segment of the 1968 U.S. general population. The all causes SMRs for ages 40-64 and 40-84 are, respectively, 89 and 93. The corresponding SMRs for the other cohort, just referred to, are 93 and 99, respectively. The slight differences in the overall SMRs between the two cohorts cannot be explained with any degree of confidence at present. The fact that the overall SMRs for the cohort reported here are lower than for the other does 'Inkrnaliaul Cbniatim d Diseases. Eighth Revision t = p < 0.05. + = p < 0.01, # = p < 0.001 (ObSancdupsted)' X2Sbtistial tat ised to determine probability, p. was Chi Square: z ,- cxp#td not necessarily imply that male workers due to differences in environmental ex- in this company are healthier than male posure, or in race composition of the workers in the other company. We feel workforce, or perhaps in the method of that a direct comparison of SMRs be- standardization. In addition, the rather tween the two cohorts should not be broad grouping "all other causes" made at this point because of the should be examined in greater detail, method of standardization used. and the analysis should be made race- However, some similarities of the specific for a more accurate comparison. cause-specific SMRs between the two As previously mentioned, further work cohorts of rubber workers are of interest, will be required to complete an inter- such as: (a) An excess of observed plant comparison of the mortality ex- deaths was found for the 40-64 and 40- perience of rubber workers. 84 age ranges in both cohorts for the Deaths Among White Male Hourly following causes: cancers of the stomach Workers. - The well recognized dif- (151) *, large intestine (153), and ference in mortality experience between prostate (185); and lymphosarcoma white, black, and other races prompted (200), diabetes mellitus (250), and us to perform race specific analyses. suicides (E950-E959); (b) A lower than Since the general U.S. population data expected number of deaths was ob- for black and other non-white mal6 served for the two age ranges in both separately were not available at the time cohorts for these causes: cancer of the of this study, we restricted the analysis to respiratory system (160-163), chronic white males only. Calculation of SMRs respiratory disease (490-493), accidents for white male hourly workers was (E800-E949 plus E960-E999), and the based on the 1968 U.S. white male group "all other causes". population. Although the mortality experience in As seen in Table 3, the SMR for all the two cohorts appears to be generally causes of death for white male members similar, the differences should be further of the 1964 Cohort is 92 for the 40-M investigated. These differences may be age group, 95 for the 65-84 age group * International Classification d Diseases. Eighth Revision. 390 Mortality Experience of a Cohort of Rubber WorkerslAndjelkoviC et 'International Clauification d Diseases. E i i t h Revision. tMaligant Neoplasms. - +$ p < 0.05. = p < 0.01. Z Ip < 0.001 x -Statistical test used to determine probability, p. was Chi Squarc: 2 (obsemd.expectedY npectd 'International Clasifiatim d Diseases. E i i RMbon. *Malignant Ncoplasrm. * *='p < 0.05. p < 0.001 xStatistical test used to determine probability. p, was Chi Squaw 2 $b$mdsxpected)z expected journal of Occupational MedicinelVol. 18, No. 6/June 1976 and 94 for the entire cohort. While this table presents SMRs for all causes of death and for the usual grouping of causes suggested by the International Classification of Diseases, Tables 4-6 show SMRs for selected smaller groupings and individual causes. There is an excess of observed over expected deaths in one or both age ranges (40-64 and 65-84) in the PAR for certain neoplasms, discussed in (a) through (d) below: (a) From Table 3, there i s an excess of deaths due to malignant neoplasms (M.N.) of the digestive organs and peritoneum (150-159)in the 65-84 age range. Within this group of diseases (see Table 4),an excess of deaths is observed in both age ranges for M.N. of the stomach (151),large intestine (153),and pancreas (157). For M.N. of the liver (155),there is an excess of deaths in the upper age range. Other causes within this group show a deficit of deaths. (b) We see from Table 3 that in both age ranges there is an excess of deaths due to M.N. of the genitourinary organs (180-189).Within this group of diseases, according to Table 4, there is an excess of deaths due to M.N. of the prostate (185) in both age ranges, and due to M.N. of the bladder (188)and other and unspecified urinary organs (189)in the 65-84 age range. (c) The SMR is elevated for M.N. of other and unspecified sites (190-199)in the 40-64 age range (see Table 3). from Table 4,the same is true for M.N. of the brain and central nervous system (19l192). (d) From Table 3,there is a statistically significant excess of deaths in the 65-84 age range due to neoplasms of the lymphatic and hematopoietic tissue (200209). However, it is more important to notice from Table 5 that each individual cause within this group shows an excess of deaths in at least one of the two age ranges. In the 65-84 age range, the excesses are statistically significant for monocytic leukemia (206) and for the other neoplasms in this group (202, 208, 209),for which the US. figures were not available by individual categories. We shall now consider non-neoplastic diseases for which an excess of deaths was observed. These are discussed in (a) through (9) below. I (a) The SMR is elevated for endocrine, nutritional, and metabolic diseases (240279).in the 40-64 age range (see Table 3). Table 6 shows that for diabetes 391 mellitus (250), which accounts for over 90% of the deaths in this disease group, the SMR i s elevated for both age ranges. (b) From Table 3 we see that the SMR is elevated for diseases of the blood and blood-forming organs (280-289) for the 65-84 age range. (c) For diseases of the nervous system and sense organs (320-389), the SMR is elevated in the 40-64 age range. (d) Also, we see that the SMR is elevated in one or both age ranges for the following: chronic rheumatic heart disease (393-398) (with the excess being statistically significant in the upper age range), hypertensive diseases (400-404), and cerebrovascular disease (430-438), with the excess of deaths in this last group being statistically significant in the 40-64age range. (e) The group "other respiratory diseases" (500-519) shows an excess of deaths in the 40-64 age range. (f) From Table 6, we see that the SMR is elevated for acute myocardial infarction (410) in both age ranges, and for arteriosclerosis (440) in the upper age range. (g) Also, the SMR for suicide (E950E959) is elevated in both age ranges. A deficit of deaths, however, was observed for some diseases of interest discussed below. From Table 3, a deficit of observed deaths for M.N. of the respiratory system (160-163) for both age groups is seen, the deficit in the upper age range being statistically significant. However, if one considers the age range 40-49, an SMR of 148 is found which emphasizes the masking effect of using broad age ranges. Statistically significantly fewer than expected deaths were observed in both age ranges for influenza and pneumonia (470-486) and in the 4064 age range for diseases of the digestive system (520-577). There was also a deficit of deaths due to cirrhosis of the liver (571) in both age ranges, as shown in Table 6. An unpublished study2 of a cohort of white male rubber workers in Akron, Ohio showed results basically similar to those found in this study. The most noticeable difference is that the other study found a deficit of deaths from suicide, while an excess was observed in this study. Deaths Among Retired White Male Hourly Workers. - In the manner of some other studies, we have presented and discussed SMRs for the cohort of 392 'lnternatiml Q a s s i f i d n of Disuses. Eibth Rmsion. - -t p < 0.05, + p < 0.01, f = p < 0.001 xStatistical t & - i d to dderrnine prohbility. p. was Chi Square: >- 2 (obsemdupected)z = erpccted 1 1 ,'Internrtnnal CIass~cstiooai Dim.Eighth Revision. - -t p 0.05, C p < 0.001 2=(0bscrucd+xwcted)z xStatistical test used l o determine probability. p. was Chi Square: GXpeCted 1 hourly workers without differentiating between active and retired work status. However, the possibility of masking effect from using broad categories of age and cause of death should be recognized. This masking effect, which Mancuso12 called a "dilution factor", should also be-considered with respect to other factcrs besides age and cause of death. As he pointed out, an excess of deaths due to a certain cause among workers in a single department of an industry may not be apparent from examination of the death rate due to that cause in the industry as a whole, because of the dilution factor. Accordingly, we shall analyze mortality by work exposure as soon as the necessary data, which have been collected, are made suitable for analysis. We also became concerned that analyzing mortality experience without regard to whether the workers were active or retired might mask important information. For this reason we elected to analyze separately the active and retired mortality experience of white male workers to learn whether mortality ex- perience differs between the two groups. Differences might be manifested in several ways. First, if active workers in the 40-64 age range have a different mortality experience than workers who retiret in this age range, the differentiation of SMRs between the two groups would lead to useful conclusions. Our feeling was that any assessment of the healthy worker effect might better be based on the mortality experience of active workers only. rather than on the combined active and retired population. Hence, the true magnitude of this effect is based on a population which separates presumabb healthier active workers from the nonworking retired individuals who, most likely, have high mortality rates. Fur- thermore, separation of early from disability retirement status will permit assessment of mortality experience among workers who retire prior to age 65, the age for normal retirement, This analysis.will be reported in a subsequent paper. The present study deals with the examination of the above mentioned differences by separating only mortality ex- Mortality Experience of a Cohort of Rubber Workers/Andjelkovic et a1 perience of active workers from that of all retired workers, under 65 years of age. We shall briefly discuss first the mor- tality experience of the active portion of the cohort, then we discuss that of the retired portion. Since workers in the 65&1 age range are virtually all retired, discussion in this section will be limited to the 40-64 age range unless otherwise noted. The overall SMR for active white male hourly workers in the cohort is 61,based on a total of 320 deaths from 35,586 person-years experience. Since the overall SMR for the 40-64 age range is 92 (see Table 3), we conclude that a. less favorable mortality experience among non-active workers in this age range is masking the true SMR for active workers. The healthy worker effect reported in other studies has usually been reflected by an SMR of 80 to 90 for the "working population". The SMR of 61 reported here for the active workers, on which we feel the healthy worker effect could be based, is substantially lower than these other figures and may, in part, be due to selecting out those most heavily exposed or unable to adapt to the conditions through early-voluntary and disability retirement. However, elevated SMRs have been found for certain causes of death in this active worker population. The causes for which an excess of deaths were observed in the active group, based on three or more deaths, are listed below. Cause of Death (1.CD.A Code) SMR M.N. of the stomach (151) 101 2V.N. of the respiratory system (160-163)125 Lvrnphosarcoma (200) Leukemia (204-207) Suicide (950-959) 117 109 115 Note that all SMRs shown above are. for white male active workers aged 40- 64, except the SMR for M.N. of the epiratory system (160-163),which is based on the 40-49 age range. Our expectation that workers who retired prior to the normal retirement age, 65,would experience less favorable mortality than active workers of the Same age range seems to be correct. From Table 7,the overall SMR for retired workers 40-64years of age i s 202, indicating a highly statistically significant excess of deaths (p<O.OOI). The SMR is based on 299 deaths from 5822 person Years experience. With respect to cause specific mor- tality among retired workers, SMRs for the distribution of MRD between two grouped and individual causes of death groups: (1) retirees in the cohort who are presented in Tables 7-10.These are died prior to age 65 during the period of organized in exactly the same manner as observation, and (2) retirees who were Tables 3-6 to facilitate comparison of alive at the end point of the study, or SMRs for retired workers with SMRs for survived to at least age 65. We found all workers. Comparison of results in the that the distribution of MRD was 65-84 age range between the two sets of statistically significantly different be- tables is not very useful, since that com- tween the two groups. We do not feel ponent of the two groups i s virtually that the difference can be presently identical. Comparison of results in the associated with specific causes of death. 40-64age range emphasizes that retired Of course, more detailed analytic studies workers in this age range have a will have to be done to examine this markedly less favorable mortality ex- preliminary finding. perience than do active workers in this age range. This was to be expected, but the magnitude of the excess, an SMR of 202 as compared with 61 for the active Summary workers in the same 40-64age range, is The analysis of mortality was carried the important point. We note again that out for all males regardless of race to these SMRs are for white males and ad- allow comparison of results with those justed by the indirect method for dif- from a similar study of Akron rubber ferences in age distribution. workers.5 Death rates for certain neo- With the exception of influenza and plasms, diabetes mellitus, cerebro- pneumonia (470-486) and those causes vascular disease, arteriosclerosis, and of death for which no deaths were ob- suicide were shown to be in excess for served, there was an excess of observed at least one of the two age ranges, com- deaths for the age group 40-64 for each pared with the U.S. general population. disease category listed in Tables 7-10. The death rate due to all causes was For about half of these disease lower than in the standard population, categories, the excess is statistically as reflected by SMRs of 89 and 93 for significant The most striking of the ex- age ranges 40-64and 40-84,respectively, cesses are for malignant neoplasms of Because of the different mortality ex- the respiratory system (160-163), perience of different racial groups, we genitourinary organs (180-189),and performed race-specific mortality other and unspecified sites (190-199), analyses for this cohort The analysis for and for endocrine, nutritional and white males was done first, as they are metabolic diseases (240-279),chronic by far the largest race-sex group in the rheumatic heart disease (393-398),and cohort. Unfortunately, suitable standard cerebrovascular disease (430-438). population data were not available for These findings emphasize the fact that blacks. The death rate for all causes for white males, after adjusting for age dif- further studies of this population are ferences, was lower in the cohort than in highly important. We plan to investigate the U.S. general population for both the the relationship of a large number of fac- 40-64and 65-84 age groups, as reflected tors to the health and -mortality ex- by SMRs of 92 and 95, respectively. perience of this population- These factors For some causes of death, however, include the following: work history, an SMR of greater than 100 was found reason for retirement prior to age 65,and for one or both age ranges, indicating an causes of disability. We are currently in- excess of deaths. These include the vestigating these relationships in a larger following: malignant neoplasms (M.N.) population involving a cohort of 31,000 of the digestive organs and peritoneum, workers employed any time since M.N. of the genitourinary organs, M.N. January 1, 1951. of other and unspecified sites, and We have made a preliminary in- neoplasms of the lymphatic and vestigation of the relationship between hematopoietic tissue. Among non- work experience- and age at death neoplastic diseases, an excess of. deaths among retirees. Summarizing the work was found for these: endocrine, history by Most Representative Depart- nutritional, and metabolic diseases, a ment (MRD), defined as the department group which includes diabetes mellitus; in which the worker spent the greatest diseases of the blood and blood forming portion of working time, we compared organs; diseases of the nervous system; h n a l of Occupational MedicineNol. 18, No. 6/June 1976 393 L' . , .. and the group other respiratory diseases. almost every cause shown in Tables 7- References Again after adjusting for age, the death 10, the most important result here is the 1.US. Department of Health, Education rate due to suicide was found to be higher for white males in the cohort than in the US. general population. Another study2 of white male rubber workers in great overall excess. As stated in the previous section, detailed studies of the relationship between .this excess of deaths and other factors related to the -and Welfare: Mortality in 1950 by Occupation and Industry. Vital Statistics Special Repom 53: Nos. 1-5:91, 397,June 1961 - September 1963. 2.Monson RR, Nakano KK Mortality Akron, Ohio, showed similar results, except that a lower death rate for suicide was observed in that study population than in the US. general population. worker are necessary. The results of this study show that for certain causes of death and certain groups of workers, the death rate is among rubber workers. Paper presented at Society for Epidemiologic Research Con. ference, Berkeley, GI.,June 1974. 3.Society of Actuaries: 1967 Occupation Study. Chicago, Society of Actuaries, p. 8, We felt it necessary to separately higher than in the US. general popula- 1967. analyze active and retired mortality experience in the 40-64 age range because of the possible masking effect or dilution factor which would be present if no tion after adjustment for age difference among white males. No association of this fact with work exposure can be made at present. However, a com- 4. Fox AI, Lindars DC, Owen R: A survey of occupational cancer in the rubber and cablemaking industries: Results of five year analysis, 1967-1971. B-rir / Ind Medicine 3:140-151, 1974. distinction were made between the two parison of work exposure summarized 5.McMichael AI, Spirtas R, Kupper L L An groups. Of course, all mortality ex- by Most Representative Department epidemiologic study of mortality within a perience beyond age 65 occurs in the (MRD) between two groups of retirees cohort of rubber workers, 1964-72. /OM 16~458-464,1974. i non-working population, since the man- indicates ;that such an association may 6.Taylor FH: The relationship of mortality datory retirement age i s 65. exist. A noticable difference in work ex- and duration of employment as reflected by a The all-causes SMR for white male ac- perience measured by MRD was found cohort of chromate workers. Am / Pub Hlth tive workers 40-64 years of age i s 61, probably due to selecting out the weaker between these two groups: (1) retirees in the cohort who died in the period of 56:218-229,1966. 7.Wagoner JK, Archer VE, Carrol BE, Holaday DA, Lawrence P A Cancer mortality individuals by disability retirement lead- observation prior to age 65, and ( 2 ) patterns among US. uranium miners and ! ing to early death. This figure i s con- retirees who were alive at the end point millers, 1950 through 1962.I Nat Cancer lnst siderably lower than the SMR of 80 to 90 which is generally held to reflect the "healthy worker effect". In spite of this of the study, or survived to at least age 65. The fact that a difference in work experience between these two groups can 32~787-8011,964. 8. Enterline PE: Mortality among asbestos products workers in the United States. Am New York Acad Science 132:156-165,1965. SMR reflecting a low all-causes death be observed indicates to us a need for 9. Lloyd JWC,iocco A: Long-term mortality rate among white male active workers, the death rate for some causes was elevated above that for the US. general population, after adjusting for age dif- further study of the relationship between work experience and health of the worker. study of steelworkers. D M 11:299-310,1969. 10.Heuper WC, Conway W D Chemical Carcinogenesis and Cancers. Springfield, 111: Charles C. Thomas, p. 88, 1964. 11.Mancuso Tf: Tumors of central nervous ferences. These causes for which an ex- system; Industrial considerations. Acta Union cess of deaths was observed in the 40-64 lnternationale contre le Cancer 19:488-489. age range include the following: M.N. of 1963. 12.Mancuso TF, Ciocco A, El-Attar AA: An I the stomach, M.N. of the respiratory epidemiological approach to the rubber in- I system (in the 40-49 age range only), 'Numbers in parentheses which immediatelyfollow dustry. JOM 10:213-232,1968. lymphosarcoma, leukemia, and suicide. In contrast to this, an all-causes SMR of 202 was observed for white male a cause of death indicate the ICDA code(s) c o r e sponding to the stated cause(s) of death. tThese workers include disability retirees, who have a health problem which prevents them from 13. Mancuso TF,Brennan MI: Epidemiologi- cal considerations of cancer of the gallbladder, bile ducts, and salivary glands in the rub ber industry. /OM 12:333-347,1970. retired workers 40-64 years of age. This working, and early retirees, who may have retired 14.Case RAM, Hosker ME Tumour of the indicates a highly statistically significant volunfarily or because of a nondisabling health urinary bladder as an occupational disease in excess of deaths as compared to the US. general population. The excess of deaths among retired white males 40-64 years of age is even greater in comparisori with the active worker population as the standard, with an all-causes SMR of 398. ,Although there is an excess of deaths problem. We thank the following persons for their help in this study: Or. Herman A Tyroler for epidemiologic advice in the conduct of this study; Ms. Anna Colosi and Ms. Mary Donelan and their staff for &la pocessinp and programming; Ms. lacqueline Brown for typing the manuscript; Ms. Pamela Wolf for assistancewith data analysis; and Dr. Frank lones for his contribution in the study design and its irn- the rubber industry in England and Wales. Brit / Prev 5oc Med 8:39-50.1954. 15. Veys C A Aetiology of tumours of the urinary bladder. Urol lnt 24:276-288,1969. 16.Adelstein AM. Occupational mortallw. Cancer. Ann Occupat Hyg 15:53-57,1 9 2 17.Cole P, Hoover R, friedell CH: CX- cupation and cancer of the lower urlnan tract. Cancer 29:1250-1260,1972. among retirees 40-64 years of age for plementation. 18.Cuira AC Bladder carcinoma in rubber workers. / Urol 106:543-552,1971. 394 Mortality Experience of a Cohort of Rubber Workershndjelkovk et