In 1991 and 2000, the National Institute for Occupational Safety and Health (NIOSH) published the results of a retrospective cancer incidence study and a mortality study, respectively, of the 1,749 workers employed at a rubber chemical manufacturing plant in NY. These workers were believed to have been occupationally exposed to ortho-toluidine (o-toluidine) and aniline, suspected bladder carcinogens in animals. Thirteen cases of bladder cancer were observed versus 3.61 expected [standardized incidence ratio (SIR) = 3.6; 90 percent confidence interval (CI) = 2.1-5.7]. Among workers considered to have definite exposure (DE) to o-toluidine and aniline, the bladder cancer excess risk was 6.5-fold, while for those with possible exposure (PE), the excess risk was nearly 4-fold. The risk of cancer was strongly associated with increased duration of employment in the department where o-toluidine was used, with a 27-fold excess risk among workers with 10 or more years of employment. The original exposure characterization for the cohort utilized a surrogate measure of exposure based on departments in which each worker was ever employed, comparing bladder cancer incidence among exposure groups within the cohort. The exposure assignments included: 1) Definitely Exposed (DE), workers who had ever worked in the department where o-toluidine and aniline were used, even if they had periods of employment outside of that department; 2) Possibly Exposed (PE), workers who had ever worked in maintenance, shipping, janitorial, or yard work; and 3) Probably Not Exposed (PNE), all other workers who were not likely to have been exposed to o-toluidine and aniline. NIOSH is in the process of updating the bladder cancer mortality and incidence studies among current and former employees at the plant, since nineteen new bladder cancer cases within the original study cohort have been reported. To support the epidemiology studies, the exposure assessment has also been updated. The goal of the exposure assessment update is to provide the most accurate possible account of which job titles were exposed to o-toluidine, aniline, and/or nitrobenzene, and approximate the extent of exposure during defined eras of production. The updated assessment includes additional years of employment (1988-2005) and additional cohort members were also identified since the original study. In support of the exposure assessment revision, documents related to the project on file at NIOSH were reviewed. A site visit to the company plant was conducted, and additional exposure information was obtained from a plant walkthrough, interviews with employees, management and union representatives, and review of additional records. In addition, the company provided available job title descriptions and electronic exposure data on aniline, o-toluidine, and nitrobenzene exposure from 1976 to present (2005). Finally, replies to a list of 64 questions seeking clarification of worker-reported jobs and departments, differences between similar job titles in the same department, and confirmation about the possible exposures of some jobs were provided by Company and Union representatives. The original exposure groups were adjusted to reclassify some departments based on information identified after the initial publications. Alternate revised exposure groups (and codes) were also created based on each department-job title combination as opposed to only department. The alternate groups account for exposure intensity and regularity of exposure as defined by the following codes: 1) Definitely exposed moderate/high and regularly (DER); 2) Probably exposed low and regularly (PER); 3) Probably exposed low and irregularly/occasionally (PEI); and 4) Probably not exposed (PNE). An approximated rank of “relative” exposure level to each exposure combination (i.e., for each department-job-year) was also assigned. The ranks are based relatively on the quantitative exposure levels, as available or by professional judgment. A ranking scale of 0 to 10 was qualitatively selected to provide enough latitude to characterize exposures of different groups based on relative exposures interpreted between jobs and departments over time. The numerical ranking scale will be applied to each cohort member by multiplying the exposure rank by duration for job held based on comprehensive work histories, to obtain individual cumulative exposure estimates. Cohort members will be ranked by cumulative estimated exposure, from lowest to highest, and tertiles or quartiles of cohort members will be compared. Inclusion of job title in addition to department is considered to be a necessary modification to improve surrogate exposure measures. Job duration will be included in the exposure characterization, to build a cumulative measure of exposure. Because exposure is ongoing, but to a lesser degree, the four level grouping scale might allow for better differentiation between exposure groups without reducing statistical power from having too many exposure groups. The cumulative numerical scale has the advantage of providing a continuous spectrum of exposure approximations for the entire cohort. These revised exposure classification schemes will be used to analyze the updated bladder and other cancer incidence and mortality data.
Susan M. Viet, PhD, CIH, WESTAT, 1650 Research Boulevard, Rockville, MD 20850