Bladder cancer incidence among workers exposed to o-toluidine, aniline and nitrobenzene at a rubber chemical manufacturing plant

Abstract

Background: An earlier investigation found increased bladder cancer incidence among workers at a rubber chemical manufacturing plant that used o-toluidine, aniline and nitrobenzene. The cohort was expanded to include additional workers (n=1875) and updated through 2007 to assess bladder cancer with improved exposure characterisation.

Methods: Work histories were updated and exposure categories and ranks were developed for o-toluidine, aniline and nitrobenzene combined. Incident cancers were identified by linkage to six state cancer registries. Residency in time-dependent cancer registry catchment areas was determined. SIR and standardised rate ratios for bladder cancer were calculated by exposure category and cumulative rank quartiles for different lag periods. Cox regression was used to model bladder cancer incidence with estimated cumulative rank, adjusting for confounders. Indirect methods were used to control for smoking.

Results: Excess bladder cancer was observed compared to the New York State population (SIR=2.87, 95% CI 2.02 to 3.96), with higher elevations among workers definitely exposed (moderate/high) (SIR=3.90, 95% CI 2.57 to 5.68), and in the highest cumulative rank quartile (SIR=6.13, 95% CI 2.80 to 11.6, 10-year lag). Bladder cancer rates increased significantly with estimated cumulative rank (10-year lag). Smoking only accounted for an estimated 8% elevation in bladder cancer incidence.

Conclusions: Bladder cancer incidence remains elevated in this cohort and significantly associated with estimated cumulative exposure. Results are consistent with earlier findings in this and other cohorts. Despite other concurrent chemical exposures, we consider o-toluidine most likely responsible for the bladder cancer incidence elevation and recommend a re-examination of occupational exposure limits.

Figure 1

Bladder cancer HRs from different dose-response models including log-linear (solid curve), log-square root (short-dashed curve), categorical (based on quintiles of the case distribution), 3-knot restricted cubic spline (medium-dashed curve, with knots at the 10th, 50th and 90th percentiles of the case distribution), and 4-knot restricted cubic spline (long-dashed curve, with knots at the 5th, 35th, 65th and 95th percentiles of the case distribution). Open circles indicate the distribution of cumulative rank for bladder cancer cases (n=37). RCS, restricted cubic spline.