Temporal variation in the association between benzene and leukemia mortality

BACKGROUND: Benzene is a human carcinogen. Exposure to benzene occurs in occupational and environmental settings. OBJECTIVE: I evaluated variation in benzene-related leukemia with age at exposure and time since exposure. METHODS: I evaluated data from a cohort of 1,845 rubber hydrochloride workers. Benzene exposure-leukemia mortality trends were estimated by applying proportional hazards regression methods. Temporal variation in the impact of benzene on leukemia rates was assessed via exposure time windows and fitting of a multistage cancer model. RESULTS: The association between leukemia mortality and benzene exposures was of greatest magnitude in the 10 years immediately after exposure relative rate (RR) at 10 ppm-years = 1.19; 95% confidence interval (CI), 1.10-1.29]; the association was of smaller magnitude in the period 10 to < 20 years after exposure (RR at 10 ppm-years = 1.05; 95% CI, 0.97-1.13); and there was no evidence of association > or = 20 years after exposure. Leukemia was more strongly associated with benzene exposures accrued at > or = 45 years of age (RR at 10 ppm-years = 1.11; 95% CI, 1.04-1.17) than with exposures accrued at younger ages (RR at 10 ppm-years = 1.01; 95% CI, 0.92-1.09). Jointly, these temporal effects can be efficiently modeled as a multistage process in which benzene exposure affects the penultimate stage in disease induction. CONCLUSIONS: Further attention should be given to evaluating the susceptibility of older workers to benzene-induced leukemia.