OBJECTIVE: To evaluate cancer mortality among United States workers exposed to toluene diisocyanate (TDI) in the manufacture of polyurethane foam. METHODS: This cohort mortality study included 4611 men and women employed in four polyurethane foam plants for at least three months between the late 1950s and 1987. The mortality experience of the cohort was then compared with that of the general United States population. RESULTS: Current and past industrial hygiene data indicated that air concentrations in 1984-5 were below the current United States standard of 0.04 mg/m3 but exceeded the standard before 1980. Mortality ratio (SMR) 2.78, 95% confidence interval (95% CI) 0.57-8.13) and non-Hodgkin's lymphoma (SMR 1.54, 95% CI 0.42-3.95) were increased, but not significantly. There was one male breast cancer. However, breast cancer was not increased in women (SMR 0.74). No other cancer category had an increased number of deaths compared with the general population. Only non-Hodgkin's lymphoma and Hodgkin's disease showed a possible relation with time since first employment and no cancer death category showed a strong relation with duration of employment. Mortality from non-malignant respiratory disease was not increased (SMR 0.86). CONCLUSIONS: This young cohort has few deaths and short follow up. The findings are therefore not conclusive. Further years of follow up will enable better evaluation of mortality.