OBJECTIVE The purpose of this study was to evaluate the risk of spontaneous abortion among the wives of male workers occupationally exposed to benzene. METHODS The wives of 823 men working in two chemical plants at the time of the study were asked to complete a questionnaire describing their pregnancies. The analysis of the 1739 pregnancies that ended in a spontaneous abortion or a birth is presented. The firms' payroll records provided all workers' employment history, including dates. Benzene exposure, graded at two levels (< 5, > or = 5 ppm), was determined for every job, so that benzene exposure for each worker's entire professional life (at these companies) could be assessed. This information was linked to the dates of the pregnancies reported in the questionnaires to enable the exposure status of each pregnancy to be defined (1270 non-exposed and 274 exposed). The frequency of spontaneous abortion, defined as the number of spontaneous abortions divided by the total of spontaneous abortions and births was evaluated. RESULTS When adjusted for tobacco consumption, mother's age and pregnancy order, the odds ratio of the association between paternal exposure to approximately 5 ppm of benzene and the risk of spontaneous abortion was close to and statistically not different from unity (OR = 1.1; 95% CI (0.7-1.8). CONCLUSIONS In this study paternal exposure to benzene did not increase the risk of spontaneous abortion.