In a population-based study survival after recurrence was compared in three cohorts of patients with a primary diagnosis of breast cancer in 1959, 1969 and 1979, respectively. The use of chemotherapy after recurrence in these cohorts was either none, sporadic or widespread. This allowed a retrospective analysis of the survival impact of chemotherapy. Given the basic assumption that the natural history of breast cancer and the influence of endocrine therapy have not changed significantly during the 20-year period covered by the study, our data suggest that chemotherapy in recurrent breast cancer prolongs survival by 9.5 months in patients who survive more than 2 weeks from the start of treatment for this recurrence.