We investigated the possible etiological role of adjuvant chemotherapy and postoperative radiation therapy in the development of leukemia. A case-control design with four controls matched to each case of leukemia from a cohort of women who had been treated for breast cancer during the years from 1970 to 1985 was used. Thirteen (0.23%) of the women in this cohort developed leukemia over varying lengths of follow-up time, ranging from 3 to 14 years. A higher percentage of the leukemia cases previously had adjuvant chemotherapy compared to their matched controls (54% versus 13%). The relative odds estimate of developing leukemia after chemotherapy compared to no chemotherapy was 14.8 (95% C.I. (1.8; 125.3) p < 0.01). This estimate and the test of statistical significance was based on the likelihood function for matched sets with one case and more than one control. Approximately the same percentage of leukemia cases as their controls had received postoperative regional radiation therapy (28% versus 23%). No significant association was found between postoperative radiation therapy and development of leukemia. A combination of adjuvant chemotherapy and postoperative radiation therapy was found more frequently in the leukemia cases than in their matched controls (33% versus 9%). The leukemia developing in patients having received adjuvant chemotherapy was frequently therapy resistant, resulting in a short survival.