The prognostic role of drug-induced amenorrhea (DIA) was restrospectively evaluated in 221 out of 254 consecutive premenopausal patients treated with adjuvant CMF or a CMF-containing regimen; 33 patients were eliminated because of lack of menstrual data. All patients had metastatic axillary nodes; drug regimens were: CMF x 9 courses +/- Tamoxifen (TM) and CMF x 6 courses; median age was 43 (range 26-54). Premenopausal status was defined as last normal menses within the 6 weeks preceding initiation of chemotherapy: DIA as cessation of menses for at least 3 months not later than 3 months from the end of chemotherapy. DIA occurred in 166,221 (75.1%) patients and was strictly related to the age of the patients; also, the older the patients the shorter the time required to develop DIA. At median follow up of 69 months, Mantel-Byar analysis showed a longer disease free survival (DFS) for patients who developed DIA as compared with non amenorrheic women (P less than 0.001). DIA prognostic value was independent of age, number of involved nodes, tumour size and number of CMF cycles, as assessed by the Cox model (RH 0.43, 95% C.I. 0.24-0.77), in which DIA was entered as a time dependent covariate.