We analysed the incidence of amenorrhoea and its association with outcome in a cohort of 1127 premenopausal women with breast cancer randomized to International Trial V (formerly Ludwig V). For 552 patients without axillary lymph node involvement, one course of perioperative cytotoxic drugs was compared with no-adjuvant chemotherapy. For 575 patients with node-positive disease, a single course of cytotoxic chemotherapy was compared with a prolonged treatment (6 or 7 courses). Amenorrhoea was defined as having no menstrual bleeding for a 3-month interval within the first 9 months after surgery. Amenorrhoea was observed in 21% of the 199 patients with node-negative breast cancer who received no adjuvant therapy, 31% of the 353 node-negative patients who had a single course of cytotoxic therapy, 31% of the 188 patients with node-positive disease who had the same short-duration therapy, and 68% of the 387 node-positive patients who had a prolonged adjuvant therapy. Amenorrhoea was associated with an increased disease-free survival (DFS) only in the patients with prolonged cytotoxic therapy: 4-year DFS % (+/- s.e.) was 68% +/- 3% vs. 61% +/- 5% for the amenorrhoea and the no-amenorrhoea groups, respectively, (p = 0.05). In contrast, the comparison between prolonged therapy and one single course among node-positive patients showed a much larger treatment effect (4-year DFS 66% vs. 38%, p less than 0.0001). We conclude that although cytotoxics-induced amenorrhoea is associated with a better outcome, it is unlikely that this form of endocrine manipulation is the main mechanism of response to adjuvant systemic chemotherapy.