OBJECTIVE To evaluate the efficacy of a flexible low-intensity combination chemotherapy (FLICC) protocol in a multicentre, phase II study of elderly patients with acute myeloid leukaemia (AML). METHOD Twenty-five patients aged 61-78 years (median 70 years) with de novo (n = 17) or secondary (n = 8) AML (cytogenetic risk: favourable 2, intermediate 18, adverse 2, unknown 3) from eight Australian centres were enrolled. Treatment comprised mitoxantrone 6 mg/m(2) intravenously daily for 3 days, cytarabine 10mg/m(2) subcutaneously every 12 hours for 7-14 days and etoposide 100mg orally for 7-14 days. RESULTS The treatment was generally well tolerated, and 13 patients (52%) achieved a complete remission (CR). One patient achieved a partial remission but died on day 28 due to pneumonia. Five patients (20%) had no response, whilst six (24%) died on or before day 30 and so were not evaluable. The median overall survival (OS) was 6.5 months, and the median remission duration was 7.7 months. Estimated 1-year survival was 32%, but patients achieving CR had an estimated 1-year survival of 64%, whereas none in the non-CR group survived to 1 year. Two of the CR patients have survived beyond 2 years. OS was significantly shorter in the adverse cytogenetic risk group of patients compared with the favourable- and intermediate-risk groups, with the rates of death relative to the adverse group being 0.02 and 0.08 in the favourable- and intermediate-risk groups, respectively. There was no significant association between CR rate and pre-existing myelodysplasia or the presence of multilineage dysplasia. The median durations of significant neutropenia (<0.5 x 10(9)/L) and thrombocytopenia (<20 x10(9)/L) with the first course of treatment in the 19 evaluable patients were 19 days (range 12-26) and 11 days (range 1-25), respectively. The median duration of stay in the hospital was 27 days (range 14-42). These values were much shorter for the second course of treatment: 6 days, 5 days and 15 days, respectively. CONCLUSION The findings of this multicentre, phase II study validate the previously reported single-institution experience with the FLICC protocol in elderly patients with AML. The clinical outcome with this protocol is comparable to those reported with more aggressive anti-leukaemia protocols.