It is still controversial how to treat elderly patients with acute myeloid leukaemia (AML), and results have been poor with most regimens. We report the long-term results of a randomised study performed by the Leukaemia Group of Middle Sweden during 1984-88 comparing two intensive chemotherapeutic drug combinations. Ninety patients >or=60-yr old with untreated AML were randomly allocated to treatment with daunorubicin, cytosine arabinoside (ara-C), and thioguanine (TAD) (43 patients) or a combination in which aclarubicin was substituted for daunorubicin (TAA) (47 patients). Forty-four patients (49%) entered complete remission (CR), 22/43 (51%) in the TAD group and 22/47 (47%) in the TAA group (ns). The CR rate in patients <or=70 yr of age was 30/42 (71%) and in patients >70 yr 14/48 (29%) (P<0.0001). Early death within 30 d after treatment initiation was more often seen in patients >70 yr than in patients <or=70 yr of age, 40% and 12%, respectively (P<0.005). The median cause-specific survival time was 178 d in the total patient group, and the 2-, 5-, and 10-yr survivals were 22%, 11%, and 8%, respectively. The cause-specific survival was not significantly different between the two treatment arms. At long-term follow-up >or=10 yr after inclusion of the last patient, 5/90 patients (one in the TAD group and four in the TAA group, respectively) were still alive, four in continuous complete remission and one in second complete remission. Thus, both treatment regimens appear to have similar efficacy, with a relatively high complete remission rate, and a reasonable survival as compared to other studies including some long-term survivors. However, early deaths are still numerous, particularly in patients above 70 yr of age, and the relapse rate is substantial.