This aim of the acute myeloid leukaemia (AML)-R trial was to compare sequential (Seq) ADE (cytarabine, daunorubicin, etoposide) with standard (Std) ADE as remission re-induction treatment and to assess any benefit of cyclosporine (CSA) as a multidrug resistance modulator in refractory/relapsed AML patients. Seq ADE, based on the concept of Timed Sequential Therapy, comprised the same drugs as Std ADE but given at higher doses and in a different sequence. Between 1992 and 1997, 235 patients with relapsed (175) and refractory (60) AML were entered: 170 were randomized between Std versus Seq ADE and 213 between CSA versus no CSA. CSA was initially given at a dose of 5 mg/kg/d and increased to 10 mg/kg/d in the latter part of the trial. Overall, the complete remission (CR) rate was 43%, with Std ADE being significantly better than Seq ADE (54% versus 34%, P = 0.01). CR rates did not differ between the CSA and no CSA arms (41% versus 45%, P = 0.6). Overall, 3 year disease-free survival (DFS) of remitters was 16%, with a relapse risk of 70%. DFS was not significantly different between the chemotherapy or the CSA arms. Overall, 3 year survival was 8%. Survival with Std ADE was significantly better than with Seq ADE (12% versus 6%, P = 0.03). CSA did not affect overall survival, except in patients > or = 60 years, who fared worse on CSA (P = 0.0003). No difference in haematological toxicity between the chemotherapy or CSA arms was seen. Survival was better with longer first CR duration (P < 0.0001). We conclude that Std ADE was superior to Seq ADE for CR achievement and survival, with no benefit with CSA, at the doses used in this study.