CHOP (cyclophosphamide, doxorubicin, vincristine and prednisolone) has been the standard chemotherapy regimen used for diffuse large cell lymphomas for over 30 years. Idarubicin is a 4-demethoxy-anthracycline analogue of daunorubicin that has proven activity in non-Hodgkin lymphoma, and has been reported to cause less cardiotoxicity. We therefore initiated a randomised trial of standard dose CHOP versus CIOP (cyclophosphamide, idarubicin, vincristine and prednisolone), in which doxorubicin 50 mg/m2 was substituted by idarubicin 10 mg/m2, a dose thought to have equivalent anti-lymphoma activity. This trial was closed prematurely after 211 patients had completed therapy when a lower complete remission (CR) rate was noted in the CIOP arm. The formal results with long-term follow up are now reported. Overall response rate was 84% in the CHOP arm and 78% in the CIOP arm, CR rates were 70% and 52% respectively in all patients (P=0.013) and 73% and 52% respectively for the eligible patients (P=0.0084). At a median of 8 years follow-up, 4-year progression-free survival for all patients was 56% in the CHOP arm and 40% in the CIOP arm (P=0.0096). Overall survival for all patients was 65% in the CHOP arm and 56% in the CIOP arm (P=0.14). Results for eligible patients were comparable. CIOP containing idarubicin at a dose of 10 mg/m2 is clearly inferior to standard CHOP.