Between 1987 and 1991, the British National Lymphoma Investigation randomized 459 patients with non-Hodgkin's lymphoma with a large-cell component to either CHOP or the PACEBOM regimen.Four hundred fifty-nine eligible patients were included in this trial, four hundred one with diffuse large-cell lymphoma and fifty-eight with diffuse mixed-cell lymphoma according to the Working Formulation. Two hundred twenty-six patients were randomized to receive CHOP and two hundred thirty-three to receive PACEBOM. The two arms of the trial were well balanced for all potential prognostic factors.The complete remission rate with PACEBOM was 64% and with CHOP 57% (NS). At eight years, the actuarial cause specific survival (CSS) in the PACEBOM arm is 59% compared to 49% in the CHOP arm (P = 0.09). Patients < 50 years of age fared significantly better in the PACEBOM arm both for CSS and overall survival (P = 0.002) and the CSS was also significantly improved in stage IV disease (P = 0.02).The mature data from this trial suggest that an etoposide-containing multi-agent weekly regimen can be superior to CHOP.