Long-term results of a multicentre randomised, comparative phase III trial of CHOP versus CNOP regimens in patients with intermediate- and high-grade non-Hodgkin's lymphomas. Novantrone International Study Group.

59 previously untreated patients with intermediate- or high-grade, stage II-IV non-Hodgkin's lymphoma (NHL) were entered into an open-label, randomised, multicentre study to compare the efficacy and safety of CHOP (cyclophosphamide, doxorubicin, vincristine and prednisolone) with that of CNOP (cyclophosphamide, mitoxantrone, vincristine and prednisolone). 10 patients refused treatment following randomisation. The remaining 349 patients received either the CHOP or CNOP regimen every 3 weeks for a maximum of six to eight cycles. The randomisation procedure was violated for 34 patients treated at two study centres. Data from these 34 patients were analysed separately for efficacy and survival. Data from the remaining 325 patients, 164 assigned to CHOP and 161 to CNOP, were used in the major efficacy and survival analyses. Of these 325 patients, 263 (81%) met the eligibility criteria of the protocol. Supplementary analyses of data from these 263 patients, 132 assigned to CHOP and 131 to CNOP, were conducted for efficacy and survival. Data from all 349 treated patients were analysed for safety. In the 325 randomised patients, the overall objective response rate was not significantly different between the two groups (chi 2 test, P = 0.35). The CHOP regimen had a 51% (83/164) complete remission (CR) rate compared with 40% (64/161) for the CNOP regimen (P = 0.05). Among those with CR, the median time to response was 104 days with the CHOP regimen and 77 days with the CNOP regimen, and the median duration of CR was 667 and 1833 days, respectively. The median time to progression was 449 days for CHOP patients and 564 days for CNOP patients. The median survival time was 932 days for CHOP patients and 1801 days for CNOP patients, with a risk of death on CNOP relative to CHOP of 0.93% (95% confidence interval 0.68-1.27). After 5 years, 50% of patients in the CNOP arm and 40% of patients in the CHOP arm were still alive; these differences between treatment groups were not statistically significant. The median time to treatment failure (TTF) was 285 days for patients on the CHOP arm and 282 days for patients on the CNOP arm. Separate analyses of 263 eligible randomised patients, and 34 patients in whom the randomisation procedure was not followed, yielded similar results for remission rate, TTF, duration of CR and estimated overall survival. The incidence of non-haematological events, such as severe nausea and vomiting (P < 0.01), mucositis (P < 0.05) and alopecia (P < 0.001), were significantly lower in were significantly lower in patients treated with CNOP as compared with those who received the CHOP regimen. The incidence of cardiovascular toxicity of any severity was similar in the two groups. While severe and potentially life-threatening neutropenia occurred more frequently in patients treated with CNOP compared with CHOP (0.05 > P > 0.10), the incidence of infection of any severity was similar in both arms. We conclude that CHOP and CNOP regimens were both efficacious in patients with previously untreated aggressive NHL.