One hundred thirty-four assessable patients with stage II-IV large-cell lymphoma (LCL) were treated with the combination chemotherapy regimen methotrexate with leucovorin, bleomycin, doxorubicin, cyclophosphamide, vincristine, and dexamethasone (m-BACOD) between July 1981 and May 1986. The m-BACOD regimen substituted moderate-dose methotrexate (200 mg/m2 x 2) for the high-dose methotrexate used in the preceding M-BACOD regimen; all other drugs were administered as with m-BACOD. Eighty-two patients (61%) in the completed m-BACOD trial achieved a complete response (CR). With a median follow-up of 3.6 years, 62 patients (76%) continue in CR. Predicted survivals of 1, 3, and 5 years for the entire m-BACOD group are 80%, 63%, and 60%, respectively, with a 5-year disease-free survival (DFS) of 74% for the patients who achieve CR. The results obtained with m-BACOD are comparable with those obtained in the preceding M-BACOD trial, which now has a median follow-up of 8.0 years. The reduction in methotrexate dosage in m-BACOD patients was not associated with an increased incidence of CNS relapse. Long-term follow-up of the 215 M/m-BACOD patients indicates that the regimens are not associated with an increased incidence of secondary malignancy. Prolonged follow-up also indicates that advanced-stage patients have a persistent rate of late relapse of about 7.0% per year for years 2 to 5 of their follow-up and that stage II patients have an approximate 2.1% per year rate of late relapse. Application of the previously described prognostic factor model to the 215 M/m-BACOD patients from the completed trials identifies a high-risk group of patients with a CR rate and predicted 5-year survival (38% and 24%, respectively) that are significantly worse than those of the group as a whole (65% and 57%, respectively).