The purpose of this study was to compare four methods of treatment for stage III-IV Hodgkin's disease. Between January 1972 and September 1976, 266 patients with stage IIIB, IVA, and IVB Hodgkin's disease from 21 cancer treatment centers across Canada were registered as eligible; 40 were found to be ineligible. Of the 226 remaining patients, only seven were followed for less than 10 years. All patients received three courses of mechlorethamine, vincristine, procarbazine, and prednisone (MOPP) chemotherapy, which induced a complete response (CR) in 36%; an additional 42% obtained adequate disease control. Patients were randomly assigned to (1) treatment with radiation to the abdomen and mantle (group AX3, 62 patients) or (2) continue their treatment with an additional three courses of MOPP (group A, 105 patients). For the A group, a second randomization took place 3 months later (regardless of status at that time) to (1) no further treatment (AC6, 23 patients), (2) radiotherapy to the abdomen and mantle (AX6, 48 patients), or (3) maintenance chemotherapy at 3-month intervals for 1 year (AC10, 26 patients). The survival of AX3 patients was somewhat better than for the A group, but the difference was not significant (P = .0565). However, there was a significant interaction (P = .0029) between age and treatment, so that among patients less than 30 years of age, the survival of the A group was better, whereas for older patients, treatment with AX3 resulted in improved survival. Age itself remained a significant prognostic factor for survival after controlling for the amount of radiotherapy delivered to the abdomen and the dose intensity of vincristine for the first three courses of chemotherapy. The addition of radiation therapy to MOPP significantly reduced the frequency of nodal relapses. These results suggest that combined modality therapy may be beneficial for some patients with Hodgkin's disease and that age must be carefully considered in interpreting the results of clinical trials in Hodgkin's disease.