One hundred twenty-five assessable patients with advanced-stage Hodgkin's disease were randomized to receive mechlorethamine, vincristine, procarbazine, and prednisone (MOPP) or MOPP alternating with lomustine (CCNU), doxorubicin, bleomycin, and streptozocin (CABS). The median follow-up is 7.7 years. The complete response rate was 60 of 66 MOPP-treated patients (91%) and 54 of 59 MOPP/CABS-treated patients (92%) (difference not significant). The level of the disease-free survival curve at longest follow-up is 65% for MOPP-treated patients and 72% for MOPP/CABS-treated patients (difference not significant). The overall survival at 12 years is projected at 68% for MOPP-treated patients and 54% for MOPP/CABS-treated patients (difference not significant). Thus, there were no significant differences in efficacy between MOPP and MOPP/CABS. However, MOPP/CABS was more emetogenic than MOPP, and four MOPP/CABS-treated patients went on to develop secondary acute leukemia. No MOPP-treated patients developed leukemia. High initial erythrocyte sedimentation rate (ESR) and high platelet counts adversely affected treatment outcome. MOPP-treated patients who received greater than 81% of the projected dose intensity of vincristine over the first three cycles had significantly improved disease-free survival rates over those receiving less than 81%. MOPP/CABS-treated patients who received greater than 82% of the projected dose intensity of vincristine had significantly better overall survival than those who received less than 82%. Disease-free survival on both arms was significantly better in patients who received greater than 84% of the projected dose intensity of all agents. The effect of dose intensity was particularly apparent in patients with poor prognostic factors where those who received greater than 84% of the projected dose intensity of all agents had significantly improved disease-free and overall survival.