The outcome of treatment for a first relapse of Hodgkin's disease after primary chemotherapy was analyzed in 80 patients. They were divided into four groups: group 1 (n = 24) had initially been treated with three cycles of (mechlorethamine, vincristine, prednisone, and procarbazine [MOPP]) and wide-field irradiation therapy; group 2 (n = 25) had six cycles of MOPP; group 3 (n = 15) and group 4 (n = 16) both initially received MOPP/ABVD (MOPP plus doxorubicin, bleomycin, vinblastine, and dacarbazine) or MOPP/ABV hybrid, but group 3 received conventional salvage regimens whereas group 4 was treated with high- dose chemotherapy and autologous bone marrow transplantation as salvage therapy (n = 16). Freedom from second failure (FF2F) was used as the major endpoint. Actuarial FF2F for all patients was 38% after a median follow-up of 75 months for patients who were alive. Risk factor analysis was performed on the 71 patients who had been treated with curative intent. The presence or absence of any one of three risk factors had a strong negative impact on outcome: stage IV disease at primary diagnosis, B symptoms at relapse, or a time from primary treatment to relapse of less than 1 year. Actuarial FF2F at 5 years was 17% in the group of patients with one or more of these three factors present (n = 49). If none of these factors was present, FF2F was 82% (n = 22) (P less than .001). Even high-dose chemotherapy and autologous bone marrow transplantation were not able to overcome the negative impact of one or more risk factors (FF2F = 19%, n = 12). The outcome of salvage treatments depends most on the presence or absence of these three risk factors and less on the type of salvage treatment. Patients with none of these risk factors present have an excellent outcome if they are treated with non-cross-resistant chemotherapy, or radiotherapy, or both. Novel approaches are needed for patients with one or more of these factors present. Reports on salvage treatments for Hodgkin's disease in first relapse after primary chemotherapy should include data on the proportion of patients having stage IV disease at diagnosis, B symptoms at relapse, and a time from primary treatment to relapse of less than 1 year.