Between 1990 and 2001, 68 patients with advanced Hodgkin's disease (HD) and 86 patients classified as low-/intermediate-grade B non-Hodgkin's lymphoma (NHL) were reported to the Austrian Stem Cell Transplantation Registry (ASCTR). Following autologous stem cell transplantation (SCT) for HD, overall survival was 56% [95% confidence interval (CI): 40-72%] with a disease-/progression-free survival of 49%, reaching a plateau at 5 years. Using multivariate Cox regression analysis BEAM conditioning (carmustine, cytarabine, etoposide and melphalan) was predictive for favourable outcome, better disease-/progression-free survival and a significantly lower risk for relapse. The cumulative incidence of relapse was 30%, even for patients in complete remission at time of SCT. The cumulative risk for developing a secondary malignancy increased continuously over time, achieving 20% at 7 years and 46% at 10 years with previous radiotherapy as the only risk factor in the multivariate analysis. Overall survival for NHL patients was 45% (95% CI: 26-64%) with a disease-/progression-free survival of 26% at 7 years. In the multivariate Cox regression analysis stage of disease at time of SCT was the most powerful parameter for overall survival, disease-/progression-free survival and relapse. Mantle cell lymphoma, greater than or equal to three lines of previous therapy, and a conditioning regimen other than BEAM were also predictive for death. The main reason for treatment failure was relapse (cumulative incidence 54-75%). Because of the high risk of relapse/progression in both disease categories and the additional high rate of second malignancies in HD patients, allogeneic stem cells should be considered a valuable alternative for selected patients. The efficacy of allotransplantation following reduced-intensity conditioning should be tested in randomised trials.