To evaluate high-dose therapy and autologous or allogeneic blood or marrow transplantation (BMT) for mantle cell lymphoma, patients receiving BMT for newly diagnosed or relapsed mantle cell lymphoma were identified through the registry at Johns Hopkins. The pathologic diagnostic criteria were reviewed, and details of the presentation, transplant procedure, and survival outcomes were determined. Fifty-eight patients were identified, of whom 64% underwent transplantation in first remission and 12% had primary induction failure. Nineteen patients (one third) received an allograft. Preparative regimens consisted of cyclophosphamide in combination with either busulfan or total body irradiation. On multiple regression analysis, transplantation after 1 or more relapses (hazard ratio, 2.98; P = .02), primary induction failure (hazard ratio, 5.39; P = .002), and allogeneic transplantation (hazard ratio, 3.03; P = .007) were associated with an inferior event-free survival (EFS). However, EFS curves were not statistically different for autologous and allogeneic BMT performed in first remission, with an estimated 3-year EFS approaching or equaling 70%. Primary induction failure and residual bone marrow involvement were the only statistically significant predictors of relapse on multiple regression analysis. At 3 years, the estimated EFS for the entire cohort after BMT was 51%, the probability of relapse was 31%, and the overall survival was 59%. The benefit of autologous or allogeneic BMT for mantle cell lymphoma is thus most apparent when transplantation is performed in first remission. Whether allogeneic BMT ultimately confers an advantage because of a graft-versus-lymphoma effect remains to be determined.