Bone marrow transplantation is useful in AML. Results of chemotherapy and transplantation are compared in Table 4. Transplantation is the preferred treatment in individuals who fail chemotherapy. Transplantation is also likely to be superior or comparable to chemotherapy in individuals less than 30 years of age in first remission. Transplantation in older individuals in first remission is controversial, but it is unlikely that these results are inferior to results with chemotherapy. Transplants from donors other than HLA-identical siblings are less well investigated but may be a reasonable alternative in young individuals in first relapse or second remission, particularly if recipient and donor share most HLA antigens. Autotransplants are difficult to evaluate critically and should be considered investigational in individuals in second or later remission for whom a suitable donor is unavailable. Autotransplants in first remission should be restricted to controlled clinical trials, since their efficacy is otherwise inevaluable. It is uncertain whether in vitro treatment of the bone marrow is necessary in the context of autotransplantation; again, controlled trials are required. Bone marrow transplantation from an HLA-identical sibling is a useful therapy in individuals with ALL who relapse despite chemotherapy. Individuals undergoing transplantation in second or later remission or in relapse have a survival rate superior to those treated with chemotherapy; these data are summarized in Table 5. One important unresolved issue in ALL is whether children with high-risk ALL and adults should receive transplants in first remission. This answer will, to a considerable extent, depend on results achieved with chemotherapy alone. If intensive chemotherapy produces 50-70% disease-free survival in these individuals, it is unlikely that transplantation will be superior. If, however, alternative chemotherapy results are inferior, transplantation may be useful. Clearly, controlled clinical trials are needed. Results of transplants from donors other than HLA-identical siblings are less certain but this approach may be considered in selected young individuals who fail chemotherapy. Autotransplants should also be considered in this setting but not in individuals in first remission. It is likewise uncertain if in vitro treatment of the bone marrow is useful; this must be addressed in controlled trials. The data reviewed indicate an important role for bone marrow transplantation in the therapy of the acute leukaemias. We have attempted to precisely define its use in various disease states.(ABSTRACT TRUNCATED AT 400 WORDS)