Allogeneic BMT is presently recognized as one of the reference anti-leukemic treatments for AML patients. It could be a very useful therapy in CR1 AML if it is efficient as an early form of consolidation after induction therapy. Later procedures may be of lesser importance since they concern a population which may have already been cured by chemotherapy. Alternatively it is not solved if very early BMT gives patients sufficient therapy and provides lower toxicity. To answer this question we analyzed data from the Société Française de Greffe de Moelle (SFGM) on a sub-group of patients according to the following criteria: (1) Allogeneic BMT carried out on patients with AML in CR1; (2) after January 1985; (3) conditioning with CyTBI; (4) GVHD prophylaxis with MTX-CsA; (5) interval between diagnosis and BMT < 100 days. Forty-two fulfilled all criteria. Age was 31 +/- 8 years and M/F ratio was 19/23. WBC at diagnosis were 36 +/- 49 x 10(9)/l and four patients needed two induction courses to achieve CR1. Median follow-up is now 51 (24-116) months. Twenty-two patients developed a grade > or = 2 acute GVHD. Early transplant mortality at 1 year is no higher than 14%. Two patients died of secondary malignancies at 3.5 and 6.5 years after BMT. Finally 5-year (and 7-year) probabilities for relapse, survival, leukemia and event-free survival are respectively 17% (17%), 71% (62%), and 71% (71%) and 68% (59%). These data indicate the feasibility of such an approach with low mortality and low relapse rates providing a good long-term outcome. This should prompt an invitation to initiate the search for an HLA-identical sibling when AML is diagnosed in a young patient eligible for allo-BMT.