From March 1988 to March 1991, 19 French bone marrow transplant (BMT) centers participated in a prospective randomized trial comparing two conditioning regimens for patients with chronic myeloid leukemia transplanted in first chronic phase with an HLA identical sibling donor. A total of 120 consecutive patients were randomized to receive either 120 mg/kg of cyclophosphamide followed by total body irradiation (CY-TBI; n = 55) or 16 mg/kg of busulfan followed by 120 mg/kg of CY (BU-CY; n = 65). Two different TBI regimens were used. Thirteen patients received a 10-Gy single-dose TBI (SDTBI), and 42 received a fractionated TBI (FTBI). Median time between diagnosis and BMT was 315 days. Overall 5-year actuarial survival was 62.9% (65.8% +/- 12.5% for CY-TBI and 60.6 +/- 11.7% for BU-CY; P = .5), and overall disease-free survival was 55% (51% +/- 14% for CY-TBI and 59.1% +/- 11.8% for BU-CY; P = .75). All patients conditioned with CY-TBI experienced sustained engraftment; in contrast, 4 of 65 patients conditioned with BU-CY rejected the graft (P = .18). There was no significant statistical difference between the two groups regarding transplant-related mortality (29% for CY-TBI and 38% for BU-CY; P = .44). So far, with a median follow up of 42 months, 11 patients have relapsed; 9 relapses occurred after CY-TBI, mostly after FTBI (8 of 9) and 2 after BU-CY (P = .02). The actuarial risk of relapse was 4.4% +/- 6.7% after BU-CY, 11.1% +/- 20.8% after SDTBI, and 31.3% +/- 18.1% after FTBI (P = .039). In addition, independently of the conditioning regimen, the increase of posttransplant immunosuppression in 16 patients with an anti-interleukin-2 receptor monoclonal antibody (MoAb) in addition to a short course of methotrexate and cyclosporine was shown to increase the actuarial risk of relapse (57% +/- 30% with MoAb v 9% +/- 7.3% without MoAb; P = .001). We conclude that BU is an acceptable alternative to TBI for patients with chronic myeloid leukemia in first chronic phase receiving BMT from HLA identical sibling donors. Both BU-CY and CY-TBI regimens gave similar transplant-related mortality, and the antileukemic efficiency of BU-CY regimen was either similar or even higher than that of CY-TBI.