PURPOSE The aims of this study were to quantify the prospects of salvage treatment of patients who did not undergo transplantation in first complete remission (CR1) and to assess the contribution of allograft in second complete remission (CR2) with respect to major risk groups. This evaluation can inform the decision whether to offer a transplant in CR1. PATIENTS AND METHODS Of 8,909 patients who entered the Medical Research Council AML10, AML12, and AML15 trials, 1,271 of 3,919 patients age 16 to 49 years who did not receive a transplant in CR1 relapsed. Of these patients, 19% are alive beyond 5 years compared with 7% of patients who relapsed after an allograft in CR1. Overall survival and the contribution of a transplant in CR2 were assessed overall and by cytogenetic risk group by using Mantel-Byar analysis. Results Fifty-five percent of patients who relapsed entered CR2. This percentage varied by risk group as follows: favorable (82%), intermediate (54%), adverse (27%), and unknown (53%), which resulted in 5-year survivals of 32%, 17%, 7%, and 23%, respectively. Sixty-seven percent of remitters received an allotransplant that delivered superior survival compared with patients who did not receive a stem-cell transplant (42% v 16%). A more-stringent assessment of a transplant by using delayed-entry (Mantel-Byar) analysis confirmed the benefit of transplant overall and within intermediate and adverse risk groups but not the favorable subgroup. CONCLUSION Successful salvage treatment of patients who do not undergo transplantation in CR1 and relapse can be achieved in 19% of patients, which is improved by a transplant except in favorable risk disease. This result suggests that, for intermediate-risk patients in particular, equivalent overall survival can be achieved by delaying transplantation until after relapse, which would require many fewer transplants.