Re: Consolidation therapy with autologous bone marrow transplantation in adults with acute myeloid leukemia: a meta-analysis.

The comprehensive meta-analysis reported by Nathan et al. (1) used comparable methods and produced results similar to those in our recent meta-analysis (2), which dealt with the same group of patients and clinical trials. Despite these similarities, the authors of the two studies reached different conclusions. Nathan et al. concluded that the data did not support the routine use of autologous bone marrow transplantation (ABMT) in adult acute myeloid leukemia patients in first complete remission. Indeed, Nathan et al. concluded that their data supported the use of non-myeloablative chemotherapy for patients who do not have a matched sibling donor because there was no statistically significant difference in overall survival between the treatment groups. By contrast, we believe the data do support the routine use of ABMT in adult acute myeloid leukemia patients in first complete remission. With the current increasing use of peripheral blood stem cells in ABMT, the development of new and safer transplant techniques, and the availability of supportive therapy, toxicity and treatmentrelated mortality is now substantially lower [i.e., 0%–6% (3)] compared with the high rate of 14% reported in 1998 by Cassileth et al. (4). As a result, we expect that disease-free survival and overall survival will be longer for patients who receive ABMT than for those who do not. To test this assumption, we calculated the expected death rates for the ABMT arms of each of the six studies included in both meta-analyses (1,2). Our calculations were designed to assess three different simulations concerning treatment-related mortality: The first analysis assumed similar death rates in both groups (ABMT versus chemotherapy or no further treatment); the second and third simulations assumed death rates of 3% and 5%, respectively, for patients who received ABMT with peripheral blood stem cells (3). The number of expected deaths in this arm was calculated according to the following formula: (number of patients receiving ABMT) (assumed percentage of treatmentrelated mortality) (actual number of patients who died in complete remission) (actual death rate of relapsed patients) (actual number of patients