Post-remission therapy in acute myeloid leukemia (AML) remains problematic. It has been demonstrated that younger patients can maintain longer complete remissions (CR) with aggressive post-remission therapies after induction treatment: allogeneic (allo), autologous (auto) stem cell transplantation (SCT), or intensive chemotherapy (ICC). The purpose of our study was to identify the most important randomized and controlled studies comparing these three therapeutic options, in order to draw conclusions and possible suggestions for post-remission therapy of AML, according to the evidence based medicine (EBM) rules. We performed an exhaustive analysis of the literature, searching either in electronic databases or among the references of the identified articles (hand searching). We searched the MEDLINE computer database for reports from 1985 through January 2005 and selected for analysis the clinical trials conducted over adults affected by newly diagnosed AML aged less than 65 years. The study design had to satisfy strict methodological criteria and must consider global mortality and/or disease free survival as primary outcomes. Overall we found 7750 papers; by using the limits "clinical trial" as publication type, "all adults 19+ years", we were able to select 344 papers. Among these, a further selection was made, based on two main clinical queries: 1) is auto-SCT superior to ICC/no other therapy in improving DFS and/or OS in adult AML patients in first CR? 2) is allo-SCT superior to auto-SCT/other therapeutic options in improving DFS and/or OS in adult AML patients in first CR? Concerning the first query, a possible advantage of auto-SCT over ICC was not clearly supported by data from clinical trials; there is no evidence that auto-SCT is superior in terms of OS to chemotherapy. Nevertheless, the reported TRM has been significantly reduced within the past years. Thus, the percentage of patients suitable for auto-SCT in CR has increased. Moreover, the scarce data concerning the comparison between auto-SCT and chemotherapy in different subsets of patients are unable to suggest a differentiated approach in patients with high-risk, standard-risk or low-risk AML. Data from the literature show that patients with unfavorable risk disease are more often addressed to allo-SCT and patients with low-risk disease receive more often intensive consolidation chemotherapy. Concerning the second query, interpretation of data from the main prospective studies about the role of allo-SCT in previously untreated AML is not easy. The first problem is the lack of real randomized clinical trials; in fact, according to the reported studies, AML patients generally receive allo-SCT on the basis of donor availability (the so called "genetic randomization"). The second problem is the frequent absence of intention to treat analysis. Despite methodological limitations, it was possible to compare allo-SCT with auto-SCT on a donor versus no-donor analysis and within risk groups. No overall benefit of allo-grafting on survival was demonstrated by any trial. In conclusion, the EBM approach highlighted the limitations observed in the published studies concerning consolidation therapy in AML; some suggestions, emerging from non-randomized, as well as randomized studies, are adequate, but not conclusive. This point, coupled with the intrinsic complexity to study AML biological heterogeneity, is probably a major obstacle to draw conclusive evidences for consolidation therapy in AML. These observations should plan to address new randomized studies on AML therapy; however, due to the emergence of genetic subgroups and new drugs targeting specific abnormalities, these trials should probably be designed directly focusing on the single entities. In this way, the cure of AML could eventually become the cure of each specific AML subset with its peculiar biological, molecular and prognostic features.