In their interesting paper, Schlenk et al 1 report that patients age 61 and over with untreated AML lived longer if randomized to receive all-trans retinoic acid (ATRA) in addition to chemotherapy. Since 85% of the 242 participants have died and since the median follow-up time in the remaining 15% is almost 3 years, the conclusions seem unlikely to change substantively. The result was independent of covariates such as cytogenetics and whether the patient had de novo or secondary AML. Since relapse rates were similar in the two arms while event free-survival was superior in the ATRA arm, the major cause underlying the improvement in survival seems to have been a higher CR rate, which reflected primarily an increased sensitivity to chemotherapy; the treatment-related mortality rate was also somewhat lower in the ATRA arm. Although less stressed by the authors, patients had longer remissions and, after covariate adjustment, longer survival if, in a second randomization performed after completion of first consolidation therapy, they were assigned to receive one cycle of i.v. etoposide and idarubicin rather than 1 year of the same drugs given orally and at lower dose. The reduction in mortality was relatively greater in patients given i.v. etoposide/idarubicin (rather than oral etoposide/idarubicin) than in patients given ATRA (rather than no ATRA). The effects of the two randomizations were independent, although it is not altogether clear whether the principal benefit of i.v. etoposide/ idraubicin was in patients not randomized to ATRA.A principal focus of the authors' informative and excellent discussion is the difference in results between their trial and two other randomized trials, which did not find that administration of ATRA was beneficial. The authors note that their patients were older than those of Burnett et al (Blood 2002; 100: 155a, abstract # 582) and were less likely to have secondary AML than those of Estey et al. 2