BACKGROUND. Most clinical trials for acute leukemia have reported results after 2-3 years of follow-up. Comparisons between the original data and longer-term follow-up data may be of interest, particularly with regard to promising new therapies. METHODS. In 1996, survival data were updated from three prospective, randomized comparisons of idarubicin and daunorubicin that began in 1984 and 1985. These were trials of the Memorial Sloan-Kettering Cancer Center (MSKCC), the U.S. Multicenter Study Group, and the Southeastern Cancer Study Group (SEG). The original results of these trials were reported in 1991 and 1992. RESULTS. The original results of the SEG trial demonstrated no significant difference between idarubicin and daunorubicin. The updated survival analysis showed similar results. The MSKCC trial revealed a significant advantage of idarubicin compared with daunorubicin in both the original and the updated analyses. The U.S. Multicenter trial found a significant difference favoring idarubicin in the original analysis, but the difference was not significant in the updated analysis. CONCLUSIONS. It is essential that the median length of follow-up be dearly stated in any clinical trial. When the results obtained with a particularly promising new drug or procedure are presented early in the course of study (within 1-2 years), the investigators should strongly consider a repeat evaluation after an additional 3-5 years of follow-up.