A SYSTEMATIC COLLABORATIVE OVERVIEW OF RANDOMIZED TRIALS COMPARING IDARUBICIN WITH DAUNORUBICIN (OR OTHER ANTHRACYCLINES) AS INDUCTION THERAPY FOR ACUTE MYELOID-LEUKEMIA

A collaborative overview using individual patient data, has been perfo rmed to compare idarubicin versus daunorubicin or other anthracyclines , when used with cytosine arabinoside as induction chemotherapy for ne wly diagnosed acute myeloid leukaemia. There were 1052 patients in fiv e trials versus daunorubicin, 100 in one trial versus doxorubicin, and 745 in one trial versus zorubicin. In the trials of idarubicin versus daunorubicin, early induction failures were similar with the two trea tments (20% idarubicin v 18% daunorubicin; P = 0.4), but after day 40 the later induction failures were fewer with idarubicin (17% v 29%: P < 0.0001). Therefore complete remission rates were higher with idarubi cin (62% v 53%: P = 0.002). Among remitters, fewer of the patients all ocated to idarubicin relapsed (P = 0.008) but slightly more died in re mission, leading to a non-significant benefit (P = 0.07) in disease-fr ee survival. Overall survival in these five trials was significantly b etter with idarubicin than with daunorubicin (13% v 9% alive at 5 year s; P = 0.03). There was a trend (P = 0.006 for remission rate) for the benefit of idarubicin over daunorubicin to decrease with increasing a ge. There were no significant differences in outcome in the small tria l comparing idarubicin versus doxorubicin, or in the large trial compa ring idarubicin versus zorubicin. The induction regimens based on idar ubicin achieved, in the particular circumstances of the trials reviewe d here, better remission rates and better overall survival than those based on daunorubicin.