A systematic collaborative overview of randomized trials comparing idarubicin with daunorubicin (or other...

A collaborative overview, using individual patient data, has been performed to compare idarubicin versus daunorubicin or other anthracyclines, when used with cytosine arabinoside as induction chemotherapy for newly diagnosed acute myeloid leukaemia. There were 1052 patients in five 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 treatments (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 idarubicin (62% v 53%; P = 0.002). Among remitters, fewer of the patients allocated to idarubicin relapsed (P = 0.008) but slightly more died in remission, leading to a non-significant benefit (P = 0.07) in disease-free survival. Overall survival in these five trials was significantly better with idarubicin than with daunorubicin (13% v 9% alive at 5 years; P = 0.03). There was a trend (P = 0.006 for remission rate) for the benefit of idarubicin over daunorubicin to decrease with increasing age. There were no significant differences in outcome in the small trial comparing idarubicin versus doxorubicin, or in the large trial comparing idarubicin versus zorubicin. The induction regimens based on idarubicin achieved, in the particular circumstances of the trials reviewed here, better remission rates and better overall survival than those based on daunorubicin.