Preliminary independent reports of the Italian GIMEMA and the Spanish PETHEMA trials for newly diagnosed acute promyelocytic leukemia (APL) indicated a similarly high antileukemic efficacy in terms of complete remission and disease-free survival rates. To better investigate these studies and the prognostic factors influencing relapse risk, this study analyzed the updated results of 217 patients with PML/RAR alpha-positive APL enrolled in GIMEMA (n = 108) and PETHEMA (n = 109). All patients received identical induction (AIDA schedule) and maintenance. For consolidation, GIMEMA patients received 3 courses including idarubicin/cytarabine, mitoxantrone/etoposide, and idarubicin/cytarabine/thioguanine, whereas PETHEMA patients received the same drugs and dose schedule of idarubicin and mitoxantrone with the omission of nonintercalating agents. Depending on whether molecular relapses were classified as censored or uncensored events, the 3-year Kaplan-Meier estimates of relapse-free survival (RFS) for the combined series were 90 +/- 2% and 86 +/- 2%, respectively. Minor differences observed between the 2 patient cohorts were negligible. Multivariate regression analysis of RFS showed that initial leukocyte (WBC) and platelet counts were the only variables with independent prognostic value. The resulting predictive model for RFS demonstrated its capability of segregating patients into low-risk (WBC count </= 10 x 10(9)/L, platelet count > 40 x 10(9)/L), intermediate-risk (WBC count </= 10 x 10(9)/L, platelets </= 40 x 10(9)/L), and high-risk (WBC count > 10 x 10(9)/L) groups, with distinctive RFS curves (P <.0001). The conclusions are that omission of nonanthracycline drugs from the AIDA regimen is not associated with reduced antileukemic efficacy and a simple predictive model may be used for risk-adapted therapy in this disease. (Blood. 2000;96:1247-1253)