In standard-risk acute promyelocytic leukemia, recent results have shown that all-trans retinoic acid plus arsenic trioxide combinations are at least as effective as classical all-trans retinoic acid plus anthracycline-based chemotherapy while being less myelosuppressive. However, the role of frontline arsenic trioxide is less clear in higher-risk acute promyelocytic leukemia, and access to arsenic remains limited for front-line treatment of standard-risk acute promyelocytic leukemia in many countries. In this randomized trial, we compared arsenic, all-trans retinoic acid and the "classical" cytarabine for consolidation treatment (after all-trans retinoic acid and chemotherapy induction treatment) in standard-risk acute promyelocytic leukemia, and evaluated the addition of arsenic during consolidation in higher-risk disease. Patients with newly diagnosed acute promyelocytic leukemia with a white blood cell count <10x10 <sup>9</sup> /L, after an induction treatment consisting of all-trans retinoic acid plus idarubicin and cytarabine, received consolidation chemotherapy with idarubicin and cytarabine, arsenic or all-trans retinoic acid. Patients with a white blood cell count >10x10 <sup>9</sup> /L received consolidation chemotherapy with or without arsenic. Overall, 795 patients with acute promyelocytic leukemia were enrolled in this trial. Among those with standard-risk acute promyelocytic leukemia (n=581), the 5-year event-free survival rates from randomization were 88.7%, 95.7% and 85.4% in the cytarabine, arsenic and all-trans retinoic acid consolidation groups, respectively (P=0.0067), and the 5-year cumulative incidences of relapse were was 5.5%, 0% and 8.2%. (P=0.001). Among those with higher-risk acute promyelocytic leukemia (n=214), the 5-year event-free survival rates were 85.5% and 92.1% (P=0.38) in the chemotherapy and chemotherapy plus arsenic groups, respectively, and the corresponding 5-year cumulative incidences of relapse were ...