Osteonecrosis is one of the most common, serious, toxicities resulting from the treatment of acute lymphoblastic leukemia. In recent years, pediatric acute lymphoblastic leukemia clinical trials have used discontinuous rather than continuous dosing of dexamethasone in an effort to reduce the incidence of osteonecrosis. However, it is not known whether discontinuous dosing would compromise antileukemic efficacy of glucocorticoids. Therefore, we tested the efficacy of discontinuous dexamethasone against continuous dexamethasone in murine models bearing human acute lymphoblastic leukemia xenografts (n = 8 patient samples) or murine BCR-ABL+ acute lymphoblastic leukemia. Plasma dexamethasone concentrations (7.9 to 212 nM) were similar to those achieved in children with acute lymphoblastic leukemia using conventional dosages. The median leukemia-free survival ranged from 16 to 59 days; dexamethasone prolonged survival from a median of 4 to 129 days in all seven dexamethasone-sensitive acute lymphoblastic leukemias. In the majority of cases (7 of 8 xenografts and the murine BCR-ABL model) we demonstrated equal efficacy of the two dexamethasone dosing regimens; whereas for one acute lymphoblastic leukemia sample, the discontinuous regimen yielded inferior antileukemic efficacy (log-rank p = 0.002). Our results support the clinical practice of using discontinuous rather than continuous dexamethasone dosing in patients with acute lymphoblastic leukemia.