Treatment of childhood acute lymphoblastic leukemia has included glucocorticosteroids for almost 50 years. Glucocorticoids are the subject of renewed interest. In one randomized trial, deferral of glucocorticosteroids from the initial month of induction therapy to the second month of therapy decreased event free survival despite preservation of remission induction rate. Dexamethasone in induction and maintenance provides a better event free survival than prednisone for standard risk patients in an isotoxic comparison even though all patients received dexamethasone in Delayed Intensification (protocol II). In a third report, patients with prior glucocorticosteroid therapy who achieved remission with subsequent multiagent therapy had a relapse rate similar to that of patients in second remission after failure of multiagent therapy. In vitro and in vivo response of leukemic cells to glucocorticosteroids is highly predictive of outcome. At relapse, loss of in vitro sensitivity to glucocorticosteroids is common and out of proportion to the loss of sensitivity to other agents. Glucocorticoid induced cell kill does not require p53 function. Investigation of leukemic cell lines finds that glucocorticosteroid resistance is most commonly linked to altered receptor number or function. Not all ligands are equivalent. Cortivazol, a pyrazolosteroid, may bind to altered receptor in some cases and induce apoptosis in dexamethasone resistant leukemic cells. Host response to exogenous glucocorticosteroid also varies. Associations between host sensitivity, disease sensitivity, and glucocorticosteroid side effects like avascular necrosis of bone remain to be investigated.