Exogenous corticoids are known to be potent inhibitors of linear growth in children. We investigated the mechanisms underlying growth failure by evaluating growth hormone (GH) release during short-term high-dose prednisone treatment (40 mg/m2/day given orally in 3 divided doses) and 7 days after steroid withdrawal in 7 prepubertal children (4 males, 3 females, age range 3-12 years), affected by acute lymphoblastic leukemia. Patients also received weekly administrations of vincristine (1.5 mg/m2 i.v.), daunomycin (20 mg/m2 i.v.) and L-asparaginase (6,000 IU/m2 i.m.). Corticoid therapy suppressed GH secretion during deep sleep as well as in response to arginine, insulin and GH-releasing hormone (GHRH) administration. A significant recovery of GH responsiveness after drug discontinuation was observed during deep sleep (14.03 +/- 3.47 vs. 1.49 +/- 0.43 ng/ml, p less than 0.025) as well as in response to arginine (13.63 +/- 2.73 vs. 4.95 +/- 1.54 ng/ml, p less than 0.025) and GHRH (32.62 +/- 4.59 vs. 7.27 +/- 3.52 ng/ml, p less than 0.005) but not to insulin (7.12 +/- 0.88 vs. 4.47 +/- 0.96 ng/ml, p = NS). Insulin-like growth factor 1 levels during deep sleep (0.61 +/- 0.13 IU/ml/min) were found to be low in the course of steroid therapy and did not increase after drug withdrawal (0.41 +/- 0.07 IU/ml/min). Our preliminary data suggest that recovery of adrenergic response to insulin does not immediately follow corticosteroid discontinuation.