High dose dexamethasone is frequently used for the treatment of neonatal respiratory conditions and to facilitate weaning from mechanical ventilation in preterm, very low birth weight infants. However, very little is known about the severity, site, and duration of steroid-induced hypothalamic-pituitary-adrenal axis suppression in this category of patients. Twenty-three preterm, very low birth weight infants who received a full 3-week dose-tapering course of dexamethasone were prospectively studied, with a human CRH stimulation test performed at three different times: before the start of steroid treatment (week 0), immediately after the course (week 3), and 4 weeks after stopping dexamethasone (week 7). Plasma ACTH and serum cortisol concentrations were measured at 0 (baseline), 15, 30, and 60 min. Immediately after the steroid course (week 3), both basal and poststimulation plasma ACTH and serum cortisol concentrations were markedly suppressed. The hormone concentrations at 0, 15, 30, and 60 min in week 3 were significantly lower than their corresponding levels in week 0 (P < 0.0001 for both ACTH and cortisol) and week 7 (P < 0.0001 and P < 0.005 for ACTH and cortisol, respectively). In contrast, when the hormone levels in week 7 were compared to their corresponding concentrations in week 0, only the 60 min serum cortisol concentration in week 7 was significantly lower (P = 0.02). The currently used dosage of dexamethasone caused severe pituitary-adrenal suppression immediately after treatment, but substantial recovery of the endocrine axis was observed 4 weeks after discontinuation of therapy. Although the recovery appeared to be earlier with the pituitary center, both pituitary and adrenal glands were capable of mounting a biochemically adequate response to exogenous human CRH stimulation at this stage. Steroid replacement therapy may be desirable at a time of stress in the immediate posttreatment period, but it would seem unnecessary 1 month after stopping dexamethasone treatment.