Very low birth weight (VLBW) infants frequently are treated with dexamethasone (DEX) to decrease the duration of ventilator dependency. Clinical trials of this intervention have been limited by small sample size or by allowing patients initially assigned to the control group to “cross over” (i.e., be treated with DEX). To address these limitations we randomized 118 ventilator-dependent infants to either a 42-day tapering course of DEX (as described by Cummings et al. NEJM, 1989) or saline, as placebo. All infants were 15 to 25 days old and had birth weight < 1500 gms. At randomization they did not have a patent ductus arteriosus and did not have sepsis. The primary study outcome was days on assisted ventilation after study entry. DEX-treated infants and controls (CON) were similar in terms of birth weight (data are medians with range in parentheses) [DEX: 765 gms (456-1362); CON: 726 gms (515-1324)], gestational age [DEX: 26 wks (23-29); CON: 25 wks(23-31), gender [DEX: 51% male; CON: 54%], and race [DEX: 67% white; CON: 56% white]. Survival rates through one year of age (adjusted for prematurity) were 50/57 (88%) in the DEX group and 45/61 (74%) in the CON group [p = 0.07; Fisher's exact test]. DEX treatment was associated with fewer days on assisted ventilation (data are medians with range in parentheses) [DEX: 14 (1-155); CON: 25 (1-196); p = 0.01; logrank test], and fewer days on supplemental oxygen [DEX: 76 (23-497); CON: 114 (16-668); p = 0.01; logrank test], but was not associated with fewer days of hospitalization [DEX: 93 (28-171); CON: 98(16-226); p = 0.3; logrank test]. The effects of DEX on days of assisted ventilation and days on oxygen persisted in multivariate analyses controlling for birth weight, gestational age, gender, and race. Rates of sepsis after study entry were similar [DEX: 8/57 (14%); CON: 5/61 (8%); p = 0.4; Fisher's exact test].
