Very low birth weight infants are frequently treated with a prolonged course of dexamethasone to decrease the duration of ventilation. The efficacy of dexamethasone treatment in individual babies is variable. We hypothesized that early pulmonary function change after dexamethasone would distinguish responders from non-responders, thereby allowing dexamethasone to be discontinued in the non-responders. Ventilated infants 15 to 25 days old and with birth weights < 1500 gms were randomized to a 42 day tapering course of dexamethasone (DEX) or placebo (CON). At randomization, patent ductus arteriosus and sepsis were not present. Pulmonary function tests for mechanical breaths were performed with a commercially available machine on day of enrollment (day 0) and day 1 and 2 after receiving the drug in 22 DEX and 22 CON infants. The groups were similar in terms of birth weight (DEX: 777.7± 41.2, CON: 796.6 ± 49.0), gender (DEX: 64% male, CON: 50%), race (DEX: 59% white, CON: 45%). Days of ventilation (DEX: 23 ± 6, CON: 33 ± 8) and days on oxygen (DEX: 85 ± 16, CON: 127 ± 30) after enrollment were decreased in the DEX patients. On day 2 versus day 0, dynamic compliance (mL/cmH20/kg) was increased in DEX (0.495 v 0.711) but not CON (0.493 v 0.516), p < 0.005; tidal volume (mL/kg) increased in DEX (6.5 v 8.3) but not CON (7.1 v 7.3), p < 0.005; and peak inspiratory pressure minus end expiratory pressure (cm H2O) decreased in DEX (14.3 v 12.8) but not CON (15.4 v 15.0), p < 0.005. Other pulmonary function parameters were unchanged. In the DEX group, 15/22 patients had a > 33% increase and 6/22 had no improvement in dynamic compliance by day 2. Change in pulmonary function tests on day 2 did not predict which patients would respond to dexamethasone; there was no correlation found between change in compliance on day 2 and subsequent days of ventilation or oxygen. Pulmonary compliance should be monitored closely after starting dexamethasone because of the large increase in most infants; however, an early increase in compliance cannot be used to identify infants responding to dexamethasone.