We compared the response to prolonged treatment with dexamethasone in two groups of ventilated preterm infants: one whose chest radiographs showed homogenous opacity, and one whose radiographs showed cystic changes and hyperinflation. Forty-nine infants were treated with dexamethasone for 42 days, beginning when they were 15 to 27 days old, had no evidence of sepsis or patent ductus arteriosus, and had experienced no decrease in ventilator support for 24 hours. Forty-nine controls were selected who met these criteria for dexamethasone treatment. All had birthweights of 500 to 1250 g. Two radiographs made between 14 and 28 days of age were reviewed. Among infants with homogeneous opacity (19 dexamethasone, 26 controls), dexamethasone was associated with fewer days on assisted ventilation (median [interquartile range]: 7 [3-11] versus 23 [9-40]; p = 0.001). Among those with cystic changes and hyperexpansion (30 dexamethasone, 23 controls), no difference was found between dexamethasone treated infants and controls (17 [7-34] versus 32 [16-47]; p = 0.9). Thus, the effect of dexamethasone on days of ventilation was attenuated in infants with cystic changes and hyperinflation.