Two different ventilation techniques were compared in a seven-centre, randomised trial with 181 preterm infants up to and including 32 completed weeks gestational age, who needed mechanical ventilation because of lung disease of any type. Technique A used a constant rate (60 cycles/min), inspiratory time (IT) (0.33s) and inspiratory: expiratory ratio (I:E) (1:2). The tidal and minute volume was only changed by varying peak inspiratory pressure until weaning via continuous positive airway pressure. Technique B used a lower rate (30 cycles/min) with longer IT (1.0 s). The I:E ratio could be changed from 1:1 to 2:1 in case of hypoxaemia. Chest X-rays taken at fixed intervals were evaluated by a paediatric radiologist and a neonatologist unaware of the type of ventilation used in the patients. A reduction of at least 20% in extra-alveolar air leakage (EAL) or death prior to EAL was supposed in infants ventilated by method A. A sequential design was used to test this hypothesis. The null hypothesis was rejected (P = 0.05) when the 22nd untied pair was completed. The largest reduction in EAL (-55%) was observed in the subgroup 31-32 weeks of gestation and none in the most immature group (< 28 weeks). We conclude that in preterm infants requiring mechanical ventilation for any reason of lung insufficiency, ventilation at 60 cycles/min and short IT (0.33 s) significantly reduces EAL or prior death compared with 30 cycles/min and a longer IT of 1 s. We speculate that a further increase in rate and reduction of IT would also lower the risk of barotrauma in the most immature and susceptible infants.