The objective of the study was to determine whether administering doxapram by infusion to the very low birthweight infant, prior to extubation during the first 3 weeks of life, would increase the incidence of successful extubation. The study patients, 56 infants of less than 1251 g birthweight and less than 30 weeks' gestation, were entered in the first 3 weeks of life when lung disease had started to improve. A randomized blinded trial was performed, with infants receiving 3.5 mg kg(-1) doxapram bolus, followed by an infusion at 1 mg kg(-1) h(-1), or placebo. Weaning from positive pressure ventilation was standardized and extubation occurred after a 12 h trial of an intermittent mandatory ventilation (IMV) rate of 6 breaths min(-1), if PCO2 < 55 mmHg, pH > 7.26, and FiO2 < 0.45. Study drug was continued for 48 h postextubation, and the infants were placed on nasopharyngeal continuous positive airway pressure (CPAP) for 72 h postextubation. Extubation failure within the first 72 h after extubation was objectively defined in terms of acidosis (pH < 7.26), hypercarbia (PCO2 > 55 mmHg), excessive oxygen requirement (FiO2 > 0.8) or frequent apnoea (more than three in 12 h, or more than two requiring face mask IMV in 24 h). No difference was noted in the frequency of successful extubation between the groups. Fifteen infants in each group were successfully extubated before the 10th day of the study. In conclusion, when given in accordance with this protocol doxapram does not increase the likelihood of successful extubation in the very low birthweight infant. Increasing successful extubations in this group of infants will require other strategies.