The aim of this study was to assess if administration of bronchodilator via a metered dose inhaler (MDI), rather than by a nebulizer, avoided the early paradoxical deterioration in lung function but, resulted in equally effective late bronchodilation. Fifteen children were studied at a median postnatal age of 9 months (range 9–18), all had been born prematurely at a median gestational age of 27 weeks (range 23–31). Lung function was measured by plethysmography before and 10 min after normal saline and 5 and 15 min after salbutamol given via an MDI and a nebulizer in random order. At 5 min, compared to baseline values, airways resistance (Raw) deteriorated by 16% after nebulized salbutamol but improved by 3% following salbutamol by the MDI (P<0·03). At 15 min Raw improved by 14% following nebulized salbutamol and 15% after salbutamol via the MDI, there was no significant difference in the magnitude of bronchodilation between the two methods of administration. Our results therefore suggest, that as the early paradoxical deterioration in lung function is usually avoided by administering salbutamol via an MDI, this should be the preferred method of administration.