OBJECTIVES: To compare change in lung volume (DeltaV(L)), using respiratory inductive plethysmography, time to recover pre-suction lung volume (t(rec)) and the cardiorespiratory disturbances associated with open suction (OS) and closed suction (CS) in ventilated infants. DESIGN: Randomised blinded crossover trial. SETTING: Neonatal intensive care unit. PATIENTS: Thirty neonates, 20 receiving synchronised intermittent mandatory ventilation (SIMV) and 10 high-frequency oscillatory ventilation (HFOV, four receiving muscle relaxant). INTERVENTIONS: OS and CS were performed, in random order, on each infant using a 6FG catheter at -19 kPa for 6 seconds and repeated after 1 minute. OUTCOME MEASURES: DeltaV(L), oxygen saturation (Spo(2)) and heart rate were continuously recorded from 2 minutes before until 5 minutes after suction. Lowest values were identified during the 60 seconds after suction. RESULTS: Variations in all measures were seen during CS and OS. During SIMV no differences were found between OS and CS for maximum DeltaV(L) or t(rec); mean (95% CI) difference of 3.5 ml/kg (-2.8 to 9.7) and 4 seconds (-5 to 13), respectively. During HFOV t(rec) was longer during OS by 13 seconds (0 to 27) but there was no difference in the maximum DeltaV(L) of 0.1 mV (-0.02 to 0.22). A small reduction in SpO(2) with CS in the SIMV group mean difference 6% (2.1 to 9.8) was the only significant difference in physiological measurements. CONCLUSIONS: Both OS and CS produced transient variable reductions in heart rate and Spo(2). During SIMV there was no difference between OS and CS in DeltaV(L) or t(rec). During HFOV there was no difference in DeltaV(L) but a slightly longer t(rec) after OS.