ObjectiveHigh-flow nasal cannula (HFNC) therapy is increasingly used in preterm infants despite a paucity of physiological studies. We aimed to investigate the effects of HFNC on respiratory physiology.Study designA prospective randomised crossover study was performed enrolling clinically stable preterm infants receiving either HFNC or nasal continuous positive airway pressure (nCPAP). Infants in three current weight groups were studied: <1000 g, 1000–1500 g and >1500 g. Infants were randomised to either first receive HFNC flows 8–2 L/min and then nCPAP 6 cm H2O or nCPAP first and then HFNC flows 8–2 L/min. Nasopharyngeal end-expiratory airway pressure (pEEP), tidal volume, dead space washout by nasopharyngeal end-expiratory CO2 (pEECO2), oxygen saturation and vital signs were measured.ResultsA total of 44 preterm infants, birth weights 500–1900 g, were studied. Increasing flows from 2 to 8 L/min significantly increased pEEP (mean 2.3–6.1 cm H2O) and reduced pEECO2 (mean 2.3%–0.9%). Tidal volume and transcutaneous CO2 were unchanged. Significant differences were seen between pEEP generated in open and closed mouth states across all HFNC flows (difference 0.6–2.3 cm H2O). Infants weighing <1000 g received higher pEEP at the same HFNC flow than infants weighing >1000 g. Variability of pEEP generated at HFNC flows of 6–8 L/min was greater than nCPAP (2.4–13.5 vs 3.5–9.9 cm H2O).ConclusionsHFNC therapy produces clinically significant pEEP with large variability at higher flow rates. Highest pressures were observed in infants weighing <1000 g. Flow, weight and mouth position are all important determinants of pressures generated. Reductions in pEECO2 support HFNC’s role in dead space washout.