Objectives: To conduct an in vitro evaluation of a humidified high-flow nasal cannula (HFNC) system at different flows, cannula sizes, and air leaks and also an in vivo analysis of mean end-expiratory esophageal pressure (EEEP) from nasal continuous positive airway pressure at 6 cm H2O (NCPAP+6) versus HFNC. Study design: In the in vitro study, we measured HFNC system pressure and flow, with varying degrees of leak and with and without the use of a pressure-limiting valve. In the in vivo study, we measured EEEP in 15 newborns on NCPAP+6 and then on HFNC at 6 L/minute, with flow decreased by 1 L/minute every 30 minutes. Heart rate, respiratory rate, fraction of inspired oxygen, arterial oxygen saturation, respiratory distress syndrome score, and EEEP were recorded for each intervention. Data analysis was done using repeated-measures analysis of variance and linear regression. Results: In the in vitro study, in the absence of leaks, the pressures were limited by the pressure-limiting valve only at flows ≥ 2 L/minute. With leaks of 30% and 50%, delivered pressures were always < 3 cm H2O. In the in vivo study, respiratory rate increased from baseline (NCPAP+6) as flow decreased (P < .02). Intrapatient and interpatient coefficients of variation were always high. Conclusions: A pressure-limiting valve is necessary in a HFNC system. Although mean EEEP levels were similar in NCPAP+6 and HFNC, tachypnea developed as flow diminished. This system apparently cannot predict EEEP, because of interpatient and intrapatient variation.