Airway pressures generated by high flow nasal cannula in patients with acute hypoxemic respiratory failure: a computational study.

Abstract

Introduction and objectives: High flow nasal cannula (HFNC) therapy is an increasingly popular mode of non-invasive respiratory support for the treatment of patients with acute hypoxemic respiratory failure (AHRF). Previous experimental studies in healthy subjects have established that HFNC generates flow-dependent positive airway pressures, but no data is available on the levels of mean airway pressure (mP_aw) or positive end-expiratory pressure (PEEP) generated by HFNC therapy in AHRF patients. We aimed to estimate the airway pressures generated by HFNC at different flow rates in patients with AHRF, whose functional lung volume may be significantly reduced compared to healthy subjects due to alveolar consolidation and/or collapse.

Materials and methods: We developed a high-fidelity mechanistic computational model of the cardiopulmonary system during HFNC therapy using data from healthy subjects, and then measured the mP_aw and PEEP levels produced when different amounts of alveolar consolidation/collapse were incorporated into the model.

Results: When calibrated to represent normal lung physiology in healthy subjects, our model recapitulates the airway pressures produced by HFNC at different flow rates in healthy volunteers who were breathing normally, with their mouths closed or open. When different amounts of alveolar consolidation/collapse are implemented in the model to reflect the pathophysiology of AHRF, the mP_aw and PEEP produced by HFNC at all flow rates increase as the functional lung volume decreases (up to a mP_aw of 11.53 and a PEEP of 11.41 cmH₂O at 60 L/min with the mouth closed when 50% of the model's alveolar compartments are non-aerated). When the model was matched to individual patient data from a cohort of 58 patients with AHRF receiving HFNC at 60 L/min, the mean (standard deviation) of the mP_aw / PEEP produced by HFNC in the models of these patients was 8.56 (1.50) / 8.92 (1.49) cmH₂O with mouths closed, and 1.73 (0.31) / 1.36 (0.36) cmH₂O with mouths open.

Conclusions: Our results suggest that the airway pressures produced by HFNC in patients with AHRF could be higher than is currently assumed based on experimental data from healthy subjects, particularly in patients whose mouths remain closed. Higher levels of PEEP could be beneficial if they lead to alveolar recruitment and improved lung compliance, but could cause alveolar overdistension if they do not, motivating the close monitoring of the effects of HFNC on lung mechanics. Further clinical studies are warranted to directly measure the airway pressures produced by HFNC in patients with different severities of AHRF.