Review the physiology, clinical evidence, and implementation challenges of NIV, CPAP, and HFNC in ARDS, with emphasis on patient selection and early recognition of treatment failure.
Approach:
Scope: The review uses the 2024 Global Definition of ARDS for contemporary practice while describing individual studies according to their enrolled populations, which include varied AHRF and ARDS definitions.
Physiology: Examines how positive pressure can improve oxygenation and reduce respiratory effort, while excessive spontaneous effort or pressure support may increase lung stress and injury risk.
Clinical application: Discusses NIV and CPAP selection, interfaces, titration, and the need to monitor comfort, effort, tidal volume, oxygenation, and gas exchange.
Key Findings:
NIV and CPAP can improve oxygenation through PEEP-mediated lung recruitment; pressure support may further reduce inspiratory workload.
Sustained high inspiratory effort and tidal volumes above 9 mL/kg predicted body weight have been associated with NIV failure or unfavorable outcomes in acute hypoxemic respiratory failure.
Evidence for NIV or CPAP in ARDS is inconsistent and influenced by disease severity and patient characteristics.
Some evidence suggests NIV or CPAP may be feasible in carefully selected patients with mild ARDS, preserved mental status, and no hemodynamic instability.
Helmet NIV allows higher PEEP with minimal leakage, but tidal-volume monitoring is difficult.
Interpretation:
Non-invasive support may be appropriate for selected patients, but its use requires individualized settings and close assessment for excessive effort, large tidal volumes, and treatment failure.
Limitations:
Much of the evidence predates the 2024 Global Definition of ARDS and includes heterogeneous AHRF and ARDS populations.
Bedside assessment of inspiratory effort and P-SILI risk, including esophageal pressure monitoring, is limited and not routinely performed.
Choosing PEEP and pressure support to control tidal volume and transpulmonary pressure is difficult; interface tolerance, leaks, and asynchrony can affect treatment.
