Criteria for stopping high-flow nasal oxygen for acute hypoxemic respiratory failure: a systematic review of randomized controlled trials

High-flow nasal cannula (HFNC) has been widely promoted during the COVID-19 pandemic to circumvent invasive mechanical ventilation. While there are several reported benefits, randomized trials demonstrate inconsistent variable success. We hypothesize that this is due to variable stopping criteria. This systematic review's purpose is to review these criteria and investigate any associations with HFNC outcomes. We searched PubMed and EMBASE for all English-language randomized controlled trials (RCTs) published from January 1, 2007, to December 31, 2022, focusing on respiratory rate as a threshold for escalation of respiratory support. Subgroup analysis was conducted based on trial failure criteria, and intubation and mortality benefits were studied. Fisher's exact test was performed following a 5% level of significance. Of the 22 RCTs included, 4 (18.2%) reported significant intubation benefits and 1 (0.05%) reported significant mortality benefit. The presence of objective failure criteria with a prespecified high respiratory rate threshold (35 breaths per minute or higher) had a significant effect on intubation rate reduction (P = 0.02). However, this result might be limited by the heterogeneity of the included studies. Further RCTs are required to confirm this conclusion. Given that a high respiratory rate threshold was associated with a reduction of intubation without increasing mortality, we hypothesize that among patients receiving HFNC who were eventually not intubated, the avoidance of intubation led to better clinical outcomes, while among eventually intubated patients, delays led to poorer outcomes.