What Explains the Reported Benefit of High-flow Nasal Cannula in High-risk Extubated Patients?

We read with great interest the randomized trial by Patil et al. comparing high-flow nasal cannula (HFNC) with noninvasive ventilation (NIV) after extubation in patients at high risk of extubation failure.¹ The authors address a critical clinical question where the optimal strategy remains debated. However, some methodological considerations merit discussion.

First, the trial compared HFNC to NIV delivered with passive humidification. While this may reflect local practice, passive humidification is associated with lower humidification performance, impaired secretion clearance, and reduced patient comfort compared with active humidification during NIV.² Because secretion management and tolerance heavily influence postextubation success, it is unclear whether the observed benefits reflect HFNC's inherent superiority or the limitations of the specific NIV strategy. These findings may not generalize to centers routinely utilizing actively humidified NIV.

Second, the definition of “high-risk extubation failure” was broad, requiring only one of multiple risk factors (e.g., age >65, obesity, chronic obstructive pulmonary disease (COPD), hypercapnia, APACHE II score >12). This creates a highly heterogeneous population; a patient eligible solely due to advanced age differs vastly from one with hypercapnic COPD and prolonged ventilation. Given that current guideline recommendations favor NIV in selected high-risk populations, particularly those with COPD and hypercapnic respiratory failure, subgroup analyses would have provided valuable insight into whether the observed advantage of HFNC was consistent across all risk categories.³

Third, reintubation within 48 hours was the primary endpoint of the study, yet explicit protocolized criteria for reintubation were not described. While the manuscript provides detailed criteria for spontaneous breathing trial failure and extubation readiness, the decision to reintubate appears to have been left to the discretion of the treating team. In an unblinded trial comparing two visibly different respiratory support modalities, variability in clinicians’ thresholds for reintubation may introduce performance and ascertainment bias. This is particularly relevant because reintubation was the primary outcome on which the study's conclusions were based.

Fourth, both interventions were strictly discontinued at 48 hours regardless of clinical status. In standard intensive care unit (ICU) practice, postextubation support is individualized based on a patient's respiratory trajectory. This protocol-mandated cessation limits external validity and may have impacted subsequent respiratory outcomes.

Finally, the reported nearly 15% absolute reduction in mortality requires cautious interpretation. Such a substantial survival benefit is unexpected for a postextubation support strategy in a relatively small, single-center study. Critically ill patients’ mortality depends on numerous variables beyond immediate respiratory management, including baseline physiological reserve, ICU complications, and subsequent care. While measured baseline characteristics were comparable, small-scale randomization may not balance all unmeasured prognostic variables. This dramatic mortality signal likely reflects residual confounding, post-randomization care variations, or random chance rather than a true treatment effect.

We congratulate the authors on this valuable contribution. Future multicenter trials incorporating contemporary NIV practices, standardized reintubation protocols, and phenotype-specific subgroup analyses will help firmly establish the optimal postextubation approach for high-risk patients.

Orcid

Vijay Sundarsingh https://orcid.org/0009-0009-3303-2328

Pramela R Rodrigues https://orcid.org/0009-0009-4128-7654

Manoj Kumar https://orcid.org/0000-0002-4990-3708