Association of prolonged pre-intubation high-flow nasal cannula duration with survival and successful V-V ECMO liberation in severe ARDS: a retrospective multicenter study.

BACKGROUND

In patients with severe acute respiratory distress syndrome (ARDS), prolonged invasive mechanical ventilation (IMV) prior to veno-venous extracorporeal membrane oxygenation (V-V ECMO) is a known risk factor for poor outcomes. However, the clinical impact of high-flow nasal cannula (HFNC) duration before intubation remains poorly defined. We examined the association between pre-intubation HFNC duration and clinical outcomes, including survival and ECMO liberation.

METHODS

This retrospective multicenter study using the J-CARVE registry analyzed 209 adults with severe ARDS who received HFNC before V-V ECMO (2012-2022). The continuous risk of prolonged HFNC for 60-day in-hospital mortality was evaluated using multivariable logistic regression. Subsequently, patients were stratified into Short (< 3 days) and Long (≥ 3 days) HFNC groups using an exploratory ROC-derived cutoff. Clinical outcomes (60-day in-hospital mortality and successful ECMO liberation) were further analyzed using multivariable Cox and Fine-Gray models.

RESULTS

Multivariable logistic regression identified prolonged HFNC duration (analyzed continuously) as an independent risk factor for 60-day in-hospital mortality (adjusted odds ratio 1.22 per 1-day increase; 95% CI 1.07-1.39; P = 0.0035). Sixty-nine patients (33.0%) were in the Long HFNC group. After adjustment, prolonged HFNC (≥ 3 days) was independently associated with a twofold increase in 60-day in-hospital mortality (adjusted Hazard Ratio 2.27; 95% CI 1.24-4.15; P = 0.008). Furthermore, successful ECMO liberation was significantly lower in the Long HFNC group (adjusted Subdistribution Hazard Ratio 0.68; 95% CI 0.50-0.94; P = 0.018).

CONCLUSIONS

Extended pre-intubation HFNC duration was associated with increased mortality and impaired lung recovery in severe ARDS. Rather than a direct cause of injury, prolonged HFNC likely serves as a complex prognostic marker reflecting disease trajectory and delayed escalation. Clinicians should recognize prolonged HFNC as a continuous trajectory of accumulating risk, rather than relying solely on IMV duration to guide ECMO initiation.