Treatment Failure in Non-Intubated Acute Respiratory Distress Syndrome: A Retrospective Cohort Study

Abstract Background Acute respiratory distress syndrome (ARDS) is a severe respiratory condition associated with high mortality, traditionally requiring invasive mechanical ventilation. However, the recent global redefinition of ARDS now encompasses non-intubated patients receiving high-flow nasal oxygen (HFNO) or non-invasive positive pressure ventilation (NIPPV), emphasizing the need for non-invasive management strategies. Despite their widespread use, the factors contributing to treatment failure in these modalities remain poorly understood. The aim of this study was to analyse the characteristics related to the treatment failure of HFNO and NIPPV in patients who met the new global definition of non-intubated ARDS. Methods Patients who satisfied the newly defined criteria for non-intubated ARDS receiving HFNO or NIPPV were included in the MIMIC-IV database. Clinical characteristics and parameters were analysed, including respiratory rate–oxygenation (ROX) value at 1, 6, and 12 h after applying HFNO or NIPPV. Of the 238 patients, 114 and 124 received HFNO and NIPPV, respectively. Results In the HFNO group, the ROX value 1 h after application was significantly lower in patients with treatment failure (AUROC 0.702, cut-off value 4.059). In the NIPPV group, the ROX values at 1, 6, and 12 h were significantly lower in patients with treatment failure. However, only that at 12 h was consistent (AUROC 0.696, cut-off 7.827) after adjusting for confounders. The ROX index demonstrated its utility in predicting treatment failure in non-intubated ARDS, with the timing of assessment varying based on the respiratory support modality. For HFNO, the ROX index at 1 h was the most predictive, whereas, for NIPPV, the ROX index at 12 h showed the strongest association with treatment failure. Conclusions These findings underscore the importance of tailored monitoring strategies based on the type of respiratory support provided.