| Sumario: | Background: The new global definition of nonintubated subjects with ARDS can be applied to patients receiving high-flow nasal cannula (HFNC), but stratification with PaO2/FIO2 has not been validated. We aimed to investigate the differences in oxygenation assessment between HFNC and CPAP for predicting outcomes. Methods: We performed a post hoc analysis of a multi-center randomized controlled trial conducted in Japan, focusing on oxygenation assessments in subjects receiving HFNC or CPAP. Subjects with PaO2/FIO2 < 300 on CPAP at 5 cm H2O were assigned to receive either HFNC or CPAP. Subjects were stratified into mild (PaO2/FIO2 ≥ 200 but <300), moderate (PaO2/FIO2 ≥ 100 but < 200), or severe (PaO2/FIO2 < 100) hypoxemia categories based on oxygenation levels during CPAP at 5 cm H2O. The primary outcome was treatment failure (intubation or in-hospital death). Discriminative performance for treatment failure was evaluated using the area under the receiver operating characteristic curve (AUROC) for each device. Results: Of the 85 subjects analyzed, 31, 48, and 6 were classified as having mild, moderate, and severe hypoxemia, respectively. Treatment failure occurred more frequently in subjects with moderate and severe hypoxemia than in those with mild hypoxemia (mild 1/31 [3.2%], moderate: 12/48 [25.0%], severe: 3/6 [50.0%], P =.004). Compared with 1 of 38 subjects (2.6%) treated with CPAP, 21 of 47 subjects (44.7%) with HFNC experienced worsened oxygenation even after 30 min of treatment (P <.001). Subjects with HFNC had a lower AUROC of PaO2/FIO2 for treatment failure (0.66, 95% CI 0.54–0.79) than those with CPAP (0.82, 95% CI 0.74–0.90, P =.041). Conclusions: PaO2/FIO2-based stratification may help predict treatment failure in patients with nonintubated subjects with ARDS. However, a thorough oxygenation assessment is necessary because of the variability introduced by different respiratory support devices. CPAP might be superior to HFNC when predicting treatment failure.
|
|---|