ABSTRACT
BACKGROUND: To assess the effectiveness of non-rebreather mask combined with low-flow nasal cannula (NRB + NC) compared to high-flow nasal cannula (HFNC) in improving oxygenation in patients with COVID-19-related hypoxemic respiratory failure (HRF). METHODS: This retrospective study was conducted in emergency departments of two tertiary hospitals from June 1 to August 31, 2021. Consecutive patients aged >18 years admitted for COVID-19-related HRF (World Health Organization criteria: confirmed COVID-19 pneumonia with respiratory rate > 30 breaths/min, severe respiratory distress, or peripheral oxygen saturation < 90% on room air) requiring NRB + NC or HFNC were screened for enrollment. Primary outcome was improvement of partial pressure arterial oxygen (PaO2) at two hours. Secondary outcomes were intubation rate, ventilator-free days, hospital length of stay, and 28-day mortality. Data were analyzed using linear regression with inverse probability of treatment weighting (IPTW) based on propensity score. RESULTS: Among the 110 patients recruited, 52 (47.3%) were treated with NRB + NC, and 58 (52.7%) with HFNC. There were significant improvements in patients' PaO2, PaO2/FIO2 ratio, and respiratory rate two hours after the initiation of NRB + NC and HFNC. Comparing the two groups, after IPTW adjustment, there were no statistically significant differences in PaO2 improvement (adjusted mean ratio [MR] 2.81; 95% CI -5.82 to 11.43; p = .524), intubation rate (adjusted OR 1.76; 95% CI 0.44 to 6.92; p = .423), ventilator-free days (adjusted MR 0.00; 95% CI -8.84 to 8.85; p = .999), hospital length of stay (adjusted MR 3.04; 95% CI -2.62 to 8.69; p = .293), and 28-day mortality (adjusted OR 0.68; 95% CI 0.15 to 2.98; p = .608). CONCLUSION: HFNC may be beneficial in COVID-19 HRF. NRB + NC is a viable alternative, especially in resource-limited settings, given similar improvement in oxygenation at two hours, and no significant differences in long-term outcomes. The effectiveness of NRB + NC needs to be investigated by a powered randomized controlled trial.