ABSTRACT
RATIONALE: A high percentage of patients with Covid-19 associated acute hypoxemic respiratory failure (AHRF) receive invasive mechanical ventilation, which is associated with a high mortality. Noninvasive ventilation (NIV) offers an alternative to invasive mechanical ventilation, though its role in both de novo AHRF and pandemic viral pneumonia has been controversial. Our group has previously demonstrated that NIV delivered by helmet reduces endotracheal intubation rates and improves mortality in patients with acute respiratory distress syndrome when compared to facemask NIV in a randomized controlled clinical trial (1). Limited data are available on the comparative efficacy of noninvasive respiratory support strategies in patients with AHRF due to Covid-19. NIV by helmet interface offers a promising strategy for these patients and may avert the need for endotracheal intubation. METHODS: All patients with Covid-19 associated AHRF admitted to the intensive care unit and managed initially with noninvasive respiratory support between March 1 and July 31, 2020 were identified. Those who received helmet NIV were matched with patients who received HFNC (high flow nasal cannula) using propensity scores in a 1:1 ratio without replacement. Baseline characteristics and therapies that differed on univariable analysis were used for matching. After matching, univariable analysis was conducted comparing the HFNC and helmet NIV groups. RESULTS: 78 patients initially managed with HFNC and 31 patients managed with helmet NIV (excluding 7 patients who received helmet NIV after intubation) were identified. Matching resulted in similar groups according to baseline characteristics and therapies received. The primary composite outcome of intubation or inhospital mortality occurred in 20 (64.5%) patients in the helmet group and 20 (64.5%) patients in the control group (absolute difference, 0%;95% CI,-23.4% to 23.4%;p=1.00). In-hospital mortality in the helmet NIV group was similar to that of the HFNC control group (absolute difference-9.7%;95% CI,-34.4% to 15.0%;p=0.45). In patients treated with helmet NIV, gas exchange improved significantly following application with an increase in the ratio of arterial partial pressure of oxygen to fraction of inspired oxygen from 72 to 141 (mean difference, 69;95% CI 26 to 112;p=0.003). CONCLUSIONS: While limited by the small sample size and observational nature, there was no significant difference observed in outcomes of patients with Covid-19 associated AHRF managed with helmet NIV compared to HFNC. After application of helmet NIV, a significant improvement in gas exchange was observed. REFERENCES: 1. Patel BK, et al. JAMA, 2016.