Variation in High-Flow Nasal Cannula and Non-Invasive Ventilation Practice Patterns During the COVID-19 Pandemic: Results from the International Viral Infection and Respiratory Illness Universal Study (VIRUS)

Abstract

RATIONALE: Critical care guidelines have supported use of non-invasive respiratory support modalities in patients with acute respiratory failure from COVID-19 since the beginning of the pandemic. However, concerns surrounding viral particle aerosolization, nosocomial spread, and patient self-induced lung injury have likely influenced choice of respiratory support strategies. To date, high flow nasal cannula (HFNC) and non-invasive positive pressure ventilation (NIPPV) practice patterns have not been characterized for patients with COVID-19. METHODS: We enrolled hospitalized patients aged 18 years or older with laboratory confirmed COVID-19 infection who received supplemental oxygen, using the Society of Critical Care Medicine Discovery VIRUS Registry. The primary outcome was hospital-level variation in use of HFNC and NIPPV, summarized using the intraclass correlation coefficient and median odds ratio. Hierarchical random effects models were used to estimate patient and hospital factors associated with HFNC and NIPPV use. Risk-adjusted estimation of the association between hospital HFNC/NIPPV use and patient risk of receiving invasive mechanical ventilation (IMV) was assessed as a secondary outcome. RESULTS: Among 8,532 patients with COVID-19 receiving oxygen support across 73 hospitals, the majority were treated in the US (92.3%) and were older (median age 63 years, IQR 52-74), white (49.1%), men (56.8%) with median SOFA score of 4 (IQR 1-6) and admission PaO2:FiO2 below 300 (49.4%). Of these, 5,298 (62.1%) received low flow oxygen (nasal cannula or face mask), while 1,768 (20.7%) received HFNC, 773 (9.1%) received NIPPV and 693 (8.1%) received both HFNC/NIPPV. Patient SOFA score (OR 0.92, 95% CI 0.90, 0.95), treatment for COVID-19 after July versus March-June (OR 1.3, 95% CI 1.0, 1.6) and ICU versus floor admission (OR 10.3, 95% CI 8.2, 12.8) were associated with HFNC/NIPPV use. After adjusting for patient and hospital characteristics, the hospital of admission contributed to 27% of the variation in use of HFNC and/or NIPPV. Odds of receiving either modality at a randomly selected high vs. low HFNC/NIPPV utilization hospital was 2.9. Hospital rates of HFNC/NIPPV use were not associated with patient receipt of IMV (OR 0.87, 95% CI 0.7, 1.1). CONCLUSION: Throughout the course of the COVID-19 pandemic, use of HFNC and NIPPV varied widely across hospitals, though use of non-invasive respiratory support modalities was not associated with patient risk for invasive mechanical ventilation. Further evaluation of HFNC and NIPPV exposure, progression to IMV and subsequent mortality within these subgroups may provide additional insights regarding optimal oxygenation and ventilation strategies of patients with COVID-19.