High flow nasal cannula oxygenation in COVID-19 related acute respiratory distress syndrome: a safe way to avoid endotracheal intubation?

BACKGROUNDS: High flow nasal cannula (HFNC) is an alternative therapy for acute respiratory distress syndrome (ARDS) due to coronavirus disease 2019 (COVID-19). This study aimed first to describe outcomes of patients suffering from COVID-19-related ARDS treated with HFNC; secondly to evaluate safety of HFNC (patients and healthcare workers) and compare patients according to respiratory outcome. METHODS: A retrospective cohort was conducted in French general hospital intensive care unit (ICU). Patients were included if receiving HFNC for hypoxemia (saturation pulse oxygen (SpO(2)) <92% under oxygen ⩾6 L/min) associated with ARDS and positive SARS-CoV-2 polymerase chain reaction (PCR). Main clinical characteristics and outcomes are described in patients: (a) with do not intubate order (HFNC-DNIO); (b) who did not need intubation (HFNC-only); and (c) eventually intubated (HFNC-intubation). Medians are presented with (1st–3rd) interquartile range. RESULTS: From 26 February to 30 June 2020, 46 patients of median age 75 (70–79) years were included. In the HFNC-DNIO group (n = 11), partial arterial oxygen pressure (PaO(2))/inhaled fraction of oxygen (FiO(2)) ratio median worst PaO(2)/FiO(2) ratio was 109 (102–172) and hospital mortality was 54.5%. Except the HFNC-DNIO patients (n = 35), 20 patients (57%) were eventually intubated (HFNC-intubation group) and 15 were only treated by HFNC (HFNC-only). HFNC-intubation patients presented higher worst respiratory rates per minute in ICU [37 (34–41) versus 33 (24–34) min, p < 0.05] and worsened ICU admission PaO(2)/FiO(2) ratios [121 (103–169) versus 191 (162–219), p < 0.001] compared with HFNC-only patients. Hospital mortality was 35% (n = 7/20) in HFNC-intubation group, 0% in HFNC-only group with a global mortality of these two groups of 20% (n = 7/35). Among tests performed in healthcare workers, 1/12 PCR in symptomatic healthcare workers and 1.8% serologies in asymptomatic healthcare workers were positive. After review of each case, COVID-19 was likely to be acquired outside hospital. CONCLUSIONS: HFNC seems to be useful for COVID-19-related ARDS and safe for healthcare workers. ARDS severity with PaO(2)/FiO(2) <150 associated with respiratory rate >35/min could be regarded as a predictor of intubation. The reviews of this paper are available via the supplemental material section.