Effect of non-invasive oxygenation strategies in immunocompromised patients with severe acute respiratory failure: a post-hoc analysis of a randomised trial

Background The use of non-invasive ventilation is controversial in immunocompromised patients with acute respiratory failure, whereas the use of high-flow nasal cannula oxygen therapy is growing as an alternative to standard oxygen. We aimed to compare outcomes of immunocompromised patients with acute respiratory failure treated with standard oxygen with those treated with high-flow nasal cannula oxygen alone or high-flow nasal cannula oxygen associated with non-invasive ventilation. Methods We did a post-hoc subgroup analysis in a subset of immunocompromised patients with non-hypercapnic acute respiratory failure from a multicentre, randomised, controlled trial. In the trial, patients from 23 intensive care units in France and Belgium were randomly assigned (1: 1: 1) to receive either standard oxygen, high-flow nasal cannula alone, or non-invasive ventilation interspaced with high-flow nasal cannula between non-invasive ventilation sessions (non-invasive ventilation group). Patients with profound neutropenia, acute-on-chronic respiratory failure, cardiogenic pulmonary oedema, shock, or altered consciousness were excluded. The primary outcome was the proportion of patients who required endotracheal intubation within 28 days after randomisation. Findings Of the 82 immunocompromised patients, 30 were treated with standard oxygen, 26 with high-flow nasal cannula alone, and 26 with non-invasive ventilation plus interspaced high-flow nasal cannula. 8 (31%) of 26 patients treated with high-flow nasal cannula alone, 13 (43%) of 30 patients treated with standard oxygen, and 17 (65%) of 26 patients treated with non-invasive ventilation required intubation at 28 days (p=0.04). Odds ratios (ORs) for intubation were higher in patients treated with non-invasive ventilation than in those treated with high-flow nasal cannula: OR 4.25 (95% CI 1.33-13.56). ORs were not significantly different between patients treated with high-flow nasal cannula alone and standard oxygen: OR 1.72 (0.57-5.18). After multivariable logistic regression, the two factors independently associated with endotracheal intubation and mortality were age and use of non-invasive ventilation as first-line therapy. Interpretation Non-invasive ventilation might be associated with an increased risk of intubation and mortality and should be used cautiously in immunocompromised patients with acute hypoxaemic respiratory failure.