High-flow nasal oxygen alone or alternating with non-invasive ventilation in critically ill immunocompromised patients with acute respiratory failure: a randomised controlled trial.

BACKGROUND: Although non-invasive ventilation (NIV) is recommended for immunocompromised patients with acute respiratory failure in the intensive care unit (ICU), it might have deleterious effects in the most severe patients. High-flow nasal oxygen (HFNO) alone might be an alternative method to reduce mortality. We aimed to determine whether HFNO alone could reduce the rate of mortality at day 28 compared with HFNO alternated with NIV.
METHODS: FLORALI-IM is a multicentre, open-label, randomised clinical trial conducted in 29 ICUs (28 in France and one in Italy). Adult immunocompromised patients with acute respiratory failure, defined as respiratory rate of 25 breaths per min or more and a partial pressure of arterial oxygen to inspired fraction of oxygen ratio of 300 mm Hg or lower, were randomly assigned (1:1) to HFNO alone (HFNO alone group) or NIV alternating with HFNO (NIV group). Key exclusion criteria were severe hypercapnia above 50 mm Hg, patients who could strongly benefit from NIV (ie, those with underlying chronic lung disease, with cardiogenic pulmonary oedema, or who were postoperative), severe shock, impaired consciousness defined as Glasgow coma score ≤12, urgent need for intubation, do not intubate order, and contraindication to NIV. Patients were assigned using computer-generated permuted blocks and were stratified according to centre and to the type of immunosuppression using a centralised web-based management system. In the HFNO alone group, patients were continuously treated by HFNO with a gas flow rate of 60 L/min or the highest tolerated. In the NIV group, patients were treated with NIV with a first session of at least 4 h, and then by sessions for a minimal duration of 12 h a day, with a dedicated ventilator, targeting a tidal volume below 8 mL/kg of predicted bodyweight, and with a positive end-expiratory level of at least 8 cm H2O. NIV sessions were interspaced with HFNO delivered as in the HFNO alone group. The primary outcome was mortality at day 28 and was assessed in the intention-to-treat population. Secondary outcomes were mortality in the ICU, in hospital, at day 90 and at day 180, intubation at day 28, length of stay in the ICU and in hospital, number of ventilator-free days at day 28, number of oxygenation technique-free days at day 28, and efficacy and tolerance of oxygenation techniques. The trial is registered with ClinicalTrials.gov, NCT02978300, and is complete.
FINDINGS: Between Jan 21, 2017 to March 4, 2019, of 497 eligible patients, 300 were randomly assigned but one patient withdrew consent, leaving 299 patients included in the intention-to-treat analysis (154 assigned to the HFNO alone group and 145 assigned to NIV group). Mortality rate at day 28 was 36% (95% CI 29·2 to 44·2; 56 of 154 patients) in the HFNO alone group and 35% (27·9 to 43·2; 51 of 145 patients) in the NIV group (absolute difference 1·2% [95% CI -9·6 to 11·9]; p=0·83). None of the other prespecified secondary outcomes were different between groups except for greater decreased discomfort after initiation of HFNO than with NIV (-4 mm on visual analogic scale [IQR -18 to 4] vs 0 mm [-16 to 17]; p=0·040).
INTERPRETATION: In critically ill immunocompromised patients with acute respiratory failure, the mortality rate did not differ between HFNO alone and NIV alternating with HFNO. However, study power was limited, so results should be interpreted with caution. FUNDING: French Ministry of Health.