Noninvasive Oxygenation Strategies in Immunocompromised Patients With Acute Hypoxemic Respiratory Failure: A Pairwise and Network Meta-Analysis of Randomized Controlled Trials.

INTRODUCTION: Acute hypoxemic respiratory failure (AHRF) is a leading cause of intensive care unit (ICU) admission among immunocompromised patients. Invasive mechanical ventilation is associated with increased morbidity and mortality.

OBJECTIVE: To evaluate the efficacy of various oxygenation strategies including noninvasive ventilation (NIV), high-flow nasal cannula (HFNC), and conventional oxygen therapy in immunocompromised patients with AHRF.

METHODS: Electronic databases including PubMed, Embase, and the Cochrane Library were reviewed from inception to December 2018. We included all randomized controlled trials (RCTs) comparing different modalities of initial oxygenation strategies in immunocompromised patients with AHRF. Our primary outcome was the need for intubation and invasive mechanical ventilation while secondary outcomes were ICU acquired infections and short- and long-term mortality. Data were extracted separately and independently by 2 reviewers. We performed a Bayesian network meta-analysis to calculate odds ratio (OR) and Bayesian 95% credible intervals (CrIs).

RESULTS: Nine RCTs were included (1570 patients, mean age 61.1 ± 13.8 years with 64% male). Noninvasive ventilation was associated with a significantly reduced intubation rate compared with standard oxygen therapy (OR: 0.53; 95% CrI: 0.26-0.91). There were no significant reductions of intubation between NIV versus HFNC (OR: 0.83; 95% CrI: 0.35-2.11) or HFNC versus standard oxygen therapy (OR: 0.65; 95% CrI: 0.26-1.24). There were no significant differences between all groups regarding short-term (28-day or ICU) mortality or long-term (90-day or hospital) mortality or ICU-acquired infections ( P > 0.05).

CONCLUSION: Among immunocompromised patients with AHRF, NIV was associated with a significant reduction of intubation compared with standard oxygen therapy. There were no significant differences among all oxygenation strategies regarding mortality and ICU-acquired infections.