Trial of noninvasive ventilation (NIV) in the emergency department (ED) for heterogeneous acute respiratory failure (ARF) has been optional and its clinical benefit unclear.

We conducted a retrospective cohort study comparing between two periods, October 2001–September 2003 and October 2004–September 2006, i.e., before and after adopting an NIV-trial strategy in which NIV was applied in the ED to any noncontraindicated ARF patients needing ventilatory support and was then continued in the intermediate-care-unit. During these two periods, we retrieved cases of ARF treated either invasively or with NIV, and compared the patients' in-hospital mortalities and the length of ICU and intermediate-care-unit stay.

Compared were 73 (invasive 56, NIV 17) and 125 cases (invasive 31, NIV 94) retrieved from 271 and 415 emergent admissions with proper pulmonary etiologies for mechanical ventilation, respectively. Of their respiratory failures, type (hypercapnic/non-hypercapnic, 0.97 vs. 0.98) and severity (pH 7.23 vs. 7.21 for hypercapnic; PaO₂/FiO₂ 133 vs. 137 for non-hypercapnic) were similar, and the rate of predisposing etiologies was not significantly different. However, excluding those with recurrent aspiration pneumonia for whom NIV was mostly used as “ceiling” treatment, significant reductions in both overall in-hospital mortality (38%–19%, risk ratio 0.51, 95% CI 0.31–0.84), and median length of ICU and intermediate-care-unit stay (12 vs. 5 days, P<0.0001) were found.

NIV-trial in the ED for all possible patients with ARF of pulmonary etiologies, excluding those with recurrent aspiration pneumonia, may reduce overall in-hospital mortality and ICU stays.