Background: This meta-analysis compared the effects of non-invasive ventilation (NIV) with invasive mechanical ventilation (InMV) and standard oxygen (O2) therapy on mortality and rate of tracheal intubation in patients presenting acute respiratory failure (ARF).
Methods: We searched the MEDLINE, EMBASE and Cochrane Central Register of clinical trials databases between 1949 and May 2015 to identify randomized trials of NIV for ARF. We excluded the ARF caused by extubation, cardiogenic pulmonary edema, and COPD.
Results: The meta-analysis included 21 studies and 1691 patients, of whom 846 were assigned to NIV and 845 to control (InMV or standard O2 therapy). One hundred ninety-one patients (22.6%) in the NIV group and 261 patients (30.9%) in the control group died before discharge from hospital. The pooled odds ratio (OR) for short-term mortality (in-hospital mortality) was 0.56 (95% CI 0.40-0.78). When comparing NIV with standard O2 therapy, the short-term mortality was 155 (27.4%) versus 204 (36.0%), respectively. For this comparison, the pooled OR of short-term mortality was 0.56 (95% CI 0.36-0.85). When comparing NIV with InMV, the short-term mortality was 36 (12.9%) versus 57 (20.5%) patients, respectively. For this comparison, the pooled OR of short-term mortality was 0.56 (95% CI 0.34-0.90). Tracheal intubation was performed in 106 patients (22.7%) in the NIV and in 183 patients (39.4%) in the standard O2 group, representing a pooled OR of 0.37 (95% CI 0.25-0.55). There were publication biases and the quality of the evidence was graded as low.
Conclusion: Compared with standard O2 therapy or InMV, NIV lowered both the short-term mortality and the rate of tracheal intubation in patients presenting with ARF.
Keywords: Acute respiratory failure; Invasive ventilation; Non-invasive ventilation; Survival analysis; Tracheal intubation.