Prone position ventilation for acute hypoxemic respiratory failure (AHRF) improves oxygenation but not survival, except possibly when AHRF is severe. To determine effects of prone versus supine ventilation in AHRF and severe hypoxemia [partial pressure of arterial oxygen (PaO(2))/inspired fraction of oxygen (FiO(2)) <100 mmHg] compared with moderate hypoxemia (100 mmHg < or = PaO(2)/FiO(2) < or = 300 mmHg). Systematic review and meta-analysis. Electronic databases (to November 2009) and conference proceedings. Two authors independently selected and extracted data from parallel-group randomized controlled trials comparing prone with supine ventilation in mechanically ventilated adults or children with AHRF. Trialists provided subgroup data. The primary outcome was hospital mortality in patients with AHRF and PaO(2)/FiO(2) <100 mmHg. Meta-analyses used study-level random-effects models. Ten trials (N = 1,867 patients) met inclusion criteria; most patients had acute lung injury. Methodological quality was relatively high. Prone ventilation reduced mortality in patients with PaO(2)/FiO(2) <100 mmHg [risk ratio (RR) 0.84, 95% confidence interval (CI) 0.74-0.96; p = 0.01; seven trials, N = 555] but not in patients with PaO(2)/FiO(2) > or =100 mmHg (RR 1.07, 95% CI 0.93-1.22; p = 0.36; seven trials, N = 1,169). Risk ratios differed significantly between subgroups (interaction p = 0.012). Post hoc analysis demonstrated statistically significant improved mortality in the more hypoxemic subgroup and significant differences between subgroups using a range of PaO(2)/FiO(2) thresholds up to approximately 140 mmHg. Prone ventilation improved oxygenation by 27-39% over the first 3 days of therapy but increased the risks of pressure ulcers (RR 1.29, 95% CI 1.16-1.44), endotracheal tube obstruction (RR 1.58, 95% CI 1.24-2.01), and chest tube dislodgement (RR 3.14, 95% CI 1.02-9.69). There was no statistical between-trial heterogeneity for most clinical outcomes. Prone ventilation reduces mortality in patients with severe hypoxemia. Given associated risks, this approach should not be routine in all patients with AHRF, but may be considered for severely hypoxemic patients.
