Six RCTs, published between 2001 and 2006, met the inclusion criteria of the review (including 1,423 patients). All the trials achieved a score 3 points on the Jadad scale, but they were mostly underpowered to detect a significant effect by up to 40%. Three trials allowed cross over from one trial arm to the other. Three trials had to stop prematurely due to slow rate of recruiting patients.
Mortality: There was no significant difference in mortality between prone and supine position groups (OR 0.97, 95% CI 0.77 to 1.22; five trials, 1,372 patients), with very low heterogeneity (I2=9.3%). Also, there was no correlation between the odds of mortality and partial pressure of arterial oxygen to the fraction of inspired oxygen (PaO2/FiO2) ratio or duration of prone position. There was no correlation between the relative risk reduction in mortality and the control group mortality rate. The funnel plot for the primary outcome (mortality) did not suggest a publication bias.
Oxygenation: There was a significant improvement in the PaO2/FiO2 in the prone position group compared with supine ventilation (WMD 25.12mmHg, 95% CI 15.10 to 35.15; five trials, 1,178 patients), although there was moderate heterogeneity (I2=56.3%).
Ventilator-associated pneumonia incidence: Although there was a trend of prone positioning reducing ventilator-associated pneumonia, the pooled effect estimate was not statistically significant (four trials, 1,017 patients); heterogeneity between the trials was moderate (I2=48.2%).
Adverse effects: Incidence of major airway complications did not differ between prone and supine position groups (five trials, 1,372 patients). Length of ICU stay did not differ between the two positions (three trials, 1,028 patients). However, pressure sores and facial oedema increased with prone position ventilation (OR 1.35, 95% CI 1.08 to 1.69; 1,378 patients).