Objective: Positive end-expiratory pressure (PEEP) has been viewed as an essential component of mechanical ventilation in acute respiratory distress syndrome (ARDS) and acute lung injury (ALI). However, clinical trials have not yet convincingly demonstrated that high PEEP levels improve survival. The object of this study was to test a priori hypotheses that a small but clinically important mortality benefit of high PEEP did exist, especially in patients with greater overall severity of illness and differences in PEEP protocols might have affected the study results.
Methods: Meta-analysis of randomized controlled trials comparing high versus low PEEP in ARDS/ALI. Studies were identified by search of MEDLINE (1950-2008) and other sources.
Measurements and main results: Five studies including 2447 patients were identified. A pooled analysis showed a significant reduction in hospital mortality in favor of high PEEP (RR=0.89; 95% CI, 0.80-0.99; p=0.03). However, significant statistical and clinical heterogeneities such as differences in disease severity and ventilator protocols were found. The differences in PEEP protocols were not associated with differences in mortality rates. A logistic analysis suggested that the beneficial effect of high PEEP was greater in patients with higher ICU severity scores.
Conclusions: The statistical and clinical heterogeneities make proper interpretation of the results difficult. However, a small, but significant mortality benefit of high PEEP may exist. In addition, our analysis suggests the effects of high PEEP are greater in patients with higher ICU severity scores.