Although placing patients with acute respiratory failure in a prone position improves their oxygenation, it does not improve survival.

In a model of acute lung injury, we showed that positive end-expiratory pressure (PEEP) and tidal volume (VT) are interactive variables that determine the extent of lung recruitment, that recruitment occurs across the entire range of total lung capacity, and that superimposed pressure is a key determinant of lung collapse. Aiming to verify if the same rules apply in a clinical setting, we randomly ventilated five ALI/ARDS patients with 10, 15, 20, 30, 35, and 45 cm H2O plateau pressure and 5, 10, 15, and 20 cm H2O of PEEP. For each PEEP-VT condition, we obtained computed tomography at end inspiration and end expiration. We found that recruitment occurred along the entire volume-pressure curve, independent of lower and upper inflection points, and that estimated threshold opening pressures were normally distributed (mode = 20 cm H2O). Recruitment occurred progressively from nondependent to dependent lung regions. Overstretching was not associated with hyperinflation. Derecruitment did not parallel deflation, and estimated threshold closing pressures were normally distributed (mode = 5 cm H2O). End-inspiratory and end-expiratory collapse were correlated, suggesting a plateau-PEEP interaction. When superimposed gravitational pressure exceeded PEEP, end-expiratory collapse increased. We concluded that the rules governing recruitment and derecruitment equally apply in an oleic acid model and in human ALI/ARDS.