Background: We tested the hypothesis that the pressure-time (P-t) curve during constant flow ventilation can be used to set a noninjurious ventilatory strategy.
Methods: In an isolated, nonperfused, lavaged model of acute lung injury, tidal volume and positive end-expiratory pressure were set to obtain: (1) a straight P-t curve (constant compliance, minimal stress); (2) a downward concavity in the P-t curve (increasing compliance, low volume stress); and (3) an upward concavity in the P-t curve (decreasing compliance, high volume stress). The P-t curve was fitted to: P = a. tb +c, where b describes the shape of the curve, b = 1 describes a straight P-t curve, b < 1 describes a downward concavity, and b > 1 describes an upward concavity. After 3 h, lungs were analyzed for histologic evidence of pulmonary damage and lavage concentration of inflammatory mediators. Ventilator-induced lung injury occurred when injury score and cytokine concentrations in the ventilated lungs were higher than those in 10 isolated lavaged rats kept statically inflated for 3 h with an airway pressure of 4 cm H2O.
Results: The threshold value for coefficient b that discriminated best between lungs with and without histologic and inflammatory evidence of ventilator-induced lung injury (receiver-operating characteristic curve) ranged between 0.90-1.10. For such threshold values, the sensitivity of coefficient b to identify noninjurious ventilatory strategy was 1.00. A significant relation (P < 0.001) between values of coefficient b and injury score, interleukin-6, and macrophage inflammatory protein-2 was found.
Conclusions: The predictive power of coefficient b to predict noninjurious ventilatory strategy in a model of acute lung injury is high.