Effects of positive end-expiratory pressure on the predictability of fluid responsiveness in acute respiratory distress syndrome patients

The prediction accuracy of pulse pressure variation (PPV) for fluid responsiveness was suggested to be unreliable in low tidal volume (VT) ventilation. However, high PEEP can cause ARDS patients relatively hypovolemic and more fluid responsive. We hypothesized that high PEEP 15 cmH(2)O can offset the disadvantage of low VT and improve the predictive performance of PPV. We prospectively enrolled 27 hypovolemic ARDS patients ventilated with low VT 6 ml/kg and three levels of PEEP (5, 10, 15 cmH(2)O) randomly. Each stage lasted for at least 5 min to allow for equilibration of hemodynamics and pulmonary mechanics. Then, fluid expansion was given with 500 ml hydroxyethyl starch (Voluven 130/70). The hemodynamics and PPV were automatically measured with a PiCCO2 monitor. The PPV values were significantly higher during PEEP15 than those during PEEP5 and PEEP10. PPV during PEEP15 precisely predicts fluid responsiveness with a cutoff value 8.8% and AUC (area under the ROC curve) of ROC (receiver operating characteristic curve) 0.847, higher than the AUC during PEEP5 (0.81) and PEEP10 (0.668). Normalizing PPV with driving pressure (PPV/Driving-P) increased the AUC of PPV to 0.875 during PEEP15. In conclusions, high PEEP 15 cmH(2)O can counteract the drawback of low VT and preserve the predicting accuracy of PPV in ARDS patients.