Stroke volume variation as a predictor of fluid responsiveness in patients undergoing airway pressure release ventilation.

We investigated the effectiveness of stroke volume variation (SVV) shown by the Vigileo-FloTrac™ system (Edwards Lifesciences, Irvine, CA) to predict fluid responsiveness in patients undergoing airway pressure release ventilation (APRV). All 80 patients mechanically ventilated in the intensive care unit of our hospital from April to November 2010 were included in this study. After starting APRV, Ringer's lactate solution was administered for 30 minutes. Haemodynamic variables including heart rate, mean arterial pressure, cardiac index (CI), stroke volume index (SVI) and SVV were measured before and after volume loading. SVV before volume loading was significantly correlated with absolute change in SVV (ΔSVV) and percentage change in stroke volume index (ΔSVI) after volume loading (ΔSVV: P<0.05, r2=0.534; ΔSVI: P<0.05, r2=0.217). Of the 80 patients, 38 (47.5%) were responders to intravascular volume expansion (increase in CI≥15%) and 42 (52.5%) were non-responders (increase in CI<15%). Receiver operating characteristic (ROC) curves were generated for SVV and central venous pressure by varying the discriminating threshold of the variable and areas under the ROC curves were calculated. The areas under the ROC curves were 0.793 for SVV (95% confidence interval: 0.709-0.877) and 0.442 for central venous pressure (95% confidence interval: 0.336-0.549), which were significantly different (P<0.05). The optimal threshold value of SVV to discriminate between responders and nonresponders was 14% (sensitivity: 78.9%; specificity: 64.3%). We found that SVV was able to predict fluid responsiveness in patients undergoing APRV with acceptable levels of sensitivity and specificity.