Ninety-two patients were studied before coronary artery surgery. Each patient was monitored with central venous pressure (CVP), the PiCCO system (Pulsion Medical Systems, Munich, Germany), and transesophageal echocardiography. Responders were defined as those who increased their stroke volume index by greater than 15%(螖SVITPTD >15%) during passive leg raising.
Central venous pressure showed no significant correlation with 螖SVITPTD (r = 鈭?.06, P = .58), in contrast to PPV (r = 0.71, P < .0001), SVV (r = 0.61, P < .0001), GEDVI (r = 鈭?.54, P < .0001), and 螖VTILVOT (r = 0.54, P < .0001). The best area under the receiver operating characteristic curve (AUC) predicting 螖SVITPTD greater than 15%was found for PPV (AUC, 0.82; P < .0001) and SVV (AUC, 0.77; P < .0001), followed by 螖VTILVOT (AUC, 0.74; P < .0001) and GEDVI (AUC, 0.71; P = .0006), whereas CVP was not able to predict fluid responsiveness (AUC, 0.58; P = .18).
In contrast to CVP, GEDVI and 螖VTILVOT reliably predicted fluid responsiveness under closed-chest conditions. Pulse pressure variation and SVV showed the highest accuracy.