Can dynamic indicators help the prediction of fluid responsiveness in spontaneously breathing critically ill patients?

Abstract

To investigate whether the respiratory changes in arterial pulse (DeltaPP) and in systolic pressure (DeltaSP) could predict fluid responsiveness in spontaneously breathing (SB) patients. Because changes in intrathoracic pressure during spontaneous breathing (SB) might be insufficient to modify loading conditions of the ventricles, performances of indicators were also assessed during a forced respiratory maneuver. Prospective interventional study. A 34-bed university hospital medico-surgical ICU. Thirty-two SB patients with clinical signs of hemodynamic instability. A 500-ml volume expansion (VE). Cardiac index, assessed using transthoracic echocardiography, increased by at least 15% after VE in 19 patients (responders). At baseline, only dynamic indicators were higher in responders than in nonresponders (13+/-5% vs. 7+/-3%, p=0.003 for DeltaPP and 10+/-4% vs. 6+/-2%, p=0.002 for DeltaSP). Moreover, they significantly decreased after VE (11+/-5% to 6+/-4%, p<0.001 for DeltaPP and 8+/-4% to 6+/-3%, p<0.001 for DeltaSP). DeltaPP and DeltaSP areas under the ROC curve were high (0.81+/-0.08 and 0.82+/-0.08; p=0.888, respectively). A DeltaPP>or=12% predicted fluid responsiveness with high specificity (92%) but poor sensitivity (63%). The forced respiratory maneuver reproducing a dyspneic state decreased the predictive power. Due to their lack of sensitivity and their dependence to respiratory status, DeltaPP and DeltaSP are clearly less reliable to predict fluid responsiveness during SB than in mechanically ventilated patients. However, when their baseline value is high without acute right ventricular dysfunction in a participating patient, a positive response to fluid is likely.