cIVC was expressed as a percentage. In addition, relatively to be sure that the formula:(Dmax – Dmin/((Dmax + Dmin)/2 = cIVC2)could not be more informative, we also built its respective receiver operator characteristic (ROC) curves.The VTI was recorded by pulse waved Doppler on a five-chamber apical view [32]. For each step of the study, the VTI (cm) was measured in triplicate. The obtained values were averaged for its determination.In parallel, the left ventricular filling pressures were assessed using the mitral inflow coupled to tissue Doppler imaging. The transmitral diastolic inflow, or E/A velocity ratio (velocity of the E wave/velocity of the A wave in cm/s) was recorded by pulse Doppler in the apical four-chamber view at the distal extremity of the mitral leaflets [26,33].

In the same view, protodiastolic tissue Doppler velocity was recorded at the lateral annular mitral annulus (Ea wave, cm/s) [33]. The ratio between E and Ea wave velocities (E/Ea ratio) was calculated as an index of left ventriclular filling pressure [33,34].Right ventriclular dilatation was defined as a right to left telediastolic ventriclular area ratio > 1 (RV/LV area ratio) [25]. Left ventriclular systolic function was visually quantified as previously described [35]. Lastly, the shortening diameter fraction was determined in the M mode and parasternal long axis view.ProtocolAfter ruling out the exclusion criteria, a first echocardiography was performed in all spontaneously breathing patients with ACF. At this time (T0), HR, MAP, E, A, and Ea velocities, E/A ratio, E/Ea ratio, and subaortic VTI were recorded.

Then a fluid challenge was performed with 500 mL of a 6% 130/0.4 hydroxyethylstarch solution (Voluven? , Fresenius-Kabi, Louviers, France) infused over 15 minutes. After this fluid challenge (at T15), HR, MAP, E, A, and Ea velocities, E/A ratio, E/Ea ratio, and subaortic VTI were recorded. Fluid responsiveness was defined as an increase in the subaortic VTI �� 15% after the fluid challenge. This served to split the patients into responders (R) and non-responders (NR) [15,16,36]. Of note, the investigators were not blinded.Statistical analysisData are expressed as medians with the 5th and 95th percentiles. For the comparisons between R and NR, Mann-Whitney, Chi square and Fisher exact tests were performed when appropriate. ROC curves were constructed to evaluate the ability of cIVC to predict fluid responsiveness.

When the AUC was greater than 0.5, the best cutoff value was defined by the closest value to the Youden index [37]. ROC curves of E wave velocity, E/A ratio, E/Ea ratio, and CVP were compared to the ROC curve of the cIVC for each individual using the Hanley test Batimastat [38]. Statistical analysis was performed using SAS v 8.1 software (SAS Institute, Cary, NY, USA). All P-values were two-tailed and a P-value < 0.05 was considered significant.