We tested the hypothesis that the dynamic evaluation of central venous pressure (CVP) amplitude could be a reliable predictor of fluid responsiveness in patients under mechanical ventilation, similar to the variation of arterial pulse pressure (ΔPp). Thirty postcardiac surgery patients, under mechanical ventilation, were evaluated. The percentual difference between inspiratory (Ppins) and expiratory pulse pressure (Ppins) was so calculated: ΔPp (%) = 100 × (Ppins − Ppexp)/[(Ppins + Ppexp)/2]. The respiratory variation of CVP curves amplitude were calculated by determining the percentual difference between inspiratory (CVPpins) and expiratory (CVPpexp) variation using vena cava "pressure" collapsibility index according the following formula: Cvci (%) = [(CVPpexp − CVPpins)/CVPpexp] × 100. There was a correlation between ΔPp and Cvci (Pearson correlation coefficient, r2 = 0.45). Receiver operating characteristic curves showed that the Cvci value more than or equal to 5% predicted ΔPp more than or equal to 13% with 91% specificity, 89% sensitivity, and AUC of 0.90. Therefore, Cvci presented a good agreement with ΔPp (kappa = 0.76) to identify potential fluid responders (patients with ΔPp ≥13%). In 9 potential fluid responders, both ΔPp and Cvci significantly decreased from 18% ± 8% to 8% ± 6% (P < 0.004) and 23% ± 15% to 7% ± 6% (P < 0.004), respectively, after fluid replacement. Our findings suggest that vena cava "pressure" collapsibility index can be used as a marker of fluid responsiveness in postcardiac surgery patients under mechanical ventilation, such as arterial pulse pressure respiratory variation.