Long-axis images of the inferior vena cava (IVC) have limitations as surrogates for IVC morphology in grading central venous pressure (CVP) by two-dimensional echocardiography (2DE), because of the various cross-sectional morphologies and the translational motion of the IVC induced by sniffing. On the basis of the relationship between venous pressure and compliance, it was hypothesized that the cross-sectional morphology of the IVC, which was obtained using three-dimensional echocardiography, might estimate CVP more accurately compared with standard grading by 2DE.Methods:
Sixty consecutive patients who underwent right-heart catheterization studies were prospectively enrolled. Echocardiography was performed <24 hours before catheterization. From three-dimensional data sets, a cross-section of the IVC was determined that was perpendicular to the long-axis reference of the IVC. Short diameter (SD), long diameter (LD), the ratio of SD to LD (S/L) as the sphericity index, and area were measured on this cross-sectional IVC image.Results:
CVP correlated moderately with SD (r = 0.69, P < .001), strongly with S/L (r = 0.75, P < .001), and modestly with area (r = 0.47, P < .001) but not with LD (r = 0.24, P = .17). The largest areas under the curve by receiver operating characteristic analyses to detect CVP ≥ 10 mm Hg were 0.98 (95% CI, 0.97–1.0; P < .001) for S/L, 0.83 for SD (95% CI, 0.74–0.94; P < .001), and 0.70 for area (95% CI, 0.56–0.84; P = .02). If a cutoff value of 0.69 for S/L was used, the sensitivity, specificity, and accuracy to detect CVP ≥ 10 mm Hg were 0.94, 0.95, and 0.95 and for CVP grading by 2DE were 0.59, 0.98, and 0.85, respectively. Estimations of CVP were more accurately reclassified using S/L rather than grading by 2DE (net reclassification improvement, 0.38; 95% CI, 0.31–0.44; P < .001).Conclusions:
S/L of an IVC cross-section measured using three-dimensional echocardiography may be a reliable parameter to estimate CVP compared with standard grading by 2DE.