Comparison of Doppler-based and three-dimensional methods for fetal cardiac output measurement.

Abstract

Fetal cardiac output is conventionally measured using two-dimensional (2D) and Doppler ultrasound (Doppler). New methods based on 3D measurements of ventricular size in systole and diastole have been proposed. Our aim was to validate these tools against the conventional Doppler-based methods.
Fetal combined cardiac output was prospectively measured at 16, 20, and 24 weeks of gestation in 15 uncomplicated pregnancies using Doppler and three different 3D algorithms [virtual organ computer-aided analysis (VOCAL), sonographic automatic volume calculation (SonoAVC), and inversion mode]. We determined the inter- and intraobserver variability of the 3D techniques and assessed the correlation between Doppler and 3D.
The 3D techniques showed adequate inter- and intraobserver reproducibility (intraclass correlation coefficient 0.69-0.98), with the best reproducibility for SonoAVC and inversion mode. Bland-Altman analysis revealed low bias and relatively good correlations when comparing the 3D methods among each other, albeit with wide 95% confidence intervals. Doppler measurement of fetal weight-adjusted combined cardiac output (349.0 ml·min(-1)·kg(-1)) yielded significantly higher results than 3D CO measurements (177.2, 160.7, and 174.0 ml·min(-1)·kg(-1) for VOCAL, SonoAVC, and inversion mode, respectively; p < 0.0001) and correlated poorly with the 3D methods.
Although 3D volume-based cardiac output measurements are reproducible, results obtained with different methods are not interchangeable. SonoAVC and inversion have the highest intra- and interobserver reproducibility. Results of cardiac output measurement by 3D and Doppler cannot be interchanged.