Reliability analysis on ductus venosus assessment at 11-14 weeks' gestation in a high-risk population

Objective To assess the intra- and interobserver reliability of the fetal ductus venosus (DV) blood flow measurements at 11 - 14 weeks' gestation in a high-risk population for chromosomal defects. Methods Thirty-five consecutive singleton pregnancies at high risk for chromosomal or genetic anomalies were included. DV end-diastolic velocity (EDV) and the pulsatility index for veins (PIV) were determined by manual tracing in three sets of waveforms by a single sonographist to assess intraobserver reliability using transvaginal ultrasound. Subsequently, the EDV and the PIV were also determined in a further set by a second sonographist to assess interobserver reliability. Reliability was evaluated by means of the Intraclass Correlation Coefficient (ICC). The degree of agreement was also examined using the limits of agreement method. Results The intraobserver reliability was 0.81 (95% C.I. 0.69-0.89) for EDV and 0.91 (95% C.I. 0.85-0.95) for the PIV, both corresponding to almost perfect reliability. The interobserver reliability was 0.50 (95% C.I. 0.20-0.70) for EDV, and 0.74 (95% C.I. 0.54-0.86) for PIV, corresponding respectively to moderate and substantial reliabilities. The 95% interval of the EDV differences (cm/s) within and between observers were (+6.1, -4.6) and (+13.6, -9.7), respectively. The 95% interval of the PIV differences within and between observers were (+0.32, -0.36) and (+0.7, -0.67), respectively. Conclusions Interobserver reliability for EDV is only moderate, whereas for the PIV is substantial, allowing its use for clinical purposes in a high-risk population. Copyright (c) 2007 John Wiley & Sons, Ltd.