Pregnancy screening by uterine artery doppler velocimetry-Which criterion performs best?

Abstract

Objective: To test whether repeating Doppler studies of the uteroplacental circulation late in gestation will improve the test's power for predicting pregnancy-induced hypertension and fetal growth restriction (FGR), and whether analysis based on a combination of quantitative and qualitative assessments of the uterine arterial waveforms will yield better results than analysis based on either alone. Methods: A total of 358 patients considered to be at medium risk for the development of pregnancy-induced hypertension and FGR were recruited. Continuous-wave Doppler studies of the uterine arteries were performed serially at 20, 28, and 36 weeks' gestation. The values of various Doppler indices in the prediction of subsequent pregnancy complications were tested at different gestations and different cutoff levels. The overall significance of performance of the Doppler indices was assessed by the Cohen kappa index, which tests the extent of agreement between the test and the "truth" over that by random chance agreement. Results: A total of 974 examination results on 334 patients were available for analysis. We found that Doppler studies of the uterine arteries at 28 and 36 weeks were less useful than studies performed at 20 weeks. Serial studies at 20 and 28 weeks showed only marginal improvement when compared with a single study at 20 weeks. The best criteria were a mean resistance index (RI) of uterine arteries above the 90th percentile and the presence of diastolic notches in both uterine arteries at 20 weeks. Although the overall kappa index only suggested fair to good agreement beyond chance, the positive predictive value for subsequent complications was good: 57% for severe complications and 93% for any complications. Conclusion: Doppler studies of the uterine artery as a test for the subsequent development of pregnancy complications are best performed at 20 weeks with a combination of RI measurements and the assessment of the presence of diastolic notches. © 1995 The American College of Obstetricians and Gynecologists.