Objective: Our purpose was to determine whether an abnormal uterine perfusion pattern was associated with subsequent pregnancy loss after fetal cardiac activity was documented. Study Design: Pulsatility indexes of both the right and left uterine arteries were obtained by transvaginal color Doppler ultrasonography in 318 consecutive viable pregnancies between 6 and 12 weeks’ gestation. The Δ uterine artery pulsatility index value, expressed as the highest uterine artery pulsatility index value minus the lowest value, was calculated for each pregnancy. Women were subsequently classified as having continuing pregnancies or pregnancy loss before 20 weeks’ gestation. To predict subsequent pregnancy loss, Doppler findings were adjusted for maternal age, history of previous abortion, presence of subchorionic hematoma, embryonic bradycardia, and gestational age by means of multivariate logistic regression analysis. Results : Twenty-four pregnancies (8%) were spontaneously aborted before 20 weeks’ gestation. Both Δ uterine artery pulsatility index (odds ratio 2.9, 95% confidence interval 1.5-5.8) and history of previous abortion (odds ratio 3.1, 95% confidence interval 1.2-8.2) were significantly associated with pregnancy loss in the multivariate logistic regression analysis. The sensitivity and specificity of the multivariate logistic regression model to predict abortion were 75% and 85%, respectively, significantly higher than the diagnostic performances of qualitative and quantitative variables considered individually. Conclusion : Discordant uterine artery pulsatility indexes in the first trimester were strongly associated with subsequent pregnancy loss. This suggests that uterine ischemia may be implicated in certain cases of early pregnancy loss after documentation of fetal cardiac activity during the first trimester. (Am J Obstet Gynecol 1998;179:1587-93.)