Transvaginal ultrasonography at 18-23 weeks in predicting placenta previa at delivery

Objective To evaluate whether transvaginal ultrasonography at 18-23 weeks' gestation is useful in predicting placenta previa at delivery. Design We performed transvaginal ultrasonography in addition to routine tra nsabdominal ultrasonography in 3696 consecutive non-selected pregnant women with singleton fetuses and measured the distance from the placental edge t o the internal cervical os. Results In 57 of 3696 patients (1.5%), the placental edge extended to or ov er the internal cervical os. In 27 patients (0.7%), the placenta extended g reater than or equal to 15 mm over the internal cervical os; in these crise s the positive predictive value of placenta previa at delivery was 19% (95% CI, 6-38%) with 100% (95% CI, 48-100%) sensitivity. With greater than or e qual to 25 mm used as the cut-off point, ten cases (0.3%) were screen-posit ive and the positive predictive value for previa at delivery wets 40% (95% CI, 12-74%) and sensitivity was 80% (95% CI, 28-100%). The frequency of pla centa previa at delivery in this population was five of 3696 (0.14%, 95% CI , 0.04-0.31%). Conclusions We recommend confirmatory transvaginal ultrasonography if place nta previa is suspected at transabdominal ultrasonography in mid-pregnancy and re-examination at 26-30 weeks if the placental edge covers the internal cervical os by 15 mm or more.