Measuring leading placental edge to internal cervical os: Transabdominal versus transvaginal approach

We aimed to compare the value of transabdominal (TA) and transvaginal (TV) approaches for assessing the risk of a low-lying placenta. This involved a comparison of TA and TV measurements between the leading placental edge and the internal cervical os. We also assessed the intra-/interobserver variation for these measurements and the efficacy of TA measures in screening for a low placenta.

Transabdominal and TV measurements of the leading placental edge to the internal cervical os were performed on 369 consecutive pregnancies of 16-41 weeks' gestation. The difference (TA-TV) from the mean was calculated and plotted against gestational age. Bland-Altman plots and paired t-tests were used to look at the differences in TA/TV measurement. Screening performance of a transabdominal approach was compared to a transvaginal 'gold standard'. Nonparametric methods were used to calculate the area under the receiver operator characteristics (ROC) curve. Intra-/interobserver variations were also calculated.

Of the pregnancies, 278 had a leading placental edge that was visible with the TV approach. Differences (TA-TV) ranged from -50 mm to +57 mm. Bland-Altman plot shows that TA measurements overestimated the distance compared with the TV measurements; the average difference in measurement was 12.0 mm (95% confidence interval 9.9-14.1). The sensitivity, specificity and negative predictive values of a TA approach were 18.2%, 97.5% and 87.2%, respectively. The receiver operator characteristics area between gestational weeks 16-23 was 0.81 (95% CI: 0.76-0.86).

The TA approach has a low sensitivity for detecting a low-lying placenta as choosing a TA cut-off with sensitivity >90% will decrease the specificity to 50%.