Accuracy of sonographic lower segment thickness and prediction of vaginal birth after caesarean in a resourced-limited setting; Prospective study

Abstract

Objectives

To assess the accuracy of ultrasound measurement of the lower uterine segment (LUS) thickness against findings at laparotomy, and to investigate its correlation with the success rate of vaginal birth after one previous caesarean delivery (CD) in a resource-limited setting.

Design

Prospective study.

Setting

Obstetrics and Gynaecology department in a tertiary hospital in Ghana.

Population

Women with one previous CD undergoing either a trial of labour (TOLAC) or elective CD.

Methods

Myometrial lower uterine segment thickness (mLUS) and full lower uterine segment thickness (fLUS) were measured with transvaginal ultrasound (TVUS). The women were managed according to local protocols with the clinicians blinded to the ultrasound measurements. The LUS was measured intraoperatively for comparison with ultrasound measurements.

Main outcome measures

Lower uterine segment findings at laparotomy, successful vaginal birth.

Results

A total of 311 pregnant women with one previous CD were enrolled; 147 women underwent elective CD and 164 women underwent a TOLAC. Of the women that underwent TOLAC, 96 (58.5%) women had a successful vaginal birth. The mLUS was comparable to the intraoperative measurement in the elective CD group with LUS thickness <5 mm (bias of 0.01, 95% CI −0.10 to 0.12 mm) whereas fLUS overestimated LUS <5 mm (bias of 0.93, 95% CI 0.80–1.06 mm). Successful vaginal birth rate correlated with increasing mLUS values (odds ratio 1.30, 95% CI 1.03–1.64). Twelve cases of uterine defect were recorded. LUS measurement ≤2.0 mm was associated with an increased risk of uterine defects with a sensitivity of 91.7% (95% CI 61.5–99.8%) and specificity of 81.8% (95% CI 75.8–86.8%).

Conclusion

Accurate TVUS measurement of the LUS is technically feasible in a resource-limited setting. This approach could help in making safer decisions on mode of birth in limited-resource settings.