Foetal weight estimation, at term, using a multivariate algorithm of maternal characteristics, has an accuracy similar to that of ultrasonography.

Abstract

Objective: To compare the accuracy of foetal weight estimation using a multivariate algorithm based on maternal characteristics and pregnancy-specific factors to that of ultrasound.

Design: A cross-sectional hospital-based study.

Setting: Antenatal Clinics and Antenatal Labour wards of the Department of Obstetrics & Gynaecology, Federal Medical Centre, Asaba, Nigeria.

Participants: A total of 92 pregnant women were admitted for labour, elective caesarean section or elective induction of labour at 37 weeks to 41 weeks and 3 days.

Main outcome measures: Mean of absolute error, mean of absolute percentage error and proportion of estimated weight within 10% of actual birth weight.

Results: Between April and August 2021, 92 participants were included. An equation based on maternal characteristics was derived. Ultrasound weight estimation was done using Hadlock's 4 formula. Both methods positively correlated with actual birth weight, and their accuracy did not differ significantly. Overall accuracy within 10% of actual birth weight was higher for ultrasonography than multivariate algorithm 71.7% and 65.2%, respectively (χ²=0.286, p=0.60). The mean absolute percentage error was smaller for ultrasound (7.98±4.74%) than clinical formula (9.11±6.76%) p=0.11. The mean absolute error was 265.57±145.67g for ultrasonography and 304.32±203.29g for the multivariate model, with no statistical difference (p=0.09).

Conclusion: The multivariate algorithm based on maternal characteristics and pregnancy-specific factors was equally accurate as ultrasonography for foetal weight estimation at term.

Funding: None declared.