Multivariable evaluation of term birth weight: a comparison between ultrasound biometry and symphysis-fundal height

Abstract

Objective: To derive a birth weight predictive equation and to compare its diagnostic value with that of ultrasound.

Methods: A longitudinal observational cohort study, including singleton pregnancies at term, was performed at St. Orsola-Malpighi Hospital, University of Bologna (Italy). A birth weight prediction formula, including symphysis-fundal height (SFH), BMI, maternal abdominal circumference (mAC) and parity was derived from a general linear model (GLM) (retrospective study). Moreover, on a new series of patients, the fetal weight was estimated by using both GLM and ultrasound using Hadlock formula (prospective study). The residual analysis and the intraclass correlation coefficient (ICC) were used to test the accuracy of methods in predicting birth weight.

Results: Between January and November 2012, 1034 patients were included in the retrospective study and 44 in the prospective one. The following GLM was derived: estimated birth weight (g) = 1485.61 + (SFH (cm) × 23.37) + (11.62 (cm) × mAC) + [BMI × (-6.81)] + (parity (0 = nulliparous, 1 = multiparous) × 72.25). When prospectively applied, the GLM and ultrasound provided a percentage of prediction within ±10% of the actual weight of 73% and 84%, respectively. Ultrasound estimation, as opposite of GLM one, was significantly associated with neonatal weight (R(2 )= 0.388, F = 26.607, p value <0.001, ICC = 0.767).

Conclusions: Although ultrasound biometry has provided the best values in fetal weight estimation, the predictive performance of both methods is limited.

Keywords: General linear model; multivariable birth weight prediction; symphysis-fundal height; term of pregnancy; ultrasound.