# Optimized sonographic weight estimation of fetuses over 3500 g using biometry-guided formula selection

Balsyte, D; Schäffer, L; Zimmermann, R; Kurmanavicius, J; Burkhardt, T (2015). Optimized sonographic weight estimation of fetuses over 3500 g using biometry-guided formula selection. Ultraschall in der Medizin, 38(01):60-64.

## Abstract

Purpose: The Hadlock et al. formula tends to underestimate fetal weight, in particular > 3500 g. At the high end of the range, the Merz et al. formula is more accurate, but becomes less so in smaller fetuses. This study was designed to improve fetal weight estimation in fetuses > 3500 g by identifying the fetal biometric parameter providing the most reliable guidance to optimal formula selection. Materials and Methods: Regression analysis of 12 032 pregnancies showed that multiplication of abdominal circumference by femur length (AC × FL) gave the best choice of appropriate formula: Hadlock for AC × FL < 24 600, Merz for those ≥ 24 600. We then tested this rule, ('Zurich method'), prospectively in 4073 pregnancies, comparing it with the Hadlock, Merz and the Kehl formulas. Birth weights were merged into 7 categories (< 1500 to ≥ 4000 g, interval of 500 g). The percentage error (PE) and absolute percentage error (APE) were calculated. Results: The PE using the Zurich method was lower in both > 3500 g groups than with the Hadlock formula alone (3500 - 3999 g: 0.9 % vs. - 5.3 %, > 4000 g: - 3.2 % vs. - 8.6 %), similar to that with the Merz formula alone, and lower than with the Kehl formulas (3500 - 3999 g: - 9.0 % vs. - 3.2 %, > 4000g: - 5.1 % vs. 0.9 %). The Zurich method and Hadlock formula also shared the lowest PE in the < 1500 g group: 0.2 % vs. 6.8 % (Kehl) vs. 9.6 % (Merz). In terms of APE the Zurich method performed almost as well as the Merz formula in the > 4000 g group, while sharing the lowest value with the Hadlock formula in the < 1500 g group (8.2 % vs. 10.5 % [Kehl], 23.6 % [Merz]). Conclusion: The Zurich method uses a pivotal value of the biometry parameter AC × FL to switch between formulas and corrects for the errors of the Hadlock formula in fetuses ≥ 3500 g and those of the Merz formula in fetuses < 3500 g.

## Abstract

Purpose: The Hadlock et al. formula tends to underestimate fetal weight, in particular > 3500 g. At the high end of the range, the Merz et al. formula is more accurate, but becomes less so in smaller fetuses. This study was designed to improve fetal weight estimation in fetuses > 3500 g by identifying the fetal biometric parameter providing the most reliable guidance to optimal formula selection. Materials and Methods: Regression analysis of 12 032 pregnancies showed that multiplication of abdominal circumference by femur length (AC × FL) gave the best choice of appropriate formula: Hadlock for AC × FL < 24 600, Merz for those ≥ 24 600. We then tested this rule, ('Zurich method'), prospectively in 4073 pregnancies, comparing it with the Hadlock, Merz and the Kehl formulas. Birth weights were merged into 7 categories (< 1500 to ≥ 4000 g, interval of 500 g). The percentage error (PE) and absolute percentage error (APE) were calculated. Results: The PE using the Zurich method was lower in both > 3500 g groups than with the Hadlock formula alone (3500 - 3999 g: 0.9 % vs. - 5.3 %, > 4000 g: - 3.2 % vs. - 8.6 %), similar to that with the Merz formula alone, and lower than with the Kehl formulas (3500 - 3999 g: - 9.0 % vs. - 3.2 %, > 4000g: - 5.1 % vs. 0.9 %). The Zurich method and Hadlock formula also shared the lowest PE in the < 1500 g group: 0.2 % vs. 6.8 % (Kehl) vs. 9.6 % (Merz). In terms of APE the Zurich method performed almost as well as the Merz formula in the > 4000 g group, while sharing the lowest value with the Hadlock formula in the < 1500 g group (8.2 % vs. 10.5 % [Kehl], 23.6 % [Merz]). Conclusion: The Zurich method uses a pivotal value of the biometry parameter AC × FL to switch between formulas and corrects for the errors of the Hadlock formula in fetuses ≥ 3500 g and those of the Merz formula in fetuses < 3500 g.

## Statistics

### Citations

Dimensions.ai Metrics