Accuracy of sonographic fetal weight estimation of fetuses with a birth weight of 1500 g or less

Objective: To compare 35 commonly used formulae for small and average sized fetuses on their accuracy in estimating the birth weight in fetuses of 1500 g or less.

Study design: For this retrospective study a database search was performed for all singleton pregnancies without structural or chromosomal defects and with a birth weight of 1500 g or less where the last ultrasound examination was performed within seven days before delivery. Percentage error and absolute percentage error were calculated based on 35 different weight estimation formulae. Multiple regression analysis was used to determine the significant contributors to the absolute percentage error.

Results: One hundred and ninety-three cases fulfilled the inclusion criteria. The median birth weight was 990 g. The percentage error ranged between -15.2% (underestimation with the Merz I formula) and 37.4% (overestimation with the Jordaan formula) and the respective standard deviations between 10.5% (Mielke I) and 54.0% (Schillinger), respectively. The absolute percentage error was between 8.5% and 37.6%. The most accurate weight estimation was achieved with the formula from Mielke (percentage error 1.8% and absolute percentage error 8.5%). Multiple regression analysis showed that significant contributors to the percentage error of the Mielke formula were biparietal diameter (OR=-0.206, p=0.045), occipitofrontal diameter (OR=0.765, p<0.0001), abdominal circumference (OR=-2.953, p<0.0001), femur length (OR=-0.903, p<0.0001), head to abdomen ratio (OR=-1.080, p<0.0001) and fetal weight (OR=2.847, p<0.0001).

Conclusion: When estimating fetal weight in fetuses weighing 1500 g or less, one has to be aware of the great differences in accuracy among the formulae.