Performance of 36 Different Weight Estimation Formulae in Fetuses with Macrosomia

Abstract

Introduction: The aim of this retrospective clinical study was to compare the accuracy of 36 commonly used weight estimation formulae in macrosomic fetuses. Material and Methods: Ultrasound estimation of fetal weight (FW) was carried out within 7 days up to delivery in 350 singleton fetuses with a birth weight (BW) of ≧4,000 g. The accuracy of the different formulae for FW estimation was compared by, firstly, the mean percentage (MPE) and mean absolute percentage error (MAPE), secondly, by the frequency distribution of differences between estimated FW and fetal BW, and thirdly by comparing detection and false positive rates in screening for fetuses with a BW of 4,000, 4,300 and 4,500 g or more. Results: MPE ranged from –62.2 to 9.6% and was closest to 0 with the Hart formula. With 12 of 36 weight estimation formulae, MAPE was 10% or less, and was smallest with the Hart formulae (3.9%). The mean detection rate among all formulae for fetuses with a BW ≧4,000, ≧4,300 and ≧4,500 g was 29, 24 and 22%, respectively, and the false positive rate was 12% (for ≧4,300 g) and 7% (≧4,500 g). Discussion: Some formulae showed advantages as far as mean and absolute percentage errors were concerned, but none reached a detection rate and false positive rate for fetuses ≧4,500 g that could lead to clinical recommendation.