A computerised method for accurately predicting fetal macrosomia up to 11 weeks before delivery.
Gerad G Nahum, Harold Stanislaw
European Journal of Obstetrics and Gynaecology and Reproductive Biology
(2007) 133:148-156

Fetal macrosomia is associated with increased peripartum complications including shoulder dystocia, instrumental delivery and caesarean section. At present we are unable to accurately predict fetal birth weight. This current study aimed to improve the accuracy of prediction using a scientific formula which combined ultrasound results and pregnancy specific characteristics.

Retrospective data was obtained from 218 normal gravidas who had had an ultrasound performed within 11 weeks of delivery. A set of mathematical equations was derived from multiple regression looking at various combinations of variables. Women with medical complications, premature delivery or multiple pregnancies were excluded.

38 equations derived were statistically significant in improving accuracy of fetal weight prediction. 11 were sensitive enough to predict macrosomia, with 1 equation having 93% specificity and a likelihood ratio of 10.3. The overall sensitivity for detection of fetal macroosomia was 75% with positive and negative predictive values of 67% and 95% respectively if scans were performed within 3 weeks of delivery.

The authors concluded that use of combination equations improved the accuracy of fetal weight prediction in normal gravidas. Unlike ultrasonographic measurements alone, which are optimal for predicting birth weight only within 1 week of delivery, they can be used up to 11 weeks pre-delivery. Furthermore, the large number of equations increases flexibility when available data is limited, and can be used on women without ultrasound measurements as a means of screening. Future research should determine whether use of such formulas can reduce adverse outcomes in mothers predicted to deliver macrosomic infants.