Can a single preterm ultrasound accurately predict birth weight in gestational diabetes?

To evaluate the accuracy of the gestation-adjusted projection method of birth weight prediction, as compared to near delivery ultrasound estimated fetal weight, in a gestational diabetic population.

A retrospective cohort was conducted including all women with gestational diabetes who had an ultrasound estimated fetal weight (EFW) between 34(0/7) and 36(6/7) weeks and an additional ultrasound EFW within 7 d of delivery at term. The gestation-adjusted projection (GAP) method was applied to the earlier sonogram, resulting in the GAP predicted birth weight. The GAP predicted weight and the term ultrasound EFW were compared to the actual birth weight. Absolute and percent birth weight errors were compared using paired t-tests.

The mean absolute percent errors and mean absolute errors for the GAP method and term ultrasound were 7.7 ± 5.6% versus 7.1 ± 5.1% and 256 ± 184 g versus 236 ± 169 g respectively (p = 0.22 and p = 0.29). The sensitivity of predicting birth weight ≥4000 g was 22% for the GAP method and 28% for term ultrasound, with specificity reaching 97% for both the methods.

The GAP method is as accurate as term ultrasound in predicting birth weight in gestational diabetes.