Conflicting evidence exists regarding the reliability of third trimester ultrasound for guiding delivery management for fetal macrosomia. Our objective was to determine the performance of third trimester ultrasound in diabetic and non-diabetic women with fetal macrosomia.METHODS:
We performed a retrospective cohort study of fetal ultrasounds from 2004–2014 with estimated fetal weight (EFW) above 4,000 grams (gm). We determined accuracy of birth weight prediction for ultrasound performed at less than and greater than 38 weeks, in non-diabetic and diabetic women. We used student t test, Wilcoxon rank sum test, chi-square, Fisher exact test, and multivariate logistic regression.RESULTS:
There were 405 fetal ultrasounds with EFW above 4,000 gm. 112 (27.7%) were performed at less than 38 weeks, 293 (72.3%) at greater than 38 weeks, and 91 (22.5%) in diabetics. Sonographic identification of EFW over 4000 gm at less than 38 weeks was associated with higher birth weight (4183 gm vs 4019 gm, P=.001) and higher correlation between EFW and birth weight than ultrasound performed after 38 weeks (71.5 gm vs 259.4 gm, P<.001). EFW to birthweight correlation was within 1.7% of EFW for ultrasound performed less than 38 weeks. Diabetics had larger birth weight then non-diabetics (4198 gm vs 4022 g, P<.001) but the correlation between EFW and birth weight was the same after adjusting for gestational age at ultrasound (102gm vs 252 gm, P=.103). In all cohorts, average EFW overestimated average birth weight.CONCLUSION:
Identification of fetal weight over 4000 gm in patients with ultrasound performed less than 38 weeks has greater reliability of predicting fetal macrosomia at birth than measurements performed greater than 38 weeks.