Grand Rounds

Are New Predictors of Fetal Weight Superior to Traditional Ultrasound?

BACKGROUND AND PURPOSE:

2 common components used to identify SGA include
- EFW
- Growth curve to determine percentile
Hadlock (US derived growth curve) remains an ultrasound standard, however there are 2 new curves using alternate formulae that may be more generalizable
- Intergrowth-21st (INTG) method (international population)
- Salomon method (Paris, France population)
Blue et al. (Journal of Ultrasound in Medicine, 2018) compared the traditional Hadlock method with these 2 newer methods for the SGA

Retrospective review
- Singleton, non-anomalous, well-dates fetuses were examined within 2 weeks of delivery
Fetal growth restriction (FGR): EFW <10th percentile
SGA: Birth weight <10th percentile
Estimated fetal weights and percentiles were calculated by
- Hadlock method
- INTG method
- Salomon method
Outcomes: Test characteristics to predict SGA were calculated using
- EFW < 10th percentile (standard method)
- Receiver operating characteristic (ROC) – derived FGR cutoffs

831 pregnancies
- SGA: 16.7%
Plotting the ROC curves using a 10^th percentile cutoff, Hadlock performed better than the other methods in the prediction of SGA
Comparing overall performance and accuracy of birth weight percentile prediction
- Hadlock and INTG methods performed comparably (P = .08)
- Hadlock area under the curve: 0.91
- INTG area under the curve: 0.90
Optimal EFW cutoffs
- Hadlock: 15th percentile
- INTG: 22nd percentile
Salomon method performed less well than other 2 methods (P<.001)
- Area under the curve: 0.82

The traditional Hadlock had the best sensitivity and ROCs to predict SGA (using 10% cutoff)
Hadlock was comparable to INTG for the prediction of SGA when ROC‐derived cutoffs were used
Using a 15^th percentile cutoff for Hadlock EFW or 22^nd percentile for INTG EFW may improve the ability of ultrasound to predict SGA

< Previous

Next >