Validity of sonographic prediction of fetal weight and weight discordance in twin pregnancies.

OBJECTIVES

The aims of this study were (1) to assess the accuracy of estimated fetal weight (EFW) in twins and (2) to assess the accuracy of sonographic examination to predict birth weight discordance (BWD).

METHODS

We retrospectively analyzed collected data on twin pregnancies between 2004 and 2007. All twin pregnancies with at least one ultrasound (US) examination within 15 days of delivery were included in this study. EFW was calculated according to Hadlock1, Hadlock2, Ong, Shepard and Warsof formulas. Mean and SD of the standardized errors and percentage of newborns with birth weight (BW) within 10% of EFW were calculated.

RESULTS

Two hundred eighty-three twin pregnancies were included. Mean and SD (%) of the standardized errors were 1.54 +/- 12.19, 0.19 +/- 11.87, 10.93 +/- 15.55, - 1.91 +/- 14.93 and 5.37 +/- 14.91 for Hadlock1, Hadlock2, Shepard, Ong and Warsof formulas, respectively. Hadlock2's formula allowed for the highest proportion of newborns with BW within 10% of EFW and it also performed best to predict discordance of more than 25% as assessed by area under the ROC curve.

CONCLUSIONS

Sonographic prediction of inter-twin BWD within 15 days of delivery seems to be accurate enough for routine clinical use. Performance and predictive values depend on the threshold chosen to define EFW and BW discordance.