RESUMO
OBJECTIVE: Birthweight discordance amongst twins is associated with an increase in perinatal morbidity and mortality. Clinical decisions regarding the management of twins are often made on the basis of estimated fetal weight in the third trimester. This study assesses the efficacy of routine third trimester ultrasound fetal biometry, third trimester fetal growth velocity and first trimester differences in size in the prediction of subsequent birthweight discordance. STUDY DESIGN: Structurally and chromosomally normal dichorionic twin gestations resulting in two live births after 24 weeks gestation were identified. Ultrasound examinations of dichorionic twin pregnancies between 10 and 14 weeks of gestation were evaluated for inter-twin biometric disparity in crown-rump length (CRL). Estimated fetal weight (EFW) was calculated from the last scan before delivery. Fetal growth velocity (FGV) was calculated from two estimates of fetal weight between 21 and 35 days apart and incorporating the last EFW before delivery. EFW and FGV were expressed as standard deviation scores (Z scores). Disparity between these ultrasound parameters was correlated to inter-twin birthweight difference. Birthweight discordance was defined as > or =20% difference in birthweight relative to the heavier twin. Fishers exact and Spearman rho tests were used for statistical analysis. Test performance is expressed as area under the ROC curve and likelihood ratios (LR). RESULTS: One hundred and thirty-five dichorionic gestations were identified over a consecutive 3-year-period. One hundred and eight twin gestations had both crown-rump length (CRL) and sufficient ultrasound measurements in the third trimester to calculate FGV. Twenty-six pregnancies (24%) had discordant birthweight. Only EFW Z score difference was correlated with birthweight disparity (r=0.36, p = <0.001). An EFW Z score difference of 0.6 had an area under the ROC curve of 0.70 (95% C.I. 0.59-0.79) but this did not differ significantly from first trimester prediction of growth discordance which had an area of 0.55 (95% C.I. 0.44-0.66). The LR for first and third trimester prediction of discordance were low. CONCLUSION: Ultrasound measurement of fetal growth velocity and size both in the first and third trimesters is a poor predictor of birthweight discordance. Disparate fetal growth is exhibited as early as 10-14 weeks gestation and differences at this early stage in fetal development are no less accurate than EFW within 1 month of delivery. First trimester ultrasound, third trimester growth velocity and third trimester EFW do not accurately predict the delivery of twins with discordant birthweight.