Abdominal circumference discordance for prediction of small for gestational age at birth in twin pregnancies

Objective To evaluate whether single sonographic abdominal circumference (AC) discordancy estimation can predict small for gestational age (SGA) at birth in twin gestations. Methods A retrospective analysis of prospectively collected data. Cohort included all twin gestations delivered at one university-affiliated medical center between 2010 and 2018, with available sonographic evaluation from 22 gestational weeks to term. Pregnancies complicated by fetal chromosomal abnormalities, major anomalies or twin to twin transfusion syndrome were excluded. One sonographic evaluation per pregnancy was selected randomly. AC discordance was calculated as (large twin AC - small twin AC)/large twin AC*100. Prediction of SGA at birth for at least one newborn (<10% percentile for gestational age by gender-specific local curves for multiples) was evaluated using ROC statistics with calculation of Youden index to establish best AC discordance cutoff. AC discordance prediction performance was compared to estimated fetal weight discordance performance. Results were adjusted for confounders using logistic regression analysis. Results After exclusion, 236 twin gestations entered analysis. Of them, 200/236 (84.7%) were dichorionic-diamniotic twins. Mean gestational age at ultrasound evaluation and at delivery were 30.9 +/- 4.4 and 35.9 +/- 2.4 weeks, respectively. In 28/236 (11.8%) pregnancies, at least one neonate was born SGA. AC discordance predicted SGA at birth as good as sonographic estimated fetal weight (sEFW) discordance: ROC-AUC 0.76, 95% CI 0.67-0.85 vs. 0.77 95% CI 0.66-0.87,p < .001 for all. Best AC discordance cutoff for prediction of SGA at birth was 7.1% (57% sensitivity, 87% specificity), ROC-AUC 0.72 (95% CI 0.61-0.84,p < .001). Results remained significant after adjustment for maternal age, nulliparity, chorionicity and ultrasound to delivery interval (aOR 1.21 95% CI 1.1-1.32,p < .001). Conclusion According to our results, AC discordance at single sonographic evaluation can predict SGA at birth in twin gestations as good as sEFW discordance. Best cutoff for SGA prediction was 7.1%.