Patterns of discordant growth and adverse neonatal outcomes in twins

Abstract

Intertwin size discordance is an independent risk factor for adverse neonatal outcomes in twin pregnancies. However, size discordance at a given point in gestation fails to take into consideration information, such as the timing of onset and the rate of progression of discordance, that may be of prognostic value. In this study, we aimed to identify distinct patterns of discordant fetal growth in twin pregnancies and to determine whether these patterns are predictive of adverse pregnancy outcomes. This was a retrospective cohort study of women with twin pregnancies in a single tertiary referral center between January 2011 and April 2020, who had at least 3 ultrasound examinations during pregnancy that included assessment of fetal biometry. Size discordance was calculated at each ultrasound examination, and pregnancies were classified into 1 of 4 predetermined patterns based on the timing of onset and the progression of discordance: pattern 1, no significant discordance group (referent); pattern 2, early (<24 weeks' gestation) progressive discordance group; pattern 3, early discordance with plateau group; or pattern 4, late (≥24 weeks' gestation) discordance group. The associations of discordance pattern (using pattern 1 as referent) with preterm birth, preeclampsia, size discordance at birth, and birthweight<10th percentile were expressed as adjusted relative risk with 95% confidence intervals and were compared with those observed for a single measurement of size discordance at 32 weeks' gestation. Of 2075 women with a twin gestation who were identified during the study period, 1059 met the study criteria. Of the 1059 women, 599 (57%) were classified as no significant discordance (pattern 1), 23 (2%) as early progressive discordance (pattern 2), 160 (15%) as early discordance with plateau (pattern 3), and 277 (26%) as late discordance (pattern 4). The associations of discordance pattern with preterm birth at <34 weeks' gestation and preeclampsia were strongest for pattern 2 (rates of 43% [adjusted relative risk, 3.43; 95% confidence interval, 2.10-5.62] and 17% [adjusted relative risk, 5.81; 95% confidence interval, 2.31-14.60], respectively), intermediate for pattern 3 (rates of 23% [adjusted relative risk, 1.82; 95% confidence interval, 1.28-2.59] and 6% [adjusted relative risk, 2.08; 95% confidence interval, 1.01-4.43], respectively), and weakest for pattern 4 (rates of 12% [adjusted relative risk, 0.96; 95% confidence interval, 0.65-1.42] and 4% [adjusted relative risk, 1.41; 0.68-2.92], respectively). In contrast, a single measurement of size discordance at 32 weeks' gestation showed no association with preeclampsia and only a weak association with preterm birth at <34 weeks' gestation. We identified 4 distinct discordance growth patterns among twins that demonstrated a dose-response relationship with adverse outcomes and seemed to be more informative than a single measurement of size discordance.