Birth-weight discordance and neonatal morbidity in twin pregnancy: analysis of STORK multiple pregnancy cohort

Objectives To investigate the relationship between weight discordance and neonatal morbidity in twin pregnancy progressing to at least 34 weeks of gestation. The secondary aim was to determine the predictive accuracy of different weight discordance cut-offs in predicting neonatal morbidity in twin pregnancy. Methods This was a retrospective multicenter cohort study of all twin pregnancies booked for antenatal care at four hospitals in the Southwest Thames region of London Obstetric Research Collaborative (STORK) over a period of 10 years. Ultrasound data were obtained by a search of each hospital's obstetric ultrasound computer database, while outcome details were obtained from the computerized maternity and neonatal records. The primary outcome was incidence of composite neonatal morbidity in twin pregnancy with birth-weight discordance. Logistic regression was used to identify and adjust for potential confounders, while a receiver-operating characteristics (ROC) curve was used to determine predictive accuracy. Results Nine hundred and thirty-nine twin pregnancies (760 dichorionic, 179 monochorionic) were included. Gestational age at birth and birth-weight decile were significantly lower in pregnancies complicated by neonatal morbidity compared with those which were not (P < 0.001 for both). On multivariable logistic regression analysis, gestational age at birth (P < 0.001), birth-weight decile (P = 0.029) and birth-weight discordance (P = 0.019), but not chorionicity (P = 0.477) or presence of at least one small-for-gestational-age (SGA) twin (P = 0.245), were associated independently with the risk of neonatal morbidity. There was a progressive increase in the risk of neonatal morbidity with increasing birth-weight discordance. Despite this association, birth-weight discordance showed an overall poor predictive accuracy for neonatal morbidity, with an area under the ROC curve of 0.58 (95% CI, 0.53-0.63) with an optimal cut-off of 17.6%, showing sensitivity and specificity of 35.2% (95% CI, 27.8-43.2%) and 83.2% (95% CI, 80.4-85.8%), respectively. Conclusion Intertwin birth-weight discordance is associated independently with the risk of neonatal morbidity in twins born after 34weeks' gestation, irrespective of chorionicity or diagnosis of SGA in either twin. However, its predictive accuracy for neonatal morbidity is poor. Copyright (c) 2017 ISUOG. Published by John Wiley & Sons Ltd.