Twin birth weight discordance and risk of preterm birth.

Abstract

Our purpose was to determine whether birth weight discordance is a risk factor for preterm birth of twins, and to further characterize the relationships involved. Maternally linked 1978-1990 Missouri birth certificates were used to analyze gestations resulting in live twins. We used contingency tables and multiple logistic regression. The degree of discordance correlated strongly with risk for live preterm birth but only for discordances >30% and preterm birth at <32 weeks' gestation. Among 9479 pregnancies with discordance <30%, 9.5% ended in birth at <32 weeks' gestation, versus 13.7% of 326 with discordance of 30% to 40% (P =.03) and versus 34.1% of 126 with discordance > or =40% (P <. 001). There were 42 preterm twin births at <32 weeks' gestation with discordances > or =40%. Of these, 51% were attributable to fetal growth restriction and 16% to large size for gestational age in one infant; in 72% the smaller twin was the second born, and in 86% the twins were like sex. The relative association between > or =40% discordance and preterm birth at <32 weeks' gestation was strengthened (final odds ratio, 9.54; P <.0001) in a multivariate model containing other risk factors for delivery at <32 weeks' gestation: black race, either twin small for gestational age, unmarried, teenage mother, number of male fetuses, like fetal sex, education <12 years, nulliparity, and cigarette smoking. Twin birth weight discordance has now clearly been demonstrated to be a risk factor for preterm birth. The effect was found particularly with discordances > or =40% before 32 weeks' gestation. Discordance was usually attributable to fetal growth restriction, most often in the second-born twin.