The aim of the study is to examine the association between the risk of preterm delivery among women with twin pregnancies and their obstetric history.
 We designed a retrospective cohort study of live twin births in 2008 in the United States that delivered after 24 weeks. Women were categorized into nulliparas, multiparas with prior term delivery, and multiparas with prior preterm delivery. The incidence of preterm birth was compared using Chi-square test and multivariable logistic regression models.
 A total of 32,895 nulliparous and 64,701 multiparous women with twin pregnancies were included in the study. Of the multiparous women, 2,505 (4%) had a history of a prior preterm delivery. Multiparous women with prior term birth were more likely to deliver at term (: 43%): in the index twin pregnancy than nulliparous women (40%) and multiparous women with a prior preterm birth (21%;  < 0.001). Compared with nulliparous women, prior term birth was protective against preterm delivery (adjusted odds ratio [aOR] = 0.67 [95% confidence interval 0.60-0.74] for delivery <28 weeks and aOR = 0.79 [0.71-0.77] for delivery <34 weeks).
 Among multiparous women with twins, a prior term delivery appeared to be protective against preterm delivery compared with nulliparous women with twins.
· Prior term birth is protective against preterm birth in subsequent twin pregnancy.. · A prior term birth confers an OR of 0.66 for delivery prior to 28 weeks in twin pregnancies.. · A prior preterm birth renders a twin pregnancy nearly twice as likely to deliver before 28 weeks..

Thieme. All rights reserved.