Predicting preterm birth in twin pregnancy: was the previous birth preterm? A Canadian experience

Most studies determining risk of preterm birth in a twin pregnancy subsequent to a previous preterm birth are based on linkage studies or small sample size. We wished to identify recurrent risk factors in a cohort of mothers with a twin pregnancy, eliminating all known confounders.

We conducted a retrospective cohort study of twin births at a tertiary care centre in Montreal, Quebec, between 1994 and 2008, extracting information, including chorionicity, from patient charts. To avoid the effect of confounding factors, we included only women with a preceding singleton pregnancy and excluded twin-to-twin transfusion syndrome, fetal chromosomal/structural anomalies, fetal demise, and preterm iatrogenic delivery for reasons not encountered in both pregnancies. We used multiple regression and sensitivity analyses to determine recurrent risk factors.

Of 1474 twin pregnancies, 576 met the inclusion criteria. Of these, 309 (53.6%) delivered before 37 weeks. Preterm birth in twins was strongly associated with preterm birth of the preceding singleton (adjusted OR 3.23; 95% CI 1.75 to 5.98). The only other risk factors were monochorionic twins (adjusted OR 1.82; 95% CI 1.21 to 2.73) and oldest or youngest maternal ages. Chronic or gestational hypertension, preeclampsia, and insulin-dependent diabetes during the singleton pregnancy did not significantly affect risk.

Preterm birth in a previous singleton pregnancy was confirmed as an independent risk factor for preterm birth in a subsequent twin pregnancy. This three-fold increase in risk remained stable regardless of year of birth, inclusion/exclusion of pregnancies following assisted reproduction, or defining preterm birth as < 34 or < 37 weeks' gestational age. Until the advent of optimal preventive strategies, close obstetric surveillance of twin pregnancies is warranted.