Association of sociodemographic variables with risk for very preterm birth in twins

To assess the influence of maternal race, age, marital status, and education on risk for earlier and later preterm births in twin pregnancies.

We analyzed 8109 white and 1906 black liveborn twin pregnancies in the Missouri Linked Sibship files for the years 1978-1990, using contingency tables and multiple logistic regression.

Black twin gestations had 1.61-fold (95% confidence interval [CI] 1.46, 1.76) greater risk than whites for preterm birth before 34 weeks' gestation. However, there was no race difference after 33 weeks. Among whites, teen age, unmarried status, and education fewer than 12 years were independently associated with risk for delivery before 34 weeks in multivariate analysis (odds ratios [OR] 1.28-1.51, each P < or=.001). These associations were diminished or absent for preterm births after 33 weeks' gestation. White unmarried teen mothers with fewer than 12 years of education had 1.83-fold (95% CI 1.39, 2.40) greater risk for preterm birth before 34 weeks' gestation compared with white married women more than 19 years of age with at least 12 years of education. In blacks, this difference was 1.47-fold (95% CI 1.13, 1.92). In both races, these differences were absent after 33 weeks' gestation.

Traditional sociodemographic risk factors were present for twin preterm birth, but mainly before 34 weeks' gestation. This, together with previous data from Missouri Linked Sibship files, indicates that dominant pathogenic mechanisms of early preterm birth in twin gestations are likely to be different from those causing later preterm twin birth. Therefore, gestational age should be accounted for in future studies seeking to identify predictive factors or biomechanisms for twin preterm birth.