Secular trends in preterm birth: a hospital-based cohort study.

Context.—Canada and the United States have reported a recent increase in the incidence of preterm birth, but the reasons for this increase are unknown.Objective.—To assess secular trends in preterm birth and its potential determinants.Design.—Hospital-based cohort study.Setting.—Canadian tertiary care university teaching hospital, 1978-1996.Participants.—A total of 65,574 nonreferred live births and stillbirths.Main Outcome Measures.—Changes in occurrence of preterm birth, before and after adjustment for changes in method of gestational age assessment, obstetric intervention, registration of births weighing less than 500 g, and sociodemographic, behavioral, and clinical determinants.Results.—A crude secular increase in preterm births was seen for births less than 37, 34, and 32 completed weeks using 3 alternative gestational age estimation methods. Based on an algorithm incorporating both menstrual and early ultrasound gestational age estimates, rates increased from 6.6% to 9.8% for births at less than 37 weeks' gestation, 1.7% to 2.3% at less than 34 weeks, and 1.0% to 1.2% at less than 32 weeks. Exclusion of births weighing less than 500 g and those with induction or preterm cesarean delivery without labor before each of the corresponding gestational age cutoffs eliminated the secular trends for births before 34 and 32 weeks and attenuated the trend for births before 37 weeks. Nearly half of the remaining trend for births before 37 weeks was accounted for by the increasing use of early ultrasound dating. The residual trend was eliminated after controlling for secular increases in unmarried status and the proportion of women aged 35 years or older. These factors, combined with a decrease in alcohol consumption and increases in histological chorioamnionitis and cocaine use, appear to have counteracted a reduction in preterm birth since the mid-1980s that otherwise would have been observed.Conclusions.—This hospital's increase in preterm births since 1978 parallels increases reported in population-based national studies from the United States and Canada. This trend appears largely attributable to the increasing use of early ultrasound dating, preterm induction and preterm cesarean delivery without labor, and changes in sociodemographic and behavioral factors.