Development and Validation of an Algorithm to Determine Spontaneous versus Provider-Initiated Preterm Birth in US Vital Records : Spontaneous vs. provider-initiated preterm birth

Although two thirds of infant deaths in the United States occur among infants born preterm (<37 weeks of gestation), only 17% of infant deaths are classified as being attributable to preterm birth with the standard classification of leading causes of death. To address this apparent discrepancy, we sought to estimate more accurately the contribution of preterm birth to infant mortality rates in the United States. We identified the top 20 leading causes of infant death in 2002 in the US linked birth/infant death file. The role of preterm birth for each cause was assessed by determining the proportion of infants who were born preterm for each cause of death and by considering the biological connection between preterm birth and the specific cause of death. Of 27970 records in the linked birth/infant death file for 2002, the 20 leading causes accounted for 22273 deaths (80% of all infant deaths). Among infant deaths attributable to the 20 leading causes, we classified 9596 infant deaths (34.3% of all infant deaths) as attributable to preterm birth. Ninety-five percent of those deaths occurred among infants who were born at <32 weeks of gestation and weighed <1500 g, and two thirds of those deaths occurred during the first 24 hours of life. On the basis of this evaluation, preterm birth is the most frequent cause of infant death in the United States, accounting for at least one third of infant deaths in 2002. The extreme prematurity of most of the infants and their short survival indicate that reducing infant mortality rates requires a comprehensive agenda to identify, to test, and to implement effective strategies for the prevention of preterm birth.

The objective of the study was to evaluate the extent to which maternal and fetal conditions necessitate medically indicated preterm birth. A population-based, retrospective, cohort study of women who delivered a singleton live birth at 20 weeks or longer in Missouri, 1989 to 1997 was performed (n = 684,711). Maternal-fetal conditions that necessitated iatrogenic preterm birth included preeclampsia, small-for-gestational-age birth, fetal distress, placental abruption, placenta previa, unexplained vaginal bleeding, pregestational and gestational diabetes, renal disease, Rh sensitization, and congenital malformations. We examined the association between each of the aforementioned conditions and risk of medically indicated preterm birth at less than 35 weeks. Medically indicated preterm birth was defined as a labor induction or a prelabor cesarean in the absence of premature rupture of membranes at preterm gestations. Adjusted relative risk with 95% confidence interval for preterm birth was derived from multivariable logistic regression models, and population attributable fractions were calculated. The preterm birth rate (less than 35 weeks) was 4.6% (n = 31,238), with 23.5% (n = 7,347) of such births being medically indicated. Preeclampsia, fetal distress, small-for-gestational-age, and placental abruption were the most common indications for a medical intervention resulting in preterm birth, with at least 1 of these conditions present in 53.2% of medically indicated preterm births and in 17.7% of term births (relative risk 4.9, 95% confidence interval 4.7, 5.2). Preeclampsia, fetal distress, small-for-gestational-age, and placental abruption, conditions that are associated with ischemic placental disease, are implicated in well over half of all medically indicated preterm births. Although the etiology of preterm birth is heterogeneous, it is reasonable that ischemic placental disease may serve as an important pathway to preterm birth.

Despite the recent increase in preterm birth in the United States, trends in preterm birth subtypes have not been adequately examined. We examined trends in preterm birth among singletons following ruptured membranes, medical indications, and spontaneous preterm birth and evaluated the impact of these trends on perinatal mortality. A population-based, retrospective cohort study comprising 46,375,578 women (16% blacks) who delivered singleton births in the United States, 1989 through 2000, was performed. Rates of preterm birth ( /= 22 weeks plus neonatal deaths within 28 days), before and after adjustment for potential confounders, were derived from ecological logistic regression models. Preterm birth rates increased by 14% (95% confidence interval 13-15%) among whites from 8.3% to 9.4% and decreased by 15% (95% confidence interval 14-16%) among blacks from 18.5% to 16.2% between 1989 and 2000. Among whites, preterm birth following ruptured membranes declined by 23%, medically indicated preterm birth increased by 55%, and spontaneous preterm birth increased by 3%. Among blacks, preterm birth following ruptured membranes declined by 37%, medically indicated preterm birth increased by 32%, and spontaneous preterm birth decreased by 27%. The largest decline in perinatal mortality among whites was associated with increases in medically indicated preterm birth, whereas the largest decline in perinatal mortality among blacks was associated with declines in preterm birth following ruptured membranes and spontaneous preterm birth. Temporal trends in preterm birth varied substantially based on underlying subtype and maternal race. The recent increase in medically indicated preterm birth was associated with a favorable reduction in perinatal mortality.