Levels of excess infant deaths attributable to maternal smoking during pregnancy in the United States.

OBJECTIVES: The objectives of the study were: 1) To determine the risk of infant mortality associated with prenatal cigarette smoking; 2) To assess whether the relationship, if existent, was dose-dependent; 3) To explore the morbidity pathway that explains the effect of tobacco smoke on infant mortality, and 4) to compute excess infant deaths attributable to maternal smoking in the United States. METHODS: Retrospective cohort study on 3,004,616 singleton live births that occurred in 1997 in the United States using the US national linked birth/infant death data. Excess infant deaths due to maternal smoking were computed using the population-attributable risk (PAR). RESULTS: Overall, 13.2% of pregnant women who delivered live births in 1997 smoked during pregnancy. The rate of infant mortality was 40% higher in this group as compared to nonsmoking gravidas (P < 0.0001). This risk increased with the amount of cigarettes consumed prenatally in a dose-dependent fashion (p for trend < 0.0001). Small-for-gestational age rather than preterm birth is the main mechanism through which smoking causes excess infant mortality. We estimated that about 5% of infant deaths in the United States were attributable to maternal smoking while pregnant, with variations by race/ethnicity. The proportion of infant deaths attributable to maternal smoking was highest among American Indians at 13%, almost three times the national average. If pregnant smokers were to halt tobacco use a total of 986 infant deaths would be averted annually. CONCLUSIONS: Smoking during pregnancy accounts for a sizeable number of infant deaths in the United States. This highlights the need for infusion of more resources into existing smoking cessation campaigns in order to achieve higher quit rates, and substantially diminish current levels of smoking-associated infant deaths.