BACKGROUND: Climate change is expected to have adverse health effects, but the association between ambient temperatures and stillbirth is unclear. OBJECTIVES: We investigated acute and chronic associations between extreme ambient temperatures and stillbirth risk, and estimated the attributable risk associated with local temperature extremes in the United States. METHODS: We linked 223,375 singleton births >= 23 weeks of gestation (2002-2008) from 12 U.S. sites to local temperature. Chronic exposure to hot (>90th percentile), cold (<10th percentile), or mild (10th-90th percentile) temperatures was defined using window- and site-specific temperature distributions for three-months preconception, first and second trimester, and whole-pregnancy averages. For acute exposure, average temperature for the week preceding delivery was compared to two alternative control weeks in a case-crossover analysis. RESULTS: In comparison with mild, whole-pregnancy exposure to cold [adjusted odds ratio (aOR) = 4.75; 95% confidence interval (CI): 3.95, 5.71] and hot (aOR = 3.71; 95% CI: 3.07, 4.47) were associated with stillbirth risk, and preconception and first and second trimester exposures were not. Approximately 17-19% of stillbirth cases were potentially attributable to chronic whole-pregnancy exposures to local temperature extremes. This is equivalent to similar to 1,116 cold-related and similar to 1,019 hot-related excess cases in the United States annually. In the case-crossover analysis, a 1 degrees C increase during the week preceding delivery was associated with a 6% (3-9%) increase in stillbirth risk during the warm season (May September). This incidence translates to similar to 4 (2-6) additional stillbirths per 10,000 births for each 1 degrees C increase. CONCLUSIONS: Extremes of local ambient temperature may have chronic and acute effects on stillbirth risk, even in temperate zones. Temperature related effects on pregnancy outcomes merit additional investigation.
