BackgroundFew studies have examined the role of long-term (≥1 year) ambient temperature with quantitative traits of early-stage cardiovascular disease (CVD) such as carotid intima-medial thickness (cIMT). Our objective was to examine associations between temperature and cIMT, a measure of subclinical atherosclerosis. MethodsThis study examined data from 3257 participants in the Coronary Artery Risk Development in Young Adults (CARDIA) study, aged 18–30 years at baseline (1985–1986). We used North America Land Data Assimilation System data to derive 12 metrics of ambient daily temperature: Mean, minimum, maximum, and standard deviation temperature in summer, winter, and year-round. We examined associations with cIMT in separate cross-sectional multivariable models at CARDIA year 20 (2005–2006) as well as stratified analyses by self-reported race and sex. We also prospectively examined cumulative temperature by summing temperature variables from Y0-Y20. ResultsAccounting for study center attenuated most associations between cIMT and ambient temperature exposure, but the winter standard deviation remained associated (overall β = −0.0104 mm/°C, 95 % CI: −0.0150 to −0.0059). Minimum summer temperature was also associated with cIMT in the overall study population (β = 0.0020 mm/°C, 95 % CI: 0.0005–0.0035). Associations did not differ substantially by race, but women had stronger associations than men. Cumulative temperature was not associated with cIMT. ConclusionsOur findings suggest a role of geography, particularly ambient temperature in cIMT. Future research to address potential residual confounding is necessary, but if validated these findings have implications for policy and strategies to mitigate health impacts of climate change.