Abstract
BACKGROUND: Evidence of associations between daily variation in air pollution and blood pressure (BP) is varied and few prior longitudinal studies adjusted for calendar time. METHODS: We studied 143,658 postmenopausal women 50 to 79 years of age from the Women's Health Initiative (1993-2005). We estimated daily atmospheric particulate matter (PM) (in three size fractions: PM2.5, PM2.5-10, and PM10) and nitrogen dioxide (NO2) concentrations at participants' residential addresses using validated lognormal kriging models. We used linear mixed-effects models to estimate the association between air pollution concentrations and repeated measures of systolic and diastolic BP (SBP, DBP) adjusting for confounders and calendar time. RESULTS: Short-term PM2.5 and NO2 were each positively associated with DBP {0.10 mmHg [95% confidence interval (CI): 0.04, 0.15]; 0.13 mmHg (95% CI: 0.09, 0.18), respectively} for interquartile range changes in lag 3-5 day PM2.5 and NO2. Short-term NO2 was negatively associated with SBP [-0.21 mmHg (95%CI: -0.30, -0.13)]. In two-pollutant models, the NO2-DBP association was slightly stronger, but for PM2.5 was attenuated to null, compared with single-pollutant models. Associations between short-term NO2 and DBP were more pronounced among those with higher body mass index, lower neighborhood socioeconomic position, and diabetes. When long-term (annual) and lag 3-5 day PM2.5 were in the same model, associations with long-term PM2.5 were stronger than for lag 3-5 day. CONCLUSIONS: We observed that short-term PM2.5 and NO2 levels were associated with increased DBP, although two-pollutant model results suggest NO2 was more likely responsible for observed associations. Long-term PM2.5 effects were larger than short-term.