Abstract
BACKGROUND: Associations between ambient temperature and cardiovascular mortality are well established. This study investigated whether inflammation could be part of the mechanism leading to temperature-related cardiovascular deaths. METHODS: The study population consisted of a cohort of 673 men with mean age of 74.6 years, living in the greater Boston area. They were seen for examination roughly every 4 years, and blood samples for inflammation marker analyses were drawn in 2000-2008 (total of 1254 visits). We used a mixed effects model to estimate the associations between ambient temperature and a variety of inflammation markers (C-reactive protein, white blood cell count, soluble Vascular Cell Adhesion Molecule-1, soluble Intercellular Adhesion Molecule-1, tumor necrosis factor alpha, and interleukins -1beta, -6 and -8). Random intercept for each subject and several possible confounders, including combustion-related air pollution and ozone, were used in the models. RESULTS: We found a 0 to 1 day lagged and up to 4 weeks cumulative responses in C-reactive protein in association with temperature. We observed a 24.9% increase [95% Confidence interval (CI): 7.36, 45.2] in C-reactive protein for a 5 degrees C decrease in the 4 weeks' moving average of temperature. We observed similar associations also between temperature and soluble Intercellular Adhesion Molecule-1 (4.52%, 95% CI: 1.05, 8.10, over 4 weeks' moving average), and between temperature and soluble Vascular Cell Adhesion Molecule-1 (6.60%, 95% CI: 1.31, 12.2 over 4 weeks' moving average). Penalized spline models showed no deviation from linearity. There were no associations between temperature and other inflammation markers. CONCLUSIONS: Cumulative exposure to decreased temperature is associated with an increase in inflammation marker levels among elderly men. This suggests that inflammation markers are part of intermediate processes, which may lead to cold-, but not heat-, related cardiovascular deaths.