Local determinants of road traffic noise levels versus determinants of air pollution levels in a Mediterranean city

International audience; BACKGROUND: Both traffic-related noise and air pollution have been associated with cardiovascular disease (CVD). Spatial correlations between these environmental stressors may entail mutual confounding in epidemiological studies investigating their long-term effects. Few studies have investigated their correlation - none in Spain - and results differ among cities. OBJECTIVES: We assessed the contribution of urban land-use and traffic variables to the noise-air pollution correlation in Girona town, where an investigation of the chronic effects of air pollution and noise on CVD takes place (REGICOR-AIR). METHODOLOGY: Outdoor annual mean concentrations of nitrogen dioxide (NO(2)) derived from monthly passive sampler measurements were obtained at 83 residential locations. Long-term average traffic-related noise levels from a validated model were assigned to each residence. Linear regression models were fitted both for NO(2) and noise. RESULTS: The correlation between NO(2) and noise (L(24h)) was 0.62. However, the correlation differed across the urban space, with lower correlations at sites with higher traffic density and in the modern downtown. Traffic density, distance from the location to the sidewalk and building density nearby explained 35.6% and 73.2% of the variability of NO(2) and noise levels, respectively. The correlation between the residuals of the two models suggested the presence of other unmeasured common variables. CONCLUSIONS: The substantial correlation between traffic-related noise and NO(2), endorsed by common determinants, and the dependence of this correlation on complex local characteristics call for careful evaluations of both factors to ultimately assess their cardiovascular effects.