Green spaces, excess weight and obesity in Spain

Background: The epidemiological evidence on green spaces and obesity is inconsistent. Objectives: To study the association of access to green spaces and surrounding greenness with obesity in Spain. Methods: We enrolled 2354 individuals 20-85 years from urban areas of seven provinces of Spain between 2008-13. Subjects were randomly selected population controls of the MCC-Spain case-control study. We geo-coded current residences and defined exposures in a buffer of 300 m around them: i) access to green space, identified using Urban Atlas, and ii) levels of surrounding greenness, measured by the Normalized Difference Vegetation Index. We examined excess weight/obesity as binary outcomes based on body mass index and waist-hip ratio. We examined effect modification by genetic factors, sex and individual socio-economic status and mediation by physical activity and concentrations of PM2.5 and NO2. To assess potential effect modification by genetic factors, we used a polygenic risk score based on obesity polymorphisms detected in genome-wide association studies. We used logistic mixed-effects models with a random effect for catchment area adjusted for potential confounders. Results: Access to green space was associated with a reduced risk of excess weight/obesity after adjusting for confounders [excess weight: OR (95%CI) = 0.82 (0.63, 1.07), p-value = 0.143; abdominal obesity: OR (95%CI) = 0.68 (0.45, 1.01), p-value = 0.057]. In the stratified analysis, this association was only observed in women. Associations between surrounding greenness and excess weight/obesity were null or modest based on a 1 IQR increase in NDVI [excess weight: OR (95%CI) = 0.99 (0.88, 1.11), p-value = 0.875; abdominal obesity: OR (95%CI) = 0.91 (0.79, 1.05), p-value = 0.186]. The observed associations were not mediated by physical activity or air pollution. Discussion: Access to green space may be associated with decreased risk of excess weight/obesity among women in Spain. Mechanisms explaining this association remain unclear.