Highway proximity associated with cardiovascular disease risk: the influence of individual-level confounders and exposure misclassification

Environmental Health

Research

Jonathan I Levy¹  Kevin Lane¹  Wig Zamore²  Andrea Stewart³  Allison P Patton⁴  Luz T Padró-Martínez⁴  Ding Ding Wang⁵  Kyle Hoesterey⁵  David Weiss⁵  Doug Brugge⁵  Mkaya Mwamburi⁵ 

[1] Boston University School of Public Health, Boston, MA, USA;Somerville Transportation Equity Partnership, Somerville, MA, USA;Tufts University School of Arts and Sciences, Medford, MA, USA;Tufts University School of Engineering, Medford, MA, USA;Tufts University School of Medicine, Boston, MA, USA;

关键词: Highway proximity;    Air pollution;    Traffic;    Geocoding;    Inflammation;   

DOI  :  10.1186/1476-069X-12-84

 received in 2013-01-31, accepted in 2013-09-30,  发布年份 2013

来源: Springer

【 摘 要 】

BackgroundElevated cardiovascular disease risk has been reported with proximity to highways or busy roadways, but proximity measures can be challenging to interpret given potential confounders and exposure error.MethodsWe conducted a cross sectional analysis of plasma levels of C-Reactive Protein (hsCRP), Interleukin-6 (IL-6), Tumor Necrosis Factor alpha receptor II (TNF-RII) and fibrinogen with distance of residence to a highway in and around Boston, Massachusetts. Distance was assigned using ortho-photo corrected parcel matching, as well as less precise approaches such as simple parcel matching and geocoding addresses to street networks. We used a combined random and convenience sample of 260 adults >40 years old. We screened a large number of individual-level variables including some infrequently collected for assessment of highway proximity, and included a subset in our final regression models. We monitored ultrafine particle (UFP) levels in the study areas to help interpret proximity measures.ResultsUsing the orthophoto corrected geocoding, in a fully adjusted model, hsCRP and IL-6 differed by distance category relative to urban background: 43% (-16%,141%) and 49% (6%,110%) increase for 0-50 m; 7% (-39%,45%) and 41% (6%,86%) for 50-150 m; 54% (-2%,142%) and 18% (-11%,57%) for 150-250 m, and 49% (-4%, 131%) and 42% (6%, 89%) for 250-450 m. There was little evidence for association for TNF-RII or fibrinogen. Ortho-photo corrected geocoding resulted in stronger associations than traditional methods which introduced differential misclassification. Restricted analysis found the effect of proximity on biomarkers was mostly downwind from the highway or upwind where there was considerable local street traffic, consistent with patterns of monitored UFP levels.ConclusionWe found associations between highway proximity and both hsCRP and IL-6, with non-monotonic patterns explained partly by individual-level factors and differences between proximity and UFP concentrations. Our analyses emphasize the importance of controlling for the risk of differential exposure misclassification from geocoding error.

【 授权许可】

CC BY   
© Brugge et al.; licensee BioMed Central Ltd. 2013