Association of Estimated Long-term Exposure to Air Pollution and Traffic Proximity With a Marker for Coronary Atherosclerosis in a Nationwide Study in China

Key Points Question Are long-term exposure to ambient air pollution and proximity to traffic associated with subclinical atherosclerosis? Findings In this population-based cross-sectional study of 8867 Chinese participants, long-term exposure to ambient nitrogen dioxide and fine particulate matter with aerodynamic diameter less than 2.5 μm was independently associated with a higher coronary artery calcium score, a key atherosclerotic marker. Associations with ozone and proximity to traffic were less consistent. Meaning Long-term exposure to ambient air pollution may be an important risk factor for coronary atherosclerosis.


Study population
The Coronary Atherosclerosis Disease Early Identification and Risk Stratification by Noninvasive Imaging (CREATION, clinical trial registration number: NCT03518437) study is a large and newly developed cohort study with participants selected from outpatients nationwide primarily from north, southwest and southeast of China. This cohort was organized by Beijing Fuwai hospital affiliated to Chinese Academy of Medical Sciences and National Center for Cardiovascular Diseases of China, with major focus on identifying risk factors that lead to coronary artery atherosclerotic plaque progression and events of coronary artery diseases (CAD). Fuwai Hospital is the largest hospital in China, specializing cardiovascular diseases treatment, prevention and research. Nowadays, more than 6 million outpatients, from 32 provinces in the nation and 6 countries worldwide have received their cares.
The CREATION cohort recruited outpatients of adults (aged>25) that visit Fuwai hospital who are suspected of having CAD without history of coronary revascularization other heart disease and who underwent cardiac computed tomography between 2015 and 2017. Similar study design could be found elsewhere 1 . The participants had relatively low pretest probability of CHD (27.3±15.1%), with asymptomatic patients (n=2587), non-anginal chest pain (n=5721), atypical angina (n=449), typical angina (n=110), therefore, this cohort can represent the low risk natural population. Typical angina was defined as having all of the following criteria: (1) substernal chest pain or discomfort that was (2) provoked by exertion or emotional stress and (3) relieved by rest and/or nitroglycerine. Symptoms were classified as atypical angina when only two of these criteria were present, and as nonanginal chest pain if one or none of these characteristics were reported.
This study built air pollution and traffic exposures upon the CREATION cohort participants who had accurate geocoded addresses for exposure assessment at baseline. All the participants will be followed-up for CAD disease events collection and some of them will be re-examined for CT scans after 3 and 5 years of baseline survey. In addition to clinical information collected during the hospital visit and subsequent visits for follow-up treatment or physical examinations, an extensive set of cardiovascular biomarkers of was collected in this cohort. Furthermore, a subset of the participants completed detailed survey on environmental risk factors, including building environment, socio-economic status and residential mobility. The large cohort size and rich subclinical and biological database make it feasible to associate long-term air pollution and traffic exposures with outcomes that may define pathways by which air pollution contributes to the atherosclerotic cardiovascular disease. Detailed participants selection and spatial distributions of their residences exposed to air pollution were shown as below.

Clinical data
A structured interview was conducted before CT exam to collect information on demographic characteristics, the presence and duration of cardiovascular risk factors and biochemical results. Hypertension was defined as a previously established diagnosis, systolic blood pressure 140 mmHg, diastolic blood pressure 90 mmHg, or antihypertensive medication use. Hypercholesterolemia was defined according to the National Cholesterol Education Panel (NCEP) guidelines or by the current use of lipid-lowering medication. Diabetes mellitus was defined as a previously established diagnosis, insulin or oral hypoglycemic therapy. Current smoking was defined as any cigarette smoking within 1 year prior to the CT exam. Alcohol consumption was defined as drinking more than twice a week for more than a year. Meat consumption was defined as eating meat more than three times a week, more than 300 grams for each time. Physical activity was defined as frequency of aerobic exercise for more than an hour. Coronary heart disease was defined as at least 1 coronary segment with a lesion of ≥50% luminal stenosis in diameter. Any disagreements between two readers were solved by consensus.The duration of the risk factors was defined as the time between made a definite diagnosis and underwent CT. The results of the biochemical test were performed within 1 month of the CT exam.

Estimating historical PM2.5
To extrapolate PM2.5 concentrations back in time, we leveraged a large dataset with annual average satellite-derived PM2.5 estimates (10x10 km 2 ) from 2004 to 2014 2 . We converted our fine-scale point-based PM2.5 predictions in 2014 to the historical years using the ratio of the satellite-derived PM2.5 predictions between 2014 and the target year (see equation 1). This calculation is based on an assumption that, for each point-based prediction from our fine-scale model, the spatial contrast of PM2.5 is constant between the two years within a 10x10km grid cell, but varies between grid cells. We then calculated the long-term PM2.5 concentrations averaged back for 3-, 4-, 5-, 6-, and 10-year respectively, ahead of baseline since 2015, to assess the effect of different lengths of exposure on difference of CAC.