Final Report Summary - CHAI (Cardiovascular Health effects of Air pollution in Andhra Pradesh, India) The CHAI Project aimed to evaluate the exposure to air pollution and its health effects in a peri-urban area in South India. The three objectives were to 1) characterize particle exposure of residents of 28 villages in peri-urban South India; 2) use objective measurement of location, time-activity patterns based on global positioning system (GPS) and wearable cameras, and continuous personal monitoring of particles, to identify specific activities and microenvironments important for exposure; and 3) quantify the association between particles exposure and vascular biomarkers, including surrogates of subclinical atherosclerosis, arterial stiffness, and global vascular injury. CHAI fully met these objectives and addressed several additional research questions. Based on a locally-developed land use regression model, annual mean (sd) fine particles (PM2.5) outdoors at residences in the study population was 34.1 (3.2) μg/m3 and 2.7 (0.5) μg/m3 for black carbon. The LUR model for annual black carbon explained 78% of total variance and included both local-scale (energy supply places) and regional-scale (roads) predictors. The model for PM2.5 explained 58% of the variance and predictors were only regional-scale (urbanicity, vegetation). Based on data from 2 × 24-h integrated personal exposure measurements (PM2.5 and black carbon in two seasons in 402 study participants), mean (sd) PM2.5 personal exposure was 55.1(82.8) μg/m3 for men and 58.5(58.8) μg/m3 for women; corresponding figures for black carbon were 4.6(7.0) μg/m3 and 6.1 (9.6) μg/m3. Most variability in personal exposure was within participant (intra-class correlation ~20%). Predictors of personal exposure included socioeconomic position for both genders; working in construction or industry in men; and use of biomass for cooking in women. Personal exposure measurements were not correlated with annual ambient concentration at residence modeled by land-use regression. There were large differences in daily mobility patterns and time activity by gender, which influenced particle exposure patterns. Using highly time-resolved activity data derived from wearable cameras, we identified several important drivers of personal PM2.5 exposure in the study population including active and passive smoking, cooking with biomass stoves, and being present in the kitchen. Using a data-driven approach combining multiple sources of information, we found that the activity data from the wearable cameras added considerable additional information that explained short term peaks in exposure beyond what self-reported questionnaire data, ambient air pollution, and GPS location could explain. In epidemiological analyses, both ambient and household air pollution were positively associated with carotid intima-media thickness (CIMT), a surrogate marker for atherosclerosis. We identified positive associations between ambient PM2.5 at residence and CIMT, particularly among men: a 3% (95%CI 0.2 5.7%) increase in CIMT per 1 µg/m3 increase in PM2.5. Use of biomass cooking fuel was associated with CIMT in all participants, and especially among women with an unvented stove: 6.1% (95% CI 1.4-10.9%) higher CIMT compared to women not using biomass fuel. We observed that women had higher CIMT compared with men, which is not usually reported in western countries; this difference might be due to cumulative exposure to household air pollution generated by biomass fuel use. Personal predicted PM2.5 covered a wide range (13-85 µg/m3 for men, 40-120 µg/m3 for women). Predicted personal PM2.5 was positively associated with CIMT, arterial stiffness, and vascular injury among men and vascular injury in women. In additional analyses we identified a positive association between PM2.5 and blood pressure among women and no association between particle exposure and blood glucose or prevalent diabetes.
