Table 2 Articles examining interactions between air pollution and socioeconomic position on cardiovascular endpoints identified through systematic review, 2005–2016
First Author | Design | Population | Air pollutant exposure(s) | Material resource(s) or psychosocial stress measure(s) | Outcome(s) | Result(s) |
|---|---|---|---|---|---|---|
Barceló, 2009 [49] | Ecological | Residents of Barcelona, Spain | TSP, PM10, NO2, CO, SO2 | Census-tract deprivation index: unemployment, lower educational level, manual workers, temporary workers | Ischemic heart disease mortality | A positive interaction between pollutants and the deprivation index was statistically significant for NO2 and ischemic disease mortality in men. |
Bravo, 2016 [30] | Case-crossover | Residents of Sau Paulo, Brazil | PM10, NO2, SO2, CO, O3 | Individual education and area-level SEP index | CVD mortality | Significant positive interaction between pollutants and individual education. Significant inverse interaction between pollutants and SEP index. |
Chi, 2016 [40] | Prospective cohort study | Women’s Health Initiative participants from 40 US sites | PM2.5 | Individual education, family income and occupation. Area-level education, occupation, family income, poverty status, median home value, neighborhood SEP score | CVD event (including MI, stroke, CVD death, cerebrovascular death) | Statistically significant effect modification by neighborhood SEP score. Non-significant higher effect for those with lowest individual income and occupation. |
Chiusolo, 2011 [44] | Case-crossover | Adults from 10 Italian cities | NO2 | Census block group median income and median SEP indicator | Cause-specific mortality | Neither income nor SEP significantly modified the association between NO2 and mortality. Significant heterogeneity in the stratum-specific estimates among the cities. |
Dragano, 2009 [22] | Cross-sectional | Adults from 3 German cities | Roadway proximity, traffic volume | Individual education and income; Neighborhood unemployment | Coronary artery calcification | Statistically significant effect modification of main effect by education and unemployment among men and modification by income among women. |
Finkelstein, 2005 [50] | Prospective cohort | Adults from Hamilton and Burlington, Ontario, Canada | Roadway proximity, TSP and SO2 | Census tract-level deprivation index: income, education and unemployment | Circulatory disease mortality | Non-significant effect modification by neighborhood deprivation index evident in high traffic areas. |
Haley, 2009 [45] | Case-crossover | Residents of New York State with CVD discharge diagnosis | PM2.5 | Census tract percentage of adults living below poverty level | CVD hospitalizations | No effect modification |
Henderson, 2011 [46] | Repeated measures | Canadian population in the southeast corner of British Columbia | PM10, smoke | Census tract income quintiles | CVD physician visits and hospitalizations | No main effects of exposures on CVD outcomes (with 2 exceptions). No effect modification |
Hicken, 2013 [31] | Cross-sectional | Multi-Ethnic Study of Atherosclerosis (MESA) cohort from 6 U.S. cities | PM2.5 | Material Resources: Individual education and income and census tract median household income. Stress: Individual chronic stress, depressive symptoms, trait anger, trait anxiety, lack of emotional support | Blood pressure | Non-significant modification showing higher effects among higher education groups and no effect modification by income. No effect modification by stress indicators. |
Hicken, 2014 [53] | Repeatedmeasures | Adults in Detroit | PM2.5 | Stress: Individual environmental stress index, life events index | Blood pressure | Higher effect of PM2.5 on blood pressure in people living in Southwest Detroit under high stress. |
Hicken, 2016 [48] | Cross-sectional | MESA cohort, 6 U.S. cities | PM2.5, NOx | Material Resources: Individual SEP index and census tract racial segregation. Stress: Individual psychosocial adversity | Left ventricular mass index (LVMI), Left ventricular ejection fraction (LVEF) | No effect modification |
Kan, 2008 [23] | Time series | Residents of Shanghai, China | PM10, SO2, NO2, and O3 | Individual education | CVD mortality | Non-significant interaction shows that residents with lower education had an increased risk of CVD mortality compared to those with higher education for all pollutants except O3. |
Malig, 2009 [24] | Case-crossover | Residents of 15 California counties | Coarse PM | Individual education | Total and CVD mortality | Significant interaction showing that the effect of coarse PM on CVD mortality was higher in those of lower education. |
McGuinn, 2016 [42] | Retrospective cohort | CATHGEN Cohort in North Carolina | PM2.5 | Census block group education, urban/rural | CAD index >23 and MI in the previous year | No effect modification |
Medina-Ramon, 2008 [32] | Case only | Residents of 48 U.S. cities | O3 | Individual education | CVD mortality | No effect modification |
Ostro, 2008 [26] | Time series | Residents of California | PM2.5 | Individual education | CVD mortality | Statistically significant interaction with lower education increasing the effect of PM2.5 and its components. |
Ostro, 2014 [25] | Longitudinal cohort | Study of Women’s Health Across the Nation (SWAN) cohort | PM2.5 | Individual education, income, marital status | Continuous CRP; CRP > 3 mg/L; CRP >3 mg/L in high age group | Statistically significant effect modification by income and non-significant effect modification by education. |
Qiu, 2015 [47] | Case only | Residents of Hong Kong who died of circulatory/respiratory system diseases | PM10, SO2, NO2, O3 | Individual employment status | CVD mortality | Significant interaction in that the unemployed were more susceptible to pollution associated mortality for all pollutants except O3. |
Raaschou-Nielsen, 2012 [33] | Prospective cohort | Diet, Cancer and Health study participants in Denmark | NO2 | Individual education | Mortality due to ischemic heart disease, cardiac rhythm, heart failure, cerebrovascular and other CVD causes | No effect modification |
Ren, 2010 [34] | Case-crossover | Population of Eastern Massachusetts | O3 | Individual education and census tract income and poverty | CVD mortality | No effect modification |
Rosenlund, 2008 [52] | Retrospective cohort | Residents of Rome, Italy | NO2 | Census block group deprivation index | Coronary heart disease mortality and hospitalizations | No effect modification |
Rosenlund, 2009 [29] | Case-control | Residents of Stockholm County, Sweden | NO2, PM10 | Individual occupation, education, income and marital status | Fatal and non-fatal MI | Higher effects for low white collar workers and higher income, but no statistically significant effect modification |
Son, 2012 [35] | Case-crossover | Residents of Seoul, Korea | PM, NO2, SO2, CO, O3 | Individual education, marital status and occupation | CVD mortality | Greater effects for lower education as well as manual occupation and unknown occupation. |
Stafoggia, 2014 [27] | Prospective cohort | European Study of Cohorts for Air Pollution Effects (ESCAPE) multi-city participants | PM2.5 | Individual education and rural/urban residence | Incident stroke | Nonsignificant effect modification by education where the lowest education had highest effect. No effect modification by urban/rural residence. |
Wilson, 2007 [41] | Ecological | Residents of central, middle and outer Phoenix, Nevada | PM2.5 and PM10 | Zip code-level income and education | CVD mortality | Lower SEP population may be more susceptible to PM associated mortality, but it is difficult to separate spatial effect. |
Winquist, 2012 [43] | Time series | Hospital patients in greater St. Louis MSA | PM2.5 and O3 | Zip code-level poverty | Emergency department visits and hospital admissions for CVD conditions | Higher effect of poverty on O3-CVD, all outcomes. Also, poverty on O3-CHD, all outcomes. Possible, non-sig differences of poverty on PM2.5-CHF relationship |
Wong, 2008 [51] | Time series | Residents of Hong Kong, China | PM10, SO2, NO2 | Community planning unit social deprivation index | CVD mortality and hospitalizations | Higher mortality from exposure to SO2 and NO2 for areas with high deprivation index. |
Zeka, 2006 [28] | Case-crossover | Residents of 20 U.S. cities | PM10 | Individual education | CVD mortality | Statistically significant effect modification by education whereby there was a higher PM10-associated risk comparing lower to higher education. |
Zhang, 2011 [36] | Retrospective cohort | Residents of selected communities in Shenyang, China | PM10, SO2, NO2 | Individual education, income and marital status | CVD and cerebrovascular mortality | No effect modification |
Zhou, 2014 [37] | Prospective cohort | Adult men from 25 cities in China | TSP (1990–2000), PM10 (2000–2006) | Individual education | CVD mortality | No effect modification |