Association between socioeconomic status and PM₂.₅ exposure concentrations in Matlosana, South Africa

Abstract

South Africa has the highest absolute burden of HIV in the world, with 7.5 million people living with the virus—equating to a prevalence of 19% (Pillay-van Wyk et al., 2016). Despite the use of antiretroviral therapies, HIV-seropositive individuals remain at a heightened risk for cardiovascular disease compared to their seronegative peers (Paisible et al., 2015; Bavinger et al., 2013; Freiberg et al., 2013). Exposure to fine particulate matter (PM2.5) further compounds this risk, as it is a known driver of cardiopulmonary disease, a leading cause of mortality worldwide. Notably, air pollution is responsible for 20% of cardiovascular disease deaths (Yusuf et al., 2019). Given that household and ambient air pollution contribute significantly to overall PM2.5 exposure—accounting for 5.2% and 2.9% of all global deaths, respectively (Health Effects Institute, 2019)—it is crucial to assess their impact on vulnerable populations. This study is part of a larger, overarching project investigating whether HIV infection modifies the relationship between PM2.5 exposure and both systolic blood pressure and percent predicted FEV₁ (forced expiratory volume), a key indicator of lung function. Specifically, this paper seeks to determine whether individuals living with HIV and/or those experiencing socioeconomic disadvantage experience disproportionate exposure to PM2.5. To characterize this study population, we categorized sociodemographic variables, assessed fuel use behaviors, and developed a weighted asset index. We also examined the distribution of both household and personal PM2.5 exposure across various socioeconomic indicators, utilizing the weighted asset index and fuel use composite scores as proxies for household fuel behaviors and stratifying by HIV status. Additionally, we conducted linear univariate and multivariate models separately for log-transformed personal and household PM2.5 measurements to quantify the impact of each variable on geometric mean PM2.5 concentrations. No significant differences in personal or household 24-hour PM2.5 exposure were observed by HIV status nor education level, sex, employment status, total family income, living situation, subjective financial security, adult hunger, child hunger, weighted asset index category, or cooking fuel type. However, a notable association was found for lighting sources: households using both candles and kerosene for lighting exhibited a 370% (95% CI: 121%, 1130%) higher geometric mean for household PM2.5 concentrations compared to those relying solely on clean (electric) lighting. The multivariate models explained 37% of the variability in household PM2.5 concentrations and 34% of the variability in personal PM2.5 concentrations. Notably, this study categorizes this peri-urban community in Matlosana, providing critical insights into socioeconomic disparities that exist in this population. However, our findings suggest that household and personal PM2.5 exposures are not distributed differently among different socioeconomic conditions nor among HIV status. Future research should explore the broader influence of urbanization and ambient air pollution, which may play a more significant role in shaping long-term health outcomes.