Ambient air pollution exposure, plasma metabolomic markers, and risk of type 2 diabetes: A prospective cohort study

Both air pollution (AP) and impaired lipid metabolism contribute to type 2 diabetes (T2D). However, little is known about the detailed associations of AP to lipidomic markers and the specific lipid metabolomic profile that mediates the impact of AP on incident T2D. We aimed to examine the associations between long-term AP exposure, plasma metabolomic markers, and incident T2D, and subsequently determine the lipid metabolomic profile that mediates the relationship between AP and T2D. This prospective study included 82,548 participants from the UK Biobank without a history of T2D at baseline. Baseline plasma samples were analyzed using the nuclear magnetic resonance (NMR) metabolomic platform, which measured 168 metabolomic markers, including lipids, lipoprotein subclasses, and other circulating metabolites. Land Use Regression models were utilized to estimate annual average concentrations of PM2.5 and NO2. The associations among AP, metabolomic markers, and T2D were investigated using multivariable linear regressions and Cox proportional hazards models. Mediation analyses were performed to assess the role of each metabolomic marker in the AP-T2D relationship. Furthermore, principal component (PC) analysis was conducted on 168 metabolomic markers to extract metabolic patterns. These patterns were utilized to determine their associations with AP and T2D, as well as their mediating role in the AP-T2D relationship. We found that long-term AP exposure was associated with some lipid metabolites, including ApoA1, HDL concentration, HDL size, and lipid components within HDL, especially in very large, large, and medium HDL, as well as some other lipids, fatty acids, amino acids, glucose, and glycoprotein acetyls. In pairwise mediation analysis, these metabolites exhibited significant mediation effects in the AP-T2D relationship. We identified six PCS representing distinct metabolic patterns. Long-term exposure to PM2.5 and NO2 showed significantly negative associations with PC2 (characterized by high levels of ApoA1, larger HDL, other lipids, and low levels of larger VLDL). PC2 mediated 12.3% and 10.3% of the associations of PM2.5 and NO2 with incident T2D, respectively. This study revealed the associations of AP with various lipid metabolites. A lipid metabolomic profile characterized by ApoA1 and larger HDL may mediate the association between AP and incident T2D.