Do diabetes and hypertension mediate the association of air pollution with the incidence of end-stage kidney disease? Results from two large European cohorts

Air pollution has been associated with kidney disease. However, it is unknown to what extent diabetes and hypertension, themselves affected by air pollution, contribute to the association. We investigated whether the association between air pollution and end-stage kidney disease (ESKD) is mediated by diabetes or hypertension in two large cohorts, the Austrian VHM&PP and the Italian RoLS. Both cohorts were followed up for incident ESKD with dialysis and kidney transplant registries. Exposure to air pollution was assessed using annual mean concentrations at the residential address derived from Europe-wide land-use regression models. Information on diabetes and hypertension was based on measurements (VHM&PP), or hospital discharges and medications (RoLS). A four-way decomposition causal mediation analysis was conducted for pollutants associated with ESKD in the two cohorts, adjusting for individual and area-level confounders. A total of 501 and 3231 incident ESKD cases were observed with a median follow-up of 27 and 8 years in VHM&PP and RoLS, respectively. The total excess risks for associations with ESKD were 13.9% and 17.9% for NO₂, and 17.0% and 20.9% for black carbon (BC), for diabetes and hypertension, respectively. The excess risk mediated by diabetes and hypertension was 2.3% (95% CI: 0.8%, 3.9%) and 3.2% (1.9%, 4.4%) for NO₂ (increase 10 μg/m³), and 2.1% (1.0%, 3.1%) and 3.5% (2.3%, 4.7%) for BC (increase 0.5*10⁻⁵/m), respectively. No positive mediation was observed for PM_2.5. Diabetes and hypertension partly mediated the associations of NO₂ and BC with ESKD, whereas no consistent results were found for PM_2.5.