Indoor and outdoor air pollution exposure in relation to incident symptomatic knee osteoarthritis: a 10-year longitudinal cohort study in China

Currently, the evidence regarding the association between air pollution and osteoarthritis is mainly limited to cross-sectional studies. This study aims to investigate the longitudinal impact of indoor and outdoor air pollution on the risk of symptomatic knee osteoarthritis (sKOA) in a nationally representative Chinese cohort.

Baseline characteristics were described for 17,096 CHARLS participants aged ≥ 45 years with complete baseline data, among whom 1,590 had prevalent sKOA at Wave 1. After excluding these prevalent cases and 994 participants without follow-up data, 14,512 participants were included in the longitudinal analysis of incident sKOA, with follow-up from 2011 to 2020. Indoor pollution was assessed based on the use of solid fuels for cooking and heating. Outdoor pollutants (PM₁, PM₂.₅, PM₁₀, NO₂, and O₃) data were obtained from the CHAPs Dataset. Cox proportional hazards model, GBTM and DLNM were used to evaluate the associations, and spatial autocorrelation analysis was employed to map the geographical clustering of cases.

During the 9-year follow-up period, a total of 698 new cases of sKOA were identified. Cooking with solid fuels increased the risk of sKOA by 36% (HR = 1.36, 95% CI: 1.05–1.77, P = 0.021); using both cooking and heating simultaneously further increased the risk to 46% (HR = 1.46, 95% CI: 1.07–2.00, P = 0.017). The crude incidence rates of sKOA were 5.18, 6.07, and 6.61 per 1,000 person-years among participants exposed to zero, one, and two sources of indoor pollution, respectively, indicating a graded dose-response relationship. Trajectory analysis showed that compared with the High-Decreasing trajectory group, the Low-Stable PM₂.₅ trajectory group faced a higher risk of sKOA (HR = 1.66, 95% CI: 1.20–2.30, P = 0.002), and similar patterns were observed for PM₁ and NO₂. DLNM analysis indicated that the effect of PM₂.₅ emerged approximately one year after exposure, and reached its peak at a concentration of 40–60 µg/m³ (HR ≈ 2.4). Additionally, the cases showed significant geographical clustering (Moran’s I = 0.55, P < 0.001). The proportional hazards assumption was evaluated using Schoenfeld residuals; no violation was detected for any covariate (all individual p > 0.05), and the global test confirmed overall model adequacy (χ² = 13.80, df = 16, p = 0.614).

Both indoor solid fuel combustion and outdoor air pollution were associated with increased risk of sKOA, but the pathways of their effects are different. Indoor exposure appears to follow a cumulative dose–response pattern, while outdoor pollution exhibits a time-dependent effect with a lag. Switching to clean household energy and reducing environmental pollution may contribute to the prevention of knee osteoarthritis.