Abstract
OBJECTIVE: To investigate whether systemic inflammation, quantified by the Aggregate Index of Systemic Inflammation (AISI), mediates the association between uric acid (UA) and diabetic kidney disease (DKD). STUDY DESIGN: We analyzed data from 1716 adults with diabetes in NHANES 1999-2018. We used regression to assess UA-AISI-DKD associations and mediation analysis to quantify AISI's indirect effect. A random forest model, interpreted via SHAP, predicted DKD risk. RESULTS: Each 1 mg/dL increase in UA was associated with a 14 % higher DKD risk (adjusted OR = 1.14, 95 % CI: 1.04-1.26). UA was positively associated with AISI (β = 0.0356, p = 0.0058), which in turn predicted DKD (OR = 1.25 per SD increase in ln-AISI, 95 % CI: 1.10-1.42). AISI partially mediates (10.94 %) the association between UA and DKD, indicating that systemic inflammation is one of several pathways linking hyperuricemia to renal injury. The random forest model performed best, with SHAP highlighting AISI as a key positive predictor. CONCLUSION: Systemic inflammation, as measured by AISI, partially mediates the cross-sectional association between serum uric acid and diabetic kidney disease, supporting inflammation as one of several contributing pathways. The predictive performance of models incorporating AISI remains modest and does not outperform conventional clinical risk scores.