An Artificial Intelligence-Based Gout Management System Reduced Chronic Kidney Disease Incident and Improved Target Serum Urate Achievement

Abstract

Abstract Objectives Stage ≥ 3 chronic kidney disease (CKD) affects ∼25% of people with gout. The effects of urate-lowering therapy (ULT) on CKD incidence and progression have remained inconclusive. Here, we assessed the impact of a gout urate-lowering therapy (ULT) clinic intervention using artificial intelligence (AI) on CKD incidence and achievement of serum urate target. Methods An observational study compared propensity score-matched gout cohorts from an AI-based Gout Intelligent Management System (GIMS) and a standard Electronic Medical Records system (EMRS) clinic database. The GIMS included a mobile application, data fusion interface, and modules for expert consultation and laboratory results management. All patients had gout and a starting eGFR > 60 ml/min. Using a 1:2 propensity score (PS)-matched cohort study design we assessed the impact of the AI-based system on CKD outcomes and ULT effectiveness over 4 years of follow-up. Results Compared with EMRS, GIMS was associated with reduced incidence of CKD stage ≥ 3. Specifically, 169/4117 new onset CKD stage ≥ 3 (incidence 4.1 per 100 person-years) with GIMS compared with 164/2128 with EMRS (incidence 7.7 per 100 person-years) during follow-up. More participants achieved serum urate <6.0 mg/dl with GIMS vs EMRS during follow-up (49.8% vs 25.9%, p < 0.001). Conclusion Application of the AI-based GIMS was associated with lower incidence of CKD stage ≥ 3 and superior target serum urate achievement in people with gout. The AI-based GIMS represents a novel approach to improve real-world renal outcomes and ULT success in gout.