Cluster-defined gout subtypes exhibit divergent chronic kidney disease trajectories: a machine learning approach in a Chinese prospective cohort.

OBJECTIVE

To systematically evaluate the contribution of clinical heterogeneity to chronic kidney disease (CKD) progression in gout patients using data-driven phenotyping, and to assess whether incorporating genetic risk improves prediction of renal outcomes.

METHODS

In this prospective cohort study, 1497 Chinese gout patients were enrolled and followed for CKD progression. K-means clustering was applied to four core clinical variables: serum urate (SU), fractional excretion of uric acid (FEUA) (i.e. high- and low-excretion), gout duration and kidney stone burden. The primary outcome was the incidence of CKD stage ≥3, assessed using Cox proportional hazards models. Genetic risk was evaluated using an unweighted genetic risk score derived from 20 single nucleotide polymorphisms associated with gout and hyperuricaemia.

RESULTS

Over a follow-up of 4166 person-years, 153 participants (10.22%) developed CKD stage ≥3. Five clinical clusters were identified: Cluster 1 (high-excretion, low-urate), Cluster 2 (low-excretion, low-urate), Cluster 3 (long-duration), Cluster 4 (low-excretion, high-urate) and Cluster 5 (nephrolithiasis). Cluster 4 and Cluster 5 were significantly associated with increased risk of CKD progression, with adjusted hazard ratios of 2.19 (95% CI 1.20-4.01) and 3.52 (95% CI 1.91-6.48), respectively. Genetic predisposition, via risk score, further amplified renal risk in Cluster 4. Achieving target SU levels and reductions in kidney stone burden were independently associated with a lower risk of CKD progression.

CONCLUSIONS

This study proposes a novel K-means clustering-based classification of gout, identifying subgroups with distinct CKD trajectories. Integration of clinical phenotyping and genetic profiling may enhance individualized risk stratification and guide targeted prevention strategies in gout-associated CKD.