fetching data ... 
Background: Gout represents a substantial burden on healthcare systems worldwide. It is essential to address risk factors and implement timely interventions to prevent the progression from asymptomatic hyperuricemia (AHU) to gout.
Objectives: We aimed to develop a machine learning-based risk stratification model to predict gout onset in patients with AHU.
Methods: We conducted a retrospective cohort study using a Common Data Model database of Boramae Medical Center. Adults aged 18 years or older with at least one serum urate measurement exceeding 7.0 mg/dL or a diagnosis of AHU were identified between 2010 and 2022. Patients with a prior diagnosis of gout or a history of prescription of gout medications were excluded. The primary outcome was the diagnosis of gout. Machine learning models, including LightGBM, XGBoost, Random Forest, AdaBoost, Decision Tree, and Logistic Regression, were used to predict the occurrence of gout. Model performance was assessed using the area under the receiver operating characteristic curve (ROC-AUC) and the precision-recall AUC (PR-AUC). Feature importance and the Cox proportional hazards model were used to identify the relevant variables and to develop a risk-scoring model.
Results: Among the 8,139 patients with AHU, 223 (2.7%) were newly diagnosed with gout during the 10-year follow-up period. Eighty-three percent of the patients were men, most in the 30- to 40-year range. LightGBM performed best, with a ROC-AUC of 0.812 and a PR-AUC of 0.112. In the risk-scoring model, predictors for gout included serum urate levels, LDL cholesterol levels, carbon dioxide levels, ESR levels, age, chronic kidney disease, and hypertension. This scoring model yielded a C-index of 0.751.
Conclusion: We established a risk-scoring system for gout in patients with AHU using real-world data. This straightforward scoring model may help predict the progression to gout in those with AHU and guide interventions for high-risk patients.
REFERENCES: NIL.
A risk-scoring model for predicting the progression to gout in patients with asymptomatic hyperuricemia
| β coefficient | P value | Score | |
|---|---|---|---|
| Age, years | |||
| <35 | 0 | Ref | 0 |
| ≥35, <65 | 0.355 | 0.114 | 3 |
| ≥65 | 0.402 | 0.097 | 4 |
| Chronic kidney disease | |||
| No | 0 | Ref | 0 |
| Yes | 2.117 | 0.005 | 20 |
| Hypertension | |||
| Yes | 0 | Ref | 0 |
| No | 0.425 | 0.388 | 4 |
| Serum total CO2 levels, mmol/L | |||
| ≥23 | 0 | Ref | 0 |
| <23 | 0.638 | <0.001 | 6 |
| LDL cholesterol levels, mg/dL | |||
| <130 | 0 | Ref | 0 |
| ≥130 | 0.423 | 0.117 | 4 |
| ESR levels, mm/hr | |||
| <20 | 0 | Ref | 0 |
| ≥20 | 0.331 | 0.164 | 3 |
| Serum urate levels, mg/dL | |||
| <7 | 0 | Ref | 0 |
| ≥7, <8 | 0.174 | 0.641 | 2 |
| ≥8, <9 | 0.104 | 0.783 | 1 |
| ≥9, <10 | 0.262 | 0.504 | 3 |
| ≥10 | 0.716 | 0.071 | 7 |
CO2, carbon dioxide; ESR, erythrocyte sedimentation rate; LDL, low-density lipoprotein
Acknowledgements: NIL.
Disclosure of Interests: None declared.
© The Authors 2025. This abstract is an open access article published in Annals of Rheumatic Diseases under the CC BY-NC-ND license (