Machine Learning Insights Into Uric Acid Elevation With Thiazide Therapy Commencement and Intensification

Area of Science:

• Cardiovascular Medicine
• Pharmacology
• Biostatistics

Background:

• Elevated serum uric acid is linked to cardiovascular diseases, including atherosclerotic heart disease, hypertension, and heart failure.
• Thiazide or thiazide-like drugs (THZ), commonly used for hypertension, can increase uric acid levels.
• Understanding factors influencing THZ-related uric acid elevation is vital for optimizing patient care.

Purpose of the Study:

• To identify clinical determinants associated with elevated serum uric acid levels in patients treated with thiazide or thiazide-like drugs (THZ).
• To evaluate the efficacy of machine learning algorithms in pinpointing these clinical influencers.

Main Methods:

• Retrospective cross-sectional study of 143 patients initiating or escalating THZ treatment.
• Collected baseline and control uric acid levels, biochemical data, and clinical information.
• Applied Random Forest, Neural Network, Support Vector Machine, and Gradient Boosting regression models for feature selection and analysis.

Main Results:

• Key determinants identified include uncontrolled diabetes, estimated Glomerular Filtration Rate (eGFR), absence of insulin, indapamide use, and absence of statin treatment.
• Other notable factors were absence of Sodium-glucose cotransporter 2 inhibitors (SGLT2i), low-dose aspirin use, and older age.
• Gradient Boosting regression model demonstrated superior performance with the highest R2 value (0.779).

Conclusions:

• Machine learning algorithms accurately identify factors influencing THZ-related uric acid fluctuations.
• This capability supports personalized treatment strategies, potentially reducing unnecessary avoidance of THZ.
• Findings provide guidance for managing THZ therapy to prevent problematic uric acid elevations.