Clustering Analysis of Ankle-Brachial Index Related Metabolic and Body Composition Profiles Using K-Means Approach

Area of Science:

• Cardiovascular Health
• Metabolic Syndrome
• Biostatistics

Background:

• The Ankle-brachial Index (ABI) is crucial for diagnosing peripheral perfusion disorders.
• Increasing clinical parameters in ABI assessment can complicate decision-making.
• Identifying distinct patient profiles based on metabolic and body composition data is needed.

Purpose of the Study:

• To explore clustering of metabolic and body composition profiles in relation to ABI values.
• To identify patterns associated with peripheral perfusion using a K-means approach.
• To simplify complex ABI assessment through data-driven patient stratification.

Main Methods:

• A cross-sectional study analyzed data from 1093 participants.
• Thirteen clinical and metabolic parameters were assessed, including body composition indices.
• K-means clustering, with and without feature selection, was applied to identify distinct patient groups.

Main Results:

• The optimal clustering solution was found at k=2.
• Feature selection reduced parameters to 7, improving cluster quality (Silhouette Score increased from 0.503 to 0.559).
• Key parameters included visceral fat, resting metabolism, BMI, body fat, skeletal muscle, subcutaneous fat, and cell age, defining distinct metabolic risk profiles.

Conclusions:

• K-means clustering successfully identified distinct metabolic and body composition profiles associated with ABI.
• These clusters represent descriptive associations, not diagnostic categories.
• Findings suggest metabolic and body composition parameters can stratify ABI into distinct profiles, potentially aiding nursing practice.