Growth Differentiation Factor 15 Predicts Cardiovascular Events in Peripheral Artery Disease

Area of Science:

• Cardiovascular Medicine
• Biomarker Discovery
• Medical Technology

Background:

• Peripheral artery disease (PAD) significantly increases the risk of major adverse cardiovascular events (MACE), yet reliable prognostic biomarkers are scarce.
• Growth differentiation factor 15 (GDF15), a cytokine involved in inflammation and atherosclerosis, has potential but is underexplored in PAD risk stratification.

Purpose of the Study:

• To evaluate the prognostic utility of GDF15 for predicting 2-year MACE in patients with PAD.
• To assess the performance of an explainable machine learning model incorporating GDF15 for MACE prediction.

Main Methods:

• Prospective analysis of 1192 individuals (454 with PAD), measuring plasma GDF15 levels at baseline.
• Development and validation of an extreme gradient boosting (XGBoost) machine learning model to predict 2-year MACE.
• Utilized Shapley additive explanations (SHAP) for model interpretability and identification of key predictive features.

Main Results:

• Median plasma GDF15 levels were significantly higher in PAD patients compared to non-PAD controls (p < 0.001).
• The XGBoost model achieved strong predictive performance for 2-year MACE (F1 score = 0.83, recall = 83.7%).
• SHAP analysis identified GDF15 as the most influential predictor of MACE, outperforming traditional clinical factors.

Conclusions:

• GDF15 is a potent prognostic biomarker for predicting 2-year MACE in patients with PAD.
• An interpretable machine learning model integrating GDF15 enhances early identification of high-risk individuals.
• This biomarker-guided approach facilitates personalized risk reduction strategies and improved cardiovascular outcomes in PAD.