Assessment of Serum Creatinine and Serum Sodium Prognostic Potential in Heart Failure Patients Using Machine Learning.

Heart failure (HF) is the leading etiology for hospital admissions and ranks among the foremost contributors to mortality. This complex clinical syndrome with various phenotypes is categorized by left ventricle ejection fraction levels (LVEF), namely preserved (HFpEF), mid-range (HFmEF), and reduced (HFrEF). This study investigates the prognostic impact of serum creatinine and serum sodium levels in HF patients across these three classes using machine learning tools. A comprehensive dataset of HF patients' medical records including serum sodium and serum creatinine was utilized. Machine learning regression models were employed to predict the LVEF levels. Additionally, classification models were implemented to categorize patients into HFpEF, HFmEF, and HFrEF classes. Regression analyses revealed the predictive capabilities of serum sodium and serum creatinine in estimating the progression of HF severity. Furthermore, classification models successfully differentiated between the three EF classes, providing valuable insights into the classification patterns of HF patients based on these biomarkers. The results demonstrated the significance of serum sodium serum creatinine as prognostic markers in HF, and this contributes to a more nuanced approach to HF management, paving the way for targeted interventions and improved patient outcomes. Moreover, this study highlights the potential of machine learning techniques to enhance risk stratification and classification in HF patients, enabling personalized prognostication and treatment strategies.