Machine Learning to Tailor Intermittent Fasting for Blood Pressure Improvement.

Area of Science:

• Cardiovascular Health
• Metabolic Health
• Nutritional Science

Background:

• Intermittent fasting (IF) shows promise for blood pressure reduction.
• Personalized strategies are needed for optimal IF effectiveness.
• Premenopausal women without hypertension are a key demographic for study.

Purpose of the Study:

• Develop a machine learning framework to predict blood pressure improvement.
• Identify predictors of response to IF and calorie restriction protocols.
• Stratify patient response based on demographic and metabolic factors.

Main Methods:

• Machine learning model development for predicting systolic blood pressure reduction (≥5 mmHg).
• Analysis of various IF and calorie restriction protocols.
• Logistic regression and decision tree analyses to identify key predictors.

Main Results:

• Model achieved 77% accuracy and 0.8 AUC in predicting responders.
• Intermittent Energy and Carbohydrate Restriction + free Protein and Fat (IECR + FF) and IECR protocols showed highest effectiveness.
• Younger age (≤47 years) and lower waist-to-hip ratio predicted better outcomes; higher HDL was beneficial, while elevated LDL was detrimental.

Conclusions:

• Machine learning can predict blood pressure response to dietary interventions.
• Dietary intervention type and age are critical factors for success.
• Central adiposity significantly hinders blood pressure improvement; further validation in diverse populations is required.