Post‑prandial glucose (PPG) excursions are a critical determinant of long‑term glycemic control and cardiovascular risk. While dietary composition is known to modulate PPG, the synergistic impact of concurrent physical activity on glucose swings remains under‑explored, especially across the spectrum of metabolic health. We aimed to better understand how meal macronutrient composition and physical activity together influence the magnitude, timing, and variability of PPG excursions in healthy, pre‑diabetic, and diabetic adults.

We took advantage of data from CGM sensors, which track glucose levels in interstitial fluid, offer a direct physiological signal related to food intake. Moreover, comprehensive, anonymized participant data, including demographics, anthropometrics, and blood biochemistry (HbA1c, fasting glucose, insulin, lipids profiles) were obtained across the spectrum of clinically relevant groups: healthy, pre-diabetes, and diabetes.

We developed an XGBoost machine learning model to predict the postprandial glucose response (measured as incremental Area Under the Curve, iAUC) from macronutrient and health data. The model achieved a Pearson correlation of 0.88 between predicted and ground-truth Area Under the Curve (AUC). It achieved a Pearson correlation of 0.71 for the more challenging incremental Area Under the Curve (iAUC). A SHAP (SHapley Additive exPlanations) analysis identified the key drivers of PPGRs in the dataset, further confirming its alignment with known nutritional science.The most important features for predicting glucose response iAUC were fasting glucose, carbohydrate amount, cholesterol, baseline glucose, protein amount, and HbA1c. These findings align with established clinical knowledge about the primary drivers of postprandial glucose, further validating the dataset's integrity and clinical relevance.

Meal macronutrient composition and concurrent physical activity jointly shape post‑prandial glucose dynamics. Moderate‑intensity activity blunts glucose excursions, especially after carbohydrate‑rich meals, though the effect is reduced in individuals with impaired glucose tolerance. These findings underscore the need for integrated dietary and exercise strategies tailored to metabolic status to optimize glycemic control.