Abstract P3101: Predictive Modeling of Glycemic Variability and its Association with Metabolic Health: A Precision Nutrition Approach Using Korean National Data

The global burden of metabolic diseases, particularly diabetes and hyperglycemia, presents ongoing challenges for effective treatment and management. Conventional glycemic control strategies may not be suitable for healthy individuals and those with prediabetes. To enhance the personalized characterization of glycemic variability (GV), which reflects long-term glycemic trends rather than short-term responses to individual meals, this study developed a predictive model for Time in Range (TIR). This model enables monitoring of metabolic effects associated with lifestyle choices without the need for a continuous glucose monitoring system (CGMS). To estimate TIR, we applied a backward elimination linear regression model to optimize predictions based on lifestyle factors such as physical activity, sleep, stress, and dietary intake from 48 CGMS clinical study participants. The model showed a strong correlation between predicted TIR and actual metrics (R = 0.743, p < 0.001). Using data from 24,564 adults from the Korean National Health and Nutrition Examination Survey, we confirmed the association between predicted TIR values and metabolic health indicators. The results demonstrated a significant association between predicted TIR and reduced risk of metabolic diseases, particularly hypertension (Q1 vs. Q5: OR = 0.87; 95% CI = 0.79–0.95) and hypercholesterolemia (Q1 vs. Q5: OR = 0.82; 95% CI = 0.73–0.92). These findings may contribute to the development of precision nutrition-based healthcare for personalized prevention and management of metabolic diseases.