Impact of Carbohydrate Intake Fluctuations on Glucose Profiles: Insights from Continuous Glucose Monitoring-Based Patient Clustering

Background: Continuous glucose monitoring (CGM) is widely applied in daily glucose management. However, its potential to categorize individuals based on glucose profiles is not fully established. This study employed CGM-based patient clustering and examined nutritional factors influencing glucose patterns. Methods: This prospective observational study enrolled 34 individuals with diabetes. Retrospective professional CGM was conducted over 7 days, during which food intake was recorded. K-means clustering was performed using CGM-derived coefficient of variation (CV) and time in range. Macronutrient intake and its fluctuations were compared across clusters. Results: Participants were grouped into cluster 1 (well-controlled), cluster 2 (highest CV), and cluster 3 (highest mean glucose). Baseline clinical characteristics, daily energy intake (kcal), and macronutrient intake did not differ significantly among clusters. However, carbohydrate intake fluctuations were greater in cluster 3 (CV 41.0%±32.1%, standard deviation [SD] 502.1±363.4 kcal) than in cluster 1 (CV 21.9%±9.0%, SD 260.2±94.1 kcal) and cluster 2 (CV 19.2%±9.1%, SD 250.2±126.1 kcal) (P=0.123 for CV; P=0.024 for SD). The SD (kcal) of carbohydrate intake was positively correlated with mean glucose levels (rho=0.88, P=0.023). Conclusion: CGM enables categorization of individuals based on glucose profiles, and higher carbohydrate intake fluctuations are associated with poorer glycemic control. Personalized dietary strategies, particularly stabilizing carbohydrate intake, may support better glucose management in individuals with high mean glucose and low CV.