Reproducibility of continuous glucose monitoring-derived postprandial glucose features and their association with glycemic control in type 2 diabetes

Journal article, 2026

BACKGROUND/OBJECTIVES: Targeting postprandial glucose response (PPGR) is more effective than lowering fasting plasma glucose in improving glycemic control and reducing cardiovascular risk in individuals with type 2 diabetes (T2D). Continuous glucose monitoring (CGM) has the potential to uncover time-related features of PPGR. This study evaluates the within-subject reproducibility of dynamic PPGR parameters obtained by CGM and explores their potential as independent predictors of glycemic control in T2D. SUBJECTS/METHODS: A total of 102 individuals with T2D underwent a 7-day CGM and consumed a standardized breakfast twice to assess the 4-h glucose response, described by the following parameters: glucose peak-the highest glucose value; time to peak-time of peak occurrence; delta glucose max-the difference between the peak and fasting glucose; nadir-the lowest post-peak glucose value; incremental area under the glucose curve; mean postprandial glucose-the average interstitial glucose concentration. Intraclass correlation coefficients (ICCs) for both single and average measurements with their 95% confidence intervals (CIs), were calculated for PPGR parameters to estimate their within-subject reproducibility. Multivariable linear regression models assessed the independent predictive contribution of PPGR parameters, fasting glucose, and 2-h postprandial glucose on 7-day CGM metrics and HbA1c. RESULTS: Moderate to good reproducibility for single measurements was observed for mean glucose (ICC: 0.78, 95%CI 0.69-0.84), glucose peak (ICC: 0.69, 95% CI 0.57-0.78), and nadir (0.74, 95% CI 0.64-0.82). Mean postprandial glucose was the strongest predictor of 7-day time in range (β = -0.772, p < 0.001), 7-day mean glucose (β = 0.800, p < 0.001) and HbA1c (β = 0.434, p < 0.001), whereas the glucose peak was the main predictor of short-term glycemic variability, as reflected by the 7-day coefficient of variation (β = 0.258, p = 0.006) and mean amplitude of glucose excursions (β = 0.613, p < 0.001). CONCLUSION: In individuals with T2D, CGM-derived parameters of PPGR are reproducible and could represent a practical tool to uncover meaningful information about glucose control, which 2-h postprandial glucose fails to predict.