Abstract

Objective: This study aimed to evolve a multimodal physiological framework to be able to measure the dynamic association of cortisol rhythms and glycemic fluctuation in Ramadan fasting.
Methods: Thirty-four adults were assessed at pre-Ramadan, Ramadan, and post-Ramadan. High-frequency salivary cortisol sampling was combined with continuous glucose monitoring (CGM). To measure delays in cortisol action, a time-dependent coupling function was constructed to separate hypothalamic-pituitary-adrenal (HPA) axis-driven glucose excursions from dietary or behavioral effects. Glycemic variability indexes were adjusted with cortisol to complement the CGM measures. The GRU prediction model with cortisol peaks and heart-rate variability was evaluated to predict glucose in the short term.
Results: There was a clear bimodal pattern of cortisol response in fasting with peaks before suhur and after iftar, which were glucose nadirs and postprandial spikes. The cortisol-adjusted variability indices showed stronger associations with hypoglycemia risk and post-iftar hyperglycemia than conventional CGM variables. The integrated GRU model was more accurate and sensitive at predicting hypoglycemia than glucose-only models.
Conclusion: The dynamics of the HPA-axis played an essential role in glycemic fluctuations during Ramadan fasting. The given framework can offer a more physiologically based approach to interpreting glucose variations and can be used to provide safer, personalized fasting guidance to patients with type 2 diabetes (T2D).

Key words: Type 2 diabetes, Ramadan fasting, cortisol rhythms, glycemic variability, HPA axis, continuous glucose monitoring, glucose forecasting.