Computational Longitudinal Analysis of Blood Glucose Response to Kombucha–Ginger Combination in Alloxan-InducedMice Using Generalized Estimating Equations

Abstract

Background: Diabetes mellitus is characterized by hyperglycemia and requires effective supportive approaches. Kombucha and ginger contain bioactive components that may affect glucose metabolism. Objective: This study aimed to analyze longitudinal blood glucose responses after administration of kombucha (Camellia sinensis) combined with ginger (Zingiber officinale) in alloxan-induced mice using a computational approach. Methodology: Data were obtained from 25 male mice allocated to five groups: 0.5% Na-CMC, metformin 65 mg/kgBW, and kombucha-ginger at 2%, 5%, and 7%. Post-induction, day-7, and day-14 glucose values were reshaped into longitudinal format and analyzed using Gaussian Generalized Estimating Equations (GEE) with an identity link and exchangeable working correlation. Results and Discussion: The model shows a significant time effect (Wald χ²=94.04; p<0.001) and a significant group×time interaction (Wald χ²=106.74; p<0.001). Mean reductions from post-induction to day 14 are 49.0 mg/dL for Na-CMC, 100.6 mg/dL for metformin, 20.6 mg/dL for 2%, 55.6 mg/dL for 5%, and 60.0 mg/dL for 7%. The 7% concentration shows the largest descriptive response among kombucha-ginger groups but remains smaller than metformin. Conclusion: GEE provides an appropriate framework for repeated measurements and demonstrates distinct longitudinal response patterns among treatment groups.