Joint modeling of onset of microvascular complications and longitudinal fasting blood sugar among type 2 diabetes patients at Ayder and Mekelle hospitals, 2025

Abstract

Introduction: Type 2 diabetes mellitus is a major cause of microvascular complications that impact patients' quality of life and contribute to long-term health issues. Worldwide, 18.8% of individuals with diabetes experience these complications, rising to 37.9% in Ethiopia. Fasting blood sugar levels are vital for managing Type 2 diabetes and are linked to the onset of microvascular complications; however, no studies have examined the relationship between fasting blood sugar progression and the onset of these complications in the Tigray region. Objective: To jointly model the longitudinal pattern of fasting blood sugar and the time to onset of microvascular complications among type 2 diabetes mellitus patients at Ayder and Mekelle hospitals using a joint modeling approach, January 2018 - March 2024 Methods: A retrospective longitudinal study was conducted among 447 type 2 diabetes patients at Ayder and Mekelle hospitals from May 25 to July 10, 2024. Data were collected from patients' medical records using Kobo Toolbox, with training and supervision provided to data collectors. The adequacy of the random effects assumption was assessed and proportional hazards assumptions were validated. Joint modeling of Cox-PH with random intercept and random slope was fitted. Akaike Information Criteria and Bayesian information criterion were used for model comparison. Variables were interpreted via hazard ratios and 95% confidence intervals. Result: The incidence rate of microvascular complications was 6 per 1000 person-months (95% CI: 5–7) with estimated restricted mean of 62 months. For one unit increased in longitudinal mean log (FBS), the hazard of complications was 18.92 (CI: 6.050, 54.598) times higher. For one unit increase in body mass index the hazard of complications was 6.1% more likely higher. A one-unit increase in the triglyceride level, the hazard of complications was 0.3% more likely higher. Patients on injection plus oral medications had a 2.45 times higher hazard of complications compared to those on single oral medications (AHR = 2.455, CI: 1.118, 5.302). Conclusion: fasting blood sugar trajectory, baseline fasting blood sugar, Body mass index, Triglycerides level, proteinuria, and type of medication intake were significant predictors of time to onset of microvascular complications. These findings emphasize the importance of continuous monitoring and tight control of blood glucose levels in diabetes management to prolong the onset of microvascular complications.