1931-LB: Clinical Study on Reversal of Type 2 Diabetes through Closed-Loop Management of PEFAA Utilizing AI-Based Wearable Devices

Introduction and Objective: Objective: Relevant studies have shown that type 2 diabetes can be reversed through lifestyle interventions. However, increasing patient adherence and improving reversal rates remain challenges in current clinical treatment. We aim to explore ways to enhance patient reversal rates by incorporating AI devices and apps. Methods: Methods: A total of 200 patients with early-stage type 2 diabetes and obesity were randomly selected and assigned to either the control group (Group A, n=100, weight loss + conventional health education) or the intervention group (Group B, n=100, weight loss + AI-based PEFAA management). Group A received conventional health education, while Group B received AI-based online medical follow-up treatment and medical science popularization education once a week for 12 weeks, with a one-year follow-up. Relevant indicators such as blood glucose, HbA1c, pancreatic islet function, body weight, body fat percentage, and muscle mass were monitored in both groups to analyze differences in reversal rates and related clinical indicators. Results: Results: The reversal rate after 12 weeks of intervention was 43% in Group A and 65% in Group B(p<0.05). There was a difference in diabetes self-management knowledge scores between the two groups, with Group A scoring 60.1 ± 2.9 and Group B scoring 70.0 ± 1.8. Cognitive scores were positively correlated with diabetes reversal rates. Group A showed a 34% decrease in body fat percentage and an 8% increase in muscle mass, while Group B experienced a 56% decrease in body fat percentage and a 12% increase in muscle mass. These changes were positively correlated with scale scores and reversal rates. Conclusion: Conclusion: AI-based PEFAA management can improve the reversal effect in patients with type 2 diabetes, which is associated with improvements in cognitive behavior, increased muscle mass, and decreased body fat percentage. Disclosure K. Zeng: None. Y. Wang: None. G. Xu: None. Funding Basic and Applied Basic Research Foundation of Guangdong Province (2023A1515012575); National Natural Science Foundation of China (81902004); Supported by Distinguished Young Scientists projects by Guangdong Second Provincial General Hospital (2024E005)