Endocrine Abstracts

JOINT3802

Background: Type 1 diabetes leads to absolute insulin deficiency and necessitates lifelong insulin therapy. Despite advancements in diabetes care, achieving optimal glycemic control remains challenging. Automated insulin delivery (AID) systems are a breakthrough in diabetes management, improving glycemic outcomes and quality of life. While regulatory-approved systems are available, open-source systems—developed by patients for self-installation—offer an alternative that requires technical expertise and active engagement with support networks. This study uniquely evaluated the efficacy and safety of both systems in a real-world, heterogeneous population, including patients with limited technological literacy and restricted internet access— groups assumed to benefit less from open-source AID systems.

Methods: This cross-sectional study was conducted at Shaare Zedek Medical Center and involved children with type 1 diabetes. Glycemic control was assessed while the children were using an AID system, compared to a sensor-augmented pump. The change in glycemic control following the transition to the automated system was compared between two systems: the regulatory-approved Medtronic G780 and an open-source system (Android APS).

Results: Sixty-one children were included in the study- 30 using the approved system and 31 using the open-source system. Both groups showed significant improvements in glycemic control, including an increase in time in range and reduction in mean blood glucose, estimated HbA1c/Glucose Management Indicator (GMI) and time spent in hypo and hyperglycemia. A trend toward greater reduction in mean blood glucose and estimated HBA1C/GMI was observed in the regulatory approved system group. No significant differences in rates of severe hypoglycemia or diabetic ketoacidosis were observed between the two systems and before and after the transition to an AID system.

Conclusions: Real world data from a heterogeneous population demonstrates significant improvement in glycemic control with both regulatory-approved and open-source AID systems, with an acceptable safety profile. Notably, even patients with limited technological literacy achieved meaningful improvements using open-source systems. These findings highlight the adaptability and effectiveness of AID systems across diverse patient populations, emphasizing the importance of preserving patient autonomy in choosing the system that best suits their needs. However, the observed trend favoring the regulatory-approved system warrants further evaluation in larger trials.