2884-LB: Impact of Rapid Occlusion Notification on Glycemia in Users of the twiist AID System

Introduction and Objective: Interruption of insulin delivery from an automated insulin delivery (AID) device can significantly impact glycemic control. Rapid detection and notification of occlusions may enable timely user intervention and limit glucose excursions. This study evaluates glucose levels in users of the twiist AID system before and after notifications of interruptions in insulin delivery. Methods: A retrospective review of real-world data was conducted for 2,006 twiist users who initiated treatment between April 2025 and Jan 2026. Continuous glucose readings during the one-hour period before and after an occlusion alarm were analyzed. To minimize confounding factors, occlusion notifications within three hours before or one hour after a carbohydrate entry or insulin bolus were excluded. Results: Across the population, 16,190 occlusion alarms were identified for analysis. Pre-notification: Mean glucose levels were stable prior to occlusion notification, decreasing by 0.4 mg/dL (95% CI -1.6 to 0.8) from 60 to 30 minutes before the alert (146.13 ±55.68 mg/dL to 145.75 ±54.13 mg/dL). Mean glucose then increased by 9.9 mg/dL (95% CI 8.7 to 11.0) from 30 minutes prior to notification to the time of notification (145.75 ± 54.13 mg/dL to 155.64 ± 52.24 mg/dL). Post-notification: Following occlusion notification, mean glucose increased by 2.6 mg/dL (95% CI 1.5 to 3.8) during the first 30 minutes (155.64 ± 52.24 mg/dL to 158.25 ± 53.02 mg/dL)., followed by an increase of 0.2 mg/dL between 30 and 60 minutes (95% CI -0.9 to 1.4) post-notification (158.25 ± 53.02 mg/dL to 158.50 ± 55.70 mg/dL) . Conclusion: In this real-world analysis, glucose levels began to rise shortly before occlusion notification and increased modestly thereafter. These findings suggest that occlusions were detected very near the early phase of glycemic worsening but were generally associated with limited subsequent glucose excursions. Rapid occlusion detection with user notification may help limit prolonged interruptions in insulin delivery and associated glycemic deterioration. Disclosure K. Snow: Employee; Current; Sequel Medtech. M.S. Hughes: Consultant; Current; Dexcom, Inc., Sanofi, Sequel Med Tech, LLC. E. Cobry: None. R.J. Galindo: Consultant; Ended; Abbott, Boehringer Ingelheim International GmbH, Dexcom, Inc., Eli Lilly and Company, Gan & Lee Pharmaceuticals, Novo Nordisk, Medtronic. Research Support; Current; Roche Pharmaceuticals, Boehringer Ingelheim International GmbH, Novo Nordisk. J. Mitri: Employee; Current; Sequel.