Neural-Net Artificial Pancreas: A Randomized Crossover Trial of a First-in-Class Automated Insulin Delivery Algorithm.

Background: Automated insulin delivery (AID) is now integral to the clinical practice of type 1 diabetes (T1D). The objective of this pilot-feasibility study was to introduce a new regulatory and clinical paradigm-a Neural-Net Artificial Pancreas (NAP)-an encoding of an AID algorithm into a neural network that approximates its action and assess NAP versus the original AID algorithm. Methods: The University of Virginia Model-Predictive Control (UMPC) algorithm was encoded into a neural network, creating its NAP approximation. Seventeen AID users with T1D were recruited and 15 participated in two consecutive 20-h hotel sessions, receiving in random order either NAP or UMPC. Their demographic characteristics were ages 22-68 years old, duration of diabetes 7-58 years, gender 10/5 female/male, White Non-Hispanic/Black 13/2, and baseline glycated hemoglobin 5.4%-8.1%. Results: The time-in-range (TIR) difference between NAP and UMPC, adjusted for entry glucose level, was 1 percentage point, with absolute TIR values of 86% (NAP) and 87% (UMPC). The two algorithms achieved similar times <70 mg/dL of 2.0% versus 1.8% and coefficients of variation of 29.3% (NAP) versus 29.1 (UMPC)%. Under identical inputs, the average absolute insulin-recommendation difference was 0.031 U/h. There were no serious adverse events on either controller. NAP had sixfold lower computational demands than UMPC. Conclusion: In a randomized crossover study, a neural-network encoding of a complex model-predictive control algorithm demonstrated similar performance, at a fraction of the computational demands. Regulatory and clinical doors are therefore open for contemporary machine-learning methods to enter the AID field. Clinical Trial Registration number: NCT05876273.

Performance of an Artificial Pancreas System for Young Children with Type 1 Diabetes.

BACKGROUND: Young children 5-8 years old with type 1 diabetes (T1D) exhibit clear needs for improved glycemic control but may be limited in their ability to safely interact with an artificial pancreas system. Our goal was to evaluate the safety and performance of an artificial pancreas (AP) system among young children with T1D. RESEARCH DESIGN AND METHODS: In a randomized, crossover trial, children with T1D age 5-8 years were enrolled to receive on separate study periods (in random order) either the UVa AP using the DiAs Control Platform software with child-resistant lock-out screens (followed as an out-patient admission) or their usual insulin pump+continuous glucose monitor (CGM) care at home. Hypoglycemic events and CGM tracings were compared between the two 68-h study periods. All analyses were adjusted for level of physical activity as tracked using Fitbit devices. RESULTS: Twelve participants (median age 7 years, n = 6 males) completed the trial. Compared to home care, the AP admission resulted in increased time with blood glucose (BG) 70-180 mg/dL (73% vs. 47%) and lower mean BG (152 mg/dL vs. 190 mg/dL), both P < 0.001 after adjustment for activity. Occurrence of hypoglycemia was similar between sessions without differences in time <70 mg/dL (AP 1.1% ± 1.1%; home 1.6% ± 1.2%). There were no adverse events during the AP or home study periods. CONCLUSIONS: Use of an AP in young children was safe and resulted in improved mean BG without increased hypoglycemia. This suggests that AP use in young children is safe and improves overall diabetes control. ClinicalTrials.gov registration number: NCT02750267.