AID systems are indispensable in type 1 diabetes

Modern innovative technologies can significantly reduce the diabetes-related burden. Daily insulin administration is vital for type 1 diabetics. The ideal therapy would be glucose-dependent control of insulin delivery without any risk of hypoglycemia. Automated insulin delivery (AID) systems are already relatively close to achieving this goal – the risk of hypoglycemia can be demonstrably reduced and quality of life improved. Whereby there are still some challenges both in regulatory and technological terms.

For type 1 diabetics in particular, technical innovations offer excellent opportunities for individualized therapy. This progress is due to intensive research efforts, explained Prof. Dr. med. et phil. Lia Bally, Senior Physician and Head of Research, University Department of Diabetology, Endocrinology, Nutritional Medicine and Metabolism, Inselspital Bern [1]. Insulin pumps have been around since the 1980s, and the first continuous glucose monitoring(CGM) devices were introduced in the 1990s. What was missing for a long time was suitable software that combines both in a closed loop.

Artificial pancreas aka (hybrid) closed loop – a milestone

The U.S. Food and Drug Administration (FDA) first approved a closed-loop glucose control system for type 1 diabetes mellitus in 2016, and the technology has been continuously improved since then. In the most advanced systems available today, the CGM is linked to an automated insulin pump using AI** algorithms for automated insulin delivery(AID) (Fig. 1) [2]. Such partially or fully automated devices are also called artificial pancreas (box). In this process, insulin delivery is adjusted both upward and downward to the current or expected blood glucose level. Because the systems approved today require some input from the user (e.g., announcement of an upcoming meal), they are referred to as hybrid closed-loop (vs. closed-loop) systems. Insulin delivery can be individualized to a person’s daily variations in insulin requirements, minor inaccuracies in the assessment of the amount of carbohydrate ingested are corrected, and many other factors affecting blood glucose levels are compensated for [2]. “Nowadays, several hybrid closed-loop systems are approved for the treatment of type 1 diabetes,” said Prof. Bally [1]. There are different control algorithms for insulin dosage adjustment. The most adaptive systems use “Model Predictive Control” (MPC) algorithms [3–5]. Using a mathematical model of the glucoregulatory system, an MPC algorithm calculates the optimal insulin dose (infusion rate) to be delivered at a given time using a predetermined glucose target by predicting glucose progression. Thus, MPC is a dynamic model that simulates the future behavior of the control process (insulin dose) as a function of the input signals (glucose concentration) and that continuously updates model parameters, such as insulin sensitivity.

** AI = Artificial Intelligence

Patients benefit from AID systems in several ways

Overall, the use of AID systems in type 1 diabetes demonstrated improvements in glucose control and time in range. In terms of patient-reported outcomes, it should be emphasized that the fear of hypoglycemia was reduced, and sleep quality and quality of life in general were improved [6]. A clinical trial evaluated the use of a hybrid closed-loop system switched on during the day and at night over a 4-week period in adults (n=28) with type 1 diabetes and an _HbA1c level below 7.5%. This showed an improvement in glucose control and a reduction in the risk of hypoglycemia, and the system was found to be safe and well tolerated [12]. Optimal glycemic control requires a balance between food intake, metabolic demand, energy expenditure, and insulin action profiles (Table 1) [7]. To this end, lifestyle factors such as diet and physical activity must be integrated into automated insulin delivery. The use of hybrid closed-loop systems also facilitates the prevention of hypoglycemia during exercise, although this remains a challenge despite advanced technology [2].

Optimizability of AID systems

Smaller, more accurate CGM devices with longer wear time and smaller insulin pumps that transfer the user interface to a smartphone/smartwatch may improve usability and minimize device burden [8]. Interoperable devices and data management platforms will provide users with the flexibility to create their own personalized AID system.

Artificial pancreas – automated insulin dosing of the basal rate
The principle of automated insulin delivery (AID) is that continuous glucose monitoring (CGM) is linked to an insulin pump via an algorithm. Algorithms for automated control of insulin delivery can be well represented mathematically. The challenge is to adapt to each patient’s current life situation. Specifically, as part of a “hybrid closed loop,” automated insulin dosing of the basal rate is performed based on glucose values measured by sensor, and insulin delivery is interrupted if hypoglycemia is imminent. Since manual input is still required at meal times and for correction, these are hybrid closed-loop systems. For type 1 diabetics, AID systems are considered the gold standard in therapy, but insulin-dependent type 2 diabetics can also benefit from this innovative technology. In summary, automated regulation of basal insulin is combined with manual entry of meal and correction bolus.

to [1,2,11]

The introduction of even faster-acting insulin analogues would allow the performance of closed-loop systems to be further improved by a more rapid onset and decay of insulin action. And integrating additional signals into the algorithms, such as heart rate or accelerometers, to detect physical activity more quickly than with CGM alone could help reduce the risk of hypoglycemia during physical activity. This would be particularly beneficial in young children, where activity is usually spontaneous and unpredictable and hypoglycemia is a major problem.

Congress: SGAIM Spring Congress

Literature:

“Artificial Intelligence in Diabetes Management,” Prof. Dr. med. et phil. Lia Bally, SGAIM Spring Congress, 10-12 May 2023.
Boettcher C, et al: 100 years of insulin therapy. “Current challenges in the therapy of type 1 diabetes in children”. Swiss Med Forum 2022; 22(47): 767-771.
Bequette BW: Algorithms for a closed-loop artificial pancreas: the case for model predictive control. J Diabetes Sci Technol 2013; 7: 1632-1643.
Elleri D, et al: Evaluation of a portable ambulatory prototype for automated overnight closed-loop insulin delivery in young people with type 1 diabetes. Pediatr Diabetes 2012; 13: 449-453.
Hovorka R, et al: Manual closed-loop insulin delivery in children and adolescents with type 1 diabetes: a phase 2 randomised crossover trial. Lancet 2010; 375: 743-751.
Weisman A, et al: Effect of artificial pancreas systems on glycaemic control in patients with type 1 diabetes: a systematic review and meta-analysis of outpatient randomised controlled trials. Lancet Diabetes Endocrinol 2017; 5(7): 501-512.
Smart CE, et al: ISPAD Clinical Practice Consensus Guidelines 2018: Nutritional management in children and adolescents with diabetes. Pediatr Diabetes 2018; 19 Suppl 27: 136-154.
Boughton CK, Hovorka R: New closed-loop insulin systems. Diabetologia 2021: 64: 1007-1015, https://link.springer.com/article/10.1007/s00125-021-05391-w,(last accessed 07/11/2023).
Schneider L, Lehmann R: “Swiss Diabetes Guide. Swiss Med Forum 2021; 21(1516): 251-256.
Lechleitner M, et al: Diagnostik und Therapie des Typ 1 Diabetes mellitus (Update 2023) [Diagnosis and insulin therapy of type 1 diabetes mellitus (Update 2023). Wien Klin Wochenschr 2023; 135(Suppl 1): 98-105.
Kordonouri O, Kerner W: Diabetes mellitus type 1 – update. Internist (Berl) 2021; 62(6): 627-637.
Bally L, et al: Day-and-night glycaemic control with closed-loop insulin delivery versus conventional insulin pump therapy in free-living adults with well-controlled type 1 diabetes: an open-label, randomised, crossover study. Lancet Diabetes Endocrinol 2017; 5(4): 261-270.