# The ‘Designer’ Artificial Pancreas: How Does Current Technology Mimic Pancreatic Physiology?

## TL;DR

This review explores how artificial pancreas systems mimic the function of a healthy pancreas to improve diabetes management and reduce patient burden.

## Contribution

The paper evaluates the evolution and current capabilities of artificial pancreas systems in replicating pancreatic physiology.

## Key findings

- Hybrid and closed-loop systems have improved glycaemic control and reduced self-management burden.
- Dual-hormone systems offer more physiological insulin regulation.
- Future advancements in insulin analogues and machine learning may enable more effective artificial pancreas systems.

## Abstract

Type 1 diabetes (T1D) is characterised by the autoimmune destruction of pancreatic insulin-producing beta cells, leading to a lifelong dependency on exogenous insulin administration for survival. Tight glycaemic control is essential to reduce the risk of long-term complications but achieving this is challenging due to high variability in insulin requirements, creating a significant burden for individuals and families.

Automated insulin delivery systems, which detect real-time blood glucose levels and self-adjust insulin delivery rates, have long been a goal in the management of diabetes. These systems aim to mimic healthy endocrine pancreatic physiology by replicating the glucose-responsive feedback control, thereby improving glycaemic outcomes and reducing the burden of self-management.

Over the last decade, several technological advances have culminated in the development of the artificial pancreas system (APS). Since the release of the first commercially available ‘hybrid’ system in 2017, multiple innovations have improved APS accuracy, personalisation and user experience.

This review discusses the evolution of the APS and evaluates how hybrid, closed-loop and dual-hormone systems replicate pancreatic physiology. We assess clinical and psychosocial outcomes and consider what future technological advancements, such as nuanced insulin analogues, alternative delivery routes, and machine learning algorithms, might enable a truly “designed” APS.