Automated insulin delivery systems are automated (or semi-automated) systems designed to assist people with insulin-requiring diabetes, by automatically adjusting insulin delivery in response to blood glucose levels. Currently available systems (as of October 2020) can only deliver (and regulate delivery of) a single hormone—insulin. Other systems currently in development aim to improve on current systems by adding one or more additional hormones that can be delivered as needed, providing something closer to the endocrine functionality of the pancreas. The endocrine functionality of the pancreas is provided by islet cells which produce the hormones insulin and glucagon. Artificial pancreatic technology mimics the secretion of these hormones into the bloodstream in response to the body's changing blood glucose levels. Maintaining balanced blood sugar levels is crucial to the function of the brain, liver, and kidneys. Therefore, for people with diabetes, it is necessary that the levels be kept balanced when the body cannot produce insulin itself. Automated insulin delivery (AID) systems are often referred to using the term artificial pancreas, but the term has no precise, universally accepted definition. For uses other than automated insulin delivery, see Artificial pancreas (disambiguation). The first automated insulin delivery system was known as the Biostator. Currently available AID systems fall into three broad classes based on their capabilities. The first systems released can only halt insulin delivery (predictive low glucose suspend) in response to already low or predicted low glucose. Hybrid Closed Loop systems can modulate delivery both up and down, although users still initiate insulin doses (boluses) for meals and typically "announce" or enter meal information. Fully Closed Loops require no manual insulin delivery actions or announcement for meals. A step forward from threshold suspend systems, predictive low glucose suspend (PLGS) systems use a mathematical model to extrapolate predicted future blood sugar levels based on recent past readings from a CGM.
Johan Auwerx, Kristina Schoonjans
Sandrine Gerber, Qixing Zhang, Yi Wang, Léo Bühler