Résumé
Closed-loop systems, so-called artificial pancreas, aim at driving insulin delivery by blood glucose levels in patients with diabetes treated with insulin. They include a continuous glucose monitoring device, a control algorithm and an insulin infusion pump. The control algorithm is the key of the system since it commands insulin infusion in order to maintain blood glucose in a predefined target range or close to a target glucose level. Algorithm prescriptions are based on past, current and predicted glucose levels according to different designs: proportional-integral-derivative (PID), model predictive control (MPC) or fuzzy logic. The last decade has shown dramatic advances toward the use in free-life of closed-loop systems through demonstrations of feasibility, safety and efficacy in successive hospital, transitional and outpatient trials. Permanent innovation has contributed to this progress by more accurate sensors for glucose monitoring, wearable platforms for running algorithms and wireless communication between devices. The approval by the FDA of a first closed-loop insulin delivery system for routine therapy of type 1 diabetes in September 2016 illustrates the maturity reached by the artificial pancreas after about 40 years of development.