Artificial Pancreas

Connecting insulin pumps and continuous glucose monitors (CGM) in an algorithm controlled closed-loop system, provides far better monitoring of glucose and dosing of insulin than conventional, manual treatment methods.

Status

• A few systems are on the market, and more are on the way.

• Current systems only automate basal insulin dosing. A fully automated system that also dispenses insulin bolus, e.g. at meals, requires the development of faster-acting insulin, the use of glucagon, and better artificial intelligence algorithms.

• A major obstacle to adoption of the systems among patients is the difficulties associated with using them. More convenient technology is required.

In artificial pancreas systems, a CGM connected to an insulin pump has taken over the role of beta cells. The CGM constantly measures blood glucose and sends values ​​to an insulin pump. In the pump, a computer program (algorithm) determines the dosage of insulin.

In more advanced dual-hormone systems, glucagon is added to the system to counteract excessive insulin doses. See the illustration above.

The systems are developed both as patch systems, ie. the pump is mounted directly on the body, and infusion set systems, ie. the pump is carried in, for instance, the pocket and is connected to the body through an infusion set. The algorithm for controlling the system can be built into the pump, the CGM, or externally e.g. on a smartphone.

The efficiency of artificial pancreas systems are limited by the following:

• Blood glucose levels from a CGM are not entirely accurate.

• A CGM measures blood sugar in the interstitial fluid, just under the skin. There is a delay of 10-15 min. in the value measured here, relative to actual blood sugar.

• The fastest acting insulins are not effective until approx. 10 min. after injection.

• Current systems all have in common that they only regulate basal insulin dosage. The most advanced systems provide suggestions for bolus dosing e.g. at meals.