Type 1 Diabetes: towards an artificial pancreas, made in Nantes

Having worked on HIV alongside Nantes University Hospital in the 2000s, Claude Moog, Centrale Nantes researcher and Head of the Control Team at IRCCyN*, is today looking into type 1 diabetes in collaboration with several colleagues, including Nicolas Magdelaine, PhD candidate and the ADTSI Team at IRCCyN.

Type 1 diabetes
Sugar is the main source of energy within the body and insulin the hormone which allows muscles to use blood sugar. Type 1 diabetes is an auto-immune disease: the pancreas stops secreting insulin and thus the level of sugar in the blood is no longer controlled.

Current treatment involves daily insulin injections at meal times. Patients receive training in hospital, where they learn to estimate the amount of carbohydrate in a meal and assess their insulin requirement accordingly. The treatment remains empiric however and each meal turns into a math problem which can lead to hypoglycaemia (harmful short term) if an overdose is given or to hyperglycaemia (harmful long term) if the dose is insufficient.

In order to improve blood sugar control and, in more general terms, patient quality of life, the idea is to automate insulin injections - minimizing stress and long term complications - via an artificial pancreas. This takes the form of an insulin pump controlled with blood sugar readings. The hardware already exists: insulin pumps which communicate with sensors that measure blood sugar levels continuously.

It's now a question of determining the right algorithm so that the pump administers insulin as efficiently and as safely as possible. The automated system, with the facility for manual intervention, would ensure that the most appropriate dose is delivered to the patient to limit blood sugar excursions.

The Nantes approach

The IRCCyN researchers are not, by any means, alone in developing an artificial pancreas, but their original approach lies in its simplicity and its focus on the medical profession. Taking its inspiration directly from insulin therapy training, the Nantes-based solution is easily understood by doctors and patients, in contrast to other industrial automation solutions, which doctors are forced to trust without understanding how they work.

The Nantes algorithm takes its inspiration from the medical profession as the researchers learnt much from the doctors and use the same tools. Doctors would thus be comfortable advising their patients and making individually-tailored system adjustments.

"Today we have an algorithm that is likely to prove simple and efficient. It is critical and central to the solution, which can be refined over time," explained Claude Moog.

"This algorithm has just been successfully tested on the simulator approved by the Food and Drug Administration as the pre-clinical testing platform.  These encouraging results should lead to clinical trials in collaboration with Nantes and Rennes University Hospitals," added Nicolas Magdelaine.

Clinical trials to follow...

*IRRCyN - Communication and Cybernetics Research Institute of Nantes, involving Nantes University- Centrale Nantes - the National Centre for Scientific Research - Mines Nantes.