Artificial Intelligence for Diabetes Control

15 July 2010 (Last Updated July 15th, 2010 18:30)

Researchers are making steady progress as they work to develop artificial pancreases for type 1 diabetes sufferers. But, as Eric Renard of Montpellier University explains to Nic Paton, this next generation of insulin pumps may not be the best solution for everyone.

Artificial Intelligence for Diabetes Control

According to the American Diabetes Association, 23.6 million people in the US – 7.8% of the population – have diabetes. One in every 400-600 children and adolescents suffer from type 1 diabetes, the variant that destroys the body's ability to make insulin. Many struggle to cope with the regime of daily insulin injections living with this disease entails, but work is underway on both sides of the Atlantic that could change all that.

In the UK, scientists at the University of Cambridge have been using an artificial pancreas – a combination of an insulin pump and a matchbox-sized glucose monitor – to improve blood sugar control in diabetes patients aged between five and 18 years old.

Their findings, published in The Lancet in February 2010, showed that the new device was more effective than conventional treatment, and led the report's authors to hail the research as 'an important stepping stone' towards bringing such an artificial pancreas to the commercial market.

The system used a computer algorithm to determine how much insulin to deliver, based on real-time glucose readings.

"The next generation of pumps with continuous monitoring and closed-loop systems are at the experimental stage."

The researchers found that their artificial pancreas halved the amount of time patients spent in mild hypoglycaemia (where blood glucose falls to 3.9mmol per litre) compared with a standard insulin pump, and prevented significant hypoglycaemia (3.0mmol per litre).

Meanwhile, doctors at Boston University, Massachusetts General Hospital and Harvard Medical School in the US have been conducting human trials of an artificial pancreas that administered insulin and glucagon.

In the device's current form, researchers tracked blood glucose by inserting a sensor into a vein; the intention is that future versions will use commercially available continuous blood glucose monitors.

Regime change

The wider use of such devices has the potential to radically alter how type 1 diabetes is managed, particularly in children. Eric Renard, professor of endocrinology and metabolism at Montpellier Medical School in France, believes it could even herald a new dawn in pumping devices.

"It is more significant for children because the method allows for stable glucose control rather than multiple injections," says Renard, who also wrote a commentary to accompany the University of Cambridge study. "You may have to inject four or five times a day, but with the pump you only have to change the catheter once every three to four days, so it is much less aggressive. There is more freedom when it comes to eating and living, and it is possible to moderate insulin delivery and frequency."

An insulin pump, he explains, has significant potential in terms of improving the patient's comfort and control, as well as the quality of treatment it provides.

"But it remains a system that is dependent on the frequency of blood glucose monitoring," says Renard. "You need to have a lot of data, and need to check your blood glucose level with fingertip pricking quite frequently to get the best results. So the next step forward is a combination system that brings together insulin pump therapy with constant monitoring. We will see the next generation of pumps combined with continuous monitoring and closed-loop systems, though at the moment this still in an experimental stage."

Is the pump a panacea?

As is often the case with medical innovations, one of the issues surrounding the pumps is cost. Renard concedes that such devices cost significantly more than conventional needle-based management systems – not to mention the hefty R&D investment that has to go into the device in the first place.

"Insulin pumps and artificial pancreases can make condition management straightforward."

"It costs about €50 for a five-day glucose-sensing system, and then you will need a transmitting device and a monitor," says Renard. "When you add all these together, the costs are significantly more. But then you should also get less hypoglycaemia, fewer hospital visits and fewer complications, so that cost needs to be offset against the longer-term benefits."

Just because the technology is there, it does not necessarily mean devices such as these will be suitable for all diabetes patients.

"It is important to select patients that will benefit the most from systems such as this," Renard warns. "You need to prescribe this system to patients who will use the information. Continuous monitoring can help them to better control their condition, but only if they use the system the whole time and make decisions based on it. For example, adolescents may not want to wear the system all day or may not take account of the data or change their treatment – in that case, continuous monitoring is both useless and expensive."

Some diabetes patients may not even want to go down this route. Insulin pumps and artificial pancreases can make management of the condition more straightforward, but they can also be intrusive in other ways. Renard explains that some patients may not want to use pump systems, or actively prefer the more conventional ways of managing their diabetes.

"Some patients might argue that, with a pump, you can always see and feel your diabetes,"Renard says. "You cannot go to the beach, for example, or undress as easily in public."

"There should be the option of patients using discreet devices to handle their condition."

With injections the condition can be kept much more discreet, and so for some patients pumps may be not a solution that suits them. Another option could be to go further down the patch pump route, where a small pump is attached to the patient's skin, doing away with the need for a catheter.

"What this means is that manufacturers will need to offer better technology according to patient wishes," he says.

"Yes, there should be sophisticated devices based on what some people want, but for other patients there should still be an option of using more discreet devices to handle their condition.

"In the future, I suspect that diabetes technology will either be based on very sophisticated devices, with all the possibilities that they bring, or on much simpler discreet devices that give people some flexibility in manage their diabetes."