Automated “Artificial Pancreas” Controls Blood Glucose Levels in Diabetes Patients for First Time

Very exciting news from a study presented at the American Diabetes Association Scientific Sessions

New Orleans, LA, June 8, 2009 – UC Santa Barbara and Sansum Diabetes Research Institute scientists have demonstrated for the first time that an automated artificial pancreas system (APS) can safely and effectively maintain desired blood glucose levels in patients with type 1 diabetes. The clinical study results will be presented today in a late-breaking poster session[1] at the American Diabetes Association’s 69th Scientific Sessions in New Orleans.

The UCSB and Sansum researchers, working with the Schneider Children’s Medical Center of Israel, tested an automated insulin delivery system comprising the OmniPod® System and the DexCom STS7® continuous glucose monitor, linked and controlled through UCSB’s artificial pancreas software. The software’s insulin delivery algorithm, optimized for each patient, includes a unique safety feature, based on clinical parameters, which prevents insulin-induced low blood glucose levels (hypoglycemia).

Without any outside intervention, the system restored normal blood glucose levels following both induced high levels (hyperglycemia) and unanticipated meals, while avoiding hypoglycemia. This was achieved through the automatic delivery of insulin to correct for the induced high blood glucose levels, and an insulin infusion rate moderated to ensure a smooth return to normal levels and avoid low blood glucose levels.

“This study demonstrates for the first time a completely automated insulin delivery system that frees the patients from controlling their pumps manually, eliminating the question of compliance in treatment,” said principal investigator Frank Doyle, Professor of Chemical Engineering at UCSB.

Doyle continued, “We pulled together a talented team of engineers and medical doctors who created the critical element of the artificial pancreas—a unique algorithm that is robust and straightforward to implement. It’s become the gold standard.” The UCSB APS software platform is also being used by a number of other teams working on the artificial pancreas project, but no other team has advanced yet to wholly-automated clinical trials.

Closed loop trials were performed in four patients for a mean duration of 5 hours (range of 2-7 hours) and included a meal of 30 grams of carbohydrates. The mean Low Blood Glucose Index was 0.02 (range 0-0.06), the mean High Blood Glucose Index was 9 (range 4.2-15), and the median Daily Risk Range was ‘low’ (range ‘low’ to ‘moderate’).

The poster was presented in New Orleans by Matthew Percival of Doyle’s research group. Other researchers in the study included Dr. Eyal Dassau, senior investigator, and Dr. Benyamin Grossman, both of Doyle’s group; and Sansum’s Dr. Lois Jovanovič, CEO and Chief Scientific Officer, and Dr. Howard Zisser, Director of Clinical Research.

The research is part of the artificial pancreas project, which is funded by the Juvenile Diabetes Research Foundation and is being conducted by an international group of diabetes research centers. The project’s first goal is to integrate an insulin pump and continuous blood glucose monitor to closely replicate a healthy pancreas for patients with type 1 diabetes—patients whose pancreases no longer produce insulin, which is used by the body to control blood glucose levels. An artificial pancreas will allow for tighter and automated control of blood glucose levels, which would significantly help to avoid the long-term complications of the disease.

U C Santa Barbara is a leading research institution. The two groups at UCSB involved in this study are the Department of Chemical Engineering, in 2007 ranked ninth in the United States and second in the University of California by U.S. News and World Report, and the Biomolecular Science and Engineering Program, which offers a unique interdisciplinary approach to graduate training and research spanning Biochemistry, Molecular Biology, Bioengineering and Biomolecular Materials.

Sansum Diabetes Research Institute is a non-profit research center devoted to the prevention, treatment, and cure of diabetes through research and education. It is best known for its work on methods to detect and chart the progress of diabetes and its expertise in new diabetes technology.

The OmniPod Insulin Management System—the OmniPod insulin pump and its Personal Diabetes Manager, which normally allows the patient to control it—is manufactured and sold by Insulet Corporation (Nasdaq: PODD). The DexCom STS7 continuous glucose monitor is a product of DexCom, Inc.

Thanks for sharing this. It’s good to hear that they are advancing with this, but I am still so skeptical.

There are so many variables and I find that I often need to make judgements calls about when to give more insulin or not. The active insulin (insulin on board or bolus on board) are true estimations of how much insulin is active (I think that it is something that needs to be personalized at the individual level?). Experience has taught me when the pump is likely over- or under-estimating bolus and I make judgement calls in response. Although “eliminating the question of compliance in treatment” may be important, it also has high costs. I think it is not feasible to store all the information that the person with diabetes knows about their own treatment. Mind vs. machine?

Also, another question can this can correct hypoglycemia? Will it include glucagon? We would still need to check blood sugar for this I assume. In my experience, stopping the pump is not a good way to correct glucose.

Perhaps I’m just a control freak and I’m not ready to give up the control of my pump to some other device!

OH MY GOD! Want! Want! WANT! That is so awesome. (The main reason I want it is because I’m incredibly non-compliant, and it would be lovely to have something that forced me to be compliant.)

Getting closer to closing the loop! Yay!!

I am currently using the Opod and dexy both and I am still learning the ins and outs I hope I gain tighter control like that!!

I agree with your skepticism. I am glad to know they are working on this and improving their results, but I still see too much variation in sensor and meter BS to be comfortable with something else controlling my insulin intake. Sometimes my sensor BS value seems right on but definitely not all the time. When it does not match my meter, I will check my BS with my meter up to 3 times & sometimes get 3 different results! - so which one is correct?

At the moment I am more than happy with a pump and CGMS!

Do you suppose that Insurance companies might actually be more willing to pay for a closed loop system because it would help the problem of non-compliance which just having a pump and/or CGMS cannot do???

This sounds pretty amazing, but what would be even more amazing is if they could make an artificial pancreas that can do all the things a real pancreas can do. I know a young lady who has pancreatitis and of course has diabetes as a result. But she only has half of her pancreas left now, and as you know it also produces bile and other enzymes, and I don’t know what all. What I do know is she is down to 89 pounds now because she can’t hold her food down. I also have a cousin out in California who has just developed pancreatitis and they just removed a third of her pancreas last week. She will probably be diabetic too. She has also lost an exceptional amount of weight. I know the artificial pancreas is a technological breakthrough for some diabetics, but for my poor friends life is looking pretty bleak. They are doing some pancreas transplants now, but I don’t know what the prerequisites are for getting on the list to get that done. Anyway, I didn’t mean to sound like a party pooper, I just saw this post and immediately thought of two very good friends who don’t even have the positive outlook of us diabetics. I feel very bad for them. OK, I’m done,
Peace, Y’all

Wow! When will this come to market? I would prefer though, for the artificial pancreas to only cover basal insulin AND shut off for two hours when your BS hits 50 (assuming cgms was mega-accurate). I would like control over the bolus insulin. Cgms portion can tell you if you need to correct if you did not get enough insulin to cover your meals. When walking around and active, you will need control over how much IOB you need to cover. Since AP can’t factor in exercise, and since our basals are based on normal, everyday activity (which can change) I see the artificial pancreas as being great mainly for keeping blood sugars stable overnight. I would like to be alerted to high and low blood sugars and given the chance to control the information myself. Kristin is correct about disconnecting not being fast enough to correct a low, but I would like an extremely loud alarm when BS hits 50, and I would still like insulin discontinued for two hours anyway so when you are 50 you are not still getting even more insulin. No, not willing to turn over all care to an AP, but would appreciate a little more basal help at times when the AP can more safely do it (when you are at rest and there is no food in the system).

I would think that the way you would solve the problem of relying on the one system is to have redundancy in the checking of glucose levels. If you have one system checking against another and then the two comparing the results and then making the judgement call would be better. Currently from what I read is that most people check against a meter. If two system were in place and the comparison would occur between the two then you could try to get a more reliable result. that is the way that is done in software to maintain up time of servers. Sorry for getting geeky on this. But a fail safe would make me feel better. At least someone is working on a solution. cheers