Last week, I had the happy opportunity to talk with Dr. Aaron Kowalski, head of the Juvenile Diabetes Foundation (JDRF ) Artificial Pancreas Project, and I am ten steps more optimistic about the future of diabetes care as a result.
The conversation was informal, and I was not prepared with official interview questions, so the best I can do in recounting now is to paraphrase a few of the notes I scribbled while sitting outside Starbucks on the phone. The notes are heavily filtered through my memory and a weekend of interpretive digestion, so if I get anything wrong, I apologize, and, please, let me know. In any case, I was so pleased by what I learned, I thought it worthwhile to share regardless.
As some of you know, the Artificial Pancreas Project was announced officially as a partnership between JDRF and Animas Corporation, a Johnson & Johnson Company, on January 13th, 2010. Research into creating an artificial pancreas has been going on for years, but the partnership, using the resources of both JDRF and its corporate allies, aims to fast-track the creation of an artificial pancreas solution for Type 1 diabetes.
Notably, the artificial pancreas, despite the name, will likely not take the form of an implanted, hidden, robot pancreas of my dreams; the first generations of the system will likely look very similar to the currently available insulin pump/continuous glucose monitor (CGM) setup, but it will be much smarter, relieving the diabetic patient of the burden of estimating blood glucose fluctuation and deciding on insulin infusions accordingly. This more intelligent insulin delivery system, though a far cry from a true replacement pancreas, would still be a huge step forward in the quality of care and the quality of life for Type 1 diabetics.
According to Dr. Kowalski, the project was conceived some four or five years ago, when JDRF was collaborating with researchers at Yale to study hypoglycemia prevention. The researchers, in collaboration with Medtronic, were developing an insulin pump that could detect hypoglycemia (too-low blood sugars) using a CGM, and could automatically stop insulin delivery as a result. Such a system, the JDRF saw, was the first step towards an artificial pancreas, which, when fully developed, would be able to monitor not just for hypoglycemia, but for any blood glucose variation, and would be able to act accordingly, infusing insulin or glucagon (a hormone that causes the liver to release stored glucose into the blood) as needed to maintain optimal glucose levels.
At the time, however, there were several large hurdles in the way of the development of an artificial pancreas. One set of hurdles was scientific– how can we monitor glucose levels with sufficient accuracy as to avoid fatal calculations of insulin doses? How do we diffuse insulin rapidly enough to effectively control post-meal glucose spikes? And so on.
But a whole other set of hurdles came from business concerns; despite how close even early algorithms were to reliable prediction of insulin requirements, the major medical manufacturers did not see enough immediate value in the promise of the artificial pancreas, and therefore weren’t allocating the necessary resources or funds to get over the scientific hurdles.
That’s where the JDRF came in. They had the money and the connections to pull together disparate areas of academic research, small medical device companies, and larger medical powerhouses like Animas. The hope was that, united by the direction of the JDRF, the corners of the industry could move the artificial pancreas from a dream to a product.
And so the Artificial Pancreas Project was born, with several discrete milestones designed to guide progress and ensure that the FDA would be on board at each interim point along the way. Roughly speaking, the JDRF has defined the steps toward a full-featured artificial pancreas as:
- Develop a system that can reliably shut off the insulin pump during periods of severe hypoglycemia.
- Develop a system that can predict hypoglycemia such that it can take preventative measures– like turning off insulin infusion and pumping in glucagon– before the wearer gets low.
- Develop a system that can monitor and predict both hypoglycemia and hyperglycemia, such that it regulates the flow of insulin and glucagon to maintain normal blood glucose levels.
Now, some of you who are paying close attention will here say, Wait a minute! They’ve been working on step 1 forever now! And indeed, that’s what Dean Kamen and the Yale researchers were working on way back at the beginning of this story. There have even been a number of recent studies showing how successful such systems prove in aiding diabetes control.
Truth is, scientifically and algorithmically speaking, they’re already more than halfway there on all three steps. So what’s the hold up? Why don’t I have an artificial pancreas yet? Well, as you might expect, even 95% of the way there leaves a dangerous gap when in comes to medical devices. The FDA is very wary of artificial pancreases, and not without reason; you can imagine the trouble that could result from a erroneous blood glucose detection and subsequent insulin infusion.
But the FDA’s inertia, according to Kowalski, is slowing down the process more than is warranted, especially when it comes to steps 1 and 2. The FDA’s main concern is that inaccurate glucose measurements will lead to undesirable insulin events, but in the case of systems designed only to prevent severe hypoglycemia, the worst that happens is insulin delivery is shut off for a period. Even if the measured glucose value was drastically off– say, showing a dangerously low 40 mg/dL instead of an actual value of 213 mg/dL– the patient could get higher, but would be able to measure on a separate glucose meter and respond before any serious, long-lasting damage could result from the period of hyperglycemia.
The inverse, however, is not so harmless. Here, I could hear the real concern Kowalski, himself a Type 1 diabetic, felt: he explained that just the other week, he had received a call informing him that a young diabetic– athletic, a teenager– had died in the night due to severe hypoglycemia. She was wearing a pump and a CGM; the glucose monitor was base-lining, as her blood sugar had gone so low as to be outside of the monitor’s ability to monitor it. Her pump, meanwhile, switched over to her higher scheduled morning basal rate, required by FDA regulations to remain oblivious to the glucose sensor’s readings. An insulin pump that could react to extreme hypoglycemia, already scientifically possible, would have saved her. And yet the FDA says, “No,” for fear of ketoacidosis– a much more unlikely scenario than the very real risk of fatal hypoglycemia.
But Dr. Kowalski and the JDRF are working with the FDA. They are explaining, showing, researching– making incremental movements towards an artificial pancreas for diabetics like me. And Dr. Kowalski. And his brother.
And meanwhile, the researchers and scientists aren’t sleeping, either. Dual insulin and glucagon pumps! Faster insulin diffusion using miniscule, painless needles that inject insulin subdermally instead of subcutaneously! And when I asked if they were monitoring various autonomic processes, like heart rate, that could give additional information when predicting blood sugar movement? You bet they are. But Dr. Kowalski one-upped me: not just heart rate monitors; heart rate monitors plus accelerometers to ensure the pump wearer is actually moving and exercising, not just scared or anxious.
And the coolest part? When I asked what the timeline was, Kowalski said, well, the project has a four to five year funding schedule. And the JDRF wants to have it figured out by then.
So, yeah. A lot of cool stuff going on. Thank you, JDRF, for putting your money where our hope is. And thank you Dr. Kowalski, for orchestrating the development and keeping everyone focused on better care for more people, sooner.
JDRF? Dr. Kowalski? You guys rock.
Edited on 04/20/2010, to correct a reference to Dean Kamen; Dean Kamen was not working with the Yale researchers, but had worked in the past with one of the JDRF affiliated Yale investigators, Dr. Bill Tamborlane.