JDRF Partners with Animas, DexCom to Build "First-Generation" Artificial Pancreas

Reprinted with permission from Amy Tenderich of DiabetesMine.com

Very big news in the diabetes world today, Folks: the Juvenile Diabetes Research Foundation (JDRF) has announced a partnership with insulin pump makers Animas Corp. (a Johnson & Johnson company) to actually start building the first ready-for-market artificial pancreas, i.e. “a fully automated system to dispense insulin … based on real-time changes in blood sugar levels.” DexCom will provide the continuous glucose monitoring (CGM) technology for the new system being developed.

I was fortunate to have a direct line yesterday to JDRF officers William Ahearn, Vice President, Strategic Communications; and Aaron Kowalski, Research Director of the organization’s Artificial Pancreas Project, to answer many questions about this announcement.

My first question was pretty simple: Hasn’t JDRF already been working with these companies behind closed doors for a while now on tying these technologies together? What’s really happening that’s new here?

“The work we’ve been doing so far is proof-of-concept: proving that CGM can improve patients’ outcomes, proving that algorithms can automate insulin delivery, proving that computer systems can pull it all together for improved control. But today’s announcement is the first step in creating real products that will end up in the hands of people,” said William Ahearn, when I spoke with him yesterday.

“This is more than the Wright Brothers’ dream of creating an airplane; this is developing an F-14,” he added.

According to the announcement, JDRF will provide $8 million in funding over the next three years for this project, with a target of having a first-generation system ready for FDA review within the next four years (give or take).

“One reason we’re so focused on programs like this is the applicability not only to children, but to adults living with Type 1 diabetes — they want things that can help them live better in the near-term,” Ahearn says.

OK, so what kind of life improvements can we actually expect from a first-generation automated pancreas?

The official word is that this first-generation system will be “partially automated,” described as follows:

An insulin pump (will be) connected wirelessly with a continuous glucose monitor (CGM). The CGM continuously reads glucose levels through a sensor with a hair-thin sensor wire inserted just below the skin, typically on the abdomen. The sensor would transmit those readings to the insulin pump, which delivers insulin through a small tube or patch* on the body. The pump would house a sophisticated computer program that will address safety concerns during the day and night, helping prevent hypoglycemia and extreme hyperglycemia. It would slow or stop insulin delivery if it detected blood sugar was going too low and would increase insulin delivery if blood sugar was too high. For example, the system would automatically discontinue insulin delivery to help prevent hypoglycemia, and then automatically resume insulin delivery based on a specific time interval (i.e., 2 hours) and/or glucose concentration. It will also automatically increase insulin delivery to reduce the amount of time spent in the hyperglycemic range and return to a pre-set basal rate once glucose concentrations have returned to acceptable levels. In this early version … the patient would still need to manually instruct the pump to deliver insulin at times (i.e. around meals). But this ‘hypoglycemia-hyperglycemia minimizer’ system would represent a significant step forward in diabetes management, and could provide immediate benefits in terms of blood sugar control, by minimizing dangerous highs and lows.

From this description, it sounds like Animas is working on some version of a patch pump (tubeless) to rival the OmniPod system. Could this possibly explain why Medtronic was not selected as the partner of choice here, despite already having an integrated insulin pump/CGM system on the market (Medtronic just announced yet another year’s delay of its planned patch pump, originally slated to be ready for FDA evaluation in 2010)?

“This is not an exclusive arrangement by any stretch. We hope to work with multiple partners,” Ahearn says, when queried on the topic.

Aaron Kowalski explains: “DexCom and Animas are already working on a combo product that’s supposed to launch this year. It will look a lot like the Medtronic integrated system, with a traditional pump and CGM tied together. Our thinking was, could we start to build some automation into this system… to help make an impact on limiting diabetic complications? I think we can.”

How is it possible to even think about a marketable “closed-loop” system, when today’s CGMs aren’t even deemed accurate enough to determine insulin dosing without doing a fingerstick to double-check? (they are currently approved by the FDA only as “adjunctive therapy.”)

“That’s the million-dollar question,” Kowalski says. “The sensors aren’t perfect, but we’re not shooting for perfection, we’re shooting to shave off the time that people spend in the very high and low range, and get them to spend more time safely in the middle.”

According to JDRF studies, he notes, the average person with diabetes spends over 12 hours a day at BG levels over 180 (!), and that’s with an A1C of 7.8. By developing algorithms set to a target range, rather than some arbitrary target number, they hope to significantly improve patient’s well-being and their long-term outcomes.

“The people struggling most, often kids and teens, are spending a tremendous amount of time with too-high blood sugar. Our ‘treat-to-target range’ aims to provide a buffer away from the hypos and hypers.”

Part of that is the automatic shut-off if your BG drops too low overnight, for example. “If you’re at 40 and you’re asleep, do you really want your pump to continue delivering insulin?”

Of course there’s the concern that the pump will shut off unnecessarily, causing your BG to spike too high. But the JDRF has enough data from previous trials to assess how often that really happens, and set the algorithms to avoid known error points, Kowalski says.

OK, I have no doubt that an automated system minimizing BG highs and lows could be an enormous help. But will it be liveable? Aesthetics freak that I am, I wanted to know: What might this new system actually look like (and feel like, 24/7)? Do prototypes of the described system already exist, from earlier clinical trials?

Apparently not yet, beyond a primitive pump/CGM connection via a bunch of clunky tech paraphernalia. That’s what this announcement is all about, according to Ahearn.

“We’ve shown that we can do it (run a viable AP system) in hospital settings, with a lot of cables and PC monitors running. Now we’re going to take things from the concept phase and make them work in the real world. We’ll be working on miniaturization, making the system liveable, how to get it reimbursed by insurance carriers, and clinical pathways for FDA approval,” he says.

More good news is that the FDA has already committed to taking a proactive role in developing an artificial pancreas.

Clinical trials for this new system could begin as early as 8 months from now, Kowalski says. Anyone interested in submitting their name for possible inclusion in these studies can do so here.

As Ahearn notes, “this is science and nothing is a given.” Who knows what obstacles this project may encounter? Nevertheless, I think this announcement makes today a pretty happy day for people living with Type 1 diabetes!