Thursday, July 10, 2014

Roadblocks to the Artificial Pancreas

Attending the ADA Scientific Session was much like drinking from a fire hose.  Information is thrown at you at lightening speed, and before your brain has time to decipher the meaning behind one fact, another comes hurling pretentiously, perhaps even unapologetically, towards you. 

With my limited smarticals, I must admit, much of the terminology went over my head.  I’m thankful for the people that sat near me, who were more than willing to decipher the medical lingo into laymen’s terms.

There was a wealth of information on current technology, and a lot of much deserved buzz around the Artificial Pancreas.  Which begged the question: Within the confines of current technology, are we ready for the Artificial Pancreas?

As far as my swelly brain could translate, here are some of the existing roadblocks:

The current CGM’s seem to be accurate enough, more to the point, the Dexcom sensors are accurate enough.  But the problem lies in the fact that Day One of a sensor is significantly less accurate than Day Two.  And Day Two is less accurate than Day Three.  It takes at least three days for the sensor to reach optimal accuracy for the AP device.  That gap is significant considering the life of the infusion set.

The current AP Devices I’ve seen need three sites.  Insulin.  Glucagon. CGM. Real Estate is not just a problem for children; adults struggle to find prime set areas as well.  Thankfully, there are researchers trying to find a way to integrate the sites.  Medtronic just released an all-in-one sensor in Europe for its pump/CGM combo.  Exciting!  But how long can the set feasibly last?  Maybe five days?  Studies show that five days is a possibility for cannula life and insulin absorbability…but then we’re back to the it-takes-three-days-for-the-best-CGM-accuracy problem.  If less than half of the wear time is prime time, what is the bonus in that?  Even so, my understanding is the all-in-one Medtronic set still uses two needles, it’s just simply all done in one action, under one sticker.

There are studies being done on cannulas, to see what the best material is to use so the site can last longer.  The problem is we are sending our bodily fluids up into that cannula just as much as it’s spitting insulin out.  (Interesting side note: studies show that there is no difference between Humalog and Novalog in the role it plays in cannula life.)

But I think one of the biggest questions is: Are the current insulins fast enough to do the job of a working pancreas?  Turns out people without diabetes release insulin into their system BEFORE they even eat.  If I remember right, it was roughly 5 minutes before to a couple minutes after we eat.  It is only a small percentage of the total insulin produced, but this small natural pre-bolus has shown to make a huge difference. (Which is why studies have shown pre-bolusing 15 min before is optimal.)  Can algorithms be adopted to fix this?  It’s possible.  The algorithms used already are completely mind-boggling. Thankfully, faster acting insulins are coming down the pipeline. 

Lastly is glucagon itself.  I spoke with a couple friends at the Friends for Life Conference this week who were part of an AP trial.  Both of them experienced stomach upset from the glucagon, and in a couple cases…it was significant.  Thankfully this information has only helped the developers create more finite dosing to alleviate this issue.  Also, glucagon is only stable for ONE Day.  It needs to be changed out, every day.  That’s a big hurdle, but one that no doubt will be figured out.  I keep thinking we walked on the moon, my iPhone is practically magic…us humans can create a better glucagon.  It's also important to note that some believe glucagon isn't necessary for a successful Artificial Pancreas system, so this entire paragraph may be a moot point.

Whether the technology is optimal or not, the artificial pancreas is coming, and it is quite remarkable.  I don’t know if we’re going to be able to “set it and forget it” for a while, or if anyone is going to be able to afford it, but we all know that small changes allow for larger measures of freedom. 

The Artificial pancreas isn’t going to bring small changes, it’s going to bring big ones.

And that, is exciting for sure.


  1. A lot of the things I have concerns with you mentioned. Its not just 3 separate sites its also 3 separate devices. My son is already a bit self conscious about people seeing his pump I can imagine getting him to carry around more of them. I know that the AP is in the beginning processes and from there can only advance more, I'm just excited that there is progress being made to help out type ones live easier lives

  2. Meri - As exciting as the development of an artificial pancreas is, I am much more interested in where things are at terms of work towards a cure or a simple treatment that doesn't involve technology. Here are two that show some promise - Dr. Faustman's work at Mass General: and Dr. Accili at Columbia University's Diabetes Center: Were things like this also discussed at the conference?

    Don't get me wrong - if an AP was available for purchase tomorrow, I would be in line for my son and myself. I am just more interested at this point in time in something that would be a real leap forward and I don't see those things getting the same play as the high tech/high cost solutions like an AP. I already feel like a bionic woman with my OmniPod and Dexcom plastered on me and I wish for something that doesn't involve more hardware.

    Anyway - THANK YOU for sharing what you learned at the conference and your thoughts going forward. You make a difference in many lives every day with your thoughtful posts.

  3. I've been trying to write a blog post about this very topic for a while. My friends keep asking me if I've heard about this awesome new thing lol. You put it into much better words than I ever could! Thanks!

  4. Meri, you make several great points. I also think: If we still have trouble with bent cannulas, occlusions, skunky insulin, etc., an AP isn't going to help. But it's better than what we have today. Imagine where glucose meters were 30 years ago! Thanks for the thoughtful post.


Moderation now enabled, so comments will not immediately be seen.