Thursday, February 5, 2015

The question of an artificial pancreas

NPR ran a story the other day about advancements that are "inching us closer" to an artificial pancreas for diabetics.

Diabetes Technology Inches Closer To An Artificial Pancreas

Apparently this journalist hasn't been around long, because I've been seeing claims about this for probably the last eight years.

About eight years ago, I participated in a study that was comparing two different brands of continuous glucose monitors (CGMs), Dexcom and Navigator.  Even then, another brand already had an insulin pump that also had a CGM.  The CGM couldn't change your insulin dosages, though, because what if it was wrong?  You could die from letting a machine change something for you without your confirming it first.

CGMs aren't as accurate as the finger-prick kind of glucometer -- and even those aren't as accurate as you would think.  So letting a machine, which lacks the ability to think about the situation as a person would, control your insulin doses could be extremely dangerous.  What if you're running your sugar slightly high because you're just about to exercise, and the pump doesn't know this and gives you extra insulin?  Humalog takes about two hours to run its course, and since the body uses insulin more efficiently during and after exercise, you could crash if you work out in that time period.

This is why we still don't have an artificial pancreas, after all these years.  This is what the journalist who wrote that article didn't understand.  It's not that we couldn't have by now -- it's that it's dangerous.  It would take an incredibly sophisticated piece of equipment to perfectly simulate a human pancreas -- something that could accurately check blood sugar, release insulin in real time, know when you're working out, and release emergency glucose if you're crashing -- and I just don't see that level of sophistication possible anytime soon.

The media does love to misreport medical findings, though, judging on how many times they've "cured diabetes" since I was diagnosed in 2002.  I'm sure we'll see another article in a year or two about how we're "almost there!" on our quest to create a perfect artificial pancreas.

No comments: