Are Insulin Pumps Better Than Shots? (Continued)


Studying The Studies

But looking closely at the various pump vs. shots studies, what could explain the wide variety in outcomes?

Well, the studies varied a lot in their design. Clinic trials come in 31 flavors: prospective, retrospective, parallel, and more. Some had two study arms, some were cross-over studies where the arms switched in the middle (trade partners allemande left, doh-see-doh). Some looked at bad diabetics, some looked at good diabetics…


Oh. I'm sorry. I'm told I'm supposed to say some studies looked at in-control persons with diabetes and some looked at sub-optimally controlled persons with diabetes.

Most of the studies took place in a single location while only a few were multi-center. Different studies enrolled different dFolks. Some looked at dFolk who had never been on pumps before. Some took pumpers and moved them to shots.  Most of the studies were on T1 adults, but some were on kids, and some were on T2s. Most published studies took place in various locations overseas. With all these differences, is it any wonder that the results would be all over the map?

In point of fact, there was only one thing that all of the studies seemed to have in common: They had small enrollment and were short in duration. Painfully so. Most pump vs. shot studies looked at only 20-40 patients for six months.

Now, beyond individual studies, there are two large "meta analyses" of pump studies; one from 2003, crunching the data from 52 different pump vs. shot studies, and the other from 2008, crunching the data from 22 studies. Both support pumps, but again they're aggregating the data of very small, very short studies. Still, like a thousand pin-pricks, these little studies have had the cumulative effect of showing that the preponderance of the science supports the fact that pumps improve health outcomes, at least for type 1s. That's why the feds and most insurance companies will grudgingly cover pumps for us T1s, at least after a protracted fight.

What about those two studies that showed pumps were worse than shots? One was a study by Boli et al published in 2009 in the prestigious journal Diabetes Care, that found that there was no difference in diabetes control between pump and MDI users. While trumpeted as a multi-center research project, this Italian study followed only 43 type 1 patients for 24 weeks. Actually, a year earlier Simon et al in Pennsylvania found that folks on shots actually did better than folks on pumps. Consider, however, that they looked at a whopping 16 patients in total, all well controlled. So they basically compared 8 organic apples to 8 organic oranges.

I suppose I should inject (pardon the pun) a side note here about the often misrepresented 1998 Veteran's Administration study of pumps. This study actually focused on reducing cardiovascular risk factors rather than on reducing blood sugar—in fact, the blood sugar levels of the two groups were matched by design—and it didn't actually use the kind of pumps we are talking about today. It looked at shots vs. implantable pumps; the uncommon kind of pumps that are surgically stitched up inside of you. Yikes! How many people signed up for that study? Answer: 59.

For what it's worth, the largest study I could find on pumps is fairly new, the results of which were announced at the 70th Scientific Sessions of the ADA in 2010. It followed 329 adults and 156 kids moved from MDI to sensor-enhanced pumps (pumps with built-in continuous glucose monitors, a.k.a. CGM) and saw a slam dunk one-year drop in A1C from an average starting-gate score of 8.3 to a finish line score of 7.5. Oh, the folks on the MDI arm of the study crossed the finish line with a mild improvement, too. From 8.3 down to 8.1. Why didn't this end the discussion? Because yet a new element was introduced: the sensor. And there's really no way to know if it was the continuous glucose monitor or the pump that was the magic bullet here. Or the combination of both. Now instead of comparing apples to oranges, we're comparing apples to almonds. Oh, and if you thought that pumps were waaaaaaaaaay more expensive than shots, just wait until you see what happens to the price tag when you add CGM to a pump system.

But what we lacked here, and really needed, was a third and a fourth study arm: CGM and syringe; and a CGM-less pump. That might have let us sort out what the key benefit was that bought us such an amazing one-year drop in A1C. Was it the pump? Was it the CGM? Or does it need to be a CGM-enhanced pump? Until that's cleared up, the sharks aren't going to be getting out their checkbooks. In any case, sign me up for the sensor-enhanced pump arm, please.

Page: 1 | 2 | 3 | 4

Last Modified Date: June 17, 2013

All content on is created and reviewed in compliance with our editorial policy.

More on this Topic

No items are associated with this tag

Sign up for FREE dLife Newsletters

dLife Membership is FREE! Get exclusive access, free recipes, newsletters, savings, and much more! FPO

You are subscribed!
You are subscribed!
You are subscribed!
2341 Views 0 comments
by Brenda Bell
Just as years ago, the community of people living with diabetes pushed for the adjective describing us to be changed from "diabetic" to "person with diabetes", we are in the throes of another surge in Political Correctness: calling the action of monitoring our current blood glucose levels "checking" rather than "testing". Frankly, I think this is a Very Bad Idea. The argument behind the change in terms is that "testing" suggests...
  • Watch dLifeTV online now!

    Click here for more info