Misdosing Insulin

We have known for several years that we often miscode our blood glucose meters. Now we have scientific proof that this miscoding can cause big errors in insulin doses with potentially serious health complications.

I wrote about the first meters that didn’t require coding in my Diabetes Health column for October 2004 and on my website. At that time, Bayer HealthCare’s Ascentia Contour and Ascensia Breeze and Roche’s Accu-Chek Compact were the only blood glucose meters that we didn’t have to code or calibrate.

Since that article a year and a half ago, the Prodigy AutoCode meter, made by Diagnostic Devices, has appeared. And the Compact Plus, which includes a detachable lancet as well as a drum of 17 test strips, is also now available.

Coding is the process where we match each new box of test strips to our meter. We do it either when we insert a code strip or code chip into the meter or when we enter a code number into the meter.

The new Ascensia Contour, released at the beginning of this month, now takes just 5 seconds to test a 0.6 microliter drop of blood, down from the previous Contour, which took three times as long. The new Ascensia Breeze2, which similar the Compact and Compact Plus has an internal 10-test disc, takes just 5 seconds, but needs 1 microliter of blood.

The folks at Bayer HealthCare send me their new meters, which I was glad to use without coding. But in a sense I calibrated them, that is I checked their results against another meter that I use, the Accu-Chek Aviva.

This morning my fasting blood glucose was 97 mg/dl (5.4 mmol/L), and I wondered how much the level would rise two hours after a breakfast of one-third of a cup of naked barley (with my usual plain soymilk, flax seeds, nutritional yeast, and cinnamon).Two tests with the Aviva were exactly the same, 126 mg/dl (7 mmol/L). The Contour reported 129 (7.2 mmol/L), and the Breeze2 said my level was 120 (6.7 mmol/L). Nothing to complain about there.

The Contour and the Aviva have the same time and blood volume specs. I can’t decide yet which one of these meters I will continue to use, so I guess that I will have to use both of them for a whileß. I’ll take a pass on the Breeze2, because of it takes a bigger blood sample and because I don’t appreciate its larger size and complexity.

The scientific proof that those meter that require coding can cause significant errors in insulin doses came out in the March 2007 issue of the Journal of Diabetes Science and Technology. The article has seven authors (how can seven people write together?). The lead author is Charles H. Raine III of the Diabetes Control Center in Orangeburg, South Carolina. Three other authors work for Bayer HealthCare.

Spokespeople for Bayer HealthCaresend me a copy of the article as well as the meters. The abstract of the article is online.

Blood glucose values that the 116 study participants got from their miscoded meters showed an average error ranging from plus 29 percent to minus 37 percent. “The autocoded meters used in this study gave BG values that results in the lowest risk of insulin dose error,” was to me the most significant result of the study. I know that if I used insulin, I wouldn’t hesitate for a moment before switching to one of these new meters.

This article is based on an earlier version of my article published by HealthCentral.