diabetes book
Diabetes Testing

Glycemic Variability

Now I know why so many diabetes specialists hold Irl Hirsch in such high esteem. Dr. Hirsch knows diabetes. He is an endocrinologist who has had diabetes since he was 6 years old. He is also a professor at the University of Washington Medical Center in Seattle and directs its Diabetes Care Center.

I hadn’t heard him speak before I caught his talk on “Understanding the Ups and Downs of Blood Sugar” at the Children with Diabetes conference that I attended a few days ago at Disney’s Coronado Springs Resort near Orlando, Florida. Along with Michael Brownlee, professor of diabetes research at Albert Einstein College of Medicine of Yeshiva University, and Louis Monnier, professor of endocrinology at France’s University of Montpellier, Dr. Hirsch has emerged as a leader in the new field of glycemic variability.
These three leaders go beyond considering the A1C number. “Now we talk about the quality of the A1C,” Dr. Hirsch told the Children with Diabetes conference. A higher quality A1C is one where our levels are more steady. When our blood glucose is high, we have greater oxidative stress, which comes from those dangerous free radicals.

“It is suggestive — but not conclusive — that glycemic variability turns on oxidative stress,” he says. To me this seems pretty obvious, but the experts are still arguing about it.

So how do we measure glycemic variability? The only way is to download a number of our blood glucose readings to our meter software. Then the software can use a common statistical tool called the standard deviation, or SD for short, to determine how great the variability is.

The target variability that Dr. Hirsch sets is that the standard deviation multiplied by two needs to be less than the mean (or average). Ideally, he says we should multiply by three, but this is difficult for anyone with type 1 diabetes.

The first caveat of using the standard deviation is that you need to have sufficient meter data. But in any case the standard deviation is actually not a very good measure, he says. If the average blood glucose level, the sort of thing that the A1C measures, is very high or very low, his way of determining the target is “irrellevant.” That sad note is where Dr. Hirsch left it in a PowerPoint presentation (Keynote on the Mac) that he made for the Children with Diabetes convention.

But what obviously excited Dr. Hirsch and his audience, myself included, was his announcement new software just became available that incorporates a much better tool for calculating glycemic variability than the standard deviation. It uses new low, high, and combined blood glucose indices developed by Dr. Boris Kovatchev, associate director of the University of Virginia Center for Biomathematical Technology. Last year Dr. Kovatchev and his associates prepared a quite technical explanation for the journal Diabetes.
The software that incorporates this better tool to calculate glycemic variability is so new that I haven’t been able to get it yet. Roche developed it for its Accu-Chek meters. They call it the Accu-Chek Smart Pix Device Reader.

The Roche booth at the Children with Diabetes convention did have a shiny brochure about it. The device transfers data from some Accu-Chek blood glucose meters (including the Active, Advantage, Aviva, and Compact meters) or its Spirit insulin pump to a USB port of a PC.

“This novel software for the first time separates it from the rest,” Dr. Hirsch said. That’s a recommendation I can’t ignore. I have asked Roche to send me a device for testing, and I will provide my review of it here.

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

Never Miss An Update

Subscribe to my free newsletter “Diabetes Update”

I send out my newsletter on first of every month. It covers new articles and columns that I have written and important developments in diabetes generally that you may have missed.

Previous Post Next Post

You Might Also Like These Articles

  • Michael Gray at

    Mr. Mendosa,
    Thank you for your very informative articles. As a ‘newby’ to self controlling diabetes I have gleaned valuable information from your articles that has lead me to ‘normal’ A1C and average BG numbers. Recently, I came across your 2007 article on Glycemic Variability and Dr. Irl Hirsch. I calculated my variability via Dr. Hirsch’s SD formula and would appreciate some clarification, if possible, on the target variability. You state in your article that Dr. Hirsch wanted to have a target GV less than the mean. Even at that level it seems to be a large number. I have read that Bernstein prefers variability of pre and post prandial numbers to be less than 20 points (mg/dl). Is this a good overall variability using Dr. Hirsch’s formula. If you have any further information or can point me to more information I would greatly appreciate it.
    Thank you,
    Michael Gray

    • David Mendosa at

      Dear Michael,

      Good question! So good that I can’t answer it and have written Dr. Hirsch in hopes that he will respond directly to you.

      Best regards,


  • Stacey at

    There are some wonderful diet guidelines and great recipes for diabetics at http://www.lowglycemicrecipes.net. They list carb amount, calories, fat, fiber content, Glycemic Index (GI), Glyco Load (GL). I haven’t seen any other sites that offer the GI and GL together in their recipes. Good site if you’re looking for new ideas and good info.

  • Weaver at

    I can’t believe no comments have been left here! I have been keeping averages, standard deviations and A1cs for 4 years and have wondered about software for this. I have kept my numbers in a spreadsheet but I would love to hear about other software. I keep track of my carbohydrate intake and insulin doses but I would like to be able to track my blood sugars following each type of food I have eaten to see in the carbs/food item might be a problem. Maybe the problem is glycemic? Thanks,