Aiming for a Normal A1c

Gretchen Becker Health Guide
  • "My doctor told me that my A1c was too low. He said I should get it over 7 to reduce my risk of heart disease."


    Too many patients who have worked really hard to get their hemoglobin A1c levels into normal ranges have been reporting experiences like that.


    Why? Why would a doctor argue for higher A1c levels when study after study has shown the benefits of lower A1c levels, even in people who don't test diabetic?


    The answer is the ACCORD study, which was stopped early because they found that elderly patients with longstanding type 2 diabetes had greater mortality when they were randomized to the "intensive treatment" group, aiming for an A1c below 6 (normal range) than those aiming for an A1c between 7 and 7.9. There were many hypotheses about why this counterintuitive result occurred.

    Add This Infographic to Your Website or Blog With This Code:


    Now a new analysis has come up with a very different interpretation.


    In the intensive-treatment group, those patients who were able to achieve the goals had lower mortality. It was the patients who couldn't achieve lower A1cs despite intensive treatment who had greater mortality, and of course this affected the average results of the entire group.


    I've discussed the details of this new interpretation, with many links, here.


    One caveat with the new analysis is that it's not considered good statistics to separate patients into new subgroups after you've seen the results. This is because it would be easy to do so to come up with the result you wanted.


    Let's say you found that your favorite drug had no effect. You could mine the data until you found some subgroup in which the drug did work, let's say left-handed women with auburn hair and husbands named Joe. Then you could claim that your drug was effective in at least one group of patients.


    Most of us patients aren't experts in statistics, and it's difficult to know which reported results we should believe. But when the results are logical, as well as consistent with the results of other studies, as they are in this new analysis, I think we should trust them even if the analysis doesn't satisfy all the requirements of formal statistics.


    Don't let anyone tell you to get your A1c higher to reduce your risks of heart attacks!



Published On: May 14, 2010