Avandia, Vytorin, and now the ACCORD diabetes study: What's going on here?

Gretchen Becker Health Guide

    I've recently seen stories reporting that


    1. Avandia, used to lower blood glucose (BG) levels, which should reduce cardiac risk, increases cardiac risk.


    2. Vytorin, used to lower cholesterol levels, which should reduce cardiac risk, increases cardiac risk.


    3. Increased HDL levels in "diabetics" (not defined), which should reduce cardiac risk, increases cardiac risk.


    4. Increasing HDL with a drug, which should reduce cardiac risk, increases cardiac risk.


    5. Postmenopausal women taking calcium citrate supplements for bone health had twice as many heart attacks.


    And now the ACCORD study:

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


    6. Lowering BG levels close to normal increases cardiac risk.


    It seems that everything we've been told to take to increase our health has been shown in some study or other to do just the opposite.


    What are we to make of this situation?


    I think the first important lesson is not to panic. News media are hungry for stories, and they tend to rush from yesterday's scandal (Avandia and Vytorin) to today's scandal (normal BG levels), leaving the true facts about the earlier ones incompletely investigated.


    For example, some people criticized the Avandia studies because their meta-analyses had excluded all the studies in which there were no deaths. Does this make sense?


    Let's say you had 500 studies that showed that Avandia didn't increase heart attack rates because no one in either group had a heart attack and 2 studies that showed that 3 people taking Avandia had heart attacks and only 1 in the control group did.


    If you excluded all the first studies, you could then report that "People taking Avandia had 3 times as many heart attacks as those not taking the drug." But in fact, it might have been closer to 50,003 vs 50,001, a fairly insignificant difference. [I've exaggerated here to illustrate a point.]


    I think we still don't know if Avandia is really dangerous or not. The ACCORD study said they analyzed for this and found no effect. Other studies found an effect (one reason for the meta-analysis). But thousands of patients, worried by fear-mongering headlines and TV soundbites, rushed to their doctors to ask to be taken off the drug.


    A similar thing is happening with the ACCORD study. We really can't interpret it properly until we have more information. As Allison Goldfine, of Harvard and the Joslin Diabetes Center (who wrote the foreword to my book The First Year: Type 2 Diabetes), said in a blog in the Wall Street Journal, it would be premature to draw conclusions before the full results are published (which the researchers said should be in a few weeks). "We have not seen the data and it's very hard to form opinions," Goldfine said."


    Ironically, almost the same day the ACCORD study results that suggested that lowering BG levels too far was harmful appeared, the media also reported a Danish study (Steno-2) in which lower BG levels resulted in a 59% reduction in the relative risk of heart attacks.


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

    If you have access to the full text of this study, in the New England Journal of Medicine, you'll see that the A1c goals in the Steno-2 study were higher than those of ACCORD, so the two studies don't necessarily contradict each other.


    Nevertheless, the headlines reporting this study don't clarify: for example, "Intensive intervention benefits type 2 diabetics." Similarly, headlines reporting the ACCORD study, for example, Deaths Halt Part of Large Diabetes Trial, tend to frighten, by mentioning death instead of increasing cardiovascular risk.


    This is the second lesson to learn from all this: Headlines and TV news soundbites simplify and are often misleading.


    So when you hear news that could affect your health, don't panic. Try to find the source of the news, if you can. Read it critically. See if it applies to you. For example, if you're a female 60-year-old Hispanic office worker in a large city and the study was done on male 30-year-old Native American farm workers, some of the conclusions might not apply to you.


    Try to look at the statistics. Are the headlines claiming that "Glucosmiraculia kills diabetics" based on a study in which two more people taking Glucosmiraculia than those taking a dummy drug died? Remember that "twice as many people died with Glucosmiraculia as with placebo" could mean that two people in 100,000 died, compared with one person in 100,000. Even if the relative risk is double, your absolute risk would still be miniscule, and if you felt a lot better on the drug, it would make sense to keep taking it.


    We won't know what the ACCORD study results mean until the researchers release their full results. In the meantime, I'm sure not going to make a big effort to get my BG levels higher!

Published On: February 07, 2008