Type 2 Drug Avandia Linked to Increased Risk of Heart Attacks

An analysis of data from multiple clinical trials has shown that people with Type 2 diabetes who take the oral drug rosiglitazone (brand name Avandia) have a 43% increased risk of having a heart attack. The results of this study, known as a meta-analysis (which combines and analyzes data from other studies), was published in the online edition of The New England Journal of Medicine (NEJM) on May 21.


Rosiglitazone is a member of the class of oral diabetes drugs called thiazolidinediones, which makes the body’s tissues more sensitive to insulin and causes the liver to release less glucose into the bloodstream. In addition to being sold as the stand-alone drug Avandia, rosiglitazone is found in Avandamet (a combination of rosiglitazone and metformin) and Avandaryl (a combination of rosiglitazone and glimepiride, a sulfonylurea drug). It is estimated that 60 million prescriptions for rosiglitazone have been written since it was approved by the U.S. Food and Drug Administration (FDA) in 1999. About 1 million Americans and 2 million people worldwide currently take rosiglitazone.

The FDA issued a public safety alert saying that people taking rosiglitazone, especially if they have heart disease or are at high risk of having a heart attack, should talk to their doctors about whether or not the drug continues to be the best choice for them. People should not stop taking the drug without consulting their doctors.

The meta-analysis, which was performed by researchers at the Cleveland Clinic, looked at data from two large trials and several small trials covering a total of 15,560 people who took rosiglitazone and 12,283 who took either a placebo (inactive drug) or a different diabetes drug. All trials included in the meta-analysis lasted at least 24 weeks; the average age of participants was 56 and the average HbA1c level was 8.2%.

The analysis found that people who took rosiglitazone had a 43% higher chance of having a heart attack and a 64% higher chance of dying from cardiovascular causes (heart disease or stroke), though this second finding was of borderline significance. The increased risk seen with rosiglitazone was the same whether the drug was being compared to a placebo or another diabetes drug.

These findings are troublesome because having diabetes already raises a person’s risk of developing cardiovascular disease. In fact, 65% of people with diabetes eventually die from some form of cardiovascular disease.

Limitations on the study included a lack of access to the original source data that the earlier clinical trial conclusions were based on. Also, the total number of cardiovascular “events” documented in the amalgamation of trials was small (86 heart attacks and 39 cardiovascular deaths in participants who took rosiglitazone vs. 72 heart attacks and 22 cardiovascular deaths in participants in the control groups). In addition, the trials examined in this study were not originally designed to explore cardiovascular outcomes linked to taking rosiglitazone.

In a joint statement, the American College of Cardiology, American Diabetes Association, and American Heart Association said that “the overall level of the risk associated with rosiglitazone appears to be small, but nonetheless one that must be considered carefully” and that “Further research will be needed in this area to provide conclusive evidence.”

The mechanism by which rosiglitazone may cause increased cardiovascular risk is not clear at this time, although the drug has been found to raise LDL, or “bad,” cholesterol by as much as 18% in other studies. It is also known to be linked to fluid retention, which can lead to congestive heart failure. Both of these potential side effects are currently disclosed in the drug’s labeling.

GlaxoSmithKline, the pharmaceutical company that manufactures rosiglitazone, issued a statement saying that it “strongly disagrees with the conclusions reached in the NEJM article, which are based on incomplete evidence and a methodology that the author admits has significant limitations.” The company says that data it considers more reliable has shown that rosiglitazone has a similar cardiac risk profile to other diabetes drugs on the market.

The FDA is currently reanalyzing data about rosiglitazone and will soon convene an advisory panel to review the drug. Meanwhile, the organization has come under fire for failing to mandate a long-term study of rosiglitazone’s safety after its approval.

The only other approved drug in the thiazolidinedione class is pioglitazone (Actos), which was shown to reduce the risk of heart attack in a specifically-designed, large-scale trial called the PROactive study, published in the journal The Lancet in 2005. A similar study looking at rosiglitazone, called the RECORD trial, is currently under way, but results are not expected until 2009.

Learn more about the health and medical experts who who provide you with the cutting-edge resources, tools, news, and more on Diabetes Self-Management.
About Our Experts >>

  • Dr M Hanshaw, PharmD

    Keep in mind, Actos has the same limitations of use as does Avandia in patients with pre-existing heart/circulatory problems. I would expect that being in the same class as Avandia the same precautions would be considered. Cause and effect is very different from association. Removal of drugs such as Zelnorm and even Vioxx with out absolute certainty of cause as opposed to association removes life changing drugs from patients who really need them. Patient’s with pre-existing cardio-vascular problems are high risk patients, but lets prescribe drugs prudently and not put patients in harms way when it is avoidable.

  • Sandy Heise

    I went to my Doctor today and after reading the article on Avandia she took me off this med, as I was on it for the last 8 years. Thank you for all your information on this. Sandy Heise

  • Gain ed much weight

    My problems with advandia were numerous. I gained a lot of weight and my edema became much worse on this drug. After I was taken off of it, the edema is much better and I have lost some of the weight but have much more to lose. I also have much more energy. Has anyone else had this same experience? If so, I would love to hear from you.

  • deborah prettyman

    I also gained weight on Avandia but didn’t have edema problems. I stopped using it and was on byetta and actos. I recently (with Dr’s agreement) dropped the actos as well. I had major back surgery a year ago…prior to that I was on just byetta and losing weight. The surgery and recovery sent my sugar readings way off the chart so I spent the better part of the year on the three drugs, gaining wt (not being able to exercise didn’t help). With just the byetta, as long as I watch what I eat, my numbers are good and I feel so much better.

  • Amy Matthias

    I am VERY concerned about this and am glad that my sister sent this information to me. I am currently taking Avandamet (4mg/1000mg) twice a day. On August 26, 2008 I had a heart attack. I have been taking this medication for about 2 years now. They said the heart attack was from a blood clot. I am only 36 years old. Should I be concerned even more about the news I just read?? I will be discussing this with my cardiologist and diabetes doctors immediately.

  • Rashmi Bnasal

    I am one of the sufferers from Pioglit. It caused so much fluid to accumulate that I got a right heart block, developed Femoral Valve insufficiency and suffered like anything. Now that I am off it, my BG is fluctuating like anything. By the way I am a heart patient and had a bypass 21 years ago. Sadly, doctors don’t keep themselves updated. Medicines we believe are life savers, turn out to be killers. I feel traumatized after all this.

  • jim snell

    Actos/like avendia can be dangerous drugs.

    I was lucky to get off actos before it fully nailed my kidneys.

    My kidney Doctor wanted me off the actos asap as soon as kidney numbers started down hill.

    Unfortunately there does not seem to be long term studies of sufficient scope to check on the long term effects of a drug or on a wide enough sample of the population.