Avandia Increases Heart Attack Risk? I'm Not Convinced!
This week, in a widely publicized study, which is actually an analysis of a group of studies called a meta-analysis, it was concluded that use of the diabetes drug rosiglitazone (Avandia) causes a significant increase in the risk of having a heart attack. It also concludes that there is a greater risk of death from all cardiovascular diseases, which includes heart attack and stroke, while taking the drug. But is this true, or is it a trick of the statisticians? Let's look at a few of the details of this "study."
First, the author took a group of 42 studies that included the use of rosiglitazone. These studies were not specifically designed to look at heart disease risk, so this begs the question: if the studies didn't control for other causes of heart attack and stroke risk, how did the author of the meta-analysis determine that there were no other causes of these increased risks in the study?
For example, poorly controlled diabetes and elevated blood lipids both increase the risk for heart attack and stroke, so how did the author control for those factors if the original studies did not? Additionally many of these trials were small and short-term, with an inadequate number of participants and too little duration to really determine the long-term risk of developing cardiovascular disease (CVD).
Next, the author took the "publicly disclosed data summaries" from these 42 trials and used that information to draw his conclusions. This means that the author is drawing his numbers from the best "spin" on the original data, not the original study numbers themselves.
Finally, there is the little problem of the statistician's spin on the data, which is reported from analysis of the "odds ratio" rather than the "relative risk" of a particular event happening. Statistical analysis makes my eyes glaze over, but if you want a better understanding of these two ways of interpreting the data, you can read a pretty coherent explanation from Steve Simon, PhD, of Children's Mercy Hospital in Kansas City.
From my view of the world (and that's why I write this blog, isn't it?) the relative risk is the number that expresses how I think about the chance that something will actually happen to one of the many individual patients for whom I have prescribed Avandia (which is almost all of the individuals living with diabetes that I have written prescriptions for over the last three years).
In this analysis it was described that there is a 43% increase in the risk of having a heart attack while taking rosiglitazone, and a 64% increase in the risk of dying from heart attack or stroke while taking this medication. These numbers are based on the odds ratio. But if you look at the actual numbers of patients who had a heart attack or died from CVD they look like this:
By comparison, the risk of heaving a heart attack while taking a diabetes drug other than rosiglitazone is 72/11,634 or 0.0061% and of death from all CVD causes is 22/11,634 or 0.0018%
Following this line of reasoning you will note that:
There is currently a study underway called the Rosiglitazone Evaluated For Cardiac Outcomes & Regulation of Glycaemia in Diabetes ("RECORD") which is going to give us additional data on the safety of rosiglitazone as measured by heart disease and stroke outcomes. Until that data is available, I am not rushing to remove my patients from rosiglitazone based on this spin on the "evidence."
Take care,
Laurie
Related Topics: Technorati Tags: Avandia, rosiglitazone, diabets, heart disease, health and wellness
First, the author took a group of 42 studies that included the use of rosiglitazone. These studies were not specifically designed to look at heart disease risk, so this begs the question: if the studies didn't control for other causes of heart attack and stroke risk, how did the author of the meta-analysis determine that there were no other causes of these increased risks in the study?
For example, poorly controlled diabetes and elevated blood lipids both increase the risk for heart attack and stroke, so how did the author control for those factors if the original studies did not? Additionally many of these trials were small and short-term, with an inadequate number of participants and too little duration to really determine the long-term risk of developing cardiovascular disease (CVD).
Next, the author took the "publicly disclosed data summaries" from these 42 trials and used that information to draw his conclusions. This means that the author is drawing his numbers from the best "spin" on the original data, not the original study numbers themselves.
Finally, there is the little problem of the statistician's spin on the data, which is reported from analysis of the "odds ratio" rather than the "relative risk" of a particular event happening. Statistical analysis makes my eyes glaze over, but if you want a better understanding of these two ways of interpreting the data, you can read a pretty coherent explanation from Steve Simon, PhD, of Children's Mercy Hospital in Kansas City.
From my view of the world (and that's why I write this blog, isn't it?) the relative risk is the number that expresses how I think about the chance that something will actually happen to one of the many individual patients for whom I have prescribed Avandia (which is almost all of the individuals living with diabetes that I have written prescriptions for over the last three years).
In this analysis it was described that there is a 43% increase in the risk of having a heart attack while taking rosiglitazone, and a 64% increase in the risk of dying from heart attack or stroke while taking this medication. These numbers are based on the odds ratio. But if you look at the actual numbers of patients who had a heart attack or died from CVD they look like this:
- Total numbers drawn from 42 studies:
- 15,560 individuals randomly assigned to medication regimens that included rosiglitazone.
- 12,283 individuals randomly assigned to comparable groups that did not include rosiglitazone.
- Of those taking rosiglitazone and completing the study there were 14,371 individuals, 86 of whom had heart attacks and 39 of whom died from either these attacks or from stroke.
- Of those not taking rosiglitazone and completing the study there were 11,634 individuals, 72 of whom had a heart attack and 22 of whom died from that attack or from stroke.
By comparison, the risk of heaving a heart attack while taking a diabetes drug other than rosiglitazone is 72/11,634 or 0.0061% and of death from all CVD causes is 22/11,634 or 0.0018%
Following this line of reasoning you will note that:
- the risk of having a heart attack on rosiglitazone is relatively small and
- the risk is nearly identical to the risk of having a heart attack while taking one of the other diabetes drugs (0.0059% versus 0.0061%).
There is currently a study underway called the Rosiglitazone Evaluated For Cardiac Outcomes & Regulation of Glycaemia in Diabetes ("RECORD") which is going to give us additional data on the safety of rosiglitazone as measured by heart disease and stroke outcomes. Until that data is available, I am not rushing to remove my patients from rosiglitazone based on this spin on the "evidence."
Take care,
Laurie
Related Topics: Technorati Tags: Avandia, rosiglitazone, diabets, heart disease, health and wellness
