Here we go again......
A piece recently published in the New York Times and cited by others has just kicked up the Avandia controversy again. As usual, this is both unnecessary, and will likely scare patients and cause more harm than it is trying to prevent. (In many ways like the current LABA scare).
I have blogged extensively on Avandia safety.
The Avandia Scare: Why it Matters, Who's Responsible, and What to Do ,For the RECORD, Avandia does not cause heart attacks, Avandia Vindicated, and The Diabetes Conspiracy.
Before going diving into what the hubub is about it is important to note two extremely important facts:
1. There is no new data or information here. Everything being discussed is old news.
2. The one thing NOT being discussed in all these reports is that the question of Avandia safety was answered this past July at the American Diabetes Association's annual meeting when the RECORD trial was presented which definitively showed that Avandia did not cause heart hospitalizations, cardiac deaths, or any heart problems.
What's this all about?
The New York Times got a hold of a report that is now public from Senators Grassley and Baucus. What is available is the Press release and the full report of their two year investigation into the FDA and GSK's handling of Avandia safety. Their letter to the FDA states that GSK knew that that there were cardiac risks associated with Avandia and did not make these risks known to the public or the FDA soon enough. The second charge they make is that the study the FDA has requested GSK conduct to test safety differences between their product Avandia and their competitor's (Takeda) product Actos called the TIDE study is unethical, because two FDA safety officers felt that since there was no benefit of Avandia over Actos, and that Avandia had heart concerns that Actos did not, the study would needlessly harm patients. One of those safety officers was Dr. David Graham.
The full report is a 342 page document that includes publications, FDA data and internal communications at GSK. One of the main focuses of their concern is that the RECORD trial was continuing when GSK knew that the initial heart attack rates were low, so low in fact that it was unlikely that when finished, the RECORD trial would have enough events (statistical power) to show whether or not Avandia was safe. What is not highlighted in the report (which comes out 6 months after RECORD was presented) is that RECORD actually did meet criteria for statistical power. In other words, all their worry was for naught. Another focus of the report was study 211, which as a study on using Avandia in patients with heart failure. (More on that later).
You can argue whether or not the Senators have a point regarding what the FDA and GSK should have done and by when; however, from a clinical perspective there is nothing new in this report. More importantly, it was irresponsible of them not to discuss the final results of RECORD in this context.
Study 211 was actually published in the well respsected Journal of the American College of Cardiology. It showed that when giving Avandia to patients with class 1 and 2 heart failure (a relative contraindication) that Avandia did not statisctically increase the rates of heart failure, death or heart attack. The only differences seen were more edema (a known side effect of Avandia, especially for patients with heart failure) and more need for medications. The essential findings of the study is that Avandia increases fluid retention (we know this, Actos does this as well) but does not actually have any effects on the heart. Why the senators want to make a big deal of study 211 is beyond me. If anything it shows that even in patients who you shouldn't give the drug to, there were no real problems.
What about the recommendation to take Actos instead of Avandia?
This is probably the scariest and most harmful outcome of the report and the media attention surrounding it. Every media outlet has not publisehd the following quote: "if every diabetic now taking Avandia were instead given a similar pill named Actos, about 500 heart attacks and 300 cases of heart failure would be averted every month because Avandia can hurt the heart"
This statistic comes from a letter written by Dr. David Graham to the FDA expressing concerns about Avandia over Actos. What Dr. Graham did is compared two Avandia meta-analysis with one meta-analysis on Actos. The two Avandia analysis used were Nissen's and the Singh meta-analysis published in JAMA. The Singh meta-analysis looked at ADOPT, RECORD and DREAM (which the FDA did a meta-analysis on and showed NO PROBLEMS with Avandia), and added study 211, which is a focus of the Grassley report. As mentioned, study 211 was a study done in 200 patients with heart failure, and since we know that Avandia does in fact increase the risk of heart failure, the patients in this study are not ones that you would normally give Avandia. Nonetheless, the Singh meta-analysis, unlike the FDA's, showed problems with Avandia, consistent with Nissen. ADPOT, DREAM and RECORD had thousands of patients. Study 211 had only 200. In other words, the only reason the Singh meta-analysis was positive and the FDA's was negative is that Dr. Singh added the 200 patients from study 211.
The Actos meta-analysis was published in JAMA, and found no heart attacks with Actos. One needs to note that one of the authors on the Actos meta-analysis is Dr. Nissen. Also, one should note that Takeda, the makers of Actos, gave Dr. Nissen and colleagues $25,000 in funding. In other words, Dr. Nissen has two published meta-analyses, one that is not funded that finds heart attacks in Avandia, and one funded by the makers of Actos that finds no heart attacks in Actos. As Arsenio Hall used to say, this is something that makes you go "hmmmmmmm."
Dr. Graham's conclusion is that comparing the Actos and Avandia meta-analyses, there were excess heart attacks and heart failures in Avandia over Actos, so that for every 63 patients taking Avandia instead of Actos for 1 year, there would be an additional heart attack. He then looked at the number of patients taking Avandia and came up with the 500/300 cases.
In addition to jumping to major conclusions from meta-analyses, one of the bigger problems here is comparing results from separate studies. This is breaking a cardinal rule of evidence based medicine. Doctors should be familiar with this because drug company reps do this all the time. They say, "my drug's studies show that my drug does X and my competitor's studies only show that their drug does Y, so you should use my drug." The astute clinician should politely ask the drug rep about head to head studies. Most often, the response will be that there are none. It is critical when making comparisons between to therapies, that head to head studies are done. Unfortunately, head to head comparisons are often not done by the drug companies because there is a huge financial risk if the results show that your company's drug is inferior. This is why proposed legistlation is rightfully calling for funding of comparative effectiveness research. In other words, Dr. Graham is stating that Actos is safer then Avandia WHEN THERE ARE NO HEAD TO HEAD STUDIES COMPARING THE SAFETY OF BOTH DRUGS. Unfortunately, this incorrect and inflammatory statement is what gets picked up by all the media. To be clear, I don't believe that Actos is harmful. I think both drugs are safe, though they both show similar side effects (fluid retention, edema and osteoporosis) and should be used with caution in certain patients (HCF).
A Diabetes Conspiracy?
This report if anything give more credibility to my postulations.
- It was Grassely and Baucus that called for congressional hearings back in May 2007, the day of the Nissen publication was announced. It was as if they new about the controversy before the story broke. The fact that they continue to draw attention to this issue (and themselves) despite the fact that the issue has been scientifically resolved and that health care reform is on life support is a concern.
- We again hear the name of Dr. David Graham, the FDA insider on Vioxx. I contend that somehow Nissen was tipped off from someone inside the FDA to publish the Avandia data himself, since the FDA decided to keep this data from the public. After this report, Graham continues to be my likely suspect. Regardless, if anything should be looked into, it should be whether or not classified FDA documents or conversations were leaked to Dr. Nissen. I am not a lawyer, but I think this is illegal.
1. Nothing new here. Avandia does not cause heart attacks, as proven by the RECORD trial.
2. Quote on Actos safety over Avandia is based on no truly comparative (head to head) studies, and thus should not bebelieved, especially given #1.
3. Don't the Senators have other things to worry about....like fixing health care?
4. Someone should take a closer look at how Nissen "discovered' the link between Avandia and heart attacks. This is the congressional hearing I would like to see.