Dr. Nissen is going to be on CBS Nightly News tonight. He gets to talk to Katie.
In truth, I have nothing against Dr. Nissen himself, and my complaints are not about the man, or about the study he did. My real complaints are against the practice of trying these issues in the public forum. The practice now makes interpretation of scientific study secondary to running a PR machine to promote or question the legitimacy of a scientific claim. While this becomes (in my opinion) a fight between those who hate the FDA and Pharma in the public forum and those who defend their practices, the private physicians such as myself are left to actually look at the studies and deal with the emotional aftermath of patients at whom the PR machines are aimed.
This whole situation reminds me greatly of that swirling around post-menopausal hormone therapy a few years back.
In 2002, headlines were made when the Women's Health Initiative, a study of the overall effects of post-menopausal hormone replacement therapy, was stopped prematurely. Researchers noted that there was a statistically significant negative effect of the use of hormones in both cardiovascular outcomes and rates of the development of certain forms of cancer. This hit the headlines and women, who were previously told to take hormones were told to come off of them.
The purpose of this study was to back up the common belief among physicians that post-menopausal hormone replacement therapy was beneficial - especially from a cardiovascular standpoint. While there were some studies prior to the WHI that suggested HRT may be harmful, the best evidence at the time was that women who were treated with hormones had a lower risk of having cardiovascular events.
The rationale behind this was the fact that women have about a 10-year delay in the onset of atherosclerosis (when compared to men), generally with onset after menopause. It was felt that the hormones, therefore, must be cardioprotective. The case for this was bolstered when several retrospective analyses came out suggesting that women who had been on hormones in the past were less likely to have heart problems. Even when other studies suggested the possibility of increased breast cancer risk, this was negated by the perceived benefit of hormones on the heart.
Prior to the WHI, it was nearly a dogma in medicine that women should get post-menopausal hormone replacement. The WHI set things on their ear when it not only failed to prove a benefit from hormones, it actually showed harm from a cardiovascular standpoint. Suddenly the drug companies who were pitching hormone therapy had to completely change their tune (Wyeth took a huge hit on this) and women were being removed en masse from HRT.
What does this have to do with Avandia? It shows the total unreliability of retrospective studies. Even in light of the WHI, doctors were so set on the facts surrounding HRT prior to the study that it took some quite a while to stop prescribing HRT for most menopausal women. The fact is, a well-designed prospective trial has far more power in proving (or disproving) something. In the case of the Women's Health Initiative, it totally contradicted the prevailing scientific opinion of the day.
Retrospective trials and meta-analyses are different from scientific studies. People who are performing them are far more doing statistics than they are doing true scientific research. In a retrospective analysis, the object is to look at existing evidence and analyze it for significant trends. It does no blinding and no study design of its own, instead it is limited by the quality of the research that others have done. Meta-analyses are a bit better in that they (hopefully) offset any bias or poor technique by using a large number of trials and "averaging the bias out." The major problem is that anyone looking at these numbers can do so with bias themselves, trying to show something to be true (or false) using statistics to back up their hypothesis. This is very weak science, and should always be regarded as such.
So is Dr. Nissen right? The only way to know is to wait for the prospective trial to come out. Those who are impatient and want to jump on his analysis have only to look back at the lessons from the past. Raising a flag of caution is probably a good thing to do at this point, but sounding the alarm and calling for the drug to be pulled is, from a historical standpoint, an over-reaction.