Controversy about Avandia (rosiglitazone) began with the rushed publication of an article in the New England Journal of Medicine. The article, which itself has always controversial, suggested that Avandia was associated with an increased risk of myocardial infarctions1–doctor-speak for “heart attacks.”
The authors, Steven Nissen, M.D., and Kathy Wolski, used data that had been published by GlaxoSmithKline, Avandia’s manufacturer, on its website,1 without the data having been first critically examined (“adjudicated”). In fact, Nissen and Wolski’s “verification” of the data consisted of checking the data from the website against articles published in the medical literature, but, whenever there was a discrepancy, simply ignoring the discrepancy and using the data from the website anyway.1 Nissen and Wolski made no attempt to find all of the studies that might have yielded evidence.2 Avandia is not to be used in patients with congestive heart failure, but Nissen and Wolski included a study of in patients with congestive heart failure.2 Three of the studies that they included were of non-diabetics with either Alzheimer’s disease or psoriasis, and they included other studies of patients who were merely at risk for diabetes.2 The designs and assessment of outcomes of the studies were a hodgepodge; Nissen and Wolski combined them, since a statistical test for variability did not confirm that variability;1 however, the statistical test they used was not suitable for sparse data,2 such as theirs. Again, rather than using statistical techniques that would have allowed the use of data from studies without any myocardial infarctions and deaths, Nissen and Wolski ignored the studies.1 Furthermore, the statistical technique that they used requires balance between the numbers in the Avandia groups and the comparison groups, but there were two- and three-fold differences.2 Only by discarding part of the data and using improper statistical techniques were Nissen and Wolski able to conclude that the “data suggest a cardiovascular risk associated with the use of rosiglitazone.”1
Notwithstanding the problems with the studies selected by Nissen and Wolski, using appropriate statistical techniques and all of the data from those same studies yields very different conclusions: “The risk for myocardial infarction and death from cardiovascular disease for diabetic patients taking rosiglitazone is uncertain. Neither increased nor decreased risk is established.”2
Nissen and Wolski were sloppy, at best, but is that all? Dr. Nissen acknowledged receiving research support from Takeda.1 What he did not mention is that Takeda manufacturers the rival drug, ACTOS (pioglitazone), which he pointedly compared to Avandia. Was his campaign against Avandia punitive, because GlaxoSmithKline did not support his research? In addition, he was widely promoted for the top position at the Food and Drug Administration3,4 for his muckraking. Whether or not he actively sought the position, he apparently never demurred, either. Dr. Nissen’s campaign against Avandia seems incompatible with a love of the truth, but it has all the hallmarks of an ego-inspired crusade.
As will become apparent in future posts, whatever his motives, Dr. Nissen has since demonstrated conspicuous lack of objectivity and unprofessional intolerance for other points of view about Avandia.
© 2013 Myron Shank, M.D., Ph.D.
4 Mundy Alicia. Horse race begins for new leaders at FDA and HHS. Health Blog November 5, 2008.