AvandiaGate: FDA Has ‘Considerable Concern’
5 CommentsBy Ed Silverman // July 26th, 2007 // 10:13 am
FDA reviewers have “considerable concern” about Avandia data that show a signal of possible heart attack risks with the diabetes med, according to briefing documents filed today on the agency web site in advance of next Monday’s advisory committee meeting. Various data present “somewhat inconsistent findings” on the risk, the reviewers wrote.
At issue is whether the Glaxo drug should stay on the market or be slapped with more stringent warnings or restrictions after the controversial meta-analysis was published in The New England Journal of Medicine. The study set off a firestorm over Glaxo’s behavior and a smear campaign against Steve Nissen, the Cleveland Clinic cardiologist who conducted the analysis. Interestingly, the Monday meeting won’t incorporate the meta-analysis, because the FDA analysis covers the bases, the agency writes.
You can read the agenda for the Monday meeting here.
You can read the committee rosters here, here and here.
You can read conflict of interest waivers for committee members here.
You can read the FDA briefing document here.
And you can read Glaxo’s briefing document here.
If this doesn’t keep you busy for awhile, what will?
Dr. Remulac
I’ve given both the GSK and FDA documents a quick review. Looks like GSK has done a pretty decent job with a mix of metaanalysis, ongling clinical trial review, and claims database analyses to paint a full picture of the data and that it is inconsistent and inconclusive. To me, the fact that several large, appropriately designed/powered CV endpoint trials including Avandia are still ongoing and have not been stopped prematurely due to statistical excess of events on Avandia sends the clearest message that the jury is still out and that we need to await the completion of these trials before we have the definitive answer. Interestingly, the insurance claims analyses that GSK commissioned also include an analysis of Actos, and they reveal no true differences between it and Avandia. All those docs that are out there switching patients to Actos should be ashamed of themselves. The FDA review is very typical “I’m smarter than you” FDA analysis. They slice and dice the metaanalysis data in many ways, and have come up with some interesting potential interactions between Avandia and patients with previous CHD who are on nitrates, pts on ACE inhibitors, and patients on insulin. In each of these groups it appears that the risk of MI may be higher with exposure to Avandia, and the FDA recommends that these specific subgroups be teased out in the analyses of the ongoing CV outcome trials when they are completed. Incidentally, the FDA WILL NOT be critiquing Nissen’s metaanalysis of trial-level data (GSK and the FDA analyzed patient-level data), but they’ll be taking that up at a later time. I think that is to spare Nissen the embarrassment of having them slam his methods during the time of the meeting. It is very disappointing to me that the FDA doesn’t have a similar analysis/scrutiny ongoing for Actos, although they have asked the mfr for that data. This is supposed to be a review of the TZD class, but it’s really just a further grilling of Avandia. I can tell you that in my 20+ year career in medicine/research/pharma that I can think of very instances where a drug has undergone this level of analysis and things remain inconclusive. To me, this means that the risk of Avandia is actually much less than it has been portrayed to and perceived by the public. The differences in incident rates of MI, etc are so small (fractions of a percent) betwen Avandia and comparators that you have treat 500-1000 patients for 6 months to several years before you’d see a single excess MI event. There are other drugs in wide clinical use today that would not withstand this sort of scrutiny, and have much worse risk-benefit ratios than Avandia!
Dr. Remulac
Oh, I forgot to provide my best guess on the outcome (even though no one asked for it! LOL):
The panel will vote that the overall risk remains inconclusive, but that it may be prudent at this time to recommend restricted use in the subgroups that appear to be at the highest risk until the CV outcomes trials are completed. Avandia will remain on the market. They will further say that their comments only pertain to Avandia, and that similar analyses will be required for all other agents used in the treatment of diabetes, starting with Actos, the only other TZD on the market.
Dr. Remulac
After reading further, here are some key passages:
Information about increased risks of myocardial ischemia with rosiglitazone warrants prominent placement in the labeling to ensure proper patient selection and appropriate monitoring for risk factors such as rapid or excessive weight gain. DDRE recommends that the following information be included in a BOXED WARNING for rosiglitazone:
· In a retrospective analysis of data from pooled controlled clinical studies, an increased risk of myocardial ischemia was observed in patients treated with rosiglitazone (Hazard ratio 1.31; 95% confidence interval 1.01, 1.70). Higher levels of risk were noted in patients with pre-existing serious heart disease, including heart failure, as well as in patients receiving insulin therapy.
· Rosiglitazone is not recommended in patients receiving insulin.
· Rosiglitazone is not recommended in patients with heart failure, or serious heart disease, including symptomatic coronary artery disease.
In contrast, treatment with pioglitazone, although clearly associated with increased risk of heart failure, has not been shown to result in increased risk of myocardial ischemia, even in patients receiving concomitant insulin therapy. A pooled analysis of cardiovascular safety data from randomized controlled trials with pioglitazone has been submitted by Takeda and is currently under review by DMEP. Information about risk of heart failure with pioglitazone should be prominently communicated in a BOXED WARNING in order to assure proper patient selection and monitoring.
Regarding PROactive study with Actos: The statistical plan left no alpha for consideration of secondary endpoints in the event of a failed primary. Therefore, all secondary endpoints would be considered exploratory. Predefined secondary endpoints included cardiovascular mortality, and the individual components of the primary endpoint. There was no significant difference between pioglitazone and placebo for any of these individual endpoints; point estimates for most hazard ratios were
Simon Passanante - AvandiaGate: FDA has considerable concerns
[...] Pharmalot reports the FDA is reviewing data regarding Avandia, a drug used to treat Type II diabetes, and finding a considerable amount of evidence that the manufacturer knew of cardiovascular side effects. [...]
Celia Chapman
neronian xanthophyllous plagiostomous rattlingness photoemission bibacious cerebroside zwinglian
MapStats: Inyo County
http://www.bris.ac.uk/Depts/Archaeology/continuing/