|
||||||
|
|
|
|
|||||||||||
|
|||||||||||||
| ||||||||||||||||||||||||||||||||||||
Just prior to the recently completed 2007 American Diabetes Association (ADA) Scientific Sessions, which convened in Chicago in late June, we received notice that a special session would be dedicated to the rosiglitazone (Avandia, GlaxoSmithKline) controversy. The debate would feature Dr. Steven Nissen, the Cleveland Clinic cardiologist who perhaps changed type 2 diabetes treatment as we know it today with publication of his now-famous meta-analysis showing increased cardiac risks with rosiglitazone, and Dr. Philip Home, a distinguished professor from the U.K. who is the lead author for the Rosiglitazone Evaluated for Cardiac Outcomes and Regulation of Glycemia in Diabetes (RECORD) Study, due to be completed in 2009.1,2 (Interim analysis of data from RECORD, which was designed to answer the question of whether rosiglitazone improved outcomes for cardiovascular hospitalization or death, was inconclusive.)
The "debate" was the most anticipated session I've seen at an ADA meeting since the announcement of the Diabetes Control and Complications Trial (DCCT) in 1993. It was discussed in almost every conversation in the 3 days leading up to the session. However, unlike the DCCT event, the actual session was disappointing and anticlimactic.
Let's review. On May 21 the New England Journal of Medicine posted Nissen's initial report on its Web site, along with an editorial that was quite negative about the drug.3 Subsequently, there was the expected posturing by the manufacturer, Glaxo-SmithKline (GSK), which asserted the drug was safe. Not unexpectedly, the press, particularly The New York Times, ran several articles pointing fingers at numerous institutions and people, including the Food and Drug Administration (FDA). By June 6, at a special hearing, members of Congress questioned the FDA's authority and competence in protecting the public from dangerous pharmaceuticals. At the same time the FDA issued an alert to health care providers based on the Nissen meta-analysis, the agency's own similar data that had not yet been exposed, data from GSK, and perhaps public pressure to do something. We are now awaiting a final FDA ruling, which will be announced before you read this.
The debate in Chicago lasted 2 hours, as the issues are complex. We all have our opinions, and perhaps mine are unique. I should first note that I have never been a strong advocate for the class of thiazolidinediones (TZDs). We were one of the first groups to note pulmonary edema with troglitazone (Rezulin, Parke-Davis/Warner-Lambert), and I also reported clinically significant macular edema from the same drug.4 Since edema can occur in the feet and lungs, why not the macula in someone with retinopathy? Although rare, this is now a known side effect from this class and is noted on the label for rosiglitazone.
But I had a more fundamental problem: Why give a drug that makes an obese person more obese? At least with insulin we have evidence we can improve microvascular complication risk and, at worst, have no deterioration of macrovascular risk. Although we assume the same microvascular benefit with TZDs, we don't have the data to support that contention. Much more important, with TZDs, we have no data demonstrating that macrovascular disease risk is improved or at least neutral in users. What is also becoming clearer, however, is that TZDs have a detrimental effect on bone density leading to increased fracture risk.
Why am I so dismayed by Nissen's data? After all, it proves these drugs indeed are dangerous and should not be used. Right? Wrong.
Previous clinical trials, such as the Diabetes Reduction Assessment with Ramipril and Rosiglitazone Medication (DREAM) and A Diabetes Outcome Progression Trial (ADOPT) need to be considered as hypothesis generating. Although not conclusive, the best a meta-analysis can do is support a hypothesis. DREAM and ADOPT cannot be used in the subsequent analysis as then the same data would be used twice, and this is not standard statistical protocol. These two studies also disproportionately weight the analysis due to greater numbers and duration.
The major question remains: What will the final RECORD data reveal? And that, of course, brings us to the next problem. The process of how these data were released has literally led to "panic in the streets" (to quote Harvard Medical School Professor David Nathan on a different topic). How many people will now drop out of RECORD due to the fear generated by the meta-analysis and editorial? What about other large ongoing trials? For example, Action to Control Cardiovascular Risk in Diabetes (ACCORD) and Bypass Angioplasty Revascularization Investigation 2 Diabetes (BARI-2D) both include rosiglitazone and have macrovascular events as primary outcomes. Since the data-monitoring safety boards for these trials have not reported concerns with rosiglitazone, could the Nissen article cause patients to drop out, jeopardizing glucose control data in these studies?
I have other questions as well. Could the New England Journal of Medicine have published this with less hype? Could the editorial have been written by different authors who took a more balanced approach? And what about the press? Although there has been quite a bit of criticism about its role in the hysteria, I understand its job is to sell newspapers; this certainly was a good story that did just that. So I have no problem with the press.
But I do question the level of the reporting: What were noted repetitively were the relative risks of myocardial infarction and deaths from cardiovascular causes. It needs to be recalled the actual event rate was quite small. For myocardial infarction, rosiglitazone resulted in 86 events/14,371 subjects compared with 72/11,634 for the control group. This is an absolute event rate of 0.60% vs. 0.62%. These numbers are obviously small, but when the odds ratio was calculated, the analysis resulted in a 43% increased risk for rosiglitazone, and this was significant at a p-level of 0.03.
My greatest concern is that the FDA, due to pressure from both the public and Congress, will tend to overreact to possible safety concerns—not just to TZDs, but to all drugs. We have to remember that with any new agent we will never have the "hard outcome" studies we would like. Heck, for insulin with type 2 diabetes, it took 76 years to get hard-outcomes data!
Despite all the potential anti-inflammatory and antiatherogenic possibilities of TZDs, I believe these drugs are probably overused and overmarketed; as of now, all we have had is the promise of these benefits. But what we have witnessed the past few months is a system that simply doesn't work. It is my hope that these drugs remain in our arsenal of agents used to treat type 2 diabetes until further randomized prospective data are published, although stronger warnings for patients with heart failure would be welcomed, as heart failure is notably more common in virtually all studies with these drugs. The good news is that after the patient stops the drug the fluid overload generally resolves. Therefore, although current guidelines recommend appropriate use of these agents in low doses with ejection fractions <40%, I would suggest TZDs not be used at all in this population.5
There is one final point the Nissen article, to its credit, illustrates for all to see: the total publication bias from the pharmaceutical industry. Of the 35 rosiglitazone trials identified in the clinical trial registry, only nine were ever published.
That's not surprising: Why would a drug company want to publish a negative trial or a trial with significant adverse outcomes? Even smaller trials performed in smaller regions of the world with results consistent with previously published studies could have subtle, important differences. This, of course, is the reason all clinical trials need to be registered and posted on the Web site clinicaltrials. gov. Still, unpublished negative trials do not undergo peer review, are not discussed in journal clubs, and are not noted by sales representatives. They usually are not acknowledged since this Web site is not generally accessed by the practicing physician.
Let's face it: The majority of our information is filtered for the sake of sales and marketing. Perhaps the Web site posting is not enough, or perhaps we need some incentive for physicians to review this information, such as continuing medical education (CME). I am smiling as I think about the FDA becoming a CME accreditation body, but why not? It's time for us to get more creative with the issue of access to drug data, while acknowledging to the industry that we appreciate there will be bad and good with every study. But only performing registration studies for drug approval or biased marketing studies is not an option.
Modern-day diabetes history is full of study controversies. University
Group Diabetes Program (UGDP), Diabetes Mellitus, Insulin Glucose Infusion in
Acute Myocardial Infarction-2 (DIGAMI-2), Clinical Trial of Reviparin and
Metabolic Modulation in Acute Myocardial Infarction Treatment
Evaluation-Estudios Cardiologicos Latin America (CREATE-ECLA), and Prospective
Pioglitazone Clinical Trial in Macrovascular Events (PRO-active) were all
marred to some degree by disagreement. The difference is none of those trials
was a meta-analysis. While we may disagree about the potential risks and
benefits of TZDs, we owe it to ourselves to allow proper scientific methods to
prevail. 
Footnotes
|
References
2. Home PD, Pocock SJ, Beck-Nielsen H, et al.: Rosiglitazone evaluated for cardiovascular outcomes: An interim analysis. N Engl J Med [Epub ahead of print Jun 5] 2007. Available online at http://content.nejm.org/cgi/content/abstract/NEJMoa073394v1. Accessed June 29, 2007.
3. Psaty BM, Furberg CD: Rosiglitazone and cardiovascular risk.
N Engl J Med 356:2522–2524, 2007.
4. Hirsch IB, Kelly J, Cooper S: Pulmonary edema associated with
troglitazone therapy. Arch Intern Med 159: 1811, 1999.
5. Nesto RW, Bell D, Bonow RO, et al.: Thiazolidinedione use, fluid
retention, and congestive heart failure: A consensus statement from the
American Heart Association and American Diabetes Association.
Diabetes Care 27:256–263, 2004.
CiteULike 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
| ||||||||||||||||||||||||||||||||||||
|
||||||
|
| DOC News | Diabetes | Diabetes Care | Clinical Diabetes | Diabetes Spectrum |
