Analysis of several recent studies indicates that Avandia (rosiglitazone), a type 2 diabetes medication that’s been taken by more than six million people worldwide, is associated with a 43 percent increased risk of heart attack and with a borderline-significant increased risk of heart attack-related death.
In response to the findings, the FDA issued a safety alert regarding Avandia, stating that patients with heart disease or at risk of heart attack should talk to their doctors about other treatment options.
Rosiglitazone is a thiazolidinedione (TZD), a class of medication that reduces fasting glucose and A1c’s by increasing the sensitivity of tissues to insulin. It was approved in 1999 for treatment of high blood sugar in type 2 diabetes and has since become the biggest selling diabetes drug in the world. Its manufacturer, GlaxoSmithKline, makes about $3 billion on it every year, even though it’s been associated with body weight increases, adverse effects on lipids, fluid retention, and anemia. Moreover, in February GlaxoSmithKline, warned the public that the drug may increases the risk of bone fractures in the feet, upper arms, and hands.
The Avandia meta-analysis, a pooled overview of 42 pre-existing research reports, was conducted by doctors at the Cleveland Clinic and published in the New England Journal of Medicine. The primary author, a cardiologist by the name of Steven Nissen, is a prominent drug safety advocate who first sounded the alarm about the arthritis drug Vioxx, which was pulled from the market three years later for causing heart problems.
In an editorial review of the Avandia meta-analysis, also published in the New England Journal of Medicine, Drs. Psaty and Furberg indicated that the approval process for Avandia may have been insufficiently rigorous because rosiglitazone has so many effects on genes, many of which are unknown. Moreover, its approval was based upon studies that may have been too short to be really useful in identifying long-term adverse heart effects.
Previously, meta-analysis was instrumental in preventing Pargluva (muraglitazar), a drug similar to Avandia, from being approved, again because of heart problems. In 2000, another glitazone, Rezulin (troglitazone), was removed from the market due to liver toxicity. Earlier this year, the Cochrane Library published a meta-analysis of Actos (pioglitazone), a thiazolidinedione similar to Avandia, which implicated it in edema and heart failure; other experts, however, countered that those were known side effects of the medicine and that its benefits outweighed the risks. A meta-analysis is considered less convincing than a large specially designed trial, and such a trial of Actos has shown benefit with regard to heart attack, stroke, and death.
Meta-analyses do have weaknesses, in part because combining data from different studies creates statistical problems. Consequently, many experts are counseling doctors and patients not to over-react to the Avandia news until the information has been carefully verified. As Jane Jeffrie Seley, a diabetes nurse practitioner warns, “It is important not to suddenly stop taking Avandia without discussing it with your health care provider and deciding what’s best for you. Allowing your blood glucose to rise without proper treatment is potentially harmful.”
Robert J. Tanenberg, MD, FACP, has concerns about the legitimacy of the meta-analysis, noting that “the authors excluded trials where there were no cardiovascular events. This ‘cherry picking’ may be why their results reached statistical significance. It is of interest that the study author, Dr Steven Nissen, notes in his disclosure that he receives research support from Takeda, which manufactures Actos, a competing drug to Avandia.” Dr. Tanenberg was also dismayed by the manner in which the findings have been trumpeted by the media, noting, “The fact that the NEJM published this study and then released it early to the press before physicians could even read it is another sad commentary on our ‘sound bite sensationalistic society.’ [Dr. Nissen] was apparently on national TV, which makes one question his true motives to do these types of studies. The net result is a panic mentality on the part of patients who have trusted their physicians who wrote the prescriptions. Every drug has a risk/benefit ratio that needs to be addressed by physicians and patients. This process is becoming increasingly usurped by the media and by trial lawyers, who are already posting their ads.”
Dr. Steven Edelman, a well-known diabetes expert, also pointed out “questionable statistics” in the meta-analysis, noting that it “included data from 42 varying clinical trials of at least 24 weeks treatment duration that excluded data in six trials that did not report any myocardial infarctions (MI) or cardiovascular (CV) deaths. While this data may eventually pan out to be correct, the numbers are exceedingly small, with an excess of only fourteen cases of myocardial infarction with rosiglitazone treatment in more than 25,000 patients. The actual numbers are 86 MI’s in 14,371 subjects in the rosiglitazone group, and 72 MI’s in 11,634 subjects in the combined control group (sulfonylureas, metformin, insulin, placebo). Interestingly, there is no statistical difference between rosiglitazone and placebo for MI or CV deaths.”
Addressing the public reaction, Dr. Edelman noted, “Obviously, we’ve been getting lots of calls from concerned patients and providers. For now, I suggest that we do nothing proactive. If patients are concerned, it’s my suggestion that they be evaluated, as in the past, and therapeutic decisions be made on an individual basis. I’m not sure what to make of the data at this point and am more than a little concerned about this approach to data analysis and its impact on clinical care. I don’t want to overreact to such a small number of cases. I understand that this is also currently the opinion of the FDA, but opinion can sway quickly and we’ll just have to keep our finger on the pulse.”
According to Dr. Alan Marcus, “there are many interventions for lowering glucose, and it is incumbent upon us to utilize what we know is the risk/benefit ratio. If Avandia offered a unique benefit that was otherwise unavailable with alternative pharmacological intervention, I would recommend it. With this current information, the patient and the clinician will have to judge whether the glucose-lowering that Avandia will achieve cannot be reached without even the possibility of added risk. Remember that Rezulin was withdrawn not only because of the associated hepatic failure risk, but also because their were alternative choices of interventions that did not have the same risk associated with their use.”
Dr. Marcus adds, “The main issue is not only the safety of Avandia, which is a serious and necessary issue, but the way we judge the safety of drugs. Clearly Avandia is efficacious in lowering glucose levels, and there is no doubt about the benefit this gives to patients. The questions regarding cardiovascular disease revolve around the appropriate treatment of the patient with diabetes as a whole human being and the requirement to include additional protective measures such as statins, aspirin, and angiotensin-focused active drugs. When the risk of a certain drug is called into question, there needs to be a clear and scientific evaluation of not just some of the facts, but all the data. We are, after all, deciding what is the risk to living people. Was this study flawed? Yes. Have we determined that Avandia is dangerous? No. What the article in NEJM has done is begun a discussion that must be answered based on fact and not hysteria. Medicine should not be a victim of McCarthyism, but should stand strong on the scientific method that gave birth to it.”
Dr. Allen King of the Diabetes Care Center, on the other hand, had this to say: “I too would be hesitant to change patients to another therapy if this were the only data. However, note the trend (not significant) in both the DREAM and the ADOPT trials (sponsored by Avandia) of a higher incidence of myocardial infarctions for Avandia. Isn’t this a reverse in a trend we expected for the TZDs? Also, DeRosa and Ron Goldberg’s head-to-head studies, comparing Avandia to Actos lipid benefit, show an increase in LDL particle number and apoB 100 (not so with Actos) (besides the increase in triglycerides). Thus we have an explanation of why one would expect an increase in cardiovascular disease with Avandia. On that basis, I am recommending patients to discontinue Avandia.”
In spite of issuing the safety warning, the FDA noted that it is still weighing several conflicting sources of data and that it also wants to investigate whether Actos has more or fewer risks than Avandia. In light of its unanswered questions, the FDA is not asking GlaxoSmithKline to take any action as yet. Investors, on the other hand, moved swiftly: the price of GlaxoSmithKline shares dropped eight percent in one day following publication of the meta-analysis.