Rosiglitazone study sparks panic

Widely-used diabetes med linked to heart attacks. Methods under fire

By Samuel Jarjour

The popular diabetes drug rosiglitazone can cause heart attacks, according to an earth-shattering study published May 21 in the New England Journal of Medicine. But an increasingly loud din of opposition is questioning the validity and benefit of the meta-analysis, especially since the largescale post-marketing trial RECORD is currently underway. Health Canada, jointly with the manufacturer GlaxoSmithKline, released a public advisory on May 30 urging patients not to panic.

The man who dropped the bombshell on the medical community is no stranger to controversy. Dr Steven Nissen, cardiology chief at the Cleveland Clinic, was also one of the doctors who uncovered rofecoxib's heart risks. His current study is a meta-analysis of 42 previously published clinical trials, obtained from sources that include rosiglitazone manufacturer GSK's own trials registry. It revealed that patients on rosiglitazone had an increased risk of myocardial infarction (MI) and cardiovascular (CV) death.

"I wasn't surprised that the study had such an impact, considering the number of people using it and the magnitude of the hazard we detected," Dr Nissen explains.

That magnitude does appear jaw-droppingly significant. Dr Nissen's meta-analysis found that patients on rosiglitazone were 43% more likely to have an MI than controls. The news is even worse for CV death: the odds for the study group were a whopping 64% higher. All of which looks like really bad news for the millions of type II diabetics who've taken the drug, and their doctors who are now wondering whether the risks are worth it.

"This kind of news can be scary for patients," says Dr Ian Blumer, an internist in London, ON, and a member of the clinical and scientific section of the Canadian Diabetes Association, and co-author of the timely book Understanding Prescription Drugs For Canadians For Dummies published this month. "I think it will also be tough for family physicians, who will be inundated with countless phone calls."

"I'd like physicians to read our paper and make their own mind up, at this point," says Dr Nissen. "It's up to them to decide what to do with it. So far, it's generally been well received by them."

TOO FAR?
But odds ratios are one thing, hard numbers another. Despite the impressive-sounding hazard ratios that came out of the data, the actual number of MIs and CV deaths Dr Nissen turned up might not be monumental enough to haul every diabetic off these meds. Eighty-six patients had MIs and 39 MIs resulted in CV deaths, out of 15,560 rosiglitazone users, while there were 72 MIs and 22 CV deaths in 12,283 controls.

Drs Psaty and Furberg's editorial throws some cold water on the hysteria: "A few events either way might have changed the findings for MI or for death from cardiovascular causes. In this setting, the possibility that the findings were due to chance cannot be excluded."

In the meantime, says Dr Blumer, FPs are left holding the bag. "They'll have to discuss the risks of the drug individually with each of their patients, who need to be reassured that the additional danger of rosiglitazone in terms of absolute risk, as opposed to the relative risk measure in the NEJM study, is miniscule," he says. "Only time will tell if the authors' concerns are ultimately found to be substantiated."

SAFETY BY NUMBERS
The findings are particularly worrying because ideally when treating for diabetes, CV risks actually go down. For Dr Nissen the fact that he found the opposite represents a huge red flag. "The risk of MI and CV death suggests safety problems," he says. "This drug's been on the market for eight years, but hasn't been shown to have any favourable impact on actual health outcomes [other than glycemic control]."

The editorial that accompanies the study, written by University of Washington's Dr Bruce Psaty and Dr Curt Furberg of Wake Forest, notes that though rosiglitazone appears to do a good job at controlling blood sugar, the absence of any other benefits should make physicians ask themselves if it's worth the risks.

Dr Blumer says that a patient who discontinues the drug because of safety fears could be worse off if their blood glucose levels start to go up. "It's always a trade off," Dr Blumer says, "any time we prescribe a drug and any time a patient takes a drug we have to balance the potential risks of taking the drug vs the risks of not taking it."

METHOD OR MADNESS
Physicians and the drug's manufacturer have questioned the validity of Dr Nissen's meta-analysis, which combined data from trials studying wildly different outcomes. The FDA has since reproduced his results, with the same conclusions.

Dr Nissen is very upfront about his study's limitations, including its dependence on second-hand data from studies not intended to assess CV outcomes, which were inconsistently defined between trials.

Still, he feels strongly that rosiglitazone's approval came from studies of its effects on surrogate endpoints, like blood glucose levels, rather than health outcomes. "In general, I think that surrogate endpoints should be used very carefully, and should be relied upon only when you're not able to assess other outcomes," he says.

In the meantime, there's every indication diabetics are strongly considering ditching rosiglitazone in droves. GSK says its RECORD study might be in jeopardy as a result of the the NEJM paper. Apparently a troubling number of frightened subjects have dropped out of the study.