ESC/WCC: Meta-Analyses Find Increased Death and MI with Cypher Stent

BARCELONA, Spain, Sept. 3 -- Two Swiss meta-analyses have found an increased rate of myocardial infarction and death with the Cypher (sirolimus-eluting) coronary stent, which is likely to put a serious chill in interventional cardiology's infatuation with drug-coated devices.

Long term data just becoming available suggest that drug-eluting stents, which excel at opening stenosed coronaries and keeping them open, may increase the risk of non-fatal myocardial infarction and death. One of the meta-analyses showed a rise in non-cardiac deaths, from cancer, sepsis and stroke.

A meta-analysis of all available data from published randomized trials of Cypher (sirolimus-eluting stent) and Taxus (paclitaxel eluting stent) found a 2.4% increase in the incidence of death or MI in patients who received Cypher stents compared versus patients treated with bare metal stents (P=0.03).

Results of the two meta-analyses were reported today at the European Society of Cardiology/World Congress of Cardiology meeting here,

One of them, according to Edoardo Camenzind, M.D., an associate professor of cardiology at the University of Geneva in Switzerland, found late stent thrombosis the likely cause of the adverse events.

But moments later, researchers from the Basel Institute for Clinical Epidemiology in Basel, Switzerland, said their meta-analysis of 17 trials that compared Cypher and or Taxus, or both, with bare metal stents found that the real problem is not stent thrombosis, but non-cardiac deaths.

And again it was Cypher, not Taxus, that raised the red flag, with an odds ratio for non-cardiac mortality of 2.74 (95% CI 1.22-6.74) (not significant) at two years and 2.04 (95% CI 1.00-4.15) (not significant) at three years, said Alain Nordmann, M.D., of University Hospital in Basel.

Shrugging off the lack of significance, Dr. Nordmann concluded that his "preliminary data suggest that sirolimus- but not paclitaxel-eluting stents may lead to increased non-cardiac mortality."

The analysis by Dr. Camenzind included data from four Cypher studies-RAVEL, SIRIUS, E-SIRIUS and C-SIRIUS-and five Taxus trials, TAXUS I, TAXUS II, TAXUS III, TAXUS IV, TAXUS V, and TAXUS VI.

All of those trials reported results that favored the drug-eluting stents over bare metal stents on the basis of highly significant reductions in the restenosis rate.

The relative risk of death or MI was 38% higher for Cypher patients than patients who received bare metal stents, while the relative risk was 16% higher for patients implanted with Taxus stents versus those who received bare metal stents.

In both studies the absolute differences were small, but with an estimated six million drug-eluting stents implanted worldwide, and about a million implanted annually in the U.S., a small percentage can add up to many patients.

"These two studies may be the most important papers of this meeting," said Salim Yusuf, M.D., a professor of medicine and director of cardiology at McMaster University in Hamilton, Ontario. Dr. Yusuf served as discussant for the papers.

He cautioned that the data are "not convincing, but they are disconcerting."

They were so disconcerting that he called for the creation of a blue-ribbon panel of interventional and non-interventional cardiologists, health economists, and government agencies to "re-evaluate the real role of stenting" in coronary disease. Moreover, Dr. Yusuf said the panel should exclude industry representatives.

At a press conference, Dr. Yusuf said that he plans to ask interventional cardiologists at his hospital to re-evaluate their use of drug-eluting stents so that they are used "in a limited way" until more data are available to determine whether the stents are "a panacea-as many of our patients and residents believe-or an expensive placebo, or--as I suspect--a Trojan horse."

Dr. Yusuf argued that stents may be closest to a Trojan horse, carrying hidden long-term danger in the guise of a useful therapy to avoid restenosis.

Philippe Gabriel Steg, M.D., of Hospitalier Bichat-Claude-Bernard in Paris, a co-author of the meta-analysis with Dr. Camenzind, likened long term drug-eluting stent data to long term data on Cox-2 inhibitors such as Vioxx (rofecoxib), which was a heralded arthritis treatment until long term studies found that the drug increased the risk of MI and stroke. In that case, too, the absolute increase in risk was small, but the drug was used by millions of patients.

With drug-eluting stents the increased adverse event rate is particularly chilling because restenosis is a common but relatively benign process, while stent thrombosis is "a rare but life-threatening disease," Dr. Steg said.

But Mayo Clinic cardiologist Ray Gibbons, M.D., who is president of the American Heart Association, cautioned against over interpreting the data from the meta-analyses. He said the late thrombosis rate may reflect a problem with medical therapy, rather than a prothrombotic effect of the stents.

Dr. Gibbons pointed out that registry data from Berne and Rotterdam-registries that include 3,875 Cypher patients and 4,271 patients implanted with Taxus stents-found that 26% of patients who had late stent thrombosis were taking no antiplatelet therapy. Only 51% were taking either aspirin or Plavix (clopidogrel), a finding he called "shocking."

He said current ACC/AHA guidelines recommend three months of dual antiplatelet therapy for the Cypher stent and six months for the Taxus stent. But the guidelines add the recommendation that patients who have no known excess bleeding risk, should be continued on dual antiplatelet therapy for a year.

And although 26% of patients who suffered late thrombosis were not receiving antiplatelet therapy, the same registry data found that 23% of late thrombosis occurred in patients who were taking dual antiplatelet therapy.

So while he agreed that the late thrombosis risk might reflect inadequate antiplatelet therapy, Steven Nissen, M.D., president of the American College of Cardiology and director of cardiology at the Cleveland Clinic, said there could be another explanation-- "the law of unintended consequences."

The drug-eluting stents prevent restenosis because their drug coatings have an anti-proliferative effect, that same action may prevent the stents from becoming re-endothelialized.

Dr. Nissen said he reviewed one paper that suggested that nine months after implantation-long after the drug coating is depleted-the stent surface was not endothelialized. The now "bare" surface was potentially prothrombotic, he said.

Dr. Yusuf also suggested that an unforeseen long-lasting effect of sirolimus or paclitaxel might also explain the non-cardiac deaths reported by Dr. Nordmann.

The agents, he said, could "cause some rapid impairment of the immune system, which causes conditions that we simply don't understand."

Cypher is made by Cordis, a Johnson & Johnson company, while Taxus is marketed by Boston Scientific. Both companies were prepared for the negative news.

Each prepared their own meta-analyses of their own long term data. Those company reports are already in the hands of the FDA.

But unlike the studies reported here-neither of which had industry funding-the company analyses concluded that there were no significant problems.

Dennis Donohoe, M.D., vice president of worldwide regulatory and clinical affairs at Cordis, said, "We looked at total mortality over four years, and it was 5.1% with bare metal stents, and 6.5% with Cypher, an absolute difference of 1.4% (P=0.22), representing a relative difference of 22%."

He said the MI rate was also not significantly different-6.3% with Cypher and 6.1% with bare metal stents.

Meanwhile, Dr. Yusuf and other cardiologists said the data also suggested the need to continue dual antiplatelet therapy in drug-eluting stent patients beyond the one year mark, and perhaps for a lifetime.