Both of the meta-analyses combined all of the Cordis/J&J-sponsored Cypher randomized trials, as well as the Boston Scientific-sponsored Taxus program: one found an increased incidence of death and Q-wave MI with the Cypher stent and a trend toward increased death/Q-wave-MI with the Taxus, while the second found no differences in cardiac mortality but an increase in noncardiac mortality, again with the Cypher stent.

Dr Edoardo Camenzind

The separate presentations, which shared a single hotline slot—necessitating rushed synopses on the part of the presenters—spurred discussant Dr Salim Yusuf (McMaster University, Hamilton, ON) to deliver a thundering indictment of what he later described to the press as an "epidemic of madness" over misuse of PCI for stable angina in general and drug-eluting stents specifically.

"As clinicians we seem to have lost our clinical judgment, let alone our ability to view data and evidence," Yusuf stated. "We therefore need a thoughtful and selective approach to PCI, complementing full medical therapy. . . . The whole field of angioplasty has been led astray by a preoccupation with restenosis, for which study after study has shown has no prognostic value. We're chasing problems that are iatrogenic that naturally would not exist in people. We've had a perverse financial incentive on the practice of cardiology. It is time to stop and reevaluate."

As clinicians we seem to have lost our clinical judgment, let alone our ability to view data and evidence.

Camenzind's meta-analysis was based on two separate analyses of the sirolimus and paclitaxel data. In the first, the investigators examined death and Q-wave MI in the published or presented papers, pooling them by time of follow-up. From eight to nine months of follow-up, out to one, two, and three years of follow-up, death/MI rates increased at rates that ranged from 30% to 40% higher in the Cypher-treated patients than those of the bare-metal-stent controls. A similar trend was seen, over time, for the paclitaxel-eluting stent, but here the relative difference between the Taxus and the bare-metal stent was only about 5% difference over the three years of follow-up.

In the second analysis, all of the randomized trials within each stent's program were stratified by last follow-up data. In this analysis, serious adverse events in sirolimus were 6.3% compared with 3.9% in the bare-metal-stent group (p=0.03) and in the paclitaxel trials were 2.6% vs 2.3% (p=NS).

"We conclude that death and Q-wave MI [as the] clinical presentation of stent thrombosis have a higher incidence in first-generation DES as compared with bare-metal controls," Camenzind stated. "Excess events appear to occur with both types of stents, although the magnitude seems to be higher with sirolimus. A risk/benefit analysis of systematic use of first-generation drug-eluting stents is warranted."

Dr Alain J Nordmann

Nordmann's findings, while raising the specter of increased deaths, actually clashed somewhat with those of Camenzind. Nordmann et al combined data from 17 randomized controlled trials of paclitaxel- or sirolimus-eluting stents to evaluate total mortality, cardiac mortality, and noncardiac mortality. While total mortality at one year trended toward a benefit of DES, at two, three, and four years the investigators saw a trend toward increased mortality with DES. For cardiac mortality, however, there was no statistically significant difference between DES and bare-metal stents or for either sirolimus- or paclitaxel-eluting stent compared with bare-metal stents. Most striking, however, was the data for noncardiac deaths, which at two and three years pointed to an association between sirolimus stent implantation and increased noncardiac mortality. Separate analyses identified these deaths as cancer, stroke, or lung disease.

"DES for the treatment of coronary artery disease do not reduce mortality when compared with bare-metal stents," Nordmann concluded. "Preliminary evidence suggests that sirolimus but not paclitaxel may lead to an increase in noncardiac mortality. Long-term follow-up and assessment of cause-specific deaths in patients receiving DES are mandatory to determine safety of patients receiving these devices."

Dr Salim Yusuf

To the press, Nordmann pointed out that obtaining raw data on mortality from the stent manufacturers had been "extremely difficult," highlighting the need for non-company-sponsored, large randomized clinical trials with ample follow-up.

At the very least, said Yusuf, large registries should be mandated to track adverse events in DES recipients. But Yusuf also made a plea to the major cardiovascular organizations to step up and revisit not only the use of DES but the role of PCI in the treatment of stable, non-drug-refractory angina. And to be clear, he added, PCI and drug-eluting stents play a key role in the treatment of unstable angina and acute coronary syndrome—it is as a treatment for stable angina to treat non-life-threatening restenosis that Yusuf singles out as a "myth" and a "man-made disease."

Long-term follow-up and assessment of cause-specific deaths in patients receiving DES are mandatory to determine safety of patients receiving these devices.

As for the meta-analyses themselves, Yusuf stated: "These new studies raise concern. I do not believe these trials are convincing, but they are disconcerting given that we have no data that this procedure is useful. There is a significant excess in noncardiac deaths, and we need to find out if this is real."

Pausing to assure a tittering audience that he was dead serious in his comments, Yusuf added, "I call on the ESC to [convene] a balanced and independent working group, and not just of interventionalists. Certainly you can bring them in, but also noninterventionalists, health economists, patient representatives, and government representatives, and have a committee to find out what the real role of PCI is, of these stents, and keep industry out of it."

Camenzind, for his part, stopped short of denying a role for drug-eluting stents, insisting that his study, and his misgivings, apply only to the two first-generation drug-eluting stents. "We need stents that can control restenosis, that don't totally abolish the healing process but that are able to control it."

Yet another study, presented earlier in the day by Dr Peter Wenaweser (Thorax Center, Rotterdam, the Netherlands) also highlighted the stent thrombosis risk with DES. In the study, Wenaweser and colleagues examined rates of early and late stent thrombosis in patients enrolled in the SIRTAX and Post-SIRTAX registries in Bern and the RESEARCH and T-SEARCH registries in Rotterdam, between April 2002 and December 2005. In Bern, patients received clopidogrel and aspirin for three to six months, while in Rotterdam, patients received dual antiplatelet therapy for three to 12 months. Only angiographically documented stent thromboses were included in the analysis.

All of the presentations are pointing to the fact that stent thrombosis is there and needs a solution. It exists, but it's not terrifying.

In all, 152 stent thromboses occurred in 8146 patients. The cumulative incidence of stent thrombosis was 2.9%, yielding a rate of 1.3 per 100 patient-years. The rate of stent thrombosis was 1.2% at 30 days, 1.7 at one year, 2.3 at two years, and 2.9% at three years, "an almost linear increase of 0.6% per year between 30 days and three years," Wenaweser commented.

In interviews with heartwire, experts tried to put the findings in perspective, offering the oft-repeated calls for longer clopidogrel duration. Dr Antonio Colombo's group (Columbus Hospital, Milan, Italy) has a forthcoming paper examining rates of stent thrombosis in patients who quit dual antiplatelet therapy at one year, compared with patients who stayed on the drug out to three years.

Dr Peter Wenaweser

"I think all of the presentations are pointing to the fact that stent thrombosis is there and needs a solution," Colombo told heartwire. "It exists, but it's not terrifying. My problem with this issue is that we did not use bare-metal stents in situations where we now use DES, so I doubt we can do a fair comparison of stent thrombosis between DES and bare-metal stents."

He also questions whether stent thrombosis rates continue to climb after two years. "I'm not convinced that we have a continuous rate, I think it probably grows up to two years, but from two years on, tends to level off. . . . We should not forget that so many patients in the randomized trials have been off antiplatelet therapy for a long time, and those patients are not dying and having MIs all the time."

Also commenting on the Wenaweser study for heartwire, Dr Anthony Gershlick (Nuffield Leicester Hospital, UK) insisted that it is important to see published papers before jumping to conclusions. For one thing, Wenaweser did not present data on percentage follow-up, making it difficult to appreciate stent thrombosis rate increases over time. As well, there are no studies tracking the "cumulative" stent thrombosis rate for bare-metal stents. "The absolute difference between DES and bare-metal stents will not be 2.9%, it will be 2.9% minus the cumulative rate for bare-metal stents," he noted. "What we need to know is the excess cumulative risk."