Philadelphia, PA - A new retrospective cohort analysis of more than 75 000 Medicare patients who were treated with a Cypher drug-eluting stent (DES) within the first nine months of the device's FDA approval suggests that DES not only reduce the risk of revascularization procedures but also death and MI [1].

According to the study authors, who used propensity matching to compare outcomes in DES-treated patients with both historical and contemporary controls treated with bare-metal stenting, the findings at the very least suggest that there is no increased mortality risk with drug-eluting stents.

The study is published in the May 27, 2008 issue of the Journal of the American College of Cardiology.

In an interview with heartwire, lead author on the study, Dr Peter Groeneveld (Philadelphia Veterans Affairs Medical Center, PA), highlighted some of the unique aspects of the study. For one, the study population was older than that typically included in the randomized trials, with more comorbidities. Two, the period studied was a time that predated the stent-thrombosis scare, so physicians who were early adopters of the technology were enthusiastically using the devices both on and off label and without extended dual antiplatelet therapy.

"What we found, much to our surprise, is that there is a very strong statistically significant association with improved survival for DES patients," he said. "And that, of course, wasn't what the clinical trials found."

Like investigators before him of other cohort analyses or meta-analyses that have hinted at a mortality benefit—or the reverse—Groeneveld acknowledges that it is impossible to control for all potential biases that could have guided choice of stent and also explained the mortality differences. "On the other hand, it would have had to have been a very strong influential factor, yet it wasn't something that made it into the records. So it's a little hard to imagine just what that could have been, since we matched on all kinds of factors that we know influence survival."

Age, comorbidities, higher risk of bleeding, and upcoming surgeries are all factors that drive use of bare-metal stents today due to the risks of long-term clopidogrel required for DES. But back in 2003-2004, when these data were collected, concerns about late stent thrombosis and the need for clopidogrel long-term were unheard of, Groeneveld points out.

One possible explanation for the reduced mortality among DES-treated patients is that restenosis, more common among bare-metal-stent-treated patients, is "not as benign a process in a 78-year-old as it is in a 58-year-old," he suggests. It's a theory that has gained increased traction in recent months. Another possibility proposed by the authors is that early adoption of DES may also be a marker for hospitals that are providing cutting-edge care and are more likely to provide evidence-based medicine.

It doesn't mean that there are no thrombosis risks to DES, but gosh, what we see is a disturbing trend to suggest that they might actually be effective.

"There are hospitals that are more keyed in than other hospitals to the drumbeat of new technologies," Groeneveld commented. "That can be good and bad, of course, because if your hospital is constantly adopting the newest best thing, every now and then they are going to get burned. People who used Vioxx first, and used a lot of it, are probably feeling kind of bad about themselves these days," he said. But it is at least plausible, the authors argue, that hospitals swift to adopt DES were, overall, providing higher-quality care than hospitals that stuck with bare-metal stents.

Groeneveld agrees that longer-term data will be essential, but he points out that he and his colleagues will have to see "an awful lot of events to counteract" the survival benefit associated with DES. "This at least provides some encouraging evidence in hundreds of thousands of patients that if there was some kind of strong mortality signal associated with DES at least in the first 18 months, we didn't see it."

Medicare is only now releasing three- and four-year follow-up data, he added.

"I suspect that there will be thrombosis events in there, but the results would have to be of an enormous volume to counteract whatever effect it was that we've already seen. It doesn't mean that there are no thrombosis risks to DES, but gosh, what we see is a disturbing trend to suggest that they might actually be effective."

Asked to comment on the study for heartwire, Dr William Weintraub (Christiana Care Health System, Newark, DE) reiterated that it is never possible to completely eliminate selection bias and that having a very large cohort does not solve the problem of confounding.

"You won't overcome it, even if you have two zillion patients. Your results looks a lot better, your p value looks strong, and your confidence intervals get tighter and tighter, but size does not eliminate the problem of bias," he said.

A strong association does not prove causality, Weintraub stressed, noting that even the authors are "pretty discreet" in their conclusions.

For his part, Weintraub takes this study, on top of the ones that have come before it, as reassuring. "I think it's likely that DES are safe, and we can stop worrying so much that, on a population basis, we're killing people," he said. The unanswered questions about clopidogrel duration and how best to prevent stent thrombosis remain.

"I feel pretty confident about using DES," he concludes. "After a period of worry, they've made a return and most people are pretty comfortable. I'm not going to tell people we can prevent them from getting a heart attack or save their life with this, but I don't think we're putting their lives at risk."