In Cold Blood
An expert panel considers a chilling possibility: The most popular type of stent may be unduly risky
For 4 million Americans whose intimate wear includes a tiny, drug-coated cylinder of metal wire, lodged in a coronary artery to prop it open, these are anxious times. Could coated stents, which in just a few years have become the treatment of choice in people with heart blockages, be putting them at higher-than-expected risk of a killer clot? This week, a Food and Drug Administration panel will convene to consider the evidence, which began trickling out in the spring and galvanized a global meeting of heart specialists in September.
It was not news that the devices can promote clots–with either a drug–coated stent or an uncoated, "bare metal" stent, blood-thinning medication is necessary for a time. What took clinical researchers aback was the prospect of a continuing toll from coated stents, year after year, in lives and heart attacks. When Johnson & Johnson's coated Cypher stent was approved by the FDA in 2003, it was hailed as a potent weapon against restenosis, the tendency of a blocked coronary artery to close up again shortly after being reopened. Historically, vessels opened by an angioplasty balloon-jammed into the blockage and inflated to squash it against the artery walls-reblocked within a year or less as often as 40 percent of the time. That meant another procedure, or surgery.
Slow growth. Bare-metal stents, introduced in the 1990s as scaffolding to keep reopened vessels from closing, cut the restenosis rate to between 15 and 30 percent. And the Cypher squeezed it below 10 percent–in most trials, far below. A coating of rapamycin, a powerful medication leaked at a controlled rate into the artery wall, successfully discouraged tissue regrowth. Together with Boston Scientific's paclitaxel-coated Taxus stent, which arrived in 2004, coated stents quickly relegated bare metal to a token 10 percent or less of the market.
In hindsight, the revolution may have happened too fast. A loud alarm was sounded at the September heart meeting by cardiologist Edoardo Camenzind, a researcher at the University Hospital Geneva in Switzerland. He reported that in patients with Cypher stents in Europe and elsewhere who were tracked for three years, death and heart attack rates were far higher than in patients who got bare-metal stents–6.3 percent vs. 3.9 percent. The rates for Taxus patients also rose, if less dramatically. The cause, Camenzind concluded, was clotting: late thrombosis. A just-published study in the American Journal of Medicine seems to buttress that conclusion: Among 6,675 stented patients in a variety of trials worldwide, none with a bare-metal stent experienced clotting after the first year, while 0.5 percent of coated-stent patients did.
Camenzind's revelation spurred researchers and manufacturers to dive back into their data, and the FDA–which declined to comment for this article because of the pending meeting–decided to put everything back on the table this week. The panel might even reach a conclusion as soon as the meeting is adjourned.
A few in the medical community had seen this coming. Approval of medical devices often rests on data generated by a few hundred patients who may not represent the real world–they're relatively easy to treat. With coated stents, that was reflected in the FDA approval letter: They were deemed appropriate for patients with a single, compact blockage. It should be neither too long nor too short, and the artery neither too wide nor too narrow.
But physicians are free to go "off label" and use the stents in just about anybody. And they have enthusiastically done just that. A study presented in October at a cardiology meeting in Washington, D.C., found that more than half of some 3,300 patients at hospitals around the country had received coated stents off-label. In the next six months, the off-label patients died more often, had more heart attacks, and had a higher rate of restenosis than the "on label" patients did. "Clinicians should be cautious in extrapolating the results of pivotal clinical trials to routine clinical practice," the presentation concluded.
No one can say exactly why coated stents seem unusually plagued by clotting. Beyond the first month or two, bare-metal stents rarely clot, possibly because regenerated tissue has covered the stent and created a smooth channel for blood flow. But coated stents may clot for years. Cardiac pathologist Renu Virmani, medical director of a Maryland pathology lab and an authority on the interplay of forces inside stented arteries, has found that the potent drugs–paclitaxel is actually a cancer medication–seem to interfere with the healing process on the inner artery wall. Months after a coated stent is set in place, it is often not fully covered by new tissue. But Virmani is sure there are other reasons for the latest findings.
One that everyone agrees upon: Patients don't always follow orders. Because the body may react strongly to the foreign object, all stent patients are advised to take anticlotting drugs–aspirin and clopidogrel (Plavix)–for a month following a bare-metal implant, three months for Cypher, and six months for Taxus. Many cardiologists routinely keep coated-stent patients on clopidogrel for a year, often beyond.
But almost 1 coated-stent patient in 7 discontinues the drug in the very first month, according to a June paper in the journal Circulation. Early dropouts in the study were almost twice as likely to be rehospitalized during the rest of the first year and were more than 10 times as likely to die. Research published this week in the Journal of the American College of Cardiology found that in stent patients who took clopidogrel for six months and then stopped, the coated-stent group had far higher rates of deaths and heart attacks in the next 18 months.
"I have a little speech I give my patients," says Neal Kleiman, director of cardiac catheterization at Methodist DeBakey Heart Center in Houston and an expert on clopidogrel. "I tell them they'll be on it for a minimum of two years, maybe more. To make a big point of it, I say, 'Look, if you want to stop, talk to me personally first.'"
Cordis Corp., the Johnson & Johnson unit that produces the Cypher stent, and Boston Scientific, which manufactures the Taxus stent, concede that late thrombosis occurs more often with coated stents than with bare metal. "When it happens, it's a catastrophe," says Campbell Rogers, chief technology officer for Cordis. "We need to find ways to make it less likely." But both companies reject the claim that the rate is excessive, especially given that the patients might have fared worse if they hadn't had a stent put in at all. "Would they have ultimately had to have bypass surgery? If so, what if they had had one of the many complications of surgery?" says Rogers. "It's not just a matter of looking at the risk of one kind of stent compared to another. It's also about looking at all the alternatives."
Back to basics. One of those alternatives, says cardiologist Robert Califf, director of the Duke Translational Medicine Institute in Durham, N.C., is a return to basics. "Bare-metal stents aren't used often enough," he says. Newer versions, such as the Driver from Medtronic and Vision from Abbott Laboratories, are made from cobalt-chromium wire rather than the usual stainless steel. The stronger material allows thinner wires. That seems to cause less injury to the artery wall, which in turn apparently lowers the incidence of rampant tissue regrowth.
Meanwhile, however, some 80,000 people a month are joining the coated-stent ranks. The FDA, manufacturers, and physicians all desperately want them to take their prescribed clopidogrel. But the drug, which has no good alternative, costs $3 to $4 per daily tablet. Easy bruising is fairly common. And while the drug's antiplatelet ability lowers thrombosis risk, it raises the risk of excessive bleeding. Kenneth Kent, a Washington, D.C., cardiologist who helped test the first bare-metal stents in the late 1980s, recalls a patient who, for fear that his stent would clot, talked his gastroenterologist into letting him stay on clopidogrel prior to a colonoscopy rather than discontinuing it the usual five days beforehand. The doctor snipped out a small polyp for biopsy during the procedure; the patient bled profusely and had to be hospitalized.
While Kent believes that any added risk from coated stents is very small, he and his colleagues at the Washington Hospital Center have cautiously dropped the percentage of coated stents they implant from the high 80s to the low 60s over the past few months and are likely to drop it further. "There really should be a good reason for using these devices," he says–such as in a small vessel, which is more likely to reblock if a bare-metal stent is employed.
"I try to push off any elective procedure for at least a year after putting in a coated stent," says Phillip Horwitz, an interventional cardiologist at the University of Iowa Hospitals and Clinics in Iowa City. The possibility of future surgery pushes some cardiologists entirely into the bare-metal camp.
Both camps will get the chance to state their case this week. Indications are that, at a minimum, the panel will recommend stronger FDA language about selecting patients suitable for coated stents and extending the aspirin-plus-clopidogrel regimen to a year. Those present are likely to witness sharp questioning by panel members of manufacturers and physicians who use coated stents on their patients. The stakes, in lives and dollars, are high.
This story appears in the December 11, 2006 print edition of U.S. News & World Report.