Friday, September 08, 2006

Were Late Stent Thromboses Really So Unexpected?

A flurry of report from the World Cardiology Conference in Barcelona, Spain have raised doubts about the safety of the currently very popular drug eluting stent (DES) as a treatment for coronary artery disease (CAD). Stents are used to maintain the patency of coronary arteries that are expanded after angioplasty (use of a balloon to expand the artery). The drugs eluted by the stents are meant to slow the growth of new tissue over the metal stents. Such growth can lead to narrowing of the artery.

However, apparently two studies reported at the Conference suggested that DES are more likely to incur blood clots than the older "bare metal" stents. Per the Newark Star-Ledger, "A Swiss-Dutch study tracked 8,146 patients and found that recipients of drug-coated stents were at increased risk of thrombosis, or blood clots, that occasionally can result in death. Two other Swiss studies, analyses of presented and published information discussed at the cardiology conference, also found that first-generation drug-coated stents had higher links to thrombosis compared to bare-metal stents."

Also, data has begun to surface from the stent manufacturers raising the same concerns. "Denis Donohue, vice president of clinical and regulatory affairs at J&J's (Johnson & Johnson) Cordis devices unit, said late stent thrombosis was an issue but there was not a clear safety signal. The company's own data showed long-term mortality rates of 6.5 percent for patients with a Cypher stent against 5.1 percent for bare-metal stents, but this was not a statistically significant difference. J&J submitted its latest data to the U.S. Food and Drug Administration last week, Donohue said."

In addition, according to the Boston Globe, Boston Scientific just revealed that its DES is associated with "a statistically significant increase in such clotting, or 'late thrombosis,' in patients receiving its stent. Compared with patients getting bare-metal stents, an additional one in 200 is likely to develop a blood clot, the company said."

The Globe also noted that Boston Scientific's Taxus stent "has been implanted in more than 2 million people worldwide," and that "the drug-coated stent is one of the most profitable medical devices in the word, bringing in more than $5 billion a year for its two main manufacturers, Boston Scientific and Johnson & Johnson."

It seems like almost every week we hear about some popular drug or device is found to have late adverse effects which no one heretofore expected. But should they have?

I took a quick look at the major published study that compared the Boston Scientific TAXUS stent with the bare-metal stent. (Stone GW, Ellis SG, Cox DA et al. A polymer-based, paclitaxel-eluting stent in patients with coronary artery disease. N Engl J Med 2004; 350: 221-231.)

The study was a randomized trial. Follow-up was nine-months. Major outcomes included death from cardiac causes, myocardial infarction (heart attack), and the need for a repeat procedure to re-open the vessel. The rate of all cardiac deaths was actually higher in the TAXUS group, 1.4% than in the bare-metal group, 1.1%. The paper did not report total mortality in either group. The article did not specifically address any other possible adverse effects of the stents, the procedure to implant them, or the drugs given to prevent blood clots. The study was funded by Boston Scientific. Five of its authors had financial arrangements with Boston Scientific, and another worked for the company.

So a strict evidence-based medicine reading revealed the study did not follow patients for a long time, so could not assess long term outcomes, good or bad. Furthermore, the article failed to report the total mortality rates of patients receiving different stents. This is a particularly curious omission in a study of a treatment meant to prevent some catastropic events that can be fatal. Thus, not only did the study fail to prove that the DES decreased mortality relative to the bare-metal stent, but it could not refute a hypothesis that the DES increased mortality. Finally, by failing to report various possible adverse effects of therapy, it could not refute a hypothesis that the DES had more adverse effects than the bare-metal stent. In summary, a strict EBM reading of one important article about drug-eluting stents would have raised doubts about whether the benefits of using DES outweighed its possible harms, and about whether DES are superior to other treatments of coronary artery disease.

Furthemore, in retrospect, one wonders why such an incomplete article appeared in print? One wonders whether unintentional bias was introduced because the company which had the most to gain from success of the DES sponsored the study, and because so many of the authors had financial ties to that company?

The lessons, again (see previous post), are that physicians must be very, very skeptical about the results of research sponsored by those with vested interests in the research turning out a certain way, and/or by research done by investigators with financial interests in organizations with such vested interests.

1 comment:

Lee Green MD MPH said...

This issue is discussed on HeartWire (, free registration required) this week too. It's a familiar smell: researchers had exceptional difficulty obtaining the actual data from the manufacturers who funded the studies.

It's human nature. These companies and people have invested hundreds of millions of dollars in developing DES. They stand to make a fortune or lose everything depending on how the studies come out. They will do everything they can to put their thumbs on the scale.

An accurate picture of the value of drug-eluting stents will be obtained only when studies are absolutely firewalled against influence by either manufacturers or procedure-oriented specialists whose incomes, prestige, or both are at stake. The people who don't want that to happen are moneyed and politically connected, and it will be a real fight to bring it about.