CardioExchange Archive

In a recent article in Archives of Internal Medicine, researchers performed an analysis of current use of drug-eluting stents (DES) in patients at various levels of risk for target-vessel revascularization (TVR), and estimated the cost and clinical outcomes of using BMS rather than DES in patients at low risk (see News). To gauge reaction to this report, Rick Lange and David Hillis asked a panel of experts to respond to the following questions:

1. In patients who need PCI, are drug-eluting stents (DES) overused?

2. Why or why not?

3. What drives DES use: physicians, hospitals, and/or patients?

This blog, Part III of a three-part series, contains responses from two of the Archives study authors to the panel’s remarks in Parts I and II. Part I contains the responses of four U.S. physicians, and Part II contains a European perspective.

John Spertus, MD, MPH

I appreciate and applaud the opinions of Drs. Stone, Ellis, Bittl, Kandzari and Serruys. Our recent article did not criticize the phenomenal advances in percutaneous coronary revascularization that you and others have made. I believe that the evolution in stent technology has been phenomenal, and the reductions in the need for repeat procedures have plummeted as treatment has progressed from balloon angioplasty to BMS to DES and now to new-generation stent platforms. Although current restenosis rates are about 50% lower with DES than with BMS (which is terrific), the absolute risk reduction is highly dependent upon patients’ underlying risk for restenosis. With a predicted TVR rate of >20%, you need to treat only 6–10 patients for one not to have to undergo a repeat procedure within a year. However, when the risk is <10%, you have to treat 25–130 patients for one not to return for another procedure (average NNT, ≈35) — all of whom would need to take prolonged dual antiplatelet therapy (DAPT) — for the one to benefit. If there were no downside of DES, then this would be a “no-brainer,” and all patients should get a DES.

The critical issue, however, is that there is a downside. DAPT is needed for much longer after DES than after BMS. Some may argue that DAPT should be used in all stented patients (or even all patients with coronary artery disease, if you believe the subset analysis from CHARISMA), but few would disagree that if a patient were having trouble taking DAPT, we would be far more comfortable stopping the thienopyridine in a BMS recipient than in a DES recipient.

From the patients’ perspective, the need for prolonged DAPT is a big deal. It is associated with a 1%–2% risk for major bleeding (requiring hospitalization and transfusion) and a MUCH higher rate of nuisance bleeding (the kind that happens when you nick yourself shaving and have to hold a paper towel against your chin for half an hour to get the bleeding to stop). We have recently reported that in patients on DAPT vs. aspirin alone, the absolute difference in nuisance bleeding is ≈15%, meaning that 1 out of every 6–7 patients on DAPT, as opposed to aspirin, will complain of easy bleeding or bruising. Moreover, this is independently associated with significantly worse quality of life, as captured by the visual analogue scale of the EQ-5D.

DAPT poses other challenges to patients. For example, if we assume medical insurance coverage and a $30 copay (choose $10 if you believe that generic clopidogrel is now in the lowest tier of copays — my lowest tier is $20/month), then a stable angina patient is looking at $360 for the first year of therapy with DES, as compared with $30 for a month of DAPT after a BMS. For many patients, this extra $330 (or $110) cost over 1 year is important. Moreover, if they need elective surgery, it must be delayed.

It is my contention that if you are really interested in providing the best treatment for a specific patient, you need to explicitly weigh the benefits (lower rates of TVR, based upon the person’s individual risks, e.g., as estimated by the TVR Risk Model that we published last year) against the costs and hassles of prolonged DAPT. Most importantly, these decisions need to reflect patients’ preferences, not physicians’. We are guides for our patients. We need to transparently explain the options, and the risks and benefits of those choices, based upon patients’ individual risks. If you told a patient that he or she has a 97% chance of a durable, successful procedure with a DES versus a 94% chance with a BMS, but that it would require the costs, bleeding hassles, and potential delays in surgery associated with DAPT over the next year, and he or she chose a DES, then that is terrific. On the other hand, if he or she would prefer a BMS, then we should respect that opinion and treat accordingly.

Our paper lays the foundation for advancing the field. We have defined lots of variation among physicians in the use of DES and demonstrated little absolute difference in rates of DES use as a function of patients’ individual risks for TVR. This suggests that the use of DES varies more by whom patients see than by who they are. Our findings suggest that if you presented the options of DES and BMS, along with a clear articulation of the benefits and risks, to patients at the lowest risk for TVR, and if half chose a BMS, then the country could save almost a quarter of a billion dollars a year. This would not be a result of “rationing” healthcare resources, or a “violation of our Hippocratic oath,” but rather a reflection of the preferences and values of our patients. What is wrong with that?

Robert W. Yeh, MD

Many of the comments put forth by Drs. Stone, Ellis, Bittle, Kandzari and Serruys resonate strongly with me. They are precisely the things that I think about as a practicing interventional cardiologist trying to make the best decisions for my patients.When it comes to balancing our obligation to individual patients with what I believe to be our duty to consider the societal impact of our decisions, I very much agree that there are few “no brainers.”  On the contrary, these difficult decisions require the full engagement of our minds.

The Hippocratic Oath that I swore during my graduation from medical school included the following: “I will remember that I remain a member of society, with special obligations to all my fellow human beings, those sound of mind and body as well as the infirm.” And although determining how to integrate dual obligations to individual patients and society is daunting, I don’t believe that doing so should be considered a violation of that oath. I would contend that it is precisely because of the likelihood that it will be administrators behind desks, not cardiologists at the bedside, who decide how we can care for our patients that physicians, as a group, should willingly engage in the difficult decisions that might improve healthcare sustainability.

Many have pointed out the limitations of our analysis. It is bound in time. It is subject to a number of assumptions. It did not consider all events or costs (which, as Dr. Spertus has pointed out, fall on both sides of the ledger). But I would submit that there are several things that we can agree on:

1. DES significantly reduce restenosis, but the absolute benefit is significantly less in patients at low risk for restenosis.

2. Despite the wording of the guidelines or what data from recent studies (see here and here) might show about EES, in practice, the duration of DAPT is longer in DES recipients than in BMS recipients — something that seems to be particularly ingrained in U.S. cardiologists. Thus, in practice, DES recipients have faced and continue to face the excess costs as well as the excess bleeding risks associated with DAPT.

3. We cannot perfectly predict adverse events — including not only TVR and related events, but also the development of bleeding, the need for urgent surgery, the requirement for warfarin, and the patient’s failure to adhere to medical therapy.

4. DES cost more than BMS.

5. There exists some threshold of cost per TVR avoided above which using a newer stent technology may not be worth it, whether the comparison involves DES versus BMS or newer-generation DES versus existing DES.

If we can agree on these things, isn’t it conceivable that a reduction in DES use isolated to the lowest-risk patients might actually be both consistent with patient goals and also be cost-saving for society?

Finally, I think it is possible that DES are both overused and underused. As we’ve written, they may be overused in patients who are not predicted to derive a large absolute benefit from them (of note, the mean TVR rate in the low-risk group was 6%–7%; 10% was just the upper limit). But they may also be underused in moderate- and high-risk patients who receive BMS (≈20% in our study).

Thanks for everyone’s great comments.