July 25th, 2012
DES in Patients at Low Risk for TVR: Is the Benefit Worth the Cost? (Part I)
Gregg W Stone, MD, John A. Bittl, M.D., David Kandzari, MD, Richard A. Lange, MD, MBA and L. David Hillis, MD
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?
Decisions of whether to use DES versus BMS are, in almost all cases, directed by physicians. The decision, like all in medicine, is driven by the physician’s desire to achieve the best health outcome for his or her patient. In this regard, the principal considerations for most physicians (at least in the U.S.) are the safety and efficacy of the alternative therapies. Reducing societal costs is a laudable and important goal, but it is very difficult to factor in when taking care of an individual patient (and nowhere in the Hippocratic oath does it state that physicians should consider societal costs when caring for individual patients). Most studies show that first-generation DES decrease TVR by 50–60% compared with BMS, regardless of baseline risk. And current-generation DES are even more effective (and safer). Thus, if I have a patient at relatively low risk for TVR with BMS (10%), and I can reduce the risk to 3–5% with DES at no excess safety risk, then — to me — it is unethical not to use the most effective device for that individual patient, as long as it is at least as safe as the alternative. The Archives article did not consider numerous factors, including improvements in current-generation DES, the incremental cost-effectiveness of multiple DES for progressive atherosclerosis over time (the benefits are exponential), the fact that guidelines recommend dual antiplatelet therapy (DAPT) for 1 year after BMS as well as DES (and that clopidogrel now costs ≈25–50 cents a day, not 1 dollar, as modeled in the paper), and other issues. Yes, in some cases concerns about safety or relative lack of incremental efficacy should lead to BMS use (e.g., in a patient likely not to comply with DAPT, or for a large focal lesion in a saphenous vein graft). In summary, I do not believe that DES as a class are overutilized. Although there is no doubt that some physicians are using them in too high a percentage of cases (100%), others are clearly underutilizing these devices.
Stephen G. Ellis, MD. Professor of Medicine and Section Head, Invasive and Interventional Cardiology, The Cleveland Clinic Foundation
Very nice job, Gregg!
I am concerned only about your contention that first-generation DES reduce TVR by 50–60% compared with BMS. This is relative rather than absolute — hence potentially misleading — and in part driven by protocol-mandated follow-up angiography.
Also, but of lesser concern, did the authors of the Archives article get current stent costs right (DES<$1350, BMS<$500)?
Finally, in the current environment one might argue that we should balance ethical obligations to the patient with those to society (the Centers for Medicare and Medicaid Services [CMS] have set a precedent for this with dialysis), and hence strike a different balance than the one you submit. Different societies might choose different cost effectiveness ratios to best meet their needs.
Almost all trials (ENDEAVOR II being the one exception) have shown similar relative reductions with DES compared with BMS in patients in the angiographic follow-up and non-angiographic follow-up groups. I do agree, of course, that absolute reductions must be the principal consideration. Thus my example of a patient with a predicted TVR rate of 10% after BMS, which would be reduced to ≈3–5% by DES, a number needed to treat to prevent one TVR of ≈14–20. I would find it unethical to tell a patient that he may have a 1 in 15–20 incremental chance of clinical restenosis (TVR) if I have an equally safe alternative. And if I asked him if he wouldn’t mind accepting this additional risk to reduce societal costs that he would be incurring — he’d probably wonder why, since he is the patient, I am not putting his interests first.
I don’t know what the average DES and BMS selling prices are in the U.S.
If ethical obligations to the patient should be balanced with those to society, then the Hippocratic oath should be changed. CMS may have a societal obligation, but physicians have to care about the patients whose welfare rests in their hands above all else. If the physician is not the principal advocate for the patient, who will be? And consider how this change would erode the doctor-patient relationship, trust, and confidence!
A rate of 76.9% sounds correct. DES are not overused because:
1. They do what they are supposed to do. They strikingly reduce restenosis.
2. Each patient wants the best care. Almost all patients needing PCI will want a stent that will last as long as possible. The preponderance of data suggest that DES, especially everolimus-eluting stents, are safer and more efficacious than BMS.
3. Theoretical physician-doctor conversations about using BMS in the setting of low potential restenosis will go nowhere. No patient ever wants to hear a statistical discourse or a public-health discussion about “the potential for significant cost savings for the U.S. health care system while minimally increasing restenosis events”.
4. Costs are less of an issue since competitive bidding has driven down DES costs, and generic clopidogrel became available on May 17, 2012 (known as Independence Day in our medical community).
5. The reasons that DES should not be used in 100% of patients are obvious and include intolerance of prolonged DAPT, preexisting need for anticoagulation, scheduled noncardiac surgery, and several other clinical scenarios.
Physicians and patients drive DES use. Hospitals have different financial incentives, but they condone DES as part of their mission to deliver high-quality healthcare and achieve high rates of patient and physician satisfaction.
I do not believe the average utilization of approximately 75% is too low; in fact, I believe that a slightly higher percent would be more appropriate (e.g., 85%, but definitely not 100%). What is more relevant than utilization rate is utilization variability; at some institutions, DES are still used in only 40–50% of cases and at others, in nearly 100% of cases — neither of which is appropriate. The perceptions that DES are not effective in larger-caliber vessels and shorter lesions are quite old and have recently been discounted. Remember, a stent is for life, and TVR remains meaningfully lower with DES than with BMS, now through 5-year follow-up and beyond. The real driver of BMS use at present is concern over the need for DAPT adherence for 1 year with DES, and the efficacy of that mandate has never been established.
Although TVR may be more common (and therefore predictable) in patients with certain characteristics, there are no absolutes in medicine; the identification of which patients will develop clinical restenosis is never certain. This uncertainty is juxtaposed with clear evidence that DES reduce clinical restenosis by at least 50%. Most patients themselves would therefore select DES, and this discussion is part of the process of informed consent. Thus, both the doctor and the patient often participate in the decision to use DES.
Regarding the Archives paper, the wide variation in DES utilization is not unexpected and has been previously described. The analysis would probably have been more informative if it were limited to data from 2007-8 to the present, as practice patterns varied widely before 2006-7.
The authors’ estimate of outcomes of reduced DES use is limited to cost and does not account for patient-related quality of life or events related to restenosis (e.g., MI), neither of which should be discounted. The model did account for costs of repeat procedures, but not for those of potential clinical sequelae. The conclusion that the absolute rate of TVR would increase only by 0.5% seems like an artifact of the model or overfitting. Clinically, no interventionalist who practiced in the era of BMS only would claim that he or she would treat restenosis only 0.5% more often if BMS were used 50% of the time. BMS represented a marked advance over balloon angioplasty, but the results were not that good!
Finally, the costs of DES (and even clopidogrel) continue to erode, offsetting the cost equation further. As novel DES are introduced, and as we are seeing the best outcomes ever with existing DES and existing practices, perhaps a better question than whether to use DES versus BMS is whether newer, more expensive DES and antiplatelet agents are necessary.
It’s easy for us (and even seems logical) to produce commentary about value-based medical decision making — unless we are directly engaged in patient care or, more importantly, with the patients. The editorial perspective should remind us that what is a “no brainer” (as described by the editorialist) behind a desk may be very different to a practicing cardiologist at the bedside.
- A More-Effective Aldosterone Synthase Inhibitor for Resistant Hypertension?
- Can Digital Tools Improve Hypertension and Cholesterol Control in Diverse Populations?
- Statins vs. Supplements for Reducing Low-Density Lipoprotein Cholesterol
- CHEST Perioperative Guidelines for Antithrombotic Therapy
- Real-World Evidence in Anticoagulation of Atrial Fibrillation
Sign up for NEJM Journal Watch Cardiology Alerts
Receive a free weekly email with the latest summaries on Cardiology.
- ACS AF AHA anticoagulation aortic valve replacement apixaban aspirin atrial fibrillation CABG cardiovascular risk cholesterol clopidogrel dabigatran diabetes diet drug-eluting stents epidemiology ESC exercise FDA FDA approvals Fellowship training guidelines HDL heart failure hypertension ICDs MI myocardial infarction obesity PCI Primary PCI risk factors rivaroxaban statins STEMI stents stroke stroke prevention TAVI TAVR type 2 diabetes venous thromboembolism warfarin women