Nearly two years after the widespread introduction of drug-eluting stents (DES) on the U.S. market, it is perhaps time to step back and evaluate the impact of this revolutionary technology on the practice of cardiology. Financial modeling has suggested that the introduction of DES has turned PCI procedures from profit generators for hospitals into sources of loss.¹ The penetration of DES within the U.S. is approximately 70%, but varies from region to region.² In many metropolitan areas, adoption is nearly universal. The data for the use of DES are strong and incontrovertible, and the clinical benefit in trial populations is beyond debate. However, the use of DES now extends far beyond the clinical criteria of the randomized trials and includes patients who are at very low baseline risk for restenosis. In this issue of the Journal, Gottschall et al. report a simplified predictive score for target vessel revascularization (TVR) following bare metal stenting. In this prospective cohort study of 848 patients, a risk score for TVR was developed. The risk score was calculated by assigning a value to each of three variables: 1) diabetes mellitus: present = 1 point; absent = 0 points; 2) reference vessel diameter: < 3.0 mm = 2 points; 3.0–3.5 mm = 1 point; > 3.5 mm = 0 points; and 3) lesion length: < 10 mm = 0 points; 10–20 mm = 1 point; > 20 mm = 2 points. A linear and statistically significant increase was found between each score level (0 = 1.4%, 1 = 4.5%, 2 = 7.1%, 3 = 10.4% and 4/5 = 15.7%; r = 0.90; p < 0.001).
The arguments for near-universal use generally fall into one of three categories: data-driven arguments, arguments based on business considerations, and arguments that could be characterized as emotion-driven. Data-driven arguments are generally based on either logical extension of the data to populations not studied, or arguments that consider cost-effectiveness. Among the business considerations is a general feeling that not using DES universally will lead to a competitive disadvantage in the marketplace. Fear of litigation is also commonly discussed as an argument for their universal use. Finally, arguments regarding DES often lead to the emotion-laden question: Which stent would you use if it were your mother?
Even with the lack of DES in large sizes, operators routinely implant smaller stents in large vessels and postdilate them to match vessel size. Kuntz et al.³ analyzed data from the SIRIUS trial and found that regardless of baseline diabetic status, lesion length or reference vessel size, the treatment effect of rapamycin (event rate control – event rate SIRIUS/event rate control) is preserved. Some operators have interpreted this data to assume that DES will benefit all patients regardless of whether the patient’s lesion characteristics would have qualified for inclusion in the original trials. However, absolute reduction in target lesion revascularization was greatest in those patients at highest risk for restenosis based on clinical characteristics (25.4%), and was relatively small (5.7%) for those patients at lowest risk (lesions < 12 mm long in vessels > 3.0 mm in nondiabetic patients).3 Similar issues surround the cost-effectiveness of DES. Taking into account such factors as the incremental cost of DES over bare metal stents and the reduction in cost from repeat procedures, Greenberg et al.⁴ utilized data from the SIRIUS trial and calculated that the cost of DES is approximately $27,500 per quality-adjusted year of life gained. While this is an acceptable cost compared to other procedures, it is based on an assumption that patients treated will be confined to those who fit inclusion criteria for the trial. Using the predictive score described in the article by Gottschall et al., it would appear highly unlikely that a similar cost-effectiveness argument could be made for a patient population with a predictive score of 0 and a one-year risk of TVR of 1.4%. Assuming a treatment effect of 65%, approximately 200 patients in this population would need to be treated to prevent one case of repeat revascularization.
An issue that bears heavily on this discussion is the enigmatic economics of medical devices compared with other advanced technologies. Computers, cell phones, PDAs, plasma televisions get better and cheaper with each passing year. In contrast, medical devices in general get more expensive as they improve. The high cost of clinical research drives these expenses, motivated, in part, by a tort-conscious society. This anomaly may be a byproduct of an overall environment in which there is the expectation of absolutes in the delivery of healthcare, a philosophy that is certainly magnified by the medical liability issue.
However, even the most cynical of cardiologists would likely agree that fear of litigation or concerns regarding marketplace competition are not the best of criteria to consider in making healthcare decisions. Questions regarding what one would do for one’s own family member are not as easy to dismiss. However, despite the immediacy that these emotional questions bring to the discussion, individual decisions by necessity cannot be the critical issues in informing specific healthcare policy. Should efforts such as the simplified score described by Gottschall et al. in this issue of the Journal reinvigorate this debate? We suggest it should.