How Appropriate for Assessing Quality Are the 2009 Appropriateness Criteria for Coronary Revascularization?

From Advocate Illinois Masonic Medical Center, Chicago, Illinois. The author reports no conflicts of interest regarding the content herein. Address for correspondence: Lloyd W. Klein MD, Advocate Illinois Masonic Medical Center, Professional Office Building Suite #625, 3000 North Halsted Avenue, Chicago, IL 60614. E-mail: lloydklein@comcast.net

“Science is the father of knowledge, but opinion breeds ignorance.” — HIPPOCRATES Whether it is most appropriate to treat coronary artery disease (CAD) patients with medical therapy alone, or to prescribe medical therapy and also perform revascularization — either by percutaneous coronary intervention (PCI) or bypass surgery (CABG) — depends on a myriad of clinical, angiographic and physiologic factors that vary widely from patient to patient. In some cases the best decision is readily apparent based on objective evidence from clinical trials; but often, the best course to take is more nuanced and dependent on one’s interpretation of that individual’s clinical and angiographic findings, then placed in the context of existing clinical trial data. Concept of appropriateness. It seems intuitive that if one could create a matrix of a large number of relevant factors and submit each scenario to a panel of thoughtful, impartial practitioners from the various fields of cardiology, that their dispassionate consensus opinion would provide a practical tool for weighing each of those factors to arrive at the principles leading to the right treatment decision. This is the concept behind the development of the ACCF/SCAI/STS/AATS/AHA/ASNC 2009 Appropriateness Criteria for Coronary Revascularization.¹Appropriateness Criteria for Coronary Revascularization.¹ A technical panel consisting of cardiologists, surgeons, and numerous other stakeholders developed a ranked score of 1–9 for each of 180 clinical scenarios, scoring each according to whether revascularization was appropriate, inappropriate or uncertain. These Appropriateness Criteria were not intended to diminish the importance of clinical judgment in evaluating individual patients, nor was it possible to include every possible variable or influential fact which might correctly sway a clinical decision. The reasons for the development of these criteria include an initial attempt at objective measurement of what constitutes quality practice, to provide a framework for evaluating patterns of care and to help reduce the large variation in rates of revascularization that has been observed. Appropriateness criteria differ from clinical guidelines both in their purpose and their format. Clinical guidelines provide a summary and discussion of evidence from trials; appropriateness criteria focus on the presenting features and clinical status of patients of the type seen in contemporary daily practice. Clinical trials and published studies frequently exclude high-risk patient subgroups and, therefore, existing scientific evidence may not be directly applicable to therapeutic decisions. The Appropriateness Criteria begin by characterizing patients according to four critical features: 1) the severity and nature of symptoms; 2) the angiographic severity and location of coronary stenoses; 3) how much of the myocardium, according to noninvasive testing, is ischemic; and 4) whether the patient is receiving an optimal medication regimen at optimal dosages. Revascularization is considered appropriate if the expected improvement in survival, symptoms, functional status and/or quality of life outweigh the risks as compared to medical therapy. The panel considered revascularization by either CABG or PCI for most of the scenarios presented, without specific consideration of the method. The relative merit of CABG versus PCI was considered only in a subset. The “rules” that were mandated at the onset include: acceptance of the size of a thallium defect as the accurate means to assess the degree of myocardium at risk; presumption that the designation of the number of vessels diseased accurately depicted large branches of that vessel; supposition that there were no relevant medical problems or patient preferences that might influence such decisions; and allowance that the clinical evaluation of symptom severity was precise. Surprisingly to many interventionists, CABG was considered the more appropriate revascularization method for patients with 13 of 14 conditions and anatomic descriptors, including two-vessel disease with proximal left anterior descending (LAD) stenosis. CABG was also favored for patients with three-vessel coronary disease regardless of left ventricular (LV) function or presence of diabetes. In only one scenario — patients with native three-vessel disease with failure of multiple bypass grafts, left interior mammary artery patent to a native coronary artery, and depressed LV ejection fraction (LVEF) — was stenting considered appropriate and bypass surgery inappropriate. Given the actual utilization of these procedures in practice, the question arises whether the apparent “disconnect” between modern interventional practice and the “objective” analysis of the technical panel accurately depicts an overperformance of PCI. Conversely, it is possible that the mandated rules guiding interpretation resulted in a skewed outcome. Current study. In this context, the study in this issue of the Journal of Invasive Cardiology by Brener and colleagues² is important because it addresses the question of whether the appropriateness class for those patients who received PCI predicted 3-year outcomes. In their practice of over 2,000 PCIs performed in the years 2005–2007, 80% were classified as appropriate, 19.4% were uncertain and just 0.6% were inappropriate. The 900-day survival rate was 92.6% in the appropriate category, 91.3% in the uncertain group and 66.9% in the inappropriate group. Multivariate analysis showed that appropriateness class did not predict outcomes; only age and comorbidities were predictive. They conclude that, at least in those years “at a large-volume, academic and private practice,” there were very few inappropriate PCI procedures and those with “uncertain” appropriateness had similar outcomes to those with appropriate indications. The authors interpret these results, I think correctly, to mean that perhaps there are other important considerations in proper case selection than the anatomic extent of disease, particularly as technical proficiency and interventional tools improve in complex subsets. However, it must be stated that there are several very significant deficiencies in this study, a statement which is not intended as a negative appraisal of how the authors conducted the study, but rather a critical evaluation of how difficult it is to retrospectively evaluate this problem. The authors appear to have included only unprotected left main stenosis PCI as inappropriate, and classified all three-vessel CAD as uncertain; this is not how the Appropriateness Criteria were constructed: ACS, anginal class, diabetes status and LVEF are critical considerations. Only patients without prior CABG and with stable angina were included, and completeness of follow up was not given for the study cohort (but was 98% for the entire group). The only endpoint studied was survival, when obviously improvement in anginal status, freedom from repeat procedures and other events may be important as well. Further, these results are in conflict with those of Anderson and colleagues.^3,4 Using the National Cardiovascular Data Registry (NCDR), the authors showed in a large survey of contemporary PCI practice, as do Brener et al, that most procedures were performed for Class I indications. However, they found a significant relationship between evidence-based indications and in-hospital outcomes. Their conclusion, that “closer adherence to guidelines can reduce variations in care, can improve quality, and may ultimately result in better outcomes,”³ is one of the critical underpinnings of the Appropriateness concept. A follow-up study demonstrated that the ACC-NCDR risk-adjusted mortality model, combined with the PCI guideline5 indication scheme, produces mortality risk estimates by indications classes that are close to actual observed values.⁴ Hence, the current study runs counter to an established relationship and further raises suspicions regarding the classification of their cases. For these reasons, the authors’ conclusions should be regarded as interesting and provocative, but require much more detailed study and far more sophisticated analyses with a larger, multicenter database. It may be that the authors have identified some important problems with the criteria, but it is also possible that their case selection and/or categorization do not reflect current practice everywhere. A 7.4% mortality rate in PCI cases deemed “appropriate” over a 3-year follow-up period seems exceedingly high, and may represent a more aggressive case selection than the appropriateness ratings suggest. Are anatomic-based criteria the best structure for making clinical decisions? Despite the clear limitations of the study, the issue raised by the authors of whether subsequent versions of the Appropriateness Criteria should be anatomic-based, however, is a serious question that should be considered. The antiquated view that single-vessel involvement is not “surgical” disease, but that most three-vessel disease is preferentially a surgical problem, is simply obsolete. Yet, existing guidelines^5,6 sustain this approach regarding which revascularization method is best employed. The reason is simple: all of the surgical and medical literature up to now has been based on this simplified version of coronary disease extent. In theory, the severity and location of a coronary stenosis and the amount of myocardium at jeopardy should correlate closely with symptoms and angiographic and functional testing; unfortunately, in practice, these relationships are not infallible. The problem of course is that in the “real world,” no one knows which test result to rely upon when there are disparate results, clinical manifestations are often difficult to characterize and angiographic interpretation is a subjective art. Although the “rules” of the Appropriateness Committee were set so that such technical issues could not derail the process, the fact is that at least three factors are almost always open to interpretation in making a clinical judgment: 1) the limitations of evaluating an angiogram in determining a “significant” stenosis, including the difficulties of defining the reference-vessel size and the presence of diffuse disease; 2) the definitions of medically refractory angina and optimal medical therapy; and 3) the role of functional tests in demonstrating the extent and severity of ischemia among patients with minimal or difficult-to-interpret symptoms. There are several situations in which everyone agrees that the angiographically derived number of vessels diseased is a secondary question. In the modern management of ST-elevation myocardial infarction (MI) and acute coronary syndromes (ACS), the usefulness of PCI overcomes all simplistic anatomic categorization, regardless of LV function and number of vessels diseased. In the setting of cardiogenic shock, there are only rare situations in which opening the ischemic vessel as rapidly as possible is not the primary concern. When patients are admitted with acute ischemia, especially with hemodynamic or electrical instability, the most appropriate revascularization method is determined primarily by the technical aspects of opening the ischemia-related artery, not the presence or absence of disease in the other vessels. The reason is that reestablishing normal flow and relieving ischemia quickly is the important consideration, and speed is a major benefit of PCI. In contradistinction, when managing stable angina patients, the key questions are symptom relief and long-term survival. Even when the stenosis responsible for chest pain or the largest territory at risk can be identified, it’s approachability by PCI or CABG is a secondary concern to how amenable the coronary arteries are to complete revascularization and whether LV function will improve after ischemia is relieved compared to medical therapy. However, there are no large-scale contemporary studies that compare CABG to medical therapy in stable angina. The existing CABG versus medical therapy randomized trials were performed over 3 decades ago, used a comparator “medical therapy” group that today, in a post-COURAGE7 world, would be considered inadequate and presumed that angiographic lesion severity was the best — indeed only — accurate diagnostic method. The indications for CABG were first defined by the results of the Coronary Artery Surgery Study (CASS).⁸ The marked degree of extrapolation from CASS and other early surgical studies to present-day practice becomes apparent when one reviews these studies. In CASS, patients 35%, had evidence of anatomically significant CAD, the coronary artery to be bypassed supplied viable myocardium and at least 3 weeks had passed since the last MI. All patients were considered operable and had mild stable angina pectoris or were free of angina after infarction. Just 780 patients with stable ischemic heart disease were randomly assigned to receive surgical (n = 390) or nonsurgical (n = 390) treatment. Annual mortality rates in patients with single-, double- and triple-vessel disease who were in the surgical group were 0.7%, 1.0% and 1.5%, respectively; the corresponding rates in patients in the medical group were 1.4%, 1.2%, and 2.1% (p = NS). There was also no difference in rate of MI at 5-year follow up. A post-hoc analysis (which was not prespecified) was then performed, which identified two subgroups that had a significant (p 50% diameter narrowing in branches of undefined size in all three arteries based on a subgroup of 11 suggested that the benefits of revascularization remain formidable compared to modern medical therapy, no correlation with disease extent and severity was attempted, and no distinction between CABG and PCI formulated. There is no way to go back now to repeat these studies, confusing the situation irreparably. Although all of the existing randomized trials over the past 15 years comparing PCI to CABG in multivessel CAD concluded that CABG was associated with fewer long-term events, the improvement in survival is limited to the diabetic population, and the benefits demonstrated are decreased angina and fewer repeat procedures.^12,13 The reality is that for most multivessel CAD patients, the decisive factor ought to be the relative merits and limitations of each procedure for that individual patient. The major advantage of CABG is the ability to achieve complete revascularization, particularly in the setting of chronic total occlusion. This feature allows for the application of CABG to anatomic subsets which are not approachable with PCI. A second major advantage of CABG, namely the superior durability of its results, is at least partly structural, in that bypass conduits protect territories, whereas stents treat lesions.¹³ Its drawbacks include a relatively long recuperation period and a significant incidence of morbidity. In contrast, the major advantages of contemporary PCI are speed of achieving normal or near-normal perfusion and relatively minimal morbidity.¹² Because of these factors, patients with an acute MI can often be rapidly stabilized by a single-lesion procedure, and more completely revascularized by either PCI or CABG during subsequent procedures, which they undergo at much lower risk (because ischemic, or hemodynamic, or electrical instability have been relieved). More residual angina and more repeat procedures are the critical shortcomings of PCI. Ideally, the informed patient should decide which set of advantages and disadvantages to select when long-term mortality is otherwise similar. Eliminating this element from the Appropriateness framework, in my opinion, is its most significant limitation. Is angiographic anatomy the optimal framework? The quality practice of cardiology is not as straightforward as strict adherence to the guidelines as determined by number of vessels diseased. There are frequently complexities that cannot be addressed in randomized trials or resolved by committees. How much to weigh various additional clinical and angiographic factors in one’s decisions is the foundation of good clinical judgment. How does one apply clinical trials performed in the most favorable cases to the elderly patient, to those with substantial comorbidities, recent MI, or with severely impaired LV function? Judgments are routinely necessary concerning how amenable to bypass are particular small and intermediate-sized vessels, vessels located on the posterior surface or near the atrio-ventricular groove that may be difficult for a surgeon to locate, those that are totally occluded and are not well visualized, or that fill by collateral vessels (and hence may be severely diffusely diseased or be underfilled). How many vessels will a surgeon actually be able to bypass, versus how many can an interventionist treat? A frequent challenge arises in interpreting focal borderline lesions in the setting of severe diffuse disease. Then, there is the question of regional LV function: how large is the territory of myocardium being revascularized? Is the dysfunction complete and not reversible, or is it partially viable and/or ischemic? How much improvement in LV function will be attained with complete revascularization, and is it worth the risk to the patient? Clearly, these considerations were not evaluated in CASS or the other surgical studies of the 1970s, when the simplistic single-vessel versus two-vessel versus three-vessel categorization was developed. The Appropriateness Criteria are intended to apply when these details are not substantial considerations, but the committee did not evaluate the proper course when they are a subject of concern, which is a substantial cohort of the CAD population. Although the designation of the number of vessels diseased is now understood to be an oversimplification of anatomic extent of disease, the value of the SYNTAX score1⁴ had not been published at the time of the deliberations of the panel, and thus could not be included in the discussions. Nevertheless, it was entirely predictable¹³ that CABG would remain the revascularization procedure of choice in patients with very extensive disease, especially in diabetics, while PCI would be at least equivalent for patients with less extensive disease. However, the divergence between the existing anatomic-based practice guidelines and patient-centered clinical decision making could not be more apparent than upon reflection of recent reconsiderations of unprotected left-main PCI.¹⁵ A problem that future studies need to address is how to describe the extent of disease in objective, quantitative terms that are intuitive and clinically accessible, that is, available on a hand-held device¹⁶ or easily calculated without a complex formula. Anyone who attends a weekly cardiac catheterization conference has probably observed the quandary that anatomy-based decision making encourages: the interpretation of angiographic lesion severity is appallingly conditional. I have watched in astonishment as angiograms are presented that are interpreted by cardiac surgeons (who wanted to operate) as showing three > 70% stenoses, while noninterventional cardiologists (who wanted to treat medically) describe three Is there a better alternative? It is time to seriously consider a different model to regulate revascularization decision making. Instead of emphasizing coronary anatomy based solely on subjective angiographic interpretation, more credence should be given to other anatomic, functional and pathophysiologic methods.^12,17 The complementary, if not definitively diagnostic, roles of intravascular ultrasound, fractional-flow reserve, magnetic resonance imaging and multislice computed tomographic scans should become as crucial a basis for decision making as the angiogram. Although the PCI guidelines⁵ and the Appropriateness Criteria do incorporate a few of these modalities to some degree, they do not establish clear indications when each should be used, or how to evaluate results that are divergent from the angiogram. Although the traditional reliance on stress imaging continues to have significant value,18 its subjective, nonquantitative estimate of size of jeopardized myocardium and its relative inability to distinguish significance of lesions other than the most severe, must be recognized in future criteria. Equally important, there must be recognition of the advances in medical therapy and risk-factor modification. The old-school notion that medical therapy is indicated only when revascularization is too risky, and hence the last resort, is out-dated and not grounded in current realities. If medical therapy is the comparator against which more aggressive options have to prove themselves, it must be acknowledged that the days when medical therapy was merely a fancy appellation for a control group are long over.7 For example, the appropriateness of PCI of a single, significant proximal LAD lesion with objective evidence of ischemia in the absence of optimized medical therapy is “uncertain” because there is no study demonstrating that PCI saves lives or improves angina compared to optimized medical therapy, even though not a single interventionist can be found who does not think such a situation is highly “appropriate” for PCI. In part, COURAGE demonstrated what interventionists were slow to perceive; surgeons likewise will be dismissive, but if CASS were repeated today, I suspect a somewhat different outcome would be found. One of the most important principles implicit in clinical decision making is the paradox that many of the same descriptors of likely benefit from revascularization (including extent of ischemia, ventricular function and extent of anatomic coronary disease) are also descriptors of risk of procedural failure and/or complications. Based on this probably absolute fact of therapy, coupled with the widely touted notion that excellence as a PCI operator derives from doing large volumes of procedures with low rates of complications, as advocated in the PCI Guidelines,⁵ there is implicit encouragement for operators to perform large numbers of low-risk procedures and to avoid high-risk cases. If these principles become the profession’s ultimate statement of quality of care, the ensuing consensus documents can be used as tools to subject operators who choose high-risk cases, and those who work in community hospitals, to extinction. One of the potential benefits of Appropriateness Criteria is to correct the misleading concept that quality = high volume = low complication rate; appropriate case selection is also a critical determinant of quality, and any instrument used to compute quality must acknowledge that complications are more apt to occur in high-risk subgroups, including those who are appropriately treated. That is why a risk-adjusted rate of complications is the crucial measurement in assessing PCI quality, and is the standard promulgated by the NCDR^3,19 and overtly endorsed by the PCI Guidelines.⁵ Yet many interventionists are skeptical of the risk-adjustment process, to some degree appropriately, because any model is limited by data accuracy and the assumption that the collected variables are predictive in all patient subsets. For these reasons, it is my fervent hope that the study of Brener et al, for all of its faults, will help stimulate further in-depth study and lead to better, more “appropriate” Appropriateness Criteria.

References

1. Patel MR, Dehmer GJ, Hirshfeld JW, et al. ACCF/SCAI/STS/AATS/AHA/ASNC Appropriateness Criteria for Coronary Revascularization. J Am Coll Cardiol 2009;53:530–553. 2. Brener SJ, Haq SA, Bose S, Sacchi TJ. Three-year survival after percutaneous coronary intervention according to appropriateness criteria for revascularization. J Invasive Cardiol 2009;21:554–557. 3. Anderson HV, Shaw RE, Brindis RG, et al. Relationship between procedure indications and outcomes of percutaneous coronary interventions by American College of Cardiology/American Heart Association Task Force Guidelines. Circulation 2005;112:2786–2791. 4. Anderson HV, Shaw RE, Brindis, RG, et al. Risk-adjusted mortality analysis of percutaneous coronary interventions by American College of Cardiology/American Heart Association Guidelines recommendations. Am J Cardiol 2007;99:189–196. 5. Smith SC, Feldman TE, Hirshfeld JW Jr, et al. ACC/AHA/SCAI 2005 Guideline Update for Percutaneous Coronary Intervention-Summary Article: A Report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (ACC/AHA/SCAI Writing Committee to Update the 2001 Guidelines for Percutaneous Coronary Intervention). J Am Coll Cardiol 2006;47:216–235. 6. Eagle KA, Guyton RA, Davidoff R, et al. ACC/AHA 2004 guideline update for coronary artery bypass graft surgery: A report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (Committee to Update the 1999 Guidelines for Coronary Artery Bypass Graft Surgery). Circulation 2004;110:e340. 7. Boden WE, O’Rourke RA, Teo KK, et al for the COURAGE Trial Research Group. Optimal medical therapy with or without PCI for stable coronary disease. N Engl J Med 2007;356:1503–1516. 8. Myocardial infarction and mortality in the coronary artery surgery study (CASS) randomized trial. N Engl J Med 1984;310:750–758. 9. Chaitman BR, Ryan TJ, Kronmal RA, et al. Coronary Artery Surgery Study (CASS): Comparability of 10 year survival in randomized and randomizable patients. J Am Coll Cardiol 1990;16:1071–1078. 10. Passamani E, Davis KB, Gillespie MJ, Killip T. A randomized trial of coronary artery bypass surgery. Survival of patients with a low ejection fraction. N Engl J Med 1985;312:1665–1671. 11. Rogers WJ, Bourassa MG, Andrews TC, et al for the ACIP Investigators. Asymptomatic Cardiac Ischemia Pilot (ACIP) Study: Outcome at 1 year for patients with asymptomatic cardiac ischemia randomized to medical therapy or revascularization. J Am Coll Cardiol 1995:26:594-605. 12. Morrison DA. Extent of voronary snatomy and left ventricular function in choosing between PCI and CABG, Chapter 4. In: Topol EJ (ed). Textbook of Interventional Cardiology, 5th Edition. WB Saunders: Philadelphia, 2007. 13. Klein LW. Are drug eluting stents the preferred treatment for multivessel coronary artery disease? J Am Coll Cardiol 2006:47:22–26. 14. Sianos G, Morel MA, Kappetein AP, et al. The SYNTAX Score: An angiographic tool grading the complexity of coronary artery disease. EuroInterv 2005;1:219–227. 15. Moussa ID, Wong SC, Feldman T. Coronary revascularization for patients with unprotected left main coronary artery disease: Evidence, guidelines, and judgment! Making clinical decisions in 2009. Catheter Cardiovasc Intervent 2009;74:448–458. 16. Accessed at https://www.syntaxscore.com/August 19, 2009. 17. Kern MJ, Lerman A, Bech JW, et al. Physiological assessment of coronary artery disease in the cardiac catheterization laboratory: A scientific statement from the American Heart Association Committee on Diagnostic and Interventional Cardiac Catheterization, Council on Clinical Cardiology. Circulation 2006;114:1321–1341. 18. Shaw LJ, Berman DS, Maron DJ, et al. Optimal medical therapy with or without percutaneous coronary intervention to reduce ischemic burden: Results from the Clinical Outcomes Utilizing Revascularization and Aggressive Drug Evaluation (COURAGE) Trial Nuclear Substudy. Circulation 2008;117:1283–1291. 19. Klein LW, Kolm P, Dey SK, et al. Assessing coronary interventional program quality. The CathPCI Registry mortality ratio: A preliminary report from the American College of Cardiology-National Cardiovascular Data Registry. JACC Cardiovasc Interv 2009;2:136–143.