Harmonizing the Paradigm With the Data in Stable Coronary Artery Disease: A Review and Viewpoint
2017; Wiley; Volume: 6; Issue: 11 Linguagem: Inglês
10.1161/jaha.117.007006
ISSN2047-9980
AutoresJoshua D. Mitchell, David Brown,
Tópico(s)Cardiac Imaging and Diagnostics
ResumoHomeJournal of the American Heart AssociationVol. 6, No. 11Harmonizing the Paradigm With the Data in Stable Coronary Artery Disease: A Review and Viewpoint Open AccessReview ArticlePDF/EPUBAboutView PDFView EPUBSections ToolsAdd to favoritesDownload citationsTrack citations ShareShare onFacebookTwitterLinked InMendeleyReddit Jump toOpen AccessReview ArticlePDF/EPUBHarmonizing the Paradigm With the Data in Stable Coronary Artery Disease: A Review and Viewpoint Joshua D. Mitchell, MD and David L. Brown, MD Joshua D. MitchellJoshua D. Mitchell Cardiovascular Division, Washington University School of Medicine, St. Louis, MO and David L. BrownDavid L. Brown Cardiovascular Division, Washington University School of Medicine, St. Louis, MO Originally published13 Nov 2017https://doi.org/10.1161/JAHA.117.007006Journal of the American Heart Association. 2017;6:e007006Stable coronary artery disease1 (CAD) or stable ischemic heart disease2 are terms preferred in Europe and the United States, respectively, that generally refer to the same patients––those with angina, its equivalent, or no symptoms who experience episodes of reversible myocardial supply:demand mismatch in the absence of acute myocardial infarction (MI) or unstable angina. Although the exact number is unknown, nearly 300 000 percutaneous coronary interventions (PCIs) for stable CAD are performed annually on inpatients in the United States.3, 4 A significant and increasing number of additional PCIs for stable CAD are performed on outpatients, but these numbers are not available. Despite randomized controlled trials and meta‐analyses of these trials demonstrating that an initial strategy of PCI for stable CAD does not improve survival or prevent MI beyond what is achieved by optimal medical therapy (OMT),5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18 more than half of these procedures are performed on patients not treated with OMT.19 Numerous explanations related to physician behavior or the healthcare environment have been advanced to explain the deviation of practice from the evidence.20 However, a broader perspective indicates that the current stable CAD paradigm wherein the epicardial stenosis is the proximate cause of angina, ischemia, MI, and death and those outcomes can be prevented by revascularization is the product of 500 years of scientific thought.History of the Current ParadigmAlthough atherosclerosis has been described since antiquity, the intellectual birth of the epicardial stenosis paradigm can be traced to the early 1500s when Leonardo da Vinci recognized the degeneration of blood vessels with age during autopsies he personally performed. He described the continuous narrowing of the lumen associated with the "thickening of the coats" of arteries.21 While he determined that this process led to death in some individuals, he attributed it to the natural aging process rather than disease.21 Unfortunately, his observations were recorded in private journals that were not published until the 1800s.In 1698, 70 years after Harvey conceptualized the circulatory system,22 Chirac demonstrated the importance of coronary blood flow on cardiac function when he observed that the canine heart stopped beating after the coronary arteries were ligated (Figure 1).23 William Heberden subsequently characterized the syndrome of angina pectoris in 176824 but was unable to connect it to Chirac's previous work or to determine the underlying cause.Download PowerPointFigure 1. Historical development of the epicardial stenosis paradigm. CABG indicates coronary artery bypass graft; CAD, coronary artery disease; PTCA, percutaneous transluminal coronary angioplasty.The link to atherosclerosis, that da Vinci had initially hinted at in the 1500s, would be discovered by Edward Jenner when he first ascribed angina to underlying atherosclerosis after correlating clinical symptoms with autopsy findings.24 In a 1777 letter to Heberden, Jenner described "a kind of firm, fleshy tube, formed within the vessel, with a considerable quantity of ossific matter dispersed irregularly through it."24 Jenner's schoolmate, Caleb Parry, reasoned that a supply:demand mismatch during exertion caused angina, likely related to flow limitations through the obstructed coronary arteries that Jenner described.25 Ironically, Jenner's mentor, the renowned anatomist John Hunter, developed angina and died suddenly after an argument at St. George's Hospital in 1793.24 Jenner noted marked atheroma during Hunter's autopsy.25In 1809, Allen Burns, a lecturer in anatomy in Glasgow, elaborated on Parry's ideas and asserted that myocardial ischemia was the likely cause of angina pectoris.25 This theory would be linked to electrocardiographic findings in 1928 when Feil and Siegel attributed ST‐ and T‐wave changes during exercise in patients with angina to a decrease in blood flow to the heart.26The primacy of the epicardial obstruction was emphasized in a review of 355 cardiac autopsies in 1941 that noted that every patient who developed angina had an occlusion or marked narrowing of a coronary artery.27 The serendipitous discovery of coronary angiography by Sones in 1958 allowed visualization of epicardial CAD in living patients for the first time.28 Subsequent landmark angiographic studies documented the natural history of epicardial CAD and its adverse impact on survival.29 Around the same time, Robb and coworkers demonstrated reduced survival in patients with an abnormal stress test.30 Other studies correlated abnormal exercise stress test findings with the presence of obstructive CAD.31With the first human coronary artery bypass by Goetz in 196032 and the first percutaneous transluminal coronary angioplasty (PTCA) by Grüntzig in 1977,33 physicians could, for the first time, bypass or dilate the epicardial stenosis and, consistent with the existing paradigm, alter the course of the disease. By the early 1980s, through a unique collaboration of physicians, hospitals, and industry that has been referred to as the medical‐industrial complex,34 an efficient system had been created to identify and treat the millions of patients with obstructive CAD. Despite early warnings that the enthusiasm for revascularization surpassed the evidence,35, 36 coronary bypass surgery and PCI soon became two of the most commonly performed procedures in the United States and throughout the world.Challenges to the ParadigmPrevention of Death and MIThe first trial comparing PTCA with medical therapy in patients with stable CAD, the 1992 ACME (Angioplasty Compared to Medicine) study randomized 212 patients with single‐vessel CAD to PTCA or medical therapy.5 PTCA resulted in greater freedom from angina (64% versus 46%) and exercise tolerance at 6 months but did not reduce mortality or MI, although the study was not powered for these outcomes. Subsequent trials were consistent in demonstrating greater but far from universal short‐term angina relief with PTCA than medical therapy but no reduction in mortality or MI.6, 7, 8, 9, 10The results of the early studies led some to argue that the failure of PTCA to improve hard outcomes (death or MI) was attributable to the inclusion of patients at low risk. MASS (Medicine Angioplasty and Surgery Study), therefore, only included patients with proximal left anterior descending coronary artery (LAD) stenosis, yet still found no reduction in mortality or MI with PTCA.6 These early PTCA trials came before the advent of stenting; medical therapy in this era was limited to antianginal medications such as β‐blockers and nitrates.Among these early trials, the RITA‐2 (second Randomised Intervention Treatment of Angina) was the first large (>1000 patients) investigation to compare PTCA with medical therapy. Despite medical therapy that was rudimentary by today's standards, death or definite MI were less frequent with medical therapy than PTCA (3.3% versus 6.3%) (relative risk, 1.92; 95% confidence interval [CI], 1.08–3.41 [P=0.02]). Angina improved in both groups but more so with PTCA. Compared with the PTCA group, there was a 16.5% absolute excess of grade 2 or worse angina in the medical group 3 months after randomization (P 70% stenosis).12 Patients randomized to OMT were permitted to cross over to PCI for refractory angina. After 4.6 years, there was no difference in death or nonfatal MI between the two groups (hazard ratio [HR] for the PCI group, 1.05; 95% CI, 0.87–1.27; P=0.62), with a 30% cross‐over rate from the OMT group to PCI.None of the trials comparing an initial strategy of PCI with stents and medical therapy to medical therapy alone in stable CAD11, 12, 13, 14, 15, 16 nor a 2012 meta‐analysis17 have shown that PCI improves survival (combined odds ratio [OR], 0.98; 95% CI, 0.83–1.15)17 or reduces MI (combined OR, 1.12; 95% CI, 0.93–1.34).17 We performed an updated meta‐analysis of all published trials randomizing patients to an initial strategy of PCI (with or without stents) in addition to medical therapy compared with medical therapy alone using the longest published follow‐up available.8, 9, 12, 13, 14, 15, 37, 38, 39, 40, 41, 42, 43 The analysis continues to show no improvement in survival (Figure 2A) and no reduction in nonfatal MI (Figure 2B). These results persisted even after limiting the analysis to the trials that only included stents (data not shown).Download PowerPointFigure 2. A, Difference in survival in randomized controlled trials comparing initial percutaneous coronary intervention (PCI) vs medical therapy (MT). A systematic search of published studies in any language in MEDLINE, Cochrane, and PubMed from 1970 to October 2017 was performed independently by both authors using the following search terms: stent, medical therapy, stable angina, coronary artery disease (CAD), and combinations of these terms. Patient outcomes (death from any cause and nonfatal myocardial infarction) were systematically reviewed and recorded independently by both authors. A meta‐analysis of summary statistics from individual trials was performed using Comprehensive Meta‐Analysis software, version 2 (Biostat Inc). Summary odds ratios (ORs) were calculated using a random‐effects model. Results from the longest reported follow‐up are shown.8, 9, 13, 14, 15, 37, 38, 39, 40, 41, 42, 43 All included studies are listed by name along with point estimates of the ORs and respective 95% confidence intervals (CIs). The red squares represent the overall findings in each plot. B, Difference in nonfatal myocardial infarction in randomized controlled trials comparing initial PCI vs MT. Results from the longest reported follow‐up are shown.8, 9, 12, 13, 14, 15, 37, 38, 39, 40, 41, 42 All included studies are listed by name along with point estimates of the ORs and respective 95% CIs. The red squares represent the overall findings in each plot.AnginaPCI has been shown to provide incremental, short‐term relief of angina compared with medical therapy that is far from universal as would be expected according to the prevailing paradigm. In MASS‐II,11 the BARI 2D (Bypass Angioplasty Revascularization Investigation 2 Diabetes) trial44 and COURAGE45 only 7% to 17% more patients in the PCI arm were free from angina at 12 months compared with patients randomized to medical therapy alone. In MASS‐II, the incremental benefit of PCI persisted for 120 months.41 By 24 months in BARI 2D44 and by 36 months in COURAGE,45 freedom from angina was not significantly different between PCI and medical therapy groups. The percentage of patients treated with PCI who continued to have angina at 1 year ranged from 45% in MASS‐II (class II or III angina) to 60% in BARI 2D to 68% in COURAGE.One month after relieving the flow‐limiting stenosis by stent placement in COURAGE, 79% of patients with baseline angina still had angina45 (Figure 3), suggesting that, in most COURAGE patients, the epicardial stenosis was not the cause of angina. There was only an 11% advantage of PCI over the medical therapy arm in angina relief at 1 month with a diminishing benefit over time. The recently reported ORBITA (Objective Randomised Blinded Investigation With Optimal Medical Therapy of Angioplasty) trial, a randomized double‐blind, sham‐controlled trial of PCI and medical therapy in patients with angina and single‐vessel CAD refractory to OMT found no benefit of PCI in change in exercise time, angina relief, angina frequency, angina stability, time to 1 mm of ST depression, quality of life or treatment satisfaction.46 These results suggest that prior studies that reported greater angina relief from PCI compared to medical therapy were confounded by the placebo effect associated with unblinded PCI and raise the question of whether PCI is associated with any incremental improvement in angina relief compared to OMT.46Download PowerPointFigure 3. The incremental benefit of percutaneous coronary intervention (PCI) and optimal medical therapy (OMT) compared with OMT alone in patients with baseline angina in the COURAGE (Clinical Outcomes Utilizing Revascularization and Aggressive Drug Evaluation) trial. The percentage of patients with angina over time from baseline to 36 months is displayed. One month after PCI, 79% of patients with baseline angina still had angina in the PCI arm, an 11% incremental benefit relative to the OMT arm. At 36 months, there was no significant difference between groups. High‐Risk SubgroupsAs was the case with the early PTCA trials, the results of randomized trials of PCI and medical therapy were downplayed because it was believed that patients at high risk had not been enrolled. However, patients at high risk including those with diabetes mellitus, proximal LAD lesions, reduced ejection fraction, multivessel CAD, increased age, chronic kidney disease, and significant myocardial ischemia have been included in recent trials.Diabetes mellitus.BARI 2D randomized 2368 patients with type 2 diabetes mellitus and stable CAD to prompt revascularization with intensive medical therapy or intensive medical therapy alone.15 After 5 years, there was no difference in mortality or in major adverse cardiovascular events between initial PCI and medical therapy groups. Subsequent analyses found no difference between initial PCI and medical therapy regardless of the number of diseased vessels, the amount of jeopardized myocardium, the number of stenotic lesions, the presence of a total occlusion, proximal LAD disease, prior revascularization, or an abnormal left ventricular ejection fraction.47 While the patients at higher risk with diabetes mellitus had more adverse cardiovascular events, those events were not prevented or reduced by PCI.Proximal LAD.In COURAGE, although survival free of death, MI, or acute coronary syndrome was reduced in all patients with a proximal LAD stenosis >90%, PCI did not improve these outcomes (P=0.79).48 These results align with the earlier findings of MASS,6 which included patients with >80% proximal LAD stenosis, and MASS‐II,11 which required >70% proximal LAD stenosis and documented ischemia. Neither study found that PCI resulted in any reduction in cardiovascular events.Three‐vessel CAD and low ejection fraction.In a subgroup analysis of COURAGE, an increasing number of diseased vessels and/or the presence of reduced ejection fraction were associated with significant reductions in event‐free survival.49 In the patients at highest risk with 3‐vessel CAD and reduced ejection fraction (n=124), there was no benefit from initial PCI and OMT compared with OMT alone (P=0.59).Other high‐risk groups.Additional analyses of COURAGE have found no benefit for an initial strategy of PCI in patients with a recent acute coronary syndrome or class III angina,50 older patients51 or patients with chronic kidney disease.52 As expected, each high‐risk subgroup experienced worse outcomes, but PCI did not improve these outcomes for any subgroup. Although these hypothesis‐generating post hoc analyses are limited by their small sample size, the sample sizes are larger than the anecdotal experience of most individual physicians, and unlike anecdotal experience, they are not subject to selection or recall bias. Across the numerous high‐risk subsets evaluated, there has never been any signal suggestive of an improvement in outcomes with PCI that would justify an appropriately powered randomized trial.CAD before vascular surgery.Given the hemodynamic stress of vascular and other major noncardiac surgeries, prophylactic coronary revascularization is often pursued in the hope of reducing postoperative ischemic events. No benefit of prophylactic PCI (or CABG) was shown in the CARP (Coronary Artery Revascularization Prophylaxis) trial in which patients with >70% stenosis of at least one coronary vessel (n=510) were randomized to revascularization before vascular surgery for an expanding abdominal aortic aneurysm (33%) or arterial occlusive disease of the legs (67%).53 In the revascularization group, 59% underwent PCI and 41% underwent CABG. There was no difference in postoperative mortality between the revascularization and medical therapy groups. No high‐risk subset that benefited from revascularization could be identified, including patients with a moderate or large degree of ischemia, a greater revised cardiac risk index, or 3‐vessel CAD and left ventricular dysfunction.The Ischemia‐PCI Reflex: Pathophysiology or Mythology?Myocardial ischemia on stress testing has long been associated with increased mortality.54 In a recent review it was asserted that "the presence and extent of ischemia is the most important factor related to outcome and that all functionally significant stenoses should be revascularized to relieve ischemia."55 However, closer examination of the evidence base used to support this opinion finds it is of insufficient quality to justify such a reflexive approach to treatment of patients with stable CAD and ischemia. Furthermore, closer review of the data suggests that ischemia is a marker for adverse outcomes rather than the cause of the adverse outcomes (analogous to ventricular ectopy after MI56).The belief that ischemia necessitates revascularization to improve outcomes can be traced, in part, to a retrospective analysis of the CASS (Coronary Artery Surgery Study) registry that reported the benefit of bypass surgery was greatest in patients who were unable to exercise beyond stage 1 due to angina and displayed 1 mm of ST depression.57Hachamovitch et al58 subsequently performed a single‐center, retrospective review of 10 627 patients who underwent an exercise or adenosine stress myocardial perfusion scan between 1991 and 1997. After a mean follow‐up of 1.9 years, Cox proportional hazards modeling indicated that the 671 (6.3%) patients who underwent revascularization (PCI in 346, CABG in 325) within 60 days experienced improved cardiac survival if their ischemic burden was ≈10% to 12.5%, but worse cardiac survival if their ischemic burden was <10%.As a retrospective analysis, this study cannot assess causality. Furthermore, it was subject to methodological flaws including short follow‐up and significant selection bias that severely limited the interpretation of its findings. Their model included degrees of myocardial ischemia up to 50% that are not seen clinically. In addition, there was no uniformly prescribed medical therapy that would be considered optimal by modern standards. Although this article is widely quoted to justify revascularization in the presence of ischemia, to our knowledge it has never been used to justify avoidance of revascularization in patients with an ischemic burden 5% reduction in ischemia on follow‐up imaging (regardless of treatment strategy) had a lower risk for death or MI in an unadjusted analysis (P=0.037).59 However, this finding did not persist in an adjusted analysis (P=0.26). The unadjusted outcome data from this small cohort of patients continue to be used to justify the use of PCI for patients with ischemia.The COURAGE investigators subsequently reviewed the 1381 randomized patients who underwent myocardial perfusion scans at baseline.60 Of these, 486 had moderate to severe ischemia, evenly divided between the randomized treatment groups. There was no difference in outcomes in patients with moderate to severe ischemia randomized to treatment with PCI and OMT compared with OMT alone. Similarly, the nuclear substudy of 1505 patients from BARI 2D found no relationship between ischemic myocardium and subsequent death or MI.60A 2014 meta‐analysis reported outcomes of patients with baseline ischemia or an abnormal fractional flow reserve (FFR) in the randomized trials comparing PCI and medical therapy with medical therapy alone.18 There was no difference in death, MI, unplanned revascularization or angina between groups (Figure 4). The strong association of ischemia with impaired event‐free survival coupled with the lack of benefit of PCI in patients with ischemia suggests that ischemia is a marker of adverse outcomes rather than the cause of the adverse outcomes. The extent of ischemia is correlated with atherosclerotic plaque burden61 that, as the substrate for acute coronary syndromes, may be the underlying determinant of prognosis and the explanation for why PCI offers no incremental benefit beyond optimal disease‐modifying therapy. In further support of this theory, in COURAGE, atherosclerotic burden was independently predictive of adverse outcomes, whereas ischemic burden was not.62Download PowerPointFigure 4. Comparison of percutaneous coronary intervention (PCI) and medical therapy (MT) vs MT alone in patients with documented myocardial ischemia. Each graph illustrates an outcome. A, Death; B, nonfatal myocardial infarction; C, unplanned revascularization; and D, angina during follow‐up. All included studies are listed by name along with point estimates of the odds ratios (ORs) and respective 95% confidence intervals (CIs). The sizes of the squares denoting the point estimate in each study are proportional to the weight of the study. The diamonds represent the overall findings in each plot. See text for full trial names. Reproduced from Stergiopoulos et al18 with permission. Copyright©2014 American Medical Association. All rights reserved.Invasive Physiologic Tools to Assess Hemodynamic Significance of Epicardial CADFFR was developed as an invasive tool to quantify the hemodynamic significance of an angiographic stenosis under the assumption that revascularization of hemodynamically significant lesions will improve outcomes––in other words, to replace the "oculostenotic" reflex with an "ischemia‐PCI" reflex. FFR estimates flow across a lesion by measuring the change in pressures between the aorta and the coronary artery distal to a lesion during pharmacologically induced maximal coronary flow.63 The impetus for its development was recognition of the limited accuracy of angiographic assessment of lesion severity. FFR was calibrated against stress testing to derive a cutoff value that correlates best with ischemia. Originally an FFR <0.75 was found to accurately predict at least one positive stress test in patients undergoing an exercise test, thallium scan, and dobutamine stress echocardiogram.64 However, contemporary stress tests were validated against coronary angiography whose limitations stimulated the development of FFR in the first place. Thus, the logic justifying FFR appears circular, at best.In the FAME (Fractional Flow Reserve versus Angiography for Multivessel Evaluation) trial, 1005 patients were randomized to angiographically guided PCI or PCI guided by an FFR <0.80 (it is not clear why 0.75 was not used). FFR‐guided PCI was associated with fewer PCIs and a reduction in the combined end point of death, MI, or revascularization at 2 years.65 Often overlooked, the FAME trial was the first study to demonstrate harm from unnecessary PCI in patients with stable CAD. It left unanswered where the line is drawn between necessary and unnecessary PCI.The FAME 2 trial provided insight into where that line should be drawn. This unblinded trial randomized 888 patients with FFR 10 times the upper limits of normal or >5 times upper limits of normal with an associated occluded artery or new Q waves, a threshold clearly associated with increased 30‐day mortality.63, 64FAME 2 was initially designed to randomize 1600 patients with a planned follow‐up of 5 years. The study was terminated prematurely at the recommendation of the data safety monitoring board after 213 days of follow‐up with only 888 patients randomized because the composite end point favored the FFR‐guided PCI group, but the end point difference was driven solely by urgent revascularization. Notably, 52% of the "urgent" revascularizations in the initial analysis14 and 49% of urgent revascularizations at 2‐year follow‐up40 were not associated with either ischemic ST‐T–wave changes or positive cardiac biomarkers. The increase in urgent revascularization in the medical therapy group appears to have been largely driven by the subset of patients with FFR <0.65 (P=0.01 for interaction)14 indicative of such a severe stenosis that passage of the flow wire across the lesion without stent placement may have resulted in plaque disruption.62 Patients in the medical therapy arm were not treated with dual anti‐platelet therapy which may have protected against events caused by iatrogenic plaque injury. In addition, the unblinded nature of the trial may have prompted more aggressive evaluation and treatment of patients with very low FFR values who were randomized to medical therapy.There were 42 more urgent revascularizations in the medical therapy arm than the FFR‐PCI arm (7 versus 49). By suggesting that the results of FAME 2 support PCI of every lesion with an FFR <0.8 is to suggest that to prevent urgent revascularizations (but not MI and death) in the additional 9.5% of patients who experienced them in the medical arm, 100% of patients with an abnormal FFR should undergo PCI. This is not a wise use of resources and recalls the lesson of FAME: unnecessary stenting leads to worse outcomes. Finally, it must be acknowledged that FFR is not a benign procedure. Coronary artery dissection, abrupt vessel closure, and death have been caused by measuring FFR. A French trial of FFR versus angiographic guidance for PCI in multivessel disease was terminated prematurely in 2016 because of excess deaths (17 versus 7) in the FFR arm at 12 months.66Concern for HarmEven with the lack of documented benefit for an initial PCI strategy in patients with stable CAD, some argue that it is reasonable to still pursue initial PCI since randomized trials have shown that it can improve or relieve angina in some patients with no increase in MI or death. It is important to note that PCI does carry with it a low, but real, risk of complications including death (0.65%), MI (15%), renal injury (13%), stroke (0.2%), contrast allergy (≤1%), and vascular complications (2–6%).4, 67, 68 While there was no difference in overall death or MI in the randomized trials of PCI versus medical therapy, the studies were not powered to detect an increase in uncommon but unavoidable procedural complications. These seemingly low risks are put in perspective when they are multiplied by the estimated 300 000 PCI procedures performed annually in the United States for stable CAD. Given the added cost69 and procedure‐related risks associated with PCI, as well as the limited improvement in angina compared with OMT, it is difficult to argue that equipoise exists in the selection of an initial treatment strategy
Referência(s)