Cardiac Enzyme Elevation After Successful Percutaneous Coronary Intervention Is Not an Independent Predictor of Adverse Outcomes
2005; Lippincott Williams & Wilkins; Volume: 112; Issue: 6 Linguagem: Inglês
10.1161/circulationaha.104.478347
ISSN1524-4539
AutoresDonald E. Cutlip, Richard E. Kuntz,
Tópico(s)Cardiac Imaging and Diagnostics
ResumoHomeCirculationVol. 112, No. 6Cardiac Enzyme Elevation After Successful Percutaneous Coronary Intervention Is Not an Independent Predictor of Adverse Outcomes Free AccessArticle CommentaryPDF/EPUBAboutView PDFView EPUBSections ToolsAdd to favoritesDownload citationsTrack citationsPermissions ShareShare onFacebookTwitterLinked InMendeleyReddit Jump toFree AccessArticle CommentaryPDF/EPUBCardiac Enzyme Elevation After Successful Percutaneous Coronary Intervention Is Not an Independent Predictor of Adverse Outcomes Donald E. Cutlip, MD and Richard E. Kuntz, MD, MSc Donald E. CutlipDonald E. Cutlip and Richard E. KuntzRichard E. Kuntz Originally published9 Aug 2005https://doi.org/10.1161/CIRCULATIONAHA.104.478347Circulation. 2005;112:916–923The debate over the importance of periprocedural creatinine kinase-myocardial band (CK-MB) elevation after percutaneous coronary intervention (PCI) has captured the attention of cardiologists since the early days of balloon angioplasty, when the phenomenon was first recognized and reported to have no clear association with adverse clinical sequelae.1,2 These early reports were limited by the small numbers of patients and follow-up that extended only to hospital discharge. Since then there have been numerous published studies documenting the frequency and clinical outcomes of periprocedural CK-MB elevation, most of which have implicated it in a clear relationship to later adverse outcome including increased mortality.3–9 Multiple additional reports of this adverse association have been presented only in abstract form. Moreover, previous controversies about whether these enzyme elevations actually represent myocardial infarction (MI) have been dismissed by clear evidence demonstrating myocardial necrosis in the zones of the target vessel after PCI accompanied by CK-MB elevation.10 The consistency of these findings has led to widespread acceptance of a deleterious role for periprocedural MI,11 and moderate elevations of CK-MB (>3 times the upper limit of normal [ULN]) have been regarded by many as appropriate surrogate end points for studies of coronary interventional devices and antithrombotic drug therapy. Still, some investigators have not found an association or have reported adverse outcomes only after large MI (CK-MB >5 to 10 times the ULN) or with a concurrent procedural complication for which the survival consequences are not debated.12–15The fact that the debate continues at national meetings and has been included here in this series of controversies in cardiovascular medicine points to a need for a critical reappraisal of this literature to determine a thesis that the available data can support, while providing a clear explanation for the discrepant findings of the opposing view. As in all analyses, the task must begin with a clear statement of the question, which we pose as "Does any elevation of CK-MB after an otherwise successful procedure have an independent association with subsequent mortality?" We frame our antagonist position within the following 3 arguments: (1) Many studies purporting a relationship between low-level CK-MB elevation and mortality have faulty designs that limit their conclusions; (2) the data do not support periprocedural CK-MB elevation as a surrogate end point for mortality; and (3) the definition of "otherwise successful" has been inconsistent and in many cases may not pertain to the current stent era.Finally, we address specific situations in which the effect of periprocedural MI may be modified by higher baseline risk, such as saphenous vein graft intervention or patients with significant left ventricular systolic dysfunction.Limitations of Previous StudiesThe evaluation of late clinical outcomes after periprocedural CK-MB elevation is restricted to a study of potentially unmatched cohorts because it is by definition a post hoc event. As such, its occurrence selects groups that are likely to be heterogeneous for multiple baseline and procedural characteristics that influence not only the choice of the interventional procedure itself but also many pre- and post-treatment management decisions. Any of these measured or unmeasured confounding factors could contribute significantly to late outcome. Indeed, the postprocedure measurement of CK-MB itself may be based on other clinical concerns and thus adds selection bias as a limitation. Finally, many PCI populations on which previous reports have been based included patients presenting with acute coronary syndromes, and it is unclear to what degree so-called periprocedural CK-MB elevations actually represented events occurring preprocedure and thus were secondary to spontaneous MI.3–6 Efforts to reduce this limitation by merely excluding patients with periprocedural CK-MB elevation are incomplete given the inherent delay between time of MI and elevation of CK-MB.Some of these limitations can be overcome by appropriate study design. For example, a prospective analysis within a clinical trial that has strict inclusion and exclusion criteria and a requirement for systematic collection of cardiac enzyme data in all patients at specific time intervals should minimize the potential for major baseline differences between the MI and no MI groups and avoid bias caused by incomplete ascertainment. If acute coronary syndrome patients are included, then the protocol must either exclude these patients unless a normal CK-MB value is obtained >8 hours after presentation and before PCI or require careful adjudication by an independent events committee to determine whether periprocedural CK-MB elevation is likely caused by a periprocedural spontaneous event.Instead, most of the studies reporting an increased mortality risk for periprocedural CK-MB elevation after otherwise successful procedures have been retrospective reviews of single-center databases.3,4,6,7 It is not possible for any retrospective analysis to approximate equality between the study groups or to adequately control for all unmatched factors influencing outcome. Moreover, even slight deficiencies in ascertaining of postprocedure CK-MB data that are invariably a part of retrospective analyses may significantly bias the results. In fact, such potential difficulties are clearly identified by the authors of these studies. In 1 study, patients who had CK >2 times the ULN were also more likely to have had recent MI, unstable angina, thrombotic or complex lesions, vein graft intervention, and directional coronary atherectomy (DCA) procedures.3 It is not surprising that these risk factors may identify a group at higher risk for subsequent death. The same authors reported similar baseline differences as well as an increased risk for subsequent MI and revascularization in patients identified only by CK-MB elevation without elevated CK,4 further suggesting that CK-MB may be an epiphenomenon in otherwise identified high-risk patients.16Differences between the MI and no-MI groups of other studies have been even less subtle. In a report from the Coronary Angioplasty versus Excisional Atherectomy Trial (CAVEAT), 3 of 13 cardiac deaths were associated with abrupt closure during or shortly after the procedure and large Q wave MI. These events represented 3 of the 6 deaths assigned to the group with elevated CK-MB, and contributed to the conclusion that elevated CK-MB was a significant predictor of 1-year mortality.5,17 In another study with a matched cohort design, the matching of groups with and without CK elevation based only on date of procedure and interventional device failed to account for significantly worse procedural success measures in the CK elevation group, including final diameter stenosis >50% (15% versus 7%, P=0.003) and final Thrombolysis in Myocardial Infarction (TIMI) grade flow 800 000 procedures performed annually in the United States. If it is true that a procedure performed only for relief of symptoms actually has an increased mortality risk in at least 10% to 20% of the patients undergoing the procedure, then this topic should receive even greater attention and the continued performance of the procedure should be strongly questioned in all but a limited number of patient groups. Support for having not adopted this stance can be developed by an examination of the value of CK-MB elevation as a surrogate end point for cardiovascular mortality.The significance of cardiovascular mortality for the public health has led to the performance of numerous interventional clinical trials aimed at preventing or delaying these events. Such trials require large numbers of patients and long periods of follow-up at considerable cost with the attendant potential pitfalls of lack of validity by the time results are available or serious delays in availability of efficacious therapies. Furthermore, as clinicians, we value reliable predictors of mortality to help guide treatment decisions for our patients. The availability of markers that are tightly linked to disease pathogenesis and outcome would be valuable as potential replacements for the hard clinical morbidity and mortality end points. Care must be exercised, however, to distinguish risk markers that are only statistically related to disease outcome from true surrogate end points. To meet the requirement for surrogate status, a marker must track with the frequency of disease both as an epidemiological risk factor and a therapeutic responder.18 That is, there must be a plausible cause-and-effect mechanism as well as evidence to support an equal directional effect on disease outcome for interventions that either increase or decrease the frequency of the surrogate marker. An intervention that decreases (or increases) the frequency of a surrogate marker for mortality should have a corresponding decrease (or increase) in mortality.Studies reporting an association of CK-MB elevation and late mortality have argued, at least implicitly, for surrogate status for periprocedural CK-MB and have led to 2 significant modifications in the practice of interventional cardiology: demise of DCA as an alternative to balloon angioplasty and approval and widespread use of glycoprotein (GP) IIb/IIIa inhibitors during PCI. Although both of these judgments by the interventional community may have been correct choices from the perspective of the modern stent era, a surrogate role for periprocedural CK-MB elevation should not be credited with providing the convincing evidence. Indeed, it is worth considering the data from the major DCA studies and mortality data from GP IIb/IIIa clinical trials as antithetical arguments for CK-MB elevation as a mortality surrogate.The CAVEAT study was the first large randomized trial of DCA versus balloon angioplasty and raised concerns regarding the continued use of DCA. The concerns were caused by the lack of efficacy in preventing restenosis and, more important, serious safety issues, including increased 1-year mortality in the DCA group.17 This 1-year mortality difference (2.2% versus 0.6%, P=0.035) was attributed mostly to the increased frequency of periprocedural MI, defined as CK-MB elevation >3 times the ULN, after DCA (15.2% versus 6.8%, P=0.001).5,17 If we apply the test for surrogacy based on these results, then we may conclude that because the potential surrogate (CK-MB elevation) is associated with increased risk of the outcome (mortality) and is increased in response to the intervention (DCA), the test criteria are fulfilled. A review of additional data proves this conclusion to be erroneous.Of the 13 cardiac deaths in CAVEAT, 5 (36%) occurred in the group with CK-MB elevation. Of these, however, 3 were associated with large Q wave MI and major procedural complications, 2 of which were actually in the balloon angioplasty group. Whereas DCA was associated with increased mortality, only 2 of the 11 DCA deaths could be related to CK-MB elevation after otherwise successful DCA procedures.Additional evidence against CK-MB as a surrogate marker after DCA comes from the Balloon Angioplasty versus Optimal Atherectomy Trial (BOAT). The BOAT investigators had the advantage of increased experience and knowledge of the potential complications of DCA from CAVEAT and other studies, resulting in many technical refinements to the procedure to achieve higher device success and fewer acute angiographic complications. Similar to CAVEAT, the frequency of CK-MB elevation was still significantly higher for DCA as compared with balloon angioplasty (16% versus 6%, P 1 year after the procedure suggests the identification of a group with a confounding factor that is associated with worse late prognosis, such as more severe underlying atherosclerotic disease. Although earlier reports did not suggest that the prognostic significance of periprocedural CK-MB elevation was explained by greater underlying disease,11 more recent studies have found evidence for an association. Kini et al observed that diffuse coronary disease, determined by lesion length >20 mm or multiple single lesions in ≥1 vessels, and systemic atherosclerosis, defined as a history of peripheral vascular, cerebrovascular, or aortic disease, were both independent predictors of periprocedural CK-MB elevation among 1675 consecutive patients.31 In a study of preintervention intravascular ultrasound analysis of 2780 lesions, Mehran et al reported increases in plaque volumes at both the reference segments and lesion sites for patients with progressively higher periprocedural CK-MB.32 Plaque volumes at the reference segments and lesion sites were both independently associated with periprocedural CK-MB elevation.32 To the extent that these studies identify more severe atherosclerosis associated with the probability of periprocedural CK-MB elevation, it is not surprising that a higher mortality risk may also exist.The available data thus do not support a role for periprocedural CK-MB elevation as a surrogate marker for morbidity and mortality after PCI. There is a reported statistical association between elevated periprocedural CK-MB and late mortality, but there is no convincing evidence of a cause-and-effect relationship and the supposition is supported mostly by retrospective studies from the prestent era. It is likely that the relationship is explained in part by the identification of a group with baseline characteristics that increase both the risk for periprocedural CK-MB elevation and late mortality.Previous reports have shown that the frequency and severity of periprocedural CK-MB elevation are strongly related to the choice of percutaneous device and that outcomes after serious procedure-related complications or Q wave MI are consistently poor. As we study this question in the title in 2005, we must consider these issues in relation to modern stent procedures, in which the risk for periprocedural CK-MB elevation is high, but recognized serious procedure-related complications have been dramatically reduced.Periprocedural CK-MB After an Otherwise "Successful" Stent ProcedureDuring the last 10 years, coronary stenting has dramatically changed the landscape of PCI, providing increased safety and durability as compared with balloon angioplasty. The major safety benefits are the result of significant decreases in acute major complications of abrupt vessel closure and emergency CABG, with combined rates of 5% to 10% in the balloon angioplasty-and-atherectomy era to 20% of patients, including elevations >3 times the ULN in at least 8% of patients.14,15,35,36 Based on extrapolation of data from many of the studies of balloon angioplasty and atherectomy discussed above, periprocedural non–Q wave MI, defined as CK-MB elevation >3 times the ULN or total CK >2 times the ULN, has been included as a component of the primary safety end point for recent coronary stent clinical trials evaluated by the US Food and Drug Administration. Whether the findings from these studies of devices with differing pathogenesis of CK-MB elevations—even accepting their other limitations, can be applied to coronary stent patients has not been clearly demonstrated.Several studies have provided some insight into this question.13–15 Saucedo et al studied 900 consecutive patients undergoing successful stenting at a single institution between 1994 to 1995 and noted large periprocedural CK-MB elevation (>5 times the ULN) in only 67 (0.7%) patients. These patients represented a uniquely high-risk group with significantly more complex lesion characteristics and increased in-hospital ischemic complications including subacute stent thrombosis and repeat revascularization as compared with patients without CK-MB elevation. In this small study, 1-year mortality was significantly higher for patients with CK-MB >5 times the ULN as compared with patients without CK-MB elevation (6.9% versus 1.7%, P=0.01) but was not different for patients with CK-MB elevation 3600 stent patients. Periprocedural CK-MB elevation >3 times the ULN occurred in 16.9% of patients and was >8 times the ULN or associated with Q wave MI in 7.8%. By 2 years, mortality was independently and significantly higher for patients with periprocedural Q wave MI (38.6%, HR 9.9, P 8 times the ULN without Q waves (14.5%, HR 2.2, P<0.0001). Lesser degrees of periprocedural CK-MB elevation had no effect on 2-year mortality compared with no CK-MB elevation.The Stone et al study has the strength of a consecutive series of patients with complete ascertainment of periprocedural CK-MB data and representing variable periprocedural risk and procedure outcomes. Although the authors excluded patients with recent MI (<72 hours) and those with subacute vessel closure or who required emergency bypass surgery within 24 hours, an accompanying editorial in Circulation speculated whether this study adequately addressed the question we have stated in this debate, namely whether there is an effect of periprocedural CK-MB elevation after an otherwise successful stent procedure.37 This raises the issue of how a successful stent procedure should be defined in 2005. In the previous studies described in this review, "otherwise successful" was broadly defined according to National Heart, Lung, and Blood Institute registry standards as <50% diameter stenosis without in-hospital death, Q wave MI, or emergency CABG.38 Even though some of the studies included procedures that were unsuccessful even by this definition, it can be argued that a more conservative definition should be adopted for stent procedures in which unstented dissections and final slow flow also connote failure. Furthermore, although the risk of emergent CABG has been reduced dramatically in concert with improved acute lesion outcomes and the near elimination of abrupt closure, stenting is associated with a unique risk of acute and subacute thrombosis. We have reported that in the modern stent era most of these thrombotic complications occur within the first 24 to 48 hours after the procedure and are associated with a 6-month mortality rate of 20%.39 Because this is the time interval during which postprocedure cardiac enzyme measurements are also determined, it is possible that enzyme levels caused by stent thrombosis may have been included among otherwise successful procedures. Obviously, it would be a mistake to equate the increased mortality related to stent thrombosis with outcomes of periprocedural CK-MB elevation after otherwise successful procedures.We recently evaluated the differential impact on the 1-year mortality effect of periprocedural CK-MB elevation according to success or failure of the stent procedure in a pooled series of 5850 coronary stent patients from 6 coronary stent clinical trials.15 For this analysis, an unsuccessful procedure was defined as final diameter stenosis <50%, final TIMI grade flow 0.20). CK-MB was elevated >8 times the ULN or new Q wave MI was present after only 2% of successful procedures. In contrast, CK-MB was elevated after 70% of unsuccessful procedures including >8 times the ULN or new Q waves in 32%, and was associated with a significantly higher risk of death in the first year as compared with no CK-MB elevation (13.1% versus 0%, P=0.03). This included 19.7% mortality among patients with unsuccessful procedures and CK-MB >8 times the ULN or new Q waves.These studies demonstrate that periprocedural CK-MB elevation is common after coronary stenting but is limited to low-level elevations in most cases with infrequent occurrence of large non–Q wave and Q wave MIs. When these larger MIs do occur, they are almost always the result of procedural or early postprocedure serious complications, which would not indicate a successful procedure by most perceived standards. Regardless, these large MIs are consistently and uniquely associated with an increased mortality risk at 1 and 2 years, with no apparent significant increase in either short- or long-term mortality for low- and moderate-level ( 5 times the ULN) elevation. There was significantly increased 1-year mortality for minor (6.5%) or major (11.7%) levels of periprocedural CK-MB elevation as compared with no elevation (4.8%).40 Distal embolization is likely to be more severe during saphenous vein graft PCI and more likely to result in major reperfusion abnormalities, which may account for this observed increased risk. For this reason, we recommend routine use of embolic protection devices in the treatment of these lesions with a goal of avoiding these complications and associated periprocedural CK-MB elevation.Ellis et al30 noted an increase in the magnitude of risk for periprocedural CK-MB elevation depending on the presence of LV systolic dysfunction. For patients with LV dysfunction, 4-month mortality rates increased from 1.9% to 14.0% if periprocedural CK-MB elevation >5 times the ULN occurred. A smaller but important increase to 3.0% was approximated for patients with CK-MB elevation <5 times the ULN.30 It is reasonable to suspect that patients with abnormal baseline LV function will have poorer tolerance of any additional insult and additional concern about periprocedural CK-MB elevation in this group is warranted.ConclusionsDespite widespread acceptance of a direct association between any level of periprocedural CK-MB elevation and subsequent mortality, a critical review of the available studies does not support this position. Instead, most studies are seriously flawed on the basis of unavoidable retrospective design and failure to adequately account for significant other differences between patients with and without CK-MB elevation. Furthermore, there are convincing data that periprocedural CK-MB elevation is not an acceptable surrogate marker for later mortality, and that any true statistical association is most likely caused by the identification by periprocedural CK-MB elevation of a group with increased mortality risk before the PCI procedure was performed. In the current stent era, available studies suggests an increased risk only for large non–Q wave and Q wave MI, the occurrence of which is almost always caused by major procedure-related complications or early postprocedure clinical events rather than by unexpected events after otherwise su
Referência(s)