Author + information
- Received June 8, 1998
- Revision received September 18, 1998
- Accepted October 30, 1998
- Published online March 1, 1999.
- ↵*Reprint requests and correspondence: Dr. Eric J. Topol, Department of Cardiology, F25, The Cleveland Clinic Foundation, 9500 Euclid Avenue, Cleveland, Ohio 44195
We sought to determine whether periprocedural myocardial infarction complicating percutaneous coronary revascularization is associated with subsequent clinical restenosis, as judged by the need for target vessel revascularization.
Although myocardial enzyme elevation following angioplasty is associated with increased late mortality, its effect on subsequent clinical restenosis, as assessed by the need for late target vessel revascularization (TVR), is unknown.
Serial myocardial enzyme determinations were performed on 2,099 patients who underwent angioplasty or atherectomy in the Evaluation of IIb/IIIa platelet receptor antagonist 7E3 in Preventing Ischemic Complications (EPIC) trial. Thirty-day survivors were prospectively followed for three years for adverse clinical events including death and need for TVR.
Within the study population, periprocedural creatine kinase (CK) elevation was a predictor of late mortality. Among patients with elevated CK, however, a paradoxical decrease in the need for late TVR was present. This relationship became progressively more profound as the magnitude of CK release increased. Late TVR occurred in 29.8% of patients with no CK elevation, 24.8% with CK elevation to >3 times normal, and 16.9% with >10 times elevation (hazard ratio 0.51, 95% CI 0.29, 0.91).
In the EPIC study, patients with periprocedural MI were less likely to develop clinical restenosis as measured by the need for TVR. Mechanistically, although it is unlikely that CK elevation prevents vascular renarrowing per se, myocardial necrosis impairs the clinical manifestation of restenosis, thereby reducing the need for ischemia-driven TVR. This novel finding 1) highlights the potential discordance between angiographic and clinical measures of restenosis, and 2) has implications for clinical trials, as therapies that reduce periprocedural MI may be associated with a perceived excess of restenosis when measured by the need for TVR.
Periprocedural myocardial infarction (MI) that occurs with percutaneous coronary revascularization has been identified as a strong predictor of late mortality (1–9). Using a threshold value of a threefold increase in creatine kinase (CK) or creatine kinase, MB fraction (CK-MB), periprocedural MI occurs in approximately 5% to 9% of patients undergoing balloon angioplasty (4,6). Approximately 75% of these events are associated with ischemic chest pain, electrocardiographic changes, or both. Supporting the association between procedural MI and late clinical events, adjunctive therapies that reduce the incidence of periangioplasty MI, most notably antagonists of the platelet glycoprotein IIb/IIIa receptor, have been associated with reductions in late mortality in patients with acute coronary syndromes (7).
Although the relationship between procedural MI and mortality following coronary intervention has been investigated in detail, the impact of early elevations in myocardial enzyme levels on subsequent clinical restenosis, as judged by the need for repeat target vessel revascularization (TVR), has not been examined. Whereas periprocedural CK elevation per se would not be expected to influence directly the process of arterial renarrowing, we hypothesized that MI in the territory of an angioplastied vessel might impair the clinical recognition of restenosis. Apart from the clinical ramifications of a potential relationship between periprocedural MI and restenosis, an interaction between postintervention MI and the need for repeat TVR could have profound implications for the interpretation of clinical trials of therapies that are able to reduce the incidence of MI following percutaneous revascularization. Even though such therapies would be expected to exert a favorable influence on late mortality, they may be associated with a paradoxical increase in the need for target vessel revascularization, thus diluting their perceived efficacy as judged by a composite clinical end point. The present study was therefore undertaken to determine whether such a relationship exists between periprocedural MI and restenosis in a large group of patients with prospective long-term follow-up.
In the EPIC study (Evaluation of IIb/IIIa platelet receptor antagonist 7E3 in Preventing Ischemic Complications trial) 2,099 patients undergoing coronary angioplasty or directional coronary atherectomy with clinical or angiographic features that placed them at high risk for acute vessel closure were randomized to receive abciximab or placebo at the time of the intervention. Full details and major results of this study have been published previously (10,11). High-risk features included at least one of the following: 1) acute myocardial infarction (MI) within 12 h of symptom onset; 2) early postinfarction angina with concomitant electrocardiographic (ECG) changes; or 3) clinical or angiographic characteristics indicating high risk, according to the modified American Heart Association/American College of Cardiology criteria (12). Patients were ineligible for enrollment if they were greater than 80 years old, had a known bleeding diathesis, had undergone major surgery within the preceding six weeks, or had experienced a stroke within the past two years.
All patients were treated with aspirin 325 mg at least two h before the procedure. Heparin was administered during the procedure in incremental boluses to achieve a target-activated clotting time between 300 and 350 s, and heparin was continued for at least 12 h following the procedure to maintain the activated partial thromboplastin time at 1.5 to 2.5 times the control value.
Prior to the procedure patients were randomized to one of three treatment arms in a double-blind manner: 1) abciximab bolus of 0.25 mg/kg body weight, followed by an abciximab infusion at 10 μg/min; 2) abciximab bolus of 0.25 mg/kg followed by a placebo infusion; or 3) placebo bolus followed by a placebo infusion. All infusions were continued for 12 h unless a clinical contraindication developed. Vascular sheaths were maintained for at least 6 h after the end of the infusion.
Detection of periprocedural myocardial infarction
Myocardial enzyme determinations were performed immediately prior to the intervention, and then routinely at 6-h intervals over the first 24 h following the intervention. The current analysis focused on total serum CK level; however, to confirm a cardiac source for elevated CK, the diagnosis of myocardial infarction also required a positive qualitative CK-MB assay. For patients with elevated baseline (preintervention) CK, an appropriately timed increase in serum CK of at least 33% following a 25% decrease was required to establish the occurrence of a procedural MI. The ratio of maximal postintervention CK elevation to the upper limit of normal was termed the CK ratio.
The aim of the present study was to determine the influence of periprocedural MI on late events, notably target vessel revascularization (TVR) and overall mortality. Therefore, as a means to eliminate the potential confounding influences of the acute clinical sequelae of periprocedural MI, only patients who survived for ≥30 days following the index procedure (n = 2,049) were analyzed.
Patients were contacted at 30 days, 6 months, and 3 years following enrollment to determine their clinical status and the occurrence of study end points. Long-term follow-up data were collected by study coordinators who remained blinded to the treatment assignments of the individual patients. The National Death Index (Public Health Service, Centers for Disease Control and Prevention, National Center for Health Statistics) was searched to verify mortality in patients who died and to determine mortality in patients who were otherwise lost to follow-up. Target vessel revascularization (TVR) during follow-up was said to have occurred if either a repeat percutaneous intervention or coronary artery bypass surgery involving the index vessel was undertaken. Source documentation regarding the need for TVR and other adverse clinical events were reviewed in a blinded fashion by an independent end-point committee. All patients were reviewed and end points adjudicated by this committee.
Median follow-up for mortality was 3.1 years, with a 99.5% completeness of follow-up at 1 year, 99.1% at 2 years, 97.0% at 2.5 years, and 62.7% follow-up at 3 years. Median follow-up for other clinical events, including the need for TVR, was 3.1 years, with 96.1% follow-up at 1 year, 95.4% at 2 years, 95.3% at 2.5 years, and 76.2% at 3 years. Follow-up data were collected in the latter half of 1995; thus, the drop-off after 2.5 years is related to eligibility and not loss of follow-up.
Descriptive statistics are expressed in terms of percentages for categorical factors. All results are reported with the use of two-sided p values. The Cox proportional-hazards regression model was used to examine possible interactions between the baseline demographic characteristics presented in Table 1and the occurrence of periprocedural CK elevation to >3 times the upper limit of normal. The effects of CK elevation on time to TVR were examined graphically by Kaplan-Meier curves. Hazard ratios with 95% confidence intervals (CI) for both target vessel revascularization and overall mortality were determined for patients falling above versus below each of the predefined enzyme cut points (CK ratio >1, >2, >3, >5, >10).
Patient and procedural characteristics
Baseline characteristics of patients enrolled in the EPIC study who were available for long-term follow-up are demonstrated in Table 1. Overall, 72% of patients were male, 57% had experienced a previous MI, and 88% underwent treatment with balloon angioplasty alone. The incidence of periprocedural MI among 30-day survivors (n = 2,049), using various CK cut points, is displayed graphically in Figure 1. As demonstrated, 13.1% of patients experienced a postintervention CK elevation >3 times the upper limit of normal.
Stepwise multivariable analysis was undertaken to determine which, if any, of the baseline characteristics listed in Table 1served as an independent positive or negative predictors of postintervention CK rise to >3 times normal. Among these variables, recent MI (within eight days of the angioplasty procedure) was independently associated with a decreased likelihood of postprocedural CK elevation (hazard ratio = 0.40, 95% CI: 0.27, 0.59). No other variable acted as an independent positive predictor of postangioplasty MI.
Periprocedural MI and late events
The relationship between periprocedural MI and 3-year mortality is outlined in Table 2and Figure 2. With respect to mortality, there was a progressive “dose-dependent” relationship between periprocedural CK elevation and late death among 30-day survivors. Three-year mortality rose from 6.2% among 30-day survivors without procedural CK elevation, to 10.4% among patients with CK elevation >10 times the upper limit of normal.
Among patients with elevated CK, however, a decrease in the need for late TVR was present (Fig. 3). This relationship was likewise CK “dose-dependent,” becoming more profound as the magnitude of CK elevation increased. The need for late TVR, for example, fell progressively from 29.8% among patients without CK elevation, to 27.9% among patients with CK elevation >1 times the upper limit of normal, 25.4% for >2-fold elevations, 24.8% for >3-fold elevations, 20.8% for >5-fold elevations, and 16.9% for >10-fold elevations (hazard ratio = 0.51, 95% CI 0.29 to 0.91). Figure 4demonstrates the time-to-event curves relating CK elevation and late TVR. As is evident, most revascularization events occurred in the early (3 to 6 months) postprocedural period, corresponding with the typical time frame for restenosis.
The current study is the first to demonstrate a relationship between periprocedural MI, as determined by elevated CK detected during routine monitoring following coronary balloon angioplasty or directional atherectomy, and clinical restenosis, defined as the need for TVR. The need for TVR over the 3 years following the index procedure fell from 29.8% for patients without periprocedural CK elevation to 16.9 for patients with large (>10 times control) CK releases, corresponding to a 43% relative reduction.
Furthermore, the inverse relationship between CK and TVR was “dose-dependent” in nature, as greater magnitudes of CK elevation were associated with a progressively lower incidence of TVR. This finding holds important implications for the interpretation of clinical trials involving interventions that effectively decrease the incidence of periprocedural MI during percutaneous coronary interventions, as such therapies may be associated with a perceived excess of restenosis when measured by the need for target vessel revascularization.
Clinical versus angiographic restenosis
Restenosis can be defined in one of two ways: 1) angiographic restenosis, which is based on quantitative measurements of absolute and relative luminal dimensions at the site of angioplasty 6 months following the procedure, and 2) clinical restenosis, which is based on the occurrence of adverse clinical events, usually represented by recurrent angina or ischemia necessitating repeat revascularization of the original target vessel. Each definition is associated with potential advantages and disadvantages. Angiographic definitions of restenosis provide highly objective quantitative information regarding the degree of luminal renarrowing following angioplasty, and, because they allow the use of powerful parametric statistical models, they permit adequate statistical power to be achieved in the clinical trial setting with smaller numbers of patients. Determination of angiographic restenosis, however, mandates 6-month follow-up angiography, which is costly, associated with a small but finite risk of complications, and makes patient enrollment into clinical trials more difficult. Furthermore, the physiologic significance of “restenotic” lesions in the 50% to 70% diameter stenosis range is variable.
Whereas angiographic definitions focus strictly on vessel morphology, clinical definitions of restenosis are perhaps more pragmatic, as they focus on procedural success from the standpoint of whether or not the patient derived benefit from the intervention. Clinical definitions are concerned only with the ability of interventional approaches to influence favorably the composite clinical end point of death, MI, or TVR (10,11,13–15). Because the incidence of TVR typically far exceeds that of late MI or death, composite end points are typically heavily weighted in favor of this particular event (16).
Incidence and predictors of CK elevation
Depending on the population studied and the cutoff value selected to denote enzyme elevation, the incidence of increased CK or CK-MB following percutaneous revascularization has ranged from 6% to 20% in a variety of studies (1–6,17). Troponin T, thought to represent a more sensitive marker for myonecrosis, has been detected in up to 44% of patients following successful angioplasty (18–20). Various clinical and angiographic variables have been identified as risk factors for periprocedural MI, including recent myocardial infarction or unstable angina, saphenous vein graft intervention, and the presence of thrombus at the treatment site. Procedural events including dissection, distal embolization, side branch compromise, in-lab abrupt closure, and hypotension requiring the administration of vasopressors have also been associated with myocardial enzyme elevation (2,3). In addition, the performance of directional coronary atherectomy instead of balloon angioplasty has been identified as a strong predictor of postprocedural MI in a number of prospective and retrospective analyses (3,11,21).
Periprocedural MI and late mortality
Although early analyses involving small numbers of patients yielded inconclusive results (17,22), abundant evidence in the form of large retrospective and prospective studies now exists to support a relationship between both symptomatic and silent MI at the time of coronary revascularization and late mortality. Among a cohort of 4,664 consecutive patients who underwent either balloon angioplasty or directional coronary atherectomy at the Cleveland Clinic and were followed clinically for 36 ± 22 months, Abdelmeguid et al. (2)found that cardiac enzyme elevation following the procedure was a strong predictor for cardiac death (relative risk = 2.2, p < 0.0001). Whereas large elevations in CK (>5 times the upper limit of normal) were associated with increased late cardiac mortality, small to moderate CK elevations (2- to 5-fold increases) were likewise a harbinger of late death. In a subsequent analysis by the same investigators (3)of 4,484 patients who underwent successful angioplasty or atherectomy, even very small elevations in CK (1 to 2 times control) were associated with an increased incidence of cardiac death, future MI, and need for cardiac rehospitalization over the ensuing 3 years.
Kong et al. (4)performed a case-control analysis that included 253 patients with, and 120 patients without, total CK and CK-MB elevation following percutaneous coronary revascularization using a variety of devices. Periprocedural MI, regardless of the magnitude of CK rise, was a strong independent predictor of late (3.5 years) mortality. Within this population, a “gradient of effect” was present, whereby the relative risk for cardiac mortality increased incrementally by a relative risk of 1.05 for each additional 100 U/liter rise in serum CK. A similar graded increase in ischemic events proportionate to level of periprocedural CK-MB rise was observed among 2,432 patients enrolled in the Integrelin to Manage Platelet Aggregation to Combat Thrombosis (IMPACT II) trial (6). In a series of 1,897 patients treated via ablative new-device angioplasty (rotational atherectomy, directional atherectomy, or excimer laser angioplasty) at the Washington Hospital Center, “minor” (1 to 4 times control) CK-MB elevations were associated with a 2.5-fold increase in late mortality (5). Likewise, among 1,012 patients enrolled in the Coronary Angioplasty Versus Excisional Atherectomy Trial (CAVEAT), a CK-MB rise to ≥3 times control was associated with adverse 30-day clinical outcome (23). Preliminary analysis of data from the Balloon vs Optimal Atherectomy Trial (BOAT) demonstrated no adverse mortality effects related to periprocedural MI; however, follow-up data beyond one year has yet to be reported (21).
In the most detailed prospective analysis to date, Topol et al. (7)documented the relationship between periprocedural MI and late (three-year) mortality among the 2,099 patients enrolled in the EPIC trial. Concordant with the larger retrospective analyses, CK elevation to any degree was associated with excess mortality during follow-up, and the survival curves for patients with normal and abnormal postprocedural CK values diverged progressively with time. A graded relationship between absolute CK level and mortality was present, with the risk ratio for death increasing from 1.5 for CK levels ≥1 × control, to 2.4 for patients with peak CK levels ≥10 × control.
Periprocedural mi and late mortality
Various mechanisms have been postulated in attempt to provide a biologic basis for the relationship between periprocedural MI, which is in most instances small in magnitude and clinically inapparent, and late mortality. Strong evidence exists to support the concept that elevated serum levels of CK-MB serve as a marker for profound myocardial cell injury or death (24,25). The recognized association between postangioplasty CK elevation and intraprocedural events including distal coronary embolization, side branch occlusion, and transient abrupt vessel closure likewise support myocyte necrosis as the etiology of enzyme elevations following percutaneous revascularization procedures (1).
It has been suggested that periprocedural MI, reflected by either small or large magnitude elevations in serum CK, may, in a fashion analogous to spontaneous MI, predispose to the development of reentrant circuits, which in turn predispose to ventricular arrhythmia and death. In the Washington Hospital Center series, for example, 42% of deaths related to “minor” CK elevations (CK peak 1 to 4 times control) were arrhythmogenic in etiology (5). The occurrence of macro- or microscopic distal embolization, which may become manifest as an otherwise uncomplicated rise in CK following a coronary intervention, could potentially compromise collateral channels in the distal coronary bed (26,27). By this mechanism, the potential for collateral flow from the treated vessel may be reduced should MI subsequently occur in the territory contralateral vessel, increasing the severity of the future MI. Furthermore, it is possible that individuals prone to myocardial necrosis during percutaneous coronary revascularization are especially predisposed to future spontaneous cardiac events.
Periprocedural MI and target vessel revascularization
Several potential explanations are available to explain the association between periprocedural MI and the reduced incidence of target vessel revascularization (TVR) observed in the present study. First, although no data exist to suggest that MI reduces the actual likelihood of vascular renarrowing (e.g., angiographic restenosis) following angioplasty, it would be expected that patients who experience myocyte necrosis in the territory of a restenotic vessel would be less likely, owing to the presence of nonviable myocardium, to experience angina as a result of recurrent narrowing (e.g., clinical restenosis). Consequently, these patients would be less likely than noninfarct patients to undergo referral for ischemia-driven TVR. Second, because periangioplasty MI is associated with increased late mortality, slightly fewer patients who experience transient CK elevations following revascularization survive to experience restenosis. Finally, the reduced incidence of TVR in patients who experience a periprocedural MI may in part reflect a subtle bias on the part of the treating physician, who may be less inclined to attempt a repeat revascularization procedure given the complicated nature of the initial procedure.
IIb/IIIa receptor antagonism and restenosis
In EPIC, the first randomized trial to assess the clinical utility of platelet glycoprotein IIb/IIIa receptor antagonists as an adjunct to percutaneous revascularization, abciximab therapy was associated with a significant reduction in the need for target vessel revascularization at both 6 months and 3 years (7,11). These results led to speculation that potent platelet inhibition may have beneficial effects in terms of restenosis. In EPILOG (Evaluation of PTCA to Improve Long-term Outcome by 7E3 GP IIb/IIIa Receptor Blockade Trial) and other subsequent trials of IIb/IIIa receptor blockade, however, no reductions in clinical or angiographic restenosis have been detected (15,28,29). The findings of the present study may have implications with respect to this discrepancy.
In the EPIC trial, abciximab therapy was associated with a 39.5% relative reduction in the incidence of periprocedural MI, whereas a much larger 58% relative reduction in MI was observed in conjunction with abciximab in EPILOG (10,15). This increased proportionate reduction in MI observed in EPILOG compared to EPIC, based on the inverse relationship between periprocedural MI and TVR, may explain, at least in part, why a reduction in TVR was apparent in the EPIC but not the EPILOG trial.
Among patients enrolled in the EPIC trial, myocardial enzyme elevation following coronary intervention was associated with a significant reduction in subsequent clinical restenosis, determined by the need for subsequent TVR. The validity of this association is further supported by the presence of a “gradient of effect,” whereby larger elevations in CK portended progressive decreases in the risk of TVR. This finding raises a potential flag of caution with respect to the interpretation of composite end points in clinical trials of agents that favorably influence the incidence of postintervention MI. Because of this relationship, such agents, while expected to improve procedural safety and clinical outcome, may be inappropriately perceived as having a deleterious effect on clinical restenosis.
☆ The EPIC study was supported by a research grant from Centocor, Malvern, Pennsylvania.
- Coronary Angioplasty Versus Excisional Atherectomy Trial
- Evaluation of IIb/IIIa platelet receptor antagonist 7E3 in Preventing Ischemic Complications trial
- Evaluation of PTCA to Improve Long-Term Outcome by 7E3 GP IIb/IIIa receptor blockade trial
- Integrelin to Manage Platelet Aggregation to Combat Thrombosis trial
- myocardial infarction
- target vessel revascularization
- Received June 8, 1998.
- Revision received September 18, 1998.
- Accepted October 30, 1998.
- American College of Cardiology
- Abdelmeguid A.E.,
- Whitlow P.L.,
- Sapp S.K.,
- Ellis S.G.,
- Topol E.J.
- Abdelmeguid A.E.,
- Topol E.J.,
- Whitlow P.L.,
- Sapp S.K.,
- Ellis S.G.
- Redwood S.,
- Popma J.,
- Kent K.,
- et al.
- Tardiff B.,
- Califf R.,
- Tcheng J.,
- et al.
- Topol E.J.,
- Ferguson J.J.,
- Weisman H.F.,
- et al.
- Califf R.M.,
- Abdelmeguid A.E.,
- Kuntz R.E.,
- et al.
- Ryan T.J.,
- Faxon D.P.,
- Gunnar R.M.,
- et al.
- Weintraub W.,
- Ghazzal Z.,
- Douglas J. Jr.,
- Morris D.,
- King S. III.
- Baim D.,
- Cutlip D.,
- Sharma S.,
- et al.
- Klein L.W.,
- Kramer B.L.,
- Howard E.,
- Lesch M.
- Harrington R.A.,
- Lincoff A.M.,
- Califf R.M.,
- et al.
- Ahmed S.A.,
- Williamson J.R.,
- Roberts R.,
- Clark R.E.,
- Sobel B.E.
- Pfisterer M.,
- Rickenbacher P.,
- Kiowski W.,
- Muller-Brand J.,
- Burkart F.
- Ellis S.,
- Serruys P.,
- Popma J.,
- et al.