We determined the timing of ischemic complications within 30 days after percutaneous coronary intervention (PCI) in patients enrolled in the Enhanced Suppression of the Platelet IIb/IIIa Receptor with Integrilin Therapy (ESPRIT) trial. Complications (death, myocardial infarction [MI], target vessel revascularization) occurred in 178 of 2064 patients (8.6%) within 30 days. More than 85% of complications occurred within the 24 hours following randomization, with the greatest risk hazard at 12–18 hours. Unexpectedly, 31% of patients who ultimately met criteria for an endpoint MI within 24 hours of PCI had completely normal CK-MB concentrations at the first 6-hour measurement. There was no “rebound” increase in events after cessation of eptifibatide. Treatment benefit persisted to 30 days. Post-procedural MI is often not detected until >= 12 hours after PCI. Treatment with a glycoprotein IIb/IIIa inhibitor is the only modifiable parameter that reduces the risk for early ischemic complications.

       Percutaneous coronary intervention (PCI) is an effective treatment for myocardial ischemia in patients with coronary artery disease. More than one million PCI procedures were performed worldwide in 1998.1 A major limitation of percutaneous revascularization is the risk of acute ischemic complications associated with the procedure, including myocardial infarction (MI), urgent target vessel revascularization (TVR), and death. Approximately 10% of angioplasty procedures are complicated by one or more of these events.2 The timing of such events is crucial to understanding the pathophysiology of ischemic complications and determining the optimal duration of antithrombotic therapy. By using data from the Enhanced Suppression of the Platelet IIb/IIIa Receptor with Integrilin Therapy (ESPRIT) trial, we sought to determine the incidence, timing, and predictors of ischemic complications after non-urgent stent PCI.

Materials and Methods
       Study Population. The design and results of the ESPRIT trial have been previously published.3,4 Briefly, 2,064 patients undergoing non-urgent coronary stenting in the United States or Canada were randomly assigned to receive eptifibatide, given as two 180 µg/kg boluses 10 minutes apart and a continuous infusion of 2.0 µg/kg/min for 18–24 hours, or placebo.        Randomization was undertaken in the cardiac catheterization laboratory immediately prior to study drug administration and balloon inflation. A weight-adjusted heparin regimen was recommended; in patients not on heparin prior to PCI, an initial bolus of 60 U/kg was to be given to achieve a target activated clotting time of 200–300 seconds. Administration of heparin after the PCI was discouraged. Per protocol, a thienopyridine was to be given, with 98% of patients receiving thienopyridine treatment as an adjunct to the PCI procedure. By protocol, blood samples were collected in all patients at baseline (within 4 hours before study drug administration), and every 6 hours after randomization up to 24 hours. Although the 24-hour sample was not required if patients were discharged home earlier than 24 hours, the 6-hour, 12-hour, and 18-hour samples were required in all patients. Assays for levels of creatine kinase (CK) and CK-MB cardiac markers were performed by a central core biochemical laboratory. In order to include results from samples drawn within 2 hours of the scheduled time, the reported CK and CK-MB results for the 6-hour time point includes samples drawn in the 6- to 8-hour window after PCI.

Figure 1

Timing of the composite of death, myocardial infarction, or urgent target vessel revascularization. The cumulative event rate for patients who had events within 30 days is plotted against the number of days after randomization. Of the patients who had an event within 30 days, 86% had at least one event within 24 hours.

Figure 2

Timing of the composite of death, myocardial infarction, or urgent target vessel revascularization by treatment group for patients who had events within 30 days. The cumulative event rate for each treatment group is plotted against the time after randomization for the first 48 hours.

       Endpoints and timing of events. Death, MI, and urgent TVR within 30 days after the index PCI were evaluated. To identify an endpoint MI, we used an algorithm that evaluated the core laboratory CK-MB assay results and clinical criteria.4 Enzymatic criteria for an endpoint MI required at least two of the core laboratory CK-MB values within 24 hours of PCI to be >= 3, the upper limit of normal (with at least a 25% increase in CK-MB if the last pre-randomization value was abnormal). MI events fulfilling enzymatic criteria did not undergo adjudication or require clinical corroboration. MI events reported on the case report forms but not fulfilling the enzymatic criteria were adjudicated by the Clinical Events Committee by evaluating source documents. Adjudication of MI required corroboration of a clinical syndrome consistent with MI, as well as supportive electrocardiographic or cardiac marker data. All reported TVR events were adjudicated to identify urgent TVR events. TVR procedures were considered urgent if performed within 24 hours of documented ischemic episodes that included one of the following: rest pain, presumed ischemic in origin and lasting at least 5 minutes; new ischemic electrocardiographic changes; acute pulmonary edema; ventricular arrhythmias; or other signs or symptoms presumed secondary to coronary ischemia.

Figure 3

Hazard plot of death, myocardial infarction, or urgent target vessel revascularization. The hazard function, based on 6 hours as the unit of time, for the composite of death, myocardial infarction, or urgent target vessel revascularization is plotted against the time, in 6-hour intervals, after randomization.

       The following criteria were used to determine timing of events: for death, the time reported on the case report form; for urgent TVR and for adjudicated MI, the time reported on the case report form and confirmed or modified by the Clinical Events Committee; and for MI fulfilling core-laboratory enzymatic criteria, the time based on the first elevation of CK-MB to >= 3, the upper limit of normal (even though two elevations were required to meet the definition of an MI).
       Comparisons were performed between groups with early events (within 24 hours of randomization), late events (24 hours, 30 days after randomization), and no events. The cutoff for early events was 24 hours, as this was the maximum duration for study drug infusion. The timing of events within the first 24 hours, particularly for MI, was examined in detail.
       Statistical Analysis. Baseline, procedural, and angiographic characteristics are summarized as medians with 25th and 75th percentiles for continuous measures and as percentages for discrete measures. Event rates are summarized as percentages or odds ratios with 95% confidence intervals (CIs). For comparisons between groups, the Pearson chi-square test was used for categorical variables, and the Wilcoxon rank-sum test was used for continuous variables.

Figure 4

Relative increase in CK-MB values over time. The change in CK-MB (expressed as percent increase over baseline CK-MB value) for serum samples sent within 24 hours of percutaneous coronary intervention (PCI) is shown for the 126 patients who developed myocardial infarction (fulfilling core laboratory CK–MB criteria) within 24 hours.

       Patient records were repeated for the following intervals: 0–6 hours on drug, 6–12 hours on drug, > 12 hours on drug, 0–6 hours off drug, 6–12 hours off drug, 12–18 hours off drug, and > 18 hours off drug. Patients were counted as being event free during each interval they experienced without an event. Thus, a patient could have up to seven records. If the patient had an event (death, MI, and/or repeat revascularization), then the endpoint was attributed to only the period in which it occurred and the repetitions of records were stopped. By using these multiple intervals for each patient, a generalized linear model of the 30-day endpoint was developed. This model could test the treatment effect overall, the event rates by time interval, and the difference in change in event rates over time for placebo versus treated patients. The rates on versus off drug for the two treatment groups were also used to determine rebound from the drug.
       Cumulative distribution plots of death, MI, and revascularization to 30 days were generated for the patients. Other methods were used to illustrate the change in risk over time. Cumulative distribution plots of death, MI, and revascularization to 30 days were generated for the patients who had an event overall and by treatment. An actuarial life table was created across all patients, dividing the data into 6-hour intervals. This method calculated the hazard at each interval along with the corresponding 95% confidence interval. Finally, for the timing of MI, the percent change in CK-MB from baseline was plotted at each time of the recording.
       Although the protocol designated the time points at which the CK-MB measures were to occur, there was some fluctuation in actual blood specimen collection times. For comparisons of CK-MB values across specific time points, all CK-MB values recorded within a 2-hour window of the designated time point (e.g., 6 and 12 hours after PCI) were considered to have been performed at the designated time.

Figure 5

Assessment for rebound effect after cessation of eptifibatide. Generalized linear modeling was used to test for rebound effect between the two treatment groups. The estimated probability with 95% confidence intervals for death, myocardial infarction (MI), or urgent target vessel revascularization (TVR) is displayed on the y-axis for successive time periods. Patients who had events were dropped from succeeding time blocks. The very similar pattern of events for both treatment groups, both on and off the study drug, indicates no discernible rebound effect.

       An intention-to-treat analysis was used for all comparisons between treatment groups. All p values < 0.05 were considered statistically significant. All analyses were performed using SAS® statistical software (SAS Institute, Inc., Cary, North Carolina).

Results
       Ischemic complications occurred in 178 patients (8.6%) during the first 30 days, with 10 deaths, 163 patients with MIs, and 28 patients with urgent TVR procedures. The timing of the endpoints for both treatment groups is shown in Figure 1 and in Table 1. The results by treatment are shown in Figure 2. Of the patients who had ischemic events within 30 days, 66% had the first event within 12 hours, 82% within 18 hours, 86% within 24 hours, and 88% within 48 hours after intervention (Table 1, Figure 1). For the composite of death and MI, the first event occurred within 24 hours in 86% of patients who had an event by 30 days. However, only 54% of patients who required urgent TVR within 30 days underwent TVR within 24 hours, and an additional 35% required their first urgent TVR between 24 hours and 7 days of the initial procedure. The risk of death, MI, or urgent TVR throughout the first 48 hours is illustrated in Figure 3, with the hazard function for the composite endpoint plotted for each 6-hour interval.
       During the first 24 hours, 126 patients (6.1%) experienced MI, as determined by the core laboratory enzymatic criteria (Figure 4). In 9 of these patients (7%), there were no CK-MB values within the first 8 hours, and in 38 patients (31%), the CK-MB values at 6 hours after PCI were within normal limits (<= 1, the upper limit of normal). In these patients, CK-MB above the upper normal limit was not detected until the second CK-MB levels were drawn (12 hours after PCI). Of the 38 patients with delayed CK-MB elevation, 8 had at least one procedural complication (major dissection, no reflow, side branch closure, and distal embolization); the remaining 30 patients had no procedural complications.
       Of the 2,064 patients enrolled in the ESPRIT trial, 153 (7.4%) experienced ischemic events within 24 hours, 25 (1.2%) had events between 24 hours and 30 days, and 1886 (91.4%) had no events. Baseline clinical and procedural characteristics of the three groups are listed in Tables 2 and 3. In the univariate analysis, the following characteristics were found more frequently in patients in whom an early (within 24 hours of randomization) ischemic event developed: increasing age, higher body weight, prior PCI, absence of diabetes, recent acute coronary syndrome, randomization to receive placebo, and lower initial heparin doses. Patients with early events were more likely to have coronary thrombus and impaired coronary flow both before and after intervention; they were also more likely to develop a major dissection, abrupt closure, no reflow, thrombus formation, side branch closure, or distal embolization event before the ischemic event.
       By use of stepwise multivariable logistic regression, we identified six independent predictors of early events (within 24 hours): pre-interventional thrombus (odds ratio [OR] 2.05 [1.07, 3.92]), absence of diabetes (OR 1.96 [1.18, 3.26]), prior PCI (OR 1.78 [1.23, 2.57]), lower grades of pre-interventional Thrombolysis in Myocardial Infarction (TIMI) flow (OR 1.74 [1.17, 2.57]), randomization to placebo (OR 1.52 [(1.08, 2.14]), and advanced age (OR 1.26 per 10 years [1.08, 1.47]). There were no significant interactions between treatment group assignment and the other variables in the model, indicating a significant benefit with eptifibatide treatment regardless of the presence or absence of other risk factors.
       For the 178 patients who had an event within 30 days, generalized linear modeling was used to test for a difference in the timing of the endpoint between treatment groups. Although fewer patients randomly assigned to receive eptifibatide had ischemic events within the first 24 hours (6.0% versus 8.9%; p = 0.01), modeling indicated no difference in the timing of events between treatment groups. The results of linear modeling to test for rebound effects are shown in Figure 5. The similar pattern of events for both treatment groups indicates no discernible evidence of a rebound effect when the study drug was discontinued.

Discussion
       This study characterizes the timing of ischemic events after contemporary coronary stenting, with and without the use of eptifibatide. More than 85% of first ischemic events occur within 24 hours after PCI. Surprisingly, in our study a substantial proportion of patients experienced ischemic events beyond the first 6 hours after intervention, with the highest risk in the 12- to 18-hour interval. Three clinical characteristics (absence of diabetes, prior PCI, and advanced age) and two angiographic factors (presence of thrombus prior to intervention and reduced TIMI flow) were risk factors for ischemic events within 24 hours. Eptifibatide significantly reduced the incidence of ischemic complications within the first 24 hours — a benefit that was sustained at 30 days. Eptifibatide was effective in reducing early ischemic events, it did not influence the timing of ischemic events, and there was no rebound effect after the drug was stopped.
       Standard balloon angioplasty, without stent placement or use of glycoprotein IIb/IIIa inhibitors, is associated with an approximate 12% incidence of death, MI, or urgent revascularization at 30 days.5–7 Coronary stenting does not affect rates of death or MI, but does reduce the incidence of abrupt vessel closure and the need for bypass surgery after PCI.8–12 The composite of death, MI, or urgent TVR occurs in 10–11% of patients undergoing routine stenting without glycoprotein IIb/IIIa inhibitors,2–4 but use of these agents reduces the risk for ischemic events by 35–50%.2–4
       Our findings indicate that most ischemic events (> 80%) occur within the first 24 hours after randomization, corresponding to the maximum duration of study drug infusion in both arms of ESPRIT. A small number of events occur up to 30 days after randomization; however, there is a rise in the event curves after 12 hours, with a “plateau” at 18–24 hours.
       The highest hazard for ischemic events occurred 12–18 hours after the first balloon inflation, raising the possibility that post-procedural MI may occur well after the PCI procedure. These findings appear contradictory to those of the Integrilin to Minimize Platelet Aggregation and Coronary Thrombosis-II (IMPACT-II) study in which most ischemic events were reported to have occurred within the first 6 hours after PCI;13 however, the IMPACT-II investigators assumed that any MI detected within 24 hours of PCI without clinically evident ischemia occurred at the time of balloon inflation. In all ESPRIT patients, CK-MB was first measured 6 hours after intervention. Given the early “washout” kinetics of CK-MB release after successful primary angioplasty,14 myonecrosis that is directly attributable to the balloon inflation and stent deployment should be detectable at 8 hours. In ESPRIT, 37% of patients who developed MI (per the core laboratory CK-MB criteria) within 24 hours of PCI had normal CK and CK-MB levels at 6 hours. The delayed rise in cardiac enzymes may thus reflect distal embolization and microvascular plugging that occur during PCI with slow washout of the microvasculature. Alternatively, ongoing platelet adhesion and embolization or delayed side-branch closure may contribute to the pathophysiology of early ischemic events after PCI. In any event, these findings indicate that early measurement of CK-MB may not identify all peri-procedural myocardial infarctions, and they highlight the importance of measuring CK-MB levels 12 to 18 hours after PCI.
       Cessation of heparin can result in a rebound effect characterized by increased thrombin activity and myocardial reinfarction.15 In our study, the similar pattern of events for the eptifibatide and placebo groups, both on and off the study drug, indicates no discernible rebound effect after the discontinuation of eptifibatide. These findings are consistent with a previous study showing no evidence of rebound ischemia after stopping eptifibatide.16 Furthermore, the similar timing of events between the two treatment groups provides additional evidence that eptifibatide prevents, rather than simply delays, the occurrence of ischemic complications. Event curves from the primary ESPRIT analysis reveal a benefit seen with eptifibatide at 48 hours that is still significant at 30 days4 and 6 months.17
       Baseline characteristics may assist clinicians in targeting patients for extended in-hospital monitoring and possibly more potent or prolonged antithrombotic therapy. In our analysis, the baseline characteristics that predicted early ischemic events were advanced age, absence of diabetes, prior PCI, lower TIMI grade flow, and presence of thrombus prior to intervention. Patients with advanced age have been shown to be at higher risk for complications after PCI in some studies, but not in others.18–20 Elderly patients in general have more complex coronary artery disease, with a higher incidence of multivessel disease and more calcified lesions.21 Similarly, patients with prior angioplasty may have more multivessel disease. Angiographic thrombus has been documented as a risk factor for abrupt closure and other thrombotic complications in numerous studies.22–25 Reduced coronary flow may represent either acute or chronic occlusion. Reduced TIMI grade flow prior to primary angioplasty for acute MI has been associated with lower procedural success rates and higher mortality rates at 30 days.26 A higher complication rate after angioplasty for chronic occlusions has been well documented.27
       The lower rate of ischemic complications in patients with a history of diabetes was an unexpected finding in this analysis. Diabetes usually increases the risk for ischemic events,27 yet it appeared to have a protective influence in the IMPACT-II7 and Do Tirofiban and ReoPro Give Similar Efficacy? (TARGET) trials.28 This paradox may be related to a bias in enrolling only diabetic patients who were otherwise at low risk for complications. In our multivariable model, there was no significant interaction with treatment assignment, indicating that the benefit of eptifibatide applied to all ESPRIT patients, regardless of risk profile. Furthermore, eptifibatide treatment was the only modifiable risk factor identified in the model. A previous analysis from the ESPRIT trial showed only a minimal and nonsignificant reduction in angiographic complications (including thrombus formation and reduced flow) with eptifibatide.29 The main benefit of eptifibatide is a significant reduction in myocardial infarction after otherwise uneventful PCI procedures.
       There are several limitations of our analysis. The definition of MI used in this study excluded small CK-MB elevations, which have been associated with adverse outcomes in some studies.30,31 Use of a different definition of MI might have altered these findings. Blood samples for CK-MB analysis were collected every 6 hours. More frequent sampling might provide a clearer picture of the temporal profile of ischemic events after PCI. This study showed no evidence of rebound phenomenon for the composite endpoint of death, MI, and urgent TVR, but was not adequately powered to assess for rebound affecting only urgent revascularization. By study protocol, eptifibatide was to be used for 18–24 hours after PCI for all patients enrolled in ESPRIT. In this study, we cannot determine whether a rebound phenomenon may occur when eptifibatide is used for a shorter duration.

Conclusions
       Ischemic complications are common after stent PCI procedures, usually occurring within 24 hours after the procedure. The early hazard for such complications extends to approximately 18 hours after PCI. Post-procedural MI is frequently not detected until >= 12 hours after PCI. Although baseline clinical and angiographic characteristics can help identify the highest risk patients, glycoprotein IIb/IIIa inhibitors significantly reduce the incidence of ischemic complications across the spectrum of risk.
       Acknowledgment. The authors wish to thank Rebecca Teaff for her editorial assistance in reviewing and preparing the manuscript.

1. Topol EJ and Serruys PW. Frontiers in interventional cardiology. Circulation 1998;98:1802–1820.
2. The EPISTENT Investigators. Randomised placebo-controlled and balloon-angioplasty–controlled trial to assess safety of coronary stenting with use of platelet glycoprotein-IIb/IIIa blockade. Lancet 1998;352:87–92.
3. O’Shea JC, Madan M, Cantor WJ. Design and methodology of the ESPRIT trial: Evaluating a novel dosing regimen of eptifibatide in percutaneous coronary intervention. Am Heart J 2000;140:834–839.
4. The ESPRIT Investigators. Novel dosing regimen of eptifibatide in planned coronary stent implantation. ESPRIT: A randomised, placebo-controlled trial. Lancet 2000;356:2037–2044.
5. The EPIC Investigators. Use of a monoclonal antibody directed against the platelet glycoprotein IIb/IIIa receptor in high-risk coronary angioplasty. N Engl J Med 1994;330:956–961.
6. The EPILOG Investigators. Platelet glycoprotein IIb/IIIa receptor blockade and low-dose heparin during percutaneous coronary revascularization. N Engl J Med 1997;336:1689–1696.
7. The IMPACT-II Investigators. Randomised placebo-controlled trial of effect of eptifibatide on complications of percutaneous coronary intervention: IMPACT-II. Integrilin to Minimise Platelet Aggregation and Coronary Thrombosis-II. Lancet 1997;349:1422–1428.
8. Lincoff AM, Topol EJ, Chapekis AT. Intracoronary stenting compared with conventional therapy for abrupt vessel closure complicating coronary angioplasty: A matched case-control study. J Am Coll Cardiol 1993;21:866–875.
9. Schömig A, Kastrati A, Mudra H. Four-year experience with Palmaz-Schatz stenting in coronary angioplasty complicated by dissection with threatened or present vessel closure. Circulation 1994;90:2716–2724.
10. Roubin GS, Cannon AD, Agrawal SK. Intracoronary stenting for acute and threatened closure complicating percutaneous transluminal coronary angioplasty. Circulation 1992;85:916–927.
11. George BS, Voorhees WD, Roubin GS. Multicenter investigation of coronary stenting to treat acute or threatened closure after percutaneous transluminal coronary angioplasty: Clinical and angiographic outcomes. J Am Coll Cardiol 1993;22:135–143.
12. Cantor WJ, Peterson ED, Popma JJ. Provisional stenting strategies: systematic overview and implications for clinical decision-making. J Am Coll Cardiol 2000;36:1142–1151.
13. Thel MC, Califf RM, Tardiff BE. Timing of and risk factors for myocardial ischemic events after percutaneous coronary intervention (IMPACT-II). Am J Cardiol 2000;85:427–433.
14. Laperche T, Steg PG, Benessiano J. Patterns of myoglobin and MM creatine kinase isoforms release early after intravenous thrombolysis or direct percutaneous transluminal coronary angioplasty for acute myocardial infarction, and implications for the early noninvasive diagnosis of reperfusion. Am J Cardiol 1992;70:1129–1134.
15. Granger CB, Miller JM, Bovill EG. Rebound increase in thrombin generation and activity after cessation of intravenous heparin in patients with acute coronary syndromes. Circulation 1995;91:1929–1935.
16. Schulman SP, Goldschmidt-Clermont PJ, Topol EJ. Effects of Integrelin, a platelet glycoprotein IIb/IIIa receptor antagonist, in unstable angina: a randomized multicenter trial. Circulation 1996;94:2083–2089.
17. O’Shea JC, Hafley GE, Greenberg S. Platelet glycoprotein IIb/IIIa Integrin blockade with eptifibatide in coronary stent intervention: The ESPRIT trial. A randomized controlled trial. J Am Med Assoc 2001;285:2468–2473.
18. Kern MJ, Deligonul U, Galan K. Percutaneous transluminal coronary angioplasty in octogenarians. Am J Cardiol 1988;61:457–458.
19. Buffet P, Danchin N, Juilliere Y. Percutaneous transluminal coronary angioplasty in patients more than 75 years old: Early and long-term results. Int J Cardiol 1992;37:33–39.
20. Simpfendorfer C, Raymond R, Schraider J. Early and long-term results of percutaneous transluminal coronary angioplasty in patients 70 years of age and older with angina pectoris. Am J Cardiol 1988;62:959–961.
21. Lincoff AM. Patients at high risk for ischemic complications. J Invas Cardiol 1994;6:13A–18A.
22. Tan K, Sulke N, Taub N. Clinical and lesion morphologic determinants of coronary angioplasty success and complications: current experience. J Am Coll Cardiol 1995;25:855–865.
23. Ellis SG, Roubin GS, King SB III. Angiographic and clinical predictors of acute closure after native vessel coronary angioplasty. Circulation 1988;77:372–379.
24. Mabin TA, Holmes DR Jr., Smith HC. Intracoronary thrombus: Role in coronary occlusion complicating percutaneous transluminal coronary angioplasty. J Am Coll Cardiol 1985;5:198–202.
25. Detre KM, Holmes DR Jr., Holubkov R. Incidence and consequences of periprocedural occlusion. The 1985–1986 National Heart, Lung, and Blood Institute Percutaneous Transluminal Coronary Angioplasty Registry. Circulation 1990;82:739–750.
26. Brodie BR, Stuckey TD, Hansen C. Benefit of coronary reperfusion before intervention on outcomes after primary angioplasty for acute myocardial infarction. Am J Cardiol 2000;85:13–18.
27. Ellis SG, Vandormael MG, Cowley MJ. Coronary morphologic and clinical determinants of procedural outcome with angioplasty for multivessel coronary disease. Circulation 1990;82:1193–1202.
28. Topol EJ, Moliterno DJ, Hermann HC. Comparison of two platelet glycoprotein IIb/IIIa inhibitors, tirofiban and abxicimab, for the prevention of ischemic events with percutaneous coronary revascularization. N Engl J Med 2001;344:1888–1894.
29. Blankenship JC, Tasissa G, O’Shea JC, et al., the ESPRIT Investigators. Effect of glycoprotein IIb/IIIa receptor inhibition on angiographic complications during percutaneous coronary intervention in the ESPRIT trial. J Am Coll Cardiol 2001;38:653–658.
30. Califf RM, Abdelmeguid AE, Kuntz RE, et al. Myonecrosis after revascularization procedures. J Am Coll Cardiol 1998;31:241–251.
31. Tardiff BE, Califf RM, Tcheng JE, et al. Clinical outcomes after detection of elevated cardiac enzymes in patients undergoing percutaneous intervention. IMPACT-II Investigators. Integrilin (eptifibatide) to Minimize Platelet Aggregation and Coronary Thrombosis-II. J Am Coll Cardiol 1999;33:88–96.