Benefits and Risks of Abciximab Use in Primary Angioplasty for Acute Myocardial Infarction
The Controlled Abciximab and Device Investigation to Lower Late Angioplasty Complications (CADILLAC) Trial
Background— Trials of platelet glycoprotein IIb/IIIa inhibitors as adjuncts to primary percutaneous coronary intervention for acute myocardial infarction (MI) have shown improved early clinical and angiographic outcomes with treatment. However, variations in trial designs, modest sample sizes, and limited long-term follow-up have precluded these studies from being definitive.
Methods and Results— As a prespecified secondary analysis of the CADILLAC trial, we compared early and late outcomes by abciximab assignment among 2082 patients randomized in an open-label, 2×2 factorial-design trial of primary stenting versus angioplasty and abciximab treatment (n=1052) versus no abciximab treatment (n=1030). Baseline characteristics were balanced between groups. Abciximab treatment was associated with a significant reduction in the composite end point of death, MI, ischemia-driven target-vessel revascularization (TVR), or disabling stroke at 30 days (4.6% versus 7.0%; relative risk, 0.65; 95% CI, 0.46 to 0.93; P=0.01). Subacute thrombosis also was significantly reduced with abciximab treatment. At 12 months, however, rates of the composite end point did not differ significantly (18.4% for controls versus 16.9% for abciximab-treated patients; relative risk, 0.92; 95% CI, 0.76 to 1.10; P=0.29), reflecting a decrease in the relative difference in TVR rates (ie, no effect of abciximab on reducing restenosis). In an angiographic substudy (n=656), myocardial salvage, restenosis, and infarct-artery reocclusion at 7 months were unaffected by abciximab treatment. There was no significant interaction between stenting and abciximab treatment.
Conclusions— Adjunctive abciximab treatment during primary percutaneous coronary intervention significantly enhanced 30-day event-free survival, predominantly by reducing ischemia-driven TVR. Abciximab treatment did not affect the composite end point at 1 year, reflecting a lack of effect on restenosis.
Received October 12, 2002; de novo received April 10, 2003; revision received June 24, 2003; accepted June 25, 2003.
Trials of platelet glycoprotein (GP) IIb/IIIa integrin blockers with acute coronary syndromes have demonstrated improved outcomes in patients with unstable thrombotic coronary lesions undergoing percutaneous coronary intervention (PCI).1,2 This has suggested the potential for benefit in acute ST-segment elevation myocardial infarction (MI). However, only few modestly sized studies have examined GP IIb/IIIa blockade as adjunctive therapy to primary PCI in acute MI.3–5 The Controlled Abciximab and Device Investigation to Lower Late Angioplasty Complications (CADILLAC) trial was a large, randomized, open-label, multicenter study that examined the efficacy and safety of abciximab (ReoPro, Centocor and Eli Lilly) and coronary stenting in primary PCI for acute MI.6 The present analysis describes the abciximab-specific end point results at 30 days (as defined in the protocol) and extends these observations to 1 year.
The details of the CADILLAC trial have been described previously.6 In brief, 2082 patients sustaining an acute MI were randomized at 76 sites in 9 countries to receive 1 of 4 reperfusion strategies in an open-label, 2×2 factorial design, as follows: (1) balloon percutaneous transluminal coronary angioplasty (PTCA), (2) PTCA with abciximab, (3) MultiLink stent implantation (Guidant Corporation), or (4) MultiLink stent implantation with abciximab. Enrollment criteria included clinical symptoms of MI for <12 hours and either ≥1 mm ST-segment elevation in 2 contiguous leads or high-grade angiographic stenosis with an associated regional wall-motion abnormality. Major exclusion criteria included cardiogenic shock, a saphenous vein graft infarct lesion, an infarct artery reference vessel diameter <2.5 mm, lesion length >64 mm, and need for urgent bypass surgery. Patients received aspirin, ticlopidine, and heparin before angiography. Heparin was adjusted by nomogram to achieve an activated clotting time of >350 seconds in patients not receiving abciximab and 200 to 300 seconds in abciximab-treated patients. Abciximab was given as a 0.25-mg/kg intravenous bolus followed by a 0.125-μg/kg per min infusion for 12 hours. Patients randomized to PTCA were allowed to crossover and undergo stent implantation for an unacceptable angiographic result after balloon PTCA. Similarly, patients randomized to not receive abciximab were allowed to receive provisional or bail-out abciximab for refractory slow or no reflow or persistent thrombosis.
Patients randomized to abciximab were scheduled for discharge at 2 days (low-risk patients) or 3 days (high-risk patients) after successful intervention. Control patients were discharged at the discretion of the treating physician. Aspirin was continued indefinitely in all patients. Ticlopidine was given for 4 weeks in patients who underwent stenting and was optional in patients undergoing balloon PTCA alone.
Study Hypothesis and End Points
The primary hypotheses, end point definitions, and results of the CADILLAC trial have been previously described.6 A major secondary protocol-specified hypothesis was that in patients undergoing PCI (either PTCA or stent implantation) with abciximab and managed with an accelerated discharge strategy, the composite 30-day end point of death, reinfarction, urgent repeat revascularization, or disabling stroke would not be inferior to those of all patients undergoing PCI without abciximab. This hypothesis and associated analyses are the principal focus of this report. Prospectively defined tertiary efficacy end points included the 6- and 12-month composite rates of major adverse cardiac events (MACE)—death, reinfarction, disabling stroke, and all target vessel revascularization (TVR)—and rates of MACE components at 30 days, 6 months, and 12 months by abciximab treatment. Safety assessments included hemorrhagic and hematological complication rates. Angiographic outcomes included periprocedural thrombolysis in myocardial infarction (TIMI) flow grade, quantitative measures of lesion severity, and rates of dissection, new thrombus formation, distal thromboemboli, side branch occlusion, and no-reflow. Procedural success was defined as the achievement of TIMI grade 3 flow, a residual diameter stenosis <50%, and freedom from MACE within 7 days. Other end points included length of hospitalization, angiographic restenosis, and myocardial recovery from baseline to follow-up angiography at 7 months in a prospectively defined substudy of 900 patients.
For analysis of noninferiority in the 30-day composite end point rates between abciximab-treated and control patients in the pooled analysis, assuming a 12.5% event rate in patients not assigned to abciximab, the Blackwelder method provided that a sample size of 1700 randomized patients had 80% power to detect an absolute 4.0% difference between the 2 groups (ie, a relative risk noninferiority boundary of 1.32) using a significance level of 0.05. With 2082 patients randomized, 85% power was present to detect this difference or, conversely, 80% power was present to detect a delta of 3.7% (ie, a relative risk noninferiority boundary of 1.30). Testing for superiority was permitted if noninferiority was established. The trial was not powered to show a difference (relative to abciximab treatment) in events at 6 months or 1 year.
Categorical variables were compared by the likelihood-ratio χ2 test or the Fisher exact test. Continuous variables are presented as medians with 25th and 75th percentiles and were compared by 1-way ANOVA or the Kruskal-Wallis test. Pairwise comparisons were performed with the Wilcoxon 2-sample test. Time-to-event data were analyzed using survival techniques and compared by the log-rank test. The influence of baseline demographic and angiographic variables on the 12-month composite MACE end point was evaluated with logistic regression using the Wald χ2 test, and the results were expressed as odds ratios with 95% CI. Logistic regression models also were performed for the 12-month composite MACE end point to assess for an interaction between abciximab treatment with stenting and the infarct vessel. Statistically significant tertiary end points were considered hypothesis-generating only, and corrections for multiple comparisons were not performed. All analyses were by intention-to-treat unless otherwise stated, and all P values are 2-sided. Statistical tests were performed with SAS for Windows, v8.02.
Baseline clinical and angiographic characteristics did not differ significantly between patients randomized to abciximab versus control therapy, except that slightly fewer patients assigned to abciximab had infarction in the left anterior descending artery. The time from symptom onset to hospital arrival tended to be slightly shorter in abciximab-assigned patients (Table 1).
The number, length, and diameter of stents implanted were similar regardless of abciximab assignment (Table 2). Altogether, <10% of patients assigned to control received abciximab as bailout or rescue therapy for refractory thrombosis or no-reflow. By core laboratory analysis, rates of distal embolization, no-reflow, and postprocedural residual thrombus were similar in both treatment groups (Table 2).
In a multivariable Cox regression model that examined 30-day and 1-year MACE, there was no significant interaction between stenting and abciximab treatment (P=0.44 for 30 days; P=0.93 for 1 year). At 30 days, the primary abciximab composite end point was reduced from 7.0% to 4.6% with abciximab treatment (P<0.0001 for noninferiority; P=0.01 for superiority; relative risk, 0.65; 95% CI, 0.46 to 0.93; Table 3 and Figure). Abciximab significantly reduced the rates of recurrent ischemia, subacute thrombosis, ischemic TVR, and all TVR within 30 days (Table 3). Among the 4 treatment groups, ischemic TVR within 30 days was required in 5.6% of patients after PTCA alone, 3.2% after stenting alone, 3.4% after PTCA with abciximab, and 1.6% with stenting with abciximab (4-way P=0.004; compared with stenting with abciximab, P<0.001 versus PTCA alone and P<0.05 versus PTCA with abciximab). Abciximab also reduced the 30-day incidence of subacute thrombosis (2.0% after PTCA only, 1.0% after stenting only, 0.8% after PTCA with abciximab, and 0.0% after stenting with abciximab [P=0.01]). As reported,6 the 30-day rates of the component events (death, reinfarction, or disabling stroke) did not differ significantly among the 4 treatment strategies relative to abciximab therapy. Results were similar, but somewhat less strong, when analysis was performed by actual treatment received (Table 4).
By 12 months, abciximab was no longer associated with significant differences in the composite incidence of MACE (16.9% with abciximab versus 18.4% with control, P=0.29; relative risk, 0.92; 95% CI, 0.76 to 1.10; Figure) because late events accrued almost equally in both groups. A weak trend was still present for fewer TVR procedures in abciximab-assigned patients (Tables 3 and 4⇑); the absolute difference in TVR rates between the 2 treatment groups at 30 days (1.9% by intention-to-treat analysis) persisted at 1 year (2.2%). No major subgroups seemed to benefit preferentially from abciximab treatment, except for possibly patients with right coronary artery MI (Table 5). In a logistic regression model for 1-year MACE that incorporated infarct size, abciximab treatment, and their interaction term, this benefit remained statistically significant (P=0.0485).
Abciximab therapy was safe; no patient sustained an intracranial hemorrhage, and moderate or major bleeding was not increased (Table 3). Mild thrombocytopenia was increased by abciximab treatment (Table 3). Transfusion requirements were increased with abciximab treatment in the as-treated analysis (Table 4), possibly related to the use of bailout abciximab in patients treated with higher levels of heparin anticoagulation, but not by intention to treat (Table 3).
Of 900 patients eligible for follow-up protocol angiography at 7 months, 656 (73%) underwent the procedure. Compared with patients not randomized to abciximab, the rates of restenosis, infarct artery reocclusion, and degree of myocardial salvage were unaffected by abciximab treatment.
The CADILLAC trial, the largest randomized trial investigating mechanical reperfusion strategies with platelet GP IIb/IIIa integrin blockade in the management of acute MI, showed that abciximab clinically and statistically significantly reduced adverse ischemic events at 30 days. Abciximab treatment reduced rates of recurrent ischemia requiring repeat TVR and essentially eliminated subacute thrombosis. Given that the 30-day event rate was lower than expected, with the same relative risk boundary (1.32), the 4.6% composite event rate among abciximab-treated patients established noninferiority (P<0.0001), allowing the demonstration of superiority compared with control therapy (P=0.01). In longer-term (12-months) follow-up, the absolute benefit of abciximab therapy present at 30 days persisted, although the relative efficacy of abciximab during follow-up was offset by late TVR events, resulting in a loss of statistical significance. Even with the early (30-day) reduction in the composite end point, abciximab treatment did not significantly reduce the composite occurrence of death, reinfarction, or TVR at 6 months or 1 year.
The results of this trial thus are generally concordant with 2 of the 3 earlier trials of adjunctive GP IIb/IIIa inhibition with primary PCI for acute MI. In the 483-patient ReoPro and Primary PTCA Organization and Randomized Trial (RAPPORT),4 abciximab as an adjunct to PTCA (with bailout stenting) improved 30-day outcomes, primarily by reducing rates of recurrent ischemia requiring urgent TVR compared with placebo (5.6% versus 11.2% for the 30-day composite of death, reinfarction, or urgent TVR; P=0.03). At 6 months, however, the event-free survival curves of the 2 groups converged; as in CADILLAC, TVR became the predominant component of the MACE end point and eliminated the overall difference in 30-day outcomes. Similarly, in the 401-patient, randomized Intracoronary Stenting and Antithrombotic Regimen (ISAR)-2 trial,5 abciximab as an adjunct to stenting improved 30-day clinical outcomes compared with placebo (5.0% versus 10.5%; P=0.038 for the composite of death, reinfarction, or TVR). Again, by 6 months, the event curves converged, with the early benefits of reduced ischemic TVR surrendered to late restenosis. In both ISAR-2 and CADILLAC, angiographic restenosis rates were similar among stented patients randomized to abciximab versus control.
Alternatively, the 300-patient, randomized Abciximab before Direct angioplasty and stenting in Myocardial Infarction Regarding Acute and Long-term follow-up (ADMIRAL) trial suggested more sustained efficacy with abciximab treatment than was observed in CADILLAC.3 In ADMIRAL, abciximab treatment was associated with a higher incidence of postprocedural TIMI grade 3 flow (95.1% versus 86.7%; P=0.04) and less than half the rate of the primary composite end point (30-day death, reinfarction, or urgent TVR; 14.6% versus 6.0%; P=0.01). Benefits of abciximab were maintained at 6 months (primary end point rates of 34.7% versus 23.3%; P<0.05). In ADMIRAL, infarct-artery reocclusion (TIMI grade 0 or 1 flow) at 6 months occurred in 12.0% of patients randomized to stent-PCI with placebo compared with 2.8% in the stent-PCI with abciximab group (P<0.05). Whether the high reocclusion rate in patients not treated with abciximab in ADMIRAL relates to the specific stent used or the fact that ≈15% of patients received PTCA or medical therapy (rather than stenting) is unknown. In CADILLAC, the rates of reocclusion in the infarct vessel at 7-month angiographic follow-up did not differ relative to treatment with abciximab.
Protocol differences could also account for the results in ADMIRAL versus CADILLAC. In ADMIRAL, there was no option for crossover to open-label abciximab in the control group, potentially enhancing differences between treatment groups. The ADMIRAL protocol specified repeat angiography at 24 hours in all patients, which could have indirectly led to less-delayed ischemic TVR (via preemptive PCI). Patient entry criteria also slightly differed, with cardiogenic shock excluded in the CADILLAC trial (in contrast to nearly 8% of patients in ADMIRAL), whereas ≈12% of patients in the CADILLAC trial did not present with ST-segment elevation or a new diagnosis of left bundle branch block. Perhaps most importantly, in CADILLAC, study drug was started in the catheterization suite, typically within minutes before angioplasty. A key protocol element of ADMIRAL was the initiation of study drug (abciximab or placebo) before arterial vascular access was established and well before the time of PCI. The mean time from study drug to angiography in ADMIRAL was 36 minutes, corresponding to abciximab administration ≈1 hour before angioplasty. TIMI flow grades at baseline increased as the time from study drug initiation increased, with this enhanced early reperfusion perhaps accounting for improved longer-term outcomes with abciximab treatment.
Given the loss of benefit at 6 months in CADILLAC, RAPPORT, and ISAR-2—trials where abciximab was initiated only in the catheterization suite—optimization of outcomes may thus depend on treatment before the catheterization laboratory environment. The advantage of the early establishment of reperfusion also is congruent from analysis of the 2507 patients in the Primary Angioplasty in Myocardial Infarction (PAMI) trials, which showed that early and late mortality were strikingly reduced in patients with spontaneous reperfusion independent of the final TIMI flow grade,7 and the randomized Primary Angioplasty Compatibility Trial (PACT), in which early infarct-artery reperfusion with reduced-dose alteplase resulted in greater early recovery of left ventricular function.8 These observations support ongoing trials of pharmacologically facilitated reperfusion before definitive mechanical revascularization.
Finally, unlike previous trials demonstrating improvements in left ventricular function with abciximab treatment,3,9 similar findings of myocardial recovery were not observed in CADILLAC. These data suggest that abciximab treatment does not influence the quality and eventual extent of left ventricular recovery when administered as in the CADILLAC trial, although the disparity in myocardial recovery among these studies again might reflect differences in patient selection or methods.
In conclusion, when abciximab is started immediately before primary PCI in acute MI, the likelihood of a successful procedural outcome and freedom from 30-day MACE is enhanced. The predominant effect of abciximab is in reducing early ischemia, abrupt vessel closure, and subacute thrombosis over the short term. Because the late composite event rate is dominated by restenosis, the relative benefit of this early, fixed treatment effect diminishes over time with a consequent loss of statistical significance; between 1 and 12 months, TVR events in both treatment arms predominate and are not affected by abciximab treatment. The absolute benefits are nonetheless maintained over longer-term follow-up (Figure). Accordingly, these results confirm and clarify the complementary effects of coronary stenting and adjunctive GP IIb/IIIa inhibition with abciximab in reducing adverse events after primary PCI.
Additional study is needed to characterize the use of abciximab for patients with cardiogenic shock, vein graft occlusion, or other high-risk conditions that were excluded from enrollment. Although the interaction between infarct vessel and abciximab effect was significant, this interesting finding still may reflect chance alone and therefore should be considered hypothesis-generating. Ultimately, the decision to use abciximab should be individualized—the benefits of improved procedural and short-term outcomes must be balanced against financial costs and the slightly higher risks of thrombocytopenia and blood-product transfusion. Finally, future studies should examine whether very early treatment (in emergency departments or earlier), rather than in the catheterization suite, might additionally improve clinical outcomes and myocardial salvage.
Dr Tcheng receives research support and speaker’s honoraria from Guidant and Eli Lilly. Dr Kandzari receives research support and speaker’s honoraria and serves as a consultant to Eli Lilly. Dr Effron is an employee and a stockholder of Eli Lilly. Dr Stone is a consultant to and receives research support from Guidant. He also receives research support from Eli Lilly. Dr Grines receives research support from Guidant and Eli Lilly.
Kong DF, Califf RM, Miller DP, et al. Clinical outcomes of therapeutic agents that block the platelet glycoprotein IIb/IIIa integrin in ischemic heart disease. Circulation. 1998; 98: 2829–2835.
Madan M, Berkowitz SD, Tcheng JE. Glycoprotein IIb/IIIa integrin blockade. Circulation. 1998; 98: 2629–2635.
Brener SJ, Barr LA, Burchenal JE, et al. Randomized, placebo-controlled trial of platelet glycoprotein IIb/IIIa blockade with primary angioplasty for acute myocardial infarction. Circulation. 1998; 98: 734–741.
Stone GW, Cox D, Garcia E, et al. Normal flow (TIMI 3) before mechanical reperfusion therapy is an independent determinant of survival in acute myocardial infarction: analysis from the primary angioplasty in myocardial infarction trials. Circulation. 2001; 104: 636–641.
Neumann FJ, Blasini R, Schmitt C, et al. Effect of glycoprotein IIb/IIIa receptor blockade on recovery of coronary flow and left ventricular function after the placement of coronary-artery stents in acute myocardial infarction. Circulation. 1998; 98: 2695–2701.