Author + information
- Received December 11, 1997
- Revision received March 3, 1998
- Accepted March 16, 1998
- Published online July 1, 1998.
- C.Michael Gibson, MS, MD, FACC†,*,
- Mukesh Goel, MD†,
- David J. Cohen, MD, MSc‡,
- Robert N. Piana, MD, FACC§,
- Lawrence I. Deckelbaum, MD, FACC∥,
- Katherine E. Harris, DrPH∥,
- Spencer B. King III, MD, FACC¶,
- for the RESTORE Investigators#
- ↵*Present address and address for correspondence: Dr. C. Michael Gibson, Allegheny General Hospital, 320 East North Avenue, Pittsburgh, Pennsylvania 15212-4772
Objectives. This study sought to investigate the effects of tirofiban versus placebo on the incidence of adverse cardiac outcomes and coronary artery restenosis at 6 months.
Background. Tirofiban is a highly selective, short-acting inhibitor of fibrinogen binding to platelet glycoprotein IIb/IIIa. In a recent clinical study, tirofiban reduced the incidence of adverse cardiovascular events at both 2 and 7 days after coronary angioplasty or directional coronary atherectomy. This reduction persisted but was no longer statistically significant at 30 days.
Methods. The Randomized Efficacy Study of Tirofiban for Outcomes and Restenosis (RESTORE) trial was a randomized, double-blind, placebo-controlled trial of tirofiban in patients undergoing balloon angioplasty or directional atherectomy within 72 h of presentation with either unstable angina pectoris or acute myocardial infarction. All patients received an initial bolus (10 μg/kg body weight over 3 min), followed by a 36-h infusion (0.15 μg/kg per min) of either tirofiban or placebo.
Results. At 6 months the composite end point (either death from any cause, new myocardial infarction, bypass surgery for angioplasty failure or recurrent ischemia, repeat target vessel angioplasty or stent insertion for actual or threatened abrupt closure) occurred in 1,070 placebo group patients (27.1%) and 1,071 tirofiban group patients (24.1%, p = 0.11). Analysis of 6-month coronary arteriograms by means of quantitative coronary arteriography showed no significant difference between placebo- and tirofiban-treated patients in either the incidence of a ≥50% diameter stenosis (57% vs. 51%, p = NS), a loss of ≥50% of lumen diameter gained (50% vs. 50%, p = NS) or a loss of ≥0.72 mm of lumen diameter (44% vs. 42%, p = NS).
Conclusions. The 3% absolute reduction in the incidence of the composite end point at 6 months (27.1% placebo vs. 24.1% tirofiban) was similar to that previously reported at 2 days (8.7% vs. 5.4%, p < 0.005), and there does not appear to be any late effect of tirofiban on clinical end points between day 2 and 6 months. Tirofiban did not reduce the incidence of restenosis at 6 months when defined in a number of ways.
Percutaneous coronary interventions expose many of the highly thrombogenic contents of the arterial wall, and coronary thrombosis remains a persistent problem after conventional balloon angioplasty and directional coronary atherectomy (DCA). Despite standard therapy with aspirin and heparin (1,2), thrombotic occlusion after interventional procedures still occurs in 4% to 12.8% of patients (3–5)and has prompted the search for more effective antiplatelet agents (5–14).
In high risk patients undergoing percutaneous transluminal coronary angioplasty (PTCA), the monoclonal antibody abciximab, directed against the platelet glycoprotein (GP) integrin receptor IIb/IIIa, has been shown to significantly reduce the composite incidence of death, myocardial infarction, emergency repeat angioplasty, emergency coronary artery bypass graft surgery (CABG) or stent implantation by 35% at both 2 and 30 days (5,15–17). Tirofiban (Aggrastat, Merck & Co.) is a synthetic, short-acting, highly selective nonpeptide inhibitor of fibrinogen binding to platelet GP IIb/IIIa (18–22). Potential advantages of this drug include a rapid onset of action, rapid reversal of antiplatelet activity after drug discontinuation, suitability for multiple repeat administrations and high specificity for the GP IIb/IIIa receptor.
The Randomized Efficacy Study of Tirofiban for Outcomes and REstenosis (RESTORE) trial (23)was a randomized, double-blind, placebo-controlled trial of tirofiban in patients undergoing PTCA or DCA within 72 h of presentation with either unstable angina pectoris or acute myocardial infarction. Previously, the RESTORE study group has reported (23)that tirofiban compared with placebo reduced the incidence of the composite end point of death; myocardial infarction; any target vessel repeat PTCA or CABG for recurrent ischemia; and stent implantation for abrupt closure by 38% at 2 days (p < 0.005) and by 27% at 7 days (p = 0.022) largely due to a reduction in nonfatal myocardial infarction and the need for repeat angioplasty. The primary composite end point at 30 days was reduced from 12.2% in the placebo group to 10.3% in the tirofiban group, a 16% relative reduction (p = 0.16). When only repeat angioplasty and CABG performed on an urgent (i.e., emergency) basis were included as components of the 30-day composite, the 10.5% event rate in the placebo group was reduced to 8.0% in the tirofiban group, a 24% relative reduction (p = 0.052) (23). The incidence of bleeding and thrombocytopenia did not differ significantly between tirofiban- and placebo-treated patients.
The effect of tirofiban and other GP IIb/IIIa inhibitors on angiographic restenosis is unknown. The goal of the 6-month clinical and angiographic substudies was to investigate the effects of tirofiban compared with placebo on the incidence of adverse cardiac outcomes and on coronary artery restenosis as measured by quantitative angiography.
The RESTORE trial (23)was a randomized, double-blind, placebo-controlled trial of tirofiban in aspirin- and heparin-treated patients undergoing PTCA or DCA within 72 h of presentation with acute coronary syndromes (either unstable angina or an acute myocardial infarction). Unstable anginawas defined as typical anginal pain at rest or with minimal effort and either 1) electrocardiographic changes; 2) hemodynamic changes suggestive of myocardial ischemia; or 3) angiographic evidence of thrombus in the target vessel immediately before PTCA or DCA (characterized by a stenosis >70% with a hazy appearance, intraluminal filling defect, overhanging edge, high degree of eccentricity or reduced Thrombolysis in Myocardial Infarction [TIMI] flow grade). Acute myocardial infarctionwas defined as typical ischemic pain lasting >20 min with ST-T wave changes or pathologic Q waves and a serum creatine kinase elevation greater than twice the upper limit of normal or an elevated creatine kinase-MB fraction value.
Treatment and adjunctive medical therapy
Patients received 300 to 325 mg of aspirin orally within 12 h of PTCA or DCA. Guidelines for heparin administration during PTCA were a maximal initial bolus of 10,000 U before the procedure (weight adjusted to 150 U/kg for patients <70 kg), and intraprocedural heparin was administered as required to maintain an activated clotting time of 300 to 400 s. After the lesion was crossed with the guide wire, the patient was randomized to receive either a bolus of tirofiban (10 μg/kg body weight) or placebo intravenously over 3 min. Each patient then received an intravenous infusion of tirofiban (0.15 μg/kg per min) or placebo for 36 h. Operators were urged to place intracoronary stents only in urgent “bailout” situations (e.g., actual or threatened abrupt closure). The decision to perform DCA or PTCA was left to the discretion of the operator. Investigators were advised to discontinue heparin administration at the conclusion of the PTCA or DCA procedure and to remove sheaths when the activated clotting time was <180 s.
Coronary arteriographic substudy
Selected study sites enrolled all consecutive patients in the angiographic substudy until a total of ∼500 patients had been enrolled. At the beginning of the PTCA or DCA procedure, nitroglycerin (200 μg intracoronary) was administered into each coronary artery (left and right coronary arteries) before repeat angiography. At the completion of the procedure, nitroglycerin (200 μg intracoronary) was readministered into each coronary artery (left and right coronary arteries) before repeat angiography. For the 6-month follow-up angiogram, the catheterization laboratory, the equipment used, the procedure followed (including angles and magnifications used) and the therapy given were identical.
For the baseline preprocedural and postprocedural angiograms the angiographically “optimal” view (i.e., without vessel overlapping or foreshortening) and the view orthogonal to it were recorded on cine film at 4- to 5-in. magnification. When two good views were not available, a single view was used that showed the stenosis in its greatest severity without foreshortening, motion blur or overlapping of branches. A portion of the non–contrast-filled catheter shaft was visible during some phase of the injection to allow for calibration and was centered in the image field whenever possible to minimize the impact of pincushion distortion.
Follow-up angiography was performed as close to 6 months after the index procedure as possible; however, angiography performed between 17 and 30 weeks after the initial PTCA or DCA was accepted. If repeat cardiac catheterization was necessary before the end of 16 weeks, and there was evidence of restenosis, this early angiogram was used as the follow-up angiogram and a repeat 6-month follow-up angiogram was not necessary. However, if there was no evidence of restenosis, angiography was repeated between 17 and 30 weeks. All follow-up angiography was completed before the end of study week 30. Patients who underwent intracoronary stent placement at the time of the initial procedure were not required to return for coronary arteriography.
All physicians and technicians in the angiographic core laboratory were blinded to treatment group assignment, the investigative center’s interpretation of the angiogram and the clinical outcome of the patient. A previously described and validated automated edge detection algorithm was utilized for quantitative angiographic analysis (24). Restenosisin the target culprit lesion was prospectively defined as follows: 1) ≥50% diameter stenosis at the time of follow-up angiography in those patients who had a <50% stenosis after the initial intervention; 2) late loss in minimal lumen diameter ≥0.72 mm (25); 3) late loss ≥50% of the initial gain in minimal lumen diameter. Flowbefore and after PTCA was assessed according to both the conventional TIMI flow grade classification scheme (26)and the corrected TIMI frame count (27). Thrombus gradewas assessed using the standard TIMI definitions (28).
Clinical end points
The clinical end points of the study were death from any cause; new myocardial infarction, CABG for angioplasty failure or recurrent ischemia; repeat target vessel revascularization for recurrent ischemia; implantation of an intracoronary stent because of actual or threatened abrupt closure of the target vessel; and a composite end point, which was the occurrence of any of these events (23). End points were evaluated at 2, 7 and 30 days, and at 6 months. The prespecified primary hypothesis of the study was that tirofiban would result in a reduction in the 30-day composite end point compared with placebo (23). All end points were adjudicated by an independent, blinded end point classification committee according to previously reported definitions (23).
The statistical significance of the differences between treatment groups with respect to the composite end point and its components was assessed using logistic regression analysis (PROC LOGISTIC, SAS) (29). The dependent variable was an indicator for whether the patient experienced the specific end point. The analysis was based on the number of patients, not events; any patient experiencing one or more of the composite events within the specific time periods was counted as having a primary event. The independent variables included an indicator of treatment group and the following two covariates: inclusion criteria (unstable angina, acute myocardial infarction within 3 days, acute myocardial infarction treated with primary PTCA) and the primary procedure (DCA or PTCA). The incidence of restenosis was assessed using chi-square analysis. All tests were two-sided, and statistical significance was declared if p ≤ 0.05. Cumulative event rates over time were plotted using Kaplan-Meier curves.
The trial included 2,212 randomized patients. The study drug (tirofiban or placebo) was not administered in 71 patients (either because the angioplasty procedure was not performed or the indication for angioplasty changed); therefore, a total of 2,141 patients received study drug infusion and were included in the efficacy and safety analyses as prespecified in the protocol (23). Of these patients, 619 (315 for the placebo group, 314 for the tirofiban group) were enrolled in the 6-month angiographic restenosis substudy. Paired serial angiograms were available for 416 (67.2%) of these 619 patients (205 for the placebo group, 211 for the tirofiban group). Among tirofiban-treated patients, 1,057 (99%) of 1,071 were available for 6 month follow-up, and among placebo patients, 1,050 (98%) of 1,070 were available.
The baseline characteristics of all patients included in the RESTORE trial have been previously reported (23)and were similar between patients treated with tirofiban or placebo. For patients in the 6-month angiographic substudy, baseline clinical characteristics again did not differ between the placebo and tirofiban groups (Table 1)and were similar to those of patients in the overall trial. Fifty-nine percent of substudy patients in the placebo group and 50.2% in the tirofiban group had one diseased vessel, 25.9% and 32.7% had two diseased vessels, and 11.2% and 15.2% had three diseased vessels, respectively (p = NS for treatment group difference). Unstable angina pectoris was the most common inclusion criteria in the placebo (67.3%) and tirofiban (66.8%) groups. The intervention was performed during acute myocardial infarction (i.e., as primary angioplasty) in 2.0% of the placebo group and 2.4% of the tirofiban group. The intervention was performed as a nonprimary procedure within 3 days of acute myocardial infarction in 30.7% of the placebo group and 30.8% of the tirofiban group. The initial procedure performed was most frequently conventional PTCA (92.2% in the placebo group, 91.9% in the tirofiban group), whereas DCA was performed in the remainder of patients.
Clinical end points
The event rate for the occurrence of the composite end point from the time of randomization through 6 months for all patients is shown in Table 2. At 6 months, the composite end point occurred in 27.1% of placebo-treated patients and 24.1% of tirofiban-treated patients (p = 0.11). These same trends were observed in patients treated both with conventional PTCA (267 [26.8%] of 997 vs. 237 [24.1%] of 985) and those treated with DCA (23 [31.5%] of 73 vs. 21 [24.4%] of 86). The event rates for the individual components of the composite end point are also shown in Table 2. Except for death, all differences in the incidence of end point components were in the same direction, favoring tirofiban, but none was statistically significant. There was a 7.6% rate of acute myocardial infarction in the placebo group versus a 6.3% rate of acute myocardial infarction in the tirofiban group (p = 0.22) and a 1.4% mortality rate in the placebo group versus a 1.8% mortality rate in the tirofiban group (p = 0.49). Tirofiban did not significantly reduce the risk of repeat angioplasty (15.7% for tirofiban vs. 17.1% for placebo, p = 0.38) or CABG procedures (5.5% for tirofiban vs. 6.8% for placebo, p = 0.20).
All baseline (before and after intervention) angiographic lumen dimensions were similar in the two groups (Table 3). The reference diameter for arteries was 2.70 ± 0.57 mm for the placebo group and 2.77 ± 0.74 mm for the tirofiban group. Initial minimal lumen diameter was 0.54 ± 0.33 mm for the placebo group and 0.57 ± 0.39 mm for the tirofiban group. The postprocedural minimum lumen diameter was 1.89 ± 0.48 mm for the placebo group and 1.89 ± 0.56 mm for the tirofiban group.
There was no difference between the placebo and tirofiban groups in either the preinterventional CTFC (47.8 ± 33.0 vs 52.0 ± 34.5, respectively, p = NS) or the TIMI flow grade distribution (41% TIMI grade 3 flow and 19% TIMI grade 0 or 1 flow for placebo vs. 37% TIMI grade 3 flow and 19% TIMI grade 0 or 1 flow for tirofiban, p = NS for both comparisons) (Table 4). After PTCA or DCA there was no significant difference in the CTFC between the two groups (18.2 ± 11.3 for placebo vs. 20.2 ± 12.5 for tirofiban), but there was a difference in the incidence of TIMI grade 3 flow (94% for placebo vs. 85% for tirofiban, p < 0.001). This difference can be accounted for by the greater incidence of TIMI grade 2 fast flow (minimally delayed flow) in the tirofiban group (4% for placebo vs. 14% for tirofiban, p < 0.001) such that when TIMI grade 2 fast flow was combined with TIMI grade 3 flow, there was no significant difference between the two groups (98% for placebo vs. 99% for tirofiban, p = NS). There were no significant differences between groups in TIMI grade 0 or 1 flow (0.5% for placebo vs. 0.0% for tirofiban, p = NS) or TIMI grade 2 slow flow (markedly delayed flow) (2% vs. 1%, respectively, p = NS). Moreover, there was no difference in the change in CTFC from before to after angioplasty between the two groups (28.1 ± 32.0 for placebo vs. 30.1 ± 34.4 for tirofiban, p = NS).
There was no difference between the tirofiban and placebo groups when 6-month restenosis was defined in a number of ways (Table 5). A ≥50% loss of lumen diameter gained after the initial procedure occurred in 50% of both the placebo and tirofiban groups (p = 0.99). A ≥50% stenosis at the time of follow-up angiography was present in 57% (n = 193) of the arteries in the placebo group and in 51% (n = 196, p = 0.26) of those in the tirofiban group (only arteries with <50% stenosis at the initial postprocedural angiographic analysis were included in this analysis). A minimal lumen diameter loss ≥0.72 mm was present in 44% of arteries in the placebo group and in 42% of those in the tirofiban group (p = 0.69). The late loss for tirofiban-treated patients did not differ from that of placebo-treated patients (0.67 ± 0.72 mm vs. 0.70 ± 0.73 mm, p = NS), and likewise, the loss index (late loss divided by acute gain) for tirofiban-treated patients did not differ from that of placebo-treated patients (0.50 ± 0.57 vs. 0.51 ± 0.58, p = NS). Figure 1shows virtually superimposable cumulative distribution functions for the minimum lumen diameters before and after PTCA and at follow-up. Patients with angiographic thrombus (TIMI flow grades 2 to 4) (28)were not found to have a higher risk of restenosis than patients without angiographically apparent thrombus (54% vs. 56%, respectively, p = NS).
Acute coronary artery occlusion and delayed restenosis remain major limitations of percutaneous coronary revascularization. Recently, the RESTORE study group reported (23)that a 36-h infusion of tirofiban reduced the relative incidence of the prespecified composite end point of adverse cardiac outcomes by 38% (3.3% absolute reduction) at 2 days (p < 0.005), and by 27% at 7 days (2.8% absolute reduction) (p = 0.02). However, by 30 days the relative risk reduction decreased to 16% (1.9% absolute reduction) and was not statistically significant (23). Although the 3% absolute reduction in the incidence of the composite end point at 6 months (27.1% placebo vs. 24.1% tirofiban) was similar to that previously reported at 2 days (8.7% vs. 5.4%, p < 0.005), this 11% relative reduction in the composite end point did not reach statistical significance (p = 0.11). These benefits were obtained without a significant increase in thrombocytopenia or major bleeding complications in the tirofiban group versus the placebo-treated group (23).
It could be hypothesized that residual thrombus at the completion of an intervention may form a nidus for cellular organization and may contribute in part to the risk of restenosis. However, in the present study patients with angiographically apparent thrombus were not found to have a higher risk of restenosis than patients without thrombus (54% vs. 56%). There was also no difference in clinical restenosis between the tirofiban and placebo groups. The 3% absolute reduction in the composite end point was achieved early and was already apparent at 2 days. The cumulative event curves remained relatively parallel beyond this time point out to 6 months (Fig. 2). Moreover, in the present, and to our knowledge the first, angiographic substudy to be reported for a GP IIb/IIIa inhibitor, we found that tirofiban did not reduce the incidence of angiographic restenosis when defined in a number of ways. There was no significant difference in the number of arteries with ≥50% stenosis, ≥50% loss of lumen diameter gained or ≥0.72-mm loss of lumen diameter between the two groups. Likewise, there was no difference in either the late loss or the loss index between the tirofiban and placebo groups. Indeed, Figure 2shows virtually superimposable cumulative distribution functions for the minimum lumen diameters before and after PTCA and at follow-up.
Previous reports have suggested (15)that abciximab, an antibody with a longer duration of action and other potential non GP IIb/IIIa receptor blocking activities, may reduce the incidence of clinical restenosis. The Evaluation of IIb/IIIa Platelet Receptor Antagonist 7E3 in Preventing Ischemic Complications (EPIC) trial (15)of abciximab in high risk patients undergoing angioplasty reported a 26% decrease in repeat target vessel revascularization at 6 months. An angiographic substudy was not performed in that trial to assess restenosis rates. This effect of abciximab was not reproduced in two subsequent trials of angioplasty in patients with refractory unstable angina (Chimeric c7E3 Antiplatelet Therapy in Unstable Angina Refractory to Standard Treatment [CAPTURE] trial ) at lower risk (Evaluation of PTCA to Improve Long-Term Outcome by c7E3 GPIIb/IIIa Receptor Blockade [EPILOG] trial ). In the CAPTURE trial, when abciximab was infused predominantly before PTCA, there was no difference in death, myocardial infarction or repeat revascularization at 6 months between abciximab- and placebo-treated patients. In the EPILOG trial, a 3% absolute difference was evident in this composite end point between abciximab- and placebo-treated patients, similar to that observed in the RESTORE trial. There was no difference in the rates of repeat revascularization. No angiographic follow-up data have yet been reported from the EPILOG trial. In the only other large-scale trial of GPIIb/IIIa inhibitor in PTCA, the Integrelin to Manage Platelet Aggregation to Combat Thrombosis (IMPACT) II trial of Integrilin (eptifibatide) (32), the absolute reduction in composite end point seen at 30 days (2.2%, low dose eptifibatide vs. placebo) persisted at 6 months but was not statistically significant.
The main limitation of this study was the relatively low rate of follow-up angiography (67%) in the angiographic substudy. That the substudy cohort was enriched with symptomatic patients may account, in part, for the relatively high rates of restenosis that were observed. Nonetheless, the angiographic substudy cohort was representative of the overall study group, and there was no difference in follow-up rates between the two treatment groups. Thus, the low rate of angiographic follow-up is unlikely to have been a source of treatment bias.
Six-month follow-up of patients undergoing high risk PTCA or DCA in the RESTORE trial demonstrated no significant benefit of tirofiban in reducing the composite clinical end point of death, myocardial infarction, CABG or repeat PTCA of the target vessel. The 3% absolute reduction in this composite end point that was observed within 48 h of initial treatment persisted over the 6-month follow-up period, but there was no evidence of additional late clinical benefit. Angiographic follow-up of a subset of patients demonstrated that tirofiban did not reduce angiographic restenosis.
↵# A list of the principal investigators of the Randomized Efficacy Study of Tirofiban for Outcomes and Restenosis (RESTORE) trial appears in reference .
☆ This study was supported in part by a grant from Merck & Co., Inc., Whitehouse Station, New Jersey.
- coronary artery bypass graft surgery
- Chimeric c7E3 Antiplatelet Therapy in Unstable Angina Refractory to Standard Treatment (trial)
- corrected Thrombolysis in Myocardial Infarction (TIMI) frame count
- directional coronary atherectomy
- Evaluation of PTCA to Improve Long-Term Outcome by c7E3 GPIIb/IIIa Receptor Blockade
- percutaneous transluminal coronary angioplasty
- Randomized Efficacy Study of Tirofiban for Outcomes and Restenosis
- Received December 11, 1997.
- Revision received March 3, 1998.
- Accepted March 16, 1998.
- American College of Cardiology
- Harrington R.A.
- De Feyter P.J.,
- van den Brand M.,
- Jaarman G.,
- van Domburg R.,
- Serruys P.W.,
- Suryapranata H.
- Lincoff A.M.,
- Popma J.J.,
- Ellis S.G.,
- Hacker J.A.,
- Topol E.J.
- Théroux P.,
- Kouz S.,
- Roy L.,
- et al.
- Willerson J.T.
- Coller B.S.
- 15.↵Topol EJ, Califf RM, Weisman HF, et al., on behalf of the EPIC Investigators. Randomized trial of coronary intervention with antibody against platelet IIb/IIIa integrin for reduction of clinical restenosis: results at six months. Lancet 1994;343:881–6.
- Tcheng J.G.
- Gates E.R.,
- McGillem M.J.,
- Mickelson J.K.,
- et al.
- Hartman G.D.
- Peerlinck K.,
- DeLepeleire I.,
- Goldberg M.,
- et al.
- Lynch J.J.,
- Cook J.J.,
- Sitko G.R.,
- et al.
- Kereiakes M.,
- Kleiman S.,
- Ambrose A.,
- et al.
- The RESTORE Investigators
- Reiber J.H.C.,
- Surreys P.W.,
- Barth J.D.
- Gibson C.M.,
- Cannon C.P.,
- Daley W.L.,
- et al.
- The TIMI IIIA Investigators
- SAS Institute Inc