Author + information
- Received December 4, 2002
- Revision received March 11, 2003
- Accepted March 13, 2003
- Published online September 17, 2003.
- Mitchell J Ross, MD*,
- Howard C Herrmann, MD, FACC*,* (, )
- David J Moliterno, MD, FACC†,
- James C Blankenship, MD, FACC‡,
- Laura Demopoulos, MD§,
- Peter M DiBattiste, MD, FACC§,
- Stephen G Ellis, MD, FACC†,
- Ziyad Ghazzal, MD, FACC∥,
- Jack L Martin, MD, FACC¶,
- Jennifer White† and
- Eric J Topol, MD, FACC†
- ↵*Reprint requests and correspondence:
Dr. Howard C. Herrmann, Hospital of the University of Pennsylvania, 3400 Spruce Street, Philadelphia, Pennsylvania 19104, USA.
Objectives We sought to assess whether pre-procedural angiographic characteristics are associated with adverse clinical outcomes after coronary stenting with glycoprotein IIb/IIIa inhibition.
Background Ischemic complications after balloon angioplasty are associated with pre- and post-procedural angiographic variables. However, in the current era of stenting with IIb/IIIa inhibition, it is unknown whether angiographic features assessed before intervention confer an increased risk of adverse procedural and subsequent clinical outcomes.
Methods In the Do Tirofiban and ReoPro Give Similar Efficacy Outcomes? Trial (TARGET), 4,809 patients undergoing planned stenting were randomized to tirofiban or abciximab. Baseline demographic, clinical, and angiographic variables were obtained. Clinical end points were recorded at 30 days and six months. The relationship between angiographic variables and adverse clinical outcomes was assessed.
Results Patients with the combination of thrombus, lesion eccentricity, and lesion length >20 mm had a 21.4% composite incidence of death, myocardial infarction, or urgent target vessel revascularization (TVR) at 30 days, compared with 4.2% in those patients without these high-risk features (hazard ratio [HR] 3.24, p < 0.001). After adjustment, the risk was independently associated with thrombus (HR 1.40, p = 0.034), eccentricity (HR 1.67, p < 0.001), and lesion length >20 mm (HR 1.89, p < 0.001). The risk of six-month TVR was independently associated with left anterior descending coronary artery lesions (HR 1.46, p < 0.001), restenotic lesions at baseline (HR 1.58, p = 0.006), and lesion length (HR 1.19, p = 0.03).
Conclusions Patients with thrombus, eccentric lesions, or lesion length >20 mm are at high risk for ischemic outcomes after coronary stenting, despite IIb/IIIa inhibition. Further research into novel anti-thrombotic therapies or procedural strategies is necessary for these patients.
Ischemic complications after percutaneous transluminal coronary angioplasty (PTCA) are associated with certain high-risk morphologic features of the treated lesion, includ- ing lesion length, angulation, and thrombus (1–4). However, these studies examined only post-procedure angiographic variables or were performed before the advent of glycoprotein IIb/IIIa inhibitors and routine stent usage. The glycoprotein IIb/IIIa inhibitors (5–11)and thienopyridines (12–17), as well as the use of intracoronary stents (18–20), have significantly reduced the incidence of adverse clinical events after percutaneous coronary intervention (PCI). In this current era, little is known about the value of angiographic features, assessed before intervention, in predicting post-procedural clinical outcomes. For this reason, we examined the recent TARGET (Do Tirofiban and ReoPro Give Similar Efficacy Outcomes?) Trial study population of 4,809 patients undergoing planned stenting with either intravenous tirofiban or abciximab, to determine whether pre-procedural angiographic characteristics are associated with an increased risk of adverse ischemic events.
The rationale and design of the TARGET trial have been previously described (21,22). In brief, patients undergoing PCI with planned stenting were considered for enrollment. Patients with acute myocardial infarction (MI), cardiogenic shock, unprotected left main coronary artery lesions, decompensated heart failure, renal insufficiency, thrombocytopenia, or bleeding disorders were excluded. The Human Investigational Review Board at each site approved the study, and all patients provided written informed consent.
Patients were randomized in a double-blind, double-dummy design to tirofiban (10 μg/kg intravenous bolus, followed by 0.15 μg/kg/min infusion for 18 to 24 h) or abciximab (0.25 mg/kg bolus, followed by 0.125 μg/kg/min infusion, maximum 10 μg/min, for 12 h). Randomization was stratified according to the presence or absence of diabetes mellitus. All patients received 250 to 500 mg of aspirin within 24 h before the procedure and 75 to 325 mg orally once daily for the duration of the study. Clopidogrel was recommended for all patients (300 mg oral loading dose within 2 to 6 h of the procedure, followed by 75 mg daily for 30 days). Unfractionated heparin was administered according to a predefined nomogram at the beginning of the procedure to a target activated clotting time of 250 s. Serum creatine kinase (CK) and CK-MB levels were measured every 6 h post procedure for 24 h. Clinical follow-up was performed at 30 days and six months, and mortality was assessed at one year.
Definitions, end points, and statistics
Baseline clinical and demographic variables were recorded at the time of enrollment. In addition, angiographic characteristics were recorded at the time of the procedure, based on the individual operator's assessment of lesion morphology. The primary end point of the study was the composite of death, MI, or urgent target vessel revascularization (TVR) within 30 days. A new peri-procedural MI (within 48 h post procedure) was defined as an elevation of the CK-MB to three times the upper limit of normal in two separate blood samples or by the presence of new Q waves in two or more contiguous electrocardiographic leads. Secondary clinical end points included each component of the primary end point; six-month rates of death, MI, and any TVR; and hematologic and hemorrhagic complications. The sample size was selected to test the hypothesis that tirofiban would not be inferior to abciximab.
Univariable analyses were performed to examine baseline and angiographic characteristics within treatment groups. Analyses were also performed on baseline characteristics by angiographic findings and on angiographic findings by acute coronary syndrome (ACS). Pearson chi-square tests were used for categorical variables, and the Wilcoxon test was used for continuous variables. For outcome variables, univariable analyses were performed using Kaplan-Meier product limit estimation. Survival was stratified by angiographic factors with log-rank testing. New variables were formed for patients with none, any one, any two, or all three of the identified angiographic risk factors. Figures are provided that demonstrate the estimated cumulative hazards for 30-day and six-month end points stratified by angiographic findings.
Cox proportional hazard modeling was used to describe 30-day and six-month end points. Variables considered for analysis included baseline, clinical, procedural, and angiographic factors. The approach to modeling was to first adjust for baseline and clinical factors. Once these adjustments were made, significant angiographic factors were determined. Stepwise selection techniques were used to identify significant factors. Significance was determined at p < 0.05. Model validation was performed via bootstrapping. All analyses were intention to treat.
Baseline angiographic characteristics
A total of 4,809 patients were randomized to tirofiban or abciximab between December 30, 1999, and August 25, 2000. The baseline clinical, demographic, and angiographic characteristics of these patients are provided in Table 1. There were no significant angiographic differences between the two treatment arms of the trial. Intracoronary thrombus was noted in 9.7% of patients, and two-thirds of patients were noted to have eccentric lesions. The majority of lesions were in native vessels, but a small number of interventions (6.1%) were performed on bypass grafts. Most lesions (59%) were between 10 and 20 mm in length, and 98% of patients demonstrated TIMI 3 flow after stenting.
Patients with intracoronary thrombus were significantly more likely than those without thrombus to be current smokers and to have an ACS as the indication for PCI (Table 2). In addition, patients with intracoronary thrombus were nearly three times as likely to have a PCI performed in a bypass graft and had lower rates of TIMI 3 flow both before and after the procedure. Patients with thrombus were less likely than those without thrombus to have restenotic lesions. Finally, patients with thrombus were more likely than those without thrombus to have lesions longer than 20 mm (23.7% vs. 15.1%, p < 0.001).
Compared with patients whose lesions were concentric, patients with eccentric lesions were more likely to be current smokers (22.9% vs. 19.9%, p = 0.02) and were more likely to have lesions longer than 20 mm (17.5% vs. 12.8%, p < 0.001). There were no other significant demographic or clinical differences between these groups.
Patients with lesions >20 mm, when compared with patients whose maximal lesion length was ≤20, were more likely to be male (76.6% vs. 72.9%, p = 0.03), and were less likely to have TIMI 3 flow both before (69.8% vs. 83.3%, p < 0.001) and after the procedure (95.4% vs. 98.5%, p < 0.001). There were no other significant differences between these two groups.
Relationship between clinical presentation and angiographic findings
Patients with an ACS indication for PCI (unstable angina pectoris, non–Q-wave MI or recent Q-wave MI) were two times more likely to have intracoronary thrombus than those presenting with non-ACS (stable angina or abnormal noninvasive testing) indications (12.0% vs. 5.9%, p < 0.001). Those patients with recent Q-wave MI had the highest incidence of thrombus (28.6%). There were no differences in the incidence of lesion eccentricity or long lesions (>20 mm) between ACS patients and non-ACS patients (Table 3).
Univariable angiographic predictors of outcome
The presence of thrombus, lesion eccentricity, or lesions >20 mm in length was found to be strongly associated with adverse clinical events at 30 days and six months (Table 4, Fig. 1). Furthermore, these three angiographic findings had an additive association with the primary end point. In patients with all three high-risk characteristics, the composite incidence of death/MI/urgent TVR at 30 days was 21.4%, compared with 4.2% in those without any of the three characteristics (Log rank chi-square p value <0.001) (Fig. 2). In keeping with the primary results of the TARGET trial, patients within individual angiographic subgroups tended toward lower incidences of the 30-day primary composite end point when treated with abciximab versus tirofiban. However, among the 86 patients possessing all three high-risk angiographic findings, the incidence of the primary 30-day composite end point was increased substantially in both groups to a similar level (20.5% in patients treated with tirofiban vs. 21.3% in patients treated with abciximab, p = 0.9).
Angiographic variables were also associated with the secondary six-month composite end point of death, MI, or any TVR. In patients with all three high-risk characteristics, the composite six-month end point was 29.7%, compared with 12.9% of those patients possessing no high-risk characteristics (Fig. 3).
Cox proportional hazard modeling was performed on all angiographic variables, in conjunction with clinical and demographic variables previously shown to be independent correlates of adverse outcome in this trial (15). High-risk angiographic variables identified as independent correlates of adverse outcome at 30 days included intracoronary thrombus, lesion eccentricity, and lesions >20 mm in length (Table 5). The presence of multiple high-risk variables was associated with an additive risk, with a hazard ratio (HR) of 3.24 for the 30-day primary composite end point in the presence of all three variables. Hazard ratios associated with these angiographic variables were similar for the secondary combined end point of death or MI at 30 days.
Multivariable analysis of the six-month composite end point of death, MI, or any TVR identified five independent correlates (p < 0.03): 1) country of procedure (U.S. vs. non-U.S., HR 1.65); 2) ACS (HR 1.33); 3) PCI on left anterior descending (LAD) coronary artery (HR 1.41); 4) presence of thrombus (HR 1.33); and 5) maximum lesion length >20 mm (HR 1.42). Age, lesion eccentricity, prior clopidogrel use, and—notably—treatment arm were not independently associated with the six-month composite end point in this model. Multivariable analysis of the six month end point of any TVR identified three independent angiographic correlates (p < 0.05): 1) PCI on LAD (HR 1.46); 2) maximum lesion length (HR 1.19); 3) and restenotic lesion (HR 1.58).
In this analysis of a large, prospective multicenter trial of stenting with IIb/IIIa inhibition, we have identified pre-procedural angiographic characteristics associated with a higher risk of adverse clinical outcomes. This effect is independent of clinical and demographic variables known to be associated with ischemic events in this population, including advanced age, ACS, and thienopyridine use.
Previous studies have identified angiographic predictors of adverse outcome, but they either examined only post-procedural angiographic variables or were performed before the advent of IIb/IIIa inhibition and routine use of coronary stents (1–4). The glycoprotein IIb/IIIa inhibitors have been shown to significantly reduce the incidence of ischemic complications after PCI in seven large, randomized, placebo-controlled trials (5–11). In a combined angiographic analysis of the EPIC (Evaluation of 7E3 for the Prevention of Ischemic Complications) and EPILOG (Evaluation in PTCA to Improve Long-Term Outcome with Abciximab GP IIb/IIIa Blockade) trials (23), 4,154 patients undergoing angioplasty or atherectomy were treated with abciximab or placebo in addition to aspirin and heparin. Abciximab was shown to reduce the 30-day composite incidence of death/MI/TVR from 12.1% to 6.1%. This risk reduction was seen across all lesion morphologies, with the exception of lesions in degenerated saphenous vein grafts. These studies, however, differed from the TARGET trial, in that coronary stenting was used only for patients with suboptimal results or for intraprocedural complications. Stenting was performed in 0.9% and 14% of the EPIC and EPILOG patients, respectively (5,6).
In trials of coronary stenting, angiographic characteristics have been shown to affect clinical outcomes. In a German study of 2,894 consecutive procedures, the presence of angiographic features (including long lesions, high-grade stenosis, vein graft lesions, or acute thrombosis) was shown to be associated with increased risk of procedural failure in percutaneous stent implantation (24). In the EPISTENT (Evaluation of Platelet IIb/IIIa Inhibition in Stenting) trial, the presence of thrombus, major dissection, or residual stenosis >10% afterstent implantation was associated with an increased incidence of adverse 30-day events, both with and without concomitant IIb/IIIa inhibition (25). In the ESPRIT (Enhanced Suppression of the Platelet IIb/IIIa Receptor with Integrilin Therapy) trial, patients with angiographic complications after stenting with IIb/IIIa inhibition had a fivefold increase in a primary composite clinical end point at 48 h (26). Similarly, in the IMPACT-II (Integrilin to Minimize Platelet Aggregation and Coronary Thrombosis-II) trial, patients with angiographic complications after stenting and IIb/IIIa inhibition had an increase in the combined incidence of death/MI/urgent TVR at 24 h (14.8% vs. 3.9%, p < 0.001) (27). However, none of these trials examined whether high-risk angiographic features assessed before intervention confer post-procedural risk in patients undergoing stent placement with IIb/IIIa inhibition.
Our analysis demonstrates a striking association between thrombus, lesion length, eccentricity, and adverse ischemic outcomes. The combination of all three features was associated with a 21.4% event rate at 30 days, more than three times the event rate seen in the TARGET trial as a whole. Although the rates of adverse outcomes were driven primarily by increased rates of post-procedure biomarker release, the other components of the 30-day and six-month composite end points were affected to a similar degree. Despite the TARGET trial results demonstrating a reduction in MI at 30 days in ACS patients treated with abciximab (vs. tirofiban) (15), we did not note outcome differences based on the drug used in those patients possessing high-risk angiographic characteristics. We did, however, see a general trend toward abciximab superiority across angiographic subgroups, but the trial may not have had adequate power to detect significant treatment differences in these smaller subgroups. In this regard, it is possible that suboptimal platelet inhibition in the tirofiban arm of the trial contributed to the adverse outcomes observed in patients with high-risk anatomy. A recent randomized study of platelet inhibition after administration of abciximab and tirofiban at TARGET doses to patients with ACS suggested that the bolus of tirofiban used in this trial may have resulted in significantly less early platelet inhibition than that achieved in the patients treated with abciximab (28). Whether improved dosing regimens of tirofiban can lead to a reduction in clinical events in high-risk populations will require further study (29). Newer anti-thrombotic therapies, including the low-molecular-weight heparins and direct thrombin inhibitors, may be necessary to further reduce complications in high-risk patients (30,31).
Angiographic follow-up was not performed in the TARGET trial; therefore, in-stent restenosis cannot be directly assessed. However, the six-month TVR rate (8.3%) may represent a surrogate for this end point. This rate is similar to that seen in patients treated with stenting and abciximab in the EPISTENT trial (8.7%) (32). We identified three angiographic characteristics that were independently associated with increased rates of six-month TVR (LAD intervention, increased lesion length, and the stenting of restenotic lesions at baseline). Prior restenosis and increased length of vessel stenting have been previously identified as predictors of in-stent restenosis (33), and LAD intervention is associated with restenosis after PTCA (34). Although lesions in saphenous vein grafts have been shown previously to be associated with increased risk of in-stent restenosis (33), we did not observe a significant increase in six-month TVR in patients undergoing stenting of saphenous vein grafts.
In the TARGET trial, clopidogrel pre-treatment 2 to 6 h before the procedure was recommended for all patients. In the CREDO (Clopidogrel for Reduction of Events During Observation) trial, patients pre-treated with clopidogrel (vs. placebo) before elective stenting had a nonsignificant 18% reduction in the 28-day composite end point of death, MI, or urgent TVR. However, the subset of patients pre-treated >6 h before PCI had a 39% reduction in the incidence of this end point (5.8% vs. 9.4%, p = 0.05) (35). Although most of our patients were not elective, it remains possible that earlier pre-treatment with clopidogrel could have reduced adverse events.
Our analysis and findings may be limited by the fact that this was a retrospective analysis. In addition, angiographic assessment was performed at the time of PCI by individual operators and not by a core angiographic laboratory. It is possible that the assessment of lesion characteristics had an impact on procedural technique, or that the immediate results of the procedure influenced the operator's assessment of lesion characteristics. Finally, the small number of events precludes sub-analysis of the angiographic predictors within clinical subsets.
Despite these limitations, the TARGET study provides a unique opportunity to study the relationship between pre-procedural angiographic findings and outcomes after coronary stenting with platelet IIb/IIIa inhibition. In this large prospective study, we have identified a population of patients that are at high risk for adverse ischemic events after a contemporary invasive strategy incorporating stents, IIb/IIIa inhibition, and thienopyridine use. These patients may represent a useful study population for future trials designed to assess the role of new anti-thrombotic therapies and procedural strategies.
☆ The TARGET trial was funded by Merck and Co., Inc. Dr. Herrmann has received honoraria and research funding from and is a consultant for Merck. Dr. Moliterno has received research funding from and is a consultant for Merck. Drs. Demopoulos and DiBattiste are employees of Merck and may potentially own stock and/or hold stock options in the company. Dr. Ellis has received research funding from Merck.
- acute coronary syndrome
- creatine kinase
- hazard ratio
- left anterior descending coronary artery
- myocardial infarction
- percutaneous coronary intervention
- percutaneous transluminal coronary angioplasty
- Do Tirofiban and ReoPro Give Similar Efficacy Outcomes? trial
- target vessel revascularization
- Received December 4, 2002.
- Revision received March 11, 2003.
- Accepted March 13, 2003.
- American College of Cardiology Foundation
- Ellis S.G.,
- Roubin G.S.,
- King S.B.,
- et al.
- Ellis S.G.,
- Vandormael M.G.,
- Cowley M.J.,
- et al.
- Myler R.K.,
- Shaw R.E.,
- Stertzer S.H.,
- et al.
- Singh M.,
- Reeder G.S.,
- Ohman E.M.,
- et al.
- The CAPTURE Investigators. Randomised placebo-controlled trial of abciximab before and during coronary intervention in refractory unstable angina: the CAPTURE Study. Lancet 1997;349:1429–35
- 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;349:1422–8
- The RESTORE Investigators
- Steinhubl S.R.,
- Lauer M.S.,
- Mukherjee D.P.,
- et al.
- Steinhubl S.R.,
- Ellis S.G.,
- Wolski K.,
- et al.
- Stone G.W.,
- Moliterno D.J.,
- Bertrand M.,
- et al.
- Ellis S.G.,
- Lincoff A.M.,
- Miller D.,
- et al.
- Schuhlen H.,
- Kastrati A.,
- Dirschinger J.,
- et al.
- Mazur W.,
- Kaluza G.L.,
- Sapp S.,
- et al.
- Blankenship J.C.,
- Tasissa G.,
- O'Shea J.C.,
- et al.
- Herrmann H.C.,
- Swierkosz T.A.,
- Kapoor S.,
- et al.
- Lincoff M.A.,
- Kleiman N.S.,
- Kottke-Marchant K.,
- et al.
- Lowe H.C.,
- Oesterle S.N.,
- Khachigian L.M.
- M-HEART Investigators,
- Hirshfeld J.W. Jr.,
- Schwartz J.S.,
- Jugo R.,
- et al.