Frequency and Outcomes of Provisional Glycoprotein IIb/IIIa Blockade in Patients Receiving Bivalirudin during Percutaneous Coronary Intervention

ABSTRACT: Objectives. This study sought to evaluate the frequency and efficacy of combination of bivalirudin and provisional glycoprotein (GP) IIb/IIIa blockade compared with bivalirudin monotherapy in current clinical practice of percutaneous coronary intervention (PCI) with drug-eluting stents (DES). Background. Previous randomized trials have demonstrated that a strategy of bivalirudin with provisional (bailout) GP IIb/IIIa inhibition was non-inferior to unfractionated heparin (UFH) plus planned GP IIb/IIIa blockade for the prevention of acute and long-term adverse clinical events. However, the frequency and efficacy of provisional GP IIb/IIIa inhibition in addition to the full-dose bivalirudin in current practice is not well established. Methods. Using the 2004/2005 Cornell Angioplasty Registry, we studied 1,340 consecutive patients undergoing urgent or elective PCI with periprocedural use of bivalirudin. We excluded patients presenting with an acute ST-elevation myocardial infarction (MI) within ≤ 24 hours, hemodynamic instability/shock, thrombolytic therapy within ≤ 7 days, or renal insufficiency. Mean clinical follow up was 24.2 ± 7.7 months. Results. Of the study cohort, 1,184 patients (88.4%) received bivalirudin alone and 156 (11.6%) received bivalirudin plus bailout GP IIb/IIIa blockade. DES were used in 86% of PCIs. The incidence of in-hospital mortality (0% vs. 0.3%; p = 1.000), MI (7.1% vs. 6.6%; p = 0.864), and the combined endpoint of death, stroke, emergent coronary artery bypass graft surgery (CABG)/PCI, or MI (7.1% vs. 6.9%; p = 0.868) were similar in the bivalirudin-plus-bailout GP IIb/IIIa inhibitor versus the bivalirudin-alone arm. There was a higher incidence of bleeding complications (16.0% vs. 9.6%; p = 0.018) in the bivalirudin-plus-bailout GP IIb/IIIa versus the bivalirudin-alone group. At follow up, there were 4 (2.6%) deaths in the bivalirudin-plus-GP IIb/IIIa inhibitor group versus 83 (7.0%) deaths in the bivalirudin-alone arm (HR 0.36, 95% confidence interval [CI] 0.13–0.98; p = 0.044). After multivariate Cox regression analysis, bailout GP IIb/IIIa use in addition to bivalirudin was associated with similar long-term survival when compared to bivalirudin monotherapy (HR 0.41, 95% CI 0.15–1.12; p = 0.081). Conclusions. Provisional GP IIb/IIIa use in bivalirudin-treated patients is higher in contemporary non-emergent PCI practice than that seen in randomized trials and is associated with similar in-hospital ischemic events, but more frequent bleeding events. These data suggest that a strategy of bivalirudin monotherapy is preferable in order to reduce bleeding complications, and GP IIb/IIIa blockade should be reserved for patients with periprocedural complications in bivalirudin-treated patients undergoing PCI.

J INVASIVE CARDIOL 2009;21:258–263

A combination of antithrombotic and antiplatelet therapies with unfractionated heparin (UFH), aspirin, thienopyridine and glycoprotein (GP) IIb/IIIa inhibitors has been used to reduce the incidence of ischemic complications during percutaneous coronary interventions (PCI).1,2 The direct thrombin inhibitor bivalirudin (Angiomax, The Medicines Company, Parsippany, New Jersey) was approved in 2000 for periprocedural use in PCI for patients with unstable angina.3–5 Recent randomized trials have demonstrated bivalirudin’s efficacy and safety in both elective PCI and PCI for acute coronary syndromes (ACS).6-11 In the Randomized Evaluation of PCI Linking Angiomax to Reduced Clinical Events-2 (REPLACE-2) trial, bivalirudin monotherapy with GP IIb/IIIa inhibition on a provisional basis, was shown to be non-inferior to UFH plus planned GP IIb/IIIa inhibitors in preventing acute composite ischemic and hemorrhagic complications following PCI.6-9 Importantly, GP IIb/IIIa inhibitors were administered to 7.2% of patients receiving bivalirudin therapy for complications during PCI.7,10 Furthermore, in this small subgroup of patients receiving both bivalirudin and provisional GP IIb/IIIa blockade, there was a higher rate of ischemic and bleeding endpoints at 30 days when compared to patients receiving bivalirudin monotherapy.10 In the Acute Catheterization and Urgent Intervention Triage strategY (ACUITY) trial, bivalirudin plus a planned GP IIb/IIIa inhibition versus a bivalirudin-alone strategy (with rescue GP IIb/IIIa inhibition) was associated with similar composite ischemia and more bleeding events at 30 days in patients with moderate- and high-risk ACS undergoing PCI.11–14 The goal of this retrospective study was to investigate the frequency of provisional GP IIb/IIIa use in bivalirudin-treated patients and whether procedural outcomes and long-term survival after provisional use of GP IIb/IIIa inhibitors is safe and efficacious in a broad spectrum of patients undergoing elective PCI or PCI for ACS.

Methods

Data collection. The Cornell Angioplasty Registry includes all patients undergoing PCI at our institution. A standard case report form delineating comprehensive patient demographics, pre-intervention clinical status, procedural findings and in-hospital complications was completed for each PCI performed. Patients undergoing PCI between January 1, 2004 and December 31, 2005 who met the inclusion criteria were enrolled in this study. Patient follow up was obtained by means of publicly available mortality data through the Social Security Death Index15 as well as through regularly scheduled phone contacts. Baseline creatine kinase (CK), CK-MB, and Troponin I levels were obtained routinely prior to PCI and 8, 16 and 24 hours after PCI. Only the initial PCI was included in this analysis for patients undergoing multiple PCIs during the defined study period. The study was approved by the institutional review board of the Weill Medical College of Cornell University.

Procedural data. Bivalirudin was administered as a 0.75 mg/kg intravenous (IV) bolus followed by an infusion of 1.75 mg/kg/hour for the duration of the PCI procedure. Upstream administration of bivalirudin prior to PCI was not practiced during the reported study period. Provisional or “bailout” GP IIb/IIIa inhibition in addition to bivalirudin was provided during the PCI at any time for angiographic or procedural complications such as new or suspected thrombus, distal embolization, decreased thrombolysis in myocardial infarction (TIMI) flow (0–2) or slow reflow, dissection with decreased flow, or prolonged myocardial ischemia. Abciximab was administered as a 0.25 mg/kg bolus and a 0.125 µg/kg per minute (maximum, 10 µg per minute) infusion for 12 hours. Eptifibatide was given as two 180 µg/kg boluses 10 minutes apart, followed by a 2.0 µg/kg per minute infusion for 18 hours. Aspirin (325 mg) was routinely administered prior to PCI. Clopidogrel (300 or 600 mg loading dose) was administered before or immediately after the PCI, followed by 75 mg/day for at least 3–6 months. The administration of IV GP IIb/IIIa inhibitors, the choice of bare-metal stents (BMS) or drug-eluting stents (DES), and the timing of thienopyridine pretreatment were at the discretion of the physician.

Data analysis. Patients presenting with an acute ST-elevation MI ≤ 24 hours, hemodynamic instability/shock, use of thrombolytic therapy ≤ 7 days, or with renal insufficiency (serum creatinine ≥ 4 mg/dL) were excluded. We excluded patients who received UFH alone or UFH plus GP IIb/IIIa inhibition during PCI, or those who received upstream GP IIb/IIIa inhibitors within 12 hours of PCI. Patients receiving upstream low-molecular-weight heparin (LMWH) or upstream UFH and having an activated clotting time ≥ 175 seconds were excluded, as the institutional practice for those patients would be to continue receiving either LMWH or UFH ± GP IIb/IIIa inhibition during PCI. Baseline patient clinical characteristics and angiographic data were compared between patients undergoing a strategy of periprocedural bivalirudin use versus bivalirudin plus bailout GP IIb/IIIa blockade. The primary endpoints analyzed were in-hospital mortality after PCI and long-term all-cause mortality. The secondary endpoints included the incidence of in-hospital major adverse cardiac events (MACE), MI, major and minor bleeding. Long-term mortality data were obtained for 100% of patients, with a mean follow-up period of 24.2 ± 7.7 months.

Definitions. MI before PCI was defined as any elevation of CK-MB or troponin I level greater than the laboratory upper limits of normal (ULN). MI after PCI was defined as CK-MB level ≥ 3 times the upper limit of normal (ULN) within 24 hours post PCI and at least 50% increase over the preprocedural levels. Multivessel disease was defined as the presence of > 70% lesion in ≥ 2 major coronary arteries/branches or a left main coronary artery lesion of > 50%. Multivessel PCI was defined as a coronary intervention in ≥ 2 major coronary arteries/branches or in the left main coronary artery. Multilesion PCI was defined as a coronary intervention in ≥ 2 lesions of a single coronary artery or multiple coronary arteries. Congestive heart failure (CHF) referred to patients having New York Heart Association Class III or IV heart failure during admission. A major adverse cardiac event (MACE) was defined as post-PCI death, emergency cardiac surgery, emergency PCI, cerebral vascular accident or MI. Vascular injury referred to an access-site complication requiring mechanical intervention. Peripheral vascular disease included patients with carotid, aorto-femoral or lower-extremity vascular disease documented by a radiological study, history of a vascular intervention or of a cerebral vascular accident. Major bleeding was defined as a drop in hemoglobin ≥ 4 g/dl. Minor bleeding was defined as a drop in hemoglobin ≥ 2 g/dl and < 4 g/dl. Angiographic success was defined as a final stenosis of ≤ 20% of the vessel diameter with the use of any percutaneous intervention.

Statistical analysis. Data management and analysis were performed using SPSS version 12.0 (SPSS, Inc., Chicago, Illinois). Data are presented as the mean value ± standard deviation (SD) for continuous variables or as proportions for dichotomous variables. Differences in baseline characteristics and comorbid conditions between groups were compared with the chi-square test or Fisher’s exact test for dichotomous variables, and mean values for continuous variables were compared with the Student’s t-test. Stepwise multivariate logistic regression was performed to determine the independent effect of GP IIb/IIIa inhibitor use on occurrence of in-hospital events. Mortality rates were calculated and plotted according to the Kaplan Meier methods, and comparisons between the two treatment groups were performed using the log-rank statistic. The relation of GP IIb/IIIa blockade to the risk of long-term mortality was assessed with Cox proportional hazards models. To test the independence of GP IIb/IIIa inhibition as a predictor of long-term mortality, bivalirudin use was entered into a multivariable Cox proportional hazards model that also included as covariates: age, diabetes mellitus, body mass index, current CHF status, chronic obstructive pulmonary disease, prior stroke, peripheral vascular disease, prior CABG, baseline hemoglobin level, creatinine clearance, left main PCI, DES use, sirolimus-eluting stent use and bivalirudin use. For all tests, a two-tailed value of p < 0.05 was required for statistical significance.

Results

Patient population. Between January 1, 2004 and December 31, 2005, 1,340 consecutive patients underwent urgent or elective PCI with periprocedural bivalirudin use: 1,184 patients (88.4%) received bivalirudin only, and 156 patients (11.6%) received bivalirudin-plus provisional GP IIb/IIIa blockade. The baseline clinical characteristics and angiographic data are listed in Table 1. Femoral vascular access was used in > 99% of the procedures. Most patients (63%) presenting for PCI had unstable angina or non-ST-segment elevation MI (UA/NSTEMI) at study entry. Patients presenting with NSTEMI (positive CK-MB or Troponin I markers) were more likely to be treated with provisional GP IIb/IIIa inhibitors in addition to bivalirudin. Presentation with NSTEMI was independently associated with the need for provisional GP IIb/IIIa inhibition (odds ratio [OR] 1.82, 95% confidence interval [CI] 1.26–2.62; p = 0.001). In contrast, age was a negative predictor for provisional GP IIb/IIIa addition (OR 0.97, 95% CI 0.95–0.98; p < 0.001). There was a high incidence of multivessel coronary disease (54%) in both groups. Multilesion and multivessel PCI were performed in 43.3% and 13.3% of patients, respectively. BMS were used only in 8.1% of patients undergoing PCI. The use of DES was similarly high, 87% in both groups.

In-hospital outcomes. Angiographic success was similarly high in both groups (Table 2). The overall mortality for the entire study sample was 0.30%. The unadjusted incidence of in-hospital mortality (0.3% vs. 0%; OR 0.88, 95% CI 0.87–0.90; p = 1.000), MI (6.6% vs. 7.1%; OR 1.08, 95% CI 0.56–2.07; p = 0.864) and MACE (6.9% vs. 7.1%; OR 1.02, 95% CI 0.53–1.96; p = 0.868) were similar in the bivalirudin and bivalirudin-plus GP IIb/IIIa inhibitor groups. After adjustment for baseline differences with multivariate logistic modeling, provisional use of GP IIb/IIIa agents was not predictive of in-hospital mortality, MI or MACE.

There was a trend towards a lower incidence of major (0.6% vs. 1.9%, OR 3.30, 95% CI 0.84–12.88; p = 0.100) and minor bleeding events (9.0% vs. 14.1%, OR 1.65, 95% CI 1.01–2.71; p = 0.059) post PCI in the bivalirudin group in comparison with the bivalirudin-plus-GP IIb/IIIa inhibitor group. Overall, there were fewer hemorrhagic complications in the bivalirudin-only versus bivalirudin-plus-GP IIb/IIIa group (9.6% vs. 16.0%; OR 1.79, 95% CI 1.12–2.86; p = 0.018). After correcting for baseline differences with multivariate logistic regression analysis, provisional use of GP IIb/IIIa inhibition in addition to bivalirudin was an independent predictor of in-hospital bleeding events (OR 2.02; 95% CI 1.18–3.44; p = 0.010).

Long-term outcomes. At 1 year, there were 50 (4.2%) deaths in the bivalirudin versus 2 (1.3%) in the bivalirudin plus GP IIb/IIIa inhibitor group (HR 0.30; c2 = 2.83; 95% CI 0.07–1.22; p = 0.093). At follow up (mean 24.2 ± 7.7 months), all-cause mortality was significantly higher in the bivalirudin group, 83 deaths (7.0%) in the bivalirudin versus 4 (2.6%) in the bivalirudin plus GP IIb/IIIa inhibitor arm (p log-rank = 0.036) (Figure 1). In a univariate Cox analysis, GP IIb/IIIa inhibitor use in addition to bivalirudin was associated with a decreased risk of long-term all-cause mortality (HR 0.36; c2 = 4.04; 95% CI 0.13–0.98; p = 0.044). The independent relation of GP IIb/IIIa blockade and long-term mortality was examined by using a multivariate logistic regression model. After adjustment for confounding variables with multivariate Cox analysis, GP IIb/IIIa inhibition was no longer an independent predictor of long-term mortality (HR 0.41; c2 = 3.04; 95% CI 0.15–1.12; p = 0.081). Table 3 lists other independent predictors of long-term all-cause mortality identified in the multivariate Cox model. Post-procedural bleeding (combined major and minor bleeding) was found to be a strong independent predictor of long-term all-cause mortality (hazard ratio [HR] 2.28; 2 = 9.12; 95% CI 1.34–3.90; p = 0.003).

Discussion

This study represents a contemporary evaluation of the direct thrombin inhibitor bivalirudin and provisional use of GP IIb/IIIa inhibition for angiographic complications during PCI. A broad spectrum of 1,340 patients undergoing urgent (mild to moderate risk ACS) or elective PCI was evaluated. The results demonstrate that bailout GP IIb/IIIa use in addition to bivalirudin in patients with procedural complications is more common (11.6%) in current interventional practice than previously reported in randomized clinical trials (7.2% bailout in the REPLACE-2 trial, 9.1% bailout in the ACUITY-PCI trial). In addition, provisional use of GP IIb/IIIa inhibition in bivalirudin-treated patients results in significantly greater number of hemorrhagic complications when compared with bivalirudin monotherapy, with similar suppression of peri-procedural ischemic endpoints. Importantly, provisional use of GP IIb/IIIa blockade with bivalirudin was associated with comparable long-term all-cause mortality to bivalirudin monotherapy in our current analysis.

Important lessons were learned from a subgroup of patients in the REPLACE-2 trial that received bivalirudin and provisional GP IIb/IIIa inhibition.10 In the REPLACE-2, a strategy of bivalirudin with bailout GP IIb/IIIa inhibition was found to be noninferior to heparin-plus-planned GP IIb/IIIa with respect to suppression of 30-day ischemic endpoints.7 By 30 days, the composite endpoint of death MI, or urgent repeat revascularization occurred in 7.6% of patients receiving bivalirudin versus 7.1% of patients in the heparin-plus-planned GP IIb/IIIa inhibitor group (p = 0.40), with no difference in 1-year mortality (1.9% vs. 2.5%, respectively, p = 0.16).7,9 The trial had a double-blind design allowing for provisional use of GP IIb/IIIa inhibitors in the bivalirudin group or a matched placebo in the heparin-plus-GP IIb/IIIa inhibitor group. Importantly, 7.2% (217/2,994) of patients in the bivalirudin group required provisional GP IIb/IIIa inhibitors for complications during PCI, whereas 5.2% (157/3,011) of patients in the heparin-plus-GP IIb/IIIa group received provisional placebo. The incidence of short-term ischemic and hemorrhagic complications was higher in patients requiring bailout GP IIb/IIIa inhibition compared to those who did not require provisional GP IIb/IIIa inhibition. Exaire et al suggested that inferior outcomes in the small subgroup of the REPLACE-2 patients requiring bailout GP IIb/IIIa blockade in addition to bivalirudin is due to the fact that bailout drug was administered once patients developed procedural complications, rather than from a deleterious effect of GP IIb/IIIa inhibition per se given during PCI.10 This is supported by the fact that similarly high ischemic and bleeding complications were seen among 5.2% of patients in the heparin-plus-planned GP IIb/IIIa inhibitor group and who, due to the blinded design of the trial, were thought to require a bailout drug (placebo) despite receiving a planned GP IIb/IIIa inhibition. Our analysis, demonstrating similar rates of post-procedural ischemic complications in the bivalirudin-only versus bivalirudin-plus-bailout GP IIb/IIIa blockade group, suggests that rescue use of GP IIb/IIIa inhibitors may have been beneficial in terms of reduction of post-PCI ischemic endpoints in patients developing angiographically documented procedural complications.

In the ACUITY-PCI trial, patients received one of three anticoagulation strategies for PCI in moderate and high-risk ACS: bivalirudin alone, bivalirudin plus planned GP IIb/IIIa inhibition, and heparin (unfractionated or enoxaparin) plus planned GP IIb/IIIa inhibition.12,13 In the bivalirudin-alone group, 9.1% (238/2,619) of patients required provisional GP IIb/IIIa inhibition for complications during PCI. The combined 30-day endpoint of death, MI or unplanned revascularization for ischemia was similar in patients receiving bivalirudin alone or bivalirudin-plus-planned GP IIb/IIIa inhibition (9% in both groups). The results of the ACUITY-PCI trial suggest that routine addition of GP IIb/IIIa agents to bivalirudin is of no benefit in terms of reduction of post-procedural ischemic events. In contrast, our study suggests that the provisional addition of GP IIb/IIIa inhibition to bivalirudin, once procedural complications occur, may be beneficial. If treated with a bivalirudin-only strategy, patients developing procedural complications would be expected to have higher rates of ischemic events compared to patients without procedural complications.10 Therefore, similar rates of ischemia in patients with procedural complications receiving rescue GP IIb/IIIa agents versus complication-free bivalirudin-treated patients, suggest an anti-ischemic benefit of bailout GP IIb/IIIa therapy. However, this anti-ischemic benefit comes at a price of increased post-procedural bleeding complications.

Bleeding events occurred more frequently in patients receiving provisional GP IIb/IIIa inhibition in our study. This finding of increased hemorrhagic events with addition of rescue GP IIb/IIIa blockade to bivalirudin (major bleeding rate, 6.5%) was also demonstrated when compared to those that did not receive a provisional drug in the REPLACE-2 trial (major bleeding rate, 3.0%).10 In the ACUITY-PCI trial both major (8% vs. 4%) and minor bleeding rates (28% vs. 15%) were increased in the bivalirudin-plus-GP IIb/IIIa inhibitor group versus the bivalirudin-alone group, respectively.10 Therefore, our analysis supports the randomized trial data that addition of either provisional or routine GP IIb/IIIa inhibition to bivalirudin leads to the increased hemorrhagic risk post PCI.

This report represents the longest follow-up data on patients receiving bivalirudin ± provisional GP IIb/IIIa inhibition in the current PCI era. Our 1-year mortality rate (1.3%) in the bivalirudin-plus-bailout GP IIb/IIIa group is comparable to the 1-year mortality rate (2.3%) reported in patients who received a bailout GP IIb/IIIa agent in the REPLACE-2 trial. Similar to our study, Exaire et al reported comparable 1-year mortality rates in patients who required a provisional drug versus those who did not.10 Recent studies suggested that post-procedural bleeding is as strong a predictor of long-term mortality as a post-procedural MI.16–18 This report is consistent with these data, as the post-procedural bleeding was found to be an independent predictor of long-term mortality (HR 2.28, p = 0.003). However, the short-term increase in bleeding with provisional use of GP IIb/IIIa agents in this analysis did not translate into a difference in long-term outcomes, possibly due to a small proportion of bivalirudin-treated patients who required provisional GP IIb/IIIa inhibition as well as low event rates in both groups. It is also plausible that an addition of rescue GP IIb/IIIa inhibitors in patients with angiographic complications resulted in a reduction in short-term ischemic events, therefore offsetting the negative impact of the increase in bleeding events after PCI. A trial evaluating bivalirudin monotherapy versus bivalirudin plus rescue IIb/IIIa inhibition in patients with angiographic complications during PCI would be needed to conclusively determine if the addition of GP IIb/IIIa agents results in improved short-term ischemic events and long-term outcomes. However, it would not be ethical to conduct such a trial and in the absence of such data, the addition of GP IIb/IIIa inhibitors should be strongly considered for angiographic complications during PCI.

Study limitations. We acknowledge several limitations in this study. First, our analysis was derived from a single high-volume tertiary care center population and thus might not apply to other institutions. Second, this study was not randomized, and the decision to use bivalirudin or a rescue GP IIb/IIIa agent was made by the individual physician and might be influenced by the procedure, operator and other confounding factors that could not be identified in our registry. Third, our analysis was limited by the retrospective design of the study.

Conclusions

The current study suggests that bailout GP IIb/IIIa use is higher in clinical practice than that seen in randomized trials, but the overall rate of GP IIb/IIIa bailout still remains < 15% of all PCI procedures, despite a broad spectrum of patients studied in this registry. The addition of bailout GP IIb/IIIa inhibition to bivalirudin-treated patients developing angiographic procedural complications results in an increase in hemorrhagic events, but may neutralize the negative impact of periprocedural ischemia. These data indicate that a strategy of bivalirudin monotherapy is preferable in order to reduce bleeding complications, and GP IIb/IIIa blockade should be reserved for patients with periprocedural complications in bivalirudin-treated patients undergoing PCI.

_______________________

From the Greenberg Division of Cardiology, New York Presbyterian Hospital, Weill Cornell Medical College, New York, New York.
The authors report no conflicts of interest regarding the content herein.

Address for correspondence: Dmitriy N. Feldman, MD, FACC, Assistant Professor of Medicine, New York Presbyterian Hospital, Weill Cornell Medical College, Greenberg Division of Cardiology, 520 East 70th Street, Starr-434 Pavilion, New York, NY 10021. E-mail: dnf9001@med.cornell.edu