Introduction

Abciximab is a chimeric mouse/human monoclonal antibody with high binding affinity for the platelet glycoprotein IIb/IIIa (GP IIb/IIIa) receptor, causing inhibition of platelet aggregation [13]. Abciximab has previously been shown to be safe and effective in reducing thrombotic complications associated with percutaneous coronary intervention (PCI), as well as in reducing ischemic complications of acute myocardial infarction [1, 47]. Abciximab and other GP IIb/IIIa receptor inhibitors (GPIs) are not currently licensed for any indication in cerebrovascular diseases. However, as a result of its success in the field of cardiology, abciximab is being used as an adjunct to neuroendovascular procedures to both prevent and treat ischemic sequelae in the setting of intracranial aneurysms, intracranial vascular stenting, and acute ischemic stroke [818]. Experience with abciximab in this setting is limited; its safety and efficacy in association with neuroendovascular procedures has not been clearly defined. Major bleeding complications, including fatal intracranial hemorrhage (ICH), have been reported with the use of GPIs [1, 9, 12, 13, 16, 19, 20]. In the current study, we relate our multicenter experience with ICH following administration of abciximab during neuroendovascular procedures.

Methods

Following Institutional Review Board (IRB) approval, we retrospectively reviewed our electronic patient databases and identified neuroendovascular procedures (including diagnostic cerebral angiograms, aneurysm coiling procedures, angioplasty and vascular stenting procedures, and emergent revascularization procedures for acute vascular occlusions) that used abciximab intravenously, intra-arterially, or both, at Mayo Clinic Hospitals in Rochester, Jacksonville, and Phoenix between November 2000 and April 2009. Abciximab was studied instead of other GPIs because our institutions use abciximab exclusively in the neuroendovascular setting. All procedures were performed by experienced neurointerventionalists utilizing a spectrum of standard materials and practices. Abciximab was administered as intravenous (IV) bolus, intra-arterial (IA) bolus, IV infusion, or as a combination of routes in all patients. From this pool of patients, all cases of periprocedural ICH were identified and the following data were collected: age, gender, comorbidities, smoking history, type of procedure, intraprocedural complications, periprocedural blood pressure, periprocedural medications, route and dose of abciximab administered, laboratory values (including coagulation parameters), and time to identification of ICH. Clinical outcome was measured either at death or discharge by the Glasgow Outcome Scale (GOS) [21]: 1 = death; 2, = vegetative state; 3 = severe disability; 4 = moderate disability; 5 = good recovery. Radiographic data reviewed included all available periprocedural CT and MRI studies. All cranial imaging was initially reviewed and interpreted by experienced neuroradiologists. Fisher’s exact test was used to assess the association between procedural variables and the development of ICH.

Results

Abciximab was used in 51 neuroendovascular procedures during the study period (Table 1). Twenty-eight of the patients presented acutely (ex. stroke, ruptured aneurysm, etc.), and 23 of the patients presented electively (ex. non-urgent diagnostic angiogram, incidental/unruptured aneurysm, etc.). Nine ICH cases were identified (Fig. 1), 5 from the acute procedure group and 4 from the elective procedure group (Table 2). There were 6 males and 3 females, with a median age of 68 years (mean 66.6, range 42–94). The indications for abciximab administration among these cases were varied and included distal propagation of thrombus (n = 2: 1 during attempted revascularization of acute internal carotid artery occlusion, 1 during coiling of anterior communicating artery aneurysm); prophylaxis during stent-assisted coiling of unruptured aneurysms (n = 2: 1 vertebral artery, 1 internal carotid artery); attempted emergent thrombolysis of internal carotid artery occlusion presenting with acute ischemic stroke (n = 2); thrombus formation at aneurysm-parent artery junction during coiling of aneurysm (n = 2: 1 anterior communicating artery ruptured aneurysm, 1 vertebrobasilar junction unruptured aneurysm); and spastic vessel occlusion during coiling of ruptured aneurysm requiring balloon angioplasty (n = 1). In the setting of a ruptured aneurysm, abciximab was used only after the aneurysm was secured.

Fig. 1
figure 1

Representative head CT images from Patients 1–9 with ICH. a Patient 1, demonstrating SAH, IVH, IPH, and SDH. b Patient 2, demonstrating SAH and IVH. c Patient 3, demonstrating SAH, and IVH (not shown in current slice). d Patient 4, demonstrating IPH. IVH was also present though more apparent on other slices (not pictured). e Patient 5, demonstrating IPH which occurred along the EVD tract. SAH and IVH were also present, though not visualized on pictured slice. f Patient 6, demonstrating IPH and IVH. Also evident is diffuse hypodensity corresponding to a large left hemispheric infarction. g Patient 7, demonstrating SAH, IPH, and IVH. h Patient 8, demonstrating SAH and IVH. i Patient 9, demonstrating SAH and IVH. CT computed tomography, EVD external ventricular drain, ICH intracranial hemorrhage, IPH intraparenchymal hemorrhage, IVH intraventricular hemorrhage, SAH subarachnoid hemorrhage, SDH subdural hemorrhage

Table 1 Data from all patients undergoing neuroendovascular procedure with abciximab
Table 2 Data from patients experiencing intracranial hemorrhage following neuroendovascular procedures in which abciximab was used

Route of abciximab administration included IV bolus only (n = 4), IA bolus followed by IV infusion (n = 3), IV bolus followed by IV infusion (n = 1), and IV infusion without preceding bolus (n = 1). The route of administration was chosen according to physician preference in each case. All IV boluses were administered over 5 min as 0.25 mg/kg doses, except for Patient 7 who received half-dose (0.125 mg/kg). Doses varied for the 3 patients that received IA boluses; Patient 5 received 10 mg (0.167 mg/kg), Patient 6 received 21.6 mg (0.25 mg/kg), and the dose for Patient 1 was not clearly documented. All IV infusions of abciximab were administered at a rate of 0.125 mcg/kg/min and discontinued at time of ICH discovery. All but 1 of the patients (Patient 6) received concomitant periprocedural antiplatelet, anticoagulant, or thrombolytic agents. Two patients received such medications prior to procedure: Patient 1 was treated with recombinant tissue plasminogen activator (r-tPA) 0.6 mg/kg IV (65 mg total, with 10% given as bolus and remainder administered as an infusion over 1 h), and Patient 4 received Aspirin 325 mg and Clopidogrel 300 mg enterally. All but one of the patients (Patient 6) received an intraprocedural IV bolus of heparin (range 3,000–5,500 units; mean 4,062 units). One patient also received intraprocedural Aspirin (Patient 2—325 mg sublingually). Postprocedural Aspirin was administered in 7 patients (Patient 1—325 mg enterally; Patient 2—600 mg rectally; Patient 3—650 mg enterally; Patient 4—325 mg enterally; Patient 5—650 mg enterally; Patient 7—325 mg enterally; Patient 8—81 mg enterally). Three patients additionally received postprocedural enteral clopidogrel (Patient 1—300 mg; Patient 4—75 mg; Patient 7—600 mg).

The intraprocedural-activated clotting time (ACT) was recorded in 6 cases, with a median value of 317 s (mean 346, range 239–550). Patient 7 did not have a clearly recorded intraprocedural ACT, but it was noted to be greater than 250 s. Patient 1 did not have a recorded intraprocedural ACT, but there was a post-procedural partial thromboplastin time (PTT) of 93.4 s. The sole unheparinized patient, Patient 6, did not have a recorded intraprocedural ACT or post-procedural PTT available.

All cases of ICH were detected within 7 h of abciximab administration, (except for Patient 4; 40 h); 5 cases were detected within 3 h. The pattern of ICH varied and included subarachnoid hemorrhage (SAH) with intraventricular hemorrhage (IVH) (n = 5); intraparenchymal hemorrhage (IPH) with IVH (n = 2); SAH, IVH, and IPH (n = 1); and a combination of SAH, IVH, IPH, and subdural hemorrhage (SDH) (n = 1). It should be noted that both Patient 5 and Patient 9 had preexisting ICH, but it became worse following the neuroendovascular procedure with abciximab administration. Patient 5 had SAH and IVH present prior to endovascular procedure, but IPH developed along the tract of the external ventricular drain (EVD) (that was placed prior to the endovascular procedure) following the coiling procedure. Patient 9 had SAH and IVH present prior to the endovascular procedure, but had worsening of SAH and IVH following the coiling procedure. Four patients died following ICH (i.e., GOS score of 1); GOS scores at discharge for the remaining cases were 3 (n = 3), and 5 (n = 2).

From the initial group of 51 patients treated with abciximab, there was no significant difference in ICH risk between those patients who presented acutely and those who presented electively (17.9 vs. 17.4%, Fisher’s exact test, P = 1.0). Moreover, we did not identify an increased risk of ICH in the setting of concomitant mechanical thrombectomy for acute stroke. Of the 28 patients who presented acutely, 12 patients presented with acute arterial occlusion and stroke. Seven (58%) of these patients were treated with both abciximab and mechanical clot retrieval with 1 occurrence of ICH, and 5 patients (42%) received abciximab without concomitant mechanical clot retrieval with 2 occurrences of ICH (P = 0.52).

Discussion

Abciximab is a mouse/human chimeric monoclonal Fab fragment (c7E3 Fab), created via genetic reconstruction, consisting of mouse-derived variable regions linked to human-derived constant regions from immunoglobulin IgG [2, 3, 22]. Abciximab inhibits platelet aggregation by engaging the GP IIb/IIIa receptor, thereby preventing the binding of fibrinogen and von Willebrand factor to activated platelets [2, 3]. It has also been shown to inhibit thrombin formation [23]. The biological half-life of abciximab is prolonged (6–12 h or more); however, the plasma half-life is much shorter (15–30 min) [24]. Abciximab’s ability to block GP IIb/IIIa receptors and inhibit platelet aggregation occurs in a dose-dependent manner. Tcheng et al. [3] demonstrated that after an IV bolus dose of 0.15 mg/kg of abciximab, a median of 54% of GP IIb/IIIa receptors were blocked, increasing to a median of 87% blockade at a dose of 0.25 mg/kg. Platelet aggregation in response to 20 μmol/l of ADP was 46% of baseline after an IV bolus dose of 0.15 mg/kg of abciximab, decreasing to just 18% at a dose of 0.25 mg/kg. These data helped to establish a dose of 0.25 mg/kg as an effective bolus dose of abciximab; a dose that causes >80% GP IIb/IIIa receptor blockade and suppresses platelet aggregation to <20% of baseline that is used in trials and clinical practice.

Mascelli et al. [2] studied the pharmacodynamics of abciximab in subjects who received a 0.25 mg/kg IV bolus followed by 12 h infusion of either 10 μg/min or 0.125 μg/kg/min of abciximab. Maximum suppression of platelet function was observed at 30 min after IV bolus, with a median platelet aggregation response of 2% of baseline. Partial recovery from both GP IIb/IIIa receptor blockade and inhibition of ADP-mediated platelet aggregation was observed in the majority of patients at 12 h after cessation of abciximab therapy, with a median GP IIb/IIIa receptor blockade of 68%, and a platelet aggregation response median of 38% of baseline. Therefore, partial recovery of blockade of both GP IIb/IIIa receptors and platelet aggregation was observed in the majority of subjects within 12 h after cessation of abciximab treatment. However, ex vivo flow cytometric evaluations revealed that a substantial amount of abciximab was bound to circulating platelets in the majority of subjects up to 15 days after abciximab treatment.

Experience with abciximab in the setting of neuroendovascular procedures is limited; its safety and efficacy in association with such procedures remains to be clearly defined. Major bleeding complications associated with abciximab, including fatal ICH, have been reported in the cardiology, stroke, and neuroendovascular literature. The EPIC study, a multicenter randomized double-blind trial of 2,099 patients randomized to PCI with heparin, Aspirin and either placebo bolus and infusion, abciximab bolus and placebo infusion, or abciximab bolus and infusion, reported a substantially increased rate of bleeding complications in the group receiving abciximab bolus and infusion, and a more moderate increase in the abciximab bolus-only group [1]. Six patients experienced ICH: 2 assigned to placebo, 1 assigned to abciximab bolus, and 3 assigned to abciximab bolus and infusion (though 1 of these did not receive study drug as ICH occurred after randomization but before angioplasty). The CAPTURE study, a subsequent randomized placebo-controlled trial that assigned 1,265 patients with refractory unstable angina to IV infusion with abciximab or placebo for 18–24 h prior to PCI with heparin and aspirin, documented similar findings with major bleeding occurring more frequently in the abciximab-treated group (3.8 vs. 1.9%, P = 0.043) [6]. Interestingly, however, an excess of ICH was not found in the abciximab-treated group. This difference in ICH between the EPIC and CAPTURE studies may be attributable to differences in heparin dosing. The EPIC study used non-weight-adjusted dosing (initial bolus of 10,000–12,000 U, followed by additional boluses to keep the ACT between 300 and 350), while the CAPTURE study used weight-adjusted dosing (initial bolus of 100 U/kg, maximum 10,000 U, with additional boluses to maintain ACT > 300). In a pooled analysis (8,555 patients) combining data from EPIC, CAPTURE, and two other major double-blind placebo-controlled randomized trials of patients undergoing PCI with heparin (EPISTENT [5] and EPILOG [7]), Akkerhuis et al. [25] showed that the rate of hemorrhagic stroke was not significantly different between abciximab-treated patients and placebo-treated patients (0.16 vs. 0.10%, P = 0.56). However, there was a trend toward a higher rate of hemorrhagic stroke in patients receiving abciximab and standard-dose heparin compared to those receiving abciximab and low-dose heparin (0.27 vs. 0.04%, P = 0.057).

Recently, the AbESTT-2 study, an international phase 3 trial evaluating the safety and efficacy of abciximab use in the setting of acute ischemic stroke, was terminated prematurely due to an increased risk of symptomatic or fatal ICH in the abciximab-treated group when compared to placebo (5.5 vs. 0.5%, P = 0.002) [19]. In a previous multicenter series, Qureshi et al. [9] identified 7 patients who developed fatal ICH that was associated with using a combination of antithrombotic agents including IV abciximab as an adjunct to neurointerventional procedures. In a comparison of GPIs with emboli prevention devices (EPDs) during carotid stenting in 305 consecutive patients, Chan et al. [13] found that the composite end point of neurologic death, nonfatal stroke, and major bleeding including ICH, was significantly higher in patients treated with GPIs (5.1 vs. 0%, P = 0.02). Two patients in the GPI group (1 treated with abciximab and the other with eptifibatide) experienced periprocedural fatal ICH. Two more patients treated with abciximab experienced delayed ICH, 1 being fatal and 1 resulting in partial neurologic deficit. Velat et al. [11] reported on a series of 29 patients treated with abciximab for thromboembolic complications during neuroendovascular procedures; ICH occurred in 3 cases (1 being fatal). All cases of ICH were felt to be secondary to iatrogenic dissection created during mechanical clot disruption, and 2 of the 3 received concomitant r-tPA administration.

Several series have evaluated the safety and efficacy of abciximab administration for treatment of thromboembolic events occurring during aneurysm coiling procedures (Table 3) [1218, 26]. These series included small numbers of patients and offered varying results. No occurrences of post-procedural ICH were reported by Ries et al. [26] (series of 42 patients), Jones et al. [14] (series of 38 patients), Aviv et al. [15] (series of 13 patients), Song et al. [17] (series of 7 patients), or Mounayer et al. [18] (series of 13 patients). In a series including 32 patients who received IA abciximab for treatment of acute thromboembolic complications during aneurysm coiling, Park et al. [12] found that 3 patients experienced post-procedure ICH; 2 of these patients died and the other was left severely disabled. Gralla et al. reported on a series of 63 patients treated with IA or IV abciximab during aneurysm coiling procedure (41 patients) or treated with IV infusion shortly following procedure (22 patients) for treatment of ischemic complications; 2 patients suffered fatal ICH (1 from the group that received intraprocedural administration of abciximab, and 1 from the group that received post-procedure administration) [16]. The authors attributed these occurrences of ICH to bleeding into pre-existing subacute infarcted tissue rather than re-rupture of aneurysms. In the present study, 3 patients (Patients 1, 6, and 7) experienced ICH in the setting of recent infarction. Patients 1 and 6 had hemorrhagic transformation of ischemic stroke following abciximab administration. Abciximab use likely increased risk of hemorrhagic transformation in these 2 patients, however, additional factors likely contributed. Patient 1 received IV thrombolytic therapy that clearly increased the risk of ICH. Additionally, Patient 1 had SAH, IVH, and SDH in addition to IPH that cannot be explained by hemorrhagic transformation of ischemic stroke alone. Patient 6 underwent attempted mechanical clot retrieval that may have increased the risk of hemorrhage in association with abciximab use. Finally, although Patient 7 experienced ICH in the setting of recent ischemic stroke, there was no hemorrhagic transformation, and SAH was felt to be secondary to internal carotid artery dissection that was present prior to the procedure or mechanical trauma related to the procedure. Patient 7’s deficits at discharge were felt to be primarily secondary to ischemic stroke rather than ICH.

Table 3 Published series studying the neuroendovascular use of abciximab

There is no known direct antidote to abciximab. Platelet administration, however, has been shown to diminish its antiplatelet effect [2, 3, 24]. Tcheng et al. [3] noted that abciximab’s platelet suppression could be partially reversed by the administration of fresh platelets, with recovery of normal hemostasis. Mascelli et al. [2] demonstrated that abciximab continuously redistributes between platelets and equilibrates onto new platelets entering the circulation. Since free circulating abciximab in the plasma is rapidly cleared by protease degradation, platelet transfusion can be expected to result in redistribution of abciximab to transfused platelets, causing dilution of bound GP IIb/IIIa receptors and, therefore, a decrease in inhibition of platelet aggregation [24]. Unfortunately there are no consensus guidelines regarding the use of platelets in this setting, and the optimal amount of platelets to administer is unknown. The authors recommend administering one single donor unit of platelets (which is equivalent to a 4–6 packs of whole-blood derived pooled donor platelets) as soon as ICH is identified. In the present study, 3 patients received platelet transfusion following identification of ICH (Patients 3, 4, and 9). Despite platelet transfusion, however, 2 of these patients (Patients 3 and 9) died.

Given the high morbidity and mortality conferred by ICH associated with abciximab administration, and the fact that there is no direct antidote available, efforts have been focused on preventing ICH from occurring in the first place. A variety of strategies have been suggested, including: the partial reversal of concomitant heparin with protamine following completion of the procedure [10]; decreasing the intra-procedural heparin effect with maintenance of a lower intra-procedural ACT [14]; the use of lower doses of abciximab [1217]; the avoidance of further coil placement following abciximab administration during coiling procedures [15]; the avoidance of abciximab use in the setting of subacute infarcted tissue [16]; and the avoidance of concomitant thrombolytics or mechanical clot disruption [11, 14]. Notably, we did not identify an increased risk of ICH in those patients presenting with acute stroke who underwent concomitant mechanical thrombectomy compared to those who received abciximab alone (Fisher’s exact test, 14.3 vs. 40%, P = 0.52). Moreover, no significant difference was found in ICH risk for the administration of abciximab during acute versus elective procedures (17.9 vs. 17.4%, P = 1.0). The use of bolus-only rather than bolus-plus-infusion protocols [15, 17, 18] and the use of IA rather than IV abciximab [17, 18] has also been suggested. Notably, we did not discover ICH in any patients treated with IA bolus alone (n = 6) from the initial pool of 51 abciximab-treated patients. Another potential strategy is the use of one of the other commercially available GPIs (eptifabitide or tirofiban) instead of abciximab. These agents have shorter half-lives (2.5 and 2–2.5 h, respectively) [24] and, as such, could have the potential for a lower risk of major bleeding and ICH in the setting of neuroendovascular use. However, although these agents have shorter half-lives than abciximab, the issue of reversal may be of greater concern. Their anti-platelet effects follow classic competitive pharmacokinetics, and their stoichiometry is such that the number of molecules of drug overwhelms the GP IIb/IIIa receptors by several orders of magnitude [24]. Thus, unlike abciximab, transfusion of platelets has little effect on these agents. Nevertheless, direct comparisons between low-dose, IA infusions of these GPIs and abciximab should be performed, as these agents may prove to have an advantage over abciximab in the neuroendovascular setting.

Conclusions

Abciximab is used in neuroendovascular procedures for both the prevention and treatment of ischemic complications. Although this agent can be very useful in selected neurointerventional situations, experience with abciximab in this setting remains limited and complications are a clear concern. ICH leading to death or serious disability is a risk associated with abciximab use during neuroendovascular procedures. In the present study, ICH following abciximab administration during neuroendovascular procedures occurred often (approximately 18% of cases), with a fatality rate of 44%. Future management strategies should focus on early recognition of GPI-related-ICH; direct comparisons of abciximab with other, shorter-half-life GPIs (eptifabitide and tirofiban); comparisons of low-dose IA bolus therapy with typical 0.25 mg/kg bolus IV treatment, both with and without post-bolus IV infusions; development of GPI antidotes; and evaluation of alternative surgical and/or endovascular procedures and devices (such as heparin-coated intracranial stents) that may reduce the risk of ischemic complications associated with neurointerventions.