Rescue Treatment of Thromboembolic Complications during Endovascular Treatment of Cerebral Aneurysms: A Meta-Analysis

The authors evaluated the results reported in 23 studies that included 516 patients in whom different drugs were used for rescue therapy in thromboembolic complications during endovascular treatment of intracranial aneurysms. The literature shows that rescue therapy with thrombolytic agents resulted in significantly more morbidity than rescue therapy with glycoprotein IIb/IIIa inhibitors. Tirofiban/eptifibatide resulted in significantly higher recanalization rates compared with abciximab. BACKGROUND AND PURPOSE: Intraprocedural thrombus formation during endovascular treatment of intracranial aneurysms is often treated with glycoprotein IIb/IIIa inhibitors and, in some instances, fibrinolytic therapy. We performed a meta-analysis evaluating the safety and efficacy of GP IIb/IIIa inhibitors compared with fibrinolysis. We also evaluated the safety and efficacy of abciximab, an irreversible inhibitor, compared with tirofiban and eptifibatide, reversible inhibitors of platelet function. MATERIALS AND METHODS: We performed a comprehensive literature search for studies on rescue therapy for intraprocedural thromboembolic complications with glycoprotein IIb/IIIa inhibitors or fibrinolysis during endovascular treatment of intracranial aneurysms. We studied rates of periprocedural stroke/hemorrhage, procedure-related morbidity and mortality, immediate arterial recanalization, and long-term good clinical outcome. Event rates were pooled across studies by using random-effects meta-analysis. RESULTS: Twenty-three studies with 516 patients were included. Patients receiving GP IIb/IIIa inhibitors had significantly lower perioperative morbidity from stroke/hemorrhage compared with those treated with fibrinolytics (11.0%; 95% CI, 7.0%–16.0% versus 29.0%; 95% CI, 13.0%–55.0%; P = .04) and were significantly less likely to have long-term morbidity (16.0%; 95% CI, 11.0%–21.0% versus 35.0%; 95% CI, 17.0%–58.0%; P = .04). There was a trend toward higher recanalization rates among patients treated with glycoprotein IIb/IIIa inhibitors compared with those treated with fibrinolytics (72.0%; 95% CI, 64.0%–78.0% versus 50.0%; 95% CI, 28.0%–73.0%; P = .08). Patients receiving tirofiban or eptifibatide had significantly higher recanalization rates compared with those treated with abciximab (83.0%; 95% CI, 68.0%–91.0% versus 66.0%; 95% CI, 58.0%–74.0%; P = .05). No difference in recanalization was seen in patients receiving intra-arterial (77.0%; 95% CI, 66.0%–85.0%) or intravenous GP IIb/IIIa inhibitors (70.0%; 95% CI, 57.0%–80.0%, P = .36). CONCLUSIONS: Rescue therapy with thrombolytic agents resulted in significantly more morbidity than rescue therapy with glycoprotein IIb/IIIa inhibitors. Tirofiban/eptifibatide resulted in significantly higher recanalization rates compared with abciximab.

P eriprocedural thromboembolic complications from endovascular treatment of intracranial aneurysms occur in 2%-15% of patients. 1 Intraprocedural thrombus formation is often treated with pharmacologic rescue, by using intra-arterial or intravenous administration of glycoprotein IIb/IIIa (GP IIb/IIIa) inhibitors or fibrinolytics. The primary goal of rescue therapy is to recanalize the artery to avoid permanent neurologic deficits. Many studies have described intraprocedural administration of GP IIb/IIIa inhibitors and thrombolytic agents during endovascular treatment of intracranial aneurysms. Some studies have suggested that patients receiving GP IIb/IIIa inhibitors have better outcomes than those treated with fibrinolytic therapy. Studies have also examined the comparative efficacy of irreversible GP IIb/IIIa inhibitor agents (abciximab) and reversible agents (tirofiban/eptifibatide) and the comparative efficacy of intra-arterial and intravenous administration of GP IIb/IIIa inhibitors, but they have not demonstrated any significant differences in recanalization rates and outcomes. 2,3 Overall however, the evidence is limited to small case series. We performed a meta-analysis of the literature examining angiographic and clinical outcomes in patients treated with GP IIb/IIIa inhibitors and fibrinolytic therapy for intraprocedural thrombus formation during intracranial aneurysm treatment. 4 We also performed subgroup analyses to compare outcomes of patients treated with abciximab versus tirofiban/eptifibatide and those treated with intra-arterial (IA) and intravenous GP IIb/IIIa rescue therapy.

MATERIALS AND METHODS
A comprehensive literature search of the data bases PubMed, Ovid MEDLINE, and Ovid EMBASE was designed and conducted by an experienced librarian with input from the authors. The key words "intracranial aneurysm," "endovascular therapy," "fibrinolytic," "tPA," "glycoprotein IIb/IIIa," "abciximab," "coil," "thromboembolic," and "rescue" were used in both ''AND" and ''OR" combinations. The search was limited to articles published from 1990 to January 2014 in the English language only. All studies reporting patients treated with GP IIb/IIIa inhibitors or fibrinolysis during endovascular treatment of intracranial aneurysms were selected. The inclusion criteria were the following: a series of Ͼ5 patients, with available data on periprocedural complications and angiographic outcomes. Studies performed by using administrative/insurance data bases were excluded. Two reviewers independently selected the included studies. For each study, we independently extracted the following information: recanalization rate, perioperative clinical stroke, perioperative infarct on imaging, perioperative hemorrhage, perioperative morbidity and mortality from stroke/hemorrhage, long-term good neurologic outcome, and long-term morbidity. Perioperative complications were identified as those that occurred within 30 days of the procedure. The recanalization rate was defined as a Thrombolysis in Cerebral Infarction score of Ն2 or a Thrombolysis in Myocardial Infarction (TIMI) score of Ն2. In cases in which patients underwent rescue therapy for a TIMI/TICI score of 2, recanalization was defined as a 1-point improvement in the recanalization rate. In series in which TIMI/TICI scores were not provided, recanalization was defined by using the words "near-complete or complete recanalization." Patients included in the long-term outcome analysis required a minimum of 3 months of follow-up. Good clinical outcome was defined as an mRS score of Յ2.
The primary objective of this study was to compare outcomes among patients treated with GP IIb/IIIa inhibitors and those treated with fibrinolytics. For this comparison, we did not control for the method of administration (IV or IA) or the drug used (ie, abciximab, tirofiban, eptifibatide, recombinant tissue plasminogen activator, streptokinase, and so forth). We performed 2 subgroup analyses: 1) comparing outcomes of patients treated with large-molecule/irreversible GP IIb/IIIa inhibitors (abciximab) with those treated with small-molecule/reversible GP IIb/IIIa in-hibitors (tirofiban/eptifibatide), and 2) comparing outcomes of patients receiving IA GP IIb/IIIa inhibitors versus IV GP IIb/IIIa inhibitors.

Statistical Analysis
Most of the included studies were noncomparative. Therefore, we estimated from each cohort the cumulative incidence (event rate) and 95% confidence interval for each outcome. Event rates for each intervention were pooled in a meta-analysis across studies by using the random-effects model. 5 In this model, studies are weighted by the inverse of their variance (ie, larger studies will have larger weight). The incidence rates of the different outcomes were compared between GP IIb/IIIa cohorts and fibrinolytic treatment cohorts by using an interaction test as described by Altman and Bland. 6 Additional comparisons were made between abciximab and tirofiban/eptifibatide and between IA versus IV GP IIb/IIIa inhibitors. Heterogeneity of the treatment effect across studies was evaluated by using the I 2 statistic. 7

Literature Review
Studies included in our literature review are summarized in the On-line Table. Our literature search yielded 623 articles; 562 articles were excluded after reading the abstract alone. Twenty-one studies were excluded for not meeting our minimum size criterion of Ն5 patients. Seventeen studies were excluded because they reported rescue therapy during endovascular procedures other than aneurysm treatment. Twenty-three studies met our inclusion criteria. In total, 533 patients were included in this study. Four hundred eighty patients received GP IIb/IIIa inhibitors (90.0%), and 63 patients (11.8%) received fibrinolytic therapy. Eighteen studies provided the proportion of patients during the study time period receiving rescue therapy for thromboembolic complications. From these studies, 444/7276 (6.1%) received rescue therapy. Among patients receiving Gp IIb/IIIa inhibitors, 322 patients received abciximab and 145 patients received either tirofiban or eptifibatide. One hundred forty-nine patients received IV GP IIb/IIIa inhibitors, and 284 patients received IA GP IIb/IIIa inhibitors.

Intra-Arterial versus Intravenous GP IIb/IIIa Inhibitors
Comparison of outcomes of IA versus IV GP IIb/IIIa inhibitors is summarized in Table 3. No difference in the recanalization rate was seen on the basis of the method of administration of GP IIb/IIIa inhibitors because patients treated with IA therapy had a recanalization rate of 77.0% (95% CI, 66.0 -85.0) and those treated with IV therapy had a recanalization rate of 70.0% (95% CI, 57.0 -80.0; P ϭ .36). Clinical outcomes and complication rates did not differ by method of GP IIb/IIIa inhibitor administration.

DISCUSSION
Our systematic review of rescue therapy for the treatment of intraoperative thromboembolic complications in the setting of aneurysm coiling demonstrated higher rates of short-and long-term morbidity among patients treated with fibrinolysis relative to those receiving GP IIb/IIIa inhibitors. In addition, we found a trend toward improved recanalization rates with GP IIb/IIIa inhibitors compared with thrombolytics. Patients receiving tirofiban/eptifibatide had significantly higher recanalization rates compared with those receiving abciximab; however, no difference was seen in clinical outcomes. No difference in outcome was seen among patients receiving intra-arterial and intravenous rescue therapy with GP IIb/IIIa inhibitors. These data should prompt practitioners to avoid fibrinolytics in place of GP IIb/IIIa agents for rescue therapy during coil embolization. In most case series, the rate of rescue therapy for thromboembolic complications ranged between 5% and 10%. 1,3,[8][9][10][11][12][13][14][15][16][17][18][19] Our systematic review found that approximately 6% of patients received rescue therapy during the treatment of intracranial aneurysms. Clinical outcomes of patients receiving GP IIb/IIIa inhibitors and fibrinolytics are variable. In a study of 592 patients undergoing rescue therapy in a large administrative data base, Brinjikji et al 4 found that patients with unruptured aneurysms receiving fibrinolytics had significantly higher rates of discharge to institutions other than home (37.5% versus 7.4%). Patients with ruptured aneurysms receiving fibrinolytics had significantly higher rates of in-hospital mortality (26.0% versus 14.5%) and discharge to institutions other than home (59.4% versus 36.5%). Our study found significantly higher rates of short-and long-term morbidity among patients receiving fibrinolytic therapy for rescue treatment. Furthermore, there was a trend toward improved recanalization rates among patients treated with GP IIb/IIIa inhibitors.
The infarction rate among patients receiving rescue therapy ranges between 10% and 40%, with higher rates generally seen in studies with postoperative MR imaging. 1,3,[8][9][10][11][12][13][14][15][16][17][18][19] In a series of 477 patients with 515 intracranial aneurysms, Ries et al 1 reported that  48 patients (10%) had thromboembolic events, 42 of these patients received rescue therapy with GP IIb/IIIa inhibitors, and 1 patient received rescue therapy with rtPA. Of the patients who did not receive rescue therapy, 3/5 had infarcts, while 31% of patients receiving GP IIb/IIIa inhibitor rescue therapy had infarctions on CT. Linfante et al 15 found that approximately 10% of patients undergoing endovascular embolization of intracranial aneurysms required intra-arterial GP IIb/IIIa inhibitor rescue therapy, with an infarct and hemorrhage rate of 0%. Many prior studies that have demonstrated high infarct rates on postoperative imaging also show that many of these infarcts are clinically silent. 1,8,14 In our study, there was no significant difference in the rate of clinical stroke and infarction on imaging between patients treated with GP IIb/IIIa inhibitor and those receiving thrombolytics.
The efficacy of small-molecule/reversible GP IIb/IIIa inhibitors (tirofiban/eptifibatide) relative to the large-molecule/irreversible inhibitor (abciximab) has not been well-established. One might assume that reversible, competitive inhibitory agents such as tirofiban/eptifibatide may lose potency, thus leading to decreased recanalization rates. In a meta-analysis of abciximab versus small-molecule GP IIb/IIIa inhibitors among patients receiving percutaneous coronary interventions, Gurm et al 20 found no difference in clinical and angiographic outcomes between these 2 agents. In a large randomized controlled trial of abciximab versus eptifibatide for coronary revascularization, Zeymer et al 21 found similar safety and efficacy rates between the 2 drugs. We found significantly higher recanalization rates, with no difference in clinical outcome, when comparing small-molecule GP IIb/IIIa inhibitors and abciximab.
The efficacy of intravenous-versus-intra-arterial administration of GP IIb/IIIa inhibitors for rescue therapy is also not wellestablished. One might assume that local intra-arterial administration of GP IIb/IIIa inhibitors at the site of thrombus formation would yield higher recanalization rates and thus improved clinical outcomes due to increased local concentrations of the therapeutic agent. A number of studies in the cardiology literature, in which GP IIb/IIIa inhibitors are routinely used in percutaneous coronary intervention, have examined the safety and efficacy of both the route and type of GP IIb/IIIa inhibitor administration. In a meta-analysis of randomized controlled trials of intracoronary versus intravenous administration of glycoprotein IIb/IIIa inhibitors during percutaneous coronary intervention for acute coronary syndrome, Friedland et al 22 found that intracoronary administration resulted in higher recanalization rates and short-term mortality after intervention. Fu et al 23 found that intracoronary administration increased coronary flow and myocardial reperfusion but did not improve outcomes compared with intravenous administration. While intra-arterial administration of GP IIb/IIIa inhibitors may lead to higher concentrations of the agent at the site of the thrombus, an intravenous bolus of GP IIb/IIIa inhibitors has been found to reduce platelet aggregation to Ͻ20% within 10 minutes of administration. 24 It may be that intra-arterial administration is beneficial only in the short term, which may not be sufficient to provide clinical benefit. Our study found no differences in recanalization or clinical outcomes between patients receiving IA or IV GP IIb/IIIa inhibitors.

Limitations
Our study has limitations. The data on which this meta-analysis is based are affected by publication bias. Studies with favorable results are more likely to be published. Moreover, the limitations of single-center experiences for selection bias and potential complication under-reporting are well-known. Many details were lacking from the studies included in our analysis. For example, many studies did not provide TICI/TIMI grading for vessel recanalization. We did not stratify outcomes on the basis of aneurysm rupture status. Reporting of long-term neurologic outcomes was inconsistent across studies. In addition, many of the series analyzed and included in our analysis were cases collected during several years, and it is possible that complication rates have improved as a result of increased operator experience and skill and improved devices and technology. Regarding the method of GP IIb/IIIa administration, there was variation in how intravenous doses were administered (loading dose alone, loading dose plus drip, or drip alone). The comparisons reported in this meta-analysis are made across studies, not within studies; this difference greatly weakens inference. Most of the studies in our analysis were noncomparative, and groups were not randomized. An ideal meta-analysis comparing variable pharmacologic therapies would include only comparative studies with randomized groups. Overall, the quality of the studies included in this analysis was low. If one used the Grading of Recommendations, Assessment, Development, and Evaluation framework, the quality of evidence (confidence in estimates) is very low because of imprecision, heterogeneity, and methodologic limitations of the included studies. 25-27

CONCLUSIONS
In this study, rescue therapy with GP IIb/IIIa inhibitors is associated with less long-and short-term morbidity compared with fibrinolytics for rescue therapy during intracranial aneurysm coiling. Our meta-analysis suggests that eptifibatide/tirofiban treatment results in significantly higher recanalization rates than treatment with abciximab. No difference in outcome was seen on the basis of the method of administration of GP IIb/IIIa inhibitors. Further clinical trials might study the relative safety and efficacy of various GP IIb/IIIa inhibitors and methods of administration to help determine the best treatment for intraprocedural thromboembolic complications during intracranial aneurysm treatment.