Thrombectomy Outcomes of Intracranial Atherosclerosis-Related Occlusions

Introduction

Endovascular thrombectomy has become the standard of care for acute stroke because of large vessel occlusion (LVO).¹ Most LVOs are secondary to emboli of cardiac or carotid origin, and current techniques of stent-retriever and aspiration thrombectomy are highly effective in removing these emboli. However, these techniques are less efficacious in LVOs with underlying intracranial atherosclerosis (ICAS).^2,3 Intraprocedural reocclusion has been commonly reported and rescue treatment with intraarterial thrombolysis, balloon angioplasty, or stenting may be required for successful revascularization.⁴ In addition, patients with ICAS-related occlusions (ICAS-O) demonstrate different risk factor profiles and unclear therapeutic outcomes. To understand the clinical features and thrombectomy outcomes of ICAS-O, we performed a systematic review and meta-analysis of studies comparing ICAS-O and non–ICAS-O treated with endovascular thrombectomy.

Methods

The data that support the findings of this study are available from the corresponding author on reasonable request.

Literature Search Strategy and Study Selection

A systematic search in the English literature with Ovid Medline, Pubmed, and Embase from January 2010 to December 2018 was performed. The following terms and their combinations were used as keywords or MeSH terms: intracranial atherosclerosis, stenosis, stenting, angioplasty, mechanical thrombectomy, endovascular, and stroke. We also manually searched the reference lists of the 18 relevant articles to identify additional studies reporting on the clinical features and thrombectomy outcome of ICAS-O that were not included in the initial literature search.

The identified studies were then evaluated with the following inclusion criteria: (1) studies comparing clinical features and risk factors of ICAS-O and non–ICAS-O and (2) studies reporting separately the thrombectomy and clinical outcomes in ICAS-O and non–ICAS-O groups. The exclusion criteria were as follows: (1) noncomparative studies reporting outcomes only on ICAS-O without a control group of non–ICAS-O, (2) case reports or studies with <5 patients in the ICAS-O group, and (3) studies that reported ICAS treatment in a subacute, nonemergent setting. Both randomized and observational studies were included. This meta-analysis was performed according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines. As this is a meta-analysis of published studies, formal approval by an ethics committee was not required.

Risk of Bias Assessment

The risk of bias of the included articles was assessed by 2 independent readers with the Newcastle Ottawa scale for cohort studies.⁵ The scale assesses the selection, comparability, and ascertainment of outcomes of the study groups, with a higher score indicating a lower risk of bias. Studies that used well-defined selection criteria, with comparable baseline stroke severity, clearly defined diagnostic criteria for ICAS-O, and those that had an independent assessment of clinical and technical outcomes are considered to have a low risk of bias.

Outcome Variables

Patients were divided into ICAS-O and non–ICAS-O groups. For the purpose of this study, patients were considered ICAS-O if they fulfilled the diagnosis criteria adopted by the authors of the respective article. Patients with LVO who received thrombectomy in the same study period and not diagnosed as ICAS-O were considered non–ICAS-O.

The primary outcome was the rate of successful reperfusion, defined by a final modified Thrombolysis in Cerebral Infarction (TICI) score of 2b/3. Secondary outcomes include good functional outcome defined as modified Rankin Scale (mRS) score of 0–2 at 90 days, 90-day mortality, symptomatic intracerebral hemorrhage, baseline demographics, and prevalence of cerebrovascular risk factors. Other technical outcomes studied include intraprocedural reocclusion, need for rescue endovascular treatment with balloon angioplasty or stent, the time from groin puncture to reperfusion, and time from symptom onset to reperfusion.

Statistical Analysis

We extracted from each study a 2×2 table for binary outcomes and the mean group sample size and a variability measure for continuous outcomes. The pooled outcomes were meta-analyzed using a random-effects model.⁶ Heterogeneity of the studies not attributable to chance was quantified with the I² statistic.⁷ The 95% CI of the odds ratio (OR) for binary outcomes and weighted mean difference for continuous outcomes were reported. Outcomes with median and interquartile range were converted to a mean and SD value based on the assumption of a log-normal distribution of the original measure.

Sensitivity analyses were performed by studying the comparative outcomes including only those studies that include predominantly (>85%) anterior circulation thrombectomy. Meta-analysis and statistical analysis were performed with OpenMeta-Analyst.⁸

Results

Literature Search

The initial literature search yielded 125 articles. The titles and abstracts of these were read and 101 articles were excluded for irrelevance. Of the remaining 24 articles, 4 were excluded for being case reports or conference abstracts and 2 were excluded for being review or editorial articles. After review, 5 studies were excluded for not reporting separately outcomes of ICAS-O, and 2 articles were excluded for overlapping patient population. In total, 11 eligible studies were included for meta-analysis.^9–19 The Preferred Reporting Items for Systematic Reviews and Meta-Analyses flow diagram is provided in Figure 1.

• Download figure
• Download PowerPoint

Study Characteristics and the Proportion of ICAS-O

A total of 1967 patients (496 ICAS-O and 1471 non–ICAS-O) in the 10 retrospective and 1 prospective observational studies were included. Ten studies were performed in Asia (6 in Korea, 3 in China, and 1 in Hong Kong), and 1 in the United States. Five studies reported on predominantly anterior circulation thrombectomy (>85%),^9,12–14,16 and 3 reported exclusively on posterior circulation thrombectomy.^10,11,15 ICAS-O accounted for 27.7% (95% CI, 18.7%–36.7%) of all thrombectomies for LVO stroke. Five of the studies had a low risk of bias, and 6 had a moderate risk of bias. The included studies are summarized in Table 1.

Patient Characteristics and Risk Factors

Comparing ICAS-O and non–ICAS-O groups, there were statistically significant differences in the age (63.7 versus 67.2; mean difference, −3.2; 95% CI, −4.68 to −1.67) and proportion of male patients (70.4% versus 51.8%; OR, 1.84; 95% CI, 1.45–2.34).

The patients with ICAS-O had significantly more hypertension (71.4% versus 63.1%; OR, 1.46; 95% CI, 1.10–1.93), diabetes mellitus (31.9% versus 22.5%; OR, 1.68; 95% CI, 1.29–2.20), dyslipidemia (36.0% versus 28.6%; OR, 1.94; 95% CI, 1.04–3.62), and smoking history (44.6% versus 21.8%; OR, 2.11; 95% CI, 1.40–3.17) but less atrial fibrillation (16.4% versus 54.1%; OR, 0.20; 95% CI, 0.13–0.31) and a lower National Institutes of Health Stroke Scale score at presentation (14.5 versus 17.0; mean difference, −2.23; 95% CI, −2.98 to −1.48). There was no difference in the prevalence of coronary artery disease (CAD), the location of the occlusion, or the use of intravenous thrombolysis (Table 2).

Thrombectomy and Clinical Outcomes

The ICAS-O group had significantly higher intraprocedural reocclusion rate (36.9% versus 2.7%; OR, 23.7; 95% CI, 6.96–80.7), need for rescue balloon angioplasty (9.0% versus 1.3%; OR, 9.49; 95% CI, 4.11–21.9), and rescue intracranial stenting (37.8% versus 2.6%; OR, 14.9; 95% CI, 7.64–29.16). The puncture-to-reperfusion time (80.8 versus 55.5 minutes; mean difference, 21.3; 95% CI, 11.3–31.3) and onset-to-reperfusion time (401.5 versus 333.4 minutes; mean difference, 56.4; 95% CI, 18.7–94.1) were also longer in the ICAS-O cohort.

There was no difference in the rate of final recanalization (TICI 2b/3) and in the rate of symptomatic intracerebral hemorrhage between ICAS-O and non–ICAS-O. There was also no significant difference in the functional outcome (mRS score of 0–2) and in the mortality rate at 90 days between groups. These results are summarized in Figure 2 and Table 3.

• Download figure
• Download PowerPoint

Study Heterogeneity

There was low heterogeneity (I²<50%) for the following outcomes: proportion of male patients (I²=2.8%), hypertension (I²=21.3%), diabetes mellitus (I²=13%), atrial fibrillation (I²=48.6%), coronary heart disease (I²=49.1%), age (I²=25.4%), baseline National Institutes of Health Stroke Scale (I²=20.4%), need for rescue balloon angioplasty (I²=0%), need for rescue intracranial stenting (I²=36.3%), symptomatic intracranial hemorrhage rate (I²=0%), functional outcome (mRS score of 0–2) at 3 months (I²=44%), and mortality rate (I²=4.4%). There was moderate substantial heterogeneity (I²>50%) for the following outcomes: dyslipidemia (I²=69.5%), smoking (I²=58.6%), occlusion location (I²=58.3%), use of intravenous thrombolysis (I²=51.7%), intraprocedural reocclusion (I²=78.2%), final rate of TICI 2b/3 (I²=68.1%), puncture-to-reperfusion time (I²=78.4%), and onset-to-reperfusion time (I²=53%).

Sensitivity Analysis

Because of the notion that anterior and posterior circulation ICAS-O may have a different prognosis, we performed a subgroup analysis to determine whether the outcomes were different in anterior circulation versus posterior circulation ICAS-O. In the 5 studies that include predominantly (>85%) anterior circulation ICAS-O, there was no statistically significant difference in final modified TICI 2b/3 rate (OR, 0.76; 95% CI, 0.52–1.11) and good functional outcome (mRS score of 0–2) at 90 days (OR, 1.12; 95% CI, 0.74–1.69) between groups. In the 3 studies that exclusively include posterior circulation ICAS-O, there was also no statistically significant difference in final TICI 2b/3 rate (OR, 0.47; 95% CI, 0.08–2.78) and good functional outcome (mRS score of 0–2) at 90 days (OR, 0.99; 95% CI, 0.58–1.68).

Discussion

ICAS is an important cause of LVO stroke which poses unique challenges to endovascular thrombectomy. Although it is more prevalent in Asia, it can affect patients of any ethnicity and is also often found in the black and Hispanic populations.²⁰ The literature on endovascular thrombectomy in ICAS-O is mainly comprised Asian studies, and our review suggests that ICAS-O accounts for up to a quarter of the LVO stroke burden in Asia. The exact proportion of ICAS-O in non-Asian populations has not been systematically studied but is likely lower, accounting for 5.5% and 8.3% of LVOs in single-center cohort studies from France and the United States, respectively.^17,21 This disparity is likely because of the different risk factor profiles of patients with ICAS-O and non–ICAS-O, as highlighted by the current study.

Compared with LVOs of other causes, patients with ICAS-O are more likely men and of a younger age. The prevalence of hypertension, diabetes mellitus, dyslipidemia, and smoking is also higher compared with patients with non–ICAS-O. This coincides with the different distribution and prevalence of cardiovascular risk factors between Asian and white populations demonstrated in the Global Burden of Disease study.²² In particular, the prevalence of atrial fibrillation, a major culprit of LVO stroke, is lower in Asian populations when compared with other ethnicities.²³ It was estimated that 8% of the white elderly population has atrial fibrillation, while the prevalence is only 3.9% in the elderly population of Asian origin.²⁴ There was no difference in the CAD prevalence observed between ICAS-O and non–ICAS-O group, which could be due to underdiagnosis of asymptomatic CAD in these patients. Hoshino et al²⁵ systematically studied a cohort of patients with ischemic stroke with no prior history of CAD with computed tomography coronary angiogram and identified asymptomatic CAD in 37.5% of patients. Similar results were found by Wu et al,²⁶ in a Taiwanese population. In addition, ICAS in at least the Chinese population does not seem to be associated with the typical risk factors for CAD, such as hypertension, diabetes mellitus, and hyperlipidemia.²⁷ Knowledge and understanding of this unique risk factor profile are important for the clinician to consider the possibility of an underlying ICAS lesion when performing emergent thrombectomy for patients in this ethnic group.

There are both diagnostic and therapeutic challenges in the management of ICAS-O. Currently, there is no reliable way to diagnose an ICAS-O using preoperative imaging in the setting of an acute LVO stroke.⁴ Various imaging predictors had been suggested to be associated with ICAS-O, including the degree of calcification of the intracranial carotid arteries on computed tomography, clot burden as assessed by gradient-echo magnetic resonance imaging, and the pattern of ischemic lesions on magnetic resonance imaging.^16,28 While these factors may suggest an underlying ICAS, they are by no means definitive and may be present in other causes of arterial occlusion such as a fibrous embolus. In the absence of universally accepted diagnostic criteria, most centers consider an occlusion to be because of underlying ICAS when there is (1) residual stenosis of 50% or more after initial thrombectomy or (2) intraprocedural restenosis or reocclusion or (3) evidence of hypoperfusion in territories downstream to the stenosis; and (4) other differential diagnosis such as vasospasm or vessel dissection have been ruled out. This is typically established by repeating the angiogram 10 to 20 minutes after a successful thrombectomy attempt. The interrater reliability using these criteria appeared to be good with a κ value up to 0.9 in a Korean study,⁹ and similar criteria are adopted by most of the included articles in this review.

The major therapeutic difficulty of ICAS-O is the tendency of intraprocedural reocclusion, which occurred in over one-third of patients compared with only 2.7% in the non–ICAS-O group. Previous autopsy studies on post-thrombectomy patients with underlying ICAS showed histological evidence of fibrous cap disruption, intraplaque hemorrhage, and subintimal dissection of the involved vessel segment, which presumably led to early reocclusion of the recanalized vessel.^29,30 It is important and reassuring to note that endovascular rescue therapy with balloon angioplasty, local infusion of glycoprotein IIb/IIIa inhibitors, and ultimately stent deployment (or detachment of a stent-retriever) was successful in revascularizing the cerebral circulation in most cases. Indeed, despite the longer puncture-to-reperfusion time in the ICAS-O cohort, there was no difference in the rate of final TICI 2b/3 revascularization as well as good functional outcome between the 2 groups.

The optimal rescue therapy in case of early reocclusion in ICAS-O remains unclear and different first-line therapies were used. Among the studies included in this review, intraarterial infusion of glycoprotein IIb/IIIa inhibitors such as Tirofiban or Abxicimab via the distal access catheter or the microcatheter was commonly used as the first-line therapy or as an adjunct to intracranial angioplasty/stenting, although detailed information about the dosage and duration was not available for analysis. Emergent intracranial stenting was necessary in one-third of the patients with ICAS-O, and another 9% received balloon angioplasty without stenting. Kang et al compared the outcomes of emergent angioplasty versus intraarterial glycoprotein IIb/IIIa inhibitor infusion in a recent 2-center prospective study of 140 patients.³¹ They showed that both approaches achieved a high revascularization rate of 95% with no difference in functional outcome and mortality. Additionally, the parenchymal and subarachnoid hemorrhage rate was nonsignificantly higher in the center which primarily used rescue balloon angioplasty and stenting, although most of these hemorrhages were asymptomatic.³¹ Further comparative studies are needed to delineate the safety profile and efficacy of these endovascular rescue approaches.

Intracranial stenting and angioplasty for ICAS in acutely symptomatic patients with stroke are controversial because of the high complication rate demonstrated in previous randomized trials.^32,33 Although the recent WEAVE trial (Wingspan Stent System Post-Market Surveillance) showed a low periprocedural complication rate if intracranial stenting was performed 8 days or more from the last stroke, stenting during emergent thrombectomy as a rescue procedure may carry a higher risk.³⁴ Similarly, intensive antiplatelet therapy after acute cerebral ischemia may increase the hemorrhagic risk.³⁵ Nevertheless, the risk-benefit profile has to be reconsidered in the context of LVO stroke with an underlying ICAS lesion that has a high reocclusion rate, as the degree of reperfusion is a strong predictor of functional outcome.³⁶ This is supported by a recent study by Baracchini et al³⁷ that showed patients with ICAS who received urgent intracranial stenting to rescue a failed LVO thrombectomy had superior functional outcome and survival than those whose artery was left occluded. Likewise, the present meta-analysis shows that a high revascularization rate in ICAS-O can be achieved with judicious use of rescue endovascular therapies and that complication rates were not increased. Indeed, as shown in Table 3, there was no significant difference in the symptomatic hemorrhage rate between the ICAS-O group and non–ICAS-O group, and the rate of functional independence or mortality was similar. There is at present no consensus about the antiplatelet management in the acute phase after rescue strategies, such as stenting and angioplasty. In the authors’ center, a computed tomography scan is routinely performed immediately to exclude intracranial hemorrhage after rescue stenting, and intravenous glycoprotein IIb/IIIa inhibitor infusion is commenced. This is then switched to standard oral antiplatelet agents after 24 hours if no major hemorrhagic transformation occurs.

This study has limitations. First, most of the included studies were performed in Asia, and the outcomes and clinical profile of Western patients with ICAS-O may be different. A recent case series of rescue stent angioplasty in patients with LVO with early reocclusion in Germany found less favorable outcomes and a higher rate of symptomatic intracranial hemorrhage was found.³⁸ It is possible that the collateral perfusion status of patients with ICAS-O may be different and contributed to the favorable clinical outcome despite the longer revascularization time. However, the lack of collateral status data in the included articles precludes detail analysis of this factor. The heterogeneity of thrombectomy techniques and rescue therapeutic approaches preclude detail comparison. Finally, the long-term outcomes of restenosis or recurrent strokes after thrombectomy in the ICAS group were not reported in most of the studies included.

Conclusions

ICAS-O is an important and challenging entity and can account for up to a quarter of LVO strokes receiving endovascular treatment. Patients with ICAS-O display a unique risk factor profile compared with non–ICAS-O. There are technical difficulties in endovascular thrombectomy of ICAS-O as evidenced by the longer puncture-to-reperfusion time and high intraprocedural reocclusion rate. Although the optimal rescue treatment remains to be defined, successful revascularization may be achieved by intraarterial glycoprotein IIb/IIIa inhibitors infusion, balloon angioplasty, or intracranial stenting. The final successful reperfusion, favorable functional outcome, and mortality rates were comparable between ICAS-O and non–ICAS-O.

Footnotes