Functional Outcome and Safety of Intracranial Thrombectomy After Emergent Extracranial Stenting in Acute Ischemic Stroke Due to Tandem Occlusions

Background and Purpose: Various endovascular approaches to treat acute ischemic stroke caused by extra- intracranial tandem occlusions (TO) exist: percutaneous transluminal angioplasty with or without emergent extracranial carotid stenting (ECS) due to high-grade stenosis preceded or followed by intracranial mechanical and/or aspiration thrombectomy (MT). Which treatment strategy to use is still a matter of debate. Methods: From our ongoing prospective stroke registry we retrospectively analyzed 1,071 patients with anterior circulation stroke getting endovascular treatment within 6 h of symptom onset. ECS prior to intracranial MT for TO (n = 222) was compared to MT as standard of care (control group; acute intracranial vessel occlusion without concomitant ipsilateral ICA-occlusion or high-grade stenosis [C; n = 849]). Good functional outcome (mRS ≤ 2 at 3 months), mortality rates, frequencies of symptomatic intracranial hemorrhage (sICH) and successful recanalization (Thrombolysis in Cerebral Infarction Score [TICI] 2b or 3) were assessed. In subgroup analyses we tried to detect possible influences of stroke etiology, dual inhibition of platelet aggregation (IPA; clopidogrel [CLO]: n = 83; ticagrelor [TIC]: n = 137; in combination with Aspirin) and intravenous thrombolysis (IVT). Results: Functional outcome was superior in TO (mRS 0–2: 44.6%) when compared with controls (36.0%; OR [95% CI]: 3.49 [1.59–7.67]; p = 0.002). There was no difference in all-cause mortality at 3 months (TO: 21.6%; C: 27.7%; 0.78 [0.47–1.29]; p = 0.324), in-hospital mortality (0.76 [0.45–1.30]; p = 0.324), sICH (TO: 3.2%; C: 5.0%; 0.70 [0.30–1.59]; p = 0.389), and TICI 2b/3 (TO: 89.1%; C: 88.3%; p = 0.813). In subgroup-analysis, TIC and CLO did not differ in functional outcome (TIC: 45.3%; CLO: 44.6%; 1.04 [0.51–2.09]; p = 0.920) and mortality rates (all-cause mortality: TIC: 23.4%; CLO: 16.9%; 0.75 [0.27–2.13]; p = 0.594). sICH was more frequent in TIC (n = 7 [5.1%]) vs. CLO (n = 0; p = 0.048). Conclusion: In our pre-selected cohort, ECS prior to intracranial MT in TO allowed for a good functional outcome that was superior compared to a control population. Mortality rates did not differ. Despite a dual IPA in TO, there was no increase in sICH. CLO and TIC for dual IPA did not differ in terms out outcome and mortality rates. A significant increase in sICH was observed after initial loading with TIC.


INTRODUCTION
Mechanical thrombectomy and/or aspiration thrombectomy (MT) in acute ischemic stroke due to embolic large vessel occlusion is effective and safe (1)(2)(3)(4)(5). Specific recommendations and national guidelines for indication, implementation, and patient selection exist (6). A considerable number of patients present with extracranial-intracranial tandem occlusions (TO; occlusion or high-grade stenosis of an extracranial internal carotid artery [ICA] with a concomitant ipsilateral intracranial large vessel occlusion) (7). Currently, there is a lack of guidance on how to treat those patients. Several endovascular treatment strategies are proposed: percutaneous transluminal angioplasty (PTA) with or without emergent stenting of the extracranial ICA (ECS) preceded or followed by MT (7)(8)(9)(10). Which technique to use is still a matter of debate (11)(12)(13)(14). We report data on TO where initial ECS is followed by MT (extracranial first).

Study Population
Consecutive patients from our prospective single-center stroke registry treated with MT between January 2010 and December 2017 were screened and retrospectively analyzed. Patients with an anterior circulation ischemic stroke caused by an occlusion of the ICA, the carotid-T, an M1-or M2-branch of the middle cerebral artery (MCA) were included. We did not consider distal MCAocclusions, occlusions of the anterior cerebral artery or posterior circulation stroke. Proximal vessel occlusions (on initial imaging) that were found recanalized during angiography (spontaneously or as an effect of intravenous thrombolysis [IVT]) were removed from further analysis. Patients treated after 6 h of symptom onset or presenting with wake-up stroke or unknown symptom onset were excluded. We did not include cases of primary stentangioplasty without MT (due to high-grade intra-or extracranial stenosis or dissection). Datasets without a 3-month follow-up as well as datasets including inconsistent information that could not be confirmed were excluded. Local institutional review board approval was obtained.
Patients either presented primarily in the emergency department of our neurovascular center or via hospitals within or surrounding the city of Stuttgart (secondary transfer) (15). Irrespectively, endovascular therapy was based upon the initial intention to treat patients (based on a shared decision-making concept including stroke-neurologists and interventional neuroradiologists) without further triage or additional imaging procedures prior to the intervention. General anesthesia was performed on a regular basis.
Information on baseline characteristics, current medication, symptom onset, stroke severity (e.g., National Institutes of Health Stroke Scale [NIHSS], modified Rankin Scale [mRS]) or periprocedural information (e.g., Thrombolysis in Cerebral Infarction Score [TICI]) were extracted from admission notes, internal documentation, referral, or discharge papers. Imaging modality and imaging times were stored in our Picture Archiving and Communication System. Follow-up information were collected by our study nurse (via telephone calls).
TO were defined as an occlusion or high-grade stenosis (NASCET [North American Symptomatic Carotid Endarterectomy Trial] >70%) of the extracranial ICA with a concomitant ipsilateral occlusion of the intracranial ICA, the carotid-T or the MCA (M1 or M2 branch). We used the extracranial first approach (emergent extracranial stenting followed by intracranial MT). The following patients defined the control group: (1) embolic intracranial large vessel occlusion; (2) absence of a high-grade intra-or extracranial stenosis requiring emergent stenting; (3) treatment within 6 h of symptom onset; (4) MT only without additional or primary stent application.
In subgroup analyses we tried to detect possible differences in TO due to etiology or medical treatment. (1) Baseline characteristics and outcome in extracranial atherosclerosis (LAD), dissection and in patients with competing etiologies (LAD/CE; both cardiac embolism [e.g., atrial fibrillation] and LAD being possible etiologies) were compared. All patients with atrial fibrillation were summarized in LAD/CE unless a definite cause of stroke (e.g., LAD) could be determined. (2) Loading. Prior to emergent stenting, aspirin (500 mg IV) in combination with either clopidogrel (CLO; 600 mg PO via a nasogastric tube) or ticagrelor (TIC; 180 mg PO via a nasogastric tube) was given to inhibit platelet aggregation. The choice of the respective drug was at the discretion of the interventional neuroradiologist.
Due to a faster inhibition in platelet aggregation (IPA), TIC is currently preferred if emergent stenting is required (16). To secure an immediate IPA (until the expected effect of TIC or CLO), we implemented a bridging concept with the glycoprotein IIb/IIIa inhibitor eptifibatide (given as a single body weightadapted IV bolus) (17). Post intervention, there was strict blood pressure control (systolic blood pressure <130 mm Hg) for a minimum of 3 days. Outcome and hemorrhagic complications in TIC and CLO were compared. After initial loading, dual IPA was continued for a minimum of 3 months (CLO [75 mg/day] or TIC [90 mg twice a day] in combination with aspirin 100 mg/day) followed by monotherapy with aspirin lifelong. Platelet function was assessed with Multiplate R and/or VerifyNow R tests. In case of incomplete IPA, CLO was replaced by TIC. Anticoagulation (in case of atrial fibrillation) was begun depending on stroke severity (NIHSS) and the size of the infarcted tissue (18). (3) Effects of IVT in both the control group and TO on outcome and hemorrhagic complications. (4) Potential factors influencing outcome and mortality in both controls and TO.

Outcome Measures
Good functional outcome (mRS 0-2 at day 90) was the primary outcome parameter. Secondary outcome measures were: (1)

Statistical Analysis
Numerical baseline characteristics were described in mean (standard deviation). Categorical baseline parameters were described in frequencies. Comparing groups, the Fisher's exact test (categorical parameters), the Kruskal-Wallis-test or the Mann-Whitney-U-test (numerical parameters) were used as appropriate. Analyzing more than two groups [e.g., subgroup analysis (1)] outcome in the respective group (e.g., LAD) was compared to the outcome in the remaining groups (e.g., LAD/CE and dissection). A multivariate logistic regression model (considering possible confounders [based on literature research; p < 0.05 in baseline characteristics]) tried to detect factors influencing outcome and mortality. A p-value below 0.05 was considered statistically significant. Stata/IC 13.1 for Windows (StataCorp LP, College Station, Texas, USA) was used for statistical analysis.

RESULTS
Between 2010 and 2017, n = 2450 acute ischemic stroke patients received endovascular recanalization therapy by us. N = 1071 (43.7%) met the predefined inclusion criteria, n = 1379 (56.3%) had to be excluded from further analysis (Figure 1). N = 222 patients were included in TO, n = 849 in the control group (C).
In univariate analysis in controls, good functional outcome was observed in 40.9% of patients treated with IVT prior to endovascular therapy vs. 33 Table 4.

DISCUSSION
The main finding of our study was that outcome after endovascular therapy in TO using the extracranial first approach was shown to be superior compared to controls (patients with an acute ischemic stroke due to intracranial large vessel occlusion without concomitant high-grade extracranial stenosis of the ipsilateral ICA). Mortality rates did not differ. Despite the need of dual IPA after emergent stenting in TO, we did not observe an increase in sICH when compared to controls. Evidence on endovascular treatment strategies in TO is mixed and inconsistent (7,8,10). Some authors suggest intracranial MT to be done before ECS (intracranial first) in order to minimize the time of critical hypoperfusion (8,14). Currently, this seems to be the most widespread approach (9). However, ECS requires dual IPA which-especially in acute ischemic stroke-might increase the rate of hemorrhagic complications. Therefore, it is argued that a PTA should be performed in the acute setting while ECS in general (or vascular surgery) has to follow later during the course of the disease (9). Besides in an increase in (asymptomatic) SAH we did not observe additional hemorrhagic complications in our cohort. Most importantly, dual IPA did not increase the frequency of sICH. A third proposed treatment option is extracranial first (7,9,10). Immediately recanalizing the proximal ICA might improve or uphold crucial collateralization (21,22). Asymptomatic ICA-stenosis can be compensated through a change in cerebral blood flow via existing or established collaterals (e.g., the Circle of Willis). A consistent reduction in cerebral perfusion (due to ICA-stenosis) might induce additional collaterals. Atherothrombotic stroke therefore is said to have greater collateral recruitment compared to stroke due to other etiologies (23). Together with possible effects of ischemic preconditioning of the brain, this could have a positive effect on outcome in TO (24). Yet, in progressive ICA-stenosis or acute occlusion, those compensation mechanisms might fail (25-29). In our cohort, time to recanalization was significantly longer in TO. Unlike the control population, where an association between symptom-onset to groin puncture time and good functional outcome was shown, symptom-onset to groin puncture time (and time to recanalization) did not significantly influence outcome in TO. This might also emphasize the importance of pre-existing collaterals. So far, superiority of any endovascular approach in TO could not be demonstrated (10). An average symptom-onset to groin puncture time of 3.6 h (TO; 3.7 h in controls) is attributed to the high number of secondary referrals (15).
The use of dual IPA in TO did not lead to an increase in symptomatic intracranial hemorrhages. The frequency of asymptomatic SAH (as seen on follow-up imaging) was significantly higher in TO. CLO is widely used in dual IPA after coronary, extra-or intracranial stenting. After loading, it takes 6 to 12 h until a sufficient inhibition in platelet activity is established (16). IPA in TIC is established after about 3 h (16). A faster IPA might reduce early complications (e.g., stent thrombosis) and therefore add an extra benefit. However, especially in dual IPA, TIC might introduce additional hemorrhagic complications (30,31). From 2016/2017 onwards, we loaded patients with TIC and aspirin on a regular basis. Comparing CLO and TIC, we did not observe differences in outcome. Similar to previous data, there was a significant increase in major and minor hemorrhagic complications (sICH, SAH) after pre-medication with TIC (30,31).
In our dataset, outcome after cervical ICA-dissection and LAD was superior compared to patients with LAD/CE (both LAD and cardioembolic etiology possible). Mortality in dissection was significantly lower. Favorable outcome in dissections-patients are younger without a typical cardiovascular risk profile-is wellknown (32). In contrast, patients with cardioembolic stroke due to atrial fibrillation are older, stroke is more severe and has a higher risk of recurrence (33). Outcome in cardioembolic stroke, even after MT in a reasonable time window-is inferior to stroke caused by other etiologies (34,35). The combination of LAD (or vascular disease in general) and atrial fibrillation has a further negative effect on functional outcome (36,37). Other potential factors such as contraindications (e.g., major stroke and necessity of anticoagulation) or triple therapy are open for discussion.
Age, NIHSS and successful reperfusion have been identified as prognostic parameters in TO before (20). In our cohort, age and NIHSS showed a similar correlation in both TO and controls. TICI 2b/3 was associated with good functional outcome and a reduction in all-cause mortality in controls. In TO, TICI 0-2a was a perfect predictor for a poor outcome as all patients with TICI 0-2a got mRS 3-6 after 3 months. IVT is said to be associated with successful reperfusion in patients with TO (38,39). In univariate analysis, we observed better functional outcome and reduced allcause mortality in controls treated with IVT prior to MT (but not in TO). In TO, IVT was safe and did not lead to an increase in sICH. As shown before, we observed a beneficial effect of high cholesterol levels (in TO) on short-term outcome (40).
The main limitation of the study is its retrospective design. There is no information on patients where MT was not considered which might introduce selection bias. A certain inconsistency in decision-making has to be expected. Small sample sizes in subgroup analyses might lead to a power problem resulting in a potential underestimation of therapeutic or predictive effects. As there was no common study protocol, there is no information on potential contraindications or off-label decisions (e.g., IVT, dual IPA, individual healing attempts).

CONCLUSION
Endovascular therapy in acute ischemic stroke due to TO using the extracranial first approach allows for a good functional outcome. In our cohort, dual IPA after emergent stenting in acute ischemic stroke patients was safe and did not lead to an overall increase in the frequency of sICH. Sufficiently powered randomized-controlled trials are needed for direct comparison of the different therapeutic strategies. In subgroup analysis focusing on dual IPA, TIC, and CLO did not differ in terms of outcome and mortality. However, TIC lead to a significant increase in sICH and SAH when compared to CLO. IVT-admission did not cause additional hemorrhagic complications.

DATA AVAILABILITY STATEMENT
All relevant data is contained within the manuscript.

AUTHOR CONTRIBUTIONS
PB: study concept and drafting of the manuscript. All authors contributed to the acquisition, analysis and interpreting of data, and to the critical revision and final approval of the manuscript.