Introduction
Endovascular therapy (EVT) is the well-established and guideline-recommended treatment for patients with acute ischemic stroke due to intracranial large-vessel occlusion (LVO) [1-3]. While there is excellent evidence on EVT in intracranial LVO [4], evidence on EVT in isolated, i.e., non-tandem cervical internal carotid artery occlusion (c-ICA-O) is sparse, and guideline recommendations are lacking [1,2]. Several observational studies, including the recently published large retrospective Endovascular Treatment for Isolated Cervical Internal Carotid Artery Occlusion study (ETIICA) and Acute Carotid Occlusion BelOw Circle of Willis study (ACOBOW) cohorts, favored neither EVT nor best medical therapy (BMT) [5-7]. The effectiveness of EVT may be influenced by various procedural techniques, e.g., emergent carotid artery stenting (CAS), which is performed in approximately half of patients with isolated c-ICA-O [5-9]. CAS is associated with a favorable technical and clinical outcome in patients with tandem lesions (i.e., intracranial LVO with concomitant extracranial occlusion or highgrade stenosis) [10,11]. However, these positive results are hampered by high rates of any intracerebral hemorrhage (ICH) (23%- 48%) [11-13] and early stent re-occlusion (17%-21%), the latter of which is associated with worse outcomes [11,14,15]. In patients with isolated c-ICA-O, data on risk and benefit of emergent CAS during EVT are scarce [16]. We hypothesized that CAS may improve technical and clinical outcomes in patients undergoing EVT for isolated c-ICA-O without compromising safety. To test this assumption, we performed a subgroup analysis of the large, multinational, retrospective cohort study ETIICA.
Methods
Study design and population
This study comprised a subgroup of the ETIICA study, an investigator-initiated, multinational, retrospective cohort study conducted at 42 sites in Europe and North America. Its methods have been described in detail previously [5]. ETIICA included consecutive patients with acute ischemic stroke and ipsilateral isolated c-ICA-O admitted to the participating hospitals between January 2018 and December 2022. Briefly, all patients presenting with acute ischemic stroke with symptoms attributable to an isolated ipsilateral c-ICA-O within 24 hours from the time last seen well were eligible for the study [5]. Patients with concomitant intracranial (tandem) occlusions were excluded [5]. Rigorous imaging assessment was performed by the local centers, using a predefined protocol provided by the lead investigators to ensure the exclusion of patients with concomitant intracranial occlusions as well as carotid pseudo-occlusions. Details on imaging assessment, as well as inclusion and exclusion criteria for computed tomography angiography (CTA), magnetic resonance angiography (MRA), and digital subtraction angiography, have been reported previously [5]. The original ETIICA cohort included patients with either BMT alone or EVT and BMT (with or without intravenous thrombolysis [IVT], respectively).
In the present sub-study, only patients who underwent EVT were analyzed. To reduce bias by indication, the population was further restricted to patients with larger-artery atherosclerosis (LAA) and dissection, as CAS is typically performed in these patients, whereas it is rarely indicated for other etiologies. We compared patients who underwent carotid stenting +/- angioplasty during EVT (CAS) with those who did not (NO-CAS). Patients who underwent carotid angioplasty without subsequent stent-ing were analyzed separately as they constituted a distinct entity. These data are presented in the Supplementary Material. Treatment techniques, selected devices, periprocedural medications, and immediate antithrombotic regimens after EVT were performed according to the treating physicians’ local standards of care. This observational study was conducted in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines.
Data availability
Anonymized data for this article will be made available upon reasonable request from a qualified investigator.
Study variables
We analyzed demographics, vascular risk factors, comorbidities, pre-existing medication (oral anticoagulants, antiplatelets, statins), stroke severity at admission and at 24 hours (assessed by the National Institutes of Health Stroke Scale [NIHSS] score), systolic and diastolic blood pressure at admission, initial imaging modality, and baseline Alberta Stroke Program Early CT Score (ASPECTS) [6]. One point was deducted if ASPECTS was assessed on magnetic resonance imaging [16]. Additionally, data on 24-to-36-hour follow-up imaging (presence of hemorrhagic transformation/ICH), stroke etiology, mortality, and mRS at 3 months were collected [5]. Stenting strategy (immediate vs. rescue), intracranial embolization, consecutive rescue thrombectomy, and final intracranial extended treatment in cerebral infarction (eTICI) score were assessed [5]. Carotid patency at the end of procedure and at 24 hours was recorded in three categories according to the North American Symptomatic Carotid Endarterectomy Trial (NASCET) criteria: (1) vessel patency (VP), defined as unrestricted perfusion or stenosis <70%, (2) stenosis >70%, or (3) complete occlusion [17,18]. Intraprocedural and postprocedural antithrombotic regimens, such as single antiplatelet therapy, dual antiplatelet therapy (DAPT), and glycoprotein (GP) IIb/IIIa inhibitors, were assessed.
Outcome measures
The primary outcomes included the following: (1) c-ICA VP on follow-up imaging (CTA/MRA/Doppler sonography) at 24 hours, defined as unrestricted perfusion or stenosis <70%; and (2) symptomatic ICH (sICH), defined as a parenchymal hemorrhage with an increase in the NIHSS score by at least 4 points [19]. As additional safety outcomes, we assessed (1) any ICH (including hemorrhagic transformation); (2) intracranial embolization during EVT; and (3) early neurological deterioration (END), defined as an increase in the NIHSS score of at least 4 points between admission and 24-hour follow-up. Functional status (mRS ordinal shift) and mortality at 3 months were assessed as secondary clinical outcomes.
Ethics
All participating centers anonymized their data before sending them to a coordinating center (Department of Neurology, Charité-Universitätsmedizin Berlin, and Center for Stroke Research Berlin, Germany). Each center was responsible for obtaining ethical approval for data collection and sharing in accordance with local rules and regulations. The requirement for informed consent was waived because of the retrospective nature of this study. This study was conducted in accordance with the principles of the Declaration of Helsinki. Statistical analysis was performed on anonymized data by a statistician (P.G.) who did not participate in data collection or interpretation [5].
Statistical analysis
Data are presented as means±standard deviations (SDs) or as medians with interquartile ranges (IQRs) for continuous variables, depending on the distribution, and as absolute numbers and percentages for categorical variables. We compared the baseline characteristics between patients with and without CAS using a chi-square test for categorical variables and an independent ttest, Mann-Whitney U test, or Kruskal-Wallis test for continuous variables, as appropriate. The inverse probability of treatment weighting (IPTW) method was applied to adjust for unbalanced baseline characteristics between the CAS and NO-CAS groups. The weighted propensity scores were estimated using a logistic regression model including the following clinically preselected variables: age, sex, pre-stroke mRS score, NIHSS score at admission, time from last-seen-well to hospital admission, ASPECTS, and IVT. The balance of baseline characteristics before and after IPTW adjustment was evaluated using the standardized differences approach. Standardized differences <±0.1 were considered to indicate an adequate balance between the CAS and NO-CAS groups [20]. For binary outcomes, logistic regression analysis was performed. The distribution of the mRS categories was assessed using shift analysis (ordinal logistic regression), with common odds ratios of >1 indicating a better functional outcome associated with CAS. All models were adjusted using IPTW. Additional adjustments were made for unequally distributed factors, namely, diabetes mellitus and perfusion-based imaging upon hospital admission (yes vs. no). To assess the association between functional outcome (mRS at 3 months) and (1) c-ICA VP at 24 hours and (2) re-occlusion after successful EVT within 24 hours, we conducted ordinal regression analyses, adjusting for the above-mentioned variables. To address potential bias by indication related to withholding CAS in patients considered at high risk for ICH, we defined the following high-risk subgroups, in which technical, safety, and clinical outcomes were examined separately in sensitivity analyses: age >75 years, IVT, oral anticoagulation, diabetes mellitus, admission systolic blood pressure >160 mm Hg, and admission blood glucose >10.0 mmol/L. Additionally, we accounted for etiology-specific aspects of CAS by reporting the details of patients with atherosclerosis and dissection as separate groups (supplementary sensitivity analyses). Statistical testing was conducted within an exploratory framework at a two-sided significance level of α=0.05; therefore, no multiple correction was done. Statistical analyses were performed using Stata MP/18 (StataCorp, 2023, College Station, TX, USA).
Results
Of the 998 patients included in the ETIICA study, 487 underwent EVT, of which 341 had an LAA or dissection stroke etiology. Of these, 24 underwent balloon angioplasty only (without subsequent stenting) and were excluded from the outcome analysis. The final sample consisted of 317 patients, with 219 patients (69.1%) who underwent emergent CAS and 98 patients (30.9%) in whom CAS was not attempted (NO-CAS) (Figure 1).
Baseline characteristics
Mean age was 68.6 years (SD, 12.9), 85 patients (26.8%) were female, and 279 patients (88.0%) were independent before the index stroke (pre-stroke mRS 0-2). The median NIHSS was 11 (IQR, 6-17), and the median ASPECTS was 10 (IQR, 9-10). Age, sex, pre-stroke dependency, stroke severity (NIHSS), and early ischemic changes (ASPECTS) were balanced between the CAS and NO-CAS groups. Stroke etiology was LAA in 83.0% and dissection in 17.0% without any difference between groups. Diabetes mellitus was more common in patients with CAS (32.0% vs. 20.4%, P=0.04). Additionally, patients with CAS were more likely to undergo magnetic resonance imaging (13.7% vs. 5.1%, P=0.04) and perfusion-based imaging (44.8% vs. 32.7%, P=0.04) upon hospital admission. All other baseline variables did not differ between groups and are detailed in Table 1.
Procedural variables
IVT was administered to just over 40% of patients in both the CAS and NO-CAS groups. In patients undergoing CAS, microwire clot disruption was more common at first EVT pass (47.0% vs. 25.0%), whereas aspiration was the predominant treatment approach in the NO-CAS group (62.5% vs. 37.4%).
Stent re-occlusion during EVT occurred in 21 patients (9.6%) who underwent CAS. Of these, only 8 (38.1%) were successfully recanalized afterward. At the end of EVT, c-ICA VP (defined as no residual stenosis or residual stenosis <70% [NASCET]) was achieved in 201 patients (91.8%) who underwent CAS and 43 patients (43.9%) who did not (p<0.01). In the CAS group, early antithrombotic therapy during or shortly after EVT (<24 h) more often included GP IIb/IIIa inhibitors (+/- other antiplatelets) (31.3% vs. 7.1%) or DAPT (47.9% vs. 19.4%). Single antiplatelet therapy was more common in the NO-CAS group (15.2% vs. 40.8%, P<0.01). In supplementary analyses evaluating etiologyspecific aspects of CAS, the rate of stent re-occlusion during EVT was numerically higher in patients with dissection (15.8% vs. 8.3%, P=0.15). Consistently, the rate of c-ICA VP at the end of EVT was numerically lower in patients with dissection than in those with LAA (86.8% vs. 92.8%, P=0.20). For further technical details on CAS with regard to stroke etiology, see Supplementary Table 1.
Co-primary outcomes
At 24 hours, follow-up imaging was performed with CTA in 20.6%, MRA in 11.5%, and Doppler ultrasound in 67.8% of patients; no differences were observed between groups (Table 1). c-ICA VP was present in 83.5% of patients who underwent CAS and 40.7% of those who did not (adjusted odds ratio [aOR], 9.45; 95% confidence interval [CI], 4.91-18.17; P<0.01). The odds of sICH at 24 hours did not increase, with rates of 2.3% (CAS) and 3.1% (NO-CAS) (aOR, 0.92; 95% CI, 0.18-4.73; P=0.92) (Table 2).
Secondary outcomes
Any ICH was more common in the CAS group (19.3% vs. 9.3%; aOR, 2.50; 95% CI, 1.12-5.60, P=0.03). Overall, clinical outcome at 3 months (mRS shift) did not differ between the groups (adjusted common odds ratio, 0.98; 95% CI, 0.62-1.56; P=0.94), and mortality was similar between the CAS and NO-CAS groups (21.8% vs. 17.2%; aOR, 1.18 [0.61-2.30]; P=0.62). The distribution of clinical outcomes across the mRS is shown in Figure 2.
Technical/safety outcomes
Intracranial embolization as an adverse event occurred in 37.0% of patients who underwent CAS and 19.6% of those who did not (aOR, 2.80; 95% CI, 1.51-5.17; P<0.01). Rescue thrombectomy was performed in 60.5% (CAS) and 73.7% (NO-CAS) of patients who experienced embolization (P=0.28). Overall, partial or complete intracranial reperfusion (eTICI 2b/2c/3) was achieved in 91.4% (CAS) and 89.5% (NO-CAS) of patients who experienced embolization (P=0.80). Rates of complete intracranial reperfusion (eTICI 3) in patients with embolization were low in both groups, at 16.0% in CAS and 21.1% in NO-CAS (P=0.60). Despite higher intracranial embolization during CAS, rates of END did not differ, reaching 22.0% in the CAS and 17.3% in the NO-CAS group (aOR, 1.28; 95% CI, 0.67-2.44; P=0.45).
Subgroup analyses
Patients who underwent CAS were divided according to the use of GP IIb/IIIa inhibitors. Sixty-eight patients (31.1%) received GP IIb/IIIa inhibitors during or shortly after EVT, whereas 151 patients (68.9%) did not. Patients treated with GP IIb/IIIa had a higher proportion of VP at 24-hour follow-up (91.9% vs. 79.7%; aOR, 3.02; 95% CI, 1.08-8.42; P=0.04). The rates of symptomatic or any ICH at 24 hours did not differ. Further details are provided in Table 3.
Twenty-four patients underwent balloon angioplasty, without subsequent stenting. Full VP or a stenosis <70% at 24 hours was present in 42.9% of patients; no sICH occurred. Further descriptive and outcome data are presented in Supplementary Tables 2 and 3.
Sensitivity analyses
Regarding stroke etiology, technical efficacy of CAS (VP at 24 hours) was shown for patients with LAA (85.5% vs. 41.4%; aOR, 8.26; 95% CI, 4.26-16.02; P<0.001) and those with dissection (73.5% vs. 37.5%; crude odds ratio, 4.63; 95% CI, 1.30-15.43; P=0.02). Rates of sICH did not differ between groups (LAA: 2.2% vs. 2.5%, aOR=0.81, 95% CI=0.14-4.57, P=0.81; dissection: 2.6% [CAS] vs. 6.3% [NO-CAS], crude odds ratio=0.41, 95% CI=0.02- 6.91, P=0.41). See Supplementary Table 4 for other outcomes. Sensitivity analyses in subgroups of patients who might have been withheld CAS due to an expected high bleeding risk under aggressive antithrombotic therapy (age >75 yr, IVT, oral anticoagulation, diabetes mellitus, admission systolic BP >160 mm Hg, and admission blood glucose >10.0 mmol/L) yielded consistent results for all outcomes (Supplementary Table 5).
VP at 24 hours and clinical outcome
When comparing patients with and without VP at 24 hours, VP was associated with improved clinical outcome (mRS shift; adjusted common odds ratio, 2.55; 95% CI, 1.53-4.24; P<0.01). See Supplementary Figure 1 for details on mRS distribution.
Discussion
In this study, we investigated the safety and efficacy of emergent CAS during EVT for isolated cervical ICA occlusion and report the following findings. First, CAS was technically effective, with 84% of patients reaching VP at 24-hour follow-up (compared with 41% of those who did not undergo CAS). Second, CAS appeared safe as the rates of sICH did not differ between the groups. Third, although distal embolization occurred frequently, CAS was not associated with a worse final intracranial reperfusion or END. Fourth, CAS was not associated with functional outcomes at 3 months.
VP is fundamental to the success of EVT and should be permanent. However, re-occlusion of carotid lesions is common and associated with sICH and worse functional outcomes [14,15]. Reocclusion can occur during EVT, shortly thereafter, or even later in the clinical course [17]. Therefore, re-occlusion rates should be considered initially and throughout the postprocedural period. In our study, 9.6% of patients who underwent CAS experienced stent re-occlusion during EVT, with only one-third of these patients being successfully recanalized afterward. While stent reocclusion rates during EVT were similar in a multicenter study from Korea on CAS in tandem occlusion (8.9%), final carotid patency at the end of EVT was achieved more frequently in this study (94.6%) than in ours (91.8%) [21]. The relatively low rates of re-recanalization after stent re-occlusion that we report most likely reflect the technical challenges of EVT in isolated c-ICAO and highlight the need for future procedural improvements. Within the first 24 hours after EVT, stent re-occlusion occurred in another 8% of patients who underwent CAS. Notably, c-ICA re-occlusion was also observed in 3% of patients who did not undergo CAS, indicating that c-ICA re-occlusion is not exclusively a concern for patients who underwent CAS.
VP at 24 hours was observed in 83.5% of patients who underwent CAS, but only in 40.7% of those who did not. These results favor the use of CAS in terms of technical efficacy. Notably, our study reports higher rates of persistent stent patency in patients receiving antithrombotic treatment with GP IIb/IIIa inhibitors, consistent with previous observational reports [22]. However, the use of more aggressive antithrombotic treatment may, on the other hand, go along with higher sICH rates. In our study, CAS as such was not associated with sICH (2.3% vs. 3.1% in the NO-CAS group). Comparative data on the risk of bleeding associated with CAS in patients with isolated c-ICA-O are lacking. However, our results are consistent with previous studies assessing emergent CAS in tandem occlusions, which similarly did not report safety concerns regarding sICH [10,11,13,23]. In contrast, the rates of any ICH were markedly higher in patients who underwent CAS than in those who did not in our study. This difference may be due to the more aggressive early antithrombotic treatment (DAPT or GP IIb/IIIa inhibitors) that the patients who underwent CAS received. Although data on the impact of CAS on any ICH in isolated c-ICA-O are lacking, increased rates of any ICH have been described in tandem occlusion (48% vs. 43% without CAS) [11].
Our study complemented these safety outcomes by reporting the rates of intracranial embolization. Embolization occurred in approximately 30% of our patients, corroborating two previous studies that reported 22% and 28% embolization during EVT for isolated c-ICA-O [7,24]. Embolization occurred more frequently in the CAS group (37% vs. 20%), possibly because of increased shear forces during stent deployment or post-stent angioplasty. While embolization was a frequent complication, immediate intracranial rescue thrombectomy could be performed in approximately two of three patients. Final intracranial eTICI 2b/2c/3 was achieved in 9 out of 10 patients in the CAS and NO-CAS groups. This finding is consistent with that of the ACOBOW study, which reported eTICI 2b/2c/3 in 87% of patients with embolization during EVT [7]. The high rates of embolization indicate a possible target for improving therapeutic management of patients with c-ICA-O, either by preventing embolization or by optimizing strategies for subsequent recanalization. Given the results of the recently published Intra-arterial Tenecteplase for Acute Stroke After Successful Endovascular Therapy (ANGEL-TNK) trial, adjuvant intra-arterial thrombolysis may be a promising approach [25].
Despite higher rates of intracranial embolization and any ICH, CAS was not associated with END. The lack of an early clinical impact of complications favors CAS. However, embolization and ICH may have offset the beneficial effect of c-ICA VP on 3-month clinical outcomes. After all, our study did not find an association between CAS and improved clinical outcomes at 3 months. Although CAS per se was not linked to clinical outcomes, we found a strong association between VP at 24 hours and functional recovery. This finding supports the notion that the higher rates of ICH and embolization may outweigh the clinical benefits of VP resulting from CAS. However, the strong association between VP and recovery underscores the importance of achieving and maintaining carotid patency. Technical improvements may reduce complications (embolism, re-occlusion, and bleeding) and form the basis for a future correlation between CAS and clinical outcomes.
Given the paucity of data on isolated c-ICA-O, we primarily compared our results with those of studies on tandem occlusion to provide a reference framework. However, patients with isolated c-ICA-O represent a distinct stroke subpopulation, differing from tandem occlusions in terms of their clinical presentation, pathophysiology, and treatment rationale. Although initial stroke severity in isolated c-ICA-O is typically lower than that in tandem occlusion (11-13 points [7,9] vs. 14-17 points [10,11,13]), clinical presentation is often unpredictable, and approximately one in six patients deteriorates early despite receiving BMT [26]. This instability reflects the pathophysiological dependence of cerebral perfusion on collateral flow through the Circle of Willis, which varies substantially between individuals. Consequently, assessment of collateral patency is crucial for understanding symptom causality and guiding therapeutic decisions. Advanced multimodal diagnostics such as perfusion imaging or transcranial ultrasound may help assess the risk of deterioration and guide the choice between a “watch-and-wait” strategy and immediate endovascular treatment. Noninvasive strategies to preserve or augment collateral flow, such as blood pressure elevation, may be considered alternative or adjunctive measures. We suggest that future trials consider these factors when defining eligibility criteria.
To the best of our knowledge, this is the first study to investigate the effects of emergent CAS during EVT in patients with isolated c-ICA-O. The advantages of our study include the large patient population, multinational-multicenter design, meticulous application of definitions, and a thorough statistical workup, including IPTW.
However, limitations should be considered. The acquisition of retrospective data may have facilitated selection and information bias. The observational design may have introduced bias by indication, resulting in divergent treatment decisions, which, in turn, led to substantial differences in VP at 24 hours. Stenting might have been conducted mainly in patients with c-ICA-O presenting with poor collateral circulation, whereas those with stable collaterals may have been less likely to undergo CAS. Consequently, intracranial perfusion at the end of EVT may have been similar between treatment groups, potentially accounting for the lack of improved clinical outcomes with CAS. As the technical outcome of CAS is highly dependent on treatment technique, differences in local standards of care and the absence of an imaging core lab further limit the findings of this study. Technical details of EVT in dissection-related c-ICA-O (e.g., longer stents, flow diverters) could not be obtained, which limits the generalizability of our findings. The lack of data on the use of balloon-guided catheters and embolic protection devices limits the interpretation of the mechanisms underlying intracranial embolization. Further data on the patency and configuration of the Circle of Willis relevant to collateral capacity could help interpret our findings.
Conclusions
Emergent CAS is technically effective and safe for patients undergoing EVT for isolated c-ICA-O. More frequent VP in patients who underwent CAS did not correspond with improved clinical outcomes at 3 months. Further research is needed to identify the optimal treatment for acute isolated c-ICA-O, and an ongoing randomized controlled trial (NCT05832762) will provide valuable insights.
PDF Links
PubReader
ePub Link
XML Download
Full text via DOI
Download Citation
Supplement1
Print
