Introduction Symptomatic internal carotid artery occlusion (ICAO) can lead to neurologic decline, recurrent stroke, and mortality.
Objective We sought to evaluate the safety and feasibility of endovascular revascularization for ICAO without tandem intracranial large vessel occlusion (LVO).
Design, setting, and participants This is a retrospective cohort analysis of all patients presenting to a single academic center with ischemic stroke and ipsilateral cervical ICAO from November 2003 through April 2016. Patients were excluded if pre-procedural angiography demonstrated tandem LVO or if patients were known to have chronic ICAO.
Main outcome(s) and measure(s) Study endpoints included discharge neurologic examination, post-procedural infarct burden, 3-month functional outcomes, and treatment durability.
Results A total of 107 patients with symptomatic angiographically-confirmed cervical ICAO without tandem LVO were identified. Median admission NIH Stroke Scale (NIHSS) score was 8 (IQR 11). Baseline radiographic stroke severity was assessed by ASPECT score (median 9; IQR 2), perfusion mismatch (present in 93%), and clinical imaging mismatch (42%). Median time from symptom onset to treatment was 25 hours (IQR 61). Successful revascularization was achieved in 92% of patients. At discharge, 83% had stable/improved NIHSS score, while at 3 months 65% achieved independence (modified Rankin Scale score ≤2). The most common complication was distal embolization (22%) of which 16% required intra-arterial treatment. Rate of significant restenosis (≥70%) was 15% at 1 year.
Conclusions Stenting in selected patients at risk of neurologic deterioration due to symptomatic ICAO can be performed with high rates of technical success and good clinical outcomes. Because of significant peri-procedural risks and high rates of restenosis, randomized studies are necessary to understand the benefit of this approach.
- acute stroke
- carotid occlusive disease
- angioplasty and stenting
- cerebral revascularization
Statistics from Altmetric.com
Symptomatic internal carotid artery occlusion (ICAO) without associated large vessel intracranial occlusion (LVO) is protean in its neurologic manifestation and outcome. Some patients with symptomatic ICAO present with mild to moderate deficits demonstrate little to no cerebral perfusion abnormalities, and often remain neurologically stable with low risk of recurrence.1 2 Others, however, exhaust their cerebrovascular reserve and present with either sudden severe ischemic symptoms without corresponding infarct (ie, clinical infarct mismatch) or neurologic instability characterized by recurrent transient episodes and/or progressive symptomatology. The prognosis for these patients is poor, with higher rates of persistent neurologic deficits and long-term disability as well as being at greater risk of recurrent stroke.1–3
Treatment of symptomatic ICAO has proven to be a significant challenge. Intravenous (IV) thrombolysis remains the mainstay of acute therapy; however, a minority of patients present within the eligible time frame. Moreover, thrombolysis has demonstrated poor recanalization for LVOs and has been associated with an increased risk of symptomatic intracranial hemorrhage (sICH) and death.4 Additional medical strategies have targeted augmenting collateral blood flow and/or inhibiting clot propagation through induced hypertension and anticoagulation, respectively. Surgical revascularization of ICAO (ie, carotid endarterectomy or intracranial/extracranial bypass) has similarly failed to show benefit.3 5
Alternatively, endovascular stenting and angioplasty for ICAO has been described as a feasible and relatively safe therapeutic alternative in this patient population.6–10 We sought to evaluate the safety and feasibility of endovascular revascularization for the treatment of acute and subacute symptomatic ICAO without tandem LVO.
This retrospective study was performed with IRB approval and subjects and/or their surrogate provided informed consent. The data used for this study are available from the corresponding author on reasonable request. All patients presenting with symptomatic ICAO to our institution between November 2003 and April 2016 were retrospectively identified from a prospectively maintained database. Patients were included for analysis if conventional catheter-based digital subtraction angiography (DSA) demonstrated ICAO without tandem intracranial LVO. Occlusion was defined as complete absence of anterograde flow in the ipsilateral extracranial ICA despite delayed filming. Absence of concomitant intracranial LVO was determined by review of both DSA and non-invasive pre-procedural imaging (magnetic resonance angiography (MRA) and/or computed tomography angiography (CTA)).
Clinical stroke severity was measured by the National Institutes of Health Stroke Scale (NIHSS) at baseline and daily through hospital discharge. Admission and post-procedural CT and MRI were reviewed; the severity of ischemic changes was graded using the Alberta Stroke Program Early CT Score (ASPECTS) and/or diffusion-weighted MRI ASPECTS.11 12 DSA of ipsilateral and contralateral common carotid arteries as well as the posterior circulation was performed in all patients.
Revascularization: patient selection and procedure
The decision to offer revascularization to the patient and/or their surrogate was made by the attending stroke neurologist caring for the patient in conjunction with the neurointerventionalist according to these physicians’ standard care. The clinical indications applied in this multidisciplinary decision process were: (1) sudden onset of severe ipsilateral ischemic symptoms within 24 hours of admission not explained by a corresponding infarct (ie, clinical infarct mismatch); or (2) clinical instability characterized as recurrent transient episodes and/or progressive symptomatology presenting ≥24 hours from ictus despite collateral augmentation therapy (eg, induced hypertension) with or without anticoagulation (therapeutic weight-based dosage of intravenous heparin). Most patients also had evidence of severe perfusion deficits on neuroimaging studies. Patency of the ipsilateral middle cerebral artery (MCA) and anterior cerebral artery (ACA) was part of the decision to proceed with revascularization.
Preoperatively, all patients were loaded with aspirin 325 mg and clopidogrel 600 mg per standard practice at our institution; no alternative systemic antiplatelet agents were used among this sample. Transfemoral arterial access was obtained using a 7 Fr or larger base catheter. Prior to crossing the lesion, patients were given intravenous heparin boluses titrated to achieve an activated clotting time >250. Attempts were made to cross the lesion with a 0.014 inch or 0.018 inch microwire; thereafter a microcatheter was advanced over the wire into the distal petrous ICA. Repeat angiography of the ICA distal to the occlusion was performed to assess the extent of ICA thrombus and to confirm patency of intracranial vessels. In cases where extensive thrombus was identified in the distal cervical, petrous, and/or cavernous ICA segments, strong consideration was given to aborting if manual aspiration thrombectomy failed.
When available, a distal embolic protection device (dEPD; SpiderFx, ev3) was placed distal to the occluded segment. Pre-stenting angioplasty was performed in all stented patients; angioplasty was always carried out following dEPD placement (when a dEPD was used). Stents were sized to the presumed ICA diameter based on distal runs and then ≥1 self-expanding stents were deployed. Post-stenting angioplasty was performed as routine care.
All patients were monitored post-procedure in the neurovascular ICU. Systolic blood pressure was strictly maintained at <140 mm Hg after revascularization. Follow-up imaging (CT and/or MRI) was routinely performed within the first 24 hours after stenting to assess for ICH. Once absence of ICH was confirmed, patients were initiated or continued on a dual antiplatelet regimen consisting of aspirin 81 mg and clopidogrel 75 mg.
Early neurologic recovery was measured by change from baseline NIHSS at 24–48 hours post-procedure. Significant early neurologic improvement was defined as either NIHSS reduction of ≥8 points compared with baseline or post-procedure NIHSS ≤2 regardless of the baseline score.13 Significant neurologic worsening was defined by a ≥2 point increase from baseline NIHSS (consistent with CREST criteria).14 Worsening of radiographic stroke burden was defined as a decrease in ASPECTS of ≥1 point on post-treatment CT or MRI. Functional recovery was assessed at 90 days, and favorable functional recovery was defined as a modified Rankin Scale score (mRS) of 0–2. Treatment durability was evaluated and assessed a propos of stent patency, development of significant in-stent restenosis (confirmed to be ≥50% by NASCET criteria on CTA or DSA and/or peak systolic velocity of ≥250 cm/s and ICA/CCA ratio ≥3 on ultrasonography), and need for repeat angioplasty with or without stenting through review of follow-up studies performed at 3, 6, and 12 months, then annually thereafter.15
The incidence of post-procedural complications was also analyzed. These included embolization (MCA or ACA branch occlusion with or without neurologic decline of ≥2 points on NIHSS), sICH (defined by SITS-MOST criteria of parenchymal hemorrhage type 2 with neurologic decline of ≥4 points on NIHSS), and myocardial infarction (defined according to CREST criteria as elevation of troponin beyond the upper limit of normal for the local laboratory plus ECG evidence of ischemia).15 16
Acceptable type I error was set a priori at α=0.05. Continuous demographic characteristics were assessed for normality using the Kolmogorov–Smirnov test; normally distributed data were analyzed by t-test while the remainder were compared using the the Wilcoxon rank-sum test. Categorical data were analyzed with Pearson’s χ2 or Fisher’s exact test. Univariable logistic regression analysis was performed to assess for predictors of technical success, dramatic early neurologic recovery, worsened radiographic stroke burden, and long-term functional recovery. Multivariable logistic regression adjusted for variables identified by univariate analysis to be marginally different between outcomes (the probability value for inclusion was set at P<0.20). Means are presented as mean±SD, medians as median (IQR), and odds ratios as OR; 95% CI. Data were analyzed using Stata Version 14 (StataCorp, College Station, Texas, USA).
The baseline clinical characteristics for these patients are shown in table 1. Over the study period a range of 7–13 patients were treated per annum; 41% of patients had fluctuant or blood pressure-dependent neurologic symptoms. Median pre-procedural radiographic infarct burden measured by ASPECTS was 9 (range 5–10) and 32% of the sample presented with ASPECTS ≤8. Pre-procedural CT or MR perfusion imaging was available for 83 patients, of whom 93% had ipsilateral perfusion deficits evidenced by increased mean transit time with preserved cerebral blood volume (the remaining six patients had symmetric perfusion parameters). Furthermore, 42% of patients had neurologic deficits out of proportion to the radiographic infarct burden (ie, clinical-imaging mismatch (CIM), defined as NIHSS ≥8 and ASPECTS ≥8 or NIHSS ≥4 and ASPECTS=10). Catheter-based angiography demonstrated significant contralateral stenosis (≥70% by NASCET criteria) in 13%, while 12% of the sample harbored bilateral ICA occlusions. In 67% of cases, angiography revealed retrograde filling of the supraclinoid, cavernous, and petrous segments of the occluded ICA (retrograde string sign).
Overall, mean time from symptom onset to endovascular treatment was 34±26 hours; 65% of the sample underwent revascularization ≥24 hours post-ictus. In five patients, a channel through the occluded ICA could not be established. Successful stent deployment was achieved in 88% of all cases and a dEPD was used in 65% of cases. Thirty-nine percent of patients required ≥1 stent to completely cover the target lesion (range 1–4). Reasons for refraining from stent placement included: insurmountable ICA clot burden (4 patients), non-significant residual stenosis after aspiration thrombectomy (3 patients), and intracranial perforation after attempted mechanical thrombectomy of an iatrogenic embolus (1 patient).
The most common procedural complication was distal embolization (22%). Figure 1 is a representative case. Nine of these patients suffered distal small branch (ie, M3) occlusions that were treated with intra-arterial pharmacologic thrombolysis. Significant branch occlusions (eg, A2, M2, or more proximal) occurred in 14 patients and all were successfully recanalized using the mechanical thrombectomy techniques that were clinically available at that particular time. These included (1) first-generation mechanical thrombectomy (Merci retriever system in two patients; Stryker, Kalamazoo, Michigan, USA), (2) manual aspiration thrombectomy alone (seven patients) employing either large (ie, 072 Navien, Medtronic, Irvine, California, USA) or intermediate aspiration catheters (5 Fr or narrower) depending on lesion location, or (3) stentriever medicated aspiration thrombectomy when manual aspiration alone had failed (3 patients). Other intra-procedural complications included intracranial vessel perforation (2 patients) and dissection of the petrous or cavernous ICA segment (4 patients, none were flow-limiting or otherwise symptomatic). Post-procedural complications included symptomatic intraparenchymal hemorrhage (7%), and MI (2%).
Univariable logistic regression was performed to analyze baseline characteristics that might predict successful revascularization including age, comorbidities (hypertension, hyperlipidemia, coronary artery disease, diabetes mellitus), stroke severity (clinical and radiographic), time from symptom onset to revascularization, retrograde string sign, and contralateral stenosis/occlusion. Regression analysis found none of these to be significant predictors of successful revascularization (all P>0.14).
Early neurologic and radiographic outcomes
Analysis of clinical and radiographic outcomes is presented in table 2. Univariable logistic regression showed that younger age (OR 1.05; 95% CI 1.02 to 1.09), less severe baseline stroke symptomatology (NIHSS <8, OR 3.29; 95% CI 1.49 to 7.14), and smaller baseline infarct burden (ie, higher ASPECTS, OR 1.77; 95% CI 1.16 to 2.71) were significantly associated with dramatic early neurologic improvement. CIM, absence of contralateral ICA disease, and successful revascularization all demonstrated a trend toward significance and were included in adjusted analysis. In multivariable analysis, only younger age (OR 1.06; 95% CI 1.02 to 1.12) and smaller infarct burden (OR 3.28; 95% CI 1.42 to 7.57) remained significantly associated with dramatic early improvement.
Risk factors associated with new or worsened radiographic stroke burden included more severe baseline stroke symptomatology (NIHSS ≥8, OR 3.50; 95% CI 1.31 to 9.36), pre-procedural absence of retrograde string sign (OR 3.57; 95% CI 1.39 to 9.09), and distal embolization (OR 14.1; 95% CI 4.49 to 44.2). Following adjustment for these, as well as acuity of presentation, CIM, contralateral ICA occlusion, and failure to deploy a stent (all marginal on univariable analysis, P=0.17, 0.09, 0.09, and 0.12 respectively), only distal embolization remained significantly associated with new or worsened radiographic stroke burden (OR 16.3; 95% CI 3.85 to 69.4). Post hoc sensitivity analyses showed that use of a dEPD did not significantly decrease the risk of embolic events (P=0.51); however, the use of an embolic protection device was associated with a significantly decreased risk of new or enlarging infarcts (P<0.01).
Long-term functional outcomes and durability
Functional outcome at 3 months was available for the entire sample; 65% of patients achieved functional independence (mRS 0–2; n=69/107 patients). Baseline and procedural characteristics were assessed for prediction of 3-month functional independence using univariable logistic regression. Characteristics shown to be associated with early neurologic recovery were also significant predictors of long-term functional outcome including younger age (OR 1.04; 95% CI 1.01 to 1.08), less severe baseline stroke symptomatology (NIHSS <8, OR 5.32; 95% CI 2.18 to 13.0; higher ASPECTS, OR 2.43; 95% CI 1.46 to 4.04), retrograde string sign (OR 2.29; 95% CI 1.01 to 5.29), and absent embolization (OR 2.28; 95% CI 1.42 to 3.64). Through multivariable analysis adjusting for these as well as marginal predictors (CIM and successful stent deployment, both P<0.20), only NIHSS <8 (OR 4.17; 95% CI 1.08 to 16.7), higher ASPECTS (OR 3.17; 95% CI 1.32 to 7.62), and absent embolization (OR 3.70; 95% CI 1.03 to 11.1) remained significant. Post hoc sensitivity analyses showed that, regardless of admission stroke severity (NIHSS and ASPECT) and/or distal embolization, long-term functional independence was independently associated with dramatic early neurologic improvement (OR 5.95; 95% CI 1.58 to 22.5), lower ASPECTS (OR 15.2; 95% CI 2.11 to 10.8), and embolization (OR 10.9; 95% CI 1.66 to 17.2). On multivariable regression, older age was the only significant factor associated with mortality (OR 1.10; 95% CI 1.03 to 1.17) adjusting for NIHSS ≥8, ASPECTS, CIM, time to recanalization, and distal embolization.
Treatment durability could be assessed in 65 patients. The median duration of imaging follow-up was 18 months (IQR 33; overall range 1.5–135). Imaging modalities used included ultrasonography (60%), DSA (31%), or CTA (9%). Long-term ICA and stent patency was 92%.
Post hoc analyses were performed to evaluate whether treatment outcomes varied across study epochs. Revascularization rates varied significantly between epochs; however, counterintuitively revascularization rates were lower during the most recent 3 years (16% vs 1–11% during the preceding 9 years, P=0.05). Despite this, treatment epoch did not significantly affect the likelihood of dramatic early neurologic improvement, long-term functional outcome, or mortality (P>0.68, 0.25, and 0.49 respectively).
To our knowledge, this is the largest series of patients treated for symptomatic ICAO in the absence of LVO. Recanalization was successfully achieved in 92% of our patients, similar to prior series examining endovascular treatment of symptomatic ICAO (77–100%, table 3). These recanalization rates were higher than those reported in a pooled analysis of stand-alone thrombolysis for symptomatic ICAO (92% vs 48%, respectively).17 Importantly, 83% of our sample had stable or improved post-procedure neurologic examinations, with 57% demonstrating improvement by ≥8 NIHSS points or post-procedure NIHSS ≤2. Furthermore, 65% of our sample were functionally independent at 3-month follow-up, well in line with prior series examining the endovascular treatment of symptomatic ICAO (40–88%, table 3).
Thrombectomy trials for acute stroke have demonstrated the benefits of revascularizing acute ICAO in the setting of a tandem LVO.18 Despite this, a dearth of studies exists examining symptomatic ICAO without LVO. Patients with symptomatic ICAO are at a higher risk of adverse outcomes compared with similarly presenting patients with ICA stenosis. Symptomatic ICAO has been associated with increased risk of death (12%), recurrent stroke (~7%), neurologic deterioration (~29%), and poor functional outcome (mRS >3, 67%).19 Furthermore, patients with ICAO and evidence of hemodynamic compromise are at even greater risk of subsequent stroke (up to 31%).20
All patients in our series demonstrated impending infarct growth and neurologic deterioration supported by CIM (42%) or fluctuant and/or progressive symptomatology (58%). This is in contradistinction to the Carotid Occlusion Surgery Study which excluded patients with progressive or unstable neurologic deficits and early treatment. In the current study, intervening in this high-risk group stabilized or improved neurologic deficits in 83% and only led to significant worsening in 14%, far lower than the reported deterioration in 31–35%.19 20 Moreover, 57% of our patients demonstrated significant early recovery neurologic improvement, a measure which has not been previously reported in any ICAO study. In terms of functional outcome, observational studies of symptomatic ICAO have reported that only 33% achieve an mRS of 0–2. Similarly, Mokin et al found that approximately 27% of patients with ICAO treated with only IV thrombolysis achieved a good functional outcome compared with approximately 44% of those treated with endovascular therapy.4 In line with more recent smaller series, our study found that endovascular treatment of symptomatic ICAO resulted in 65% achieving functional independence (table 3).
Concerns over the technical risks of traversing ICAO lesions (ie, dissection and distal embolization) have impeded recognition of endovascular revascularization as a viable treatment alternative, despite the known benefit of mechanical thrombectomy over medical therapy for LVOs.21 Dissection during lesion traversal occurred in fewer than 4% of our patients and was neither flow-limiting nor otherwise symptomatic. This, and the high rate of revascularization noted above, might be explained by the architecture of acutely symptomatic ICAO lesions (ie, loose organization of fresh thrombus permitting traversal of a microwire and channel re-establishment).
Moreover, vessel pathophysiology probably contributes to the risk of distal embolization, which was observed in 22% of our series and was associated with poor long-term outcome, even after multivariable analysis (OR 3.70; 95% CI 1.03 to 11.1). Most procedures in our study (65%) were performed using a dEPD, with increasing use particularly after availability of over-the-wire devices. Interestingly, use of a dEPD did not significantly decrease the overall risk of embolic events (P=0.51); however, it did significantly decrease the risk of embolization being associated with new or enlarging infarcts (P<0.01). Post hoc subgroup analysis suggests a non-signficant trend among those patients who required thrombolysis and/or thrombectomy of their distal emboli (increased infarct burden in 7/7 without dEPD vs 5/9 with dEPD, P=0.08). However, sensitivity analysis of these small subgroups is far too susceptible to statistical error to draw definitive conclusions. Proximal protection throughflow reversal may be a beneficial supplementary technique as recent reports have shown it to be associated with significant reductions in embolic load to the brain during carotid artery stenting.22 Future studies will need to focus on technical refinement to reduce the risk of peri-procedural stroke and distal embolization.23
An additional complication observed was sICH, which occurred in 6.5% of patients, a rate comparable to IV thrombolysis treatment alone. This complication might be explained by cerebral hyperperfusion and/or hemorrhagic conversion of ischemic stroke worsened by dual antiplatelet treatment. A limitation of carotid stenting is the need for short-term dual antiplatelet therapy, especially in patients at risk for hyperperfusion with sizeable acute infarcts (known risk factors for hemorrhagic transformation). In patients undergoing carotid stenting, cerebral hyperperfusion syndrome has been noted to peak within 12 hours of the procedure and is usually associated with insufficient post-procedural blood pressure control.24 Despite this, the rate of reperfusion injury in our cohort was comparable to the rate noted in successful mechanical thrombectomy of acute intracranial ICA occlusion.18
Fourteen patients died post-procedure. Older age and worse NIHSS were the only risk factors associated with mortality; however, following adjustment for relevant predictors of poor outcome (NIHSS ≥8, ASPECTS, CIM, time to recanalization, and distal embolization), only older age remained a significant predictor (OR 1.10; 95% CI 1.03 to 1.17). All but two of the patients who died post-procedure were successfully revascularized. One patient died from undetected sepsis, another from post-procedure myocardial infarction; however, most commonly, patients were transitioned to comfort-directed care when their neurologic function failed to significantly improve post-procedure.
Limitations of study
This study is limited by its retrospective nature and by the absence of a control arm. As per local treatment paradigm, all patients with symptomatic ICAO (as determined by non-invasive imaging) who presented with severe symptoms not explained by corresponding infarct or who demonstrated clinical instability underwent attempted revascularization, given the significant risks associated with optimal medical management. As such, no patients with symptomatic ICAO without LVO who underwent DSA were excluded. Nevertheless, our comparison between these findings and results from previous case series and historical controls provides more evidence of the safety and feasibility of angioplasty and stenting in these selected patients. An additional study limitation derived from its retrospective design is the lack of a standardized post-procedural imaging protocol. Instead, the decision to obtain imaging (and which modality to employ) was at the discretion of the attending physcians. Despite this, 90% underwent either post-procedure CT, MRI, or perfusion imaging.
Our results confirm the feasibility and safety of revascularization through angioplasty and stenting for the treatment of acutely symptomatic ICAO without intracranial LVO considered to be at high risk of subsequent deterioration. Our rates of recanalization and long-term independence were similar to smaller endovascular series, while for the first time we demonstrated high rates of early neurologic improvement. Since all patients were treated in this study, it is unknown whether this approach is superior to best medical therapy. Despite the lack of a control arm, this study may serve as a guideline for future clinical trial design and method, particularly selecting patients with positive retrograde sign and the importance of using EPD. Definitive determination of treatment superiority will require further investigation in the context of a prospective randomized trial.
Contributors Conception and design: TGJ, AJ, MJ. Acquisition of data: All. Analysis and interpretation of data: All. Drafting the article: AJ, DP. Critically revising the article: All. Administrative/technical/material support: DP, SMD. Study supervision: TGJ.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests TGJ has the following disclosures: Consultant, Neuravi (steering committee -modest), Codman Neurovascular (DSMB - modest), Stryker Neurovascular (PI DAWN - unpaid), Fundacio Ictus (PI REVASCAT unpaid); stock, Anaconda, Silk Road, Blockade Medical (modest). AFD has the following disclosure: Consultant, Medtronic (device proctor - modest).
Patient consent Not required.
Ethics approval Institutional review board of University of Pittsburgh.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.