Occlusion site does not impact outcome in patients with carotid stroke undergoing endovascular reperfusion

Authors


  • Conflict of interest: All authors hereby declare they have no conflicts of interest.

Abstract

Background

It remains unclear whether occlusion site impacts outcome in patients with acute carotid artery occlusions.

Methods

Patients with acute carotid artery occlusion that underwent endovascular reperfusion treatments were prospectively enrolled. Patients with extracranial carotid bifurcation occlusions were compared with those with intracranial carotid-T-occlusions. Collected data included demographics, risk factor profile, and procedure-related variables. Neurological deficits were studied with the National Institutes of Health Stroke Scale and outcome was studied with the modified Rankin Score at day 90 after stroke and dichotomized into favorable (≤2) or unfavorable (>3). Recanalization status was studied with the thrombolysis in cerebral infarction scale.

Results

We included 51 patients (33 with extracranial bifurcation occlusion and 18 with intracranial T-occlusion). Patients with T lesions were significantly older (median 74 versus 56 years, P = 0·02), more frequently had atrial fibrillation (61% versus 18%; P = 0·005) and cardioembolism (78% versus 21% P = 0·001), smoked less often (6% versus 42%; P = 0·01), and less often required stent implantation (11% versus 48%; P = 0·015). However, neurological severity, other procedure and peri-procedure-related variables including recanalization rates and percentages of symptomatic hemorrhages did not differ between the groups. Mortality rates (24% versus 23%) and chances for favorable outcomes (33% versus 24%) did not significantly differ. On multivariate logistic regression analysis, occlusion location was not a significant modifier of outcome.

Conclusions

Despite differences in stroke risk factors and treatments used between patients with extracranial bifurcation and intracranial T-occlusions, lesion location in itself does not influence outcome in patients with acute carotid artery occlusion treated with endovascular reperfusion.

Introduction

Acute carotid artery occlusion usually causes large hemispheric stroke resulting in poor outcome [1-4]. Furthermore, these strokes do not respond favorably to thrombolysis probably due to low recanalization rates [5]. The carotid artery may be occluded extracranially at the bifurcation secondary to an acute thrombus that may form on an existing plaque or to cardioembolism or to other causes such as dissection [2, 6-9]. The carotid artery may also be occluded intracranially at the T bifurcation to the middle and anterior cerebral arteries and occlusion at this site is usually secondary to cardioembolism [1, 6, 10]. Most studies investigating response to therapy and outcome in patients with acute carotid occlusion did not differ between the two sites of occlusion but rather lumped all patients together [3, 11-14]. Nevertheless, the projected differences in stroke etiology for carotid occlusions at different sites may suggest different responses to therapy.

Aims

The goal of the current study was to compare between patients with intracranial carotid-T-occlusion and those with acute extracranial occlusion at the bifurcation.

Patients and methods

We prospectively recruited consecutive patients presenting with large hemispheric stroke that underwent endovascular reperfusion treatment (ERT) over the span of 10 years from 2002 to 2012 into our stroke registry and the data were retrospectively analyzed. The institutional review board (Hadassah Medical Organization) authorized anonymous inclusion of patients into the consecutive data base without getting informed consent.

In this study, we included only patients with acute carotid occlusions that were treated with ERT. To be eligible for ERT, patients had to have an initial National Institutes of Health Stroke Scale (NIHSS) [15] score of ≥10 and to be previously independent [modified Rankin Scale score (mRS) < 2]. Furthermore, all included patients presented within eight-hours of stroke onset. The diagnosis of internal carotid artery occlusion was established according to clinical findings that included hemiparesis/hemiplegia, sensory symptoms, and evidence for cortical involvement such as aphasia, neglect, or hemianopsia in various combinations. The diagnosis of carotid occlusion had to be proven on computed tomography (CT) angiography (CTA) or digital subtraction angiography (DSA) in all patients. Some of the included patients underwent a full multiparametric stoke magnetic resonance imaging (MRI) protocol that included diffusion, perfusion, and susceptibility weighted imaging as well as magnetic resonance (MR) angiography and FLAIR. Multimodal MR studies were used in some of the patients to determine if there was tissue at risk (DWI-PWI mismatch) especially if the time of onset was not clear at first or if there was a discrepancy between the clinical picture (NIHSS) and the CTA finding of complete carotid occlusion.

Exclusion criteria for this study included evidence of large hemispheric infarction on the admission CT defined as hypodensity covering more than 1/3 of the middle cerebral artery (MCA) territory, international normalized ratio (INR) > 3, and existing disease with limited life expectancy (e.g., terminal cancer). Patients with small vessel disease were excluded as were those presenting in deep coma and those with primary intracerebral or subarachnoid hemorrhage.

In this study, we compared patients with acute extracranial carotid artery occlusion at the bifurcation with those patients with acute intracranial T-occlusions leading to occlusion of both the middle and anterior cerebral arteries. Occlusion sites were verified according to the findings on DSA.

Clinical and demographic characteristics accrued included cerebrovascular risk profile, concomitant medications, time from symptom onset to initiation of endovascular procedure, and time onset to reperfusion. Infarct etiology was classified according to TOAST [16] criteria as cardioembolic, large artery atherothrombotic, other classified (e.g., dissection) or unclassified. Of note, cardioembolic strokes included all those caused by atrial fibrillation as well as those caused by valvular thrombi and left ventricle thrombi.

All patients were admitted to the intensive care unit for at least 24 h postprocedure. Neurological deficits were determined with the NIHSS and functional deficits before admission and at 90 days postinfarct were evaluated with the mRS and good outcome was defined as an mRS ≤ 2.

Radiological parameters were evaluated on entry CT/MRI and on the diagnostic and therapeutic angiography and follow-up CT/CTA. Patients with concomitant occlusion of the MCA by a separate thrombus necessitating endovascular reperfusion were considered to have tandem lesions.

The number and types of procedural modalities used were also documented and studied in all patients. Flow at the end of the endovascular procedure was classified with the Thrombolysis in Cerebral Infarction system (TICI; 0 – no flow, 1 – minimal flow, 2a – residual stenosis, 2b – near normal flow, and 3 – normal patent vessel) [17].

Treatment complications including postprocedure hemorrhage and clinical deterioration without hemorrhage were also documented. Symptomatic intracranial hemorrhage was defined according to ECASS-III criteria [18].

Statistical evaluations were performed with the Sigma-Stat package (SPSS Inc., Chicago, IL, USA). For univariate analysis, patients with bifurcation lesions were compared with those with T-occlusions using Student's t-test or chi-square tests. We then ran a multivariate logistic regression analysis model that included age, gender, lesion location, TOAST classification, and admission NIHSS to test the effects of those variables on the chances for having a favorable outcome at 90 days poststroke.

Results

Over the study period, 6430 patients with stroke were treated at our center. Fifty-one consecutive patients fulfilling entry criteria were recruited into this study. Of those, 33 had acute carotid occlusion at the bifurcation and 18 had acute carotid-T-occlusions. The baseline clinical and radiological characteristics are presented in Table 1. All patients were independent prior to the procedure (mRS ≤ 2). Another 32 patients with internal carotid artery (ICA) occlusions were excluded from treatment eventually based on our inclusion/exclusion criteria (mainly due to presentation later than eight-hours from symptom onset and significant existing disability prior to the current event) and these patients were not evaluated in the current study.

Table 1. Univariate analysis comparing patients with bifurcation and T-occlusions of the internal carotid artery
Variable/groupBifurcation (n = 33)T (n = 18)P
  1. MCA, middle cerebral artery; NIHSS, National Institutes of Health Stroke Scale.
Age ± SD (median)57·2 ± 14·0 (56)67·9 ± 12·9 (74)0·02
Gender (male %)21 (64)4 (22)0·01
Tandem lesion at MCA8 (24)4 (22)1
Involved hemisphere – left (%)22 (67)12 (67)1
Hypertension (%)16 (48)12 (67)0·34
Ischemic heart disease (%)10 (31)5 (28)0·89
Atrial fibrillation (%)6 (18)11 (61)0·005
Diabetes mellitus (%)6 (18)6 (33)0·30
Hyperlipidemia (%)12 (36)7 (39)0·91
Smoking (%)14 (42)1 (6)0·015
TOAST classification (%)  <0·001
Cardioembolic7 (21)14 (78) 
Large vessel17 (55)2 (11) 
Small vessel00 
Other4 (12)2 (11) 
unknown5 (15)0 
Admission NIHSS (median)20·2 ± 5·4 (22)21·8 ± 5·0 (235)0·31
Day 1 NIHSS (median)15·1 ± 8·2 (17·5)18·0 ± 6·5 (19)0·20
Day 90 NIHSS (median)9·9 ± 6·2 (11)10·3 ± 5·8 (10)0·82

Patients with bifurcation occlusions differed significantly from those with T lesions in that they were younger [median 56 (min-max 27–83) versus 74 (min-max 27–90) years, P = 0·02], more often smoked (42% versus 6%; P = 0·015), and less often had atrial fibrillation and cardioembolic stroke (P = 0·001). The percentage of males was significantly higher among patients with bifurcation lesions (64% versus 22%; P = 0·001), but other baseline variables did not differ between the groups.

All included patients had severe stroke (mean admission NIHSS 20·75 ± 1·6). Neurological disability measured with the NIHSS on admission, day 1, and discharge (Table 1) did not differ between the groups. The percentage of patients with coexisting occlusions of the MCA (tandem lesions) did not differ significantly between the groups (Table 1). Of note, in the T-occlusion group, two of the patients had contiguous thrombi extending from the carotid terminus into the M1 segment of the MCA and two others had separate thrombi in the carotid terminus and in the more distal M1 segment. The percentage of patients treated with systemic tissue plasminogen activator (tPA) prior to ERT (i.e., treated with a bridging strategy) also did not differ significantly between the groups (3% versus 11%).

Of the different treatments used, MERCI retrievers were used in three patients with T-occlusion and four patients with bifurcation occlusions; IIb-IIIa antagonists were given in 5 and 7, respectively; balloon angioplasty was used in 2 and 11 patients, respectively; and stentreiver devices were used in 8 and 12 patients, respectively. Of note, stent implantation was needed more often in treating patients with bifurcation lesions (48% versus 11%; P = 0·015). All other procedure-related variables including onset to treatment time and time to vessel recanalization, number and types of procedural modalities used, and lesion length did not differ between the groups (Table 2). The chances for achieving good-to-excellent or excellent-only target vessel recanalization or were similar for both groups (TICI 2b-3 52% versus 50%; TICI 3 37% versus 39%, respectively). However, patients that were treated with stentrievers had higher chances of achieving TICI 2b-3 (7/8 in the T-occlusion group and 9/12 in the bifurcation occlusion group). Following the procedure, six patients (18%) with bifurcation occlusion and five patients (28%) with T-occlusions underwent decompressive hemicraniectomy (P = 0·49) because of failed or futile recanalization and development of large edema volumes with impending herniation.

Table 2. Comparison of procedure and postprocedure-associated variables
Variable/groupBifurcation (n = 33)T (n = 18)P
  1. ICH, intracerebral hemorrhage; TICI, thrombolysis in cerebral infarction.
Onset to treatment (median)5·4 ± 2·1 (5)5·2 ± 2·4 (5)0·74
Time to recanalization (median)6·3 ± 1·8 (6)5·8 ± 2·6 (5·5)0·42
TICI at end of endovascular treatment (%)  0·85
0–111 (33)5 (28) 
2a5 (15)4 (22) 
2b–317 (52)9 (50) 
Any ICH (%)10 (31)5 (28)0·89
Symptomatic ICH (%)5 (15)1 (6)0·40
Decompressive hemicraniectomy (%)6 (18)5 (28)0·49
Modified Rankin Score day 90 (%)N = 31N = 170·78
0–211 (33)4 (24) 
36 (18)4 (24) 
4–56 (18)5 (29) 
68 (24)4 (24) 

Fifteen of our patients (29%) had hemorrhagic transformation of the infarcts. Six of these (12%) were classified as symptomatic. Of these, five had bifurcation lesions and one had T-occlusion (P = 0·40).

At seven-days poststroke, eight patients with bifurcation lesions (24%) and four patients with T-occlusions (23%) had died (P = 0·79).

At 90 days poststroke, outcome could be determined in 50 of the 51 patients and one was lost to follow-up. The percentage of patients that achieved favorable outcome at 90 days defined as mRS ≤ 2 was similar between the groups (11/33 versus 4/17; P = 0·78) and no significant shifts in outcome were noted.

On multivariate regression analysis controlling for age, gender, admission NIHSS, and TOAST classification, lesion location did not have an impact on outcome (Table 3). In this analysis, age and admission NIHSS scores were negatively correlated and cardioembolism was positively associated with favorable outcome.

Table 3. Multivariate analysis on chances for having a favorable outcome (mRS ≤ 2)
VariableOR95% CIP
  1. CI, confidence interval; mRS, modified Rankin Scale; NIHSS, National Institutes of Health Stroke Scale; OR, odds ratio.
Age0·920·86-0·980·018
Admission NIHSS0·750·62-0·910·004
Cardioembolism9·631·17–79·40·035
Male gender9·170·92-90·980·058
T-occlusion0·680·07-6·530·741

Discussion

Most patients with acute carotid occlusions have large strokes and poor outcome [1, 2, 6, 7, 11, 19, 20]. However, previous studies lumped all patients with acute carotid occlusions without dividing them according to the site of occlusion [3, 6, 11, 12]. The main finding of the current study is that despite differences in risk factor profile and stroke cause, the outcome in patients with acute carotid occlusion treated with ERT does not seem to be influenced by occlusion site. Furthermore, these patients respond to therapeutic interventions in a similar fashion and the chances of achieving good or excellent reperfusion were not influenced by occlusion site.

Given the poor outcome of these patients and the low rates of recanalization with systemic tPA [5, 13, 14], these patients are usually referred to ERT which reportedly result in higher rates of favorable outcome [9, 19-26]. Given the relatively low rates of mortality in our cohort (∼25%), the current results further emphasize this practice and suggest that ERT may indeed be required for both types of acute carotid occlusion. However, the results of ongoing randomized studies comparing systemic thrombolysis with new generation stentreivers are expected to answer the question of how should patients with similar large vessel occlusion be treated. In this regard, it should also be noted that ERT is constantly evolving with newer devices being introduced over time. These devices (e.g., stentrievers) offer higher chances of achieving target vessel recanalization.

Importantly, lesion location was not a significant modifier of outcome on multivariate analysis. On this analysis, age and NIHSS score on admission were negatively correlated with favorable outcome, whereas cardioembolism was positively associated with favorable outcome.

It is slightly surprising that despite the larger proportion of cardioembolism in patients with T-lesions and their older age and the known association of these factors with poor outcome, the rates of unfavorable outcome did not differ between the groups [19-21, 27, 28]. In this aspect, it should be recognized that differences in the natural history of extracranial and intracranial occlusions likely exist such that intracranial T-occlusions are usually sudden and embolic and almost always manifest clinically, whereas atherothrombotic extracranial bifurcation occlusions may develop slowly over years and in some cases even occur in the absence of clinical manifestations in lieu of good collateral circulation. Therefore, we emphasize the fact that this study only dealt with acute symptomatic occlusions that were treated with ERT and is not representative of the natural history of carotid occlusion either not treated or treated with tPA only. Furthermore, one can speculate that other factors influencing outcome such as the percentage of patients with tandem lesions involving the ICA and MCA [25, 29] and the similar stroke severity as observed in the NIHSS [30] and collateral state [31, 32] did not differ between the groups and may have negated the effect of poor prognostic markers.

Furthermore, the finding that patients with cardioembolism actually had increased chances for favorable outcome in the current series may suggest that these patients could be especially responsive to ERT. Possible explanations for this phenomenon may include a higher likelihood of embolus extraction with stent-retrievers as opposed to more difficult vessel recanalization procedures in patients with existing atherosclerotic plaque which may be more resistant to ERT and necessitate stent implantation.

Our study has several limitations. First, this was a retrospective analysis of prospectively accrued data from consecutive patients and not a true prospective study comparing patients with T-lesions with those with bifurcation lesions and as such may be open to bias. Second, we only included patients treated with an endovascular or bridging approach and did not include patients treated with systemic tPA or untreated patients and therefore, the generalization of our results to all patients with acute carotid occlusions may not be complete. Third, we did not directly look at other factors which may have had an influence on final outcome such as collateral flow across the circle of Willis, or clot size and burden. Furthermore, given the low number of patients with dissection, we could not determine whether stroke mechanism at any particular location bares an impact on outcome.

In conclusion, our results indicate that-occlusion location does not influence outcome or chances for recanalization in patients with intracranial or extracranial carotid occlusions treated with ERT.

Acknowledgements

This study was supported in part by the Peritz and Chantal Scheinberg Cerebrovascular Research Fund and by the Sol Irwin Juni Trust Fund.

Ancillary