Article Text

Download PDFPDF

Endovascular treatment of carotid embolic occlusions has a higher recanalization rate compared with cardioembolic occlusions
  1. M S Hussain1,
  2. R Lin2,
  3. E Cheng-Ching1,
  4. T G Jovin2,
  5. S I Moskowitz1,
  6. M Bain1,
  7. M Horowitz2,
  8. R Gupta1,3
  1. 1Cerebrovascular Center, Cleveland Clinic Foundation, Cleveland, Ohio, USA
  2. 2Stroke Institute, University of Pittsburgh Medical Center, Pittsburgh, Pennsylvania, USA
  3. 3Department of Neurology, Michigan State University, Michigan, USA
  1. Correspondence to R Gupta, Cerebrovascular Center, Cleveland Clinic Foundation, 9500 Euclid Avenue, S80, Cleveland, OH 44195, USA; guptar{at}


Background and purpose Treatment of large artery cerebral occlusions is rapidly evolving. We hypothesized that patients with intracranial embolic occlusions secondary to an extracranial carotid artery stenosis or occlusion have a higher probability of successful endovascular recanalization compared with those with cardioembolic occlusions.

Methods We retrospectively reviewed the databases of three institutions (University of Pittsburgh Medical Center (UPMC), Michigan State University (MSU) and Cleveland Clinic Foundation (CCF)) for acute anterior circulation ischemic strokes treated with endovascular therapies from January 2006 to July 2008. After collection of demographic, radiographic and angiographic variables, two groups were identified: artery to artery embolic occlusions and cardioembolic/cryptogenic intracranial occlusions. We defined recanalization as TIMI 2 or 3 flow. A binary logistic regression model was constructed to determine which characteristics were unique to patients with carotid embolic occlusions.

Results A total of 207 patients were identified (UPMC=100, CCF=71, MSU=36) with a mean age of 69±11 years and mean NIHSS of 17±5. Of these, 157 (75%) were due to a cardiac or cryptogenic source and 50 (25%) were from a carotid embolic source. The use of multimodal therapy (OR 2.6 (1.2–5.6), p<0.009) and the presence of a carotid embolic intracranial occlusion (OR 3.6 (1.2–7.1), p<0.012) were associated with successful recanalization, while carotid terminus occlusions were associated with unsuccessful recanalization (OR 0.35 (0.18–0.68), p<0.002).

Conclusions Patients with intracranial occlusions secondary to an extracranial carotid stenosis or total occlusion appear to have more successful recanalization rates when treated with endovascular therapy compared with those with cardioembolic occlusions.

  • Stroke
  • intervention

Statistics from

The use of endovascular techniques to treat large artery occlusions in ischemic stroke has been evolving. The location of the occlusion in the anterior circulation portends the prognosis for the patient, with poorer outcomes for occlusions of the carotid terminus than for middle cerebral artery (MCA) occlusion.1 Tandem occlusions of the extracranial carotid artery and intracranial internal carotid artery (ICA) or MCA appear to have a poor recanalization rate with intravenous thrombolysis.2 Poor recanalization may be linked to clot burden or the etiology of the thrombus. Artery to artery emboli from a ruptured plaque of the carotid artery consist of a heterogeneous composition of fibrin with aggregates of platelets, while cardioembolic lesions tend to be more uniform aggregates of fibrin.3 The red thrombi generated from the heart appear to be well organized compared wih their atheroembolic counterparts, thereby making endovascular treatment potentially more difficult.

We hypothesize that patients with proximal ICA stenosis or occlusion as the mechanism of the intracranial occlusion treated with endovascular therapy would have higher recanalization rates compared with those with isolated intracranial occlusions from a cardioembolic source.


After Institutional Review Board approval at each institution, we retrospectively reviewed the endovascular databases of three institutions (University of Pittsburgh Medical Center (UPMC), The Cleveland Clinic Foundation (CCF) and Michigan State University (MSU)) to identify patients treated for acute ischemic stroke between January 2006 and July 2008. We identified patients with anterior circulation occlusions treated with either intra-arterial thrombolysis, mechanical therapy or in combination. Pretreatment with intravenous thrombolysis was not an exclusion criterion. We identified a total of 207 patients (UPMC=100, CCF=71, MSU=36) treated among the three institutions.

Demographic information, National Institutes of Health Stroke Scale (NIHSS) and CT scans were collected and recorded. Angiographic measurements including pretreatment and post-treatment thrombolysis in myocardial infarction (TIMI) score graded as 0 for absent perfusion, grade 1 for minimal distal perfusion, grade 2 for partial perfusion and grade 3 for complete perfusion were recorded.4 TIMI grades were independently assigned by two of the authors (MSH and RL) with joined reconciliation when discrepant. Each patient underwent an evaluation for the mechanism of the stroke during their hospital stay including: a CT or MR angiogram of the head and neck, carotid duplex ultrasonography and transcranial dopplers when CTA or MRA was not possible, transthoracic echocardiogram and telemetry monitoring. The etiology of the stroke was identified as artery to artery embolus from a proximal carotid source if there was a >70% stenosis of the common or extracranial ICA at the time of angiography. One patient was included in the carotid embolic group with a 55% luminal narrowing and an ulcerated plaque without another source identified. In the absence of a proximal carotid stenosis, patients with atrial fibrillation, mechanical heart valve, ejection fraction of <35%, evidence of a thrombus in the left atrium or left ventricle on echocardiography, patent foramen ovale with a concomitant deep venous thrombus were classified as cardioembolic etiology. If no source could be identified, the patient was classified as a cryptogenic stroke. Patients treated with endovascular therapy as a rescue therapy after intravenous tissue plasminogen activator (t-PA) were imaged with a CT and CTA to assess for hemorrhage and presence of arterial occlusion prior to treatment. If there was less than 1/3 the MCA territory with hypodensity, no signs of hemorrhage and presence of a large artery occlusion, patients were taken for endovascular treatment.

Pharmacological intra-arterial treatment modalities included: t-PA (dose range 2–12 mg) and intravenous eptifibatide, a glycoprotein IIb/IIIa (GP IIb/IIIa) receptor antagonist (bolus 180 μg/kg). Doses of t-PA were administered in 2-mg aliquots with 5-min intervals between subsequent administrations through the microcatheter within the thrombus. Mechanical maneuvers were performed at the discretion of the interventionalist and included: Merci retriever system (Concentric Medical, Inc., Mountain View, California, USA), aspiration thrombectomy (Penumbra Corp., Alameda, California, USA) angioplasty or stent placement. Intracranial angioplasty was performed with standard coronary balloons with inflation to 10% below the estimated vessel diameter. Similarly, balloon-mounted intracranial stents were placed in the intracranial circulation. For extracranial ICA lesions, Acculink self-expanding stents (Guidant Corp., Indianapolis, Indiana, USA) or a Viatrac balloon angioplasty (Guidant Corp.) were used prior to treating the intracranial occlusion if the lesion was >70%. Multimodal therapy was defined as the use of mechanical therapy in conjunction with thrombolysis. Recanalization was defined as achievement of TIMI 2 or 3 grade.

CT scans 24 h post-procedure were assessed for hemorrhage by two of the authors (RL and RG) using previously published criteria.5 If a patient had a worsening of their NIHSS score by >4 points with a noted confluent hemorrhage on head CT, this was recorded as a symptomatic hemorrhage. All other hemorrhages were defined as asymptomatic.


Baseline characteristics for carotid embolic occlusions versus cardioembolic/cryptogenic occlusions and cardioembolic occlusions alone were compared using Fisher's exact test for categorical variables and Student t test for continuous variables. A p value of <0.05 was considered significant. Variables with a p value of <0.20 were entered in the binary logistic regression analysis. A binary logistic regression model using the backward elimination method to define the most conservative model was utilized to define the independent variables.


A total of 207 patients (UPMC=100, CCF=71, MSU=36) were identified that met inclusion criteria for this study. The mean age of the entire cohort was 69±11 years with a mean NIHSS of 17±5. There were 50 patients with intracranial occlusions as a result of an extracranial carotid stenosis, 125 patients due to a cardioembolic source and 32 were due to a cryptogenic etiology. Of the 50 patients with extracranial carotid artery disease, 40 (80%) were noted to have a total 100% occlusion, 9 (18%) a stenosis of 70–99% and 1 (2%) an ulcerated 55% stenosis. Overall 61 (29.5%) patients had received intravenous t-PA prior to endovascular treatment. TIMI 2 or 3 recanalization was achieved in 137 (66%) patients. Asymptomatic hemorrhage occurred in 67 (32%) patients while symptomatic hemorrhages occurred in 22 (10.6%) patients.

Table 1 summarizes the univariate analysis comparing the two groups. Patients who presented with extracranial carotid stenosis were more likely to be younger, have a lower NIHSS and were treated with stent placement along with the use of glycoprotein IIb/IIIa antagonists. There were no significant differences in rates of hemorrhage between the two groups. Table 2 summarizes the independent variables associated with successful recanalization. Patients with extracranial carotid disease as the mechanism were significantly more likely to achieve successful recanalization.

Table 1

Univariate analysis for determining differences in patients with cardioembolic occlusion in comparison with patients with carotid embolic occlusions

Table 2

Independent variables associated with successful recanalization

When patients with the defined mechanism of cardioembolic stroke (cryptogenic strokes excluded) were compared with those with carotid embolic stroke the presence of a cardioembolic occlusion was still noted to be an independent variable for poor recanalization in multivariate analysis (OR 0.27, 95% CI (0.15 to 0.86), p<0.003).


The current study demonstrates that patients with intracranial occlusions secondary to a proximal carotid artery stenosis or occlusion are more likely to undergo successful recanalization with endovascular therapy compared with their cardioemboli/cryptogenic counterparts. Studies have analyzed the effect of intravenous t-PA on patients with tandem occlusions and isolated intracranial occlusions and noted the recanalization rate to be lower in tandem occlusions.2 Moreover, patients with cardioembolic MCA occlusions tend to have more rapid and higher recanalization rates when compared with those with carotid embolic occlusion when treated with intravenous thrombolysis.6 Cardioembolic lesions are more uniform with fibrin allowing for t-PA penetration on the receptors of the thrombus. Atheroembolic clots appear to have a heterogeneous composition of fibrin and platelet aggregates thereby potentially making t-PA less effective.3 Although recent histopathological analysis of retrieved clots shows a complex pattern of fibrin, platelets and erythrocyte aggregates regardless of the stroke mechanism, the number of carotid embolic thrombi was small.7 There may have also been a selection bias within the study, as only those clots that were successfully retrieved were available for analysis, and those that were not retrieved may have a different composition.7

The effectiveness of intravenous thrombolytics during an acute myocardial infarction appears to be dependent on blood pressure. Hypotensive patients in cardiogenic shock appear to have lower rates of recanalization when compared with normotensive patients.8 A similar mechanism may be responsible for the decreased recanalization seen in tandem occlusions where a severe proximal carotid stenosis or occlusion does not allow enough exposure of t-PA to the intracranial occlusion due to a pressure gradient. Higher blood pressure has been associated with lower rates of recanalization in patients treated with intravenous t-PA.9 The authors hypothesize that this may be a result of early cerebral edema or the detrimental effects of blood pressure on fibrinolysis leading to elevations of fibrinogen levels.9 Another plausible mechanism is that the thrombus undergoes significant compaction and adherence to the cerebral vessel due to the elevations of systolic blood pressure.

In our cohort, 82% of patients with a proximal carotid lesion were treated with stenting, thereby allowing for relief of the pressure gradient to the intracranial lesion. Endovascular therapy has the additional advantage of treating the atherothrombotic lesion that will not respond to intravenous t-PA. Patients with carotid embolic occlusions were more likely to be treated with glycoprotein IIb/IIIa inhibitors due to a high proportion of patients being treated with stent placement. The use of eptifibatide in conjunction with t-PA has recently been shown to have a similar safety profile to intravenous t-PA alone.10 We have also recently demonstrated that this combination does not appear to have a higher hemorrhage rate compared with other treatment modalities.11

Other factors potentially contributing to our finding is the degree of collaterals and the severity of clot impaction in the intracranial vessel. Patients with proximal carotid lesions likely develop collaterals over time as the vessel begins to narrow. Others have found patients presenting with cardioembolic occlusions present with higher NIHSS compared with patients with occlusions secondary to MCA atheromatous disease.12 We found a similar result in our cohort with carotid embolic strokes presenting with lower initial NIHSS. The obstruction of forward flow may prevent compaction of the thrombus into the cerebral vessels. Patients with cardioembolic occlusions have constant forward flow with elevations of blood pressure on the thrombus causing a situation of mechanical compaction of the thrombus and likely adherence of the thrombus to the cerebral vessel. The difficulty in recanalizing cardioembolic occlusions may be a time-based phenomenon of how long the thrombus has had time to compact thereby preventing easy retrieval or lysis when compared with intravenous studies of <3 h. In contrast, carotid embolic occlusions may be more protected from this phenomenon due to the proximal obstruction and reduced pressure gradient.

There are several limitations of this study due to its retrospective design. The first is that we are unable to quantify the thrombus burden that may impact recanalization rates. Second, we were unable to analyze the effect of recanalization on outcome measures due to the retrospective design. Third, there were three different institutions involved in this study and endovascular strategies may have varied. All of the operators in this study used a similar paradigm of treatment and it is thus less likely that this had a significant impact on our results. Lastly, there were some patients where an etiology was not found and were determined to be cryptogenic in etiology. It is possible some patients had in situ thrombosis of the intracranial MCA or ICA as the etiology of the occlusion. Many cryptogenic embolic strokes may also be cardiac in etiology due to paroxysmal atrial fibrillation13 and thus we performed the analysis in combination as well as separately to discern any differences.

In conclusion, this study gives preliminary evidence that endovascular stroke treatment of patients with a carotid embolic source has a higher recanalization rate compared with treatment of those with a cardioembolic occlusion.



  • Competing interests Dr Jovin is a consultant for Concentric Medical, EV3, CoAxia Inc.; Dr Gupta is on the scientific advisory board for Concentric Medical; Dr Horowitz is a consultant for EV3; Drs Hussain, Lin, Cheng-Ching, Moskowitz and Bain have no competing interests.

  • Ethics approval This study was conducted with the approval of the IRB from each hospital.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

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.