Article Text

other Versions

Original research
Clinical, angiographic and radiographic outcome differences among mechanical thrombectomy devices: initial experience of a large-volume center
  1. Tareq Kass-Hout1,
  2. Omar Kass-Hout1,
  3. Chung-Huan Johnny Sun1,
  4. Taha Kass-Hout2,
  5. Samir Belagaje1,
  6. Aaron Anderson1,
  7. Michael Frankel1,
  8. Rishi Gupta3,
  9. Raul Nogueira1
  1. 1Department of Neurology, Emory University School of Medicine, Atlanta, Georgia, USA
  2. 2Humanitarian Tracker, Washington, DC, USA
  3. 3Wellstar Neurosurgery, Kennestone Hospital, Marietta, Georgia, USA
  1. Correspondence to Dr Raul Gomes Nogueira, Department of Neurology, Emory University School of Medicine, 80 Jesse Hill Drive SE, Room #333, Atlanta, GA 30303, USA; raul.g.nogueira{at}


Background and purpose Higher reperfusion rates have been established with endovascular treatment of acute ischemic stroke (AIS). There are limited data on the comparative performance of mechanical thrombectomy devices.

Methods A retrospective single-center analysis was undertaken of all consecutive patients who underwent thrombectomy using Merci, Penumbra or stent retrievers (SR) from September 2010 to November 2012. Baseline characteristics, rates of successful recanalization (modified Thrombolysis in Cerebral Infarction (mTICI) score 2b–3), symptomatic intracerebral hemorrhage (sICH), final infarct volume, 90-day mortality and independent functional outcomes at 90 days were compared across the three devices.

Results Our cohort included 287 patients. There were mild imbalances in baseline characteristics with trends towards higher National Institutes of Health Stroke Scale (NIHSS) score in patients in the Merci group (SR=18 vs Merci=21 vs Penumbra=19, p=0.06) and lower Alberta Stroke Program Early CT Score (ASPECTS) in patients in the SR group (>7: SR=51% vs Merci=61% vs Penumbra=62%, p=0.12). On univariate analysis there were no differences in the rate of sICH (SR=7% vs Merci=7% vs Penumbra=6%, p=0.921) and infarct volume (SR=61.5 mL vs Merci=69.5 mL vs Penumbra=59.2 mL, p=0.621). Trends towards better functional outcomes were found with Penumbra and SR vs Merci (41% vs 36% vs 25%, respectively, p=0.079). Complete or near complete reperfusion (mTICI 2b–3) was higher in the SR and Penumbra groups than in the Merci group (86% vs 78% vs 70%, respectively, p=0.027). Binary logistic regression showed that SR was an independent predictor of good functional outcome (OR 2.27, 95% CI 1.018 to 5.048; p=0.045).

Conclusions Although our initial data confirm the superiority of SR technology over the Merci device, there was no significant difference in near complete/complete reperfusion, final infarct volumes or clinical outcomes between SR and Penumbra thromboaspiration.

Statistics from


There has been a gradual shift in the paradigm of acute ischemic stroke (AIS) therapy over the last few years. Stroke therapy is based on the concept that early recanalization of an occluded artery will lead to preservation of the time-sensitive penumbra which theoretically leads to a better clinical outcome. A pooled analysis of the Alteplase ThromboLysis for Acute Noninterventional Therapy in Ischemic Stroke, European Cooperative Acute Stroke Study (ECASS) and National Insititute of Neurological Disorders and Stroke recombinant tissue plasminogen activator (rt-PA) stroke trials demonstrated that the time from symptom onset to treatment is an independent predictor of favorable outcome after intravenous (IV) thrombolysis.1–,6 Understanding that early reperfusion is critical for a favorable outcome in patients with AIS has led to a brief but eventful development of technologies for catheter-based treatment to achieve faster and safer rates of recanalization.

Mechanical thrombectomy was initially introduced as a complementary treatment to AIS therapy, particularly when intra-arterial (IA) or IV thrombolysis is ineffectual or associated with a high rate of reocclusion.7–13 Although more successful than IV tissue plasminogen activator (tPA), recanalization with the old generation mechanical thrombectomy devices was only moderately successful in opening blood vessels with complete recanalization (Thrombolysis in Myocardial Infarction/Thrombolysis in Cerebral Infarction (TIMI/TICI) 3), approximating only 25% in the Merci retriever (Stryker Neurovascular, Mountain View, California, USA) and Penumbra Aspiration System (Penumbra, Alameda, California, USA) single-arm trials.14–18 Reperfusion with the newer generation mechanical thrombectomy devices (ie, stent retrievers (SR) (Solitaire Flow Restoration and the Trevo Retriever devices)) was tested in two head-to-head randomized clinical trials which showed that these devices are significantly better at restoring blood flow in occluded arteries compared with the Merci retriever.19 ,20 The aim of this study is to compare the effectiveness of SRs with the previous technologies (Merci device or Penumbra Aspiration System) in treating large artery AIS.


Data collection

A retrospective study was undertaken of all patients with ischemic stroke treated with mechanical thrombectomy devices at our institution over a 2-year period (October 2010 to November 2012). Data were obtained from chart reviews of electronic medical records and review of radiographic images. All information pertaining to age, baseline demographics, past medical history, admission glucose levels, admission National Institutes of Health Stroke Scale (NIHSS) score and time from last known normal to arterial access were collated for the purposes of the analysis.

Imaging interpretation

All patients undergoing endovascular therapy had pretreatment head CT scans evaluated for signs of early ischemic changes as defined by the Alberta Stroke Program Early CT Score (ASPECTS).21 Procedure times were calculated as the duration between the start of femoral access and the time of final device pass. Reperfusion rates were graded based on the modified TICI (mTICI) scoring system which has been previously described. Scores of mTICI 2b (perfusion with >50% distal branch filling) and mTICI 3 (perfusion with filling in all distal branches) were considered successful whereas mTICI scores of ≤2a were considered unsuccessful. All intracranial hemorrhages were classified using the ECASS definition22 and evaluated on CT images obtained within 24 h of the procedure. Symptomatic hemorrhages were defined as parenchymal hematomas type 1 or type 2 with space-occupying effects within the infarcted area. Post-procedural infarct volumes were calculated by summating the total area of infarct on each image slice (DWI-MRI or 48 h head CT) and multiplying by the slice thickness.

Patient selection

All patients with acute stroke were selected for mechanical thrombectomy based on previously described protocols.23 Depending on the time period, device availability and operator preference, patients were treated with one of three reperfusion options: Merci retrievers, the Penumbra Aspiration System or a SR (either the Solitaire FR Revascularization Device; Covidien, Plymouth, Minnesota, USA or the Trevo Pro Retrieval System; Stryker, Kalamazoo, Michigan, USA). If a patient received more than one thrombectomy device, they were retained in the analysis and classified on an intention-to-treat basis according to the first device employed. All patients treated with intracranial stenting, extracranial stenting, angioplasty and/or IA tPA alone were excluded from the analysis.

Outcome measures

The modified Rankin Scale (mRS) score was used to evaluate each patient's functional independence at 3 months following treatment with endovascular therapy. The primary endpoint of ‘good outcome’ was defined as an mRS score of 0–2 determined by an independent reviewer blinded to the procedure results. All evaluations were performed during follow-up examination visits or by telephone interviews using a preset questionnaire.


IBM SPSS software V.20 was used to perform the statistical analysis. A two-sample t test and Mann–Whitney U test was performed to compare each normally and non-normally distributed continuous variable between the three devices. The Fisher exact test was used to compare categorical variables and a Spearman rank test was performed to assess colinearity of the baseline variables. Non-colinear variables with p<0.20 on univariate testing were then included in the binary logistic regression model for interpretation.


Clinical and treatment characteristics

A total of 324 consecutive patients with AIS underwent endovascular intervention during the study period. Of these, 37 patients were excluded (8 underwent intracranial stenting only, 16 received IA tPA only, 12 underwent angioplasty only and 1 patient did not require mechanical thrombectomy due to spontaneous recanalization), leaving 287 patients for analysis. The mean age was 65.9±14.3 years, median NIHSS score was 19.3 and 50% were men. The sites of occlusion included the middle cerebral artery (MCA) in 198 patients (69%) (MCA-M1 segment, n=161 (56.1%); MCA-M2 segment, n=37 (12.9%)), the anterior cerebral artery in one patient (0.34%), the internal carotid artery terminus in 65 patients (22.6%) and the basilar artery in 23 patients (8.6%).

Among the whole cohort, 115 patients (40%) were in the SR group (Solitaire FR, n=87; Trevo Pro, n=28), 71 of whom (61.7%) received some form of adjuvant therapy (IA tPA, n=40; Penumbra system for either local aspiration or to treat residual distal occlusions, n=8; both IA tPA and Penumbra, n=23); 81 patients (28.2%) were in the Merci Retriever group, 65 of whom (80.2%) received additional therapy (IA tPA, n=5; Penumbra system, n=15; both IA tPA and Penumbra, n=45); and the remaining 91 patients (31.8%) were in the Penumbra Aspiration System group, in whom adjuvant therapy with IA tPA was used in 68 patients (74.7%).

The baseline characteristics were similar in the three groups (table 1) except for a trend towards a lower predominance of men (SR=58% vs Merci=38% vs Penumbra=54%; p=0.06) and a higher baseline NIHSS score (SR=18 vs Merci=21 vs Penumbra=19; p=0.06) in the Merci group and a lower pretreatment ASPECTS score in the SR group (>7: SR=51% vs Merci=61% vs Penumbra=62%; p=0.12).

Table 1

Demographic and clinical characteristics

Clinical, angiographic and radiographic outcomes

At the time of the analysis, 98% of the patients (281/287; SR=109, Merci=81, Penumbra=91) had had their 90-day functional assessment. Tables 24 summarize the clinical, angiographic and radiographic differences among the three groups. The SR technology was associated with significantly higher rates of complete reperfusion (mTICI 3: 37.4% SR vs 13.6% Merci vs 8.8% Penumbra; p<0.001). The rates of near complete/complete reperfusion (mTICI 2b–3) were also higher in the SR group than in the Merci group (86% vs 70%; p=0.007); however, there was no significant difference in the rates of near complete/complete reperfusion between the SR and Penumbra groups (86% SR vs 78% Penumbra; p=0.251). Final infarct volumes were similar in the three groups (61.5±65.1 SR vs 69.5±70.2 Merci vs 59.2±40.4 Penumbra; p=0.541) Similarly, the rates of symptomatic intracranial hemorrhage (sICH) and the mortality rates at 90 days were not different among the three groups (7% SR vs 7% Merci vs 5% Penumbra; p=0.863 and 32% SR vs 41% Merci vs 26% Penumbra; p=0.132, respectively). The procedure time was similar in the SR and Penumbra groups (79.3±53.5 min SR vs 90.6±43.9 min Merci vs 74.5±40.4 min Penumbra; p=0.086). There was a trend towards better functional outcomes with the Penumbra system and SR than with the Merci device (90-day mRS 0–2: 36% SR vs 25% Merci vs 41% Penumbra; p=0.079; figure 1).

Table 2

Unadjusted differences in clinical, angiographic and radiographic outcomes between stent retrievers and Merci retriever

Table 3

Unadjusted differences in clinical, angiographic and radiographic outcomes between stent retrievers and the Penumbra aspiration system

Table 4

Unadjusted differences in clinical, angiographic and radiographic outcomes between the Penumbra aspiration system and Merci retriever

Figure 1

Schematic plots comparing the clinical, radiographic and angiographic unadjusted outcome differences between stent retrievers, the Merci retriever and the Penumbra aspiration system.

Multivariate analysis of predictors of good functional outcome

In view of the interrelationships between all predictors of variables associated with a good clinical outcome, multiple logistic regression analysis was used based on a total of 224 patients (57 were direct transfers to the interventional suite with no immediate pretreatment ASPECTS and 6 patients had incomplete follow-up). After excluding the procedure duration times and the reperfusion rates due to their colinear relationship with the device used, the binary logistic model showed the following independent predictors of good functional outcome at 90 days after acute stroke intervention: baseline NIHSS (OR 0.919, 95% CI 0.863 to 0.979; p=0.009), age (OR 0.970, 95% CI 0.947 to 0.993; p=0.012), favorable pretreatment ASPECTS score (OR 4.991, 95% CI 2.512 to 9.916; p<0.001) and device used. Interestingly, when the Merci retriever was used as a reference, the use of SR was found to be a significant predictor of a good outcome at 90 days (OR 2.268, 95% CI 1.018 to 5.048; p=0.045) whereas the use of the Penumbra system approached significance for a good outcome (OR 2.159, 95% CI 0.953 to 4.894; p=0.065; table 5).

Table 5

Multivariate predictors of good functional outcome at 90 days after acute stroke intervention


Our analysis reinforces the findings of the SWIFT and TREVO-2 trials by demonstrating superiority of SR over the Merci retriever. Our results also suggest that the Penumbra thromboaspiration technology is superior to the Merci device. Despite significantly higher rates of complete reperfusion (mTICI 3) with SR than with Penumbra, we were not able to elicit any other significant differences in terms of clinical, angiographic or radiographic outcomes between these two devices.

Due to the modest technical efficacy of the Merci retriever and the earlier generations of the Penumbra system in establishing recanalization along with the continued evolution of endovascular therapy to improve both efficacy and safety, a paradigm shift has occurred towards AIS therapy with newer generation thrombectomy technology including SR and the Penumbra Max aspiration system, which have shown substantially higher rates of reperfusion than earlier devices.19 ,20 ,24 ,25 Recently, the FDA granted clearance to both the Solitaire flow restoration stentriever device (ev3/Covidien Vascular Therapies, California, USA) and the Trevo Pro Retrieval System (Stryker) under 510(k) clearance. The decision for clearance of these two devices was based on the results of two randomized trials demonstrating that they are more effective than their predecessor, the Merci retriever, in the setting of AIS.19 ,20 In the SWIFT trial the Solitaire flow restoration device was compared with the Merci device in 113 patients,19 and in TREVO 2 the Trevo device was compared with the Merci device in 178 patients.20 Both trials enrolled patients with large intracranial artery occlusion within 8 h of symptom onset who had a contraindication to, or had not responded to, IV tPA. In the SWIFT trial the primary end point of recanalization (defined as a TIMI score of 2 or 3 without sICH) was more common with Solitaire (OR 4.87, 95% CI 2.14 to 11.10) while, in TREVO 2, the primary end point of reperfusion (TICI score ≥2) was more common with Trevo (OR 4.22, 95% CI 2.01 to 08.86). Safety and some secondary neurological and functional outcome measures also suggested benefits with the new devices compared with the Merci retriever. To the best of our knowledge, a comparison between SRs and thromboaspiration systems has not yet been done and our study is the first to report such an analysis.

In agreement with previously published results,19 ,20 our initial data suggest in the ‘real-world’ clinical setting that, compared with the Merci retriever, neurothrombectomy with the SR is associated with an increased rate of successful cerebral recanalization without sICH, improved final disability outcomes and reduced mortality. On the other hand, except for the significantly higher rates of complete TICI 3 reperfusion with SR, thromboaspiration with the Penumbra system had similar clinical, radiographic and angiographic outcomes to SR. All previous studies have defined successful reperfusion as near complete (mTICI 2b–3) rather than complete (mTICI 3) reperfusion. Thus, the clinical implication of higher rates of complete reperfusion with SR remains unclear, and future analyses with larger sample sizes are required to determine if such an impact on clinical outcomes exists. This is particularly important when taking into account the added cost of SR to stroke therapy, which we reported previously to be approximately $4000 more expensive than the older mechanical thrombectomy devices.26 Nevertheless, we have demonstrated in a multivariate analysis that the use of SR could potentially be an independent predictor of good functional outcome at 90 days (table 4).

Our study has some shortcomings, mostly in relation to its retrospective design and relatively small sample size. There were imbalances in the baseline characteristics. Specifically, the trend towards a higher baseline NIHSS score in the Merci group and lower pretreatment ASPECTS score in the SR group could have affected our results by benefiting the Penumbra group. As an attempt to overcome this limitation, we performed a multivariate analysis adjusting for the differences. In addition, this was an intention-to-treat analysis with ‘first device used’, with adjunctive treatments used in many of the cases. However, this more closely reflects real clinical practice than the use of any single device in isolation. Finally, both technical expertise and technology evolve over time. This analysis incorporates our learning curve of SR treatment and it is possible that more recent SR results would be superior. Similarly, the Penumbra thromboaspiration system (which now includes the 5MAX and 5MAX ACE catheters) has evolved from that used during most of the duration of our study. On the other hand, the single-center nature of our analysis with only two operators performing all procedures allowed for a more homogeneous approach to treatment, thereby improving our ability to compare the different devices.


Our initial data suggest that the SR and Penumbra thromboaspiration system is superior to the Merci device in achieving higher rates of reperfusion and better outcomes. No significant difference in near complete/complete reperfusion, final infarct volumes or clinical outcomes could be established between SR and Penumbra thromboaspiration. A randomized clinical trial comparing these two technologies is needed.


View Abstract


  • Contributors RGN and RG conceived and designed the research. C-HJS and TK-H acquired the data. C-HJS, TK-H, TahaK-H and OK-H analyzed and interpreted the data and performed the statistical analysis. RGN and RG handled the funding and supervision. TK-H, RGN and RG drafted the manuscript. All authors made critical revision of the manuscript for important intellectual content.

  • Funding None.

  • Competing interests RGN: Stryker Neurovascular (Trevo-2 and DAWN Trials Principal Instigator); Covidien (SWIFT and SWIFT Prime Trials Steering Committee; STAR Trial Angiographic Core Lab), Penumbra (3D Separator Trial Executive Committee), Rapid Medical (Stroke Trial DSMB). RG: Consultant Stryker Neurovascular, Covidien, Rapid Medical; Steering Committee for Penumbra THERAPY trial; Associate Editor Journal of Neuroimaging; Associate Editor Interventional Neurology; Royalties from UpToDate.

  • Ethics approval Approval for the study was granted by the Emory University institutional review board.

  • 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.