Article Text

Download PDFPDF

Original research
A multicenter study of the safety and effectiveness of mechanical thrombectomy for patients with acute ischemic stroke not meeting top-tier evidence criteria
  1. Nitin Goyal1,
  2. Georgios Tsivgoulis1,2,3,
  3. Donald Frei4,
  4. Aquilla Turk5,
  5. Blaise Baxter6,
  6. Michael T Froehler7,
  7. J Mocco8,
  8. Jay Vachhani9,
  9. Daniel Hoit9,
  10. Lucas Elijovich1,9,
  11. David Loy4,
  12. Raymond D Turner5,
  13. Justin Mascitelli8,
  14. Kiersten Espaillat7,
  15. Andrei V Alexandrov1,
  16. Anne W Alexandrov1,10,
  17. Adam S Arthur9
  1. 1Department of Neurology, University of Tennessee Health Science Center, Memphis, Tennessee, USA
  2. 2Second Department of Neurology, “Attikon University Hospital”, School of Medicine, National and Kapodistrian University of Athens, Athens, Greece
  3. 3International Clinical Research Center, St. Anne's Hospital, Brno, Czech Republic
  4. 4Department of Interventional Neuroradiology, Radiology Imaging Associates, Englewood, Colorado, USA
  5. 5Department of Neurosurgery, Medical University of South Carolina, Charleston, South Carolina, USA
  6. 6Department of Interventional Neuroradiology, Erlanger Hospital, Chattanooga, Tennessee, USA
  7. 7Cerebrovascular Program, Vanderbilt University, Nashville, Tennessee, USA
  8. 8Department of Neurosurgery, Mount Sinai Medical Center, New York, New York, USA
  9. 9Department of Neurosurgery, University of Tennessee Health Science Center and Semmes-Murphey Clinic, Memphis, Tennessee, USA
  10. 10Australian Catholic University, Sydney, New South Wales, Australia
  1. Correspondence to Dr Adam S Arthur, Department of Neurosurgery, Semmes-Murphey Clinic, 6325 Humphreys Blvd, Memphis, TN 38120, USA; aarthur{at}


Background While mechanical thrombectomy (MT) has become the standard of care for patients with acute ischemic stroke (AIS) with emergent large-vessel occlusions (ELVO), recently published guidelines appropriately award top-tier evidence to the same selective criteria that were employed in completed clinical trials. We sought to evaluate the safety and effectiveness of MT in patients with AIS with ELVO who do not meet top-tier evidence criteria (TTEC).

Methods We conducted an observational study on consecutive patients with AIS with ELVO who underwent MT at six high-volume endovascular centers. Standard safety outcomes (3-month mortality, symptomatic intracranial hemorrhage) and effectiveness outcomes (3-month functional independence: modified Rankin Scale scores of 0–2) were compared between patients meeting and failing TTEC.

Results The sample consisted of 349 (60%) controls fulfilling TTEC and 234 (40%) non-TTEC patients. Control patients meeting TTEC for MT tended to have higher functional independence rates at 3 months (47% vs 39%; p=0.055), while the rates of symptomatic intracerebral hemorrhage (sICH) were similar (9%) in both groups (p=0.983). In multivariable logistic regression models, adherence to TTEC for MT was not independently related to any safety outcome (sICH: OR 0.71, 95% CI 0.30 to 1.68, p=0.434; 3-month mortality: OR 1.27, 95% CI 0.69 to 2.33, p=0.448) or effectiveness outcome (3-month functional independence: OR 0.81, 95% CI 0.48 to 1.37, p=0.434; 3-month functional improvement: OR 0.73, 95% CI 0.48 to 1.11, p=0.138) after adjusting for potential confounders.

Conclusions Approximately 40% of patients with AIS with ELVO offered MT do not fulfill TTEC for MT. Patients who did not meet TTEC had high rates of good clinical outcome and low complication rates.

  • Stroke
  • Thrombectomy

Statistics from

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.


The natural history for patients with acute ischemic stroke (AIS) from emergent large vessel occlusion (ELVO) is dismal, resulting in permanent disability or death.1 ,2 The recent publication of five major randomized controlled clinical trials (RCTs) provided overwhelming evidence that timely mechanical thrombectomy (MT) in patients with AIS due to ELVO is safe and improves functional outcomes.3–7 With a number needed to treat of 2.6, MT has become the new standard of care in the USA and abroad for patients with ELVO.8 Guidelines supporting thrombectomy were recently released,9 awarding top-tier evidence to the same set of restrictive criteria that were used in recent positive MT RCTs. Data regarding the prevalence of patients with ELVO who meet these top-tier evidence criteria (TTEC) are lacking, and it remains unclear whether MT holds any benefit at 3 months for patients with ELVO with presentations outside current top-tier evidence recommendations.

We recently reported that nearly 50% of patients who underwent MT would have been denied the procedure if top-tier guidelines were strictly followed at a single center.10 Interestingly, 33% of cases treated outside these guidelines achieved functional independence at 3 months and, importantly, the procedure was safe in this cohort without excessive risk for symptomatic intracerebral hemorrhage (sICH).10 While exceptions for the use of MT in cases not meeting these guidelines might be advocated by some, it remains unknown whether other healthcare providers, payors, or centers will deny patients treatment if their cases do not meet requirements stipulated in the US guidelines. Given the potential for beneficial outcomes, we sought to examine current MT practices at high-volume centers, prospectively evaluating both safety and effectiveness in patients with AIS with ELVO not fulfilling TTEC.


We analyzed an institutional review board-approved multicenter database of MT for AIS to determine the effectiveness of MT in patients outside top-tier evidence recommendations. Data on consecutive stroke cases with ELVO who underwent MT were collected at six high-volume endovascular centers during a 2-year period (May 2013–May 2015) (Appendix 1).

Patient selection

All cases were coded as either meeting or failing TTEC (Class I, Level of evidence A) based on their conformance to the following criteria that were advocated in the recent American Heart Association (AHA) guidelines:9

  • Pre-stroke baseline modified Rankin scale (mRS) score 0–1

  • AIS receiving treatment with IV thrombolysis

  • Causative occlusion of the internal carotid artery or proximal (M1) middle cerebral artery (MCA)

  • Age 18 years or older

  • NIH Stroke Scale (NIHSS) score of ≥6

  • Treatment initiated (groin puncture) within 6 hours of symptom onset

Although the Alberta Stroke Program Early CT Score (ASPECTS) >6 was also included in the AHA guidelines, participating centers routinely excluded cases not meeting this criterion; therefore, we did not include this variable in our selection criteria. The non-TTEC group included patients with confirmed ELVO on CT angiography (CTA) who did not meet top-tier AHA evidence for MT (distal anterior circulation occlusions (M2 or M3 MCA or anterior cerebral artery), posterior circulation occlusions, NIHSS score <6, symptoms onset to groin puncture time >6 hours, and pre-stroke baseline mRS score of >1).

Baseline characteristic

Baseline data included age, sex, admission and discharge NIHSS scores, and pretreatment with IV tissue-type plasminogen activator. Procedural technical details/efficiency data included onset to groin puncture time, groin puncture to recanalization time, and type of primary endovascular therapy (stent retriever or direct aspiration).


Our primary safety endpoint was symptomatic intracranial hemorrhage (sICH), defined as the presence of a parenchymal hematoma type 2 on brain CT and/or MRI gradient recall echo sequence, accounting for deterioration with an increase in NIHSS score of ≥4 points within 36 hours from treatment.11 Other serious complications (ie, vessel dissection/perforation, vasospasm or groin hematoma) were also reported. The primary effectiveness endpoint was 3-month functional independence, defined as a mRS score of 0–2. Other secondary effectiveness endpoints were: (1) successful recanalization, defined as a Thrombolysis in Cerebral Infarction (TICI) score of 2b/3 as previously described;10 ,12 (2) neurologic improvement during hospitalization, quantified as the relative decrease in NIHSS score at hospital discharge in comparison with hospital admission ([NIHSSadm−NIHSSdis]/NIHSSadm×100%);13 ,14 and (3) shift in the distribution of 3-month mRS scores.15 Both endovascular specialists and NIHSS/mRS-certified assessors of stroke severity and functional outcomes were unaware of the purposes of the study and performed treatments and assessments as part of their clinical duties.

Statistical analyses

Statistical comparisons were performed between patient subgroups using the χ2 test, Fisher's exact test, independent Student t-test, and Mann–Whitney U test, as indicated for dichotomous or continuous variables. The distribution of the mRS score at 3 months among patients with ELVO was compared between different subgroups using both the Cochran–Mantel–Haenszel test and univariable/multivariable ordinal logistic regression (shift analysis).15 Multivariable logistic regression models were used to evaluate associations between endovascular reperfusion therapies performed in accordance with or without top-tier evidence (AHA Class I/Level A) recommendations, with 3-month functional independence, 3-month mortality, and sICH before and after adjusting for potential confounders, including admission NIHSS score, pretreatment with IV thrombolysis, location of occlusion, and onset to groin puncture time. Associations are presented as ORs with corresponding 95% CIs. Statistical significance was achieved if the p value was ≤0.05 in multivariable logistic regression analyses. The Statistical Package for Social Science V.11.5 for Windows (SPSS, Chicago, Illinois, USA) was used for statistical analyses.


Baseline characteristics

A total of 583 patients with AIS due to ELVO (mean age 62±19 years; 50% men; median (IQR) admission NIHSS score 17 (14–21) points) underwent MT during the study period. Systemic thrombolysis prior to MT was administered in 292 patients (50%).

For the overall sample, complete recanalization (TICI 2b or 3) was achieved in 69% (n=403) of patients. Three-month functional outcome data were available in 510 patients (87%), with functional independence at 3 months achieved in 44% (n=222) and death within 3 months occurring in 24% (n=120). A total of 50 patients (9%) experienced sICH.

The baseline characteristics of the two study groups are compared in table 1. A total of 349 cases (60%) were classified as controls, filling TTEC for MT (mean age 63±18 years; 47% men; median (IQR) admission NIHSS score 17 (14–21) points), whereas 234 patients (40%) were classified as non-TTEC cases (mean age 62±19 years; 53% men; median (IQR) admission NIHSS score 16 (9–21) points). The two most common reasons for failing TTEC were location of intracranial occlusion (n=144) and falling outside the 6-hour treatment window (n=108). Other reasons included 23 patients (4%) with a pretreatment NIHSS score <6, 13 (2%) with a premorbid mRS >2, while 53 (9%) subjects had two or more characteristics outside the guidelines. The locations of proximal intracranial occlusions in non-TTEC patients were: posterior circulation (n=72); M2 MCA (n=66); M3 MCA (n=2); and anterior cerebral artery (n=4).

Table 1

Baseline characteristics and outcomes (effectiveness and safety) of the study population stratified according to compliance with top-tier AHA evidence for MT


Briefly, control patients meeting top-tier evidence had higher pretreatment stroke severity (p=0.006) and were more frequently pretreated with systemic thrombolysis (p<0.001). Moreover, the median onset to groin puncture time was shorter in this group than in non-TTEC patients (p<0.001). Complete recanalization rates and median groin to revascularization time did not differ between the two groups. The rates of sICH and other serious complications were also similar in the two groups. Control patients meeting TTEC for MT experienced greater neurological improvement during hospitalization (median relative reduction in NIHSS score at discharge 75% vs 57%; p=0.018) and tended to have higher functional independence rates at 3 months (47% vs 39%; p=0.055), while 3-month mortality was higher in non-TTEC cases (30% vs 19%; p=0.006). Controls had greater functional improvement at 3 months (p=0.013, Cochran–Mantel–Haenszel test) according to the distribution of the 90-day mRS scores (figure 1A).

Figure 1

Distribution in 3-month modified Rankin Scale (mRS) scores according to (A) adherence to American Heart Association (AHA) top-tier evidence criteria; (B) location of occlusion criteria; and (C) treatment time window criteria. The reported p values have been calculated using the Cochran–Mantel–Haenszel (CMH) test.

Since the two most common reasons for non-adherence to top-tier evidence recommendations were location of occlusion and delayed treatment window, we also performed sensitivity analyses of safety and effectiveness outcomes in subgroups stratified on the basis of the AHA guidelines for location of occlusions and time from onset to groin puncture (table 2). Non-TTEC cases—due to location of occlusion—had higher 3-month mortality rates (34% vs 20%; p=0.002), despite a lower prevalence of sICH (4% vs 10%; p=0.029), in comparison with the group of patients treated according to TTEC. Basilar artery occlusion was identified in 31 of the 49 (63%) non-TTEC cases who died during the 3-month follow-up period. The two groups did not differ (p>0.05) in terms of complete recanalization, other serious complications, and 3-month functional independence rates. There was a trend towards greater 3-month functional improvement in controls fulfilling TTEC (p=0.055, Cochran–Mantel–Haenszel test; figure 1B). Stratification by adherence to the 6 hour window for onset to groin puncture time had similar complete recanalization and 3-month mortality rates; however, patients treated outside the recommended time window had higher sICH (14% vs 7%; p=0.029) and lower 3-month functional independence (34% vs 46%; p=0.040) rates. The two groups did not differ (p=0.110, Cochran–Mantel–Haenszel test) in terms of distribution in 3-month mRS scores (figure 1C).

Table 2

Baseline characteristics, safety and effectiveness outcomes of the study population stratified according to compliance with AHA recommendations for MT in relation to location of intracranial occlusion and elapsed time between symptom onset and groin puncture

Multivariable analyses

Table 3 summarizes compliance with TTEC in relationship to sICH, 3-month mortality, and 3-month functional independence using multivariable logistic regression models adjusting for potential confounders (admission stroke severity, pretreatment with IV thrombolysis, location of occlusion, and onset to groin puncture time). We also evaluated the association between compliance with TTEC and 3-month functional improvement (shift analysis) using multivariable ordinal regression analyses. Adherence to TTEC for MT was not independently related to any safety (sICH and 3-month mortality) or effectiveness (3-month functional independence and 3-month functional improvement) outcome. Similarly, there was no independent association between top-tier evidence adherence and location of occlusion, onset to groin puncture time, safety, or effectiveness outcomes.

Table 3

Association of compliance with AHA recommendations for MT with outcomes on multivariable logistic regression models and multivariable ordinal regression models adjusting for potential confounders

Subgroup analyses

We also evaluated the association between compliance with TTEC and efficacy (complete recanalization, 3-month functional independence) and safety (sICH, 3-month mortality) outcomes in patients with anterior circulation occlusions (n=551). The two groups (with and without adherence to TTEC) did not differ in any of the safety or efficacy outcomes (table 4). In addition, we compared safety and efficacy outcomes in the following subgroups using univariable and multivariable logistic regression analyses:

  1. Cases meeting TTEC recommendations for MT (n=349) with cases not meeting AHA recommendations for location of intracranial occlusion (n=144); results are presented in online supplementary table S1.

  2. Cases meeting TTEC recommendations for MT (n=349) with cases not meeting AHA recommendations for onset to groin puncture time (n=108); results are presented in online supplementary table S2.

  3. Cases meeting TTEC recommendations for MT (n=349) with posterior circulation occlusion cases (n=72); results are presented in online supplementary table S3.

  4. Cases meeting TTEC recommendations for MT (n=349) with distal MCA occlusion cases (n=68); results are presented in online supplementary table S4.

  5. Cases meeting TTEC recommendations for MT (n=349) with cases not meeting both AHA recommendations for onset to groin puncture time and for location of intracranial occlusion (n=30); results are presented in online supplementary table S5.

Table 4

Outcomes of patients with anterior circulation occlusions (n=511) stratified according to compliance with top-tier AHA evidence for MT

We detected no independent association in any of the subgroup analyses comparing safety and efficacy outcomes between cases meeting TTEC for MT (n=349) and different subgroups of patients not meeting specific TTEC recommendations (see online supplementary tables S1–S5).


Our study found that 40% of our sample would not have been offered MT had adherence to currently published guidelines citing top-tier evidence been maintained. In unadjusted analysis, adherence to TTEC for MT results in superior outcomes compared with non-TTEC cases. However, our adjusted analyses show that non-TTEC patients benefit similarly from MT and, importantly, at no increase in risk. Given the likely devastating disability associated with untreated ELVO, these findings are provocative and important to reconcile with those from the five key MT trials that support current top-tier evidence, especially when experienced neurointerventionalists are available to oversee care. Our findings are in concordance with our single-center pilot study which demonstrated that the safety and outcomes of MT were not related to strict adherence to recent top-tier evidence recommendations.10

The most common reason for not meeting TTEC was the location of the occlusions (ie, distal anterior circulation or posterior circulation occlusions). Our study indicates that MT can be performed safely and effectively in this group. Although unadjusted mortality rates were higher in this subgroup, after adjusting for potential confounders there was no association between adherence to TTEC for location of occlusion and the likelihood of 3-month mortality in multivariable analyses. In addition, the higher mortality rates for the subgroup of patients not meeting TTEC in our cohort can be largely explained by the inclusion of patients with basilar artery occlusions. Despite the higher mortality of basilar artery occlusions after MT, these patients should be offered timely intervention, considering the extremely poor natural history of the untreated disease with mortality rates up to 80%.16 Several investigators have shown that timely recanalization improves functional outcome and decreases mortality in patients with acute basilar artery occlusions.17 ,18 Our findings are also supported by a multicenter study which reported favorable outcomes in 35% of patients with AIS with posterior circulation ELVO treated with stent retrievers or primary aspiration thrombectomy.19 On the other hand, little is known about the natural history of patients with distal anterior circulation occlusions (M2 or M3 MCA). The question of the benefit of MT for patients with more distal occlusions was not answered by the recent randomized trials, which restricted the enrollment of these patients.8 However, a recent observational report by Sarraj et al20 showed that MT was associated with higher rates of good functional outcome in patients with distal (M2) MCA occlusions in comparison with standard therapy (63% vs 35%), corroborating our observations.

The second common reason for not meeting TTEC was delayed treatment window. The adjusted analysis of our data suggests that MT can be performed in this group with equal effectiveness and at no increased cost of sICH and mortality when compared with the group meeting TTEC for treatment time window. Our preliminary observations are partly supported by the findings of the Endovascular Treatment for Small Core and Proximal Occlusion Ischemic Stroke (ESCAPE) trial. Within ESCAPE, 49 patients underwent randomization at ≤6 hours after stroke onset, with the direction of effect in this subgroup favoring MT for mRS 0–2 at 90 days.5 In addition, a recent meta-analysis of individual patient data from five randomized trials demonstrated the benefits of MT in patients with ELVO randomized between 6 and 8 hours from symptom onset.8 The issue of treatment time window and eligibility for MT is of special concern, especially when significant proportions of patients with ELVO are still being transferred from non-interventional stroke centers and arrive outside the 6 hour window. Non-randomized studies have shown the utility of imaging guided selection (collateral scores or perfusion imaging) for MT, especially for patients with delayed presentation.21 Of note, patients with ASPECTS <6 were systematically excluded from treatment within our study cohort and this may have led to the selection of patients with better collaterals. Our group has previously underscored the relationship of good collaterals with favorable functional outcomes in patients with ELVO treated with MT in extended time windows.18 ,22 Unfortunately, this hypothesis cannot be verified since data on collateral status were not collected in this multicenter cohort. Moreover, the ongoing PerfusiOn Imaging Selection of Ischemic STroke PatIents for EndoVascular ThErapy (POSITIVE) trial will provide randomized data regarding the safety and efficacy of MT in patients with ELVO during an extended treatment window of 6–12 hours from symptom onset, as will the Diffusion and Perfusion Imaging Evaluation for Understanding Stroke Evolution (DEFUSE)-3 study for patients in the 6–15 hour time window after onset.

As recently as 3 years ago, MT for patients with ELVO was regarded as unproven and a call was issued for Medicare to deny procedural reimbursement in order to improve enrollment in clinical trials.23 It is important to note that the entire cohort of patients within our data sample might have been denied intervention had these calls been heeded. Although findings from the five successful MT trials3–7 have significantly improved our knowledge of the benefit of MT, we would argue that further calls to stifle the provision of MT outside clinical trials in patients not meeting top-tier evidence may be unjustified, given the grim natural history of severe strokes with persisting occlusions.1 ,16 ,24 Withholding patient access to rescue procedures would further slow the improvement of our systems aimed at reducing costly and devastating disability. Findings from both clinical trials and exploratory observational work enable a more complete understanding of what constitutes best medical practice. Evidence-based medicine is a commonly cited and laudable goal with the presumption being that judicious use of the best evidence in making decisions about the care of individual patients will help to ensure the best outcomes. However, the decision to offer specific therapies to individuals in emergency situations is often more complex than completion of an evidence-based checklist. In fact, one of the tenets of evidence-based medicine is that the results from clinical trials and also the operator's technical experience should be shared with patients and families so that a thoroughly informed decision may be reached. Patients and their families should most certainly be informed of the expected outcome if treatment is withheld.

Several limitations of our work should be acknowledged. First, our methods were observational and included self-reported safety and effectiveness outcomes lacking central adjudication of imaging or clinical outcomes. Second, because all centers excluded patients with ASPECTS <6, we did not collect data on ASPECTS from baseline CT nor did we collect collateral scores on pretreatment CTA. Third, the specific devices and methods used for MT were heterogeneous and were selected according to the treating physicians' preference. Fourth, 3-month functional outcome was available in 87% of our study population. However, our large sample size and our highly experienced multicenter investigators make our findings compelling. Additionally, we did not have a control group which was managed medically so we could not compare the safety and effectiveness of MT in the non-TTEC group with standard medical management. Finally, our multicenter approach to MT for patients with AIS with ELVO reflects everyday clinical practice experience, which plays an essential role in increasing the availability of endovascular treatment in patients with AIS. Given the high morbidity and mortality of untreated ELVO, it is fortunate that MT is now recognized as standard of care and is continuing to rapidly evolve. We believe that systems of care, patient selection, and the MT procedure itself are all poised for significant advances that will result in dramatic improvements in patient outcomes.25 ,26

Our findings provide evidence that MT may be offered to patients with ELVO who do not meet current TTEC without compromising patient safety and with reasonable effectiveness.


The authors wish to thank Andrew J Gienapp (Department of Medical Education, Methodist University Hospital, Memphis, TN and Department of Neurosurgery, University of Tennessee Health Science Center, Memphis, Tennessee, USA) for copy editing, preparation of the manuscript and figure for publishing, and publication assistance.

Appendix 1. Participating centers and numbers of patients contributed by each center

Participating centerN (%)
Radiology Imaging Associates, Denver, CO190 (33)
University of Tennessee, Memphis, TN104 (18)
Medical University of South Carolina, SC89 (15)
Erlanger Medical Center, Chattanooga, TN78 (13.5)
Vanderbilt Medical Center, Nashville, TN78 (13.5)
Mount Sinai Institute, New York, NY44 (7)



  • Competing interests ASA is a consultant for Codman, Medtronic, Microvention, Penumbra, Sequent, Siemens, Stryker and has received research support from Sequent and Siemens. BB serves on the speakers' bureaus of Medtronic, Penumbra, Pulsar, Silk Road, and Stryker. LE is a consultant for Codman Neurovascular, Medtronic, MicroVention, Penumbra, Sequent, and Stryker. DF is a consultant for Codman, MicroVention, Penumbra, and Stryker; is a member of the speaker's bureaus for Codman, MicroVention, Penumbra, and Stryker; and is a stock shareholder of Penumbra. MTF is a consultant for Blockade Medical and Medtronic and has received grant funding from Medtronic, MicroVention, NINDS/NIH, Penumbra, and Stryker. JM has received funding support from Medtronic Neurovascular, Microvention, Penumbra, and Stryker Neurovascular; is an investor with Blockade Medical, Cerebrotech, and TSP; and has served as a consultant for Cerebrotech, Endostream, Pulsar, Rebound, and TSP. AT is a consultant for Medtronic, Penumbra, MicroVention, and Stryker and has received research grants from Medtronic, Penumbra, MicroVention, and Stryker. RDT is a consultant for Blockade, Codman, Devora Medical, Medtronic, MicroVention, Penumbra, Pulsar Vascular, Q'Apel, Rebound Medical, and Stryker.

  • Ethics approval University of Tennessee Acute Stroke Registry.

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