Article Text


Original research
Higher volume endovascular stroke centers have faster times to treatment, higher reperfusion rates and higher rates of good clinical outcomes
  1. Rishi Gupta1,
  2. Anat Horev2,
  3. Thanh Nguyen3,
  4. Dheeraj Gandhi4,
  5. Dolora Wisco5,
  6. Brenda A Glenn1,
  7. Ashis H Tayal6,
  8. Bryan Ludwig7,
  9. John B Terry7,
  10. Raphael Y Gershon8,
  11. Tudor Jovin2,
  12. Paul F Clemmons9,
  13. Michael R Frankel1,
  14. Carolyn A Cronin10,
  15. Aaron M Anderson1,
  16. Muhammad Shazam Hussain5,
  17. Kevin N Sheth10,
  18. Samir R Belagaje1 ,
  19. Melissa Tian6,
  20. Raul G Nogueira1
  1. 1Department of Neurology, Emory University School of Medicine, Marcus Stroke and Neuroscience Center, Grady Memorial Hospital, Atlanta, GA, USA
  2. 2Stroke Institute, Department of Neurology, University of Pittsburgh Medical Center, Pittsburgh, PA, USA
  3. 3Departments of Neurology and Radiology, Boston University School of Medicine, Boston, MA, USA
  4. 4Department of Radiology, University of Maryland School of Medicine, Baltimore, MD, USA
  5. 5Cerebrovascular Center, The Cleveland Clinic Foundation, Cleveland, OH, USA
  6. 6Department of Neurology, Allegheny General Hospital, Pittsburgh, PA, USA
  7. 7Departments of Radiology and Neurology, Miami Valley Hospital, Wright State University Boonshoft School of Medicine, Dayton, OH, USA
  8. 8Department of Anesthesiology, Emory University School of Medicine, Atlanta, GA, USA
  9. 9Department of Radiology, Vanderbilt University School of Medicine, Nashville, TN, USA
  10. 10Department of Neurology, University of Maryland School of Medicine, Baltimore, MD, USA
  1. Correspondence to Dr Raul G Nogueira, 49 Jesse Hill Jr Drive, SE, Faculty Office Building #396, Atlanta, GA 30303, USA;rnoguei{at}


Background and purpose Technological advances have helped to improve the efficiency of treating patients with large vessel occlusion in acute ischemic stroke. Unfortunately, the sequence of events prior to reperfusion may lead to significant treatment delays. This study sought to determine if high-volume (HV) centers were efficient at delivery of endovascular treatment approaches.

Methods A retrospective review was performed of nine centers to assess a series of time points from obtaining a CT scan to the end of the endovascular procedure. Demographic, radiographic and angiographic variables were assessed by multivariate analysis to determine if HV centers were more efficient at delivery of care.

Results A total of 442 consecutive patients of mean age 66±14 years and median NIH Stroke Scale score of 18 were studied. HV centers were more likely to treat patients after intravenous administration of tissue plasminogen activator and those transferred from outside hospitals. After adjusting for appropriate variables, HV centers had significantly lower times from CT acquisition to groin puncture (OR 0.991, 95% CI 0.989 to 0.997, p=0.001) and total procedure times (OR 0.991, 95% CI 0.986 to 0.996, p=0.001). Additionally, patients treated at HV centers were more likely to have a good clinical outcome (OR 1.86, 95% CI 1.11 to 3.10, p<0.018) and successful reperfusion (OR 1.82, 95% CI 1.16 to 2.86, p<0.008).

Conclusions Significant delays occur in treating patients with endovascular therapy in acute ischemic stroke, offering opportunities for improvements in systems of care. Ongoing prospective clinical trials can help to assess if HV centers are achieving better clinical outcomes and higher reperfusion rates.

Statistics from

The development of primary stroke centers has helped to reduce door to needle time for patients eligible for intravenous tissue plasminogen activator (t-PA).1 The recent publication of proposed metrics for comprehensive stroke centers (CSC) includes assessment of time to obtain multimodal imaging, time from arrival at an institution to groin puncture and symptomatic intracerebral hemorrhage (sICH) rates.2 Currently, only limited data are available on how quickly centers are able to achieve ‘door to puncture’ times. Primary stroke centers are currently required to track the percentage of patients with ‘door to administration of intravenous t-PA’ time to <60 min, but less than one-third of hospitals achieve this metric.3 Centers that administer intravenous t-PA for acute ischemic stroke more than 20 times a year have a twofold higher rate of achieving the door to needle metric.3 The American Heart Association and American Stroke Association have developed a national quality improvement initiative called ‘Target: Stroke’ to help improve the timely delivery of thrombolytic treatment to eligible patients.4

Successful achievement of CSC metrics is highly dependent on the ability to deliver efficient endovascular reperfusion treatments. Improvements in the process will occur at serial points throughout patient assessment including hospital-to-hospital transfer, image acquisition, interpretation and decision making, activation of the endovascular team, anesthesia support and rapid reperfusion. While individual milestones will require specific solutions, an interdisciplinary team approach will be required to achieve the global metric. Endovascular stroke volume is a candidate surrogate for this global assessment. High-volume (HV) centers performing specialized procedures (or time-dependent treatments) such as percutaneous coronary interventions5 ,6 or high-risk surgical procedures7 have reported lower mortality rates.

We hypothesize that patients treated at HV centers have lower times from CT acquisition to groin puncture, procedural times and higher reperfusion rates.


This study is a retrospective analysis of consecutive patients treated with endovascular therapy at nine tertiary stroke centers from September 2009 to July 2011. Approval was obtained from each institution's institutional review board as only de-identified information was used for purposes of this analysis. Patients who presented within 8 h from symptom onset with an anterior circulation large vessel occlusion involving the internal carotid artery (ICA) or middle cerebral artery (MCA) were included in the study. Of 562 patients screened, 72 (12.8%) presented with posterior circulation strokes, 39 (8.8%) presented with a wake up stroke or unknown time of onset and nine (2.0%) had distal occlusions involving M3 MCA branches or A2 anterior cerebral artery occlusions. Thus, a total of 442 patients were included in the analysis.

Data were collected regarding demographics (age and sex), previous medical history (hypertension, atrial fibrillation, diabetes mellitus and dyslipidemia), radiographic interpretation of hemorrhages and final infarct volume, location of thrombus on angiography, reperfusion status and clinical outcomes. Additionally, the time interval for each milestone from image acquisition to reperfusion was acquired to determine how quickly each process took for the individual patient. The time from CT completion to groin puncture, time from groin puncture to guide catheter placement in the target vessel, the time from the first angiographic image revealing the treatable occluded segment to placement of the microcatheter into or distal to the thrombus and total procedural time were recorded. Total procedure time was used as a surrogate for time to reperfusion. If there was failed reperfusion, the time of cessation of the procedure was demarcated as procedural time. Centers performing more than 50 endovascular intra-arterial stroke interventions annually were designated as HV centers. This threshold was selected as the median across the nine institutions. Successful reperfusion was defined as achieving a score of TICI 2B on the final angiographic image. sICH was defined as a parenchymal hematoma type 2 using the European Cooperative Acute Stroke Study (ECASS) definition, while a sICH was defined as hemorrhagic infarction types 1 or 2 or parenchymal hematoma type 1. Given the challenges of discerning deterioration in the NIH Stroke Scale (NIHSS) score retrospectively, particularly in intubated patients, this designation was felt to best approximate rates of sICH. Patients with a modified Rankin score of ≤2 between 90 and 120 days were considered to have a good clinical outcome.

Reperfusion status, hemorrhage type and outcomes were adjudicated locally at each individual center. Given the retrospective nature of this study, centers were not aware of the time metrics being captured.

Statistical analysis

Baseline characteristics were compared between the HV and lower volume centers using the Fisher exact test for categorical variables and Student t test or Mann–Whitney U tests for non-parametric continuous variables as appropriate. A secondary analysis was performed comparing patients with good clinical outcomes to those with poor clinical outcomes and predictors of successful reperfusion. In both analyses, variables with a p value <0.20 were entered into the multivariate binary logistic regression model to determine independent predictors of a good clinical outcome and successful reperfusion. All statistical analyses were performed using SPSS V.18.0.


A total of 442 patients with a mean age of 66±14 years and a median NIHSS of 18 (IQR 14–21) were analyzed. The distribution of the vessels treated was as follows: 308 patients (70%) M1 MCA, 78 patients (18%) ICA terminus, 54 patients (12%) M2 MCA, two patients (0.4%) with isolated extracranial ICA occlusion. Forty-seven patients (11%) had tandem extracranial ICA and intracranial MCA or ICA terminus occlusion. Overall, the median time from acquisition of a CT image to groin puncture was 96 min (IQR 53–149), the median time from groin puncture to guide catheter injection of the target vessel was 13 min (IQR 8–20), the time from the guide catheter run to placement of the microcatheter in the thrombus was 18 min (IQR 11–32) and total procedure time was 95 min (IQR 65–134). A total of 218 patients (49%) were found to have a time from CT to puncture of <95 min, with 175 (80%) of these patients treated at HV centers and 43 (20%) at lower volume centers (p<0.0001). When considering patients who were not transferred from an outside facility, only 20 of 112 patients (17.8%) at lower volume centers would achieve the current 120 min door to puncture metric compared with 40 of 100 patients (40%) at HV centers. We did not have access to hospital arrival time data, so we assessed the CT to puncture time as 95 min assuming a door to CT time of 25 min which most primary stroke centers are able to achieve.

A total of 230 patients (52%) were transferred from an outside hospital for an endovascular procedure. A significantly larger proportion of patients treated at HV centers were transfers from outside hospitals (64% vs 31%, p<0.001). Patients who were transferred from an outside hospital did not have significant differences in procedure times (97±52 min vs 112±57 min, p<0.13), but did have shorter times from CT to groin puncture than patients coming from inside the institution (87±69 min vs 139±77 min, p<0.01).

Table 1 summarizes the univariate analysis comparing baseline differences between HV and lower volume centers. There were no significant differences in the two groups with regard to age, baseline NIHSS, location of thrombus and hemorrhage rates. A higher proportion of patients treated at HV centers were transferred from an outside facility, were given intravenous t-PA, had higher reperfusion rates and better clinical outcomes. Additionally, patients at HV centers had shorter times to each milestone measured from CT acquisition to reperfusion and smaller median infarct volumes. Table 2 summarizes the OR of procedure times and time from CT acquisition to groin puncture time, which was lower in HV centers after adjusting for NIHSS, age, use of intravenous t-PA, location of thrombus and if the patient was transferred from an outside facility.

Table 1

Univariate comparison of high-volume centers performing intra-arterial stroke therapies with lower volume centers

Table 2

ORs for procedure time and CT to puncture time in high-volume centers compared with lower volume centers

Univariate analysis was performed to determine predictors of good clinical outcomes and predictors of successful reperfusion. Variables with a p value <0.20 were entered into a binary logistic regression model to determine independent predictors of good clinical outcome and successful reperfusion. This analysis confirmed that patients treated at HV centers were more likely to have good clinical outcomes (OR 1.86, 95% CI 1.11 to 3.10, p<0.018; table 3) and successful reperfusion (OR 1.82, 95% CI 1.16 to 2.86, p<0.008; table 4).

Table 3

Independent predictors of a good clinical outcome after intra-arterial treatment for acute ischemic stroke

Table 4

Independent predictors of successful reperfusion during intra-arterial treatment for acute ischemic stroke


This study is the first multicenter attempt to describe the relevant times for serial metrics at each interval in endovascular stroke treatment. Centers performing higher volumes of acute stroke interventions have lower procedural times, higher reperfusion rates and better clinical outcomes. A better understanding of time delays to treatment allows for the development of realistic metrics as well as opportunities to improve efficiencies within local health systems. Given that time to treatment is an important factor in improving outcomes, a systematic approach is required to create a multidisciplinary collaborative effort at achieving lower times.

The sequence of events required to treat a patient with endovascular therapy for a large vessel occlusion in acute ischemic stroke hinges upon the coordination of an efficient transfer process to CSC, activation of the stroke team if in the emergency room, emergency room physician evaluation, neurological evaluation, acquisition and interpretation of radiographic imaging, activation of an anesthesiology team and an endovascular team to perform the procedure. If each step occurs in series, substantial time delays may result, particularly in off hours and weekends. Unfortunately, in a retrospective study design it is difficult to identify precisely the delays to treatment, but our analysis segments blocks of time in the care of the patient to better identify targets for future improvement.

A recent analysis of patients transferred from an outside facility via ambulance to a tertiary hospital found a direct correlation between delays in transferring a patient and utilization of endovascular therapy. The authors found a 2.5% decrease in offering treatment for every minute elapsed.8 Much of the focus on times to reperfusion has been on time from symptom onset to reperfusion. However, increased granularity in temporal milestones drives the development of milestone-related process improvements. One study considered time from CT imaging to microcatheter placement and found a mean time of 174±60 min, with level I trauma centers providing more efficient care.9 A second study found the median time from CT to groin puncture with arrivals in the emergency room to be 104 min, with significant delays noted on weekends and after hours.10 The current study found a similar CT to microcatheter time of 151±90 min overall (median 136, IQR 82–193), but HV centers were noted to have significantly reduced times of 119±71 min compared with 204±96 min (p<0.0001). Additionally, the recent proposed CSC metrics suggest a door to groin puncture time of 2 h.2 We did not have data on time of arrival, but if we add 25 min to estimate door to CT time, we estimate that 28% of patients met the door to puncture metric when patients arrived in the institution's emergency room, but significantly more HV centers achieved this metric than lesser volume centers (40% vs 17.8%, p<0.001).

In order to improve efficiencies of patients being brought to angiography, it will require infrastructural changes that include collaboration among various specialties with support of administrators providing real-time feedback to clinicians.11 Achieving rapid reperfusion once the patient is brought to the angiography suite will require experienced operators, innovative devices and clinical support with anesthesia and neurocritical care. One group considered angiographic procedural times in a small cohort of 34 patients and found a mean time of 101 min,12 which is consistent with our cohort, although HV centers were noted to have significantly lower procedure times. When looking at the time points closely, the greatest delay in achieving reperfusion is initiation of the procedure after imaging. The etiologies of these delays is beyond the scope of this analysis, but may be due to acquiring advanced modality imaging, activation of the endovascular team, operator experience and anesthesia delays. We were unable to collect data on transfer time, but a previous study at a single institution noted a median time of 104 min for transfer. Although this is a significant amount of time, HV centers may make up for some of this time as the median time from CT acquisition to groin puncture and CT acquisition to reperfusion was significantly lower (70 (IQR 43–120) min vs 146 (IQR 92–195) min and 158 (IQR 111–219) min vs 268 (IQR 216–338) min, respectively).

Time delays to reperfusion have a profound impact on clinical outcomes. For every 30 min delay in reperfusion there is a 12% lower probability of a good clinical outcome.13 Although there are patients with excellent collaterals who may have a longer duration of adequate perfusion to the territory affected by the occlusion, as a whole population of patients presenting with acute symptoms, efficient reperfusion is desirable. Second, successful reperfusion has been associated with a higher proportion of patients with a good clinical outcome in several studies.14 ,15 Additionally, longer durations of ischemia from time delays may lead to a larger core infarct which has also been implicated in lower recanalization rates.16 Lastly, operator experience may play a role as, with many procedurally-based specialties, experience in treating a specific disease may be associated with more timely and efficient decision making to achieve efficient reperfusion.

There are several limitations to this analysis. First, we did not assess the time of day the patient was being treated. Patients presenting during off hours or at weekends theoretically may have been more common in the lower volume centers which might account for these differences, although this is not likely. It is important to note that there was no statistical difference in time from symptom onset to treatment between the two centers. Second, we were not able consistently to identify the time the patient arrived at the hospital. This would be the ideal starting point to assess the timing of the sequence of events in a patient's care but was difficult to obtain in this retrospective analysis. Each center is a primary stroke center and would thus have a substantial number of patients in the CT scanner within 25 min of presentation. Third, patients were enrolled in clinical trials in some instances which may lead to delays in the care of the patient due to imaging protocols and the consent process, thereby potentially contributing to delays. Fourth, given the retrospective nature of this study, images were interpreted by the physicians at each individual institution and were not adjudicated by a central core laboratory. Interpretation of reperfusion and hemorrhages was thus performed by the local investigator. This would not have an impact on the final clinical outcome or the timings of the sequence of events leading to and during the procedure, but may have affected the reperfusion and hemorrhage rates reported at each center.

In conclusion, we found significant opportunities for health systems to improve time delays in treating patients with endovascular stroke therapies. Centers that currently perform higher numbers of these procedures currently appear to have lower times to reperfusion and a higher proportion of patients with successful reperfusion translating into improved clinical outcomes. A prospective study should be undertaken to examine these sequential steps and implementation of strategic parameters in order to reduce times to treatment.


View Abstract


  • Correction notice This article has been corrected since it was published Online First. The author list and the corresponding author email address have been amended.

  • Competing interests None.

  • Ethics approval Ethics approval was obtained from the IRB of each institution.

  • 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.