Author + information
- Received October 2, 2018
- Revision received November 26, 2018
- Accepted December 4, 2018
- Published online February 18, 2019.
- Rody El Nawar, MDa,b,
- Bertrand Lapergue, MD, PhDc,d,
- Michel Piotin, MD, PhDe,
- Benjamin Gory, MD, PhDf,g,h,
- Raphael Blanc, MD, MSce,
- Arturo Consoli, MDi,
- Georges Rodesch, MD, PhDi,
- Mikael Mazighi, MD, PhDe,j,k,
- Frederic Bourdain, MDc,
- Maéva Kyheng, BSTl,
- Julien Labreuche, BSTl,
- Fernando Pico, MD, PhDa,d,k,∗ (, )
- on behalf of the ETIS Investigators
- aDepartment of Neurology and Stroke Center, Centre Hospitalier de Versailles, Versailles, France
- bGilbert and Rose-Marie Chagoury School of Medicine, Lebanese American University, Beirut, Lebanon
- cDepartment of Neurology and Stroke Center, Hospital Foch, Suresnes, France
- dVersailles Saint-Quentin en Yvelines and Paris Saclay University, Versailles, France
- eDepartment of Interventional Neuroradiology, Rothschild Foundation, Paris, France
- fDepartment of Interventional Neuroradiology, Hospices Civils de Lyon, Lyon, France
- gDepartment of Diagnostic and Therapeutic Neuroradiology, University Hospital of Nancy, Nancy, France
- hUniversity of Lorraine, INSERM U1254, IADI, Nancy, France
- iDepartment of Diagnostic and Interventional Neuroradiology, Hospital Foch, Suresnes, France
- jParis Denis Diderot University, Paris, France
- kINSERM LVTS (Laboratory for Vascular Translational Science)-1148, Paris, France
- lLille University, CHU Lille, EA 2694 - Santé publique: épidémiologie et qualité des soins, Lille, France
- ↵∗Address for correspondence:
Pr. Fernando Pico, Department of Neurology and Stroke Center, Centre Hospitalier de Versailles, 177 Rue de Versailles, Versailles Saint-Quentin en Yvelines et Paris Saclay University, 78150 Le Chesnay, Versailles, France.
Objectives The aim of this study was to determine whether individual operator characteristics have an impact on reperfusion and procedural complication rates.
Background Mechanical thrombectomy (MT) is a Level IA treatment in acute ischemic stroke (AIS) patients. The operator’s effect has been found to be an independent predictor for clinical outcome and technical performance in interventional cardiology.
Methods From the ETIS (Endovascular Treatment in Ischemic Stroke) study, a prospective, multicenter, observational real-world MT registry, the authors included all AIS patients consecutively treated by MT between January 2012 and March 2017 in 3 high-volume comprehensive stroke centers by 19 operators. We assessed the effect of individual operator characteristics on successful reperfusion, defined as modified Thrombolysis In Cerebral Infarction 2b/3 at the end of MT, and procedural complications using multivariable hierarchical logistic regression models.
Results A total of 1,541 patients with anterior and posterior AIS were enrolled (mean age 67 years; median NIHSS 16). There was a significant operator effect on successful reperfusion, with an intraclass correlation coefficient of 0.036 (p = 0.046), but not on complications (intraclass correlation coefficient = 0). There was a dose–response relationship between annual operator volume and successful reperfusion rate (p = 0.003) with an adjusted odds ratio for successful reperfusion equal to 2.52 (95% confidence interval: 1.37 to 4.64) for patients treated by an operator with an annual volume ≥40 MT/year compared with those treated by an operator with <14 MT/year (first tertile). Nevertheless, this result did not translate to better clinical outcomes.
Conclusions Our data suggest that operator volume of MT/year has a positive impact on successful reperfusion in AIS patients, but not on clinical outcomes nor on complication rates. Further studies are warranted to investigate threshold procedure numbers associated with better outcomes.
Endovascular mechanical thrombectomy (MT) is considered a Class I, Level of Evidence: A recommendation for immediate revascularization of proximal intracranial arterial occlusion of the anterior circulation (1–6). According to The Joint Commission in collaboration with the American Heart Association/American Stroke Association, stroke centers must present precise standards and advanced stroke certification to be considered as MT-capable stroke centers. Moreover, The Joint Commission proposes a minimum volume of 12 MT procedures/year or 24 over the past 2 years accomplished by 2 to 3 experienced operators (7). Nevertheless, these thresholds are not supported by strong evidence. There is a need for studies assessing the relationships between operator experience, operator volume, success reperfusion, and complication rates.
Robust data support the association of a higher complication rate with low operator's experience in several cardiological interventions such as percutaneous coronary intervention, coronary stent placement, and percutaneous left atrial appendage closure (8–10). The main purpose of our study was to analyze the impact of the individual operator characteristics on reperfusion and complication rates at the end of the MT procedure.
The study population was selected from the prospective observational multicenter ETIS (Endovascular Treatment in Ischemic Stroke) registry from 3 large comprehensive stroke centers in France. All patients were identified from the ETIS registry between January 2012 and March 2017 (11) (Online Appendix). The number of operators in this registry was 19. Patients enrolled in this study presented with an acute ischemic stroke (AIS) with occlusion of the internal carotid artery or the middle cerebral artery (M1 and M2 segments) or vertebrobasilar arteries, and were treated with MT by either contact aspiration or stent retrievers. A more detailed description of the technique has been published previously (11). A local ethics committee and the French Data Protection Agency approved the use of patient data for this research protocol. In accordance with French legislation, informed consent was not needed from patients because this study analyzed anonymized data collected prospectively as a part of routine clinical care.
The following clinical characteristics of each patient were analyzed: age, sex, vascular risk factors, treatments before the procedure, vital signs before treatment, the severity of the AIS using the National Institutes of Health Stroke Scale (NIHSS) score at admission and 24 h after the initiation of treatment assessed by a certified medical professional or by a stroke neurologist implicated in the short-term management of the patient post-MT, and the degree of disability or dependence in the daily activities using the modified Rankin Scale (mRS) at day 90. Furthermore, the following radiological imaging characteristics were identified: 1) diffusion-weighted imaging–Alberta Stroke Program Early Computed Tomography Score (ASPECTS) and computed tomography (CT) brain ASPECTS—each ASPECTS region was scored 0 if abnormal and 1 if normal; 2) site of occlusion in the middle cerebral artery M1 or M2; 3) modified Thrombolysis In Cerebral Infarction (mTICI) score; and 4) imaging results 24 h after MT to evaluate the presence of a hemorrhagic transformation.
Concerning treatment and procedure, we assessed: 1) time from symptom onset to initiation of intravenous thrombolysis, to groin puncture, and to final reperfusion; and 2) the use of general or local anesthesia during the MT.
Operator's experience characteristics
Operator characteristics were defined according to the number of MT performed per year during the study and according to the year of the first MT. The name of the operator was checked on the signed interventional report. The experience of the operator was estimated by the year of first MT. Operators were divided into 2 groups according to the median year of first MT (≤2010 or >2010) in this sample.
The primary study outcome was the angiographic success rate defined by a reperfusion of the occluded artery evaluated by the mTICI scores of 2b/3 (considered as restoration of flow to more than 50% of the involved brain territory at the end of the MT procedure). Secondary outcomes include the favorable outcome and procedural complications. Favorable outcome was evaluated by the mRS at 90 days post-MT, during face-to-face interviews or via telephone conversations with the patient, their relatives, or their general practitioner.
The procedural complications during and post-MT collected from the register were: intracranial hemorrhage (ICH), subarachnoid hemorrhage, groin complication (formation of a groin hematoma, retroperitoneal hematoma, femoral artery pseudoaneurysm), distal limb artery occlusion by migrated emboli, recurrent AIS, vessel perforation or dissection, vasospasm, device fracture, infection, and death. We grouped ICH according to the classification of the ECASS 1 (European Cooperative Acute Stroke Study 1) criteria: parenchymal hematoma type 1 and 2, and hemorrhage infarction type 1 and 2 (12). Symptomatic ICH was considered in patients presenting within 24 h with an increase of 4 or more points in their NIHSS score caused by the hemorrhage.
Continuous variables are expressed as mean ± SD in the case of normal distribution or median (interquartile range) otherwise. Categorical variables are expressed as number (percentage). Normality of distributions was assessed using histograms and the Shapiro-Wilk test. The main patient-level characteristics were compared between patients with and without successful reperfusion, with and without a favorable outcome, and between patients with and without procedure complications using the Student’s t-test (or Mann-Whitney U test) for continuous variables and the chi-square test for categorical variables.
To assess the effect of operators on each outcome (successful reperfusion, favorable outcome, and procedural complications), we used a multivariable mixed (hierarchical) logistic regression model including all patient-level characteristics associated with outcome in bivariate analyses (p < 0.20) as fixed effects, and operators as a random effect. We derived from this model the intraclass correlation coefficient (ICC), which represents the outcome variability attributable to individual operators. We further investigated the association of the 2 main operator characteristics (annual volume categorized into 3 levels and operator experience categorized in 2 levels using a cutoff date of January 2010) with each outcome by including the annual operator volume as a covariate into the previous multivariable hierarchical logistic regression model. Odds ratios for middle and upper tertiles relative to the lowest tertile of annual operator volume were calculated as effect size. Statistical testing was conducted at the 2-tailed α-level of 0.05. Data were analyzed using SAS software version 9.4 (SAS Institute, Cary, North Carolina).
From January 2012 to March 2017, 1,541 AIS patients were treated by MT in 3 French comprehensive stroke centers. During the study period, a total of 19 operators performed MT with a median number of MT/year of 27 (range from 5 to 69 MT/operator/year). Patient’s characteristics are reported for the overall study population.
Successful reperfusion and clinical outcome
The overall rate of successful reperfusion was 76.6% (n = 1,175; 95% confidence interval [CI]: 74.5 to 78.7). As shown in Table 1, intravenous thrombolysis, local anesthesia, diabetes, and higher ASPECTS were significantly associated with successful reperfusion in univariate analyses. In a multivariable hierarchical logistic regression model controlling for all patient-level characteristics associated with successful reperfusion (p < 0.20), the effect of the operator on successful reperfusion was significant (p = 0.046), with an ICC of 0.037, meaning that 3.7% of the successful reperfusion variability was due to interoperator variation.
When annual operator annual volume was included in the multivariable hierarchical logistic regression model as a covariate, we found a dose–response relationship between annual operator volume and successful reperfusion (p = 0.003 associated with a 1-level increase) (Table 2). Compared with patients treated by operators with an annual volume <14 MT, for the patient-level characteristic of successful reperfusion, the adjusted odds ratio was 1.75 (95% CI: 1.02 to 2.99) for patients treated by an operator with an annual volume between 14 and 39 MT, and was 2.52 (95% CI: 1.37 to 4.64) for patients treated by an operator with an annual volume ≥40 MT. No association was found between the year of the first MT (≤2010 or >2010), taken as a marker of operator experience, and the rate of successful reperfusion (p = 0.85).
Clinical favorable outcome was achieved in 44.1% (n = 680; 95% CI: 41.7 to 46.6). In a multivariable hierarchical logistic regression model controlling for all patient-level characteristics associated with clinical favorable outcome (p < 0.20), the effect of the operator on successful reperfusion was not significant (p = 0.13), with an ICC of 1.5%. In addition, operator characteristics had no effect on the clinical favorable outcome rate (Table 2).
Two hundred and forty-seven patients (16.4%; 95% CI: 14.5% to 18.2%) presented with at least 1 procedural complication. Embolization in a new territory (n = 102) was the most frequent complication, followed by arterial perforation (n = 67), arterial dissection (n = 63), and spasm (n = 24). As shown in Table 3, a higher rate of procedural complications was associated with no prior history of antiplatelet medication, higher admission NIHSS score, lower ASPECTS, and less often M1 middle cerebral artery occlusion. In a multivariable hierarchical logistic regression model controlling for patient-level characteristics associated with procedure complications (p < 0.20), we found no effect of operator characteristics (operator volume and operator experience [year of first MT/2010]) (Table 4) (ICC = 0). In addition, when annual operator volume was included in the multivariable hierarchical regression model, we found no change in the odds ratio of procedural complications with increasing annual operator volume (p = 0.67) (Table 4) nor with the year of first MT (p = 0.91). Of note, none of the MT complications needed an urgent neurosurgical intervention directly related to the procedure.
Our large multicenter real-world MT cohort, including 1,541 consecutive patients with anterior and posterior AIS documented the correlation between higher operator volume and successful reperfusion rate, with an ICC of 0.037 (p = 0.046), but not on complications. We observed a dose–response relationship between annual operator volume and successful reperfusion (p = 0.003). This finding is in line with previous studies in other interventions. For example, characteristics of operators such as volume of procedures per year and number of years of practice have been found to be an independent clinical and radiological outcome predictor in interventional cardiology for myocardial infarction syndromes in several studies (8,9,13,14). Moreover, several studies, such as the CAPTURE 2 (Carotid ACCULINK/ACCUNET Post Approval Trial to Uncover Rare Events), analyzing the association of the physician variables and outcomes in carotid artery stenting, showed that one of the most important determinants of perioperative carotid artery stenting outcomes was the operator carotid stenting volume (15). However, this annual-volume operator effect on successful reperfusion rates was not found to have an impact on mRS ≤2 at 90 days in our study. We can hypothesize that our study was underpowered to observe an effect on clinical outcome, because the 90-day mRS scores integrate other factors than just the success of MT, such as individual and environmental factors (16).
To date, only 1 study performed on the MR CLEAN (Multicenter Randomized Clinical Trial of Endovascular Treatment for Acute Ischemic Stroke) registry, based on 313 patients, showed no significant relation between the level of experience, neurological adverse events, and poor outcomes in operators with <50 MT procedures (17).
In our study, 247 patients had complications. The most common one was embolization in a new territory (7%). The overall rate of complications was not statistically significant between the operators' experience and volume. There was no significant operator variation for complications. This finding is somewhat controversial in the published reports for other interventions. In the randomized CREST (Carotid Revascularization Endarterectomy Versus Stenting Trial), there was no association between operator volume and rate of complications in carotid revascularization (18). By contrast, experience and high volume had better prognosis on many procedures such as carotid artery stenting, coronary stenting, and percutaneous left atrial appendage closure (9,10,19–22).
Our analysis also demonstrated a better reperfusion rate associated with female sex, patients on antiplatelet therapy, high ASPECT score before MT, and occlusion of the M1 segment of the middle cerebral artery as previously reported (23–27).
Our findings should not be taken as a guideline, but we noted that a volume of at least 40 MT procedures/operator/year, that is, <1/week, was associated with a better reperfusion rate. Currently, The Joint Commission has suggested a minimum of 12 MT during the previous year, or 24 MT over the past 2 years to certificate a thrombectomy-capable stroke center (28). On the basis of the minimum number of 75 coronary interventions/operator/year required to obtain formal certification in coronary interventional procedures, it is important to establish a minimal number of procedures/operator/year to accredit each operator and MT-capable stroke center (22). Our results did not allow us to establish a threshold volume for accreditation, but have added valuable data on this topic.
First, our study was done in 3 large-volume stroke centers. It will be interesting to compare in the future our results with other stroke centers with a lesser volume of MT per year. Second, the majority of our patients enrolled in this study were treated during the first 6 h. Several studies such as the DAWN (Clinical Mismatch in the Triage of Wake Up and Late Presenting Strokes Undergoing Neurointervention With Trevo) and DEFUSE 3 (Endovascular Therapy Following Imaging Evaluation for Ischemic Stroke 3) studies showed the benefit of MT beyond 6 h of last seen well (the last time the patient was seen normal before the discovery of stroke symptoms) (29,30). It will be interesting to compare in the future the operator volume effect on reperfusion and complication rate between 6 and 24 h after AIS onset. Third, in our study, we analyzed only the volume of MT and not all kind of interventions on brain vessels such as embolization of intracranial aneurysm or arteriovenous malformation. This issue could be assessed in further studies, in particular if the volume of all interventions on brain vessels have an added value on better outcomes than just the volume of MT. Last, no formal sample size was calculated for the present study’s objectives. We calculated that with 1,500 patients, we would have a precision <2.55% to estimate the primary outcome rate (successful reperfusion).
Our data suggest that successful reperfusion in AIS patients treated with MT is dependent on the operator volume of MT per year. We did not find any relation between operator volume and complication rates, and the better procedure outcome did not translate into better clinical outcomes. Further studies are warranted to confirm these findings and investigate threshold procedure numbers associated with better outcomes.
WHAT IS KNOWN? The operator’s effect has been found to be an independent predictor for clinical outcome and technical performance in interventional cardiology.
WHAT IS NEW? Our data suggest that operator volume of MT/year has a positive impact on successful reperfusion in AIS patients, but not on clinical outcomes nor on complication rates.
WHAT IS NEXT? Further studies are warranted to investigate threshold procedure numbers associated with better outcomes.
The authors thank Mary Osborne-Pellegrin for her help in editing the final draft of the paper.
The authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- acute ischemic stroke
- Alberta Stroke Program Early Computed Tomography Score
- confidence interval
- computed tomography
- intraclass correlation coefficient
- intracranial hemorrhage
- modified Rankin Scale
- mechanical thrombectomy
- National Institutes of Health Stroke Scale
- Received October 2, 2018.
- Revision received November 26, 2018.
- Accepted December 4, 2018.
- 2019 American College of Cardiology Foundation
- Powers W.J.,
- Derdeyn C.P.,
- Biller J.,
- et al.,
- American Heart Association Stroke Council
- The Joint Commission
- Klein L.W.,
- Schaer G.L.,
- Calvin J.E.,
- et al.
- Kastrati A.,
- Neumann F.-J.,
- Schömig A.
- Reddy V.Y.,
- Holmes D.,
- Doshil S.K.,
- Neuzil P.,
- Kar S.
- Desilles J.P.,
- Consoli A.,
- Redjem H.,
- et al.,
- ETIS (Endovascular Treatment in Ischemic Stroke) Research Investigators
- Fiorelli M.,
- Bastianello S.,
- von Kummer R.,
- et al.
- Hannan E.L.,
- Zhong Y.,
- Jacobs A.K.,
- et al.
- Xu B.,
- Redfors B.,
- Yang Y.,
- et al.
- Gray W.A.,
- Rosenfield K.A.,
- Jaff M.R.,
- Chaturvedi S.,
- Peng L.,
- Verta P.,
- CAPTURE 2 Investigators and Executive Committee
- Beumer D.,
- van Boxtel T.H.,
- Schipperen S.,
- et al.,
- MR CLEAN Pretrial Investigators
- Shishehbor M.H.,
- Venkatachalam S.,
- Gray W.A.,
- et al.
- Carrafiello G.,
- De Lodovici M.L.,
- Piffaretti G.,
- et al.
- Levine G.N.,
- Bates E.R.,
- Blankenship J.C.,
- et al.,
- American College of Cardiology Foundation, American Heart Association Task Force on Practice Guidelines, Society for Cardiovascular Angiography and Interventions
- Nogueira R.G.,
- Liebeskind D.S.,
- Sung G.,
- Duckwiler G.,
- Smith W.S.,
- MERCI, Multi MERCI Writing Committee
- Luo G.,
- Mo D.,
- Tong X.,
- et al.
- Raymond S.,
- Rost N.S.,
- Schaefer P.W.,
- et al.
- Yoon W.,
- Kim S.K.,
- Heo T.W.,
- Baek B.H.,
- Lee Y.Y.,
- Kang H.K.
- Fargen K.M.,
- Fiorella D.J.,
- Mocco J.
- Nogueira R.G.,
- Jadhav A.P.,
- Haussen D.C.,
- et al.,
- DAWN Trial Investigators