Abstract
Background:
After thrombectomy has shown to be effective in acute stroke patients with large vessel occlusion, the potential benefit of secondary referral for such an intervention needs to be validated.
Aims:
We aimed to compare consecutive stoke patients directly admitted and treated with thrombectomy at a neurointerventional centre with patients secondarily referred for such a procedure from hospitals with a stroke unit.
Methods:
Periprocedure times and mortality in 300 patients primarily treated in eight neurointerventional centres were compared with 343 patients referred from nine other hospitals in a prospective multicentre study of a German neurovascular network. Data on functional outcome at 3 months was available in 430 (76.4%) patients.
Results:
In-hospital mortality (14.8% versus 11.7%, p = 0.26) and 3 months mortality (21.9% versus 24.1%, p = 0.53) were not statistically different in both patient groups despite a significant shorter symptom to groin puncture time in directly admitted patients, which was mainly caused by a longer interfacility transfer time. We found a nonsignificant trend for better functional outcome at 3 months in directly admitted patients (modified Rankin Scale 0–2, 44.0% versus 35.7%, p = 0.08).
Conclusions:
Our results show that a drip-and-ship thrombectomy concept can be effectively organized in a metropolitan stroke network. Every effort should be made to speed up the emergency interfacility transfer to a neurointerventional centre in stroke patients eligible for thrombectomy after initial brain imaging.
Introduction
Mechanical thrombectomy with stent retrievers in acute stroke patients with occlusion of large brain-supplying arteries has recently been shown in randomized controlled trials to be superior compared with intravenous thrombolysis only [Berkhemer et al. 2015; Goyal et al. 2015; Campbell et al. 2015; Saver et al. 2015; Jovin et al. 2015]. To improve access for interventional recanalization therapies, before these results were available, the Neurovascular Net Ruhr was initiated by neurology departments with an acute stroke unit and neuroradiology departments in a metropolitan area with more than 5 million inhabitants in the western part of Germany to improve access for interventional recanalization therapies [Eyding et al. 2012]. In this network, a standardized, prospective assessment was performed over 12 months of consecutive acute ischemic stroke patients with occlusion of a large cerebral artery and treated with mechanical thrombectomy according to standard criteria and an allocation concept including a teleradiology system. We herein report the periprocedural time management and outcomes at discharge and 3 months of patients directly admitted to a neurointerventional centre compared with those referred from external hospitals. Our hypothesis was that patients directly admitted and treated at a neurointerventional centre showed a better outcome (mortality and functional independence) compared with patients secondarily referred to such a centre.
Methods
Between June 2012 and August 2013, all consecutive patients with acute ischemic stroke treated with intravenous thrombolysis and/or thrombectomy from 17 participating hospitals with a stroke unit (8 of them with the possibility to perform thrombectomy 24 hours, 7 days per week by experienced neuroradiologists) of the Neurovascular Net Ruhr were prospectively documented for at least 12 months using a standardized case report form including demographics, vascular risk factors, periprocedural time management, treatment complications, and data on hospital outcome and discharge. Data quality was centrally monitored by consistency checks and queries were sent to the documenting centre. Vital status at 3 months was checked by local investigators. In addition, a follow up after 3 months, including mailed questionnaires, was performed centrally in all patients who, at baseline, had given informed consent. Functional outcome at hospital discharge was assessed using the National Institutes of Health Stroke Scale (NIHSS) and modified Rankin Scale (mRS) by local investigators who were not blinded, while the 3-months outcome was centrally rated by a blinded investigator of the Institute of Epidemiology and Social Medicine, University of Münster. The study was approved by the ethics committees of the Chambers of Physicians of Westphalia and North Rhine.
Statistics
Baseline characteristics, periprocedural times and outcome parameters were compared in directly admitted and referred patients using the Mann–Whitney U-test or chi-square test, as appropriate. A p value <0.05 was considered significant. All analyses were performed with SPSS version 21.
Results
A total of 665 patients were treated with mechanical thrombectomy during the 14-month study period. Information on referral and vital status was available in 643 (96.7%) patients at hospital discharge and in 607 (91.3%) at 3 months. A total of 479 patients (85.1% of those discharged alive) gave consent to be contacted for functional outcome assessment at 3 months and 430 (76.4%) could be reached. Of these, 200 (46.5%) patients were directly admitted to one of the eight neurointerventional centres (in-house group) and 230 (53.5%) patients were referred to one these centres from one of the nine remaining hospitals (referral group). Patients without follow-up information on functional outcome were significantly younger and less often had a history of hyperlipidaemia, while all other baseline demographics including stroke severity and onset to treatment time did not differ significantly.
Between the in-house and the referral group, the following variables did not significantly differ: age, sex, NIHSS at admission, cardiovascular risk factors, and history of cerebrovascular events. Significantly more patients in the referral group presented with carotid-T occlusion (Table 1). Bridging intravenous (IV) thrombolysis was applied in approximately 50% of all patients, with no significant difference between both treatment groups. Stent retrievers were used in 79.8%, the Penumbra aspiration system in 13.6% of all treated patients (table 1), and 98% of all interventions were performed under general anaesthesia.
Baseline characteristics and treatment details.
Multiple occlusion sites possible.
Stroke symptom onset or last time point patient seen well.
p values in bold are significant (p<0.05).
ACA, anterior cerebral artery; CT, computed tomography; IV, intravenous; MR, magnetic resonance; MCA, middle cerebral artery; NIHSS, National Institutes of Health Stroke Scale; PCA, posterior cerebral artery; SD, standard deviation; TICI, Thrombolysis in Cerebral Infarction Score.
There was no significant difference between both treatment groups regarding the median time delay between stroke symptom onset or time point the patient had last seen well and first cerebral imaging (79 minutes in the in-house group versus 90 minutes in the referral group, p = 0.58; Table 1). Patients directly admitted to a neurointerventional centre had significantly shorter time delays from symptom onset to groin puncture (150 minutes versus 233 minutes, p < 0.001), from initial brain imaging to groin puncture (61 minutes versus 139 minutes, p < 0.001), and from symptom onset to end of thrombectomy (245 minutes versus 292 minutes, p < 0.001). In contrast, the time gap between admission at the neurointerventional centre and groin puncture was significantly shorter in the referral group (82 minutes versus 34 minutes, p < 0.001; Table 1).
Symptomatic intracranial haemorrhage (3.0% versus 5.3%, p = 0.39) during hospitalisation did not significantly differ between the in-house and referral group. In-hospital mortality (14.8% versus 11.7%, p = 0.26) and mortality at 3 months (21.9% versus 24.1%, p = 0.53) were also not significantly different (Table 2).
Mortality and functional outcome (subgroup) in-hospital and at 3 months, according to referral status and age group.*
No difference in proportions between in-house and referral group reached statistical significance.
mRS, modified Rankin Scale.
In the subgroup of patients with information on functional status at 3 months, outcome was nonsignificant better (defined as a mRS score 0–2) in patients directly admitted to a neurointerventional centre (44.0% versus 35.7%, p = 0.08, Table 2). Patients ⩾80 years old had a worse outcome compared with younger patients, irrespective of treatment group (Table 2). As expected, patients ⩾80 years old had a significant higher prevalence of atrial fibrillation (64.4% versus 36.8%, p < 0.001) and arterial hypertension (84.5% versus 67.9%, p = 0.001), which might have contributed to the worse outcome.
Discussion
Our prospective multicentre study showed an insignificantly more favourable outcome and lower mortality at 3 months among directly admitted stroke patients compared with patients referred for thrombectomy from an external hospital. Potential explanations for the nonsignificant differences were the considerable shorter time delays between symptom onset and groin puncture or end of thrombectomy in the first group. The most relevant time delay between both treatment groups was observed in the time between first brain imaging and groin puncture (median time intervals of 61 versus 139 minutes). This significantly longer time delay in referred patients mainly resulted from the interfacility transfer and/or waiting for possible neurological recovery from intravenous thrombolysis, and could only be partly compensated by a significantly shorter time delay from admission at the interventional centre to groin puncture in these patients. A median time delay of 139 minutes in secondarily referred patients was also substantially longer compared with recently published randomized trials (ESCAPE, 51 minutes; EXTEND-IA, 93 minutes; SWIFT PRIME, 57 minutes; REVASCAT, 67 minutes), while the time delay in our directly admitted patients was in the range of these trials [Goyal et al. 2015; Campbell et al. 2015; Saver et al. 2015; Jovin et al. 2015].
A median time of 233 minutes from symptom onset to groin puncture in the referral group was similar to that reported in the single centre Bernese stroke registry in Switzerland and shorter compared with the 277 minutes in the multicentre SONIIA stroke registry in Catalonia, Spain [Gratz et al. 2014; Abilleira et al. 2014]. The very short median time of 34 minutes from admission at the neurointerventional centre to groin puncture in referred patients indicates that our allocation concept was efficient in early notification and preparation of the neurointerventional and anaesthesia team for an incoming referred patient. The median time of 82 minutes from admission to groin puncture in patients directly admitted was still shorter than the median time intervals achieved in the EXTEND-IA (113 minutes), SWIFT PRIME (90 minutes) and REVASCAT trial (109 minutes) [Campbell et al. 2015; Saver et al. 2015; Jovin et al. 2015].
The proportion of in-hospital favourable outcome and successful revascularization as prerequisite [Lin et al. 2015] in our study were similar to those reported from the Bernese [mRS 0–2, 39.8%; Thrombolysis in Cerebral Infarction Score (TICI) 2b-3, 70.9%) and Catalan (43.3% and 73.9%, respectively) thrombectomy registries [Gratz et al. 2014; Abilleira et al. 2014]. The observed mortality rate was also in line with these two registries.
Our study confirmed previous findings from observational studies that a favourable outcome in patients over 80 years was achieved only in every fifth patient and that this age group had a high mortality irrespective of the referral status [Abilleira et al. 2014; Kurre et al. 2013; Kleine et al. 2015; Parrilla et al. 2015]. This finding warrants careful selection among octogenarian and nonagenarian acute stroke patients who are eligible for thrombectomy.
The most important limitation of our study is that data on functional outcome were available only in 75% of all discharged patients, which might result in a selection bias. However, patients without a 3 months follow up did not differ significantly in stroke severity on admission and onset-to-groin-puncture-time, and were even younger than those with follow-up information. A strength of our study is the inclusion of consecutive acute stroke patients with no age limit, occlusions of the anterior and posterior brain circulation, and patients with wake-up stroke or unknown stroke onset, resulting in a population which is representative for stroke patients with large vessel occlusions. The inclusion of some patients with wake-up stroke or unknown stroke onset, who were treated with IV thrombolysis and mechanical thrombectomy, resulted in a relatively high upper range limit of the symptom-to-first imaging-time and the symptom-to-treatment-times (Table 1). Treatment in these patients was performed on the discretion of the local neuroradiologist, when no extensive ischemic signs were yet seen on native computed tomography (CT) imaging, or when a diffusion-weighted (DWI) and fluid-attenuated inversion recovery (FLAIR) magnetic resonance imaging (MRI) mismatch was present [Thomalla et al. 2011].So far, differences between referred compared with directly admitted patients treated with mechanical thrombectomy have not been reported. The referred patient group is of particular interest, since the majority of hospitals with stroke units to date do not have the possibility to perform mechanical thrombectomy 24 hours, 7 days per week to date. In his editorial comment to the publication of the SWIFT PRIME and REVASCAT trials, Anthony J. Furlan highlighted the importance of time and pointed out, that an ‘emergency department door-to-groin puncture time of 90 minutes was achieved in the SWIFT PRIME trial. This goal requires stroke workflow efficiencies not yet in place in many hospitals’ [Furlan, 2015]. Our study, with a median time from admission to groin puncture of 82 minutes in patients primarily admitted and only 34 minutes in patients secondarily transferred, showed that such workflow structures, including standardized patient selection, a teleradiology network and an internet based allocation system, can effectively be implemented in a metropolitan neurovascular network in daily practice.
Conclusion
Patients directly receiving thrombectomy at a neurointerventional centre had an insignificantly more favourable outcome compared with secondarily referred patients. A drip-and-ship concept in stroke patients eligible for mechanical thrombectomy can be effectively organized in a metropolitan stroke network similar to established thrombolytic treatment protocols [Tekle et al. 2012]. Every effort should be made to speed up the emergency interfacility transfer after initial brain imaging shows a large brain vessel occlusion instead of waiting for possible neurological recovery from IV thrombolysis.
Further members of the Neurovascular Network Ruhr, Germany, who contributed data for this analysis
Department of Radiology and Neuroradiology (Lothar Heuser), Knappschafts-krankenhaus Bochum; Department of Radiology and Neuroradiology (Peter Mönnigs), St Josef-Hospital Bochum; Department of Neurology (Bernhard Wallner), Knappschaftskrankenhaus Bottrop; Department of Neurology (Clemens Dobis), Knappschaftskrankenhaus Dortmund; Department of Neurology (Katja Hinrichs), Städtisches Klinikum Dortmund; Department of Neurology (Martina Nolden-Koch), Klinikum Duisburg; Departments of Radiology and Neuroradiology (Rene Chapot, Markus Heddier), Alfried Krupp Krankenhaus Essen; Department of Radiology and Neuroradiology (Marc Schlamann), University Hospital of Duisburg-Essen; Department of Neurology (Frauke Busch), St Josef-Krankenhaus Essen; Department of Neurology (Frank Tillmann), Evangelische Kliniken Gelsenkirchen; Department of Neurology (Hans Ruf), St Johannes-Hospital Hagen; Department of Neurology (Bettina Klein), St Marien-Hospital Hamm; Department of Neurology (Tobias Höhle), Evangelisches Krankenhaus Herne, Departments of Neurology (Iris Adelt) and Radiology and Neuroradiology (Mathias König), St-Marien-Hospital Lünen; Department of Neurology (Andrea Hofmann), St Josef-Krankenhaus Moers; Institute of Epidemiology and Social Medicine (Marianne Kalic), University of Münster; Department of Neurology (Barbara Fauser), St Josef-Hospital Oberhausen; Department of Neurology (Philipp Erdmann), Knappschaftskrankenhaus Recklinghausen, all in Germany.
Footnotes
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
Conflict of interest statement
The author(s) declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: R.W. received speaker honoria from Covidien and from serving on a scientific advisory board of Covidien. C.W. and H.N. received speaker honoria from Covidien. W.W. received speaker honoria from Penumbra. The other authors declare no conflicts of interest in preparing this article.