Article Text
Abstract
Background and purpose To investigate the efficacy and safety of mechanical thrombectomy in patients with acute ischemic stroke according to the oral anticoagulation medication taken at the time of stroke onset.
Materials and methods A retrospective multicenter study of prospectively collected data based on data from the registry the Madrid Stroke Network was performed. We included consecutive patients with acute ischemic stroke treated with mechanical thrombectomy and compared the frequency of intracranial hemorrhage and the modified Rankin Scale (mRS) score at 3 months according to anticoagulation status.
Results The study population comprised 502 patients, of whom 389 (77.5%) were not anticoagulated, 104 (20.7%) were taking vitamin K antagonists, and 9 (1.8%) were taking direct oral anticoagulants. Intravenous thrombolysis had been performed in 59.8% and 15.0% of non-anticoagulated and anticoagulated patients, respectively. Rates of intracranial hemorrhage after treatment were similar between non-anticoagulated and anticoagulated patients, as were rates of recanalization. After 3 months of follow-up, the mRS score was ≤2 in 56.3% and 55.7% of non-anticoagulated and anticoagulated patients, respectively (P=NS). Mortality rates were similar in the two groups (13.1%and12.4%, respectively). Among anticoagulated patients, no differences were found for intracranial bleeding, mRS score, or mortality rates between patients taking vitamin K antagonists and those taking direct oral anticoagulants.
Conclusions Mechanical thrombectomy is feasible in anticoagulated patients with acute ischemic stroke. The outcomes and safety profile are similar to those of patients with no prior anticoagulation therapy.
- thrombectomy
Statistics from Altmetric.com
Introduction
Stroke is estimated to be responsible for about 11% of deaths in Europe (14% of women and 9% of men).1 Although age-standardized rates of stroke mortality have decreased worldwide in the last two decades owing to improvements in diagnosis and therapy, the overall global burden of stroke remains huge.2 Oral anticoagulation medication is usually administered to prevent stroke in patients with atrial fibrillation and in patients with prosthetic valves.3–7
Unless contraindicated, intravenous thrombolysis administered within the first 4.5 hours of the onset of symptoms has been considered the treatment of choice in patients with acute stroke.8 9 However, intravenous thrombolysis may not be sufficiently effective in cases of large artery occlusion, and in many patients this approach cannot be applied and further treatment is needed. Thus, intravenous thrombolysis has many important contraindications that limit its use in routine practice, including treatment with vitamin K antagonists (VKAs) and an international normalized ratio (INR) value >1.7.8 9
In the last 2 years, seven major randomized clinical trials (MR CLEAN,10 ESCAPE,11 SWIFT-PRIME,12 EXTEND-IA,13 REVASCAT,14 THRACE,15 and PISTE16) have shown that combining endovascular mechanical thrombectomy with intravenous recombinant tissue plasminogen activator (rtPA) is clinically more beneficial than intravenous rtPA alone in patients with acute ischemic stroke and occlusion of a large intracerebral artery in the anterior circulation. Thus, catheter-based mechanical thrombectomy with a stent retriever within the first 6 hours of stroke onset may be considered in patients with exclusion criteria for intravenous thrombolysis or in whom reperfusion is not achieved with rtPA.17–19 Unfortunately, data on the efficacy and safety of endovascular treatment in anticoagulated patients are very scarce.20–23
Previous data have demonstrated that implementing a collaborative network for endovascular treatment of acute ischemic stroke such as the Madrid Stroke Network is reasonable, provides access to this therapy for more patients, and exhibits good results in terms of recanalization rates and clinical outcomes.24 25
The main objective of this study was to determine the efficacy and safety of endovascular therapy in patients with acute ischemic stroke according to the previous use of oral anticoagulants at the time of stroke onset. For this purpose, the frequency of intracranial hemorrhage and the modified Rankin Scale (mRS) at 3 months according to anticoagulation status were analyzed.
Materials and methods
We performed a retrospective multicenter study of prospectively collected data based on registry data from the Madrid Stroke Network. The study population comprised all patients diagnosed with large vessel occlusion stroke and considered for mechanical thrombectomy. The Madrid Stroke Network is composed of nine hospitals with stroke units (six comprehensive stroke centers and three primary stroke centers) and 17 community hospitals working within a coordinated system to guarantee specialized care for all stroke patients in the region of Madrid, which has a population of around 6.3 million inhabitants. To ensure the provision of mechanical thrombectomy during non-working hours, two nodes (Northeast and Southwest, each with three comprehensive stroke centers) have been established so that one hospital from each node works on-call on a weekly rotating shift. The characteristics and operating protocols of the Madrid Stroke Network have been published elsewhere.24 25 This study was based on the official registry of stroke of the Autonomous Community of Madrid. This is an online, centralized, and anonymized prospective registry performed with the purpose of data analysis. This registry has been authorized by the Health Council of Madrid, within the framework of Spanish Law for the Protection of Personal Data (Organic Law 15/1999 on Patient Data protection and Royal Legislative Decree 1720/2007 of December 21). In this study, we analyzed data for patients treated at the Northeast node from January 2012 to December 2016.
Secondary objectives of the study were to assess the efficacy and safety of endovascular therapy according to the following subgroups: prior use of intravenous thrombolysis (rtPA), and anterior or posterior circulation stroke. Moreover, independent factors that could be associated with the development of hemorrhagic complications were also determined.
The criteria for mechanical thrombectomy according to the consensus protocol of the Madrid Stroke Network were as follows: acute ischemic stroke due to intracranial large vessel occlusion; moderate to severe neurological deficit (National Institute of Health Stroke Scale (NIHSS) score >5 but, in cases of demonstrated acute obstruction with mild symptoms, the decision for thrombectomy was individualized)24 25; time elapsed from stroke onset ≤8 hours for anterior circulation strokes and, in cases of basilar occlusion, ≤12 hours in maximum deficit after onset or ≤24 hours for fluctuating or progressing stroke (If the exact time of symptom onset was not known, it was considered to be the last time the patient was confirmed as being asymptomatic); and failure of intravenous thrombolysis, defined as the complete absence of clinical improvement after 30 min of initiating the intravenous thrombolysis infusion within the therapeutic window for an endovascular procedure, or contraindication for intravenous thrombolysis. Initially, our protocols established a waiting period of 30 min after initiation of intravenous thrombolysis before initiation of the endovascular procedures. After publication of the pivotal randomized clinical trials on mechanical thrombectomy,10–16 protocols for endovascular treatment were initiated as soon as large vessel occlusion was confirmed and immediately after the start of intravenous thrombolysis in patients receiving bridging thrombolysis; the time from symptom onset to groin puncture was set at <6 hours. When the calculated time from symptom onset was >6 hours, endovascular treatment was considered if the patient showed a favorable radiological pattern with signs of little damage on CT perfusion scan or MRI.24 25
The main exclusion criteria included the following clinical conditions or comorbidities that might increase procedure risks: previous dependency for daily living activities; severe established lesions on neuroimaging; and demonstration of a non-favorable mismatch pattern using CT or MRI, according to PWI/DWI (mismatch perfusion-diffusion), among those patients for whom the time from symptom onset was unknown (this criterion did not apply in cases of intravenous thrombolysis). The criteria ‘previous dependency for daily living activities’ and ‘severe established lesions on neuroimaging’ applied to mechanical thrombectomy and/or intravenous thrombolysis and mechanical thrombectomy.24–26
The variables that were prospectively recorded by neurologists comprised age, sex, time from stroke onset, clinical severity at admission according to the NIHSS,27 28 type of oral anticoagulant, INR value at admission, occluded artery, localization of the lesion, early ischemic changes on pretreatment CT according to the Alberta Stroke Program Early CT Score (ASPECTS),29 presence and quantification of mismatch in CT perfusion, and severity of occlusion determined by the Thrombolysis in Cerebral Infarction (TICI) scale.30 Details associated with treatment (time from symptom onset to groin puncture and duration of treatment) were recorded, as was the type of anesthesia.
The safety of endovascular therapy in anticoagulated patients was analyzed by comparing the frequency of intracranial hemorrhage with that recorded in non-anticoagulated patients. Both symptomatic and asymptomatic intracranial hemorrhage were defined according to the SITS criteria31: symptomatic intracerebral hemorrhage was defined as local or remote parenchymal hemorrhage type 2 on the 22–36 hours post-treatment imaging scan combined with a neurological deterioration of ≥4 points on the NIHSS from baseline, or from the lowest NIHSS value between baseline and 24 hours, or leading to death. Asymptomatic hemorrhage included those patients who did not meet all the criteria. A CT scan was systematically performed within 24±4 hours after mechanical thrombectomy or when clinical impairment occurred. Thrombectomy outcomes were measured based on the recanalization rate, considering successful recanalization as TICI 2b or 3.30 Functional outcome was assessed at 3 months by comparing anticoagulated patients with non-anticoagulated patients based on the mRS, a score of ≤2 indicating functional independence.32
Statistical analysis
Quantitative variables were expressed using measures of central tendency and dispersion (mean and SD or median and IQR according to the sample distribution). Qualitative variables were expressed as absolute numbers (n) and relative frequencies (%). The Kolmogorov–Smirnov test was used to assess whether variables were normally distributed. In the bivariate analysis, means were compared using the t-test or Mann–Whitney test based on the sample distribution. Depending on the sample size, the χ2 test or Fisher exact test was used to compare qualitative variables. A Kaplan–Meier survival analysis was performed to compare mortality rates between anticoagulated and non-anticoagulated patients. A regression logistic analysis was performed to assess the predictive factors of symptomatic intracranial hemorrhage after mechanical thrombectomy. The variables included in the model were age, pretreatment NIHSS, INR >1.7, intravenous thrombolysis, and oral anticoagulation. Statistical significance was set at P<0.05. The statistical analysis was performed using SPSS Version 17.0 (SPSS Inc, Chicago, Illinois, USA).
Results
Between January 2012 and December 2016, a total of 502 patients were included, of whom 389 (77.5%) were not anticoagulated, 104 (20.7%) were taking VKAs (mean INR at admission 1.91±2.0), and 9 (1.8%) were taking direct oral anticoagulants (DOACs) (5 (1.0%) dabigatran, 3 (0.6%) rivaroxaban, and 1 (0.2%) apixaban).
Compared with non-anticoagulated patients, anticoagulated patients were older and more frequently had hypertension, diabetes, atrial fibrillation, ischemic heart disease, and a prosthetic heart valve. Smoking, previous use of antiplatelet medication, excessive alcohol consumption, and vertebrobasilar ischemic stroke were more common in non-anticoagulated patients (table 1). Among patients taking VKA, median INR was 1.0 (minimum–maximum 0.6–1.4) in non-anticoagulated patients and 1.7 (0.8–11.0) in anticoagulated patients (P<0.01).
No significant differences were recorded between non-anticoagulated and anticoagulated patients with respect to ischemic changes on the pretreatment CT according to the ASPECTS and the presence and quantification of mismatch in the CT perfusion scan (table 2).
Previous intravenous thrombolysis was performed in 59.8% and 15.0% of non-anticoagulated and anticoagulated patients, respectively (P<0.05). Of the anticoagulated patients, 17 were treated with intravenous thrombolysis and, in the remaining 96 patients, intravenous thrombolysis was contraindicated. Contraindications for intravenous thrombolysis were INR >1.7 (70 patients; 73%), recent surgery (14; 14.6%), thrombocytopenia (1 patient; 1%), coagulopathy (1 patient; 1%), previous intracranial hemorrhage (1 patient; 1%) and in nine (9.3%) cases the cause was unknown. No significant differences were found between non-anticoagulated and anticoagulated patients regarding time from onset of symptoms to recanalization or the type of anesthesia (table 2).
Rates of asymptomatic and symptomatic intracranial bleeding after treatment were similar in both non-anticoagulated and anticoagulated patients. No significant differences were found between anticoagulated patients who took VKAs and those who took DOACs (table 3). Among patients treated with intravenous thrombolysis, rates of asymptomatic and symptomatic intracranial bleeding were similar in both non-anticoagulated and anticoagulated patients (65 (28.6%) vs 4 (23.5%), respectively; P=NS). In the regression logistic analysis, only pretreatment NIHSS was independently associated with the development of symptomatic intracranial hemorrhage (OR 1.1 (for each point in NIHSS), 95% CI 1.03 to 1.16; P=0.04). However, this model had a low predictive accuracy (R2 of Nagelkerke=0.06). Rates of other complications were similar between non-anticoagulated and anticoagulated patients (table 4).
Interestingly, recanalization rates were identical in both anticoagulated and non-anticoagulated patients (TICI score >2a, 82.3% vs 82.3%).
After 3 months of follow-up, the mRS score was ≤2 in 56.3% and 55.7% of non-anticoagulated and anticoagulated patients, respectively (P=NS). No significant differences were found between anticoagulated patients who took VKAs and those who took DOACs (table 3).
Mortality rates at 3 months were similar in both non-anticoagulated and anticoagulated patients (13.1% vs 12.4%, respectively; P=NS) (table 3 and figure 1). No significant differences were found between anticoagulated patients who took VKAs and those who took DOACs.
Discussion
We analyzed the outcomes and risks of endovascular therapy in patients with acute ischemic stroke according to the use of oral anticoagulation at the time of stroke onset. Our study showed that mechanical thrombectomy was feasible in anticoagulated patients with acute ischemic stroke. This is very relevant, since effective anticoagulation is a contraindication for intravenous thrombolysis and only mechanical thrombectomy can be performed in this setting. In this context, endovascular therapy alone may be considered as first-line therapy,33 regardless of anticoagulation use.
Our study included a wide sample of anticoagulated patients (n=113) in whom the main reason for anticoagulation was atrial fibrillation, followed by the presence of a prosthetic heart valve. Anticoagulated patients were older and had more comorbidities than non-anticoagulated patients. These data are in accordance with those published by other authors.23 34 35 Given that the patients in the present study were included consecutively, more than 20% of patients with acute ischemic stroke who are candidates for endovascular therapy may be anticoagulated in clinical practice.
As in the study by Rebello et al 34 no significant differences were found between non-anticoagulated and anticoagulated patients regarding the ASPECTS. In our study, previous intravenous thrombolysis was performed in nearly 60% (n=233) of non-anticoagulated patients and in 15% (n=17) of anticoagulated patients. In their prospective observational study, Benavente et al 23 compared 30 anticoagulated patients and 109 non-anticoagulated patients who underwent primary mechanical thrombectomy. No patient received previous intravenous thrombolysis or treatment with DOACs. By applying propensity score matching, Seiffge et al 35 found that only 33 anticoagulated patients and 13 non-anticoagulated patients underwent intravenous thrombolysis and endovascular treatment, and that 47 patients and one patient had received endovascular treatment only. Of note, these studies included a limited number of patients. Therefore, our data provided relevant information about patients who received endovascular therapy according to their anticoagulation status, regardless of the use of previous intravenous thrombolysis.
As in other studies, we found that time from onset of symptoms to recanalization was similar, regardless of whether anticoagulation therapy had been administered.23 Consequently, anticoagulation status did not delay endovascular therapy. However, in patients in whom intravenous thrombolysis was not performed, primary thrombectomy was not performed earlier. Furthermore, the type of anesthesia was not influenced by the use of oral anticoagulants.
In our study, recanalization rates were similar between anticoagulated and non-anticoagulated patients. However, other studies have reported a trend towards better TICI scores among anticoagulated patients, suggesting that cardioembolic clots may be more easily removed than atherothrombotic clots or that anticoagulation could facilitate persistent recanalization.23 34
Safety is one of the major concerns in endovascular therapy in anticoagulated patients, and hemorrhagic transformation in particular is worthy of attention. Our data showed that rates of asymptomatic and symptomatic intracranial bleeding after treatment were similar, independent of the use of oral anticoagulants at the time of stroke onset. This finding is in accordance with the results of other authors, suggesting that mechanical thrombectomy is safe in both anticoagulated and non-anticoagulated patients, even in those with an elevated INR.22 23 34 35 Compared with warfarin, DOACs are associated with a lower risk of intracranial bleeding in patients with non-valvular atrial fibrillation.7 In patients undergoing intra-arterial treatment, Seiffge et al 35 reported a lower risk of intracranial hemorrhage in those taking DOACs than in those taking VKAs. However, other authors did not find significant differences between the type of oral anticoagulant among patients treated with intravenous thrombolysis.21 In our study, no significant differences were found between VKAs and DOACs, but the number of patients taking DOACs was very low.
After 3 months of follow-up, functional outcome based on the mRS score was similar in both anticoagulated and non-anticoagulated patients. In fact, a mRS score ≤2 was achieved in approximately 55% of patients, regardless of whether or not they had taken oral anticoagulants. However, other studies have shown a trend toward less favorable outcomes in anticoagulated patients.23 33 34 This was particularly evident with VKAs, but not with DOACs. By contrast, other authors have suggested better clinical outcomes with DOACs than with VKAs after endovascular therapy.34 35
In our study, mortality rates at 3 months were similar in both non-anticoagulated and anticoagulated patients. Some authors have suggested higher mortality rates in anticoagulated patients after mechanical thrombectomy while others have reported improved survival rates in this population.23 34 Interestingly, one study showed that mortality rates were higher in patients taking VKAs with an INR ≥1.7 than in patients with an INR <1.7.23 However, other authors did not find an increased risk of unfavorable outcomes in patients with an elevated INR.22 On the other hand, a trend toward lower mortality rates has been reported in patients taking DOACs than in patients taking VKAs.34 As a result, the differences between the studies can be explained by the use of different oral anticoagulants, and, among patients taking VKAs, by the different degrees of anticoagulation control.
Our study is subject to a series of limitations. First, this was an observational study, which may limit comparisons among groups. Second, patients were admitted only to experienced stroke centers. Third, the number of patients taking DOACs was very small. Fourth, the time of last intake was not available, although this limitation may not be important for VKA users when the INR value was accessible. These limitations prevent our results being applied in other populations.
In conclusion, our data suggest that mechanical thrombectomy is feasible in anticoagulated patients with acute ischemic stroke. Outcomes are similar to those of non-anticoagulated patients, and there is no increase in hemorrhagic complications. Further studies are warranted to clarify whether DOACs provide added value in this setting compared with VKAs.
Acknowledgments
Writing and editorial assistance was provided by Content Ed Net (Madrid, Spain) with funding from Bayer Hispania.
References
Footnotes
Contributors All authors contributed extensively to the work presented in this paper. All authors have contributed to the conception and design of the work; the acquisition, analysis, and interpretation of data for the work; drafted the work and revised it critically; approved the final version to be published; and agree to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.
Competing interests None declared.
Ethics approval Madrid Stroke Network.
Provenance and peer review Not commissioned; externally peer reviewed.