General anaesthesia (GA) during endovascular thrombectomy has been associated with worse patient outcomes in observational studies compared with patients treated without GA. We assessed functional outcome in ischaemic stroke patients with large vessel anterior circulation occlusion undergoing endovascular thrombectomy under GA, versus thrombectomy not under GA (with or without sedation) versus standard care (ie, no thrombectomy), stratified by the use of GA versus standard care. For this meta-analysis, patient-level data were pooled from all patients included in randomised trials in PuMed published between Jan 1, 2010, and May 31, 2017, that compared endovascular thrombectomy predominantly done with stent retrievers with standard care in anterior circulation ischaemic stroke patients (HERMES Collaboration). The primary outcome was functional outcome assessed by ordinal analysis of the modified Rankin scale (mRS) at 90 days in the GA and non-GA subgroups of patients treated with endovascular therapy versus those patients treated with standard care, adjusted for baseline prognostic variables. To account for between-trial variance we used mixed-effects modelling with a random effect for trials incorporated in all models. Bias was assessed using the Cochrane method. The meta-analysis was prospectively designed, but not registered. Seven trials were identified by our search; of 1764 patients included in these trials, 871 were allocated to endovascular thrombectomy and 893 were assigned standard care. After exclusion of 74 patients (72 did not undergo the procedure and two had missing data on anaesthetic strategy), 236 (30%) of 797 patients who had endovascular procedures were treated under GA. At baseline, patients receiving GA were younger and had a shorter delay between stroke onset and randomisation but they had similar pre-treatment clinical severity compared with patients who did not have GA. Endovascular thrombectomy improved functional outcome at 3 months both in patients who had GA (adjusted common odds ratio (cOR) 1·52, 95% CI 1·09-2·11, p=0·014) and in those who did not have GA (adjusted cOR 2·33, 95% CI 1·75-3·10, p<0·0001) versus standard care. However, outcomes were significantly better for patients who did not receive GA versus those who received GA (covariate-adjusted cOR 1·53, 95% CI 1·14-2·04, p=0·0044). The risk of bias and variability between studies was assessed to be low. Worse outcomes after endovascular thrombectomy were associated with GA, after adjustment for baseline prognostic variables. These data support avoidance of GA whenever possible. The procedure did, however, remain effective versus standard care in patients treated under GA, indicating that treatment should not be withheld in those who require anaesthesia for medical reasons. Medtronic.

Despite the great socioeconomic burden of stroke, there have been few reports of stroke statistics in Korea. In this scenario, the Epidemiologic Research Council of the Korean Stroke Society launched the "Stroke Statistics in Korea" project, aimed at writing a contemporary, comprehensive, and representative report on stroke epidemiology in Korea. This report contains general statistics of stroke, prevalence of behavioral and vascular risk factors, stroke characteristics, pre-hospital system of care, hospital management, quality of stroke care, and outcomes. In this report, we analyzed the most up-to-date and nationally representative databases, rather than performing a systematic review of existing evidence. In summary, one in 40 adults are patients with stroke and 232 subjects per 100,000 experience a stroke event every year. Among the 100 patients with stroke in 2014, 76 had ischemic stroke, 15 had intracerebral hemorrhage, and nine had subarachnoid hemorrhage. Stroke mortality is gradually declining, but it remains as high as 30 deaths per 100,000 individuals, with regional disparities. As for stroke risk factors, the prevalence of smoking is decreasing in men but not in women, and the prevalence of alcohol drinking is increasing in women but not in men. Population-attributable risk factors vary with age. Smoking plays a role in young-aged individuals, hypertension and diabetes in middle-aged individuals, and atrial fibrillation in the elderly. About four out of 10 hospitalized patients with stroke are visiting an emergency room within 3 hours of symptom onset, and only half use an ambulance. Regarding acute management, the proportion of patients with ischemic stroke receiving intravenous thrombolysis and endovascular treatment was 10.7% and 3.6%, respectively. Decompressive surgery was performed in 1.4% of patients with ischemic stroke and in 28.1% of those with intracerebral hemorrhage. The cumulative incidence of bleeding and fracture at 1 year after stroke was 8.9% and 4.7%, respectively. The direct costs of stroke were about ₩1.68 trillion (KRW), of which ₩1.11 trillion were for ischemic stroke and ₩540 billion for hemorrhagic stroke. The great burden of stroke in Korea can be reduced through more concentrated efforts to control major attributable risk factors for age and sex, reorganize emergency medical service systems to give patients with stroke more opportunities for reperfusion therapy, disseminate stroke unit care, and reduce regional disparities. We hope that this report can contribute to achieving these tasks.

Background and objectives: Post-stroke cognitive impairment (PSCI) occurs in up to 50% of patients with acute ischemic stroke (AIS). Thus, the prediction of cognitive outcomes in AIS may be useful for treatment decisions. This PSCI cohort study aimed to determine the applicability of a machine learning approach for predicting PSCI after stroke. Methods: This retrospective study used a prospective PSCI cohort of patients with AIS. Demographic features, clinical characteristics, and brain imaging variables previously known to be associated with PSCI were included in the analysis. The primary outcome was PSCI at 3–6 months, defined as an adjusted z-score of less than − 2.0 standard deviation in at least one of the four cognitive domains (memory, executive/frontal, visuospatial, and language), using the Korean version of the Vascular Cognitive Impairment Harmonization Standards-Neuropsychological Protocol (VCIHS-NP). We developed four machine learning models (logistic regression, support vector machine, extreme gradient boost, and artificial neural network) and compared their accuracies for outcome variables. Results: A total of 951 patients (mean age 65.7 ± 11.9; male 61.5%) with AIS were included in this study. The area under the curve for the extreme gradient boost and the artificial neural network was the highest (0.7919 and 0.7365, respectively) among the four models for predicting PSCI according to the VCIHS-NP definition. The most important features for predicting PSCI include the presence of cortical infarcts, mesial temporal lobe atrophy, initial stroke severity, stroke history, and strategic lesion infarcts. Conclusion: Our findings indicate that machine-learning algorithms, particularly the extreme gradient boost and the artificial neural network models, can best predict cognitive outcomes after ischemic stroke.

This project assessed performance of natural language processing (NLP) and machine learning (ML) algorithms for classification of brain MRI radiology reports into acute ischemic stroke (AIS) and non-AIS phenotypes. All brain MRI reports from a single academic institution over a two year period were randomly divided into 2 groups for ML: training (70%) and testing (30%). Using “quanteda” NLP package, all text data were parsed into tokens to create the data frequency matrix. Ten-fold cross-validation was applied for bias correction of the training set. Labeling for AIS was performed manually, identifying clinical notes. We applied binary logistic regression, naïve Bayesian classification, single decision tree, and support vector machine for the binary classifiers, and we assessed performance of the algorithms by F1-measure. We also assessed how n-grams or term frequency-inverse document frequency weighting affected the performance of the algorithms. Of all 3,204 brain MRI documents, 432 (14.3%) were labeled as AIS. AIS documents were longer in character length than those of non-AIS (median [interquartile range]; 551 [377–681] vs. 309 [164–396]). Of all ML algorithms, single decision tree had the highest F1-measure (93.2) and accuracy (98.0%). Adding bigrams to the ML model improved F1-mesaure of naïve Bayesian classification, but not in others, and term frequency-inverse document frequency weighting to data frequency matrix did not show any additional performance improvements. Supervised ML based NLP algorithms are useful for automatic classification of brain MRI reports for identification of AIS patients. Single decision tree was the best classifier to identify brain MRI reports with AIS.