Data-driven machine-learning analysis of potential embolic sources in embolic stroke of undetermined source.

BACKGROUND AND PURPOSE: Hierarchical clustering, a common 'unsupervised' machine-learning algorithm, is advantageous for exploring potential underlying aetiology in particularly heterogeneous diseases. We investigated potential embolic sources in embolic stroke of undetermined source (ESUS) using a data-driven machine-learning method, and explored variation in stroke recurrence between clusters. METHODS: We used a hierarchical k-means clustering algorithm on patients' baseline data, which assigned each individual into a unique clustering group, using a minimum-variance method to calculate the similarity between ESUS patients based on all baseline features. Potential embolic sources were categorised into atrial cardiopathy, atrial fibrillation, arterial disease, left ventricular disease, cardiac valvulopathy, patent foramen ovale (PFO) and cancer. RESULTS: Among 800 consecutive ESUS patients (43.3% women, median age 67 years), the optimal number of clusters was four. Left ventricular disease was most prevalent in cluster 1 (present in all patients) and perfectly associated with cluster 1. PFO was most prevalent in cluster 2 (38.9% of patients) and associated significantly with increased likelihood of cluster 2 [adjusted odds ratio: 2.69, 95% confidence interval (CI): 1.64-4.41]. Arterial disease was most prevalent in cluster 3 (57.7%) and associated with increased likelihood of cluster 3 (adjusted odds ratio: 2.21, 95% CI: 1.43-3.13). Atrial cardiopathy was most prevalent in cluster 4 (100%) and perfectly associated with cluster 4. Cluster 3 was the largest cluster involving 53.7% of patients. Atrial fibrillation was not significantly associated with any cluster. CONCLUSIONS: This data-driven machine-learning analysis identified four clusters of ESUS that were strongly associated with arterial disease, atrial cardiopathy, PFO and left ventricular disease, respectively. More than half of the patients were assigned to the cluster associated with arterial disease.

Ankle-brachial index long-term outcome after first-ever ischaemic stroke.

BACKGROUND AND PURPOSE: Ankle-brachial blood pressure index (ABI) is a clinical tool to identify the presence of peripheral artery disease. There is a scarcity of data associating ABI with long-term outcome in patients with IS. The association between ABI and long-term outcome in patients with first-ever acute IS was assessed. METHODS: Ankle-brachial blood pressure index was assessed in all consecutive patients with a first-ever acute IS admitted at Alexandra University hospital (Athens, Greece) between January 2005 and December 2010. ABI was considered normal when > 0.90 and ≤ 1.30. The Kaplan-Meier product limit method was used to estimate the probability of 5-year composite cardiovascular event-free (defined as recurrent stroke, myocardial infarction or cardiovascular death) and overall survival. A multivariate analysis was performed to assess whether ABI is an independent predictor of 5-year mortality and dependence. RESULTS: Amongst 653 patients, 129 (19.8%) with ABI ≤ 0.9 were identified. Five-year cumulative composite cardiovascular event-free and overall survival rates were better in normal ABI stroke patients (log-rank test: 7.22, P = 0.007 and 23.40, P < 0.001, respectively). There was no difference in 5-year risk of stroke recurrence between low and normal ABI groups (hazard ratio, HR = 1.23, 95% CI 0.68-2.23). In multivariate Cox regression analysis, independent predictors of 5-year mortality included age (HR = 2.55 per 10 years, 95% CI 1.86-3.48, P < 0.001), the National Institutes of Health Stroke Scale (per point increase HR = 1.12, 95% CI 1.08-1.16, P < 0.001), and low ABI (HR = 2.22, 95% CI 1.22-4.03, P = 0.009). Age (HR = 1.21 per 10 years, 95% CI 1.01-1.45, P = 0.04) and low ABI (HR = 1.72, 95% CI 1.11-2.67, P = 0.01) were independent predictors of the composite cardiovascular end-point. CONCLUSIONS: Low ABI in patients with acute IS is associated with increased 5-year cardiovascular event risk and mortality. However, ABI does not appear to predict long-term stroke recurrence.

The type of atrial fibrillation is associated with long-term outcome in patients with acute ischemic stroke.

BACKGROUND/OBJECTIVES: We aimed to investigate the association between the type of atrial fibrillation (AF) and long-term outcome in terms of mortality and stroke recurrence in patients with ischemic stroke and non-valvular AF. METHODS: All consecutive patients admitted with acute ischemic stroke to Alexandra Hospital between 1993 and 2010 were included in the analysis. Patients were categorized in 3 groups according to the type of AF (paroxysmal, persistent, and permanent) and were followed up for up to 10 years after the index stroke or until death. The endpoints were inhospital, 30-day and 10-year stroke recurrence, and 30-day and 10-year all-cause mortality. The Kaplan-Meier product limit method was used to estimate the probability of 10-year stroke recurrence and survival. Multivariate Cox proportional hazard models were used to identify significant predictors of stroke recurrence and all-cause mortality. RESULTS: There were 811 patients (419 females, 392 males) with non-valvular AF and mean age of 75.8 ± 9.4 years. 277 (34.2%) patients had paroxysmal AF, 165 (20.3%) persistent and 369 (45.5%) permanent. Inhospital stroke recurrence rate was low (1.8%) and similar among the 3 patient groups; on the contrary, the probability of 10-year stroke recurrence was significantly higher in patients with permanent AF (p<0.01 by log-rank test). The probability of 10-year survival was significantly higher in patients with paroxysmal AF (p<0.001 by log-rank test). The type of AF was a significant predictor of 10-year stroke recurrence and mortality. Patients with permanent AF had higher risk of stroke recurrence (HR: 1.78, 95%CI: 1.21-2.61) and mortality (HR: 1.55, 95%CI: 1.20-1.99) compared to patients with paroxysmal AF. CONCLUSIONS: Long-term outcome in stroke patients with AF is associated with the type of AF; patients with paroxysmal AF have lower rates of stroke recurrence and mortality.

Long-term outcome in posterior cerebral artery stroke.

INTRODUCTION: Previous studies on posterior cerebral artery (PCA) strokes focused mainly on topography and underlying pathophysiology. However, there are no data on long-term prognosis and its association with the localization of the infarct. METHODS: All consecutive PCA strokes registered in the Athens Stroke Outcome Project between 01/1998 and 12/2009 were included in the analysis. The New England Posterior Circulation Registry criteria were applied to classify them in relation to topography: (i) pure PCA infarcts, including pure cortical-only and combined cortical/deep PCA infarcts (groups A and B respectively), and (ii) PCA-plus strokes, including cortical-only and combined cortical/deep PCA strokes with ≥1 concomitant infarcts outside PCA territory (groups C and D respectively). Patients were prospectively followed up to 10 years after stroke. RESULTS: Amongst 185 (8.1%) PCA patients that were followed up for 49.6±26.7months, 98 (53%), 24 (13%), 36 (19.5%), and 27 (14.6%) were classified in group A, B, C, and D, respectively. Infections and brain edema with mass effect were more frequently encountered in PCA-plus strokes compared to pure PCA (P<0.05 and <0.01 respectively). At 6 months, 56% of cortical-only PCA patients had no or minor disability, compared to 37%, 36%, and 26% in the other groups (P=0.015). The 10-year probability of death was 55.1% (95%CI: 42.2-68.0) for pure PCA compared to 72.5% (95%CI: 58.8-86.2) for PCA-plus (log-rank 14.2, P=0.001). Long-term mortality was associated with initial neurologic severity and underlying stroke mechanism. CONCLUSIONS: Patients with pure PCA stroke have significantly lower risk of disability and long-term mortality compared to PCA strokes with coincident infarction outside the PCA territory.