Clinical and Neuroimaging Characteristics in Embolic Stroke of Undetermined versus Cardioembolic Origin: A Population‐Based Study

BACKGROUND AND PURPOSE: Evidence suggests that cardioembolism represents the underlying mechanism in the minority of embolic strokes of undetermined source (ESUS). In this population-based study, we sought to compare the clinical and imaging characteristics as well as outcomes in patients with ESUS and cardioembolic stroke (CE). METHODS: We included consecutive patients with first-ever ischemic stroke (IS) from the previously published population-based Evros-Stroke-Registry identified as ESUS or CE according to standardized criteria. Baseline characteristics, admission NIHSS scores, cerebral edema, hemorrhagic transformation, stroke recurrence, functional outcomes (determined by modified Rankin Scale [mRS] scores), and mortality rates were recorded during the 1-year follow-up period. RESULTS: We identified 21 ESUS (3.7% of IS) and 211 CE (37.1% of IS) cases. Patients with ESUS were younger (median age: 68 years [interquartile range [IQR]: 61-75] vs 80 years [IQR: 75-84]; P < .001), had lower median admission NIHSS scores (4 points [IQR: 2-8] vs 10 points [IQR: 5-17]; P < .001), and lower prevalence of cerebral edema on neuroimaging studies (0 vs. 33.3%, P = .002). Functional outcomes were more favorable in ESUS at 28 (median mRS score: 2 [IQR: 1-3] vs 4 [IQR: 4-5]; P < .001), 90 (median mRS score: 1 [IQR: 0-2] vs 4 [IQR: 3-5]; P < .001), and 365 days (median mRS score: 1 [IQR: 0-2] vs 4 [IQR: 2-4]; P < 0.001). At 1-year, the mortality rate was lower in ESUS (0% [95% confidence interval [CI]: 0-13.5%] vs 34.6% [95% CI: 28.2-41.0%]; P < .001); the 1-year recurrent rate was also lower numerically (0% [95% CI: 0-13.5%] vs 9.5% [95% CI: 5.5-13.4%]; P = .140) but this difference failed to reach statistical significance due to the small study population. CONCLUSIONS: The clinical and neuroimaging profiles as well as clinical outcomes vary substantially between ESUS and CE indicating different underlying mechanisms.