In a retrospective cohort study of patients hospitalized with embolic stroke of undetermined source in Taiwan, researchers found that 14.7 percent experienced a recurrent ischemic stroke over a median follow-up of 28.4 months, though no baseline clinical or metabolic characteristic reliably identified individuals at elevated risk.
The investigation addressed embolic stroke of undetermined source, a condition recognized to carry a substantial risk of secondary events whose clinical correlates remain poorly characterized among East Asian populations. To explore these relationships, researchers sought to describe recurrence patterns and investigate potential baseline clinical and metabolic factors associated with secondary stroke events in a hypothesis-generating framework.
The research was conducted as a single-center retrospective cohort study at Fu Jen Catholic University Hospital in Taiwan. The study cohort included 116 consecutive patients hospitalized with embolic stroke of undetermined source between January 2021 and December 2024, with clinical follow-up extending through December 31, 2025, to monitor subsequent ischemic occurrences.
To evaluate potential associations with recurrent stroke, the investigators applied univariable Cox proportional hazards regression for their primary evaluation. They also performed sensitivity analyses, which included Fine–Gray competing-risk regression treating death as a competing event, a 12-month logistic analysis, and Bayesian sensitivity analyses utilizing conjugate beta–binomial models with Jeffreys priors. Because of the small number of recorded outcomes, all analyses were treated as descriptive and hypothesis-generating.
Across the cohort during the median follow-up period of 28.4 months, 17 patients experienced a recurrent ischemic stroke, representing an overall recurrence rate of 14.7 percent. The cumulative incidence of recurrence was measured at 10.5 percent at 12 months and 14.0 percent at 24 months. Furthermore, among the recurrent cases with a documented date, half of the secondary events took place within six months of the index stroke.
In the primary analysis, no baseline variable demonstrated a statistically significant association with stroke recurrence. The researchers noted that with a total of 17 recurrence events, odds ratios of approximately 4 to 7 would have been necessary to attain 80 percent statistical power, highlighting the substantial statistical constraints inherent in detecting modest baseline effects within this sample size.
Recurrence was numerically more frequent among patients who presented with minor index events, defined as a National Institutes of Health Stroke Scale score of 3 or less. In this group, 12 of 65 patients (18.5 percent) experienced recurrence, compared to 5 of 51 patients (9.8 percent) presenting with more severe initial strokes. However, this comparison yielded an imprecise estimate, with a hazard ratio of 1.58 and a 95 percent confidence interval ranging from 0.55 to 4.50.
Topographical analysis revealed specific anatomical patterns among patients whose initial infarct was confined to a single arterial circulation. For nine recurrent patients with assessable neuroimaging, six experienced their recurrent stroke within the same arterial circulation that had been affected during the index event.
The authors concluded that while stroke recurrence is not uncommon in this clinical population, baseline features do not reliably identify patients at higher risk. They noted that the observed early clustering of recurrence, the numerically higher rate following minor strokes, and the recurrence topography provide hypothesis-generating clues compatible with occult atrial fibrillation, non-stenotic atherosclerosis, differences in treatment and follow-up, or chance. They called for larger prospective studies utilizing prolonged rhythm monitoring and vessel-wall imaging to clarify these potential mechanisms.
Subject of Research: Recurrence patterns and clinical correlates after embolic stroke of undetermined source
Article Title: Recurrence patterns and clinical correlates after embolic stroke of undetermined source: a hypothesis-generating retrospective cohort study from Taiwan
Article References: Hsu, H.-P., Lo, C.-Y., Kao, Y.-W., Chu, P.-H., Shia, B.-C., Chen, Y.-M., & Lin, C.-Y. (2026). Recurrence patterns and clinical correlates after embolic stroke of undetermined source: a hypothesis-generating retrospective cohort study from Taiwan. BMC Neurology. https://doi.org/10.1186/s12883-026-05516-x
Image Credits: AI Generated
DOI: 10.1186/s12883-026-05516-x
Keywords: embolic stroke of undetermined source, ESUS, ischemic stroke recurrence, stroke recurrence patterns, Fu Jen Catholic University Hospital, Taiwan neurology research, retrospective cohort study, competing-risk regression, Bayesian sensitivity analysis, National Institutes of Health Stroke Scale, NIHSS, minor stroke, arterial circulation, stroke topography, occult atrial fibrillation, non-stenotic atherosclerosis, vessel-wall imaging, prolonged rhythm monitoring
News Source: Cassandra Pierce. (November 18, 2026). Recurrent Ischemic Strokes Cluster Within Six Months in Taiwanese ESUS Study. Scienmag.



