Abstract
BACKGROUND: The predictive utility of genetic risk factors added to clinical risk factors for stroke recurrence is unclear. Furthermore, whether those predictive utility change in different age groups is uncertain. This study aimed to compare the association of genetic and clinical risk factors with recurrent stroke. METHODS: Data were collected from 5,378 patients with a history of stroke from the UK Biobank. Cox proportional hazards regression models were used to assess the associations of genetic, demographic, lifestyle, and comorbidity risk factors with stroke recurrence, both in overall and three age-based groups. RESULTS: Over a median follow-up of 11.5 (interquartile range [IQR] 10.6–12.3) years, 417 patients had recurrent stroke. The hazard ratio (HR) for recurrent stroke was most pronounced for diabetes mellitus (HR 2.28, 95% confidence interval [CI] 1.83-2.86). Additionally, current smoking, hypertension, age, and polygenic risk score for ischemic stroke (stroke PRS) were also significantly associated with recurrent stroke. Based on concordance index (C-index, 95% CI), the strongest predictors were diabetes mellitus (0.578, 0.556-0.599) followed by age, hypertension and stroke PRS. However, the stroke PRS (0.577, 0.527-0.627) was the strongest predictor in the younger 1/3 patient group (median age 54, IQR 50-57). In addition to demographic, lifestyle, and comorbidity risk factors, genetic risk factor improved C-index of the prediction model from 0.640 (0.592-0.688) to 0.652 (0.602-0.703) in the younger group. CONCLUSIONS: Recurrent stroke is significantly associated with genetic and clinical risk factors. Genetic risk factor enhances the predictive model`s accuracy, particularly in younger patients. [Figure: see text]