Development and validation of a prediction model for ischemic stroke recurrence risk in patients with ischemic stroke and atrial fibrillation.

BACKGROUND

The recurrence rate of atrial fibrillation (AF)-related ischemic stroke (IS) remains persistently high, significantly increasing patient mortality, disability, and socioeconomic burden. This underscores an urgent need for a practical tool to predict long-term recurrence risk. This study aimed to investigate the key risk factors for recurrent IS in patients with AF and IS, and to construct and validate a recurrence risk prediction model using LASSO regression.

METHODS

We retrospectively enrolled 113 patients with AF complicated by IS between 2017 and 2024, with a follow-up period of up to 7 years. LASSO regression was employed to screen predictive factors and develop a risk model. The model's performance was evaluated using the concordance index (C-index) and the area under the receiver operating characteristic curve (AUC). A restricted cubic spline analysis was conducted to examine the non-linear relationship between age and the risk of recurrent IS. Subgroup analyses were performed using Cox regression models.

RESULTS

During the 7-year follow-up, recurrent IS occurred in 45.13% of patients. The LASSO regression-based prediction model, incorporating 19 predictive factors, demonstrated high predictive power with an AUC of 0.917. Furthermore, the model's predictive ability improved over time, achieving a C-index of 0.760 at the seventh year of follow-up. A significant non-linear relationship was identified between age and recurrence risk.

CONCLUSION

A LASSO-derived model accurately estimates the long-term risk of recurrent IS after AF-related stroke. Age has a non-linear influence on recurrence, and the efficacy of anticoagulation may be modified by smoking status. These findings support the development of individualized prevention strategies.