Clinical and anatomical predictors of early in-hospital stroke after transcatheter aortic valve implantation

Stroke remains a serious complication after transcatheter aortic valve implantation (TAVI), despite reductions since early experience, and is associated with increased morbidity and mortality. Identifying patients at higher risk may inform procedural planning and prevention strategies.

To describe the clinical and anatomical characteristics of patients who developed early in-hospital stroke after TAVI for severe aortic stenosis (AS), and to identify independent predictors of this complication.

We retrospectively analysed consecutive patients undergoing TAVI at our centre between 2021 and 2024. Demographic, clinical, echocardiographic, CT-derived and procedural variables were collected, including the Society of Thoracic Surgeons (STS) risk score. Clinical and anatomical features of patients with stroke were also described. Paradoxical low-flow low-gradient AS (pLFLG) was defined according to guideline criteria. Univariate analysis identified variables associated with in-hospital stroke; those with p<0.10 were included in multivariate logistic regression to determine independent predictors.

We included 300 patients; mean age 83±5 years; 46.3% male. In-hospital stroke occurred in 8 patients (2.7%). Overall, 86% were hypertensive, 71.3% dyslipidaemic, 34.7% diabetic, 21.7% had prior atrial fibrillation, and 17% had an STS score >8. Compared with those without stroke, affected patients more frequently had STS score >8 (50% vs 16%, p=0.012), paradoxical low-flow low-gradient AS (25% vs 6%, p=0.032), and larger aortic root angles (median 52º vs 48º, p=0.049). All three variables remained independent predictors after multivariate analysis (STS >8 p=0.027; pLFLG p=0.020; aortic root angle p=0.036).

In this cohort, higher operative risk (STS >8), paradoxical low-flow low-gradient AS and larger aortic root angles were independently associated with early in-hospital stroke after TAVI. Patients with stroke showed a higher-risk profile and more complex valve anatomy. These findings may support selective use of cerebral protection devices, although confirmation in larger multicentre studies is warranted.