- Journal
- ARCHIVES OF CARDIOVASCULAR DISEASES
- Année
- 2026
- Volume
- 119
- Numéro
- 1
- Pages
- 15-24
- Mois
- JAN
- DOI
- 10.1016/j.acvd.2025.12.003
Abstract
Background: Practitioners recommending transcatheter aortic valve implantation (TAVI) currently lack reliable tools to predict periprocedural risk of ischaemic stroke. Aims: We aimed to develop and internally validate a clinical risk score to accurately stratify this risk. Methods: Using data from the nationwide, multicentre FRANCE-TAVI registry, we developed a clinical predictive risk score for 30-day ischaemic stroke post-TAVI using multivariable logistic regression analysis. The model was internally validated through cross-validation techniques. Results: Among 62,747 patients, 1712 (2.7%) experienced ischaemic stroke within 30 days. Nine clinical predictors were identified: female sex, age > 85 years, weight < 60 kg, symptomatic status, history of stroke or transient ischaemic attack, multiple (i.e. > 1) episodes of acute heart failure, severe mobility reduction, diabetes and creatinine clearance < 60 mL/min. The resulting scoring model demonstrated good accuracy (Brier score 0.18), moderate discrimination (C-index 0.63) and excellent calibration as assessed by calibration plots, calibration-in-the-large and calibration slope. The score categorized patients into low - (90.2% of the population), intermediate - (8.0%) and high-risk (1.8%) groups. Observed stroke rates increased progressively across these groups, from 2.25% in the low-risk group to 6.51% in the intermediate-risk group and 10.10% in the high-risk group. Conclusions: This newly developed STRAT score is a clinical, practical and effective tool for predicting early ischaemic stroke in patients undergoing TAVI. It was derived and internally validated in the FRANCE-TAVI registry and may help tailor preventive strategies. Further studies are necessary to externally validate this score and evaluate its impact on clinical decision-making. (c) 2025 The Authors. Published by Elsevier Masson SAS. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).