OBJECTIVES: Stroke incidence after transcatheter aortic valve implantation (TAVI) still represents a concern. This multicentre study aimed at investigating the hypothesis that CHADS2 and CHA2DS2-VASc scores may be used to predict perioperative stroke after TAVI.METHODS: The Italian Transcatheter Balloon-Expandable Valve Implantation Registry (ITER) is a multicentre, prospective registry of patients undergoing balloon-expandable TAVI using Edwards Sapien and Sapien XT prosthesis between 2007 and 2012. The primary end-point of this study was the 30-day stroke rate. Secondary safety end-points were all the major adverse events based on Valve Academic Research Consortium (VARC-2) criteria.RESULTS: One thousand nine hundred and four patients were enrolled in the registry. Mean age was 81.6 ± 6.2 years and 1147 (60.2%) patients were female; mean CHADS2 and CHA2DS2-VASc scores were 2.2 ± 0.8 and 4.4 ± 1.1, respectively. Fifty-four (2.8%) patients had a stroke within 30 days. At multivariable logistic regression analysis, CHA2DS2-VASc (OR: 1.35, 95% CI: 1.03-1.78; P = 0.031) and previous cardiac surgery (OR: 1.96, 95% CI: 1.06-3.6; P = 0.033) but not CHADS2 (OR: 1.05, 95% CI: 0.76-1.44; P = 0.77) were found to be independent predictors of in-hospital stroke. A CHA2DS2-VASc score ≥5 was strongly related to the occurrence of in-hospital stroke (OR: 2.51, 95% CI: 1.38-4.57; P= 0.001). However, CHA2DS2-VASc score showed only poor accuracy for in-hospital stroke with a trend for better accuracy when compared with CHADS2 score (area under the curve: 0.61, 95% CI: 0.59-0.63 vs 0.51; 95% CI: 0.49-0.54, respectively, P = 0.092).CONCLUSIONS: In TAVI patients, CHA2DS2-VASc provided a strong correlation for in-hospital stroke but with low accuracy. Dedicated scores to properly tailor procedures and preventive strategies are needed.