Abstract Aims To elucidate the frequency and clinical impact of left atrial appendage thrombus (LAAT) in patients set for transcatheter aortic valve implantation (TAVI). Methods and results All patients undergoing TAVI between January 2014 and June 2020 with analysable multislice computed tomography (MSCT) for LAAT were included. Baseline and procedural characteristics were collected, pre-procedural MSCT’s were retrospectively analysed for LAAT presence. The primary endpoint was defined as the cumulative incidence of any cerebrovascular event (stroke or transient ischaemic attack) within the first year after TAVI. A Cox proportional hazards model was used to identify predictors. A total of 1050 cases had analysable MSCT. Median age was 80 [interquartile range (IQR) 74–84], median Society of Thoracic Surgeons’ Predicted Risk Of Mortality (STS-PROM) was 3.4% (IQR 2.3–5.5). Thirty-six percent were on oral anticoagulant therapy for atrial fibrillation (AF). LAAT was present in 48 (4.6%) of cases. Patients with LAAT were at higher operative risk [STS-PROM: 4.9% (2.9–7.1) vs. 3.4% (2.3–5.5), P = 0.01], had worse systolic left ventricular function [EF 52% (35–60) vs. 55% (45–65), P = 0.01] and more permanent pacemakers at baseline (35% vs. 10%, P < 0.01). All patients with LAAT had a history of AF and patients with LAAT were more often on vitamin K antagonist-treatment than patients without LAAT [43/47 (91%) vs. 232/329 (71%), P < 0.01]. LAAT [hazard ratio (HR) 2.94 (1.39–6.22), P < 0.01] and the implantation of more than one valve [HR 4.52 (1.79–11.25), P < 0.01] were independent predictors for cerebrovascular events. Conclusion Patients with MSCT-identified LAAT were at higher risk for cerebrovascular events during the first year after TAVI.