Introduction Atrial fibrillation is common in patients undergoing TAVI and embolisation of left atrial appendage (LAA) thrombus is a potential mechanism of periprocedural stroke. The optimal method for diagnosing LAA thrombus is unclear and the prevalence and clinical impact of LAA thrombus in patients undergoing TAVI is unknown.

Methods We examined cardiac CT (CCT) of 124 patients undergoing TAVI at James Cook University Hospital between July 2013 and October 2015, for LAA thrombus. Two blinded, independent CCT reporters assessed each scan. The LAA was graded as definite, possible or no thrombus. We also compared CCT findings with TOE where both modalities were available.

Results Forty-two (33.9%) patients had AF with a mean CHA2DS2-VASc score of 3.7. Definite LAA thrombus was present in 10 (8.1%), probable in 9 (7.3%) and no thrombus in 105 (84.6). Fifty-five patients underwent TOE and LAA thrombus was identified in 5 (9.1%), with a further 5 (9.1%) demonstrating spontaneous echo contrast. TOE confirmed CCT findings in 52 of 55 cases in which both modalities were performed. Six patients had periprocedural CVA, of which two had a history of AF. Four of the cases had evidence of thrombus on either CT or TOE. The remaining two cases were negative on CCT, but did not undergo TOE.

Conclusion The prevalence of definite LAA thrombus in our TAVI cohort is high (8.1%) and the presence of thrombus is associated with peri-procedural CVA. Identification of thrombus on CCT may improve clinical outcomes by promoting therapy such as anti-coagulation or embolic protection devices during TAVI.