Development of a prediction model for clinical valve thrombosis after transcatheter aortic valve implantation

Aims/ Background
Transcatheter aortic valve implantation (TAVI) has become a well-established treatment option for patients with symptomatic severe aortic stenosis. The reported incidence of clinically significant valve thrombosis after TAVI ranges between 1% and 4%. Valve thrombosis may lead to premature valve deterioration. This is especially relevant for younger low-risk patients. Therefore, we aimed to identify predictors of valve thrombosis in a large cohort and develop an easy-to-use score for risk stratification.

Methods
We initially performed a systematic literature analysis in PubMed identifying possible risk factors for clinical valve thrombosis. After expert-guided pre-selection, a total of 80 variables were considered for further testing in a development cohort. The development cohort consisted of 5268 patients undergoing TAVI between January 2018 and June 2023 in a single high-volume center. The prevalence of clinically significant valve thrombosis in this cohort was 4.2% (n=221). Definition of clinically significant valve thrombosis was adapted from the VARC-3 consensus statement. Leaflet thrombosis was assessed by multidetector computed tomography. A predictive score utilizing a two-step approach with univariate logistic and subsequently multivariate logistic regression as well as receiver operating characteristics (ROC) analysis was established.

Results
As a result of the two-step regression methodology, we produced a four-variable model, calculating the pretest probability of clinical valve thrombosis with an area under the ROC curve of 0.747. The final variables of the score included: aortic valve-in-valve implantation (odds ratio [OR]=5.6; 95% confidence interval [CI] 3.9-8.2; p<0.001), intra-annular transcatheter heart valve (THV) design (OR=3.9; 95%CI 2.8-5.4; p<0.001), THV size ≤23mm (OR=1.9; 95%CI 1.4-2.6; p<0.001) and absence of oral anticoagulation prior to the index procedure (OR=1.7; 95%CI 1.2-2.2; p<0,001). The score ranged from 0 to 7 points, with an increased probability for the development of clinically significant valve thrombosis in patients with ≥ 3 points (74% of clinical valve thrombosis cases).

Conclusions
The here presented score might aid clinicians to predict the development of clinical leaflet thrombosis in patients undergoing TAVI and potentially guide the stratification of the subsequent antithrombotic regimen. Further validation of the score is to be tested in an external cohort.