Assessing the CHA2DS2-VASc score in risk stratification and management of oral anticoagulation during TAVI

Abstract Background International guidelines advise perioperative risk stratification to guide oral anticoagulation management in patients receiving long-term oral anticoagulation who require an elective procedure.1,2 In patients with atrial fibrillation (AF), thromboembolic risk is frequently stratified based on the CHA2DS2VASc score.1,2 The validity of this approach and its interaction with oral anticoagulation management in AF patients undergoing transcatheter aortic valve implantation (TAVI) has not been well studied. Methods Data were obtained from the POPular PAUSE TAVI trial (NCT04437303), a randomized clinical trial comparing interruption versus continuation of oral anticoagulation during TAVI. Items of the CHA2DS2VASc score were registered at baseline. Patients were classified into three pre-defined risk groups: CHA2DS2VASc score ≥ 7 (high risk); 5-6 (moderate risk), and 1-4 (low risk). Outcomes of interest were thromboembolic events (ischemic stroke, transient ischemic attack, myocardial infarction or systemic embolism); cerebrovascular events (ischemic stroke or transient ischemic attack); and major bleeding (type 2-4), defined according to the Valve Academic Research Consortium-3 criteria. Results A total of 818 patients were included: 63 (7.7%) with a high, 318 (38.9%) with a moderate, and 437 (53.3%) with a low CHA2DS2VASc score. The incidence of thromboembolic events was 9.5% in the high risk group, 8.5% in the moderate risk group, and 8.7% in the low risk group. Incidences of cerebrovascular events were 6.3%, 6.3%, and 6.4% across the risk groups, respectively. Major bleeding occurred in 12.7% in the high risk group, in 10.1% in the moderate risk group, and in 9.4% in the low risk group. There was no significant interaction between the randomized strategy and CHA2DS2VASc score for the occurrence of thromboembolic events, either when the CHA2DS2VASc score was modeled as a three-group ordinal variable (p=0.28) or as a continuous variable (p=0.65). Conclusions Among patients with AF undergoing TAVI, the CHA2DS2VASc score did not identify subgroups of patients at higher risk for thromboembolic events. Additionally, no effect modification was observed regarding continuation of oral anticoagulation during TAVI in patients with higher CHA2DS2VASc scores. Considering the outcomes from the main trial, we generally recommend interrupting oral anticoagulation during TAVI and found no evidence to support individualizing anticoagulation management based on the CHA2DS2VASc score.