Post-TAVI pacemaker requirement prediction - simple math?

Abstract Introduction As transcatheter aortic valve implantation (TAVI) procedures become increasingly common, predicting which patients will require permanent pacemaker implantation (PPI) is a valuable, yet still eluding capability. Purpose To evaluate predictors of post-TAVI PPI and create a risk-defining calculator. Methods Single center study of pts submitted to TAVI without prior history of cardiac device implantation, from 2012 to 2023. Clinical, electrocardiography (ECG), echocardiographic and CT-derived data were collected and analyzed. For statistical analysis T-student, Chi-square tests and logistic regression were performed. Results We included 709 pts, 56,3% of which were female, with a mean age of 82±6,5 years. Regarding baseline ECG, mean QRS was 107±23ms, with 26% of patients displaying complete bundle branch block. Of those, 65,4% were LBBB and 34,6% RBBB. Mean PQ interval was 169±44ms, with 18% of patients displaying 1st degree AV block. 24% of pts presented in AF. Mean aortic valve Agatston score was 3368±1736 Hounsfield units. Roughly half of implanted valves were balloon-expandable (50,4%) and 49,6% self-expandable devices. Regarding valve oversizing index (OI), 6% of pts had undersized valves (OI<0); 28,7% had oversized valves with an OI up to 20% and 65,3% had an OI greater than 20%. The QRS complexes were prolonged by 32±27ms at 48h post-TAVR and PQ increased 15±33ms. Regarding post-TAVR conduction disturbances – 21,7% developed complete AV block; 23,4% new-onset LBBB; 1,7% new-onset RBBB. Overall, 30% of pts required PPI – 27% during index hospitalization and 3% over a mean FUP of 38.8±26 months. On bivariate analysis, baseline QRS duration (p=.002); post-TAVI PQ interval (p=.004) and QRS duration (p=.008); post-TAVI QRS prolongation (p=.03); implanted valve size (p=.01); history of AF (p.003; OR 4,9); baseline RBBB (p<.001; OR 4,9); baseline LBBB (p <.001 OR 2,28); new onset LBBB (p.049; OR 1,45); new onset RBBB (p.005 OR 4,9); and self-expandable valves (p.01; OR 1,53) had significative associations with PPI at FUP. When a logistical regression was conducted, only baseline complete branch block, baseline QRS duration and post-TAVI QRS duration emerged as independent predictors. The prediction model derived from these results performed poorly, explained about 31% of observed variance, and is not adequate for clinical use. Conclusion Several clinical, electrocardiographical, and CT-derived factors present a significative association with post-TAVI PPI. However, in our patient cohort, no model could be derived to accurately predict device implantation at FUP. Individual case assessment and clinical surveillance remain essential in post-TAVI follow-up.