Introduction Conduction system abnormalities, including AV block, are amongst the most common complications of transcatheter aortic valve implantation (TAVI). Post-TAVI high degree AV block necessitates permanent pacemaker (PPM) implantation.
Purpose To assess the ability of standardly available pre-, intra- and post-TAVI factors to predict PPM implantation within 30-days post procedure.
Methods Demographic and clinical (pre-, intra-, and post-procedural) data including ECG parameters were collected from all patients who underwent TAVI at our centre from August 2017 to November 2020. Patients with pre-existing PPM were excluded from the study. Predictive factors were selected through univariate analysis, and selected characteristics were incorporated into a multivariate binomial logistic regression model, in order to create a 30-day PPM risk-prediction model. The Akaike information criterion (AIC) and area under receiver operating curve (AUC/C-statistic) were used to assess discriminative performance.
Results In total, data from a total of 446 patients were analysed. Of these, 40 (8.97%) received PPM implantation within 30 days of the procedure. The mean age of the patients was 81.5 (±7.3 SD) years; 99 (22.2%) had pre-existing first degree AV block, 55 (12.3%) had pre-existing left bundle branch block (LBBB) and 50 (11.2%) had pre-existing right bundle branch block (RBBB). Intra-procedurally 40 (9.0%) developed LBBB, 21 (4.7%) developed 3rd degree AV block, and 95 (21.3%) patients required temporary pacing wire (TPW) pacing. Post-procedurally, 138 (30.9%) exhibited AV block, 107 (24.0%) LBBB and 50 (11.2%) RBBB. The following factors met significance at multivariate logistic regression analysis: pre-TAVI RBBB (OR 6.62 [95% CI, 1.37-36.51]), intra-TAVI 3rd degree AV block (OR 12.80 [95% CI, 3.44-53.34]), intra-TAVI LBBB (OR 4.02 [95% CI, 1.28-12.53]), use of TPW pacing (OR 8.58 [95% CI, 3.19-25.12]) and post-TAVI LBBB (OR 7.84 [95% CI, 2.75-24.46]) (table 1). Finally, variables were incorporated into a multivariate logistic regression model with the outcome variable of 30-day PPM implantation (figure 1). A model incorporating five factors (pre-TAVI RBBB, intra-TAVI 3rd degree AV block, intra-TAVI LBBB, use of TPW pacing and post-TAVI LBBB) demonstrated excellent discriminative ability (accuracy 0.925 and an AUC of 0.952) at predicting PPM implantation (figure 2).
Conclusions Following variable selection, the best performing model incorporated five factors including pre-TAVI RBBB, intra-TAVI AV block (3rd degree), intra-TAVI LBBB, use of TPW pacing and post-TAVI LBBB. We aim to validate this model using an external cohort.
Conflict of Interest None
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.