Predictors for permanent pacemaker requirement after transcatheter aortic valve implantation with the CoreValve bioprosthesis

Abstract

Introduction: Changes in atrioventricular (AV) conduction and need for permanent pacemaker (PPM) are a recognized complication after open aortic valve replacement. We hypothesized that the need for PPM after CoreValve (Corevalve Inc, Irvine, CA) can be predicted with a combination of baseline variables.

Methods: In patients undergoing transcatheter aortic valve implantation, potential clinical, electrocardiographic and echocardiographic predictors of permanent pacing requirement were studied.

Results: Between January 2007 and March 2008, 34 patients with severe symptomatic aortic stenosis were recruited in a single center. Mean age was 84.4 years (SD 5.4, range 71-93). Of 34 cases paced at baseline, 3 (8.8%) were excluded from this analysis, as was the single periprocedural mortality. Of the remaining 30, 10 underwent permanent pacemaker implantation during the same admission (33.3%). PPM was for prolonged high-grade AV block in 4 cases, episodic high-grade AV block in 5, and sinus node disease in 1. Need for pacemaker was correlated to left axis deviation at baseline (P = .004, r = 0.508) and left bundle-branch block with left axis deviation (P = .002, r = 0.548). It was related to diastolic interventricular septal dimension on transthoracic echocardiography >17 mm (P = .045, r = 0.39) and the baseline thickness of the native noncoronary cusp (P = .002, r = 0.655). A susceptibility model was generated, and if at least one of (1) left bundle-branch block with left axis deviation, (2) interventricular septal dimension >17 mm, or (3) noncoronary cusp thickness >8 mm was present, the likelihood of PPM could be predicted with 75% sensitivity and 100% specificity and a receiver operating characteristic curve area of 0.93 +/- 0.055 (P < .001).

Conclusions: After transcatheter aortic valve implantation with CoreValve, permanent pacing was performed in around a third of patients and we present preliminary concepts towards a predictive model for this phenomenon.