TRANSCATHETER AORTIC VALVE REPLACEMENT IN BRITISH COLUMBIA. IMPLICATIONS OF VALVE SELECTION ON PACEMAKER RATES, HOSPITAL STAY, AND READMISSION

Transcatheter aortic valve replacement (TAVR) has become a routine procedure for aortic stenosis. However, transcatheter heart valves differ in terms of the risk of atrio-ventricular conduction block. We examined the effect of valve type on the requirement for a monitoring, pacemakers, length of hospital stay, and rehospitalisation. All patients undergoing transfemoral TAVR in 4 hospitals in British Columbia (BC) between 2012 and 2016 participated in a mandated Cardiac Services BC registry with linkages to provincial health databases. We evaluated 1109 patients undergoing successful transfemoral TAVR for native aortic valve stenosis with the 5 most commonly utilized valves. Valves implanted were (A) balloon-expandable Sapien XT/Sapien 3 (n=703), (B) self-expandable CoreValve/EvolutR (n=339), and (C) mechanically-expandable Lotus (n=67). Baseline clinical characteristics were similar between groups; mean age 82.6 years with multiple comorbidities (STS predicted risk of mortality 6.1 ± 4.5%). Mortality rates were not statistically different by valve type, either at 30 days (3.0%, 3.0%, and 0.0% respectively, p=0.48), or at 1 year (10.4%, 15.2%, and 8.0%, p=0.11). During the study period the rate of new permanent pacemaker implantation rose slightly, both at 30 days post-TAVR (11.8% in 2012 and 14.9% in 2015) and at 1 year (13.6% and 18.6%) for BC as a whole. However, the need for new pacemakers varied according to valve type; (A) 5.8%, (B) 24.8%, and (C) 37.5% at 30 days (p< .001). At 1 year pacemaker rates continued to rise, but also to diverge with no evidence of a ‘catch-up’; (A) 7.5%, (B) 28.2%, and (C) 45.8% (p< .001). The length of stay fell, from a median (IQR) of 3 (3-6) days in 2012 to 2 (1-4) days in 2016. However the length of stay varied according to valve type; (A) 2, (B) 4, and (C) 4 days (median, p < 0.0001). Similarly, 30-day readmission rates trended according to valve type; (A) 15.1%, (B) 23.8%, and (C) 20.0% (p=0.046). Valve type was associated with considerable variation in length of hospital stay, the need for both early and late pacemaker implantation, and re-hospitalization. These differences will have important implications for resource utilization and patients.