Propensity matched analysis of longterm outcomes following transcatheter based aortic valve implantation versus classic aortic valve replacement in patients with previous cardiac surgery

Background The aim of this study was to compare outcome of patients with previous cardiac surgery undergoing transapical aortic valve implantation (Redo-TAVI) to those undergoing classic aortic valve replacement (Redo-AVR) by using propensity analysis. Methods From January 2005 through May 2012, 52 high-risk patients underwent Redo-TAVI using a pericardial xenograft fixed within a stainless steel, balloon-expandable stent (Edwards SAPIEN™). During the same period of time 167 patients underwent classic Redo-AVR. Logistic regression analysis was used to identify covariates among 11 baseline patient variables including the type of initial surgery. Using the significant regression coefficients, each patient’s propensity score was calculated, allowing selectively matched subgroups of 40 patients each. Initial surgery included coronary artery bypass grafting in 30 patients, aortic valve replacement in 7 patients and mitral valve reconstruction in 3 patients in each group. Follow-up was 4 ± 2 years and was 100% complete. Results Postoperative chest tube drainage (163 ± 214 vs. 562 ± 332 ml/24 h, p = 0.02) and incidence of early permanent neurologic deficit (0 vs. 13%, p = 0.04) was lower in patients with Redo-TAVI and there was a trend towards improved 30-day survival (p = 0.06). Also we detected a decreased ventilation time (p = 0.04) and lower transfusion rate of allogenic blood products (p ≤ 0.05) in the Redo-TAVI group. At late follow up differences regarding incidence of major adverse events, including death and permanent neurologic deficits (25% vs. 43%, p = 0.01) statistically supported early postoperative findings. Conclusion The encouraging results regarding early and long-term outcomes following TAVI in patients with previous cardiac surgery show, that this evolving approach may be particularly beneficial in this patient cohort.