Transfemoral, transapical and transcatheter aortic valve implantation and surgical aortic valve replacement: a meta-analysis of direct and adjusted indirect comparisons of early and mid-term deaths.

Abstract

Clinical outcomes of transfemoral-transcatheter aortic valve implantation (TF-TAVI) versus surgical aortic valve replacement (SAVR) or transapical (TA)-TAVI are limited to a few randomized clinical trials (RCTs). Because previous meta-analyses only included a limited number of adjusted studies or several non-adjusted studies, our goal was to compare and summarize the outcomes of TF-TAVI vs SAVR and TF-TAVI vs TA-TAVI exclusively with the RCT and propensity-matched cohort studies with direct and adjusted indirect comparisons to reach more precise conclusions. We hypothesized that TF-TAVI would offer surgical candidates a better outcome compared with SAVR and TA-TAVI because of its potential for fewer myocardial injuries. A literature search was conducted through PUBMED and EMBASE through June 2016. Only RCTs and propensity-matched cohort studies were included. A direct meta-analysis of TF-TAVI vs SAVR, TA-TAVI vs SAVR and TF-TAVI vs TA-TAVI was conducted. Then, the effect size of an indirect meta-analysis was calculated from the direct meta-analysis. The effect sizes of direct and indirect meta-analyses were then combined. A random-effects model was used to calculate the hazards ratio and the odds ratio with 95% confidence intervals. Early (in-hospital or 30 days) and mid-term (≥1 year) all-cause mortality rates were assessed. Our search resulted in 4 RCTs (n = 2319) and 14 propensity-matched cohort (n = 7217) studies with 9536 patients of whom 3471, 1769 and 4296 received TF, TA and SAVR, respectively. Direct meta-analyses and combined direct and indirect meta-analyses of early and mid-term deaths with TF-TAVI and SAVR were similar. Early deaths with TF-TAVI vs TA-TAVI were comparable in direct meta-analyses (odds ratio 0.64, P = 0.35) and direct and indirect meta-analyses combined (odds ratio 0.73, P = 0.24). Mid-term deaths with TF-TAVI vs TA-TAVI were increased (hazard ratio 0.83, P = 0.07) in a direct meta-analysis and became significant after addition of the indirect meta-analysis (hazard ratio 0.78, 95% confidence interval 0.67-0.92, P = 0.003). In conclusion, TF-TAVI was associated with similar early and mid-term deaths compared with SAVR. The number of early deaths was not significantly different between TF-TAVI and TA-TAVI, whereas there were fewer mid-term deaths with TF-TAVI than with TA-TAVI.