Abstract 14445: Quality Assessment of Transcatheter Aortic Valve Replacement Meta Analysis: Is It Worth Reading?

Abstract

Introduction: Multiple meta-analyses have been performed since Transcatheter Aortic Valve Replacement (TAVR) and quality of the rapidly publishing data on TAVR is debatable. Methods: Descriptive statistics was performed on pooled samples of quality assessment statistics including heterogeneity I2, Q-statistics ,Tau, and Cochrane. An approach to high heterogeneity was reviewed in this meta-analysis to review the quality of content published online. We reported total n, average numbers and percentages. Continuous data with normal distribution was reported as mean with standard deviation (SD). Results: A total of 231 meta-analyses done on TAVR were included in our study. The total of 1365 outcomes were reported in all of the meta-analyses. The total outcomes with heterogeneity I2< 75% and I2 >75% were 1125 and 244 respectively. The average number of outcomes with I2 <75% was 5, while the outcomes with I2>75% was 1.19. Tau, H statistics, and Cochrane statistics were reported for 79, 1 and 42 outcomes respectively. The subgrouping was performed in 87 outcomes. The mean number of subgroups was 1 ± SD 0.06, Odds ratios, relative risk and standard mean difference were used as an effect size in 93, 63, and 36 studies respectively. The random effect model, fixed effect model, or combine models were used in 180, 67, and 53 outcomes respectively. Sensitivity analysis was performed in 109 outcomes. Meta-regression was performed in 55 outcomes with an average of 0.27 studies. 46 outcomes were able to report I2<75% after meta regression, and subgroup analysis. Trend (n) of outcomes with heterogeneity I2>75% across different studies is shown in figure 1. Conclusion: TAVR heterogeneity was high in up to 18% of meta analysis outcomes in all meta analysis performed. Meta regression was performed in 22% with high heterogeneity, and 18% of the outcomes were able to fix the high heterogeneity in the data.