LAUSR

Transcatheter aortic valve replacement (TAVR) has become the predominant approach to treating severe aortic stenosis patients with 72,991 TAVRs versus 57,626 surgical aortic valve replacement in 2019 in the United States. The transfemoral approach is utilized in 95% of TAVRs which requires large bore access (14−24 French internal/ 5.9−9.2mm external diameter) sheaths. Some permit transient sheath diameter expansion (Edwards eSheath and Terumo Solopath) as the delivery catheter is advanced. Average patient age is approximately 80 years, often with associated frailty and comorbidities. Studies reporting the Valve Academic Research Consortium (VARC‐2) standardized definition of major vascular complication (MVC) report rates of initially 15% now trending down to 4% in the most recent studies. MVCs are associated with a fourfold increase in 30 day mortality after TAVR which averages 2.2% in 2015 through 2018 in theTranscatheter ValveTherapy Registry (TVTR). Vascular closure devices (VCD) utilized include suture‐based preclosure (ProGlide/ProstarXL) and plug‐based closure for up to 22 French access (MANTA). Sedhom et al. perform a systematic review and meta‐analysis of the 10 studies (2 randomized trials (RCTs) and 8 observational studies) comparing VCD outcomes after TAVR. Overall, 3.6% of suture‐based and 3.6% of plugged‐based VCD treated patients experienced MVCs. In the eight observational studies, there were 4.0% (47/1188) MVCs with suture‐based versus 3.0% (24/794) with the plug device, odds ratio (OR) = 0.67 [95% confidence interval (CI): 0.40−1.15] suggesting nonsignificantly lower odds of MVC with the plug‐based device. In contrast, the two RCTs demonstrated 1.4% (5/362) MVCs with suture‐based versus 3.3% (12/360) MVCs with vascular plug, OR = 2.27 [95% CI: 0.81−6.32] with nonsignificantly increased odds of MVCs with the plug‐based device. Pooling the studies, OR = 0.89 [95% CI: 0.52−1.53] suggesting no significant difference in VARC‐2 MVCs. However, a significant interaction or heterogeneity was observed between the results and the study design, raising a methodological cloud over the validity of the pooled results. While contemporary results with both VCD types seem intuitively acceptable, the frequency of TAVR in the United States and consequences of MVCs should perhaps lead us to place greater trust in the RCTs. Assuming the RCT point estimates and underlying binomial distributions for their sample sizes are accurate within the expected range of sampling error, 1000 random draws from these binomial distributions result in the null hypothesis of no difference being rejected 42% of the time favoring suture‐based VCD with mean difference of 2.0% less MVCs [95% CI: −0.2% to 4.2%]. If the same binomial proportion distributions hold with doubling trial sample sizes (additional 722 patients randomized), the estimated 1000 random draw MVC difference is 1.9% [95% CI: 0.3%−3.5%] favoring suture‐based devices, translating into an estimated 1401 [95% CI: 333−2569] fewer patients with MVCs per year in the USA. While extrapolating RCT results to larger counterfactual sample sizes does not provide additional evidence of relative VCD effectiveness, it does highlight the potential utility of larger RCTs to further inform vascular closure after transfemoral TAVR. Perhaps a larger