LAUSR

Transcatheter aortic valve implantation (TAVI) can lead to paravalvular leak (PVL) in 15% to 20% of cases, which remains an important prognostic factor and an independent predictor of mortality in short and long-term follow-up. To evaluate feasibility and safety of oversizing Medtronic Self-Expandable Valve (Dvalve), calculated according to the aortic annulus maximal diameter (Dmax), on the incidence of PVL and in-hospital mortality after TAVI. We retrospectively analyzed the data of 610 patients treated with TAVI between January 2016 and December 2018. A group of 45 patients of the oversized group (October 2017 to December 2018) accordingly to the Dmax, when (Dvalve − Dmax) | <2 mm in the absence of contraindication was compared to a control group of 213 patients whose prosthesis size had been chosen according to the aortic annulus perimeter (January 2016 to September 2017). In the “oversized” group, no patient had a significant PVL after TAVI compared with the control group (0% vs. 7.51%; p=0.041). Balloon post-dilatation was significantly less frequent in the “oversized” group (0% vs. 10.3%; p=0.012). Per-procedural irradiation and the average length of in-hospital stay were significantly lower (PDS = 2,296.05±1,667.94 cGy·mm2 vs. 4,568±1,352.84 cGy·mm2; p<0.001; and 5.23±1.74 days vs. 6.33±3.23 days; p=0.029, respectively). No case of annulus rupture occurred in the “oversized” group. The incidences of high-degree atrioventricular block with definitive pacing and in-hospital mortality were similar between the two groups. Oversizing the self-expandable valve, according to the aortic annulus maximal diameter, significantly reduced PVL after TAVI, balloon post-dilatation, per-procedural irradiation, and the length of hospital stay, without increasing the risk of mechanical, rhythmic, conductive and coronary occlusion complications. It does not increase the in-hospital mortality rate either. Randomized controlled trials are needed to establish a firm conclusion about its feasibility and safety. Type of funding source: None