LAUSR

Transcatheter implantation of a prosthesis in a degenerated aortic bioprosthesis [valve-in-valve (ViV)] was initially an off-label application of transcatheter aortic valve implantation (TAVI), which rapidly appeared of interest since patients with degenerated bioprostheses are likely to be at higher risk for surgery than those with native aortic stenosis. Transcatheter ViV is now recognized as an option for patients at high risk for surgery in American and European guidelines. Transcatheter aortic ViV accounted for 3.3% of all TAVI procedures between 2012 and 2015 in the Society of Thoracic Surgeons/ American College of Cardiology (ACC) Transcatheter Valve Therapy Registry, and the rate of ViV increased from 1.7% between 2010 and 2012 to 4.4% between 2013 and 2015 in the FRANCE-2 and FRANCE TAVI registries. The number of ViV procedures is likely to increase in the future given the growing use of bioprostheses over the last decades, in particular in younger patients. 5,6 The study by Ribeiro et al. published in this issue of the journal draws attention to coronary obstruction in the particular case of transcatheter aortic ViV procedures. The rationale for studying this rare but severe complication in patients undergoing ViV was justified by a previous study showing higher rates of coronary obstruction after transcatheter ViV than after TAVI for native aortic stenosis. A strength of this study is that it is based on the largest experience in ViV from a multicentre international registry totalling 1612 procedures. The 2.5% rate of coronary obstruction after ViV found in the previous study was based on only three cases of coronary obstruction out of 121 patients. The 2.3% rate reported in the present study is consistent but more robust since it is calculated from 37 cases of coronary obstruction occurring in 1612 patients who underwent ViV procedures. The true rate of coronary obstruction is probably higher due to the possibility of clinically silent cases, in particular in patients with previous coronary artery bypass grafting. In addition to more accurate estimations, the study of a larger number of cases allows for the search of predictive factors. The only clinical factor significantly associated with the risk of coronary obstruction was the presence of a stentless prosthesis or a stented prosthesis with externally mounted leaflets. These findings are consistent with smaller series and can be interpreted by a higher risk of interaction between the bioprosthetic cusps and coronary ostia, most often involving the left coronary artery. Conversely, the 0.7% rate of coronary obstruction reported after ViV in patients who have a stented prosthesis with externally mounted leaflets is not higher than after TAVI in patients with native aortic stenosis. This is of importance since stented prostheses with internally mounted leaflets have been widely used. With regards to the prosthesis implanted during transcatheter ViV, there was no significant difference in coronary obstruction rate between balloon-expandable and selfexpanding prostheses (Figure 1). However, coronary obstruction may occur later with self-expanding than balloon-expandable prostheses. This finding raises questions regarding the mechanism of delayed obstruction and optimal patient monitoring. Another predictive factor of coronary obstruction is related to imaging. The authors describe a new measurement obtained from computed tomography (CT) scan, the virtual transcatheter valve to coronary ostium (VTC) distance, which appears to be strongly associated with the risk of coronary obstruction. The value of VTC distance to predict coronary obstruction should, however, be interpreted with caution. First, analysis of CT scans was performed in a subset of 20 of the 37 patients of the whole population. Secondly, the suggested cut-off value of VTC distance to predict coronary obstruction raises the question of its reproducibility when analysed outside a central core lab. Thirdly, multivariable analysis of CT data in this paper is a case–control study with potential inherent bias.