Aortic calcification and the risk for paravalvular regurgitation after TAVI: The importance of focusing on reliable outcomes and appropriate variables

We read with great interest the article of Hansson et al.1 about calcium load and distribution as an independent predictor of paravalvular regurgitation (PAR) after TAVR. Our group has been working on this topic for numerous months (manuscript under review). We, as Hansson et al., are firmly convinced that measurement of aortic calcification before TAVI is of utmost importance, but we would like to raise some potential flaws of the methodology used in their study so as to encourage a fruitful debate on this issue. When we looked at calcium distribution in our series, we found it very challenging with a 0.6 mm collimation width to exactly define the annulus (basal plane), which is the reference plane for device implantation, but even more so to border the “upper LVOT” zone (defined as the region from the basal plane reaching 2 mm into the LVOT). When we applied this approach, we found an unacceptably high inter observer variability in our results. As the authors state, this analysis was carried out by only one observer, which we think could be prone to bias. Assessment of PAR after TAVR is challenging, especially when it is assessed by only transthoracic echo (TTE), since it does underestimate the severity2. We would have preferred to avoid the composite outcome chosen by the authors (significant PAR [sPAR] as defined by the need for postdilatation and/or >mild PAR at TTE on discharge). Unfortunately, the authors do not provide an algorithm for performing postdilatation, but in fact, as stated in the study limitations section, postdilatation was left at the discretion of the treating operator. This again is prone to bias. In many centers, follow-up imaging with transesophageal echocardiography (TEE) is not standard practice, whereas TEE is routinely performed intraoperatively in patients undergoing TAVI.3–5 It would potentially have been more appropriate to assess PAR with intraoperative TEE at the end of the procedure (in particular in a balloon expandable device as Sapien XT, i.e. following postdilatation) and to enter either postdilatation or the implanting center into the multivariable analysis. This would have provided more reliable results, with outcomes being less affected by site and operator differences. Moreover, the reason for investigating the impact of aortic calcification in relation to the composite endpoint of sPAR remains unclear, given that survival rates at follow-up were evaluated only in association with the occurrence of PAR, thus focusing on a different outcome from the one evaluated by multivariable logistic regression analysis. Anyhow, we congratulate with Hansson and colleagues for their valuable manuscript and we hope this letter may help focus attention and encourage a fruitful debate on preoperative assessment of aortic valve calcifications.