To the Editor,

We appreciate the interest shown by Rodriguez-Bailón et al. in our articlew published in Revista Española de Cardiología,1 and we would like to make a few points regarding their comments. Two recently published reviews of several different studies2, 3 regarding treatment using the CoreValve percutaneous prosthetic aortic valve and one large study4 have shown that frequency and/or severity of paravalvular leaks tend to decrease. To our understanding, the difference in results with those from the Rodriguez-Bailón et al. study is based on methodology, with different criteria used to “quantify” these leaks. Some controversy exists among cardiac sonographers whether or not to evaluate paravalvular and central regurgitations the same, or if these should be classified into 4 or 3 grades as recommended in the most recent guidelines for prosthesis evaluations. On the other hand, the concept of “reducing” the leak also differs between studies, and some define a significant decrease as a reduction by more than 1 degree, but in other studies, as in ours, a decrease is deemed significant when it is at least 1 degree. Lastly, Rodriguez-Bailón et al. observed a decrease in the frequency of moderate regurgitations and an increase in mild regurgitations during the follow-up period, which was interpreted as an absence of changes due to high concordance (κ =0.724). In our opinion, the kappa coefficient is not the most appropriate instrument to use for evaluating the changes over time of an ordinal quantitative variable, such as the degree of regurgitation. Probably, a nonparametric test, such as the Wilcoxon test for paired samples, would be a better option. In a recent analysis performed using the data from our study (92 cases by January 2011), we obtained similar results after 1 month (significant improvement in leaks, P<.001), and with no changes after 1 year (P=.09), thus eliminating the bias introduced by the reduced number of patients. We insist on the self-expandability of the prosthesis as the probable cause of this decrease, as do other authors, basing our conclusions on the echocardiographic observation of this phenomenon within the first days following the procedure. We have not found more cases of periprosthetic thrombosis than in other series, as we have followed the antithrombotic protocols recommended by the manufacturers.

We have also confirmed an early decrease in ventricular hypertrophy (P<.05) using our most recent data, which has also been described recently by other authors,5 and so we reiterate that differences in methodology could be the cause of the differences observed between studies.

In any case, we share the sentiment expressed by Rodriguez-Bailón et al. for the need for larger studies to further elucidate these “discrepancies,” but perhaps our primary objective should be to solidify the criteria used for study methods and for the definition of variables and objectives.

.

Corresponding author: mamenl@hotmail.com