Change in Atrial Fibrillation Status, Comments to Val-FAAP Registry. Response

To the Editor,

First of all, we would like to thank Vidal-Pérez et al. for their interest in the Val-FAAP study and its results. Part of the Val-FAAP study consisted in analyzing the clinical profile of patients in whom atrial fibrillation (AF) had become permanent compared to those in which it had not.¹ Those who progressed to permanent AF were older and had more comorbidities. Due to the cross-sectional study design, it was not possible to assess the prognostic impact of the transition to permanent AF. The AFBAR study provided evidence that change in AF status increased the likelihood of death or hospitalization almost threefold.² However, so did heart disease and left ventricular dysfunction. In light of these data, the question arises whether the change to permanent AF was an independent predictor of cardiovascular events or whether it was simply more common in patients with underlying heart disease, i.e. patients in whom mortality is already higher.

As regards the Val-FAAP methodology, when we compared the clinical profile of patients who transitioned to permanent AF, we took into account all patients who had evolved to that state, regardless of whether their AF had only started recently, or whether it was paroxysmal or persistent. When assessing change in AF status, we took into account both the type of AF recorded when the patient first presented with a diagnosis of AF (data were collected from medical records), as well as the type of AF at the time of data collection. The discrepancy in the percentage of patients transitioning to permanent AF compared to the AFBAR study is probably due to the longer period between the initial diagnosis of AF and the time of data collection in the Val-FAAP, although we did not quantify the length of time.

When determining the prevalence of a particular disease, great care needs to be taken with methodological aspects. The study, for example, should be carried out in a representative sample of the population. However, the majority of published studies are carried out in clinicians’ offices and cannot therefore be considered population studies. We were surprised that the AFBAR researchers considered their methodology to be more appropriate for estimating the prevalence of the disease. In this type of study, choosing the sample is very important, as it must be representative of the population of interest. Obviously, the larger the sample, the lower the probability of a selection bias. A total of 119 526 patients were included in the Val-FAAP study, and represented the entire Spanish population.¹ With smaller sample sizes, the possibility of bias increases, and even more so if the sample is limited to a specific geographic area as it is then difficult to generalize the results to populations in other areas. Another notable feature of AF is that a significant percentage of cases are asymptomatic or “silent”, and therefore cannot be detected. That is important because of the clinical implications of subclinical AF.³ By only taking into account patients with a known diagnosis of AF, rather than attempting, as in the Val-FAAP study, to detect the arrhythmia in the whole population included, we will undoubtedly underestimate the percentage of patients with AF. For all those reasons, and despite the fact that the data were collected from individuals attending primary care centers, we believe the results of the Val-FAAP study provide a better picture of the larger reality in Spain than data obtained in studies conducted in specific regions; the Val-FAAP data also allow us to better estimate the prevalence of AF in Spain.

Corresponding author: [e-mail]