Author + information
- Laurent Fauchier, MD, PhD* ( and )
- Laurent Gorin, MD
- ↵*Service de Cardiologie B., Centre Hospitalier Universitaire Trousseau, 37044 Tours, France
We read with interest the paper by Nieuwlaat et al. (1) on patients with atrial fibrillation (AF) and heart failure (HF) from the observational Euro Heart Survey. We have recently reported a large series of patients (n = 1,269) with both AF and HF and a similar number of deaths (n = 247) during follow-up (2).
The Euro Heart Survey was a declarative multicenter registry, whereas we performed a single-center systematic continuous analysis retrieving information from the computerized codification system filled in for each patient using the International Classification of Diseases-10th Revision of the World Health Organization (3). It is interesting to note that in both studies the characteristics of the patients were very similar in many aspects (type of AF, association or not with left ventricular systolic dysfunction, medication during follow-up), although the way of collecting a large amount of data was somewhat different.
However, and in contrast to the results presented by Nieuwlaat et al. (1), we found that in unselected patients with AF and HF, treatments with beta-blocker alone or with beta-blocker plus digoxin were associated with a significant decrease in the risk of death. We think this may have some interest because very few studies have been published to date addressing the effect of beta-blockers in HF patients having AF. Beyond the older age of the patients and the continuous collection of data in our study, we do not have clear explanation for these different results. We were somewhat surprised by the listing of the multivariable determinants of all-cause mortality in patients with AF and HF in the Euro Heart Survey because it appears that a major bleeding is presented as a strong predictor of death during follow-up. The parameters included in this type of multivariate analysis should only include baseline characteristics of the patients. A severe event during follow-up is obviously associated with a higher risk of mortality. Because a history of major bleeding at baseline is not presented, we are wondering if Nieuwlaat et al. (1) have included a major bleeding during follow-up as a predictor of mortality. If this is the case, this inappropriate way of performing the analysis may have affected the final results on the predictors of all-cause mortality.
- American College of Cardiology Foundation