Sustained atrial fibrillation increases the risk of anticoagulation-related bleeding in heart failure

Juergen H. Prochaska*, Sebastian Goebel, Markus Nagler, Torben Knoepfler, Lisa Eggebrecht, Heidrun Lamparter, Marina Panova-Noeva, Karsten Keller, Meike Coldewey, Christoph Bickel, Michael Lauterbach, Roland Hardt, Christine Espinola-Klein, Hugo ten Cate, Thomas Rostock, Thomas Muenzel, Philipp S. Wild*

*Corresponding author for this work

Research output: Contribution to journalArticleAcademicpeer-review


BackgroundOral anticoagulation therapy in individuals with atrial fibrillation (AF) reduces the risk of thromboembolic events at cost of an increased bleeding risk. Whether anticoagulation-related outcomes differ between patients with paroxysmal and sustained AF receiving anticoagulation is controversially discussed.MethodsIn the present analysis of the prospective multi-center cohort study thrombEVAL, the incidence of anticoagulation-related adverse events was analyzed according to the AF phenotype. Information on outcome was centrally recorded over 3years, validated via medical records and adjudicated by an independent review panel. Study monitoring was provided by an independent institution.ResultsOverall, the sample comprised 1089 AF individuals, of whom n=398 had paroxysmal AF and n=691 experienced sustained AF. In Cox regression analysis with adjustment for potential confounders, sustained AF indicated an independently elevated risk of clinically relevant bleeding compared to paroxysmal AF [hazard ratio (HR) 1.40 (1.02; 1.93); P=0.038]. For clinically relevant bleeding, a significant interaction of the pattern of AF type with concomitant heart failure (HF) was detected: HRHF 2.45 (1.51, 3.98) vs. HRno HF 0.85 (0.55, 1.34); P-interaction=0.003. In HF patients, sustained AF indicated also an elevated risk of major bleeding [HR 2.25 (1.26, 4.20); P=0.006]. A simplified HAS-BLED score incorporating only information on age (>65years), bleeding history, and HF with sustained AF demonstrated better discriminative performance for clinically relevant bleeding than the original version: AUC(HAS-BLED): 0.583 vs. AUC(simplifiedHAS-BLED): 0.642 (P=0.004).ConclusionsIn HF patients receiving oral anticoagulation, sustained AF indicates a substantially elevated risk of bleeding.Clinical Trial Registration, identifier: NCT01809015.

Original languageEnglish
Pages (from-to)1170-1179
Number of pages10
JournalClinical research in cardiology
Issue number12
Publication statusPublished - Dec 2018


  • Atrial fibrillation
  • Anticoagulation
  • Bleeding
  • Heart failure

Cite this