Published in:
Open Access 01-12-2016 | Research article
Thromboembolic events, bleeding, and drug discontinuation in patients with atrial fibrillation on anticoagulation: a prospective hospital-based registry
Published in: BMC Cardiovascular Disorders | Issue 1/2016
Login to get accessAbstract
Background
The clinical practice of stroke prevention in atrial fibrillation (AF) with direct oral anticoagulants (DOACS) differs from anticoagulation in randomized trial patients. We investigated the risk of thromboembolism, bleeding, and drug discontinuation in a hospital-based real-world setting.
Methods
All-comer patients with non-valvular AF were recruited into a registry at an academic tertiary care center. After informed consent, patients underwent a personal structured interview including medical history, past and current anticoagulation, and returned for follow-up after 6–12 months.
Results
The registry comprised 282 patients (42% women, median age 71 years) with a median CHA2DS2-Vasc-Score of 4 (25. to 75. percentile 2.5–5), who were prospectively followed 285 days in median. At inclusion, 118 patients took vitamin-K-antagonists, 33 dabigatran, 87 rivaroxaban, 30 apixaban, 5 low-molecular-weight heparin, and 9 were on no anticoagulant. Occurrence of stroke (rate 2.8/100 patient-years), was associated with prior stroke (hazard ratio [HR] 18.5, 95% confidence interval 2.16–159), increased HbA1c (HR per 1% increase 1.71, 1.20–2.45) and borderline significantly associated with vascular disease (HR 8.33, 0.97–71.3). Further we observed a high rate of major bleeding (2.8/100 patient-years), clinically relevant non-major bleeding (4.1/100 patient-years), and venous thromboembolism (2.8/100 patient-years). Anticoagulation was discontinued by 80 patients (36.9/100 patient-years), and diabetes (HR 2.31, 1.32–4.02), history of bleeding (HR 2.51, 1.44–4.37) and elevated leucocyte count (HR per 1G/l increase 1.02, 1.00–1.05) were associated with increased risk of discontinuation.
Conclusions
In this hospital-based registry, patients with atrial fibrillation had an increased risk of thromboembolic events despite anticoagulation. The low drug persistence may be attributable to distinct comorbid conditions and bleeding complications.