Substantial differences in initiation of oral anticoagulant therapy and clinical outcome among non-valvular atrial fibrillation patients treated in inpatient and outpatient settings

Abstract

Aims Patients with atrial fibrillation (AF) are encountered and treated in different healthcare settings, which may affect the quality of care. We investigated the use of oral anticoagulant (OAC) therapy and the risk of thrombo-embolism (TE) and bleeding, according to the healthcare setting. Methods and results Using national Danish registers, we categorized non-valvular AF patients (2002-11) according to the setting of their first-time AF contact: hospitalization (inpatients), ambulatory (outpatients), or emergency department (ED). Event rates and hazard ratios (HRs), calculated using Cox regression analysis, were estimated for outcomes of TE and bleeding. We included 116 051 non-valvular AF patients [mean age 71.9 years (standard deviation 14.1), 51.3% males], of whom 55.2% were inpatients, 41.9% outpatients, and 2.9% ED patients. OAC therapy 180 days after AF diagnosis among patients with a CHADS2 ≥ 2 was 42.1, 63.0, and 32.4%, respectively. Initiation of OAC therapy was only modestly influenced by CHADS2 and HAS-BLED scores, regardless of the healthcare setting. The rate of TE was 4.30 [95% confidence interval (CI) 4.21-4.40] per 100 person-years for inpatients, 2.28 (95% CI 2.22-2.36) for outpatients, and 2.30 (95% CI 2.05-2.59) for ED patients. The adjusted HR of TE, with inpatients as reference, was 0.74 (95% CI 0.71-0.77) for outpatients and 0.89 (95% CI 0.79-1.01) for ED patients. Conclusion In a nationwide cohort of non-valvular AF patients, outpatients were much more likely to receive OAC therapy and had a significantly lower risk of stroke/TE compared with inpatients and ED patients. However, across all settings investigated, OAC therapy was far from optimal.