A clinical decision instrument to predict 30-day death and cardiovascular hospitalizations after an emergency department visit for atrial fibrillation: The Atrial Fibrillation in the Emergency Room, Part 2 (AFTER2) study

Background: In previous work, we derived and validated a tool that predicts 30-day mortality in emergency department atrial fibrillation (AF) patients. The objective of this study was to derive and validate a tool that predicts a composite of 30-day mortality and return cardiovascular hospitalizations. Methods: This retrospective cohort study at 24 emergency departments in Ontario, Canada, included patients with a primary diagnosis of AF who were seen between April 2008 and March 2009. We assessed a composite outcome of 30-day mortality and subsequent hospitalizations for a cardiovascular reason, including stroke. Results: Of 3,510 patients, 2,343 were randomly selected for the derivation cohort, leaving 1,167 in the validation cohort. The composite outcome occurred in 227 (9.7%) and 125 (10.7%) patients in the derivation and validation cohorts, respectively. Eleven variables were independently associated with the outcome: older age, not taking anticoagulation, HAS-BLED score of >= 3, 3 laboratory results (positive troponin, supratherapeutic international normalized ratio, and elevated creatinine), emergency department administration of furosemide, and 4 patient comorbidities (heart failure, chronic obstructive lung disease, cancer, dementia). In the validation cohort, the observed 30-day outcomes in the 5 risk strata that were defined using the derivation cohort were 2.0%, 6.6%, 10.7%, 12.5%, and 20.0%. The c statistic was 0.73 and 0.69 in the derivation and validation cohort, respectively. Conclusions: Using a population-based sample, we derived and validated a tool that predicts the risk of early death and rehospitalization for a cardiovascular reason in emergency department AF patients. The tool can offer information to managing physicians about the risk of death and rehospitalization for AF patients seen in the in emergency department, as well as identify patient groups for future targeted interventions aimed at preventing these outcomes. (C) 2018 Elsevier Inc. All rights reserved.