Introduction: Atrial fibrillation (AF) is the most common sustained cardiac rhythm abnormality. AF treatment is provided by either cardiologists or non-cardiologists and is based on the prevention of thromboembolic events and on heart rate control or rhythm control. There is evidence that AF management by specialists leads to improved patient outcomes. Moreover, management of complex cardiac arrhythmias by a specialized team has been considered helpful by an overwhelming majority of practicing physicians. Therefore, we designed a multidisciplinary team - Fast Approach by a Specialized Team for Atrial Fibrillation (FAST-AF) - to evaluate the treatment effectiveness of a specialized arrhythmia team compared to the standard of care in the management of patients with new-onset AF presenting in the emergency department (ED).

Methods: We propose an open-label, single-center, randomized, parallel group, superiority trial in an university hospital during a period of 36 months. 294 patients with new-onset AF presenting in the ED will be randomized in a 1:1 ratio to FAST-AF or standard of care. The primary outcome will be time to discharge from the beginning of the intervention. Secondary outcomes will be time to heart rate/rhythm control, compliance to anticoagulation after discharge, readmission rates, quality of life assessment (AFEQT) and hospitalization costs. For the primary analysis, an intention-to-treat (ITT) time to event (TTE) analysis will be conducted. The trial will be prospectively registered at ClinicalTrials.gov.

Conclusions: AF is a prevalent disease that causes several complications and morbidity. The heterogeneity of medical criteria, the lack of knowledge of international guidelines and inadequate adherence to treatment may compromise therapeutic success. This study design is readily replicable, given its simplicity and feasibility. If the study show promising results of FAST-AF intervention, future studies can be conducted  to assess cost reduction and specialized consultation by virtual meeting platforms.