The role of prognostic stratification on prescription of anticoagulants in older patients with atrial fibrillation: a multicenter, observational, prospective European study (EUROSAF)

Background Literature suggests that different risks of mortality could influence physicians in prescribing or not anticoagulants in older patients with atrial fibrillation (AF). The Multidimensional Prognostic Index (MPI) can be considered a tool for the detection of multidimensional frailty. The aim of this cross-sectional study was to evaluate whether prescription patterns of oral anticoagulants exist, based on MPI values.Methods Older hospitalised patients (age ≥ 65 years) with non-valvular AF were included across 24 European centres. MPI was calculated using validated and standardised tools derived from a comprehensive geriatric assessment. Other functional and clinical information were collected to calculate indexes specific for haemorrhagic and thromboembolic risk in AF.Results Altogether, 2,012 participants affected by AF (mean age was 83.2 ± 7.5, range: 65–104 years), with a higher presence of women (57.0%), were included. Overall, 440 took vitamin K antagonists VKAs (22.0%), 667 (33.4%) direct oral anticoagulants (DOACs), whilst 44.6% did not take any anticoagulant treatment. Prescription of anticoagulants was associated with MPI values, with people taking anticoagulants having lower mean MPI values. Anticoagulant therapy was not used in 53.1% of the group with the highest risk of mortality, compared with 32.3% of those in the group with the lowest mortality risk. People with higher scores in MPI were less frequently treated with anticoagulant therapy, after adjusting for several potential confounders.Conclusions The EURopean study of Older Subjects with Atrial Fibrillation (EUROSAF) suggested that almost half of the older persons with AF do not receive anticoagulants and that MPI is an important determinant in prescribing or not anticoagulants. Trial Registration: https://clinicaltrials.gov/ct2/show/NCT02973984KEY POINTSAtrial fibrillation is a common condition in older people. The data regarding the use of anticoagulants is mainly derived from randomised controlled trials that do not include a sufficient number of older frail people.Our study suggests that a consistent part of older people affected by atrial fibrillation was not treated with anticoagulants, in particular, older frail patients; however, it is unclear if this choice is supported or not by evidence.The prognostic evaluation through the multidimensional prognostic index could be useful information for the choice in the prescription of anticoagulants in older people affected by atrial fibrillation.