Pôle de biologie, pharmacie et santé publique,
Inserm, CIC-P 802, Centre hospitalier universitaire de Poitiers, Université de Poitiers, Poitiers, France
Preventing atrial fibrillation (AF) complications relies mainly on anticoagulant therapy. Still it is difficult to prescribe vitamin K antagonists (VKA) in geriatric patients with AF. In order to improve anticoagulation decision in this disease, we set up an algorithm. Charts of all patients with AF hospitalized between February and May 2012 were reviewed. Patients treated with anticoagulation for another indication (venous thromboembolism disease, prosthetic valve) were excluded. Algorithm was built-up with 6 criteria (past bleeding with VKA, autonomy (GIR score), MMSE score, risk of falls, co-morbidities index). Each criterion had a score (0, 0.5, 1 point) according to an intensity scale (light, moderate, high). The final algorithm composite score led to the prescription or not of VKA. Patients were followed-up during 6 months after discharge. One hundred and fifty-three patients were included, mean age 86.1 ± 5.6 years; 67.3% had a GIR score ≤3, 70.6% MMSE score < 23, and 83.7% a moderate risk of falls. According to the algorithm, 92 patients (60.1%) had a VKA prescription. Prescription was significantly less prescribed in the oldest old (p=0.02). Follow-up showed 4 bleeding events without any link with VKA prescription. Thirty-four patients died (22.2%), among 24 (34.4%) who did not have VKA (p=0.005). The algorithm improves VKA prescription according to an objective evaluation and probably prevents the prescription in the patients with the worse short term prognosis.