ABSTRACTBackgroundVitamin K antagonists (VKA) have a narrow therapeutic range, and literature analysis reveals poor quality of anticoagulation control. We sought to assess the prevalence of poor anticoagulant control in patients under VKA treatment in the prevention of stroke for atrial fibrillation (AF).HypothesisControl of anticoagulation with VKA is inadequate in a high percentage of patients with AF.MethodsPatients with AF under VKA treatment were prospectively recruited in this observational registry. The sample comprised 948 patients. The estimated time spent in the therapeutic range (TTR) was calculated, and variables related with a TTR >65% were analyzed.ResultsMean age was 73.8 ± 9.4 years, and 42.5% of the patients were women. Mean TTR was 63.77% ± 23.80% for the direct method and 60.27% ± 24.48% for the Rosendaal method. Prevalence of poor anticoagulation control was 54%. Variables associated with good anticoagulation control were university studies (odds ratio [OR]: 1.99, 95% confidence interval [CI]: 1.08‐3.64), chronic hepatic disease (OR: 8.15, 95% CI: 1.57‐42.24), low comorbidity expressed as Charlson index (OR: 0.87, 95% CI: 0.76‐0.99), no previous cardiac disease (OR: 0.64, 95% CI: 0.41‐0.98), lower risk of bleeding assessed as hypertension, abnormal renal/liver function, stroke, bleeding history or predisposition, labile international normalized ratio, elderly age, and use of drugs or alcohol (HAS‐BLED; OR: 0.81, 95% CI: 0.69‐0.95), and lower heart rate (OR: 0.99, 95% CI: 0.98‐0.99).ConclusionsPatients who receive VKA to prevent stroke for AF spend less than half the time within therapeutic range.