OBJECTIVES: To estimate monthly medical costs of controlled and inadequately-controlled non-valvular atrial fibrillation (NVAF) patients treated with Vitamin K antagonists (VKA). METHODS: Retrospective cohort study of NVAF patients newly initiating VKA between January 2009 and December 2014 using a database from a healthcare provider from Badalona (Spain) (120,000 inhabitants). We followed patients from first VKA treatment until study end, leaving database, VKA discontinuation, or death; calculated proportion of patients with adequate anticoagulation (time in therapeutic range (TTR) ≥60% using target international normalized ratio (INR) 2-3) over 12 months post-initiation; and estimated monthly mean healthcare costs linked to NVAF for controlled and inadequately-controlled patients, adopting a payer perspective. RESULTS: The study included 1,144 patients (median(IQR) follow-up 28.8 (14.2–43.1) months) with 67.1% well-controlled patients over 0-12 month period. At baseline, for controlled and inadequately-controlled patients respectively, mean(SD) age was 74.3(10.3) and 74.3(10.7) years and 48.2% and 54.3% were female. Mean Charlson index was 0.9(1.4) and 1.1(1.4), (p:0.025), and CHA2DS2-VASc scores were 2.7(1.6) and 2.9(1.6) (p:0.038) for controlled and inadequately-controlled patients, respectively. Mean(SD) monthly medical costs were €80.3(174.8) and €143.6(656.6) respectively, being 1.8-fold higher for inadequately-controlled patients (p:0.013). The largest cost components were general practitioner visits, hospitalizations, and specialist visits. INR testing was also more costly in inadequately-controlled patients (mean monthly INR cost(SD) €2.3(1.1) versus €2.7(1. 9) p:<0.001 for controlled and inadequately-controlled patients respectively) and accounted for 1.9% and 2.9% of costs, respectively. Other time periods assessed gave qualitatively similar results. CONCLUSIONS: In this study, 1/3 of NVAF patients treated with VKAs were inadequately controlled. These patients had a higher baseline stroke risk and comorbidity index and generated higher direct medical costs associated with NVAF compared to controlled patients. In this context, future research could address the medical costs associated with novel anticoagulants in NVAF patients.