OBJECTIVES: Atrial fibrillation (AF) is the most common cardiac rhythm disorder; it increases stroke risk 5 to 6 times. Prevalence in Colombia has been estimated in 3.6% in population age 60 or over (some 180 000 potential patients). The aim of this study was to estimate cost-effectiveness of dabigatran 110 mg and 150 mg BID as a therapy for non valvular AF in Colombian population. METHODS: From a third-party payer perspective (Colombian health system) we used a three-month cycle Markov model with six health states (and death): non-disabling stroke, disabling stroke, acute myocardial infarction and pulmonary embolism; two additional events were transient states: minor and mayor bleeding. Transition probabilities and proportion of events were extracted from the RELY trial; utilities were derived from the literature. Costs for medications and procedures were obtained from official government databases, all costs were in 2014 Colombian pesos (1 USD = 2000.33 COP). Annual discount rate was 5% and we used a life time horizon (close to 20 years, on average). Cost-effectiveness threshold was 3 times per capita GDP (around USD 22,500). RESULTS: Compared with warfarin, patients treated with dabigatran 150 and 110 mg gained, on average 0.37 and 0.23 life-years, respectively, or 0.38 and 0.25 QALYs. The ICER for dabigatran 150 mg was USD 13,248 per QALY, and for dabigatran 110 mg was 23,621 per QALY. CONCLUSIONS: Dabigatran 150, compared with warfarin, the standard therapy, is cost-effective for ambulatory treatment of patients with non valvular AF. Dabigatran 110 ICER is discretely over the threshold (around 1000 USD).