Warfarin dosage estimation using the pharmacogenetic algorithms has been shown to improve the quality of anticoagulation control in patients with atrial fibrillation. We sought to assess the genetic, demographic and clinical factors that determine the quality of anticoagulation in patients following aortic valve replacement (AVR). We studied 200 consecutive patients (130 men) aged 63 ± 12.3 years, undergoing AVR, in whom warfarin dose was established using a pharmacogenetic algorithm. The quality of anticoagulation within the first 3 months since surgery was expressed as the time of international normalized ratio (INR) in the therapeutic range (TTR). The median TTR in the entire cohort was 59.6% (interquartile range, 38.7 - 82.7). Ninety-nine (49.5%) patients with TTR ≥ 60% did not differ from those with poor anticoagulation control (TTR < 60%) with regard to demographic and cardiovascular risk factors. Coronary artery disease (n = 84, 42%) and previous stroke (n = 5, 2.5%) predicted higher TTR, while possession of CYP2C9*2 variant allele (n = 49, 25%) was associated with lower TTR (P = 0.01). In turn, VKORC1 c.-1639A, CYP2C9*2 and *3 variants were independently associated with actual warfarin dose (P < 0.0001). In AVR patients better anticoagulation control is observed in patients with coronary artery disease and history of stroke, which might result in part from previous lifestyle modification and therapy. Possession of CYP2C9*2 and/or CYP2C9*3 allele variants is associated with lower TTR values and warfarin dose variations in AVR patients, the latter affected also by VKORC1 c.-1693G>A polymorphism.