Warfarin is a common anticoagulant with narrow therapeutic window and variable anticoagulation effects. Single gene polymorphisms in cytochrome P450 2C9 (CYP2C9) and vitamin K epoxide reductase (VKORC1) have been shown to impact warfarin dosing in adults. Insufficient data exists on genetic and clinical factors which influence warfarin dosing in children. Pediatric patients with heart disease who received long-term warfarin therapy were tested for VKORC1 and CYP2C9 polymorphisms. Clinical and demographic data were reviewed in those children who achieved stable therapeutic international normalized ratio (INR). Multiple linear regression modeling was used to assess relationships between stable warfarin doses and genetic or clinical variables. Fifty children were tested for VKORC1 and CYP2C9 polymorphisms; 37 patients (M 26: F 11) had complete data, achieved stable therapeutic INR, and were included in dose variability analysis. There were predominance of white race 73% and male sex 70.3%. The mean age was 9.6 years (1.8-18.6 years). The mean weight was 37.8 kg (7.7-95 kg). Fontan physiology and mechanical cardiac valves were two most common indications for chronic warfarin therapy (25/37 or 67.6%). Twelve patients (32.4%) had >/=2 indications for warfarin therapy. Three patients had documented venous or arterial clots, and 5 patients had strokes. Congenital heart disease was present in 29 patients (78.4%), including Fontan physiology (20), complex biventricular physiology (4), and congenital mitral valve disease (5). Acquired heart disease was present in 8 patients (21.6%), including Kawasaki disease with coronary aneurysms (3), acquired mitral valve disease (3), and Marfan syndrome (2). Stable warfarin dose (mg/kg/day) was strongly associated with VKORC1 polymorphism (p < 0.0001) and goal therapeutic INR (p = 0.009). Negative correlations were observed between stable warfarin dose and age, weight, height, and BSA (p = 0.04, 0.02, 0.02, and 0.02 respectively). Factors which did not influence warfarin dose included CYP2C9 polymorphism (p = 0.17), concurrent medications (p = 0.85), sex (p = 0.4), race (p = 0.14), congenital heart disease (p = 0.09), and Fontan physiology (p = 0.76). The gene-dose effect was observed in children with homozygous wild type VKORC1 CC, who required higher warfarin dose compared to those carrying heterozygous TC or homozygous TT (p = 0.028 and 0.0004 respectively). The full multiple linear regression model revealed that VKORC1 genotypes accounted for 47% of dosing variability; CYPC29 accounted for 5%. Overall, the combination of VKORC1, CYP2C9, age, and target INR accounted for 82% of dosing variability. In children with heart disease, VKORC1 genotypes, age, and target INR are important determinants influencing warfarin dosing in children with heart disease. Future warfarin dosing algorithm in children should factor both genetic and clinical factors.