OBJECTIVE Our objective was to determine the influence of cytochrome P450 (CYP) 2C9 and CYP2C19 genetic polymorphisms on warfarin dose requirement and metabolic clearance. METHODS The study population consisted of 93 Italian outpatients receiving long-term warfarin anticoagulant therapy (international normalized ratio values, 2-3), divided into 3 dose groups: low (<26.25 mg/wk; n = 37), medium (26.25-43.75 mg/wk; n = 32), and high (>43.75 mg/wk; n = 24). Steady-state unbound plasma concentrations of S- and R-warfarin were measured by HPLC and equilibrium dialysis, and corresponding unbound oral clearance (CL(free)) values were calculated. Allelic variants of CYP2C9 (CYP2C9(*)2 and CYP2C9(*)3) and CYP2C19 (CYP2C19(*)2) were identified by polymerase chain reaction, followed by restriction enzyme analysis. RESULTS Fifty-four patients carried no CYP2C9 mutated alleles ((*)1/(*)1), 31 carried one ((*)1/(*)2, n = 15; and (*)1/(*)3, n = 16), and 8 carried two ((*)2/(*)2, n = 2; (*)3/(*)3, n = 2; and (*)2/(*)3, n = 4). Two subjects were homozygous and 19 were heterozygous for the CYP2C19(*)2 allele variant. The frequencies of CYP2C9 mutated alleles were 72% in the low-dose group, 36% in the medium-dose group, and 4% in the high-dose group; the corresponding mean S-warfarin CL(free) values were 307.5 mL/min, 480.3 mL/min, and 881.3 mL/min. The mean S-warfarin CL(free) values varied significantly among the CYP2C9 genotype groups (P <.0001), although most patients (72%) with no mutated alleles showed S-warfarin CL(free) values in the same range as those carrying mutated alleles (58-777 mL/min). No relationship was found between S-warfarin CL(free) and CYP2C19 genotype or between R-warfarin CL(free) and either CYP2C9 or CYP2C19 genotype. CONCLUSIONS CYP2C9 genetic polymorphisms markedly influence warfarin dose requirements and metabolic clearance of the S-warfarin enantiomer, although nongenetic factors may also contribute to their large interindividual variability.