Our objective was to prospectively study the impact of CYP2C9 polymorphism (*2 and *3) on the risk of overanticoagulation during the induction phase of warfarin therapy. Blood samples for genotyping were collected from 219 patients requiring warfarin therapy, and clinical data were prospectively collected during the first 3 weeks of medication. Patients were divided into 3 groups according to CYP2C9 genotype, as follows: *1 (homozygous), *2 (*1/*2 and *2/*2), and *3 (any genotype containing the *3 allele). During the first week of treatment, the relative risk of achieving at least 1 international normalized ratio (INR) value above the therapeutic interval (2-3) was 2.8 (95% confidence interval, 1.2-6.7) and 6.1 (2.7-13.6) in the *2 and *3 groups, respectively (with *1 used as control). During the second week, the corresponding values were 2.1 (1.2-3.7) and 3.5 (2.1-5.8), respectively. By the third week, the genetic impact was no longer evident, presumably as a result of successful dose individualization. Increased INR levels (compared with the *1 group) were already demonstrated in the *2 group on the fourth treatment day. The CYP2C9*2 and *3 single-nucleotide polymorphisms significantly increase the risk of overanticoagulation during the first 2 weeks of warfarin treatment, with increased INR levels evident after only 4 days’ treatment in *2 carriers. Our prospective data are consistent with results from previous retrospective studies and indicate that CYP2C9 genotyping may be a means of improving safety during warfarin induction.