Introduction. Benzodiazepine tranquilizers usage is the generally accepted approach to the alcohol withdrawal syndrome (AWS) treatment. The most common benzodiazepine for AWS therapy in Russia is phenazepam. This drug is metabolized by CYP3A4 and CYP3A5 enzymes. The correlation between CYP3A4 polymorphisms and phenazepam’s safety has not been previously investigated. Matherials and methods. 102 male patients with non-complicated AWS (F 10.3 by ICD-10) were involved into the study. For the 5 days of dynamic observation each participant of the study was prescribed phenazepam. To detect the CYP3A4*22 (rs35599367) polymorphism 5 ml of venous blood from each patient was collected. On the first and the sixth days of the study 10 ml of urine was obtained from each participant. The measurement of 6-beta-hydroxycortisol/cortisol ratio in the urine samples was used to define CYP3A4 activity. To evaluate the safety of therapy UKU Side-Effects Rating Scale was applied. Statistical analysis was performed with SPSS Statistics 21.0. Results. Participants in two groups were comparable in demographic, clinical and anamnestic parameters. Overall UKU rate, rates of Psychiatric and Autonomous nervous system UKU subscales and frequency and severity of side effects did not differ significantly between carriers of polymorphic and wild-type CYP3A4*22 variants. In CYP3A4*22 carriers was observed trend towards significance in increase of CYP3A4 activity (p = 0.051). However there were no statistically significant differences in CYP3A4 inducers prescription between CC-homozygous and T allele carriers. Conclusion. In this study the association between the CYP3A4*polymorphism and phenazepam’s safety in patients with AWS was not detected. The differences of CYP3A4 activity in CC carriers and T allele carriers did not achieve statistical significance.