AIM: Caffeine metabolite ratios in urine can be used to evaluate the activity of CYP1A2, CYP2A6, NAT2 and XO in vivo. In this study, an HPLC method was developed to quantify the concentrations of caffeine metabolites in 24 Chinese healthy volunteers. METHODS: 100 mg of caffeine was given to each subject as a probe drug and the urine samples of 0-12 h were collected. Urine samples were extracted with chloroform/isopropanol (9∶1, v/v) under acerbic environment (pH 3.5) and separated on a Hypersil BDS C18 column with gradient elution. The mobile phase was composed of phase A (acetonitrile) and phase B (10 mmol/L ammonium formate/formic acid (998/2, v/v) (pH 3.5) from 98/2 to 70/30, the detection wave length was 280 nm. CYP1A2 phenotype was calculated from the metabolite ratio of (AFMU+1X+1U)/17U, CYP2A6 from the ratio of 17U/(17U+17X+1U+1X+AFMU), NAT2 from the ratio of AFMU/(AFMU+1U+1X) and XO from the ratio of 1U/1X+1U. RESULTS: The calibration curves of AFMU, 1U, 1X, 17U, 17X, 137X were linear over the range of 1.25-160 μmol/L, yielding correlation of coefficients from 0.991 to 0.999. The LOD were 2.5 ng/mL for AFMU, 1X, 1U and 1 ng/mL for 17U, 17X, 137X. Recoveries for the analytes wereranged from 81%-94%, Intra-and Inter-day coefficients of variation were ranged from 2.56%-9.48% and 5.74%-10.91%, respectively. After logarithmic transformation of these metabolite ratios, phenotype of CYP1A2, CYP2A6, NAT2 were shown normal distribution, while the phenotype of XO was shown non-normal distribution with two subjects appearing with very low metabolic capacity. It was also found a negative association between the phenotype of CYP2A6 and CYP1A2 which might be explained by the fact that CYP2A6 plays an influential role in 17X metabolism. CONCLUSION: In brief, the experiment demonstrates an accurate, stable and replicable HPLC method for phenotyping of these enzymes in Chinese volunteers. AIM: Caffeine metabolite ratios in urine can be used to evaluate the activity of CYP1A2, CYP2A6, NAT2 and XO in vivo. In this study, an HPLC method was developed to quantify the concentrations of caffeine metabolites in 24 Chinese healthy volunteers. METHODS: 100 mg of caffeine was given to each subject as a probe drug and the urine samples were collected 0-12 h were collected. Urine samples were extracted with chloroform / isopropanol (9: 1, v / v) under acerbic environment The mobile phase was composed of phase A (acetonitrile) and phase B (10 mmol / L ammonium formate / formic acid (998/2, v / v) 2 to 70/30, the detection wave length was 280 nm. CYP1A2 phenotype was calculated from the metabolite ratio of (AFMU + 1X + 1U) / 17U, CYP2A6 from the ratio 17U / (17U + 17X + 1U + 1X + AFMU ), NAT2 from the ratio of AFMU / (AFMU + 1U + 1X) and XO from the ratio of 1U / 1X + 1U. RESULTS: The calibration curves of AFMU, 1U, 1X, 17U, 17X, 137X we The LOD were 2.5 ng / mL for AFMU, 1X, 1 U and 1 ng / mL for 17U, 17X, 137X. Recoveries for the reline over the range of 1.25-160 μmol / L, yielding correlation of coefficients from 0.991 to 0.999. respectively. After logarithmic transformation of these metabolite ratios, phenotype of CYP1A2, CYP2A6, NAT2 were shown normal distribution, while the phenotype of XO was shown non-normal distribution with two subjects appearing with very low metabolic capacity. It was also found a negative association between the phenotype of CYP2A6 and CYP1A2 which might be explained by the fact that CYP2A6 plays an influential role in 17X metabolism. CONCLUSION: In brief, the experiment demonstrates an accurate, stable and replicable HPLC method for phenotyping of these enzymes in Chinese volunteers.