Novel Single-Point Plasma or Saliva Dextromethorphan Method for Determining CYP2D6 Activity 1

Transcription

1 /98/ $03.00/0 THE JOURNAL OF PHARMACOLOGY AND EXPERIMENTAL THERAPEUTICS Vol. 285, No. 3 Copyright 1998 by The American Society for Pharmacology and Experimental Therapeutics Printed in U.S.A. JPET 285: , 1998 Novel Single-Point Plasma or Saliva Dextromethorphan Method for Determining CYP2D6 Activity 1 OLIVER YOA-PU HU, HUNG-SHANG TANG, HSIEN-YUAN LANE, WEN-HO CHANG and TEH-MIN HU Pharmaceutical Research Institute and School of Pharmacy (O.Y.P.H., T.M.H.), Tri-Service General Hospital (H.S.T), National Defense Medical Center, Taipei and Laboratory of Biological Psychiatry (H.Y.L., W.H.C.), Taipei City Psychiatric Center, Taipei, Taiwan, Republic of China Accepted for publication January 28, 1998 This paper is available online at 4-Hydroxylation of debrisoquin, which is mediated by cytochrome P450 2D6 (CYP2D6), is a well-known example of polymorphic drug metabolism. O-Demethylation of the cough suppressant dextromethorphan to dextrorphan has been shown to co-segregate with 4-hydroxylation of debrisoquin (Schmid et al., 1985). In recent years, this innocuous substance has been used widely and safely to detect persons with CYP2D6 deficiency in Swiss populations (Schmid et al., 1985), in white French populations (Larrey et al., 1987; Jacqz et al., 1988; Frèche et al., 1990; Duche et al., 1993), in Caucasian populations in Germany (Hildebrand et al., 1989), in white Spanish citizens (Henthorn et al., 1989), in Caucasians in the United States (Straka et al., 1995), in Jordanians (Irshaid et al., 1993), in African American subjects (Marinac et al., 1995), in Hmong subjects (Straka et al., 1995) and even Received for publication October 28, This work was supported in part by the National Science Council (grant NSC B ). ABSTRACT O-Demethylation of dextromethorphan co-segregates with 4-hydroxylation of debrisoquin and is used for CYP2D6 phenotyping. In most previous studies, 8-h urinary samples were collected for determining the dextromethorphan metabolic ratio (dextromethorphan/dextrorphan molar ratio). In addition, a salivary sampling at 3 h had been suggested for the phenotyping. To evaluate the repeatability and validity of previously reported and other potential phenotyping methods, we determined the metabolic ratios from urine samples (for various intervals), or from plasma or saliva (at varying time points) after repetitive single doses of immediate-release or repetitive multiple doses of controlled-release dextromethorphan preparations. For the single-dose study, each of 12 subjects received 15 mg of immediate-release dextromethorphan in period I and period II, respectively, with a 1-week washout period. For the multipledose study, each of 16 subjects received 60 mg controlledrelease dextromethorphan twice daily for 5 days in period I and period II, respectively, with a 2-week washout period. Dextromethorphan and dextrorphan were assayed by high-performance liquid chromatography. In the single-dose study, most metabolic ratios revealed good repeatabilities for the two periods (paired t test). The metabolic ratio from urine collected for 4h,6h,8hor12hfrom plasma at any time between 1hand 5 h or at 8 h, or from saliva at 2hor6h,could reflect that from 0- to 24-h urine or AUC. In the multiple-dose study, all metabolic ratios revealed good repeatabilities. The plasma metabolic ratio at any time between 0.5 h and 10 h or the saliva metabolic ratio at any time between 3 h and 12 h, but not the urine metabolic ratio from any interval, could predict the metabolic ratio from ACU SS.The2h,3h,4hor5hplasma metabolic ratio and 6 h saliva metabolic ratios after a single dose correlated significantly with their corresponding multipledose metabolic ratio (r 0.8, P.05). In conclusion, the plasma sample at 2 h, 3 h, 4hor5horthesaliva sample at 6 h in either the single immediate-release (15 mg) or the multiple controlled-release dose (60 mg) procedure could be used for determining the dextromethorphan metabolic ratio. in patients with liver disease (Larrey et al., 1989). In these studies determination of metabolic phenotype involves having the subjects take one dose of dextromethorphan orally and then collect urine for 8 to 10 h. This protocol is not convenient for many subjects, however. In another study dextromethorphan metabolic ratios obtained from 0 4 h, 0 6 h and 0 8 h urinary samples were found to be constant (Küpfer et al., 1986). Hou et al. (1991) suggested salivary analysis for determination of dextromethorphan metabolic phenotype. Each of 62 volunteers was given a 50-mg capsule of dextromethorphan hydrobromide and then urine (0 8 h) and saliva were collected (at 3 h). The correlation coefficient for the logarithm of urinary metabolic ratio vs. that of salivary metabolic ratio was In their studies, salivary analysis requires a larger dose, 50 mg, of dextromethorphan. Consequently, 22 of 61 subjects had mild and transient symptoms after taking 50 mg dextromethorphan. The symptoms occurred more often in ABBREVIATIONS: CYP2D6, cytochrome P450 2D6; HPLC, high-performance liquid chromatography; AUC, area under the plasma drug concentration time curve from 0 to infinity; AUC SS, steady-state area under the plasma drug concentration time curve within a dose interval. 955

2 956 Hu et al. Vol. 285 intermediate metabolizers, poor metabolizers and Chinese subjects. The occurrence of symptoms was associated with higher salivary concentrations of dextromethorphan but not with dextrorphan or body weight. Although dextromethorphan had been used widely for CYP2D6 phenotyping, the previously reported and other potential sampling methods, such as those with shorter urinecollecting intervals or appropriate saliva- or plasma-sampling time points, or those using controlled-released preparations which would have higher concentrations and fewer side effects, had not yet been extensively studied in terms of their repeatability and validity. Hence, the main aims of this study were (1) to rigorously evaluate the repeatability and validity of previously used methods for determining dextromethorphan metabolic ratios; (2) to develop a simple method that could precisely and reproducibly measure the dextromethorphan metabolic ratio from urine, plasma or saliva; and (3) to determine whether a reduction of the dosage or the utilization of a controlled-release preparation could reduce the side effects without influencing the usefulness of the methods for determining the metabolic ratio. Materials and Methods Subjects. Nineteen healthy unrelated Chinese male volunteers participated in the study. Their body weights ranged from 54.4 to 86 kg (mean S.D., kg), within 10% of their ideal body weights, and their ages ranged from 22 to 34 years (mean S.D., 26 3 years). All subjects were nonsmokers and physically healthy as determined by complete physical and laboratory examinations such as complete blood count (hemoglobin, hematocrit, red blood cell count, red cell indices, white blood cell count with differential, and platelet count), blood urea nitrogen, serum creatinine, serum glutaminic-oxaloacetic transaminase and blood sugar before the study. The subjects were instructed to abstain from any drug for at least 2 weeks before or during the study. Subjects with a history of drug or alcohol abuse or drug sensitivity were excluded. Among the 19 subjects, 9 took part in both repetitive single- and multiple-dose studies, whereas others were recruited into either. Hence, 12 subjects entered the repetitive single-dose study and 16 did the repetitive multiple-dose study. To each subject the study was explained and an informed consent was obtained from each. The study was approved by the Institutional Review Board of the National Defense Medical Center. Study protocol. We determined the metabolic ratios by both repetitive single and multiple doses of dextromethorphan. For the single-dose study, each of 12 subjects received a 15 mg immediaterelease dextromethorphan tablet (Medicone, 15 mg, Batch no. 3002, Shionogi & Co. Ltd., Japan) in period I and another in period II with a 1-week washout between the two administrations. On day 0, the subjects received a complete physical examination, a medical history was taken and laboratory tests were performed. On the morning of day 1 of each study period, a tablet containing 15 mg of dextromethorphan was given with 200 ml of water to fasted subjects. Neither food nor drink except water was permitted until 4 h after dosing. Urine specimens during 0 2, 2 4, 4 6, 6 8, 8 12 h and h after dosing were collected respectively. Urine volume over each collection interval was measured. Venous blood samples were obtained at 0.5, 1, 1.5, 2, 3, 4, 5, 6, 8 and 12 h after dosing. The blood samples were centrifuged immediately to express plasma. Saliva samples were obtained at 2, 3, 4, 5, 6, 8 and 12 h. A 5- to 10-ml saliva specimen was collected for 2 to 5 min for each sampling. All saliva samples were centrifuged and the sediment was discarded. All samples were frozen at 70 C until assay. For the multiple-dose study, each of 16 subjects received a controlled-release tablet of 60 mg dextromethorphan (tues Hustenstiller, Germany) twice a day for 5 days to achieve a steady state in period I. After a 2-week washout, each subject entered period II with the same procedure as period I. After the last dose, urine, plasma or saliva samples were collected according to the same schedule as that of the single-dose study. The saliva sample at 2 h, nevertheless, was not collected. In addition, other details were similar to those of the single-dose study. The nine subjects taking part in both studies entered the single-dose study first and then the multiple-dose study with a 1-week washout between the two studies. Analytic methods. Dextromethorphan and dextrorphan in urine, plasma and saliva at each sampling interval (or point) were assayed by HPLC. To 5-ml urine samples, 1-ml plasma samples or 3-ml saliva sample, 100 l (300 ng/ml) of internal standard levallorphan, 100 l of 28% NH 4 OH and 5 ml of 10% n-butanol in hexane were added. The mixture was rotated for 30 min and centrifuged at 4500 rpm for 10 min. The upper organic layer was collected, and 300 l of0.1nhcl was used to extract the drug from the organic layer by vortexing 20 min followed by another 5-min centrifuge. The acid layer then was separated and injected into HPLC. The injected volume was 40 l for urine and 200 l for plasma or saliva, respectively. All reagents were analytical grade. The chromatographic solvents used in the chromatographic separation were HPLC grade. The HPLC set being used was equipped with a Shimadzu LC9A pump, a Waters 717 WISP autosampler and a Shimadzu F551 fluorescence detector. Separations were performed on a reverse-phase phenyl column (Zorbax, Shimadzu, Japan) with the column maintained at a constant temperature of 40 C. Levallorphan (100 l; 300 ng/ml) was chosen to be the internal standard for this experiment. The composition of the mobile phase was 10 mm potassium phosphate and acetonitrile (1:1, ph 4.0 with 8.5% phosphoric acid). All water was Milli-Q grade. A 1.0 ml/min flow rate was used, and the eluent was monitored by a Shimadzu F551 fluorescence detector at excitation and emission 280 and 310 nm wavelengths, respectively. With the above-mentioned conditions, the retention times of dextromethorphan, dextrorphan 3-OH, 3-methoxy and levallorphan were 13.0, 6.0, 5.4, 10.3 and 8.6 min, respectively. Linearity was observed in a standard curve of dextromethorphan and dextrorphan across a range of 1 to 10,000 ng/ml, 1 to 300 ng/ml and 1 to 200 ng/ml for urine, plasma and saliva, respectively. Within-day and between-day standard curves in plasma, saliva and urine with dextromethorphan and dextrorphan (n 6) had a correlation coefficient greater than 0.99, and for dextromethorphan and dextrorphan, the coefficient of variance (%) was less than 15% for all the concentrations. The respective extraction recovery in plasma for dextromethorphan and dextrorphan were 75.5 to 97.9% and 53.4 to 72.0% from 1 to 300 ng/ml. The limit of quantitation was 1 ng/ml for both compounds in urine, plasma and saliva. Calibration curves were prepared daily for dextromethorphan and dextrorphan in respective matrixes. Figure 1 shows the representative HPLC chromatogram for the parent compound and three metabolites, dextrorphan, 3-OH- and 3-methoxydextrorphan. Figure 2 shows the within-day calibration curves (n 6) for dextromethorphan and these metabolites. Data analysis. Dextromethorphan metabolic ratios (dextromethorphan/dextrorphan molar ratios) from all urine-sampling intervals and all plasma- and saliva-sampling points were calculated. The data from 0.5 h plasma, 12 h plasma, 8 h, and 12 h saliva of the single-dose study, however, were excluded from analyses because for these sampling methods both compounds were detectable in not more than four subjects in at least one period. A paired t test was used to compare the metabolic ratio of period I and that of period II for each urinary, plasma (including AUC in the single-dose study and AUC SS in the multiple-dose study) or salivary sampling method. If no significant difference was revealed between the metabolic ratios of the two periods (P.05), the method was defined as repeatable, and the average of the metabolic ratios from the two periods was used for the consequent analyses. Pearson s productmoment correlation was used to investigate the relationships of

3 1998 New Method for Phenotype CYP2D6 957 Fig. 1. HPLC chromatogram for blank plasma and 1 ng/ml of dextromethorphan (retention time, 13.0 min), dextrorphan (retention time, 6.0 min), 3-OH-dextrorphan (retention time, 5.4 min) and 3-methoxydextrorphan (retention time, 10.3 min) in plasma. metabolic ratios from various sampling methods: the metabolic ratio from AUC or 0 24 h urine versus those from other sampling procedures in the single-dose study, the metabolic ratio from AUC SS or 0 24 h urine versus those from other procedures in the multiple-dose study and each parameter from the single-dose method versus the corresponding one from the multiple-dose method. A correlation coefficient greater than 0.8 with the P value less than.05 was regarded as a significant correlation. Results The single-dose study. For each urinary, plasma (including AUC ), or salivary sampling method, except the 3h(P.045) and 5h(P.0008) salivary sampling, no significant difference was revealed between the metabolic ratios of period I and period II by a paired t test. The P value of the paired t test for the metabolic ratios from the two periods of the salivary sampling at 4 h was.055, just over.05. Thereafter, the average of the metabolic ratios from the two periods of each method, except the 3 h, 4 h and 5 h salivary sampling, was used for further analyses as shown in tables 1, 2 and 3. With poor repeatabilities, the single-dose salivary sampling procedures at 3 h, 4 h and 5 h were not appropriate for determining dextromethorphan metabolic ratios. In table 1, correlation coefficients of the metabolic ratio from AUC vs. the ratios from all other sampling methods, except 0 2 h urine and 6 h plasma, were all greater than 0.80 (each P.05). Surprisingly, the correlation coefficient of the metabolic ratio from AUC vs. the ratio from 0 24 h urine reached Therefore, in table 2, correlation coefficients of the metabolic ratio from 0 24 h urine vs. the ratios from all other sampling methods were equal to the corresponding correlation coefficients in table 1. Nevertheless, the corresponding slopes and the intercepts in the two tables were different. For each subject, side effects were not evident during the single-dose study. The metabolic ratios of 0 8 h urine ranged from to (mean S.D.: ). No poor metabolizer was found according to the demarcation point of a dextromethorphan/dextrorphan ratio of 4.0, obtained from urinary samples untreated with -glucuronidase (Hou et al., 1991). The multiple-dose study. For each urinary, plasma (including AUC SS ratio), or salivary sampling method, no significant difference was noted between the two study periods by a paired t test. In terms of repeatabilities, all multipledose sampling procedures were suitable for determining dextromethorphan metabolic ratios. Hence, the average of the metabolic ratios from the two periods of each method was used for further analyses shown in tables 3, 4 and 5. In table 4, the correlation coefficient for the metabolic ratio from AUC SS vs. the ratio from each urine-sampling method was between 0.59 and 0.75 (P.05). Nevertheless, the ratio from AUC SS correlated highly with each of the ratios at all plasma-sampling points (from 0.5 h to 12 h) with all correlation coefficients greater than 0.95 (each P.0001). Among them, the correlation between multiple-dose metabolic ratios from AUC SS and multiple-dose plasma metabolic ratios at 4 h was shown in figure 3. Besides, the correlation coefficient for the ratio from AUC SS vs. the ratio from each salivasampling point was greater than 0.85 (each P.0005). In table 5, correlation coefficients for the metabolic ratio in 0 24 h urine vs. the ratio in 0 12 h urine, in 0 8 h urine, in 0 6 h urine, in 0 4 h urine and even in 0 2 h urine all exceeded 0.9 (each P.0001). However, the correlation coefficient for the metabolic ratio in 0 24 h urine vs. each of the ratios from all plasma-sampling points (excluding 2 h and 12 h) and from all saliva-sampling points (excluding 12 h) was below 0.8. For each subject, side effects were not evident throughout the multiple-dose procedure with the controlled-release dextromethorphan preparations. The correlation between the single-dose and the multiple-dose studies. Table 3 displays Pearson s productmoment correlation between each sampling method in the multiple-dose study and the corresponding method in the single-dose study. All corresponding methods, except plasma concentrations at 1 h, 1.5 h, 6 h and 8 h, significantly correlated with each other (r 0.8, P.05). Discussion The present repetitive single- and multiple-dose studies were conducted to rigorously evaluate the validity and repeatability of previously used methods, such as urinary sampling for about 8 h (Schmid et al., 1985; Larrey et al., 1987; Jacqz et al., 1988; Henthorn et al., 1989; Hildebrand et al., 1989; Frèche et al., 1990; Duche et al., 1993; Irshaid et al., 1993; Marinac et al., 1995; Straka et al., 1995) or 4h(Küpfer et al., 1986; Lam and Rodriquez, 1993) and salivary sampling at 3 h (Hou et al., 1991), for determining dextromethorphan metabolic ratios and to develop a method that could simply, precisely, reproducibly and safely measure the metabolic ratio from urine, plasma or saliva. In addition, the controlledrelease dextromethorphan preparations were used in the

4 958 Hu et al. Vol. 285 Fig. 2. Within-day calibration curves (n 6) of 1 to 300 ng/ml dextromethorphan, dextrorphan, 3-OH-dextrorphan and 3-methoxydextrorphan in plasma. TABLE 1 metabolic ratio of AUC (the y-axis) and those of other parameters (the x-axis) in the single-dose study 0 2 h urine h urine h urine h urine h urine h urine h plasma h plasma h plasma h plasma h plasma h plasma h plasma h plasma h saliva h saliva multiple-dose study to increase the dextromethorphan and dextrorphan concentrations and to reduce the possible side effects. In the present single-dose study, the sampling method with a salivary analysis at 3 h, 4 h or 5 h, was not as repeatable as that at 6 h after repetitive doses. All other methods in the present single- and multiple-dose studies were repeatable as well (paired t test, P.05). Therefore, a single saliva sample at 6 h, rather than one at 3 h after a single dose, might be more suitable for determining the metabolic ratio. Hou et al. (1991) also suggested further studies to evaluate the possibility of reducing the dosage from 50 mg for lessening potential side effects. In this study, a single tablet containing 15 mg dextromethorphan proved enough for the salivary analysis in most subjects (table 1). In addition, the 60-mg controlled-release dextromethorphan tablet was used for the multiple-dose procedure. Side effects were not evident throughout the single- and the multiple-dose procedures for each subject. No side effect induced even by the multiple-dose 60 mg dextromethorphan was caused by the controlled-release dosage form and, consequently, the TABLE 2 metabolic ratio of 0 24 h urine (the y-axis) and those of other parameters (the x-axis) in the single-dose study 0 2 h urine h urine h urine h urine h urine h plasma h plasma h plasma h plasma h plasma h plasma h plasma h plasma h saliva h saliva lower peak plasma concentration. Thus, the 60-mg controlled-release preparation used in the multiple-dose study has both advantages: easy laboratory measurement and the lack of side effects. The present single-dose study also demonstrates that with the metabolic ratio from AUC or 0 24 h urine as a reference, simple plasma sampling (at any point between 1 h and 5 h or at 8 h), saliva sampling (at2hor6h)orurine collection (for any interval between 4 h and 12 h) is feasible for determining the metabolic ratio (tables 1 and 2). The simple plasma or saliva sampling is a single-point method to determine dextromethorphan metabolic ratios instead of collecting urine for at least 4 h. We further tested the validity of the sampling methods with the multiple-dose procedures. In the multiple-dose study, the plasma metabolic ratio at any point between 0.5 h and 12 h or saliva metabolic ratio at any point between 3 h and 12 h, but not the urine metabolic ratio from any interval (including 0 24 h), could predict the metabolic ratio from AUC SS (r 0.85, P.0005) (table 4). Because the data from AUC SS could reflect more directly metabolic ratios than

5 1998 New Method for Phenotype CYP2D6 959 TABLE 3 Pearson s product-moment correlation between the dextromethorphan metabolic ratio from each multiple-dose sampling method (the y-axis) and that from the corresponding single-dose method (the x-axis) 0 2 h urine h urine h urine h urine h urine h urine h plasma h plasma h plasma h plasma h plasma h plasma h plasma h plasma AUC h saliva TABLE 4 metabolic ratio of AUC SS (the y-axis) and those of other parameters (the x-axis) in the multiple-dose study 0 2 h urine h urine h urine h urine h urine h urine h plasma h plasma h plasma h plasma h plasma h plasma h plasma h plasma h plasma h plasma h saliva h saliva h saliva h saliva h saliva h saliva TABLE 5 metabolic ratio of 0 24 h urine (the y-axis) and those of other parameters (the x-axis) in the multiple-dose study 0 2 h urine h urine h urine h urine h urine h plasma h plasma h plasma h plasma h plasma h plasma h plasma h plasma h plasma h plasma h saliva h saliva h saliva h saliva h saliva h saliva Fig. 3. The correlation between multiple-dose dextromethorphan metabolic ratios from AUC SS and multiple-dose plasma dextromethorphan metabolic ratios at 4h(r , P.0001, n 16). those from 0 24 h urine, the above-mentioned any-point plasma or saliva could reveal the metabolic ratio better than the urine from any interval in the multiple-dose procedure. Finally, we tested the validity of the sampling method with the correlation between the single-dose and multiple-dose procedures. Theoretically, the multiple-dose steady-state methods should be more valid than the single-dose methods in determining the metabolic ratio, simply because the steady state has been reached with the drug distribution in equilibrium among plasma, tissue, saliva and urine. In table 3, nevertheless, all corresponding methods, except plasma concentrations at 1 h, 1.5 h, 6 h and 8 h, in the single- and the multiple-dose studies significantly correlated with each other. Therefore, all sampling methods in the single-dose study, except plasma concentrations at 1 h, 1.5 h, 6 h and 8 h, could be almost as valid as the corresponding multiple-dose methods listed in table 3. However, as mentioned above, in the multiple-dose procedure the urinary metabolic ratio from any interval could not predict the metabolic ratio from AUC SS. Based upon the results of these validity tests, we suggested that the plasma sample at 2 h, 3 h,4hor5horthe saliva sample at6hineither the single immediate-release or the multiple controlled-release dose procedure could be used for determining the dextromethorphan metabolic ratio. The incidence rate of CYP2D6 deficiency in Chinese subjects is about 1% (Wang et al., 1993; Lane et al., 1996). Hence it was expected not to find any poor metabolizer in the present study with Chinese subjects. Furthermore, among extensive metabolizers, up to a thousandfold difference in metabolic ratios was reported in various populations (Schmid et al., 1985; Larrey et al., 1987; Hildebrand et al., 1989; Henthorn et al., 1989; Irshaid et al. 1993; Straka et al. 1995;

6 960 Hu et al. Vol. 285 Marinac et al. 1995; Lane et al., 1996). Hence, determining the actual metabolic ratio could provide more valuable information than merely differentiating between poor and extensive metabolizers. In most previous studies (Schmid et al., 1985; Küpfer et al., 1986; Larrey et al., 1987; Jacqz et al., 1988; Henthorn et al., 1989; Hildebrand et al., 1989; Frèche et al., 1990; Lam and Rodriquez, 1993; Marinac et al., 1995; Straka et al., 1995), metabolic ratios were determined in urinary samples that were treated with -glucuronidase (with or without sulfatase) before extraction. Because dextrorphan, not dextromethorphan, can further conjugate with glucuronide (Duche et al., 1993), the measured dextrorphan was the sum of conjugated and unconjugated dextrorphan. Hou et al. (1991) compared the urinary metabolic ratios of -glucuronidasetreated samples with those of untreated samples; a good correlation (r 0.987) was found in 19 subjects. Subsequently, metabolic ratios from untreated urinary samples were used for phenotyping in all 61 subjects. The study of Duche et al. (1993) also showed that the dextromethorphan/ free dextrorphan ratio, as well as the dextromethorphan/ total dextrorphan ratio, could be used to determine the phenotype. Theoretically, once the equilibrium has been reached, the dextromethorphan/free dextrorphan ratio should correlate well with the dextromethorphan/total dextrorphan ratio, although the absolute values of the two ratios may differ. As mentioned in the present study, several multiple-dose methods also correlated well with their corresponding single-dose methods. Therefore, the methods without deconjugation (from either single- or multiple-dose procedures) could be as valid as the methods with deconjugation for determining metabolic ratios. In conclusion, the plasma sample at 2 h, 3 h, 4hor5hor the saliva sample at 6 h in either the single immediaterelease or the multiple controlled-release dose procedure could be suggested to be widely used for determining the dextromethorphan metabolic ratio with easy laboratory measurement and negligible side effects. References Duche JC, Querol-Ferrer V, Barre J, Mesangeau M and Tillement JP (1993) Dextromethorphan O-demethylation and dextrorphan glucuronidation in French population. Int J Clin Pharmacol Ther Toxicol 31: Frèche JP, Dragacci S, Petit AM, Siest JP, Galteau MM and Siest G (1990) Development of an ELISA to study the polymorphism of dextromethorphan oxidation in a French population. Eur J Clin Pharmacol 39: Henthorn TK, Benitez J, Avram MJ, Martinez G, Llerena A, Cobaleda J, Krejcie TC and Gibbons RD (1989) Assessment of the debrisoquin and dextromethorphan phenotyping tests by gaussian mixture distributions analysis. Clin Pharmacol Ther 45: Hildebrand. M, Seifert W and Reichenberger A (1989) Determination of dextromethorphan metabolizer phenotype in healthy volunteers. Eur J Clin Pharmacol 36: Hou ZY, Pickle LW, Meyer PS and Woosley RL (1991) Salivary analysis for determination of dextromethorphan metabolite phenotype. Clin Pharmacol Ther 49: Irshaid YM, Al-Hadidi HF and Rawashdeh NM (1993) Dextromethorphan O- demethylation polymorphism in Jordanians. Eur J Pharmacol 45: Jacqz E, Dulac H and Mathieu H (1988) Phenotyping polymorphic drug metabolism in the French Caucasian population. Eur J Clin Pharmacol 35: Küpfer A, Schmid B and Praff G (1986) Pharmacogenetics of dextromethorphan O-demethylation in man. Xenobiotica 16: Lam YWF and Rodriquez SY (1993) HPLC determination of dextromethorphan and dextrorphan for oxidation phenotyping by fluorescence and ultraviolet detection. Ther Drug Monit 15: Lane HY, Deng HC, Huang SM, Hu WH, Chang WH and Hu OYP (1996) Low frequency of dextromethorphan O-demethylation deficiency in a Chinese population. Clin Pharmacol Ther 60: Larrey D, Amouyal G, Tinel M, Letteron P, Berson A, Labbe G and Pessayre D (1987) Polymorphism of dextromethorphan oxidation in French population. Br J Clin Pharmacol 24: Larrey D, Babany G, Tinel M, Freneaux E, Amouyal G, Habersetzer F, Letteron P and Pessayre D (1989) Effect of liver disease on dextromethorphan oxidation capacity and phenotype: A study in 107 patients. Br J Clin Pharmacol 28: Marinac JS, Foxworth JW and Willsie SK (1995) Dextromethorphan polymorphic hepatic oxidation (CYP2D6) in healthy black American adult subjects. Ther Drug Monit 17: Schmid B, Bircher J, Preisig R and Kupfer A (1985) Polymorphic dextromethorphan metabolism: Co-segregation of oxidative O-demethylation with debrisoquin hydroxylation. Clin Pharmacol Ther 38: Straka RJ, Hansen SR and Walker PF (1995) Comparison of the prevalence of the poor metabolizer phenotype for CYP2D6 between 203 Hmong subjects and 280 white subjects residing in Minnesota. Clin Pharmacol Ther 58: Wang SL, Huang JD, Lai MD, Liu BH and Lai ML (1993) Molecular basis of genetic variation in debrisoquin hydroxylation in Chinese subjects: Polymorphism in RFLP and DNA sequence of CYP2D6. Clin Pharmacol Ther 53: Send reprint requests to: Dr. Oliver Yoa-Pu Hu, Pharmaceutical Research Institute and School of Pharmacy, National Defense Medical Center, PO Box , Shih-Yuan St., Taipei, Taiwan, R.O.C.