Pharmacokinetics of codeine and its metabolite morphine in ultra-rapid metabolizers due to CYP2D6 duplication

Transcription

1 (2007) 7, & 2007 Nature Publishing Group All rights reserved X/07 $ ORIGINAL ARTICLE Pharmacokinetics of codeine and its metabolite morphine in ultra-rapid metabolizers due to CYP2D6 duplication J Kirchheiner 1, H Schmidt 2, M Tzvetkov 3, J-THA Keulen 4, JLötsch 2, I Roots 4 and J Brockmöller 3 1 Department of Pharmacology of Natural Products and Clinical Pharmacology, University Ulm, Ulm, Germany; 2 Pharmazentrum Frankfurt/ ZAFES, Institute of Clinical Pharmacology, Johann Wolfgang Goethe-University, Frankfurt, Germany; 3 Department of Clinical Pharmacology, Georg August University Göttingen, Göttingen, Germany and 4 Institute of Clinical Pharmacology, Charite University Medicine Berlin, Berlin, Germany Correspondence: Dr J Kirchheiner, Department of Pharmacology of Natural Products and Clinical Pharmacology, University Ulm, Helmholtzstrasse, 20, Ulm Germany. julia.kirchheiner@uni-ulm.de Codeine is an analgesic drug acting on m-opiate receptors predominantly via its metabolite morphine, which is formed almost exclusively by the genetically polymorphic enzyme cytochrome P450 2D6 (CYP2D6). Whereas it is known that individuals lacking CYP2D6 activity (poor metabolizers, PM) suffer from poor analgesia from codeine, ultra-fast metabolizers (UM) due to the CYP2D6 gene duplication may experience exaggerated and even potentially dangerous opioidergic effects and no systematical study has been performed so far on this question. A single dose of 30 mg codeine was administered to 12 UM of CYP2D6 substrates carrying a CYP2D6 gene duplication, 11 extensive metabolizers (EM) and three PM. Genotyping was performed using polymerase chain reaction-restriction fragment length polymorphism methods and a single-base primer extension method for characterization of the gene-duplication alleles. Pharmacokinetics was measured over 24 h after drug intake and codeine and its metabolites in plasma and urine were analyzed by liquid chromatography with tandem mass spectrometry. Significant differences between the EM and UM groups were detected in areas under the plasma concentration versus time curves (AUCs) of morphine with a median (range) AUC of 11 (5 17) mghl 1 in EMs and 16 (10 24) mghl 1 in UM (P ¼ 0.02). In urine collected over 12 h, the metabolic ratios of the codeine þ codeine-6-glucuronide divided by the sum of morphine þ its glucuronides metabolites were 11 (6 17) in EMs and 9 (6 16) in UM (P ¼ 0.05). Ten of the 11 CYP2D6 UMs felt sedation (91%) compared to six (50%) of the 12 EMs (P ¼ 0.03). CYP2D6 genotypes predicting ultrarapid metabolism resulted in about 50% higher plasma concentrations of morphine and its glucuronides compared with the EM. No severe adverse effects were seen in the UMs in our study most likely because we used for safety reasons a low dose of only 30 mg. It might be good if physicians would know about the CYP2D6 duplication genotype of their patients before administering codeine. (2007) 7, ; doi: /sj.tpj ; published online 4 July 2006 Keywords: CYP2D6; ultrarapid metabolism; gene duplication; codeine Introduction Received 7 March 2006; revised 28 April 2006; accepted 24 May 2006; published online 4 July 2006 Codeine is widely used as anti-tussive and analgesic drug and it is known that O-demethylation of codeine into morphine is mediated by the genetically polymorphic enzyme cytochrome P450 2D6 (CYP2D6). 1 Whereas this

2 258 Codeine in CYP2D6 gene duplication carriers O-demethylation reaction accounts only for less than 10% of codeine clearance, it is regarded to be the bioactivation reaction essential for the analgesic activity of codeine preparations. 1 A strong correlation between debrisoquine metabolic ratio (MR), reflecting CYP2D6 activity, and the urinary recovery of codeine and its metabolites formed by O- and N-demethylation has been shown. 2 Poor metabolizers (PM) lacking CYP2D6 activity had extremely low plasma concentrations and urinary recovery rates of morphine and morphine glucuronides 2 5 and very little analgesic efficacy of codeine. 6,7 Correspondingly, they may have a lower risk of developing codeine addiction because the active principle acting at the m-opioid receptors is lacking. 8 On the other hand, several case reports have described severe opioid side effects after intake of codeine in individuals later identified as ultra-rapid metabolizers (UM) of CYP2D6 substrates. 9,10 The CYP2D6 gene duplication leads to ultra-rapid metabolism if the duplicated genes are fully active and if the duplication is combined with another active CYP2D6 allele. Usually, these individuals have high CYP2D6 activity, and a nearly linear gene-dose effect depending on the number of active CYP2D6 alleles has been shown Individuals carrying this genotype are found in about 3% of the northern European White population, between 5 and 10% in southern European populations 15 and between 10 and 30% in Arabian and northeast African countries. 16 Individuals phenotyped as ultra-rapid CYP2D6 metabolizers have been reported to develop up to 45-fold higher concentrations of codeine O-demethylated metabolites than persons with poor metabolism. 2 However, the impact of this genotype on pharmacokinetic parameters of codeine and opioid side effects has never been systematically studied. One case report describes opioid intoxication in a patient later defined as a carrier of the CYP2D6 gene duplication, but this patient also had an inhibition of CYP3A4 activity and an impaired renal function. 10 A reliable estimation of the effect of the CYP2D6 gene duplication cannot be derived from the existing case reports, which however are important as they draw out attention to a potential medical problem. Therefore, the present study analyzes the pharmacokinetic differences of codeine between a group of UM carrying the CYP2D6 gene duplication and extensive metabolizers (EM) carrying the wild-type allele. A small group of PM serve as an additional reference group. Materials and methods From a large sample of healthy male volunteers, a panel of 26 healthy males of Caucasian ethnicity (between 18 and 65 years, body mass index between 20 and 30 kg m 2,no pathological findings in blood pressure, ECG and clinical chemistry) was selected including 11 carriers of a CYP2D6 duplication of a functional allele combined with another functional allele (the UM group), 12 carriers of two functional CYP2D6 alleles (the EM group) and three carriers of two functionally deficient alleles of CYP2D6 (the PM group). A sample size of at least 10 per group (EM and UM groups) was estimated to be sufficient for discriminating differences of more than 25% in total clearance of codeine between CYP2D6 EM and UM with a power of 80% assuming the variability of codeine pharmacokinetic parameters (AUC, clearance) in EMs as given by Quiding et al. 17 After an overnight fast, all volunteers received a single oral dose of 30 mg codeine (Codeine Phosphoricum, Berlin- Chemie, Berlin, Germany) with tap water and no food except water was allowed until 4 h after administration of the drug. For pharmacokinetic analysis, blood samples with ethylenediamine tetraacetic acid as anticoagulant were taken before dosing and at 0.5, 1, 1.5, 2, 3, 4, 6, 8, 10, 12 and 24 h after the volunteers had taken 30 mg codeine. A single dose of 30 mg codeine was chosen in order to minimize the risk of opioid side effects in individuals who were anticipated to have an ultrarapid activity in bioactivating codeine to morphine and morphine glucuronides. Urine was collected in two fractions between 0 and 6 h and between 6 and 12 h. For at least 3 days before the test, volunteers were not allowed to take alcohol and grapefruit juice and no medication was allowed for at least 3 weeks before the day of the study. The study protocol was approved by the ethics committee of the Humboldt University of Berlin and all volunteers had given their written informed consent before the study. The codeine effect on m-opiod receptors was assessed by measurement of pupil size in a window-less room with controlled low illumination at 0, 1, 2, 3, 4, 6, 8, 12 and 24 h. After 3 min of adaptation (the participants had to stay in that room), the pupil diameter was measured with a pupil size scale allowing discrimination of 1 mm differences. Possible side effects were monitored by a questionnaire involving dizziness, palpitations, nausea and obstipation. Assessments of the adverse effects were performed at 0, 1, 4 and 12 h after dosing. Genotyping Analyses for the CYP2D6 alleles *2, *3,*4,*5,*6, *9, *10, *35, *41 (according to the nomenclature webpage ki.se/cypalleles/) and the duplication alleles were performed with the polymerase chain reaction-restriction fragment length polymorphism-methods described earlier. 18 Identification of allele CYP2D6*41 was based on the 1584 G4C and the 4180 G4C polymorphism as described earlier. 19 In addition, we have analyzed which of the single nucleotide polymorphisms was in the duplication allele and which was in the other not duplicated allele by using a modification of the single-base primer extension method described by Fuselli et al. 20,21 Briefly, the sequence between the exon 9 of the upstream and the intron 2 of the downstream CYP2D6 gene copy of the CYP2D6 duplication allele was amplified using duplication-specific long PCR with the primers 5 0 -GCC ACC ATG GTG TCT TTG CTT TC-3 0 and 5 0 -ACC GGA TTC CAG CTG GGA AAT G and Expand Long Template PCR System (Roche, Mannheim, Germany) as follows: 27 ml reactions containing buffer-1 of the long PCR kit were supplemented with additional MgCl 2

3 Codeine in CYP2D6 gene duplication carriers 259 to a final concentration of 3.5, 0.35 mm dntps, 0.2 mm primers and 1.5 U polymerase mix, and were incubated for 2 min at 941C followed by 35 cycles (10 s at 961C, 20 s at 601C and for 7 min at 681C) and 7 min at 681C. The obtained PCR product was used as a template for the single-base primer extension reaction (SNaPshot, Applied Biosystems) designed to genotype the polymorphic position 4180G4C of the upstream gene-copy and positions 1584G4C and 100C4T of the downstream gene-copy. The following primers were used for the extension reactions: 1584G4C, 5 0 -(A) 9 CCT GGA CAA CTT GG AAG AAC C-3 0 ; 100C4T, 5 0 -(A) 31 CAA CGC TGG GCT GCA CGC TAC-3 0 and 4180G4C, (A) 37 CAA AGC TCA TAG GGG GAT GGG-3 0. The single-base primer extension reactions were performed according to the manufacturer s instructions (SNaPshot, Applied Biosystems, Weiterstadt, Germany). Assuming that always the two gene copies in CYP2D6 duplication alleles have identical constellations, this method allowed an unambiguous differentiation which of the genes was duplicated. A fine classification of CYP2D6 activity owing to genotype was used by counting with a score of 0.5 for those alleles coding for a low-activity enzyme (CYP2D6*9, *10, *41), by counting with a score of 1.0 for alleles CYP2D6*1, *2, and *35 and by counting with the same score for each gene within the CYP2D6 duplication allele. In addition to the CYP2D6 genotypes, the 118 A4G polymorphism (rs ) within the m-opioid receptor gene (OPRM1) leading to an amino-acid exchange of asparagine (Asn) to aspartate (Asp) was determined. The genotyping of the OPRM1 Asp40Asn polymorphism (rs ) was performed with a pre-made TaqMan genotyping assay (ID# C_ _1, Applied Biosystems, Darmstadt, Germany) according to the manufacturer s instructions. The genotyping results were verified by genotyping 20% of the samples twice. No discrepancy between the two genotyping results could be observed. This polymorphism with allele frequencies between 10 and 32% in different ethnic populations 23 has been shown to cause differences in opioid responses such as pupil constriction. 24,25 The latter analysis was performed to test whether there is no major imbalance in the allele distribution between the CYP2D6 genotype groups, which might influence the opioid effect parameters such as pupil diameters and side effects. Concentration analyses Aliquots of human plasma samples were extracted by solid phase extraction and the eluate was analyzed for morphine, morphine-3-glucuronide (M3G), morphine-6-glucuronide (M6G), codeine and codeine-6-glucuronide (C6G). Briefly, 50 ml internal standard (40 ng/ml morphine-d 3, 20 ng/ml M3G-d 3, 70 ng/ml M6G-d 3, 40 ng/ml codeine-d 3 and 50 ng/ml C6G-d 3 in water) and 750 ml water were added to 200 ml plasma. Solid phase extraction was performed on a system consisting of Oasis HLB extraction cartridges (1 ml, 30 mg sorbent, Waters, Eschborn, Germany) attached to a Visiprep vacuum manifold (Supelco, Deisenhofen, Germany). The extraction column was activated with 1 ml methanol and 1 ml water. After the sample was drawn through, the column was washed with 1 ml water. The analytes were eluted from the column with 1 ml methanol. The organic solvent was evaporated to dryness by a gentle stream of nitrogen at 551C. The residue was reconstituted with 200 ml of water/methanol (95:5) with 5 mm ammonium acetate and 0.001% formic acid. Aliquots of human urine samples were diluted 1:100 in water/methanol (95:5) with 5 mm ammonium acetate and 0.001% formic acid, and dilutions were directly analyzed for morphine, M3G, M6G, codeine and C6G. High-performance liquid chromatography analysis was performed using an Agilent 1100 Series binary pump (G1312A) and degasser (G1379A) connected to an HTC PAL autosampler (Chromtech, Idstein, Germany). Chromatographic separations were obtained under gradient conditions using a Synergi Hydro-RP column (150 cm L 2mm ID, 4 mm particle size and 80 Å pore size) with a C18 guard column (4 mm L 2 mm ID) (Phenomenex, Aschaffenburg, Germany). The mobile phase consisted of eluent A (water with 5 mm ammonium acetate and 0.001% formic acid) and eluent B (methanol with 5 mm ammonium acetate and 0.001% formic acid). The gradient was as follows: from t ¼ 0 to t ¼ 0.5 min 95:5 (A:B), from t ¼ 0.5 to t ¼ 7.5 min a linear gradient from 95:5 to 10:90, then from t ¼ 7.5 to t ¼ 10 min 10:90, from t ¼ 10 to t ¼ 11 min a linear gradient from 10:90 to 95:5 and finally from t ¼ 11 to t ¼ 14 min 95:5. Samples (10 ml) were injected onto the LC-MS/MS. The flow rate was set at 0.3 ml/min and the runtime at 14 min. Morphine, M3G, M6G, codeine and C6G eluted after 8.2, 6.3, 6.7, 8.7 and 7.3 min, respectively. MS and MS/MS analyses were performed on a 4000 Q TRAP triple quadrupole mass spectrometer with a TurboIon- Spray source (Applied Biosystems, Darmstadt, Germany). For measurement of the analytes, the positive ion mode was chosen. High-purity nitrogen was used as curtain gas, highpurity air was used as nebulizer and turbo gas was used. The heated turbo gas was set at 5501C. Mass spectrometry parameters were as follows: curtain gas 10 psig, IonSpray voltage 4700 V, nebulizer gas 60 psig and turbo gas 50 psig. Quantitation was performed in the multiple reaction mode (MRM) using nitrogen as the collision gas with a collision energy of 82, 45, 45, 36 and 44 ev for morphine, M3G, M6G, codeine and C6G, respectively. Precursor-to-product ion transitions of m/z for morphine, m/z for M3G and M6G, m/z for codeine and m/z for C6G were used for the MRM with a dwell time of 50 ms. All analytes were quantified by use of calibration curves (r for all analytes) of peak area ratios obtained by analysis with 0.05, 0.1, 0.2, 0.5, 1, 2, 5, 10, 20, 50 and 100 ng/ml for spiked plasma samples or with 5, 10, 20, 50, 100, 200, 500, 1000, 2000, 5000 and ng/ml for spiked diluted urine samples. Quality control samples contained 0.1, 1 and 5 ng/ml (spiked plasma samples) or 10, 100 and 500 ng/ml (spiked diluted urine samples) of each analyte. Samples with concentrations of analytes

4 260 Codeine in CYP2D6 gene duplication carriers above the calibration curves were sufficiently diluted and reassayed. Concentrations of the calibration standards, quality controls and unknowns were evaluated by Analyst software (version 1.4; Applied Biosystems, Darmstadt, Germany). The intra-day and inter-day variability was o15%. Pharmacokinetic analysis Non-compartmental pharmacokinetic parameters were estimated using the WinNonlin version 2.1software. Area under the plasma concentration time curves (AUC 0 infinity ) was determined by the linear trapezoidal rule with extrapolation to infinity (using the predicted concentration at the last measurement) and total clearance was calculated as dose/ AUC 0 infinity. Urinary MRs were calculated as concentration of codeine divided by morphine, and as the sum of concentrations of (codeine þ C6G) divided by the sum of morphine metabolites (morphine þ M3G þ M6G). Maximum concentrations were taken as measured. Statistical analysis For statistical analysis, the Jonckheere Terpstra trend test was applied as implemented in the SPSS version software with the predefined trend of zero, two and three active CYP2D6 genes and for testing the trend of increasing CYP2D6 fine activity. Results The CYP2D6 genotypes of the study participants are listed in Table 1. The volunteers were selected for this study based on differentiation between three types of alleles only, namely alleles conferring absolutely no activity, alleles coding for active enzyme and duplication alleles, but in the analysis given here, the CYP2D6 genotype groups could be further subclassified into fine-activity groups taking into consideration that activity of the protein coded by alleles *9, *10, *17 or *41 is lower than that of the proteins coded by *1, *2 or *35. This fine-activity divided the UM-group into two groups, those with a 2.5-fold CYP2D6 gene activity and those with a threefold activity compared to carriers of only one active allele. Correspondingly, the EM group, initially defined as carriers of two active allele, was subdivided into 1.5 and twofold allele activity after taking the moderate to low activity of the CYP2D6 protein variant alleles into account. In order to avoid confounding of the opioid effects data by the m-opioid receptor genotype, we also genotyped the Table 1 CYP2D6 and l-opioid receptor genotypes of the study participants and phenotypical classifications (phenotype groups UM, EM, PM and fine activity groups) CYP2D6 genotype N CYP2D6-activity group predicted by genotype a CYP2D6 phenotype b OPRM1 Asn40Asp Cl/weight (l h 1 kg 1 ) AUC (morphine) (mghl 1 ) *3/*3 1 0 PM Asn/Asn 1.7 ( ) 0.5 ( ) *4/*4 2 Asn/Asn *1/* EM Asn/Asn 1.7 ( ) 8.4 (5.0 12) *1/*10 1 Asn/Asn *2/*41 1 Asn/Asn *35/*41 1 Asn/Asn *1/*1 4 2 Asn/Asn 2.0 ( ) 12 (6.9 17) *1/*2 3 2 (Asn/Asn), 1 Asn/Asp *1/*35 1 Asn/Asn *1x2/* UM Asn/Asn 1.8 ( ) 16 (11 18) *1x2/*10 1 Asn/Asn *1x2/*41 1 Asn/Asn *2x2/*41 1 Asn/Asn *35x2/*1 1 3 Asn/Asn 2.4 ( ) 16 (10 24) *2x2/*35 2 Asn/Asn *1x2/*35 1 Asn/Asn *2x2/*1 2 Asn/Asn *1x2/*1 1 Asn/Asn Significance c P ¼ 0.01 Po0.001 Abbreviations: AUC, areas under the plasma concentration versus time curves; EM, extensive metabolizers; PM, poor metabolizers; UM, ultra-fast metabolizers. Data are given as median and range. a Predicted CYP2D6 activity: combination of alleles were calculated as follows: inactive alleles: CYP2D6*3, *4, *5, *6; alleles with decreased activity: factor 0.5 (CYP2D6*9, *10, *41); alleles with full CYP2D6 activity: factor 1 (CYP2D6*1, *2, *35). b PM: carriers of two inactive alleles CYP2D6*3, *4, *5, *6; EM: Carriers of any combinations of the active alleles CYP2D6*1, *2, *9, *10, *41 or *35; UM: carriers of one CYP2D6 gene duplication allele (*1 2, *2 2, *35 2) and one active CYP2D6 gene. c Significance was tested with the non-parametric Jonckheere-Terptra Trend test presuming a trend of decreasing AUC or C max with increasing CYP2D6 genotype-based predicted activity.

5 Codeine in CYP2D6 gene duplication carriers 261 participants for the functionally relevant 118 A4G polymorphism of the m-opioid receptor gene, which is found in Caucasians with an allele frequency of 15% 26 (OPRM1 genotype given in Table 1). However, only one subject who had the CYP2D6 genotype *1/*2 was a heterozygous carrier of the Asp amino-acid exchange. This individual was not differed from the group of CYP2D6 EMs in pharmacokinetic and opioid effect parameters, thus the m-opioid receptor genotype could be ignored in the further analyses. Codeine maximal concentrations varied between 24 and 104 mgl 1, with a median of 47 mgl 1 in all study participants and the median clearance of codeine was 161 l h 1 ranging from 75 to 211 l h 1 in the entire group studied here. Elimination half-life ranged between 3.2 and 5.7 h with a median of 3.7 h. The plasma concentrations of the O-demethylated metabolites morphine, M3G and M6G varied substantially between the subjects of this study group and much of this variability was explained by CYP2D6 genotype. Most concentrations in the three PM were around the limit of quantification of 0.05 ng ml 1. Median morphine AUCs were 0.5, 11 and 16 mghl 1 in PM, EM and UMs (P ¼ 0.02 for comparison between the EM and UM groups, U-test), median M3G AUCs were 6.4, 382 and 506 mghl 1 (P ¼ 0.02) and those for M6G were 6.5, 63 and 87 mghl 1 (P ¼ 0.036) (Tables 1 and 2 and Figure 1). Statistical testing was performed comparing only the UM and EM groups by the exact Mann Whitney U-test because the sample size of the study was chosen to test for this difference. When taking the additional reference group of three PMs into account who had no CYP2D6 activity, a highly significant trend towards higher O-demethylated codeine metabolites with increasing CYP2D6 activity was detected (Po0.001 according to the nonparametric Jonckheere Terpstra trend test). Indeed, there was an almost exact linear relationship between the number of active CYP2D6 gene copies and codeine total clearance, justifying using linear regression analysis to describe the relation between CYP2D6 gene copies and metabolic activity. The coefficient of determination (R 2 ), that is, the proportion of variability in morphine AUC explained by CYP2D6 genotype, was 60% when calculating with the three groups (PM, EM and UM) and 63% when calculating with the five groups ( fine-activity scoring ) taking additionally the information about lowactivity CYP2D6 alleles into account. The corresponding R 2 values were 66 and 68% for M3G and 60 and 63% for M6G, respectively. Thus, consideration of the low-activity alleles *9, *10 and *41 explained more of the variability within the EM and UM groups (Table 1). The effect of the CYP2D6 genotype was visible in oral total clearance of codeine only when adjusted for body weight (P ¼ 0.07, R 2 ¼ 0.13%). In Figure 2, the median ratios of AUC(morphine)/ AUC(codeine) are depicted in relation to the CYP2D6 fine activity scores. There was a significant trend towards higher AUC ratio with increasing CYP2D6 activity even if testing was performed without the PM group (P ¼ 0.002, Jonckheere Terpstra trend test). Urine was collected during the first 6 h and from 6 to 12 h after codeine intake. The MRs calculated as codeine/ morphine and (codeine þ C6G)/(morphine þ M3G þ M6G) varied significantly dependent on CYP2D6 activity. In PMs, the ratios were about 10-fold higher compared to EM or UM. A significant trend towards lower ratios with increasing CYP2D6 activity was still detected after exclusion of the PMs from the analysis (Table 3). Table 2 Pharmacokinetic parameters in plasma of codeine and its O-demethylated metabolites Codeine Codeine-6-glucuronide Morphine Morphine-3-glucuronide Morphine-6-glucuronide AUC 0 infinity (mghl 1 ) PM (n ¼ 3) 180 ( ) 4066 ( ) 0.5 ( ) 6.4 (5 18) 6.5 ( ) EM (n ¼ 11) 191 ( ) 3850 ( ) 11 (5 17) 382 ( ) 63 (50 112) UM (n ¼ 12) 192 ( ) 3385 ( ) 16 (10 24) 506 ( ) 87 (66 134) P* ¼ NS NS 0.02* 0.02* 0.036* C max (mgl 1 ) PM 45 (37 56) 628 ( ) 0.05 ( ) 0.7 ( ) 0.8 ( ) EM 51 (24 104) 652 ( ) 2.1 ( ) 39 (32 82) 9.6 (7.2 17) UM 43 (30 70) 672 ( ) 2.6 ( ) 59 (33 103) 13 (8.7 24) P ¼ NS NS NS Elimination half-life (h) PM 4.8 ( ) 4.8 ( ) 17 (15 60) 8.2 (7.6 13) 6.2 (6.2 14) EM 3.6 ( ) 3.5 ( ) 13 (7.7 30) 9.3 (7.0 17) 7.2 ( ) UM 3.7 ( ) 3.4 ( ) 14 (6.3 27) 10 (6.3 14) 7.1 (5.7 14) P ¼ NS NS NS NS NS Abbreviations: EM, extensive metabolizers; PM, poor metabolizers; UM, ultra-fast metabolizers. Data are given as median and range *Significance given for two-group difference between the EM and the UM group (non-parametric U-test); the non-parametric Jonckheere Terpstra trend test involving also the PM group resulted in much lower P-values (P ¼ for morphine 6 glucuronide and Po0.001 for morphine, and morphine 3 glucuronide).

6 262 Codeine in CYP2D6 gene duplication carriers Concentration [µg/l] Concentration [µg/l] Concentration [µg/l] Concentration [µg/l] Codeine Morphine Morphine-3-glucuronide Morphine-6-glucuronide Time [h] EM UM PM Figure 1 Concentration time curves (mean and standard deviation) of codeine and the O-demethylated codeine metabolites morphine, M3G and M6G. Red lines correspond to the mean plasma concentrations of the UM group, blue lines to the means of the EM group and green lines to the means of the PM group. With regard to the pupil diameter measured as pharmacodynamic parameter, the influence of CYP2D6 genotype was weak. Indeed, all participants had a miosis with a maximum after 2.5 h from codeine intake and the three PMs even had the strongest decline in pupil diameter, a finding AUC(morphine)/AUC(Codeine) that was however not significant and the statistical power of the present study was far too small with respect to this pharmacodynamic parameter. The low dose of 30 mg codeine was taken for safety reasons but it was too low to exert significant analgetic effects and thus analgesia was not evaluated. Twenty of the 26 participants complained at least of one adverse effect. This was mostly owing to sedation, which was reported by 18 of the 26 study participants during the first 4 h after codeine intake. Ten of the 11 CYP2D6 UMs felt sedation (91%) compared to six (50%) of the 12 EMs (P ¼ 0.069, exact Fisher s test). When combining all adverse effects, all of the UMs, five (42%) of the 12 EMs and two (66%) of the PMs had at least one adverse effect (P ¼ 0.041, exact Fisher s test for the comparison between the EM and PM groups). Discussion PM CYP2D6 fine activity Figure 2 Ratio of plasma AUC of morphine over plasma AUC or codeine in relation to the CYP2D6 activity expressed by the number of active alleles differentiating between fully active alleles, which were considered with one arbitrary activity unit, and alleles with reduced activity, with were arbitrarily considered, which 0.5 activity units. O-demethylation of codeine is mediated by the polymorphic enzyme CYP2D6 and individuals who very rapidly bioactivate codeine to morphine might be at higher risk for opioid intoxication or developing codeine addiction. In this study, we focused on differences in codeine metabolite pharmacokinetics between extensive and ultrarapid metabolizers according to CYP2D6 genotype. In addition, we also included three PM because these individuals are known not to O-demethylate codeine to morphine and we wanted to confirm that our study data would fit into the prior knowledge, which was indeed true; and similarly as described earlier, the morphine AUCs in our study varied EM UM

7 Codeine in CYP2D6 gene duplication carriers 263 Table 3 Metabolic ratios in urine of codeine and C6G and the demethylated metabolites from the 0 6 h and the 6 12 h interval of urine collection Fine activity Codeine/morphine 0 6 h (codeine + C6G)/ (morphine+m3g+m6g) 0 6 h Codeine/morphine 6 12 h (codeine+c6g)/ (morphine+m3g+m6g) 6 12 h PM 0 (n ¼ 3) 198 ( ) 363 ( ) 68 (46 278) 255 ( ) EM 1.5 (n ¼ 3) 28 (16 31) 15 (12 17) 18 (15 20) 13 (12 15) 2(n¼6) 19 (9.9 28) 10 (5.9 16) 9.4 (5.6 19) 8.4 (5.6 12) UM 2.5 (n ¼ 4) 17 (13 19) 10 (6.2 11) 14 (5.6 16) 8.6 ( ) 3(n¼7) 18 (6.3 43) 8.9 (6.2 16) 12 (4.5 30) 7.4 ( ) P* NS 0.01 NS Abbreviations: EM, extensive metabolizers; PM, poor metabolizers; UM, ultra-fast metabolizers. Data are given as median and range. Significance was tested with the non-parametric Jonckheere Terpstra Trend test with exclusion of the PM group. Test was performed only for differences between EM and UM. more than 30-fold between the PM and the UM groups (Table 1). 2 7 We observed a strong correlation between the number of active CYP2D6 genes and plasma concentrations as well as urinary recovery ratios of all O-demethylated codeine metabolites. The plasma concentrations and AUCs of morphine between UMs and EMs differed about 1.5-fold with a nearly exact linear gene-dose effect (Table 2) and, if the corresponding drug concentration analytics is available, one might even suggest that codeine in the safe low doses of 30 mg or below might be a suitable CYP2D6 in vivo phenotyping probe drug having the advantage that the diagnostic metabolite appears to be formed almost exclusively by CYP2D6. In Figure 3, the correlation between the plasma morphine/codeine ratio and the oral total clearance of metoprolol, a typical phenotyping substance for CYP2D6 activity, is depicted. The data for metoprolol were taken from Kirchheiner et al. 13 In fact, the participants of the present study were also UMs according to metoprolol CYP2D6 metabolic phenotype but there is no strict antimode between EM and UM, that is, there is an overlap of the ranges of phenotypes and there is apparently some scatter, probably partially due to interindividual variation in daily CYP2D6 activity and partially due to specific factors (other enzyme, transporters) specifically relevant for one or the other probe drug. Detailed genotyping for CYP2D6 alleles resulted in a more precise classification of the metabolizer phenotypes by attributing activity scores to the single alleles as follows: inactive alleles (CYP2D6*3, *4, *5, *6) score 0; variant alleles with low activity (CYP2D6*9, *10, *17, *41) score 0.5; alleles with full CYP2D6 activity (CYP2D6*1, *2, *35) score 1 (Table 1). The correlation of genotype with plasma concentrations of morphine and its metabolites was even more distinct when applying this fine classification (Figures 2 and 3) and according to the coefficient of determination (R 2 ) about 63% of interindividual variability in morphine AUC could be explained by CYP2D6 genotype. Of course, giving the alleles 9, 10 and 41 a score of 0.5 is arbitrary and the reduction in catalytic activity conferred by the amino-acid substitution polymorphisms may even differ between drugs. However, if we really want to use genotype information as a Plasma AUC morphine / codeine ratio Total clearance of CYP2D6 probe drug metoprolol [L. h -1 ] Figure 3 Correlation between the plasma morphine/codeine ratio and the oral total clearance of metoprolol, a phenotyping substance for CYP2D6 activity. Green dots depict the PMs, blue dots depict the EMs and red dots depict the UMs as defined in Table 1. The data for metoprolol were taken from Kirchheiner et al. 13 In fact, the participants of the present study were also UMs according to metoprolol CYP2D6 metabolic phenotype but there is no strict antimode between EM and UM, that is, there is an overlap of the ranges of phenotypes and there is apparently some scatter, probably partially owing to interindividual variation in daily CYP2D6 activity and partially owing to specific factors (other enzymes, transporters) specifically relevant for one or the other probe drug. guidance to drug therapy, we do need concrete and practical algorithms as to how to deal with genotype information and apparently such a fine-scoring system for CYP2D6 genotypepredicted activity was not only introduced by ourselves 11,13 but independently also by other groups. 27,28 The genotype-caused differences in metabolism were also reflected in MRs as determined in urine. During the 6 and 12 h interval of urine collection, UMs and EMs differed significantly in MR of (codeine þ C6G) divided by the sum

8 264 Codeine in CYP2D6 gene duplication carriers of morphine and its glucuronides fraction (morphine þ M3G þ M6G) as given in Table 3. The 0 to 6 h urine collection interval had even the higher power to differentiate between the CYP2D6 genotypes than the 6 12 h interval. The PMs, as they have very low formation of morphine and metabolites, had more than 20-fold higher ratios. Apparently, urine MRs may be used as phenotypic indicators and have the advantage that per person only one 0 6 h urine sample might have to be analyzed if one would use low-dose codeine for CYP2D6 phenotyping. The clinically important point is to what extent the enhanced codeine O-demethylation capacity observed in the ultra-um influences the opioidergic side effects. We measured sedation and miosis as opioid effect-related parameters. All volunteers had a slight decrease in pupil diameter (mean 1 mm70.5) but no CYP2D6 genotyperelated effects on miosis have been detected. Interestingly, the small PM group had a even longer period of miosis, an observation that was however statistically not significant owing to the small sample size. This finding is similar to that for dihydrocodeine, which also leads to miosis in individuals (PMs) not metabolizing dihydrocodeine to dihydromorphine, 29 and also fits to recent data where an additional effect of codeine itself on miosis was observed when comparing equal concentrations of morphine and its metabolites with or without codeine. 30 These observations suggest that single dose of codeine has an opiate receptormediated effect of its own despite its 7 14 times lower potency than morphine. 31 In addition, other non-odemethylated metabolites might exert a certain opioid effect such as norcodeine, which is recovered in urine to 4%. 32 However, similar to codeine, norcodeine did not bind to m receptors in rhesus monkey brain membranes, indicating a low m-opiate receptor affinity. 33 Alternatively, C6G has been proposed to exert analgesic effects because of its high amounts. 34 However, the K i value of C6G for the m-opiate receptor is about twice that of codeine, 35 and glucuronides are less likely to enter the brain owing to their hydrophilic properties. In addition, as no randomized placebo-controlled design was used in this study, minor placebo effects on adverse event monitoring or pupil size measurements cannot entirely be ruled out. However, this is the same in typical phase I trials in drug development, which are also usually not blinded but nevertheless considered to provide reliable data on tolerability of a drug. Sedation belongs to the central nervous opioid effects and in our study, the UM felt more often sedated than the EM. A study on analgesic effects in nine PMs and nine EMs reported analgesic effects only in EMs but the same incidence of adverse drug effects in both groups. 36 Opioid effects also depend on biotransformation within the brain, and, thus, concentrations of codeine, norcodeine and C6G in the brain might differ from those observed in plasma. Recently, a functional variant of CYP2D7 leading to expression of this pseudogene in brain and to brain metabolism of codeine to morphine has been described in four of eight individuals of whom brain autopsies were obtained. 37 The population frequency of this variant has not yet been studied so far but this mechanism as well might be a possible explanation for opioid effects in PMs after codeine intake. In conclusion, ultrarapid codeine metabolism caused by a CYP2D6 gene duplication resulted in a 1.5-fold higher morphine exposure compared to EM; in other words, for an UM (about 3% in many Caucasian populations) a dose of 30 mg codeine would make the same effects as a dose of 45 mg in an EM (about 50% in many Caucasian populations). This difference is only moderate but the risk for opioid intoxication might be increased in UMs if other additional factors such as reduction in renal function or further inhibition of other enzyme systems occur and the opioid intoxication risk owing to genotype might be substantial if physicians are dealing with carriers of CYP2D6 gene multiplications, which are however so rare that we did not identify any carrier or such a genotype in a screening of about 1000 healthy subjects. Acknowledgments This study was supported by a grant from the German Ministry of Education and Research (GG 9845/5, Pharmacogenetic diagnostics). Duality of interest Dr Brockmöller has received research grants and lecture honoraria by Roche, manufacturer of the Amplichip P450, a tool for CYP2D6 genotyping. Dr Julia Kirchheiner has no duality of interest. References 1 Dayer P, Desmeules J, Leemann T, Striberni R. Bioactivation of the narcotic drug codeine in human liver is mediated by the polymorphic monooxygenase catalyzing debrisoquine 4-hydroxylation (cytochrome P-450 dbl/bufi). Biochem Biophys Res Commun 1988; 152: Yue QY, Alm C, Svensson JO, Sawe J. Quantification of the O- and N-demethylated and the glucuronidated metabolites of codeine relative to the debrisoquine metabolic ratio in urine in ultrarapid, rapid, and poor debrisoquine hydroxylators. Ther Drug Monit 1997; 19: Johansson I, Yue QY, Dahl ML, Heim M, Sawe J, Bertilsson L et al. Genetic analysis of the interethnic difference between Chinese and Caucasians in the polymorphic metabolism of debrisoquine and codeine. Eur J Clin Pharmacol 1991; 40: Mortimer O, Persson K, Ladona MG, Spalding D, Zanger UM, Meyer UA et al. Polymorphic formation of morphine from codeine in poor and extensive metabolizers of dextromethorphan: relationship to the presence of immunoidentified cytochrome P-450IID1. Clin Pharmacol Ther 1990; 47: Mikus G, Bochner F, Eichelbaum M, Horak P, Somogyi AA, Spector S. Endogenous codeine and morphine in poor and extensive metabolisers of the CYP2D6 (debrisoquine/sparteine) polymorphism. J Pharmacol Exp Ther 1994; 268: Persson K, Sjostrom S, Sigurdardottir I, Molnar V, Hammarlund-Udenaes M, Rane A. Patient-controlled analgesia (PCA) with codeine for postoperative pain relief in ten extensive metabolisers and one poor metaboliser of dextromethorphan. Br J Clin Pharmacol 1995; 39: Sindrup SH, Poulsen L, Brosen K, Arendt-Nielsen L, Gram LF. Are poor metabolisers of sparteine/debrisoquine less pain tolerant than extensive metabolisers? Pain 1993; 53: Tyndale RF, Droll KP, Sellers EM. Genetically deficient CYP2D6 metabolism provides protection against oral opiate dependence. Pharmacogenetics 1997; 7:

9 Codeine in CYP2D6 gene duplication carriers Dalen P, Frengell C, Dahl ML, Sjoqvist F. Quick onset of severe abdominal pain after codeine in an ultrarapid metabolizer of debrisoquine. Ther Drug Monit 1997; 19: Gasche Y, Daali Y, Fathi M, Chiappe A, Cottini S, Dayer P et al. Codeine intoxication associated with ultrarapid CYP2D6 metabolism. NEnglJ Med 2004; 351: Kirchheiner J, Henckel HB, Franke L, Meineke I, Tzvetkov M, Uebelhack R et al. Impact of the CYP2D6 ultra-rapid metabolizer genotype on doxepin pharmacokinetics and serotonin in platelets. Pharmacogenet Genomics 2005; 15: Kirchheiner J, Henckel HB, Meineke I, Roots I, Brockmoller J. Impact of the CYP2D6 ultrarapid metabolizer genotype on mirtazapine pharmacokinetics and adverse events in healthy volunteers. J Clin Psychopharmacol 2004; 24: Kirchheiner J, Heesch C, Bauer S, Meisel C, Seringer A, Goldammer M et al. Impact of the ultrarapid metabolizer genotype of cytochrome P450 2D6 on metoprolol pharmacokinetics and pharmacodynamics. Clin Pharmacol Ther 2004; 76: Lundqvist E, Johansson I, Ingelman-Sundberg M. Genetic mechanisms for duplication and multiduplication of the human CYP2D6 gene and methods for detection of duplicated CYP2D6 genes. Gene 1999; 226: Aynacioglu AS, Sachse C, Bozkurt A, Kortunay S, Nacak M, Schroder T et al. Low frequency of defective alleles of cytochrome P450 enzymes 2C19 and 2D6 in the Turkish population. Clin Pharmacol Ther 1999; 66: McLellan RA, Oscarson M, Seidegard J, Evans DA, Ingelman-Sundberg M. Frequent occurrence of CYP2D6 gene duplication in Saudi Arabians. Pharmacogenetics 1997; 7: Quiding H, Anderson P, Bondesson U, Boreus LO, Hynning PA. Plasma concentrations of codeine and its metabolite, morphine, after single and repeated oral administration. Eur J Clin Pharmacol 1986; 30: Sachse C, Brockmöller J, Bauer S, Roots I. Cytochrome P450 2D6 variants in a Caucasian population: allele frequencies and phenotypic consequences. Am J Hum Genet 1997; 60: Raimundo S, Fischer J, Eichelbaum M, Griese EU, Schwab M, Zanger UM. Elucidation of the genetic basis of the common intermediate metabolizer phenotype for drug oxidation by CYP2D6. Pharmacogenetics 2000; 10: Fuselli S, Dupanloup I, Frigato E, Cruciani F, Scozzari R, Moral P et al. Molecular diversity at the CYP2D6 locus in the Mediterranean region. Eur J Hum Genet 2004; 12: Sistonen J, Fuselli S, Levo A, Sajantila A. CYP2D6 Genotyping by a multiplex primer extension reaction. Clin Chem 2005; 51: Johansson I, Lundqvist E, Dahl ML, Ingelman-Sundberg M. PCR-based genotyping for duplicated and deleted CYP2D6 genes. Pharmacogenetics 1996; 6: LaForge KS, Yuferov V, Kreek MJ. Opioid receptor and peptide gene polymorphisms: potential implications for addictions. Eur J Pharmacol 2000; 410: Lotsch J, Skarke C, Grosch S, Darimont J, Schmidt H, Geisslinger G. The polymorphism A118G of the human mu-opioid receptor gene decreases the pupil constrictory effect of morphine-6-glucuronide but not that of morphine. Pharmacogenetics 2002; 12: Bond C, LaForge KS, Tian M, Melia D, Zhang S, Borg L et al. Singlenucleotide polymorphism in the human mu opioid receptor gene alters beta-endorphin binding and activity: possible implications for opiate addiction. Proc Natl Acad Sci USA 1998; 95: Crowley JJ, Oslin DW, Patkar AA, Gottheil E, DeMaria Jr PA, O Brien CP et al. A genetic association study of the mu opioid receptor and severe opioid dependence. Psychiatr Genet 2003; 13: Steimer W, Zopf K, von Amelunxen S, Pfeiffer H, Bachofer J, Popp J et al. Allele-specific change of concentration and functional gene dose for the prediction of steady-state serum concentrations of amitriptyline and nortriptyline in CYP2C19 and CYP2D6 extensive and intermediate metabolizers. Clin Chem 2004; 50: Zineh I, Beitelshees AL, Gaedigk A, Walker JR, Pauly DF, Eberst K et al. Pharmacokinetics and CYP2D6 genotypes do not predict metoprolol adverse events or efficacy in hypertension. Clin Pharmacol Ther 2004; 76: Schmidt H, Vormfelde SV, Walchner-Bonjean M, Klinder K, Freudenthaler S, Gleiter CH et al. The role of active metabolites in dihydrocodeine effects. Int J Clin Pharmacol Ther 2003; 41: Lotsch J, Skarke C, Schmidt H, Rohrbacher M, Hofmann U, Schwab M et al. Evidence for morphine-independent central nervous opioid effects after administration of codeine: contribution of other codeine metabolites. Clin Pharmacol Ther 2006; 79: Kay DC, Gorodetzky CW, Martin WR. Comparative effects of codeine and morphine in man. J Pharmacol Exp Ther 1967; 156: Hedenmalm K, Sundgren M, Granberg K, Spigset O, Dahlqvist R. Urinary excretion of codeine, ethylmorphine, and their metabolites: relation to the CYP2D6 activity. Ther Drug Monit 1997; 19: Bertalmio AJ, Medzihradsky F, Winger G, Woods JH. Differential influence of N-dealkylation on the stimulus properties of some opioid agonists. J Pharmacol Exp Ther 1992; 261: Vree TB, van Dongen RT, Koopman-Kimenai PM. Codeine analgesia is due to codeine-6-glucuronide, not morphine. Int J Clin Pract 2000; 54: Mignat C, Wille U, Ziegler A. Affinity profiles of morphine, codeine, dihydrocodeine and their glucuronides at opioid receptor subtypes. Life Sci 1995; 56: Eckhardt K, Li S, Ammon S, Schanzle G, Mikus G, Eichelbaum M. Same incidence of adverse drug events after codeine administration irrespective of the genetically determined differences in morphine formation. Pain 1998; 76: Pai HV, Kommaddi RP, Chinta SJ, Mori T, Boyd MR, Ravindranath V. A frameshift mutation and alternate splicing in human brain generate a functional form of the pseudogene cytochrome P4502D7 that demethylates codeine to morphine. J Biol Chem 2004; 279: