7
human psychopharmacology Hum Psychopharmacol Clin Exp 2004; 19: 17–23. Published online in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/hup.539 No evidence of increased adverse drug reactions in cytochrome P450 CYP2D6 poor metabolizers treated with fluoxetine or nortriptyline Rebecca L. Roberts 1 , Roger T. Mulder 2 , Peter R. Joyce 2 , Suzanne E. Luty 2 and Martin A. Kennedy 1 * 1 Department of Pathology, Christchurch School of Medicine & Health Sciences, University of Otago, P.O. Box 4345, Christchurch, New Zealand 2 Department of Psychological Medicine, Christchurch School of Medicine & Health Sciences, University of Otago, P.O. Box 4345, Christchurch, New Zealand The polymorphic enzyme cytochrome P450 CYP2D6 is involved in the metabolism of many antidepressants, including nor- triptyline and fluoxetine. Some 7%–10% Caucasians have inactivating mutations in both alleles of the CYP2D6 gene, and are referred to as poor metabolizers (PMs). Several case reports and clinical studies suggest that CYP2D6 PMs are at a greater risk of developing adverse drug reactions (ADRs) on antidepressant medication than extensive metabolizers (EMs). However, few clinical trials have investigated whether CYP2D6 PM genotype is predictive of ADRs during antide- pressant treatment. This paper explores the link between CYP2D6 genotype and antidepressant-associated ADRs in outpa- tients being treated for major depression with either nortriptyline or fluoxetine. Patients were randomized to fluoxetine (n ¼ 65) or nortriptyline (n ¼ 60) for the 6 week trial. CYP2D6 genotypes predicted that of these patients 115 were EM and the remaining 10 were PMs. ADRs attributed to antidepressant usage were recorded over the 6-week trial. Although the type of ADR was, as expected, different between drugs, the frequency of ADRs experienced did not differ significantly between the two antidepressants or between CYP2D6 PMs and EMs. In addition, the frequency at which PMs discontinued antidepressant medication was not noticeably different from EMs, although with only 10 PMs the study is under powered to detect moderate or small differences. These findings suggest that inability to efficiently metabolize antidepressants that are CYP2D6 substrates does not necessarily lead to increased occurrence of antidepressant-associated ADRs. Thus, for clini- cians prescribing antidepressant monotherapy, CYP2D6 polymorphisms are probably not of relevance to antidepressant side effects and therapy. Copyright # 2004 John Wiley & Sons, Ltd. key words — adverse drug reaction; fluoxetine; nortriptyline; CYP2D6 genotype; major depression; antidepressants INTRODUCTION Polymorphisms within the liver enzyme cytochrome P450 2D6 (CYP2D6, debrisoquine/sparteine hydroxy- lase) lead to distinct subgroups of the population that differ in their ability to metabolize many clinically significant psychoactive drugs (Meyer, 1994). Poor metabolizers (PM) completely lack enzyme activity because of inactivating mutations in both alleles, whereas individuals with at least one functional CYP2D6 gene exhibit an extensive metabolizer (EM) phenotype and show a broad spectrum of enzyme activity from relatively low to ultra-high (Meyer, 1994). Approximately 5% to 10% of Cauca- sians are PMs. These individuals may have a heigh- tened risk of concentration-dependent adverse drug reactions (ADRs) at standard doses of drugs that are CYP2D6 substrates (Bertilsson et al., 1981; Nordin and Bertilsson, 1995; Bertilsson et al., 1997; Vandel et al., 1999; Bertilsson et al., 2002). Most antidepressants are metabolized and inacti- vated as a result of hydroxylation mediated by CYP2D6 (Eichelbaum and Gross, 1990; Kroemer Received 23 May 2003 Copyright # 2004 John Wiley & Sons, Ltd. Accepted 21 August 2003 *Correspondence to: Dr M. Kennedy, Department of Pathology, Christchurch School of Medicine and Health Sciences, PO Box 4345, Christchurch, New Zealand. Tel: (64-3) 364 1222. Fax: (64-3) 364 0009. E-mail: [email protected] Contract/grant sponsors: Health Research Council of New Zealand; University of Otago.

No evidence of increased adverse drug reactions in cytochrome P450 CYP2D6 poor metabolizers treated with fluoxetine or nortriptyline

Embed Size (px)

Citation preview

Page 1: No evidence of increased adverse drug reactions in cytochrome P450 CYP2D6 poor metabolizers treated with fluoxetine or nortriptyline

human psychopharmacology

Hum Psychopharmacol Clin Exp 2004; 19: 17–23.

Published online in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/hup.539

No evidence of increased adverse drug reactions incytochrome P450 CYP2D6 poor metabolizers treatedwith fluoxetine or nortriptyline

Rebecca L. Roberts1, Roger T. Mulder2, Peter R. Joyce2, Suzanne E. Luty2 and Martin A. Kennedy1*

1Department of Pathology, Christchurch School of Medicine & Health Sciences, University of Otago,P.O. Box 4345, Christchurch, New Zealand2Department of Psychological Medicine, Christchurch School of Medicine & Health Sciences, University of Otago,P.O. Box 4345, Christchurch, New Zealand

The polymorphic enzyme cytochrome P450 CYP2D6 is involved in the metabolism of many antidepressants, including nor-triptyline and fluoxetine. Some 7%–10% Caucasians have inactivating mutations in both alleles of the CYP2D6 gene, andare referred to as poor metabolizers (PMs). Several case reports and clinical studies suggest that CYP2D6 PMs are at agreater risk of developing adverse drug reactions (ADRs) on antidepressant medication than extensive metabolizers(EMs). However, few clinical trials have investigated whether CYP2D6 PM genotype is predictive of ADRs during antide-pressant treatment. This paper explores the link between CYP2D6 genotype and antidepressant-associated ADRs in outpa-tients being treated for major depression with either nortriptyline or fluoxetine. Patients were randomized to fluoxetine(n¼ 65) or nortriptyline (n¼ 60) for the 6 week trial. CYP2D6 genotypes predicted that of these patients 115 were EMand the remaining 10 were PMs. ADRs attributed to antidepressant usage were recorded over the 6-week trial. Althoughthe type of ADR was, as expected, different between drugs, the frequency of ADRs experienced did not differ significantlybetween the two antidepressants or between CYP2D6 PMs and EMs. In addition, the frequency at which PMs discontinuedantidepressant medication was not noticeably different from EMs, although with only 10 PMs the study is under powered todetect moderate or small differences. These findings suggest that inability to efficiently metabolize antidepressants that areCYP2D6 substrates does not necessarily lead to increased occurrence of antidepressant-associated ADRs. Thus, for clini-cians prescribing antidepressant monotherapy, CYP2D6 polymorphisms are probably not of relevance to antidepressant sideeffects and therapy. Copyright # 2004 John Wiley & Sons, Ltd.

key words— adverse drug reaction; fluoxetine; nortriptyline; CYP2D6 genotype; major depression; antidepressants

INTRODUCTION

Polymorphisms within the liver enzyme cytochromeP450 2D6 (CYP2D6, debrisoquine/sparteine hydroxy-lase) lead to distinct subgroups of the population thatdiffer in their ability to metabolize many clinicallysignificant psychoactive drugs (Meyer, 1994). Poormetabolizers (PM) completely lack enzyme activity

because of inactivating mutations in both alleles,whereas individuals with at least one functionalCYP2D6 gene exhibit an extensive metabolizer(EM) phenotype and show a broad spectrum ofenzyme activity from relatively low to ultra-high(Meyer, 1994). Approximately 5% to 10% of Cauca-sians are PMs. These individuals may have a heigh-tened risk of concentration-dependent adverse drugreactions (ADRs) at standard doses of drugs that areCYP2D6 substrates (Bertilsson et al., 1981; Nordinand Bertilsson, 1995; Bertilsson et al., 1997; Vandelet al., 1999; Bertilsson et al., 2002).

Most antidepressants are metabolized and inacti-vated as a result of hydroxylation mediated byCYP2D6 (Eichelbaum and Gross, 1990; Kroemer

Received 23 May 2003

Copyright # 2004 John Wiley & Sons, Ltd. Accepted 21 August 2003

* Correspondence to: Dr M. Kennedy, Department of Pathology,Christchurch School of Medicine and Health Sciences, PO Box4345, Christchurch, New Zealand. Tel: (64-3) 364 1222. Fax: (64-3)364 0009. E-mail: [email protected]/grant sponsors: Health Research Council of New Zealand;University of Otago.

Page 2: No evidence of increased adverse drug reactions in cytochrome P450 CYP2D6 poor metabolizers treated with fluoxetine or nortriptyline

and Eichelbaum, 1995). This includes many tricyclicantidepressants (TCAs) such as nortriptyline, monoa-mine oxidase inhibitors, and some selective serotoninreuptake inhibitors (SSRIs), including fluoxetine andparoxetine (Mellstrom et al., 1981; Venkatakrishnanet al., 1999; Hamelin et al., 1996; Fjordside et al.,1999; Morita et al., 2000).

Several case reports suggest that individuals with aCYP2D6 PM genotype may be at greater risk of severeADRs on antidepressants than CYP2D6 extensivemetabolizers (EMs). For example, an elderly womanon a low daily dose of nortriptyline was hospitalizedwith fatigue, vertigo and confusion 6 days aftercommencing antidepressant treatment. Nortriptylinelevels in the patient’s plasma far exceeded the thera-peutic range. Retrospective genotyping found thewoman was a CYP2D6 PM (Bertilsson et al., 1981).In two separate cases a 7-year-old child and a 21-year-old woman died while receiving the TCA imipramine(Swanson et al., 1997). Although neither individualhad ingested an acute overdose, a lethal concentrationof drug was found in their plasma, suggesting bothwere CYP2D6 PMs. Similarly, a 9-year-old with aCYP2D6 PM genotype died after being prescribedfluoxetine over a 10-month period (Sallee et al.,2000). At autopsy, blood, brain and other tissues con-tained toxic levels of the parent drug and the metabo-lite norfluoxetine. A range of other reports exist,suggesting that PM status can be associated withsevere or fatal ADRs (Bertilsson et al., 2002; March-ioni et al., 1996; Lessard et al., 1999; Perucca et al.,1997), although the frequency of such events isunclear.

Despite such case reports documenting severe anti-depressant-associated ADRs in CYP2D6 PMs, thereare few clinical studies examining the more generalrelationship between the CYP2D6 genotype andADRs (de Leon et al., 1998; Vandel et al., 1999; Chouet al., 2000; Chen et al., 1996). This study examinedwhether the CYP2D6 PM genotype increased suscept-ibility to ADRs with the antidepressants nortriptylineand fluoxetine, in a group of outpatients being treatedfor major depression.

MATERIALS AND METHODS

Subjects

Patients (age 18–64 years) were included for whom amajor depressive episode was the current principaldiagnosis, and for whom the treating psychiatrist con-sidered that treatment with antidepressant medicationwas indicated. Patients were excluded if current

moderate to severe alcohol or drug dependence wasdeemed to be the principal current diagnosis (althoughmany patients with alcohol or drug dependence wereincluded if the treating clinician considered that majordepression was the principal current diagnosis).Patients were also excluded if they had a history ofmania (but not hypomania) or schizophrenia. Patientswere required to be free of any major physical illnessand to be free of prescribed psychotropic medicationfor a minimum of 2 weeks, except for an occasionalhypnotic for sleep. After giving consent, the patientsattended for a detailed clinical and neurobiologicalassessment including the structured clinical interviewfor DSM (SCID) (Spitzer et al., 1992) and for clinicalratings on the Hamilton (HDRS) (Hamilton, 1960)(mean 17 item score 19.9� 4.4) and Montgomeryand Asberg (MADRS) (Montgomery and Asberg,1979) depression rating scales (mean MADRS score31.0� 6.6).

During assessment, 10 ml of peripheral blood wasdrawn from each subject for DNA analysis. GenomicDNA was extracted from peripheral blood asdescribed previously (Ciulla et al., 1988; Robertset al., 2002). Following assessment, depressedpatients were randomized to treatment with eitherfluoxetine or nortriptyline. More details on the clinicalassessment and treatment of these patients areincluded in other papers arising from this clinical trial(Carter et al., 2000; Joyce et al., 2002; Luty et al.,2002). All patients gave informed consent, and thestudy was approved by the Canterbury Ethics Com-mittee, Christchurch, New Zealand. Of the patientsrecruited for this study 96% were New Zealand Cau-casians.

Following randomization the patients were seenregularly, initially at least weekly, to optimize poten-tial benefits of the antidepressant medication, includ-ing adjustments of antidepressant dosage. If thepatients were randomized to fluoxetine, they initiallyreceived 20 mg per day for a period of 3 weeks. Afterthis, the clinician could vary the dosage, based uponclinical response and side effects. At 6 weeks, themean fluoxetine dose was 28 mg, with a range of10–80 mg, and 20 mg was the most common dosage.If the patients were randomized to nortriptyline, theyreceived 25 mg, 50 mg, and then 75 mg over the firstthree nights, and then the clinicians were free to adjustthe dosage based upon clinical response, side effectsand drug levels. At 6 weeks the mean nortriptylinedose was 90 mg, with a range of 50–175 mg.

At 3 weeks and 6 weeks, the patients were reas-sessed for depression severity by repeating the HDRS,MADRS and a global clinical rating. At these 3 and

18 r. l. roberts ET AL.

Copyright # 2004 John Wiley & Sons, Ltd. Hum Psychopharmacol Clin Exp 2004; 19: 17–23.

Page 3: No evidence of increased adverse drug reactions in cytochrome P450 CYP2D6 poor metabolizers treated with fluoxetine or nortriptyline

6 week assessments, all the patients were specificallyasked about ADRs that may have been associated withthe antidepressant medication, although a list of spe-cific side effects was not systematically completed.Any significant ADR reported was then rated as mild,moderate or severe. For the purposes of this paper, allADRs reported at either visit were combined. Patientswere able to report as many side effects as they wereexperiencing, and each was rated as mild, moderate orsevere. The most severe ADRs contributed to theseverity rating of any ADRs, and because few wererated as severe, these were combined with ADRs ratedas moderate.

CYP2D6 genotyping

The CYP2D6 gene is located adjacent to two highlyhomologous pseudogenes (CYP2D8 and CYP2D7).Of the mutations so far found in CYP2D6 several havealso been detected in CYP2D8 and CYP2D7 (Heimand Meyer, 1990). To avoid false-positive detectionof mutations through co-amplification of pseudogenesequences, initial amplification of the entire CYP2D6gene was carried out using primers complementary tounique intronic sequences of CYP2D6. The method ofSachse et al. (1997) was used with the following mod-ifications. Reaction volumes were 25 ml, and con-tained 2 mM MgCl2, 1.25 mM of each primer, 2 unitsof ELONGASE Enzyme Mix (Gibco BRL, Rockville,MD, USA) and �100 ng of genomic DNA. Thermalcycling was performed with an initial denaturation of30 s at 94�C, followed by 35 cycles of 30 s at 94�C,6 min at 68�C, and a terminal extension of 7 min at68�C.

The CYP2D6-specific PCR product from each sub-ject was tested for the CYP2D6 alleles *1 to *16, andCYP2D6*19 and CYP2D6*20. Alleles defined by apoint mutation were identified in parallel by nestedPCR-RFLP analyses (Sachse et al., 1997) and a multi-plex PCR (Roberts et al., 2000). Allelic variantsinvolving duplications (CYP2D6*M�N), dele-tion (CYP2D6*5) or rearrangements (CYP2D6*13and *16) were detected using Southern blots(Chomczynski and Qasba, 1984; Roberts et al.,2000). The 1.6 kb probe used to screen Southern blotswas generated by PCR using primers P61 (50

CCCGTTCTGTCCCGAGTATG 30) and P92 (50

CTCAGCCTCAACGTACCCCT 30). PCRs were25 ml, and contained 2 mM MgCl2, 0.625 mM of eachprimer (P61 and P92), 200 mM dNTPs, 1 unit of TaqDNA polymerase (Roche Biochemicals), and approxi-mately 100 ng of genomic DNA. Thermal cycling was

performed with an initial denaturation of 30 s at 94�C,followed by 30 cycles of 30 s at 94�C, 30 s at 58�C,2 min at 72�C, and a terminal extension of 4 min at72�C. Approximately 25 ng of purified product waslabelled with [�-32P]dCTP and used as a probe. Thisprobe contained sequences derived from CYP2D6 andthe pseudogenes CYP2D7 and CYP2D8, enablingdetection of the deletion allele CYP2D6*5 whichresults from the complete loss of CYP2D6 (Gaedigket al., 1991).

Where a 9–16 kb band was detected by Southernblotting, a long PCR (Johansson et al., 1996; Steenet al., 1995) was performed to distinguish betweenthe hybrid alleles (CYP2D6*13 and *16) andCYP2D6*5.

Statistical analysis

All clinical and genotyping data were entered into therelational database PARADOX, and transferred forstatistical analysis to SYSTAT. Statistical tests usedincluded descriptive statistics and chi-square tests,with corrections for small numbers where applicable.

RESULTS AND DISCUSSION

Although 195 depressed patients (57% female, meanage 31.6) were entered into the clinical trial, completeCYP2D6 genotypes were available on only 125patients, who are the subjects for this paper. Of the125 patients, 115 were inferred as EMs on the basisof genotype as they carried wild type (CYP2D6*1)or intermediate (CYP2D6*2, *9 or *10) alleles. Theother ten patients were classified as PMs on the basisof genotype, as they were homozygous or compoundheterozygous for two defective alleles (CYP2D6*4/*4or *4/*5). No duplicated (CYP2D6*N�M) or hybridalleles were detected in this population.

Of the 125 patients who commenced the antide-pressant to which they had been randomized, 15%(10/65) who were prescribed fluoxetine and 30%(18/60) who were prescribed nortriptyline did notcomplete an adequate 6 week trial (�2¼ 3.83,p¼ 0.05). One patient on fluoxetine and six on nor-triptyline discontinued the antidepressant within thefirst week; six on fluoxetine and seven on nortriptylinediscontinued during weeks 1–3; and three patients onfluoxetine and four on nortriptyline discontinuedbetween weeks 3 and 6. One patient on nortriptyline,with moderate postural hypotension, took the antide-pressant for 6 weeks but at an inadequate dose andwith inadequate plasma levels, and was deemed to

CYP2D6 AND ANTIDEPRESSANT ADRs 19

Copyright # 2004 John Wiley & Sons, Ltd. Hum Psychopharmacol Clin Exp 2004; 19: 17–23.

Page 4: No evidence of increased adverse drug reactions in cytochrome P450 CYP2D6 poor metabolizers treated with fluoxetine or nortriptyline

have not completed an adequate trial. Table 1 showsthe likelihood of completing an antidepressant trialby drug and metabolizer status. While patients on nor-triptyline were less likely to complete an adequatetrial, there was no significant effect of metabolizer sta-tus (i.e. PMs were not more likely to fail to completean adequate trial). Of the three PMs who did not com-plete an adequate trial one on fluoxetine and one onnortriptyline discontinued between weeks 1 and 3,and one patient on fluoxetine discontinued betweenweeks 3 and 6.

Of the PMs on fluoxetine, their dosages at 6 weekswere 10 mg, 20 mg (two patients), 40 mg, and 60 mg.It was the patient on 60 mg who reported fatigue as aside effect. For the poor metabolizers on nortriptyline,all were on 75 mg at 6 weeks, with blood levels withinthe therapeutic range 200–400 nmol/l.

Table 2 shows the number of patients by drug andmetabolizer status that reported clinically significantADRs during the first 6 weeks. From this table it

can be seen that nausea or gastrointestinal upset, anxi-ety, agitation or restlessness were the most commonADRs with fluoxetine, whereas dry mouth and pos-tural hypotension were the most clinically significantADRs associated with nortriptyline. From the sum-mary of ADRs at the bottom of Table 2, it can be seenthat PMs were no more likely to experience ADRsthan EMs, and of the 17 patients who developed mod-erate or severe ADRs, only one was a PM. As numbersare small, multiple statistical tests were not under-taken.

Several previous clinical studies investigatedwhether CYP2D6 PMs were susceptible to ADRs onpsychoactive medication. A pilot study of CYP2D6genotyping in a psychiatric hospital concluded thatpatients of PM genotype were more likely to sufferADRs than those of the EM genotype (de Leonet al., 1998). Another study examined the occurrenceof extrapyramidal ADRs in psychiatric patients andfound that those with a PM genotype were more likelyto develop ADRs when treated with CYP2D6-meta-bolized drugs (Vandel et al., 1999). The subjects ofthese studies were psychiatric inpatients being treatedwith a combination of antidepressants and antipsycho-tics, and most had a history of exposure to psychoac-tive medication. Moreover, most of the patients in thefirst study were taking two drugs that were CYP2D6substrates (Vandel et al., 1999), as were 10% of thepatients in the second study, with a similar proportiontaking a second drug known to inhibit CYP2D6 (deLeon et al., 1998). Furthermore, in the second study,

Table 1. Frequency of patients who failed to complete an adequate6 week antidepressant trial by drug and metabolizer status

Metabolizer status Percent (n) who failed to complete6 week trial

Fluoxetine Nortriptyline Total

EM 14% (8/59) 30% (17/56) 22% (25/115)PM 33% (2/6) 25% (1/4) 30% (3/10)Total 15% (10/65) 30% (18/60) 22% (28/125)

Table 2. The frequency of clincially significant antidepressant associated ADRs by drug and metabolizer status reported during the6 week trial

ADR Fluoxetine Nortriptyline

EM (n¼ 59) PM (n¼ 6) EM (n¼ 56) PM (n¼ 4)

Anxiety, agitation, restlessness 7 0 2 0Postural hypotension 0 0 7 0Nausea, gastrointestinal upset 7 0 3 0Dry mouth 1 0 20 3Sweating 1 0 4 0Blurred vision 0 0 2 0Constipation 0 0 3 0Headache 3 0 2 0Sedation 1 0 0 1Insomnia 1 0 0 0Rash 1 0 0 0Sexual dysfunction 2 0 0 0Fatigue 2 1 1 0Percentage (n) with

Any mild 27% (16) 17% (1) 34% (19) 50% (2)Any moderate/severe 14% (8) 0 14% (8) 25% (1)Any total 41% (24) 17% (1) 48% (27) 75% (3)

20 r. l. roberts ET AL.

Copyright # 2004 John Wiley & Sons, Ltd. Hum Psychopharmacol Clin Exp 2004; 19: 17–23.

Page 5: No evidence of increased adverse drug reactions in cytochrome P450 CYP2D6 poor metabolizers treated with fluoxetine or nortriptyline

8 of the 12 patients with CYP2D6 PM genotype wereon two CYP2D6-metabolized drugs (de Leon et al.,1998).

Chen et al. (1996) included, in their analysis ofCYP2D6 screening costs and influence on clinicaloutcomes, 18 depressed patients who had experiencedan ADR when treated with a CYP2D6-metabolizedantidepressant (Chen et al., 1996). A significantlygreater frequency of defective CYP2D6 allelesoccurred in this group compared with a group ofdepressed patients who had not reported ADRs, butvery few of the patients were PMs, and it is not clearhow many drugs these patients were being concomi-tantly treated with.

Differences between the design of our study andthose described above make it difficult to directlycompare findings. First, the nortriptyline PMs in ourstudy were on slightly lower doses than the EMs,because clinicians adjusted the doses based on clinicalresponse, side effect and blood levels. This may haveobscured effects that could arise in the absence of doseadjustment. Second, in contrast to the above studies,none of our patients were concurrently taking anyother medication, and 66% of our subjects had noprior exposure to any antidepressant. It is conceivablethat many of the ADRs reported by others (de Leonet al., 1998; Chou et al., 2000; Vandel et al., 1999;Chen et al., 1996) may be the result of drug–druginteractions exacerbated by CYP2D6 PM genotype,which would not have been observed in our clinicaltrial. If so, this would suggest that CYP2D6 PM gen-otype may be of most relevance in cases where morethan one drug is being prescribed. A possible excep-tion is described by Spina et al. (1997), in a prospec-tive study of 31 depressed outpatients treated solelywith desipramine, in which the two enrolled CYP2D6phenotypic PMs both experienced severe ADRsrequiring dosage reduction (Spina et al., 1997).

Other studies have also questioned the clinicalrelevance of CYP2D6 metabolizer status for predict-ing treatment outcome or ADRs. An early study ofimipramine treatment of depression in patients subse-quently phenotyped with debrisoquine and mepheny-toin (Meyer et al., 1988) found no correlation betweenPM status for either probe drug and the intensity orfrequency of side effects. However, the four debriso-quine PMs in the group of 42 phenotyped patientswere all judged to be non-responders to imipramine(Meyer et al., 1988). In contrast, we observed thatPM genotypes were not over-represented amongstnon-responders being treated with fluoxetine ornortriptyline.

De Vries et al. (2001) studied the frequency ofCYP2D6 and CYP2C19 alleles, and blood drug con-centrations, in 39 psychiatric patients who sufferedtherapeutic failure due to severe ADRs while beingtreated with drugs that were substrates for theseenzymes (de Vries et al., 2001). They found a distri-bution of PM alleles in this group, for both genes, thatreflected the normal population distribution, and con-cluded that routine genotyping would be of no clinicalrelevance for predicting therapeutic failure causedby severe side effects after treatment with CYP2D6/2C19 substrates. Finally, follow up of ADRs describedas seizures or myoclonus during treatment with anti-depressants in the Swedish national database did notimplicate CYP2D6 or CYP2C19 PM status as risk fac-tors for these ADRs (Spigset et al., 1997).

As only ten CYP2D6 PMs were detected in ourstudy, further work is needed to determine whetherour observations hold for a larger group, and for otherantidepressant drugs. Numerous case reports supportthe view that the combination of antidepressant med-ication and CYP2D6 PM genotype can lead to severeor lethal ADRs (Bluhm et al., 1993; Bertilsson et al.,1981; Swanson et al., 1997; Sallee et al., 2000). How-ever, case reports reflect extreme examples selectedfrom large numbers of treated patients, and it is diffi-cult to assess other possible genetic or environmentalinfluences in these isolated cases. A review of theexisting literature on CYP2D6 and CYP2C19 geno-type effects on antidepressant treatment led to preli-minary dose recommendations for individuals ofdifferent genotypes (Kirchheiner et al., 2001), butidentified only a small number of studies that providedefficacy data in relation to genotype, and of these nonefound a significant difference between PM and EMindividuals in the per cent improvement of HDRSscores (Baumann et al., 1986; Morinobu et al.,1997; Koyama et al., 1996). Although several of thereviewed studies reported a higher incidence of ADRsin PMs than EMs, only five studies provided explicitdata (Kirchheiner et al., 2001). In all five studies, thenumber of PMs was small, leading to low statisticalpower. Indeed, only one study showed a significantdifference between EM and PM in the frequency ofside effects on desipramine (Spina et al., 1997).

Our data strongly suggest that CYP2D6 genotypicPMs receiving solely fluoxetine or nortriptyline areat no greater risk of ADRs than genotypic EMpatients. However, an accurate assessment of the truerate of antidepressant-induced ADRs in CYP2D6 PMsremains to be carried out, and this will require sub-stantially larger groups than have been studied to date.

CYP2D6 AND ANTIDEPRESSANT ADRs 21

Copyright # 2004 John Wiley & Sons, Ltd. Hum Psychopharmacol Clin Exp 2004; 19: 17–23.

Page 6: No evidence of increased adverse drug reactions in cytochrome P450 CYP2D6 poor metabolizers treated with fluoxetine or nortriptyline

ACKNOWLEDGEMENTS

We thank Allison Miller, Isobel Stevens and RobynAbbott for contributions to this research. This workwas supported by grants from the Health ResearchCouncil of New Zealand and University of Otago.Rebecca Roberts was a Leslie Averill Fellow of theCanterbury Medical Research Foundation, and MartinKennedy is a Senior Research Fellow of the HealthResearch Council of New Zealand.

REFERENCES

Baumann P, Jonzier-Perey M, Koeb L, et al. 1986. Amitriptylinepharmacokinetics and clinical response: II. Metabolic poly-morphism assessed by hydroxylation of debrisoquine and mephe-nytoin. Int Clin Psychopharmacol 1: 102–112.

Bertilsson L, Dahl ML, Dalen P, et al. 2002. Molecular genetics ofCYP2D6: clinical relevance with focus on psychotropic drugs. BrJ Clin Pharmacol 53: 111–122.

Bertilsson L, Dahl ML, Tybring G. 1997. Pharmacogenetics of anti-depressants: clinical aspects. Acta Psychiatr Scand Suppl 391:14–21.

Bertilsson L, Mellstrom B, Sjokvist F, et al. 1981. Slow hydroxyla-tion of nortriptyline and concomitant poor debrisoquine hydroxy-lation: clinical implications. Lancet 1: 560–561.

Bluhm RE, Wilkinson GR, Shelton R, et al. 1993. Genetically deter-mined drug-metabolizing activity and desipramine-associatedcardiotoxicity: a case report. Clin Pharmacol Ther 53: 89–95.

Carter JD, Joyce PR, Mulder RT, et al. 2000. Gender differences inthe presentation of depressed outpatients: a comparison ofdescriptive variables. J Affect Disord 61: 59–67.

Chen S, Chou WH, Blouin RA, et al. 1996. The cytochrome P4502D6 (CYP2D6) enzyme polymorphism: screening costs andinfluence on clinical outcomes in psychiatry. Clin PharmacolTher 60: 522–534.

Chomczynski P, Qasba P. 1984. Alkaline transfer of DNA to plasticmembrane. Biochim Biophys Acta 122: 340–344.

Chou WH, Yan FX, de Leon J, et al. 2000. Extension of a pilotstudy: impact from the cytochrome P450 2D6 polymorphismon outcome and costs associated with severe mental illness. JClin Psychopharmacol 20: 246–251.

Ciulla T, Slar R, Hauser S. 1988. A simple method for DNA purifi-cation from peripheral blood. Anal Biochem 174: 485–488.

de Leon J, Barnhill J, Rogers T, et al. 1998. Pilot study of the cyto-chrome P450-2D6 genotype in a psychiatric state hospital. Am JPsychiatry 155: 1278–1280.

de Vries EM, Pot HJ, Conemans JMH, et al. 2001. The clinical rele-vance of routine genotyping of CYP2D6/2C19 null alleles of psy-chiatric patients. Br J Clin Pharmacol 53: 543P–559P.

Eichelbaum M, Gross AS. 1990. The genetic polymorphism of deb-risoquine/sparteine metabolism—clinical aspects. PharmacolTher 46: 377–394.

Fjordside L, Jeppesen U, Eap CB, et al. 1999. The stereoselectivemetabolism of fluoxetine in poor and extensive metabolizers ofsparteine. Pharmacogenetics 9: 55–60.

Gaedigk A, Blum M, Gaedigk R, et al. 1991. Deletion of the entirecytochrome P450 CYP2D6 gene as a cause of impaired drugmetabolism in poor metabolizers of the debrisoquine/sparteinepolymorphism. Am J Hum Genet 48: 943–950.

Hamelin BA, Turgeon J, Vallee F, et al. 1996. The disposition offluoxetine but not sertraline is altered in poor metabolizers ofdebrisoquin. Clin Pharmacol Ther 60: 512–521.

Hamilton M. 1960. A rating scale for depression. J Neurol Neuro-surg Psychiatry 12: 56–62.

Heim M, Meyer UA. 1990. Genotyping of poor metabolisers of deb-risoquine by allele-specific PCR amplification. Lancet 336: 529–532.

Johansson I, Lundqvist E, Dahl ML, et al. 1996. PCR-based geno-typing for duplicated and deleted CYP2D6 genes. Pharmacoge-netics 6: 351–355.

Joyce PR, Mulder RT, Luty SE, et al. 2002. Patterns and predictorsof response, remission and recovery in major depression treatedwith fluoxetine or nortriptyline. Aust NZJ Psychiat 36: 384–391.

Kirchheiner J, Brosen K, Dahl ML, et al. 2001. CYP2D6 andCYP2C19 genotype-based dose recommendations for antide-pressants: a first step towards subpopulation-specific dosages.Acta Psychiatr Scand 104: 173–192.

Koyama E, Tanaka T, Chiba K, et al. 1996. Steady-state plasma con-centrations of imipramine and desipramine in relation to S-mephenytoin 4’-hydroxylation status in Japanese depressivepatients. J Clin Psychopharmacol 16: 286–293.

Kroemer HK, Eichelbaum M. 1995. ‘It’s the genes, stupid’. Mole-cular bases and clinical consequences of genetic cytochromeP450 2D6 polymorphism. Life Sci 56: 2285–2298.

Lessard E, Yessine MA, Hamelin BA, et al. 1999. Influence ofCYP2D6 activity on the disposition and cardiovascular toxicityof the antidepressant agent venlafaxine in humans. Pharmacoge-netics 9: 435–443.

Luty SE, Joyce PR, Mulder RT, et al. 2002. Social adjustment indepression: the impact of depression severity, personality, andclinic versus community sampling. J Affect Disord 70: 143–154.

Marchioni E, Perucca E, Soragna D, et al. 1996. Choreiform syn-drome associated with fluoxetine treatment in a patient with defi-cient CYP2D6 activity. Neurology 46: 853.

Mellstrom B, Bertilsson L, Sawe J, et al. 1981. E- and Z-10-hydro-xylation of nortriptyline: relationship to polymorphic debriso-quine hydroxylation. Clin Pharmacol Ther 30: 189–193.

Meyer JW, Woggon B, Kupfer A. 1988. Importance of oxidativepolymorphism on clinical efficacy and side-effects of imipra-mine—a retrospective study. Pharmacopsychiatry 21: 365–366.

Meyer UA. 1994. Pharmacogenetics: the slow, the rapid, and theultrarapid. Proc Natl Acad Sci USA 91: 1983–1984.

Montgomery SA, Asberg M. 1979. A new depression scale designedto be sensitive to change. Br J Psychiatry 134: 382–389.

Morinobu S, Tanaka T, Kawakatsu S, et al. 1997. Effects of geneticdefects in the CYP2C19 gene on the N-demethylation of imipra-mine, and clinical outcome of imipramine therapy. PsychiatryClin Neurosci 51: 253–257.

Morita S, Shimoda K, Someya T, et al. 2000. Steady-state plasmalevels of nortriptyline and its hydroxylated metabolites inJapanese patients: impact of CYP2D6 genotype on the hydroxy-lation of nortriptyline. J Clin Psychopharmacol 20: 141–149.

Nordin C, Bertilsson L. 1995. Active hydroxymetabolites of antide-pressants. Emphasis on E-10-hydroxy- nortriptyline. Clin Phar-macokinet 28: 26–40.

Perucca E, Marchioni E, Soragna D, et al. 1997. Fluoxetine-inducedmovement disorders and deficient CYP2D6 enzyme activity. MovDisord 12: 624–625.

Roberts R, Joyce P, Kennedy MA. 2000. Rapid and comprehensivedetermination of cytochrome P450 CYP2D6 poor metabolizergenotypes by multiplex polymerase chain reaction. Hum Mutat16: 77–85.

22 r. l. roberts ET AL.

Copyright # 2004 John Wiley & Sons, Ltd. Hum Psychopharmacol Clin Exp 2004; 19: 17–23.

Page 7: No evidence of increased adverse drug reactions in cytochrome P450 CYP2D6 poor metabolizers treated with fluoxetine or nortriptyline

Roberts RL, Joyce PR, Mulder RT, et al. 2002. A common P-glyco-protein polymorphism is associated with nortriptyline-inducedpostural hypotension in patients treated for major depression.Pharmacogenom J 2: 191–196.

Sachse C, Brockmoller J, Bauer S, et al. 1997. Cytochrome P4502D6 variants in a Caucasian population: allele frequencies andphenotypic consequences. Am J Hum Genet 60: 284–295.

Sallee FR, DeVane CL, Ferrell RE. 2000. Fluoxetine-related deathin a child with cytochrome P-450 2D6 genetic deficiency. J ChildAdolesc Psychopharmacol 10: 27–34.

Spigset O, Hedenmalm K, Dahl ML, et al. 1997. Seizures and myo-clonus associated with antidepressant treatment: assessment ofpotential risk factors, including CYP2D6 and CYP2C19 poly-morphisms, and treatment with CYP2D6 inhibitors. Acta Psy-chiatr Scand 96: 379–384.

Spina E, Gitto C, Avenoso A, et al. 1997. Relationship betweenplasma desipramine levels, CYP2D6 phenotype and clinicalresponse to desipramine: a prospective study. Eur J Clin Pharma-col 51: 395–398.

Spitzer RL, Williams JB, Gibbon M, et al. 1992. TheStructured Clinical Interview for DSM-III-R (SCID). I:History, rationale, and description. Arch Gen Psychiatry 49:624–629.

Steen VM, Andreassen OA, Daly AK, et al. 1995. Detection of thepoor metabolizer-associated CYP2D6(D) gene deletion allele bylong-PCR technology. Pharmacogenetics 5: 215–223.

Swanson JR, Jones GR, Krasselt W, et al. 1997. Death of two sub-jects due to imipramine and desipramine metabolite accumula-tion during chronic therapy: a review of the literature andpossible mechanisms. J Forensic Sci 42: 335–339.

Vandel P, Haffen E, Vandel S, et al. 1999. Drug extrapyramidal sideeffects. CYP2D6 genotypes and phenotypes. Eur J Clin Pharma-col 55: 659–665.

Venkatakrishnan K, von Moltke LL, Greenblatt DJ. 1999. Nortripty-line E-10-hydroxylation in vitro is mediated by human CYP2D6(high affinity) and CYP3A4 (low affinity): implications for inter-actions with enzyme-inducing drugs. J Clin Pharmacol 39:567–577.

CYP2D6 AND ANTIDEPRESSANT ADRs 23

Copyright # 2004 John Wiley & Sons, Ltd. Hum Psychopharmacol Clin Exp 2004; 19: 17–23.