Clinically significant pharmacokinetic drug interactions with psychoactive drugs: antidepressants and antipsychotics and the cytochrome P450 system
Summary
Psychotherapeutic drugs (antipsychotics and antidepressants) are widely used for treating anxiety. Many psychotherapeutic drugs are metabolized mainly by cytochrome P450 (CYP)2C19 and CYP2D6, and are often administered with other drugs. Therefore, it is necessary to be careful when coadministering psychotherapeutic drugs whose metabolism might be inhibited by other drugs. In particular, selective serotonin reuptake inhibitors (SSRIs) inhibit the metabolism of psychotherapeutic drugs mediated by CYP2C19 and CYP2D6. It is useful to phenotype CYP2C19 and CYP2D6 (extensive metabolizers or poor metabolizers) before giving such medication.
Knowledge of substrates, inhibitors and inducers of CYP isoenzymes may help clinicians to anticipate and avoid psychotherapeutic drug interactions and improve rational prescribing practices. In addition, genotyping for these drugs may be also useful in preventing side-effects.
Introduction
The cytochrome P450 (P450 or CYP) isoenzymes are a group of heme-containing enzymes embedded primarily in the lipid bilayer of the endoplasmic reticulum of hepatocytes. P450 takes part in the metabolism of many drugs, steroids and carcinogens (1) and more than 30 human CYP isozymes have been identified to date. It has been estimated that ≥90% of human drug oxidation can be attributed to six main enzymes (CYP1A2, 2C9, 2C19, 2D6, 2E1 and 3A4). The activities of the CYP2C19 (2–6) and CYP2D6 (2, 4, 6–13) enzymes are bimodally distributed in the population, allowing classification of individuals as either extensive (EMs) or poor (PMs) metabolizers. PMs of CYP2D6 involve ≈5–10% of Caucasians compared with 1–3% of Orientals (2, 4). In addition, CYP2C19 exhibits a genetic polymorphism, with 20% of Orientals and 3–5% of Caucasians being reported as PMs (2, 4). When drugs are converted to an active metabolite by CYP2C19 or CYP2D6, the drug may be ineffective in PMs. Those individuals with impaired enzyme function resulting from genetic mutation (PMs) are likely to develop adverse effects from high levels of unmetabolized drugs (14–17).
Psychotherapeutic drugs used for treating depression, in many clinical guises, are referred to as antidepressant drugs. Those used for treating schizophrenia and other psychoses are termed antipsychotic drugs, while drugs for treating mania are referred to as mood-stabilizing drugs (e.g. lithium carbonate). Most antipsychotics and antidepressants are also metabolized by CYP2C19 (polymorphic mephenytoin hydroxylase) and/or CYP2D6 (polymorphic debrisoquine/sparteine hydroxylase).
In the present review, we will discuss clinically significant pharmacokinetic drug interactions involving antidepressants, but excluding benzodiazepines (18) and antipsychotics in terms of CYP levels.
METABOLISM OF ANTIDEPRESSANTS
As shown in Table 1, the P450 isozymes that catalyse a number of selective serotonin reuptake inhibitors (SSRI) are mainly CYP2C19 and CYP2D6, polymorphic CYP enzymes. Tricyclic antidepressants are also metabolized mainly by these two isozymes (CYP2C19 and CYP2D6) which produce active metabolites. Active metabolites undergo further N-demethylation and/or hydroxylation or conjugation resulting in inactive compounds that are excreted in urine. Those metabolized by CYP2C19 are nortriptyline (metabolite: E-10-hydroxynortriptyline, E-10-hydroxydidemethylÍamitriptyline) (19) and imipramine (metabolite: desipramine) (20–23) and those metabolized by CYP2D6 are amitriptyline (metabolite: l0-hydroxyamitriptyline) (24, 25), clomipramine (metabolite: 8-hydroxyÍclomipramine) (26), desipramine (metabolite: 8-hydroxylclomipramine, 2-hydroxydesipramine) (27) and imipramine (metabolite: 2-hydroxyimipramine) (22, 23, 28, 29). Since the tetracyclic antidepressants mianserin (28) (metabolite: 8-hydroxymianserin) and maprotiline (metabolite:?) (29) are also metabolized by these CYPs, they may produce severe effects in PMs following normal doses of debrisoquine. For example, Morinobu et al. reported the relationship between the genetic polymorphism of 5-mephenytoin 4-hydroxylation catalysed by CYP2C19 and the N-demethylation of imipramine in 10 depressed Japanese patients, five of whom were PMs and five EMs. The demethylation index (the desipramine/imipramine ratio) was significantly lower in patients with genetic defects. Plasma levels of imipramine and 2-hydroxyÍimipramine normalized by the daily dose (mg) per patient weight (kg) were significantly higher in patients with genetic defects (21). Imipramine metabolism in clinical genetic trials gave similar results as shown by Koyama et al. (22, 23).
METABOLISM OF ANTIPSYCHOTICS
The metabolism of most antipsychotics is governed by CYP2D6 (Table 1). Chlorpromazine is metabolized mainly by demethylation (metabolite: nor- and di-norchlorpromazine), hydroxylation (metabolite: 7-hydroxychlorpromazine) and sulfoxidation (30). Levomepromazine is similarly metabolized by demethylation, hydroxylation and sulfoxidation (sulfoxide) (31). Other phenothiazine derivatives are also metabolized by CYP2D6 isoenzyme. Haloperidol, a butyrophenone derivative, is extensively bioÍtransformed to some inactive metabolites by CYP2D6 (metabolite: reduced haloperidol, 4-fluoroÍbenzoylpropionic acid and 4-fluorophenylacetic acid) and CYP3A4 (metabolite: 4-(4-chlorophenyl)-4-hydroxypiperidine) (32, 33). However, in the case of other antipsychotics, e.g. iminodibenzyl derivatives, benzamide derivatives, it is not known what kind of CYPs catalyse their metabolism. In pharmacokinetic studies, Tyndale et al. (34) and Llerena et al. (35, 36) reported that interethnic differences in debrisoquine hydroxylation polymorphism (CYP2D6) might partly be responsible for the variation in haloperidol disposition between races. Similar results were also found for perphenazine (37) and thioridazine (38, 39) metabolism in genetic polymorphism tests.
INTERACTIONS WITH ANTIDEPRESSANTS
Selective serotonin reuptake inhibitors (Table 2)
Selective serotonin reuptake inhibitors (SSRIs), such as paroxetine, fluoxetine, fluvoxamine, sertraline, norfluoxetine and citalopram, are used concurrently because of the high prevalence of comorbid psychotic and depressive symptoms. Many SSRIs are metabolized by CYP2D6 and many drugs such as tricyclic antidepressants, antipsychotics and benzodiazepines have the potential to interact with SSRIs (40–44).
Tricyclic antidepressants
CYP2D6 is responsible for the hydroxylation of tricyclic antidepressants, while CYP1A2, CYP3A4 and CYP2C19 are involved in their N-demethylation. There is a great deal of evidence for potent inhibitory interaction between SSRIs and tricyclic antidepressants (45–55). For example, Kurtz et al. reported a study in which 12 healthy male subjects received a 50-mg single dose of either desipramine or imipramine under three conditions: (i) alone, (ii) after a single 150 mg dose of sertraline, and (iii) after the eighth daily 150 mg dose of sertraline. Treatment with sertraline significantly reduced the apparent plasma clearance (CL) and prolonged the half-life of desipramine. These changes resulted in higher plasma desipramine concentrations, as indicated by a significant increase in the maximum plasma concentration (Cmax) and area under the plasma concentration–time curve (AUC) (22% and 54%, respectively). Both single- and multiple-dose treatment with sertraline significantly reduced the CL of imipramine. This effect was stronger after multiple predoses of sertraline, when the imipramine Cmax and AUC increased by 39% and 68%, respectively (45).
Benzodiazepines
Combination therapy with benzodiazepines and SSRIs is likely in cases of comorbid anxiety and depression as well as during treatment initiation with an SSRI where benzodiazepines can relieve adverse effects. Benzodiazepines are hydroxylated mainly by CYP3A4 (e.g. midazolam), while others are directly conjugated (e.g. oxazepam) or nitro-reduced (e.g. nitrazepam). In vivo and in vitro alprazolam metabolism is inhibited by fluvoxamine (56), fluoxetine (57,58) and sertraline (59), but not by paroxetine (60). These results show that when any two psychoactive drugs are administered together, increased patient monitoring and patient education is recommended, e.g. when alprazolam and fluoxetine are prescribed concurrently. Fluvoxamine (61) and sertraline (62) have been shown to inhibit diazepam metabolism. The effect of fluoxetine (63) and paroxetine (64) on diazepam metabolism, however, is less controversial. When fluvoxamine (100 mg/day) was coadministered with a single dose of bromazepam (12 mg), the bromazepam AUC and half-life were significantly increased (65). In vitro 1′-hydroxy midazolam formation was inhibited by sertraline and desmethylsertraline, fluoxetine and norfluoxetine (66).
Anticonvulsants
Fluoxetine may be expected to interact with carbamazepine (67, 68) and phenytoin (69). For example, Shader et al. reported that, after administration of fluoxetine, the phenytoin concentration increased by 160%. In addition, fluvoxamine (53) and sertraline (70) have also been shown to increase carbamazepine concentrations. However, high doses of setraline do not appear to affect the pharmacokinetics or pharmacodynamics of phenytoin in healthy volunteers (71).
Antipsychotics
Many antipsychotic drugs are metabolized mainly by CYP2D6 (30–39). Unlike other antipsychotics, clozapine is metabolized primarily by CYP1A2 (72). Goff et al. reported a 20% increase (P< 0·05) in plasma haloperidol levels after administration of fluoxetine (20 mg/day) for 7–10 days to eight psychotic patients (73). Fluvoxamine also increases haloperidol levels (74) but the mechanism of this inhibition is unclear. There have been several reports of increased plasma clozapine concentrations after administration of fluvoxamine (75–77), fluoxetine (78, 79), paroxetine (79) and sertraline (79).
Others
Other drug interactions have been reported between psychoactive and other drugs such as warfarin (mediated by CYP1A2 and CYP2C9) (80–82), theophylline (mediated by CYP1A2 and CYP3A4) (83–85), and terfenadine (mediated by CYP3A4) (86, 87). Recently, Spina et al. reported their study of the effect of oral ketoconazole (200 mg/day for 14 days) on the kinetics of a single oral dose of imipramine (100 mg) and desipramine (100 mg) in two groups of six healthy male subjects. Ketoconazole, a relatively specific inhibitor of CYP3A4, was shown to reduce imipramine clearance and prolong its half-life. However, it does not affect the 2-hydroxylation of imipramine and desipramine (88).
Tricyclic antidepressants
Tricyclic antidepressants and benzodiazepines are often used to treat depression. Although it has been reported that alprazolam increases serum concentrations of imipramine and desipramine (89), only a few reports have been published on drug–drug interactions involving these drugs. Interactions between tricyclic antidepressants and SSRIs, and between tricyclic antidepressants and antipsychotics are covered in separate sections.
INTERACTIONS WITH ANTIPSYCHOTICS
Selective serotonin reuptake inhibitors (Table 2)
Many antipsychotic drugs are metabolized mainly by CYP2D6. However, clozapine is metabolized primarily by CYP1A2 (72) and by CYP2D6 as a secondary pathway. Centorrino et al. reported an interaction between clozapine (346 mg for at least 1 week) and fluoxetine (36 mg) in psychotic patients. Fluoxetine significantly increased the clozapine levels compared with control patients (78). Fluoxetine also inhibited haloperidol metabolism (73, 90). Llerena et al. reported a pharmacogenetic trial (91) in which the plasma levels of the reduced metabolite of haloperidol, after a single oral dose (2 or 4 mg) of parent drug, were significantly higher in PM than in EM of debrisoquine hydroxylase (CYP2D6). Fluvoxamine has been shown to increase serum concentrations of haloperidol (74) and clozapine (75). In an interaction study of paroxetine (20 mg/day orally for 10 days) and perphenazine (0·11 mg/kg) orally, paroxetine treatment resulted in a 2- to 21-fold reduction in CYP2D6 activity (P< 0·001) and, after pretreatment with paroxetine, perphenazine peak plasma concentrations increased 2- to 13-fold (P< 0·01) (92).
Tricyclic and tetracyclic antidepressants
Cooper et al. reported that, in 65 schizophrenic patients, perphenazine significantly increased steady-state nortriptyline plasma levels by 55%, probably through inhibition of the hydroxylation metabolic pathway. There was no effect on amitriptyline levels (93). Vandel et al. also obtained similar results (94). Nelson & Jatlow (95) investigated the effect of neuroleptic drugs on desipramine steady-state plasma concentrations in 30 patients. In 15 of these patients receiving a neuroloeptic drug, desipramine plasma levels were 2-fold higher than those in 15 patients who received dispiramine alone. Loga et al. have studied the concomitant use of chlorpromazine and nortriptyline (96). They found that the plasma chlorpromazine concentrations in seven male in-patients suffering from acute schizophrenia rose when nortriptyline was added, and the antipyrine plasma half-life was prolonged.
Anticonvulsants
Phenobarbital, phenytoin and carbamazepine during long-term treatment are strong inducers of several drugs. Coadministration of these drugs reduces their pharmacological effects. A pharmacokinetic interaction between haloperidol and carbamazepine has been studied in schizophrenic patients. The serum carbamazepine concentrations in patients treated without haloperidol were reduced significantly (P< 0·05), on average ≈40%, compared with those treated with both haloperidol and carbamazepine. However, the serum haloperidol concentrations in patients treated with both haloperidol and carbamazepine were significantly reduced (P< 0·05), compared with those treated with haloperidol but not with carbamazepine (97). In addition, coadministration of carbamazepine reduced plasma clozapine concentrations by about 50% (98) or 47% (99). However, there is a report of a non-interaction between carbamazepine and thioridazine (99). Linnoila et al. showed that patients who had therapeutic plasma levels of phenobarbital and/or phenytoin (diphenylhydantoin) had significantly lower plasma levels of haloperidol and mesoridazine, the active metabolite of thioridazine, compared with patients who were not given anticonvulsants. Plasma thioridazine levels per se were unaffected by concomitant anticonvulsant treatment (100).
Others
Markowitz et al. have published a comprehensive review of the pharmacokinetic and pharmacodynamic interactions between antipsychotics and antihypertensives and provided recommendations for the selection of antihypertensives in patients receiving antipsychotic therapy. They state that the combination of the beta-antagonists propranolol or pindolol with thioridazine or chlorpromazine should be avoided if possible (101).
CONCLUSIONS
Many reviews have been published on the clinical pharmacokinetic drug interactions of psychotherapeutic drugs (antipsychotics and antidepressants) (102–107). Care must be taken when coadministering psychotherapeutic drugs to avoid inhibition and induction of the metabolism of other drugs. In particular, selective serotonin reuptake inhibitors (SSRIs) inhibit the metabolism of psychotherapeutic drugs mediated by CYP2C19 and CYP2D6. Knowledge of substrates, inhibitors and inducers of CYP isoenzymes may help clinicians to anticipate and avoid psychotherapeutic drug interactions and improve rational prescribing practices, including the genotypic determination of these drugs.
