Plasma concentrations of haloperidol are related to CYP2D6 genotype at low, but not high doses of haloperidol in Korean schizophrenic patients

Abstract

Aims

This study was carried out to evaluate the influence of CYP2D6 genotype on the steady state plasma concentrations of haloperidol and reduced haloperidol in Korean schizophrenic patients.

Methods

One hundred and twenty Korean schizophrenic patients treated with various, clinically determined, doses of haloperidol (range 3–60, median 20 mg day⁻¹) during monotherapy were recruited. CYP2D6 genotypes were determined by analysis of the CYP2D6*10 allele using allele-specific PCR and the CYP2D6*5 allele by long-PCR. Steady state plasma concentrations of haloperidol and reduced haloperidol were analysed by h.p.l.c.

Results

Twenty-three (19.2%), 60 (50.0%), 1 (0.8%), 33 (27.5%) and 3 patients (2.5%) possessed the CYP2D6 genotypes *1/*1, *1/*10, *1/*5, *10/*10 and *10/*5, respectively. The allele frequencies of CYP2D6*1, *10 and *5 were 44.6%, 53.8% and 1.7%, respectively. Significant relationships between dose and plasma concentrations of haloperidol (linear; r² = 0.60, P < 0.0001) and reduced haloperidol (quadratic equation; r² = 0.67) were observed. Overall, the concentrations normalized for dose (C/D) of haloperidol were significantly different between the CYP2D6*1/*1, *1/*10 and *10/*10 genotype groups (one-way anova; P = 0.028). No significant differences between the genotype groups were found with respect to the C/D of reduced haloperidol (P = 0.755). However, in patients with daily doses less than 20 mg, significant differences in the C/D of haloperidol (P = 0.003), but not of reduced haloperidol, were found between the three major genotype groups. In patients with doses higher than 20 mg, no differences were found between the genotype groups for either haloperidol or reduced haloperidol. 68 patients (57%) used benztropine, an antimuscarinic agent. All four patients with a *5 allele (one together with *1 and three with *10) were found to use benztropine. The patients homozygous for the *1 allele seemed to need less benztropine than the patients with one or two mutated alleles (Fisher's exact test; P = 0.036).

Conclusions

The dose-corrected steady state plasma concentrations of haloperidol, but not of reduced haloperidol, were significantly different between the CYP2D6*1/*1, *1/*10 and *10/*10 genotype groups when doses lower than 20 mg haloperidol were given. No differences were found at higher doses. These results suggest the involvement of CYP2D6 in the metabolism of haloperidol at low doses of haloperidol (< 20 mg daily), while another enzyme, probably CYP3A4, contributes at higher doses.

Introduction

Haloperidol is one of the most frequently prescribed antipsychotics in patients with acute exacerbation of schizophrenia. The main metabolic pathways of haloperidol in humans include oxidative N-dealkylation [1], carbonyl reduction [2, 3], glucuronide conjugation [4], and formation of pyridinium metabolites [5, 6]. The keto group of haloperidol undergoes metabolic reduction to form reduced haloperidol, which is one of the main metabolites of haloperidol in humans [7–9]. Although the reduced haloperidol/haloperidol ratio in plasma has been reported to be related to clinical outcome [10, 11], the clinical importance of reduced haloperidol remains unclear. Reduced haloperidol is also oxidized back to haloperidol, and metabolic interconversion between haloperidol and reduced haloperidol has been observed in humans [12, 13].

A panel study in extensive (EM) and poor metabolizers (PM) of debrisoquine, whose metabolism is catalysed by polymorphic cytochrome P450 CYP2D6, showed that the mean clearance of haloperidol was significantly lower in PM than in EM after a single oral dose of haloperidol [14]. Furthermore, the mean plasma concentrations of reduced haloperidol were significantly higher in PM than in EM, indicating that the disposition of both haloperidol and reduced haloperidol after a single dose of haloperidol is related to the CYP2D6 polymorphism [15]. Recently, two studies from Japan [16, 17] suggested that the steady state plasma concentrations of both haloperidol and reduced haloperidol are at least partly dependent on CYP2D6 activity.

Three major mutated alleles of CYP2D6, CYP2D6*3, CYP2D6*4 and CYP2D6*5, account for 90–95% of the PM alleles in Caucasians [18–21]. Orientals differ from Caucasians in respect of their CYP2D6 activity. Firstly, less than 1% can be classified as PMs when using 12.6 as the antimode of the debrisoquine metabolic ratio (MR) found in Caucasians [22]. This is mainly because CYP2D6*4 is almost absent in Orientals [23]. Secondly, the distribution of the debrisoquine MR is shifted toward higher values in Oriental EMs compared wtih Caucasian EMs, indicating lower mean CYP2D6 activity in the former [22]. Molecular studies have revealed that this shift in Orientals is due to high frequency of the CYP2D6*10 allele giving rise to an unstable gene product with decreased CYP2D6 activity [23, 24, 25].

In the present study, we investigated the steady state plasma concentrations of haloperidol and reduced haloperidol in relation to the CYP2D6 genotype in Korean schizophrenic patients being treated with a wide range of haloperidol doses.

Methods

A total of 120 native Korean patients (71 males and 49 females) who met the DSM-IV diagnostic criteria for schizophrenia or schizoaffective disorder took part in this study. The patients' age ranged between 19 and 54 years (mean±s.d. 34 ± 7) and their body weight ranged from 39 to 87 kg (60 ± 10). Only patients who had taken the same dosage of haloperidol for at least 2 weeks were included to ensure that steady state had been reached. They were all undergoing monotherapy with oral haloperidol, but were allowed to use benzotropine for the treatment of extrapyramidal symptoms and/or benzodiazepines. It has been reported that these drugs have no effects on the pharmacokinetics of haloperidol [26] and other substrates of CYP2D6 [27]. The haloperidol doses administered to the patients were based on the clinical judgement of the treating physicians and varied from 3 to 60 mg day⁻¹ (mean±s.d. 21.4 ± 11.4; median 20). All patients were admitted to the Seoul National Mental Hospital in Seoul, Korea and were inpatients at the time of blood sampling. All patients gave their written informed consent before participation. The study protocol was approved by the Ethics Committee at Huddinge Hospital, Huddinge, Sweden. At the time of the study, there was no Ethics Committee at the Seoul National Mental Hospital.

Venous blood samples were taken just before the morning dose for the determination of drug plasma concentration and for genotyping. Blood (10 ml) was centrifuged and plasma was stored at −20 °C. Leukocyte nuclei were prepared from another 10 ml of blood. Plasma samples were brought to Sweden for analysis.

Plasma concentrations of haloperidol and reduced haloperidol were analysed by high performance liquid chromatography (h.p.l.c.). A 1.5 ml sample of plasma mixed with 50 µl of internal standard (50 µg ml⁻¹ of chlorsubstituted haloperidol) and 1 ml of 0.5 m sodium hydroxide in a 14 ml polypropylene tube, was extracted with 4 ml 3% isoamylalcohol in heptane by rotation for 10 min. After centrifugation for 10 min at 3000 g, 3.5 ml of the organic phase were transferred to another polypropylene tube containing 150 µl of 25 mm acetic acid. After extraction for 5 min and centrifugation for 5 min, the organic phase was discarded. Diethyl ether (1 ml) was added, and the tube was whirlmixed for 10 s. The ether phase was discarded and 15 µl of the acid phase was injected on to a 150 × 3.0 mm Zorbax SB C18 column (3.5 µm particle size). The eluent was a 50-mm potassium phosphate buffer pH 7.4 containing 2 mm N,N-dimethyl octylamine and 38% acetonitrile. The flow rate was 0.8 ml min⁻¹ and the temperature was 40 °C. Ultraviolet detection was carried out at a wavelength of 220 nm. Retention times were 4.1 min (reduced haloperidol), 6.8 min (haloperidol) and 11.0 min (internal standard). Linear standard curves with correlation coefficients higher than 0.99 were obtained in the range 2.5–320 nm for both haloperidol and reduced haloperidol. The limits of detection were about 1.0 nm for haloperidol, and about 0.5 nm for reduced haloperidol. Coefficients of variation (day to day, n = 14) of 4.7% at a concentration of 14.7 nm for haloperidol, and 8.7% at a concentration of 12.6 nm for reduced haloperidol, were obtained.

In order to identify the CYP2D6*10 allele containing the C188(r)T mutation in exon 1, DNA was prepared from leucocyte nuclei with a guanidinium isothiocyanate method, and allele specific PCR amplification was carried out according to Johansson et al. [23]. Alleles in which the C188(r)T mutation could not be detected are classified as CYP2D6*1. DNA samples from patients homozygous for the CYP2D6*1 or *10 alleles were further analysed for the CYP2D6*5 allele by long-fragment PCR as described by Johansson et al. [28].

Linear regression was used to analyse the relationship between the dose of haloperidol and the steady state concentrations of haloperidol, reduced haloperidol and the haloperidol/reduced haloperidol ratio. For the dose-concentration ratio for reduced haloperidol, the relationship was reanalysed using the quadratic polynomial equation (Y = A + Β×X + C×X²) to create a smooth curve for graphing.

One-way anova was used to compare the means of three major genotype groups (*1/*1, *1/*10 and *10/*10) in the total group, in patients with doses <20 mg and those with doses ≥20 mg. Then, the Bonferroni test was used for pairwise comparison of each genotype group. The Fisher's exact test was used to compare the difference between the patients with and without benztropine with respect to CYP2D6 genotype. The Grubb's test was used for detecting extreme outliers.

Results

Allele-specific PCR for CYP2D6*10 revealed that 23 (19.2%), 60 (50.0%), 1 (0.8%), 33 (27.5%) and 3 patients (2.5%) possessed the CYP2D6 genotypes *1/*1, *1/*10, *1/*5, *10/*10 and *10/*5, respectively. Due to analytical problems, analysis of the CYP2D6*5 alleles could be performed in only 44 out of the 60 patients homozygous for *1 or *10. Four patients were found to carry the *5 allele (one *1/*5 and 3 *10/*5). Of the remaining 16 samples not analysed for *5, only three additional patients should theoretically have had a *5 allele. The allele frequencies of CYP2D6*1, *10 and *5 in the present study population were thus 44.6%, 53.8% and 1.7%, respectively.

The daily dose of haloperidol in this study ranged between 3 and 60 mg with a mean of 21.4 mg and a median of 20 mg. For analysing the concentrations of haloperidol and reduced haloperidol, three extreme outliers (two with very high haloperidol concentration and one with extremely high reduced haloperidol concentration) were excluded by Grubb's test. A significant linear relationship between the dose and the plasma concentrations of haloperidol and reduced haloperidol were observed (r² = 0.60, P < 0.0001; r² = 0.56, P < 0.0001). However, the plasma concentrations of reduced haloperidol increased disproportionally at high doses and the data were statistically better fitted by a quadratic polynomial equation (r² = 0.67). Consequently, there was a negative relationship between the haloperidol/reduced haloperidol ratio and haloperidol dose (r² = 0.30, P < 0.0001) (Figure 1).

Figure 1.

Relationship between the haloperidol dose and the steady state plasma concentrations of (a) parent drug, (b) reduced haloperidol and (c) the haloperidol/reduced haloperidol (HP/RHP) ratio in schizophrenic patients.

The concentrations of haloperidol and reduced haloperidol divided by dose (C/D) in relation to the CYP2D6 genotypes are shown in Figure 2. To evaluate the difference in the dose-corrected haloperidol concentrations of the three major genotype groups (*1/*1, *1/*10 and *10/*10), the means of them were compared by one-way anova. In the entire patient group, the C/Ds of haloperidol were statistically different between the three major genotypes (one-way anova; P = 0.028). There was no significant difference between the genotype groups with respect to the C/D of reduced haloperidol (P = 0.755).

Figure 2.

The dose-corrected steady state concentrations of haloperidol and reduced haloperidol in relation to CYP2D6 genotype. The mean in each genotype group is indicated by the horizontal line. The three major genotype groups were compared by a one-way anova. The closed symbols designate the samples in which the *5 allele could not be determined due to analytical problems. *P < 0.05, Bonferroni test for comparison between *1/*1 and *10/*10. **P < 0.01, Bonferroni test for comparison between *1/*10 and *10/*10.

In patients on daily doses of less than 20 mg (the median dose), the mean values of the dose-corrected haloperidol concentrations were 2.0, 2.3, 1.7, 3.9 and 3.4 nm mg⁻¹ in the *1/*1, *1/*10, *1/*5, *10/*10 and *10/*5 genotype groups, respectively, and showed a statistically significant difference between the three major genotype groups (P = 0.003) (Figure 2). The corresponding median values for reduced haloperidol in these genotype groups were 0.6, 1.0, 0.4, 1.5 and 1.5 nm mg⁻¹, respectively, but showed no statistical difference between the three major genotype groups (P = 0.334) (Figure 2).

In patients on doses of 20 mg or more, the median values of the dose-corrected haloperidol concentrations were 2.5, 2.7 and 2.8 nm mg⁻¹ in the *1/*1, *1/*10 and *10/*10 groups respectively, and those of reduced haloperidol were 1.7, 2.1 and 1.8 nm mg⁻¹. No differences were found between the genotype groups either for haloperidol or reduced haloperidol (P values were between 0.667 and 0.677). No patient taking 20 mg or more haloperidol had a *5 allele.

The concentrations were normalised for absolute daily dose, but very similar results were obtained when the dose was expressed as a function of body & weight.

Sixty-eight patients (57%) used benztropine, an antimuscarinic agent, and 16 patients (13%) used benzodiazepines with or without benztropine. Thirty-five% (8/23) of the patients in the *1/*1 group were treated with benztropine, whereas a higher proportion, 62% (37/60) and 58% (19/33) in the *1/*10 and *10/*10 groups, respectively, took this drug. All four patients with a *5 allele were found to use benztropine. Patients homozygous for the *1 allele needed less benztropine than those with one or two mutated alleles (Fisher's exact test; P = 0.036). There was no significant difference in haloperidol doses between the different genotype groups, but there was a tendency towards higher doses in the *1/*1 group (P = 0.051) (data not shown). This result correlates well with the finding that patients homozygous for the *1 allele needed less benztropine than other patients, suggesting less extrapyramidal symptoms.

Discussion

Due to the large interindividual variability in the pharmacokinetics and the clinical outcome of haloperidol treatment, therapeutic drug monitoring has been advocated for optimization of drug dosage. According to the comprehensive review of the concentration-effect studies on haloperidol by Ulrich et al. [29], the maximum therapeutic benefit from haloperidol is obtained at a plasma concentration of about 10 µg l⁻¹ (26.6 nm), with effects decreasing at higher concentrations. A therapeutic range was suggested between about 14.9 and 45.0 nm, and patients with plasma haloperidol concentrations within this range showed significantly better response compared with those outside. The plasma concentrations of haloperidol in our study varied greatly (4.5–230.2 nm), with a median concentration of 47.2 nm, that is slightly outside the suggested therapeutic range.

In the present study, there was a linear relationship between the plasma concentration and the dose of haloperidol. However, the plasma concentration of reduced haloperidol increased disproportionally at high doses. This has been reported earlier [10, 30].

The dose-corrected concentrations of haloperidol differed statistically between the three major genotype groups. However, no difference was found between the different genotype groups with respect to dose-corrected reduced haloperidol concentrations. The other two genotype groups (*1/*5 and *10/*5) were not considered for these comparisons due to the small number of patients. In order to evaluate whether the effect of genotype depends on dose, the data were analysed separately for doses lower and higher than the median (20 mg).

At doses lower than 20 mg, the dose-corrected mean concentrations of haloperidol were significantly different between the three major genotype groups. However, the one-way anova showed no statistical difference in those of reduced haloperidol between the genotype groups (0.6, 1.0 and 1.5 nm mg⁻¹). At higher doses (≥ 20 mg), no differences were found between the genotype groups for either haloperidol or reduced haloperidol. These findings support the observation that CYP2D6 has high affinity and low capacity for its substrates and becomes saturated at relatively low concentrations [31], thus decreasing any pharmacokinetic differences between CYP2D6 genotype groups.

A thorough review by Ulrich et al. [29] showed that the daily doses of haloperidol administered to schizophrenic patients vary from 1 mg to 160 mg, which is very similar to that seen in the present study. Based on an analysis of positron emission tomography (PET) studies, Nyberg & Farde [32] concluded that 3–5 mg haloperidol daily gives a 70–80% dopamine D₂ receptor occupancy. They also claimed that higher doses do not seem to have additional clinical benefit. The doses used in the present study are much higher than those used in previous haloperidol studies in relation to CYP2D6. Llerena et al. [14, 15] showed that plasma concentrations of haloperidol and reduced haloperidol were significantly higher in PM of debrisoquine than in EM, after single oral doses of 2–4 mg haloperidol to healthy Caucasian volunteers. In a PET study by Nyberg et al. [33], a PM of debrisoquine showed higher plasma concentrations of haloperidol and higher dopamine D₂ receptor occupancy than EMs, indicating higher risk of extrapyramidal symptoms. This study was also conducted using relatively low doses of haloperidol decanoate (30–50 mg/4 weeks). These data suggest collectively that CYP2D6 is important for the metabolism of haloperidol at low doses.

It is well known that there are differences between Caucasian and Oriental populations with respect to CYP2D6 activity [22]. The activity of CYP2D6 in PM of debrisoquine is lower than that in Orientals homozygous for the CYP2D6*10 allele. Thus it is not surprising that the difference in concentrations of haloperidol between the *1/*1, *1/*10 and *10/*10 groups is not as great as that between PM and EM of debrisoquine in Caucasians. Suzuki et al. [16] and Mihara et al. [17] studied Japanese schizophrenic patients taking a fixed daily dose of 12 mg haloperidol and found that there was a significant difference in plasma concentrations of haloperidol and reduced haloperidol between patients homozygous for *1 and those heterozygous for *10, whereas no difference was found between patients heterozygous and homozygous for *10. In contrast we did observe a difference between the *1/*10 and *10/*10 genotype groups, but not between the *1/*1 and *1/*10 groups in patients with dose less than 20 mg. The latter finding might have been due to the small number of patients (6) homozygous for the *1 allele. Among the 7 patients assigned to the genotype CYP2D6*1/*1 and in whom the *5 allele was not analysed, on average one patient would have had the CYP2D6*1/*5 genotype.

Recently, the involvement of CYP3A4 in haloperidol metabolism has been reported. Fang et al. [34] showed that CYP3A4 and CYP2D6 metabolise haloperidol through the formation of the pyridinium metabolite and by N-dealkylation. In addition, CYP3A4 also catalyses the oxidation of reduced haloperidol back to haloperidol. Other in vitro studies have also demonstrated the involvement of CYP3A4 in the oxidation of reduced haloperidol to haloperidol, but not that of CYP2D6 [35, 36]. The large overlaps between the genotype groups might be due to the interindividual variation that was demonstrated in CYP2D6 activity within each of the three major genotype groups in the Korean population [25], or involvement of other enzymes, for example CYP3A4.

In conclusion, we found that dose-corrected plasma concentrations of haloperidol were related to CYP2D6 genotype at haloperidol doses lower than 20 mg daily, but not at higher doses in Korean schizophrenic patients. However, although statistically significant in the former, mean differences between *1/*1 and *10/*10 and between *1/*10 and *10/*10 subjects were small and there was considerable overlap between genotype groups.