Serum concentrations of sertraline and N-desmethyl sertraline in relation to CYP2C19 genotype in psychiatric patients
- First Online:
- Cite this article as:
- Rudberg, I., Hermann, M., Refsum, H. et al. Eur J Clin Pharmacol (2008) 64: 1181. doi:10.1007/s00228-008-0533-3
To investigate the impact of CYP2C19 genotype on serum concentrations of sertraline and N-desmethyl sertraline in psychiatric patients.
Patients treated with sertraline (n = 121) were divided into six subgroups according to CYP2C19 genotype: CYP2C19*17/*17, CYP2C19*1/*17, CYP2C19*1/*1, CYP2C19*17/def, CYP2C19*1/def and CYP2C19def/def (def = allele encoding defective CYP2C19 metabolism, i.e. *2 and *3). Dose-adjusted serum concentrations were compared by linear mixed model analyses using the CYP2C19*1/*1 subgroup as reference.
Subgroups carrying one or two alleles encoding defective CYP2C19 metabolism achieved significantly higher mean dose-adjusted serum concentrations of sertraline and N-desmethyl sertraline compared to the CYP2C19*1/*1 subgroup (P < 0.05). The effect of CYP2C19 genotype was expressed as 3.2-fold (sertraline) and 4.5-fold (N-desmethyl sertraline) higher dose-adjusted serum concentrations in the CYP2C19def/def subgroup compared to the CYP2C19*1/*1 subgroup (P < 0.01). The CYP2C19*17 allele had no influence on the dose-adjusted serum concentrations of sertraline and N-desmethyl sertraline.
The significantly higher serum concentrations associated with alleles encoding defective CYP2C19 metabolism might be of relevance for the clinical outcome of sertraline treatment.
Individual differences in drug metabolism are a major contributor to variability in drug response . Cytochrome P450 2C19 (CYP2C19) is an important drug-metabolising enzyme, and factors such as age, use of interacting drugs and gender are potential contributors to variability in activity of the CYP2C19 enzyme [2, 3]. However, the interindividual variability in CYP2C19 activity is mainly determined by the genetic polymorphism of this enzyme . Seven variant alleles (CYP2C19*2-*8) which encode non-functional CYP2C19 protein have been described . Carriers of two alleles encoding defective CYP2C19 metabolism are poor metabolisers (PMs) of CYP2C19 substrates. Additionally, a recent study reported that a novel variant allele, CYP2C19*17, was associated with increased CYP2C19 metabolism due to increased gene transcription .
CYP2C19 is involved in the metabolism of several antidepressants. Compared to CYP2C19 extensive metabolisers (EMs), CYP2C19 PMs are reported to achieve higher systemic exposure of imipramine [6, 7], trimipramine , clomipramine , amitriptyline , citalopram , escitalopram , sertraline , fluoxetine  and moclobemide [15, 16]. So far, however, little is known about the impact of CYP2C19*17 on the pharmacokinetics of antidepressants. In a recent study, we investigated serum concentrations of escitalopram in relation to CYP2C19 genotype among psychiatric patients. This showed a significantly lower mean serum concentration of escitalopram among homozygous carriers of the CYP2C19*17 allele compared to homozygous CYP2C19 EMs .
Sertraline is a frequently used antidepressant, and it accounted for about 20% of the daily doses of selective serotonin reuptake inhibitors (SSRIs) sold in Norway in 2007 . Metabolism of sertraline is catalysed by multiple CYP enzymes, and at least five different isoenzymes, including CYP2B6, CYP2C9, CYP2C19, CYP2D6 and CYP3A4, are involved in the formation of the primary metabolite N-desmethyl sertraline in vitro [18–20]. Although N-desmethyl sertraline exhibits less activity compared to the parent drug, the metabolite is usually included in therapeutic drug monitoring of sertraline [21–24]. A study by Wang et al. showed that after a single oral dose of sertraline administered to healthy volunteers, CYP2C19 PMs had significantly higher area under the concentration versus time curves (AUC) of sertraline (41%) and lower AUC of the N-desmethyl sertraline (35%), compared to a mixed group of heterozygous and homozygous CYP2C19 EMs . However, the impact of CYP2C19 genotype on pharmacokinetics of sertraline and N-desmethyl sertraline at steady state has not been studied. Moreover, the influence of the CYP2C19*17 allele on sertraline metabolism is unknown. Therefore, the objective of this study was to investigate the impact of genetic variation in CYP2C19, including the novel CYP2C19*17 allele, on the serum concentrations of sertraline and N-desmethyl sertraline in psychiatric patients.
The therapeutic drug monitoring database at the Department of Psychopharmacology, Diakonhjemmet Hospital, Oslo, Norway, was used to identify all patients with CYP genotype and serum concentrations of sertraline and N-desmethyl sertraline determined between January 2001 and December 2007. Serum concentration measurements were excluded if the time between drug intake and blood sampling was <10 h or >30 h (or time not specified). Samples were also excluded if information on recent initiation/dose-adjustment of the sertraline treatment (<5 days) or concurrent use of drugs known to interact with CYP2C19, CYP2C9, CYP2B6, CYP2D6 or CYP3A4 was given on the requisition form. It was searched for drugs listed as inducers or strong/moderate inhibitors according to http://medicine.iupui.edu/flockhart/table.htm  or inducers/inhibitors according to www.CYP450.no . Information on patient age and gender was obtained from the requisition forms (diagnosis not available). Ethnicity was not confirmed, but the study population is likely to consist mainly of Caucasians. Genotype results of CYP2C19 (only alleles encoding defective metabolism), CYP2C9, CYP2D6 and serum concentrations of sertraline and N-desmethyl sertraline were taken from the therapeutic drug monitoring files. Stored DNA samples were reanalysed to detect the CYP2C19*17 allele. The study was approved by the Regional Committee for Medical Research Ethics, Oslo, Norway.
The study population was divided into six subgroups according to CYP2C19 genotype: CYP2C19*17/*17, CYP2C19*1/*17, CYP2C19*1/*1, CYP2C19*17/def, CYP2C19*1/def and CYP2C19def/def (def = allele encoding defective CYP2C19 metabolism). To account for differences in drug dosage, all serum concentrations were dose-adjusted (i.e. nM/mg). In the following text, dose-adjusted serum concentrations are referred to as “serum concentrations”. The primary endpoint of the study was comparison of serum concentrations of sertraline and N-desmethyl sertraline between the CYP2C19*1/*1 subgroup and the subgroups containing CYP2C19 variant alleles. Secondarily, the relationship between serum concentrations of sertraline and N-desmethyl sertraline in individual serum samples was evaluated.
The routine genotyping procedure included detection of the single nucleotide polymorphisms (SNPs) specific to CYP2C9*2, *3 and *5; CYP2C19*2, *3, *4 and *5; and CYP2D6*3, *4, *6, *7, and *8. Additionally, CYP2D6 was analysed for gene deletion (*5) and gene multiplication [27, 28]. About one-half of the samples were analysed by polymerase chain reaction (PCR) followed by restriction fragment length polymorphism analysis or allele-specific PCR analysis, whereas the rest of the samples were analysed by TaqMan-based real-time PCR methods. Cross-validation of the two methods showed 100% agreement. The genotyping assays did not discriminate between functional and non-functional CYP2D6 multiplications.
DNA samples were reanalysed to detect the CYP2C19*17 allele. The CYP2C19*17 allele is characterised by two SNPs, -3402C>T and -806C>T, which have been reported to be in complete linkage with each other . Thus, analysis of one of the SNPs is sufficient to detect the CYP2C19*17 allele . A commercially available TaqMan-based real-time PCR kit (Assay ID C___469857_10, Applied Biosystems, Foster City, CA) was used for detection of the -806C>T SNP. Absence of the listed mutations was interpreted as presence of the *1-allele encoding functional enzyme activity (CYP2C9*1, CYP2C19*1 or CYP2D6*1).
Serum analyses of drug and metabolite
Reference substances of sertraline and N-desmethyl sertraline were provided by Pfizer (New York, USA). The internal standard protriptyline was purchased from Sigma-Aldrich (Oslo, Norway). Serum concentrations of sertraline and N-desmethyl sertraline were analysed by liquid chromatography-mass spectrometry assays developed for routine therapeutic drug monitoring analysis. Serum samples were purified by protein precipitation (0.5 ml serum and 1.0 ml acetonitrile-methanol 90/10 with 1.0 µM internal standard, centrifuged for 10 min at 1,800 g, 4°C).
In the period from which serum concentration measurements were included (2001–2007), three different analytical assays were used: HPLC-MS, HPLC-MS/MS and UPLC-MS/MS (new methods cross-validated with previous methods). In the current routine method, 5 µl of the supernatant was injected into an Acquity UPLC with Micromass Quattro micro tandem MS detector (both Waters, Milford, MA, USA). Separation was performed on an Acquity UPLC BEH Shield RP18 column (1.7 µm, 1.0 × 100 mm) (Waters, Milford, MA, USA) by gradient elution (18–45% acetonitrile in 10 mM ammonium acetate buffer, pH = 4.8). The retention times were 3.35 min for sertraline, 3.30 min for N-desmethyl sertraline and 2.90 min for protriptyline. The total run time was 5 min. Detection was performed by multiple reaction monitoring in the positive electrospray ionization mode at the following m/z transitions: 306→275 (sertraline), 292→159 (N-desmethyl sertraline) and 264→233 (protriptyline). Calibration curves ranged from 20–300 nM for sertraline and 20–600 nM for N-desmethyl sertraline and were linear for both substances (R2 values > 0.998, n = 5). Quality control samples in the upper and lower range of the calibration curve showed intra- and interday accuracy and precision < 5% for both substances (n = 5). Limit of quantification (LOQ), defined as a signal-to-noise ratio of 10, was 3 nM for sertraline and 6 nM for N-desmethyl sertraline. Accuracy and precision were <20% for both compounds at LOQ (n = 5). All serum concentrations measured were above LOQ. Serum concentrations above the upper value of the calibration curve (about 5% of the samples) were estimated by extrapolation of the linear calibration curves.
Statistics and measures
The serum concentrations were not normally distributed and were therefore logarithmically transformed prior to the statistical analyses. To account for multiple serum samples from some individuals, linear mixed model analyses were used. Age ≥ 70 years, gender and presence of variant alleles encoding impaired CYP2D6 and CYP2C9 enzyme activity were included as potential covariates when analysing the effect of CYP2C19 genotype on serum concentrations of sertraline and N-desmethyl sertraline. Statistically significant covariates were included in the final models. To describe the effect of CYP2C19 genotype and covariates, parameter estimates were transformed back to the original scale. Data are therefore presented as geometric means with 95% confidence intervals (CIs). The relationship between serum concentrations of sertraline and N-desmethyl sertraline in individual samples was evaluated by linear mixed model regression analyses. Total range and the 95/5 percentile ratio of individual serum concentrations were used to describe variability in the data material. Estimates of covariance parameters were used to describe intraindividual variability.
Age, drug dosage and time between drug intake and blood sampling were compared between the CYP2C19*1/*1 subgroup and the various subgroups with variant alleles by a two-tailed Mann-Whitney test. Group comparisons of gender and allele frequencies were performed by the Fischer’s exact test. Estimations of CIs for population frequencies of CYP2C19 alleles were calculated as described by Altman . Statistical significance was considered as P < 0.05. Mixed model analyses and Mann-Whitney tests were performed by SPSS 15.0 (SPSS, Chicago, IL). GraphPad Prism 4 (GraphPad Software, San Diego, CA) was used for the Fischer’s exact tests and graphical presentations.
Characteristics of the six study subgroups
No. of subjects (males/females)
No. of serum samples
Age (years) (mean ± SD)
55 ± 32
43 ± 20
51 ± 24
43 ± 18
47 ± 23
33 ± 18
Time between drug intake and blood sampling (h) (median)
Drug dosage (mg/day) (median and rangea)
CYP2C9 allele frequencies
CYP2C9*2 and *3
CYP2D6 allele frequencies
CYP2D6 *3, *4, *5, and *6
CYP2D6 gene duplication
Serum concentrations of sertraline and N-desmethyl sertraline in different CYP2C19 genotype subgroups
Serum concentration (95% CI)
Serum concentration (95% CI)
Twenty-four subjects (19.8%) were classified as elderly (age ≥ 70 years). Age ≥ 70 years was significant as a covariate when analysing the effect of CYP2C19 genotype on mean serum concentrations of sertraline and N-desmethyl sertraline and was therefore included in the final models used. The effect of age ≥70 years was expressed as 1.8-fold (95% CI 1.3–2.6) and 2.0-fold (95% CI 1.5–2.7) higher serum concentrations of sertraline and N-desmethyl sertraline respectively compared to younger patients <70 years (P < 0.001). Gender and presence of variant alleles encoding impaired CYP2D6 and CYP2C9 enzyme activity were not significant as covariates (P > 0.3). Of the unexplained variance in the linear mixed model analyses, 56 and 57% were due to intraindividual variability for sertraline and N-desmethyl sertraline respectively.
Frequencies of CYP2C19 alleles
Frequency (95% CI)
All subgroups carrying alleles encoding defective CYP2C19 metabolism achieved significantly higher mean serum concentrations of sertraline compared to the CYP2C19*1/*1 subgroup. The effect of CYP2C19 genotype was expressed as a 3.2-fold higher mean serum concentration of sertraline in the CYP2C19def/def subgroup compared to the CYP2C19*1/*1 subgroup. In a recent study, mean serum concentration of escitalopram ranged six-fold between the CYP2C19*1/*1 and CYP2C19def/def subgroups . Accordingly, genetic variation in CYP2C19 seems to be less important for serum concentration of sertraline than for escitalopram. Nevertheless, the 3.2-fold difference in serum concentration between CYP2C19 genotype subgroups might be of importance for the clinical outcome of sertraline treatment, either in terms of improved antidepressive effect or increased occurrence of dose-dependent side-effects.
Little is known regarding the relationship between systemic exposure and clinical response of sertraline treatment. A study by Mauri and co-workers which included patients with major depression treated with sertraline reported no significant difference in trough plasma concentrations between responders and non-responders or between patients with and without side-effects . However, the low number of patients (n = 21) was not sufficient to draw firm conclusions. Regarding another SSRI, it could be mentioned that a recent study reported a significant concentration-response relationship for fluvoxamine in a subpopulation of moderately to severely depressed patients . It might be reasonable to assume that the existence of a relationship between systemic exposure and clinical effect is a shared characteristic for the SSRIs, and further studies are needed to investigate this during sertraline treatment.
Mean serum concentrations of the metabolite N-desmethyl sertraline were also significantly higher in all subgroups with alleles encoding defective CYP2C19 metabolism. This indicates that CYP2C19 is involved in the further metabolism of N-desmethyl sertraline. The CYP2C19def/def subgroup achieved a 4.5-fold higher serum concentration of N-desmethyl sertraline compared to the CYP2C19*1/*1 subgroup. In contrast, Wang et al. reported a significantly lower AUC (35%) of N-desmethyl sertraline in CYP2C19 PMs compared to carriers of the CYP2C19*1 allele after a single oral dose (100 mg) in healthy volunteers . Another discrepancy between the two studies is a more pronounced impact of CYP2C19 genotype on the parent compound in the present study than observed by Wang et al. (3.2-fold vs. 1.4-fold higher systemic exposure respectively) . One possible explanation for the observed discrepancies between the studies may be differences in pharmacokinetics following a single drug dosage compared to the steady-state conditions. Furthermore, differences in the reference groups might have contributed to the diverging results. Three out of six subjects classified as EMs in the study of Wang et al. were heterozygous for a CYP2C19 allele encoding defective metabolism, while the present study used patients homozygous for the CYP2C19*1 allele as reference group.
In vitro studies have reported that the potency of N-desmethyl sertraline to inhibit serotonin reuptake is about 2-13% compared to the parent drug [22–24]. This indicates a limited contribution from N-desmethyl sertraline to inhibition of serotonin reuptake during sertraline treatment. In addition to inhibition of serotonin reuptake, both sertraline and N-desmethyl sertraline display inhibition of noradrenalin and dopamine uptake in vitro. Regarding these effects, N-desmethyl sertraline shows a potency of 30–100% compared to the parent drug [22–24]. Although inhibition of noradrenalin and dopamine uptake is not considered as the main target of sertraline, these effects might contribute to clinical response of sertraline treatment. Moreover, the human brain distribution and plasma protein binding of N-desmethyl sertraline relative to sertraline are unknown. Thus, it is difficult to exclude a clinical impact of the differences in serum concentrations of N-desmethyl sertraline between CYP2C19 genotype subgroups.
The effect of CYP2C19 genotype on serum concentrations of sertraline and N-desmethyl sertraline was only associated with the presence of alleles encoding defective CYP2C19 metabolism. Lack of a detectable effect of the CYP2C19*17 allele indicates that the *17 allele is of quantitatively less importance for the CYP2C19 phenotype than the alleles encoding defective metabolism. This is in accordance with studies with other CYP2C19 substrates (escitalopram and omeprazole). Compared to CYP2C19*1/*1, homozygous carriers of CYP2C19*17 had a 42% lower serum concentration of escitalopram, while the CYP2C19def/def subgroup had a six-fold higher serum concentration . For the proton-pump inhibitor omeprazole, the CYP2C19*17/*17 genotype was associated with a 52% lower AUC , while CYP2C19 PMs have been reported to obtain 5- to 15-fold higher AUC, compared to carriers of the CYP2C19*1 allele .
The variability in serum concentrations in our material was substantial, with a 95/5 percentile ratio of 10.0 for sertraline and 10.9 for N-desmethyl sertraline. In the whole study population, there was a significant positive correlation between the serum concentrations of sertraline and N-desmethyl sertraline in each individual sample, which suggests that the parent drug and metabolite share a common metabolic pathway. The increased serum concentration of both sertraline and N-desmethyl sertraline associated with alleles encoding defective CYP2C19 metabolism indicates that CYP2C19 is involved in the metabolism of both substances. Moreover, analysis of the CYP2C19*1/*1 subgroup alone showed a correlation between the serum concentration of the parent drug and metabolite similar to that observed for the whole study population. This may indicate that factors other than CYP2C19 genotype influence the pharmacokinetics of the two substances similarly. Non-genetic variation in CYP2C19 phenotype may contribute to the observed variation in serum concentrations of sertraline and N-desmethyl sertraline. For instance, CYP2C19 enzyme activity seems to decrease with age [3, 35], and this is one possible explanation for the higher serum concentrations among patients aged ≥ 70 years observed in the present study. Furthermore, variable activity of other enzymes possibly involved in the metabolism of sertraline and N-desmethyl sertraline may contribute to variability in serum concentrations. In vitro studies have showed that sertraline is a substrate for CYP2B6, CYP2C9, CYP2D6 and CYP3A4 in addition to CYP2C19 [18–20]. According to the negative findings on CYP2D6 and CYP2C9 genotype studies in vivo , CYP3A4 and CYP2B6 seem to be of greater relevance for further studies.
The allele frequencies of CYP2C19*2, *3 and *17 in the present study were in accordance with previous studies of Caucasian populations [5, 12, 37, 38]. However, the prevalence of CYP2C19 variant alleles shows marked interethnic differences. About 1–6% of Caucasians are CYP2C19 PMs . In comparison, alleles encoding defective CYP2C19 metabolism are more frequent in Asian populations, where 12–23% are CYP2C19 PMs . Hence, the impact of the increased serum concentrations of sertraline and N-desmethyl sertraline associated with CYP2C19 alleles encoding defective metabolism may be of greater relevance for the use of sertraline in Asian populations.
Use of data from therapeutic drug monitoring is associated with some methodological limitations. These include for instance lack of compliance control, reliability of the information given on the requisition form and incomplete information regarding co-medication and recent dose adjustments. However, in comparison to conventional pharmacokinetic studies, population biobanks often include large amounts of data and reflect real-world patients in a clinical treatment setting. The results of such studies are therefore valuable in the translation of basic pharmacogenetic science into practical applications in clinical everyday practice .
In summary, presence of alleles encoding defective CYP2C19 metabolism (*2 and *3) had a significant influence on the serum concentrations of sertraline and N-desmethyl sertraline in psychiatric patients. On the contrary, CYP2C19*17 had no impact on the serum concentrations of sertraline and N-desmethyl sertraline. The effect of the alleles encoding defective CYP2C19 metabolism was expressed as significantly higher mean serum concentrations of both compounds in individuals carrying one or two alleles encoding defective CYP2C19 metabolism. The differences in systemic exposure between CYP2C19 genotypes might be of relevance for the clinical response of sertraline treatment, either in terms of improved antidepressive effect or increased occurrence of dose-dependent side effects.
The authors thank Marianne Hjerpset, Lene Kristin Støten and Linda Uthus for performing the serum analyses and the genotyping.