Population pharmacokinetics model for escitalopram in Chinese psychiatric patients: effect of CYP2C19 and age

Abstract

Objective: To establish a population pharmacokinetic model in Chinese psychiatric patients to characterize escitalopram pharmacokinetic profile to identify factors influencing drug exposure, and through simulation to compare the results with the established therapeutic reference range. Methods: Demographic information, dosing regimen, CYP2C19 genotype, concomitant medications, and liver and kidney function indicators were retrospectively collected for inpatients taking escitalopram with therapeutic drug monitoring from 2018 to 2021. Nonlinear mixed-effects modeling was used to model the pharmacokinetic characteristics of escitalopram. Goodness-of-fit plots, bootstrapping, and normalized prediction distribution errors were used to evaluate the model. Simulation for different dosing regimens was based on the final estimations. Results: The study comprised 106 patients and 337 measurements of serum sample. A structural model with one compartment with first-order absorption and elimination described the data adequately. The population-estimated apparent volume of distribution and apparent clearance were 815 and 16.3 L/h, respectively. Age and CYP2C19 phenotype had a significant effect on the apparent clearance (CL/F). CL/F of escitalopram decreased with increased age, and CL/F of poor metabolizer patients was significantly lower than in extensive and immediate metabolizer patients. The final model-based simulation showed that the daily dose of adolescents with poor metabolizer might be as high as 15 mg or 20 mg and referring to the therapeutic range for adults may result in overdose and a high risk of adverse effects in older patients. Conclusion: A population pharmacokinetics model of escitalopram was successfully created for the Chinese population. Depending on the age of the patients, CYP2C19 genotype and serum drug concentrations throughout treatment are required for adequate individualization of dosing regimens. When developing a regimen for older patients, especially those who are poor metabolizers, vigilance is required.

Keywords: CYP2C19 genotype; adolescent; elderly; escitalopam; population pharmacokinetics.

FIGURE 1

(A) CL/F and (B) DRC of escitalopram with different CYP2C19 phenotype.

FIGURE 2

Goodness-of-fit plots (A) Population predicted concentration (PRED) versus observed concentrations; (B) individual predicted concentration (IPRED) versus observed concentrations; (C) population predicted concentrations versus conditional weighted residuals (CWRES); and (D) time-after last dose versus CWRES.

FIGURE 3

NPDE metrics for the PopPK model of escitalopram. The mean of normalized prediction distribution errors (NPDE) was 0.02359, variance was 0.9894, skewness was 0.04414, and kurtosis was 0.3027. The results of t-test and Fisher variance test were 0.664 and 0.911, respectively. The statistical values Shapiro-Wilk (SW) test for normality was 0.0633, and the global adjusted p-value was 0.19.

FIGURE 4

Simulated concentrations for ages ≥18 and <65 years in (A) extensive metabolizers; (B) immediate metabolizers, and (C) poor metabolizers at different daily doses. The red dash lines represented 15 ng/ml, and the red solid lines represented 80 ng/ml.

FIGURE 5

Simulated concentrations in EM, IM, and PM older patients (65 years old) at (A) 5 mg/day, (B) 10 mg/day, (C) 15 mg/day, and (D) 20 mg/day. The red dash lines represented 15 ng/ml and the red solid lines represented 80 ng/ml.

FIGURE 6

Simulated concentrations in EM, IM, and PM adolescents (16 years old) at (A) 5 mg/day, (B) 10 mg/day, (C) 15 mg/day, and (D) 20 mg/day. The red dash lines represented 15 ng/ml and the red solid lines represented 80 ng/ml.