Hyun Chul Kim1,2, Jung Sunwoo3, Seonghae Yoon4, In-Jin Jang1, Jae-Yong Chung4
1Department of Clinical Pharmacology and Therapeutics, Seoul National University College of Medicine and Hospital, 2Kidney Research Institute, Seoul National University Medical Research Center, 3Clinical Trials Center, Chungnam National University Hospital, 4Department of Clinical Pharmacology and Therapeutics, Seoul National University College of Medicine and Bundang Hospital
Objectives: Zolpidem is a non-benzodiazepine hypnotic agent used for the short-term treatment of insomnia, characterized by difficulty initiating sleep. According to the Food and Drug Administration (FDA) label for zolpidem, the recommended initial dose is 5 mg for adult women and 5 or 10 mg for adult men, reflecting the higher exposure observed in females compared to males. To better understand sex-related differences in the pharmacokinetics (PKs) and pharmacodynamics (PDs) of zolpidem, we comprehensively analyzed covariate effects in healthy male and female volunteers. Methods: The population PK-PD analysis for zolpidem was performed using a nonlinear mixed-effects modeling approach with NONMEM (version 7.5.1, ICON Development Solutions, Ellicott City, MD, USA). The model was developed using the first-order conditional estimation method with interaction and evaluated using Perl-speaks-NONMEM (PsN; version 5.4.0, Uppsala University, Sweden). Previously published data [1] from 30 volunteers (15 males and 15 females) were used for the analysis, with 325 data points for plasma zolpidem concentrations and 388 data points each for the digital symbol substitution test (DSST) score, the choice reaction test (CRT) time, and the sedation visual analogue scale (VAS) score. For the population PK model, a one-compartment model with first-order elimination and various absorption models was explored. For the concentration–effect relationship, several models linked with either a direct or effect compartment were evaluated. Baseline models for each PD measure were explored to describe pre-dose values. Inter-individual variability (IIV) was evaluated using the exponential error model, except for the baseline VAS values, which were evaluated using an additive variability model with logit-transformed VAS. Additive, proportional, and combined error models were explored to describe random residual variability. Potential covariates were evaluated through stepwise covariate modeling at significance levels of 0.05 (forward selection) and 0.01 (backward elimination). Candidate covariates included sex, age, weight, albumin, alanine aminotransferase (ALT), aspartate aminotransferase (AST), total bilirubin, creatinine, and each baseline PD measure at 1h. Results: A one-compartment model with first-order elimination and a transit compartment absorption model adequately described the PK profiles of zolpidem. The typical values (with % coefficient of variation of IIV, %CV) for the absorption rate constant, apparent clearance, apparent volume of distribution, mean transit time, number of transit compartments were 11.7 h-1 (330.8%), 18 L/h (24.1%), 64 L, 0.25 h (37.7%), and 19.4, respectively with relative bioavailability fixed at 1 (25.9%). Each PD measure was sequentially linked to the final fixed PK parameters. The PD measures were directly linked to the PK model using the sigmoid Emax model, with a two-parameter hyperbolic learning curve for the PK-DSST model and with a linear spline baseline model [2] for the PK-VAS model. The typical values (with %CV) for the concentration at the half-maximum inhibition (IC50) for the DSST score and the concentration at half-maximum response (EC50) for the CRT time and VAS score were 205 µg/L (21.8%), 282 µg/L (21.1%), and 146 µg/L (53.9%), respectively, with the corresponding Hill factors of 2.54, 6, and 1.92. Significant covariates for the DSST score were weight and albumin for IC50, and age for the baseline DSST score. For CRT time, weight, albumin, and baseline CRT time at 1h were significant covariates for EC50. Additionally, age was significant for the baseline CRT time, and albumin and baseline CRT time at 1h were significant for the Hill factor. No covariates were significantly associated with the PK parameters or the VAS score. Conclusion: Zolpidem exhibited increased potency with higher weight and lower albumin, while older age was associated with lower baseline DSST scores and higher baseline CRT times, leading to reduced response. Since weight and albumin differ significantly between males and females in the dataset, these covariates might contribute to observed differences between sexes. However, the proposed population PK-PD models suggest that weight, albumin, age, and baseline PD measures, rather than sex, are the primary determinants of zolpidem’s pharmacodynamic response.
[1] Yoon, S., Jeong, S., Jung, E. et al. Effect of CYP3A4 metabolism on sex differences in the pharmacokinetics and pharmacodynamics of zolpidem. Sci Rep 11, 19150 (2021). https://doi.org/10.1038/s41598-021-98689-z [2] Trocóniz, I.F., Boland, K. & Staab, A. Population Pharmacokinetic/Pharmacodynamic Model for the Sedative Effects of Flibanserin in Healthy Volunteers. Pharm Res 29, 1518–1529 (2012). https://doi.org/10.1007/s11095-011-0648-6
Reference: PAGE 33 (2025) Abstr 11505 [www.page-meeting.org/?abstract=11505]
Poster: Drug/Disease Modelling - CNS