To establish a population pharmacokinetic (PPK) model of duloxetine (DLXT) in hospitalized patients with depression, to investigate its influencing factors, and to provide reference for individualized medication.

A total of 325 patients who were hospitalized and received oral DLXT treatment at Shenzhen Kangning Hospital from January 2022 to October 2025 were retrospectively included, and their medication information, plasma concentrations, and other clinically relevant data were collected. Nonlinear mixed-effects modeling (NONMEM) was used to establish a PPK model of DLXT concentrations. Goodness-of-fit (GOF) plots, visual predictive check (VPC) and Bootstrap were used to evaluate the stability and predictive performance of the model, respectively. Based on the significant covariates identified in the final model, the steady-state plasma concentrations (C_ss) of DLXT over time were simulated under different dosage regimens in various populations.

A one-compartment model with first-order absorption and elimination was developed using 450 plasma concentration data points from 325 inpatients. The inter-individual variability of apparent clearance (CL/F) was 42.2%, and the proportional residual variability was 27.1%. Gender was identified as a significant covariate affecting CL/F. The final model was expressed as: CL/F(L·h^-1)=64.6×(1-0.25×Gender), where Gender is an indicator variable (male=0, female=1). The apparent volume of distribution was 1 530 L, and the absorption rate constant was fixed at 0.168 h^-1. Monte Carlo simulations showed that the C_ss in females was higher than that in males at the same doses.

A PPK model for DLXT in hospitalized patients with depression was successfully established. The final model is stable and reliable, and it can be used for individualized dosing regimens in clinical practice.