Population PK and nonlinear mixed-effects (NLME) models
A population PK model describes the typical concentration–time profile and how it varies between and within patients, estimated from the pooled, often sparse data of many patients with a nonlinear mixed-effects model.
Definition
A population PK (PopPK) model describes the drug's typical profile in a patient population together with the variability around it. It is estimated from the pooled data of many patients, which may be sparse for each patient, with a nonlinear mixed-effects (NLME) model: nonlinear because concentrations are nonlinear functions of the parameters, mixed because it has both fixed and random effects.
Parts of the model
- Fixed effects: typical values of the parameters (for example CL and V) and the effects of covariates such as body weight (covariates).
- Random effects: the between-subject variability of the parameters.
- Residual error: the remaining difference between observation and prediction.
Software
Commonly used tools are NONMEM, Monolix and nlmixr2 (R). On this site, nlmixr2 and rxode2 were used in the nimotuzumab and warfarin analyses, and published models were reimplemented from their tables, as for mosunetuzumab.
Frequently asked questions
What is the difference between fixed and random effects?
Fixed effects are the typical parameter values and covariate effects shared by the population; random effects are the deviations of individuals from them, described by a distribution.
Why can a population model use sparse data?
Because the data of all patients are analyzed together, so the typical profile and the variability are informed by the whole population, even if each patient has few samples.