Modelling approaches in pharmacokinetics Flashcards
Pop-PK features and differences from classical PK
Looks at the variability in drug concentrations within a population using statistical methods. Can identify potential causes for variation in large clinical trial populations that occur along different populations, e.g., age.
Then used as framework for optimising dosing strategies for subpopulations that have significant variability in plasma concentrations.
Pop-PK often is more unstructured with sparser sampling than classical PK. Few samples, but from many patients. it uses the target patient population. much more focus on PK/PD than in classical PK studies (which tend to be in small studies with restrictive criteria to prevent inter-individual variability)
The statistical methods in pop-PK quantify variability, whereas in classical, variability is not considered.
Types of variability
Between subject variability (BSV) and between occasion variability (BOV). between occasion would be in the same patient.
Mixed effect modelling and its parameters
Trials are often implemented in a way in which sampling would be sparser than ideal, e.g., missing time points.
In individual PK modelling, epsilon (in the equation) is the measurement of error, but it is only the difference in observed and predicted. In pop-PK, the model accounts for magnitude, and sometimes even the source of variability between individuals.
pop-PK is implemented using non-linear mixed effect (NLME) modelling. it includes all the demographic and pharmacological data for all subjects. provides an estimate in variation between PK and population PK parameters. it also estimates the residual variation that cannot be explained with the modelling. This model is repeatedly applied until the estimation no longer changes (this is minimisation).
Theta in NMLE is the population estimate of the PK parameters
Eta is the explanation for intra- and inter-individual variability
Epsilon is the unexplained residual variability
In NMLE, the structural model (which compartment model), the covariate model (the effect of individual variables such as weight on PK), and statistical models (of BSV and BOV) should be defined
what is PBPK, and example of its use
also known as mechanistic modelling - first proposed in 1937. It integrates a range of preclinical and physiochemical data to predict in vivo PK - uses LogP, pKa, solubility, permeability, PPB, and in vivo animal data.
It has physiological meaning that can be extrapolated to different patients. Has detailed descriptions of all the mechanisms that influence the PK of a compound. contains all compartments - full range of tissue and organs - 14 compartments. can predict changes in [drug] in each tissue.
Follows the law of mass balance. the cardiac output will equal the total blood flow through each compartment. needs tissue data as well. e.g., organ perfusion, organ volume, GFR, wieght+composition, enzyme and transporter levels.
PBPK was used in dosing regimes for chloroquine for the prevention of transferring Zika to offspring in pregnant women. the model predicted the SD and plasma concentrations. proved a useful guide for dosing strategies.
