Pharmacokinetic-pharmacodynamic (PK-PD) modeling captures the effects of a drug into a mathematical model with the intent to better understand, quantify, interpolate and extrapolate its pharmacology. This is a rather complete and dense description, and we will break it down into better digestable pieces. But let us start with the unique selling point of PK-PD: It allows us to accelerate decision making and design better experiments, ultimately saving money and resources.
In this blog, these properties will be discussed one by one, illustrating the methods and powers of PK-PD.
What is pharmacokinetics about?
Pharmacokinetic literally means the kinetic properties of a drug (pharmacon in Ancient Greek means ‘poison’). It is used in a narrow sense, when we talk about how much a body gets exposed to a drug and how long this exposure is. The exposure typically is changing all the time; the word kinetics refers to these changes (kinesis in Ancient Greek means movement). A tablet for example gets swallowed and slowly appears in blood as measurable concentrations. More and more gets absorbed and a maximum or peak in exposure occurs. The drug thereafter is broken down by the body and/or excreted. The exposure finally becomes so small that it cannot be detected in blood anymore. pharmacokinetics is the study of how fast these steps go and how extensive the exposure is. In short: pharmacokinetics is about what the body does with the drug.
What is pharmacodynamics about?
Drugs are designed to reach a target in the body. They change the behavior of the target. These changes at or close to the target typically happen rather fast. How much changes, depends on the exposure of the target to the drug. Above we just discussed how the exposure to a drug always changes with time. One plus is one is two – and this means that the target-related changes also change with time. Some of these changes can be measured, e.g. with biomarkers. The changes in these biomarkers depend on the drug, and vary with time. The collective of the extent and time-dependency of these changes is called pharmacodynamics, again from pharmacon and the plain english dynamics. In short: pharmacodynamics is about what the drug does to the body
So, what than is a PK-PD model?
The time course of how a body is exposed to a drug can be captured by a mathematical model. These models are not freely chosen, but follow principles from physiology and biology of relevant processes in the body. A limited amount of different pharmacokinetic models can adequately describe exposure to the fast majority drugs! Once a PK model is developed, a pharmacodynamic (PD) model can be build on top of it. PD models describe the changes in the body as response to a drug. These changes are -again- governed by general physiology and biology, and -again- a limited number of models is capable to describe the majority of PD behaviors. Together the two type of models form a PK-PD model.
The crucial point here is that PK-PD models describe experimental data, but in such a way that known physiology and biology is followed. We know that PK-PD models adequately describe the behavior of hundreds of drugs, and that they are all consistent with core mechanisms. Pk-PD models can therefore be used, and they are, to better understand your drug behavior, but also to inter- and extrapolate. Mostly because we understand the principles behind their behavior.
How to apply PK-PD models in decision making?
The discovery and development of a drug involved various stages of uncertainty, or unknown properties of drugs. Each stage the type of experiments and studies change. PK-PD models allow to plug in values for a different stage into a data-driven model and thereby predict what will be going on in the next phase. Such model predictions are than ran, and presented in a way supporting decision making. Depending on how certain the PK-PD predictions are, but also how clear the predictions stand out from the expectations, decisions on if and how to further invest in a drug can be made before the next experiment or study is performed. This saves of course a lot of time, money and internal resources. Money and resources that of course are better spent on other molecules or projects.