Chapter 6: Model-informed drug development
In a challenge study, healthy subjects will get infected with the RSV virus. After this, they will receive
the drug. These patients are monitored and their viral load is checked continuously. These data are
modelled and used for simulations for children.
PopPK model
A population PK model is made out of the results of such a challenge study. This is the data we have:
Placebo (N=12); 750 mg LD then 500 mg q12h MD (N=8); 750 mg LD then 150 mg q12h MD (N=7);
and 375 mg q12h over 5 days (no LD – N=8). So, we have 1 placebo group and three groups who will
receive different doses. Based on the modeling results (the real time analysis), the decision is made
whether or not we continue with the trial; or if we need more subjects; etc.. So, we use plasma
samples for the determination of the active
drug (ALS-8112) and the inactive, uridine
metabolite (ALS-8144). The concentrations
were obtained at 18 timepoints after
initiation of dosing. The drug plasma
concentrations were measured using a
validated LC-MS/MKS assay that detects
ALS-8112 and ALS-8144 with LLoQs 2 ng/mL
and 1 ng/mL, respectively. All these
concentrations, so the whole dataset, is
received in Excel. We now need to model
the data. As you can see in the model, all
ALS-numbers are different. This is because
you go from the prodrug to the active
metabolite in the body. You also have two
different absorption constants (Kas and Kaf: slow and fast). The software cannot read those diagrams,
but only numbers and equations. So, there are a lot of equations made which will tell you how the
drug will move in the body and how it will act inside of the body. We not only have typical values (for
the typical patients), but also interindividual
variabilities (the variance in between the
patients) and the residual errors. In the
graph on the right, you see a representation
of the model, overlayed with the data.
There is a lot of variability, but the model
will take this into account. If you use the
pcVPCs of the population PK model, then
we see that the typical trend is well
captured. The model can predict the real-
world very well! Now, we have a model, and
now we can simulate. We can perform
virtual trials!
RSV kinetics and popPD model
This is about the population
pharmacodynamics model. The
pharmacodynamics is the
relationship between the exposure
and the response. Because we are
not interested in the exposure, we
are only interested in the response of
the human body. This is different
Pagina 1 van 7
In a challenge study, healthy subjects will get infected with the RSV virus. After this, they will receive
the drug. These patients are monitored and their viral load is checked continuously. These data are
modelled and used for simulations for children.
PopPK model
A population PK model is made out of the results of such a challenge study. This is the data we have:
Placebo (N=12); 750 mg LD then 500 mg q12h MD (N=8); 750 mg LD then 150 mg q12h MD (N=7);
and 375 mg q12h over 5 days (no LD – N=8). So, we have 1 placebo group and three groups who will
receive different doses. Based on the modeling results (the real time analysis), the decision is made
whether or not we continue with the trial; or if we need more subjects; etc.. So, we use plasma
samples for the determination of the active
drug (ALS-8112) and the inactive, uridine
metabolite (ALS-8144). The concentrations
were obtained at 18 timepoints after
initiation of dosing. The drug plasma
concentrations were measured using a
validated LC-MS/MKS assay that detects
ALS-8112 and ALS-8144 with LLoQs 2 ng/mL
and 1 ng/mL, respectively. All these
concentrations, so the whole dataset, is
received in Excel. We now need to model
the data. As you can see in the model, all
ALS-numbers are different. This is because
you go from the prodrug to the active
metabolite in the body. You also have two
different absorption constants (Kas and Kaf: slow and fast). The software cannot read those diagrams,
but only numbers and equations. So, there are a lot of equations made which will tell you how the
drug will move in the body and how it will act inside of the body. We not only have typical values (for
the typical patients), but also interindividual
variabilities (the variance in between the
patients) and the residual errors. In the
graph on the right, you see a representation
of the model, overlayed with the data.
There is a lot of variability, but the model
will take this into account. If you use the
pcVPCs of the population PK model, then
we see that the typical trend is well
captured. The model can predict the real-
world very well! Now, we have a model, and
now we can simulate. We can perform
virtual trials!
RSV kinetics and popPD model
This is about the population
pharmacodynamics model. The
pharmacodynamics is the
relationship between the exposure
and the response. Because we are
not interested in the exposure, we
are only interested in the response of
the human body. This is different
Pagina 1 van 7
