T>MIC at SS to find dosage for Amoxicillin vs. E.coli with 90% PTA

Phoenix Certara Forums Modelling and Simulation
1

Hi All,

I am trying to optimize dosage of AMX against E.coli in animals. I followed these steps:

1- I analyzed the concentration-time data of AMX after IV and oral dosing and I found the best fit with two compartment model.

2- I generated a meta -population of 1000 drug disposition curves by MCS with different dosing regimen from 72 to 96 h post administration (steady state), with a step of 0.25 h.

3- I took PRED concentration table to NCA to compute the T>MIC.

4- The PKPD index for AMX vs. E.coli is 57% corresponding to 14h. It means that the concentration of AMX should be at least 14 h over MIC (8mg/l) during dosing interval (24h) at steady state in 90% of animals.

5- I took the data from NCA to phoenix statistical tool to find the 90% percentile (PTA).

The problem is:

The actual time interval is from 72 to 96 h, which is 24 h. but NCA gives me results of 0 to 96h.

For example:

mean

Time high 18

Time between 55

Time low 23

This is not correct because I simulated the interval of 72-96 and the total mean values of time over MIC should be 24h, not 96h.

Do you know what should I do to deal with the issue?

Regards

Amir

P.S, please note that it is not possible to administer AMX as BID or TID in livestock.

2

Try to work in relative time i.e instead of 72-96 you want your time to be 0-24 ( clock start at the start of your profile you want to compute T> MIC on.

Samer

3

Thanks Samer,

That is exactly what I want but I don’t how to do that in setting of MCS or NCA.

4

you change your input data let us say you do a data transform

Timerel = Time - 72

and use Timerel ( starting at zero) in your computations.

5

Is there any other way? My Input data for NCA is from simulation for 2000 animals with over than a million time points. I can export MCS data to excel and correct time points, but it takes long long time.

6

you should always test on few subjects whether your setup is what you expect before simulating million rows.

it should not take much time to linearly transform your time column.
without reproducible code or project it is hard to help you

7

Thanks Samer, I’II do that.