Formualtion covariate on absorption parameters including IV data

[Aline Fuchs]

Dear IQ team,

I have a combined model with IV and ORAL data. 

Oral data comes from 2 different experiments with different formulations. IV data also uses another formulation so data NLME contains 3 covariates levels.

I have included formulation as covariate on Fabs and ka.

The result shows a covariate effect also for IV data on Fabs and ka. I understand that the estimate actually set IV formulation effect similar to no effect.

Is it relevant / important to keep the formulation covariate estimate for IV data int he results table? I am tempted to remove it for communication purpose for non-modeler audience but I worry if someone want to use the model results for model implementation / simulation with IQR or any other software.

Thanks, Best,

Aline

Aline Fuchs

AC Immune SA

[Henning Schmidt]

Hi Aline,

It might be good to reconsider the model structure. I would believe that in this case Fabs should not be associated with IV.

Examples: a simple PK model (linear 1 cpt distribution, IV and oral admin) could be written as:

d/dt(Ad) = -ka*Ad + Fabs*INPUT1

d/dt(Ac) = +ka*Ad - CL/Vc*Ac + INPUT2

INPUT1 would be the oral dosing input and doses in the dataset with ADM=1 would go into it.

INPUT2 would be the IV dosing input and doses in the dataset with ADM=2 would go into it. The TINF column in the dataset would code for the infusion time. If 0 then bolus.

This would mean that it should not be possible to add a formulation covariate for IV on Fabs, since Fabs is not involved with IV administration but only with oral administration. 

Without having seen the model or the data I would suggest:

• Check the implementation of your model and ensure that Fabs (and ka) is only associated with oral administration and not with IV administration.
• My first assumption would be that the systemic parameters would be independent of the formulation. Thus I would initially only test one oral formulation vs the other. For this it might be better to introduce binary categorical covariates (FORMO1, FORMO2) rather than a single FORM covariate with 3 levels.
• Further considerations based on results.

Best,

Henning

________________________________________________________________________________

Henning Schmidt, PhD

Managing Director

IntiQuan GmbH

Elisabethenstrasse 23 [see map]

4051 Basel, Switzerland

Web: www.intiquan.com

Tel: +41 76 603 28 06

Confidentiality Note: This message is intended solely for the designated recipient(s). It may contain confidential or proprietary information and may be subject to attorney-client privilege or other confidentiality protections. If you are not a designated recipient you may not review, copy or distribute this message. If you receive this in error, please notify the sender by reply e-mail and delete this message. Thank you.

Data Protection & Privacy: The IntiQuan Data Protection statement applies for all communication. You can view it here: https://www.intiquan.com/data-protection.

IntiQuan GmbH | Managing Director: Dr. Henning Schmidt | Trade register: Basel Stadt CH-270.4.004.907-8 | UID: CHE-181.306.905

--
You received this message because you are subscribed to the Google Groups "IQR Tools Users Group" group.
To unsubscribe from this group and stop receiving emails from it, send an email to iqr-tools-users-...@intiquan.com.
To view this discussion on the web visit https://groups.google.com/a/intiquan.com/d/msgid/iqr-tools-users-group/ef269672-06f9-440f-8c5e-2b1ff16cc50cn%40intiquan.com.

[Aline Fuchs]

Hi Henning,

Thank You!

This is how my model is currently  written:

d/dt(Ad) = -ka*Ad + Fabs*INPUT2
d/dt(Ac)  =  ka*Ad - CL/Vc*Ac + INPUT1

TINF = 0 for all data (ORAL and IV)

Thus, It's just that INPUT1 = IV with ADM = 1 and INPUT2 =  ORAL with ADM=2: Would this change the results ?

I am not sure I get what you mean by doing this? For this it might be better to introduce binary categorical covariates (FORMO1, FORMO2) rather than a single FORM covariate with 3 levels.

should FORMO1 and FORMO2 be 2 different column ? I don't see how to create the correct dataset that contains IV data without providing IV formulation (also I agree that I don't want to test it actually)

[Henning Schmidt]

Great, so the model is perfectly fine.

I mean: instead of having a column FORM with entries 1,2, and 3 for 3 different formulations, I would suggest to have one column FORMO1 which would be 1 or the first oral formulation and 0 otherwise. A col FORMO2 which is 1 for the second oral formulation and 0 otherwise. And if you want also a col FORMIV which is 1 fo the IV formulation and 0 otherwise.

Then you could test FORMO1 or FORMO2 on Fabs and ka. Initially I would not go for testing both.

Best,

Henning

________________________________________________________________________________

Henning Schmidt, PhD

Managing Director

IntiQuan GmbH

Elisabethenstrasse 23 [see map]

4051 Basel, Switzerland

Web: www.intiquan.com

Tel: +41 76 603 28 06

Confidentiality Note: This message is intended solely for the designated recipient(s). It may contain confidential or proprietary information and may be subject to attorney-client privilege or other confidentiality protections. If you are not a designated recipient you may not review, copy or distribute this message. If you receive this in error, please notify the sender by reply e-mail and delete this message. Thank you.

Data Protection & Privacy: The IntiQuan Data Protection statement applies for all communication. You can view it here: https://www.intiquan.com/data-protection.

IntiQuan GmbH | Managing Director: Dr. Henning Schmidt | Trade register: Basel Stadt CH-270.4.004.907-8 | UID: CHE-181.306.905