Dear All,
I am trying to simultaneously model the Oral data of parent and Metabolite (formed when parent dosed) by using a 2 compartment extravascular model.
The graphical model is attached. Is the graphical model Correct for simultaneous fitting of parent and metabolite ?
with regards,
2 comp Ext_Parent_Metabolite Model.docx (25 KB)
Metabolism & Pharmacokinetics
Hi Raghav,
It appears you have 3 compartments in your model; I guess that C3 will also be the central compartment and CLPar2Met is a one way conversion of parent to metabolite. Is there also excretion of unchanged parent to urine? In that case it looks OK to me, (could you collect urine data too?)
However really i can’t say it’s ‘right’ or ‘wrong’; you have to be the judge of whether it describes your data well and you’re happy with the physiological assumptions it assumes.
Simon
The only that is not correct is that you cannot use twice the same name. Here you are using twice A0 for the urine. Change the name of one of the two. If you shift to edit as textual the pre-compiler should tell you that error. Often the program will crash only in textual mode but anyway remember never to use twice the same name. Other than that I agree with Simon. I do not see any issue with the model representation.
Best
Serge
Hi Serge , Simon,
Thank you for the reply,
Using the graphical model above, I tried to fit the plasma data of parent and metabolite (parent dosed). I have attached the project.
the units of Vmet , Clmet are different and are represented in ug and ug/h respectively. I have a question whether the parameters Vmet and Clpar2met and Clmet are identified as there is no change in the concentration time profile when i use initial estimates tab to adjust the values ?
Will the graphical model above work if there is any metabolism in the gut ? How to model the parent-metabolite data when there is a significant gut first pass ? and when there is a significant liver metabolism without gut metabolism ?
The other question i have is, How to define MDV when i have two columns one for parent (CObsPar) and one for metabolite (CObsMet)
With best regards,
R
Parent_Metabolite_M & S.phxproj (1.07 MB)
A lot of questions here, first have you looked at the NLME user guide, Help>Documents>
on p116;
MDV:
The MDV flag indicates that there is a missing dependent variable. This flag is Optional for Phoenix models as the tool recognizes which records contain observed values and which do not. However, it can be optionally used, for example, to exclude observations from certain analyses.
The MDV flag needs to be numeric with values of nonzero numeric, indicating that the dependent variable value is missing (typically a value of 1 is used) or 0/blank indicating that the dependent variable value is present. In other words, if the MDV flag is not blank or zero then the dependent variable observation is treated as a missing. If the MDV flag is zero or blank then the observation is present.
The user can indicate that an MDV column is needed by checking the MDV option in the Input Options tab. This action will add a new column, MDV, in the Main and Dosing mapping panels. Any column name can be mapped to MDV but …
Secondly I saw you have got a BLQ Wizard step, where you are censoring the BLQ values to 0; I wouldn’t do that. Either leave it as blank (or text) e.g. missing numeric, or set it to LoQ and use the BLQ option in NLME ( you can read more about this in the User manuals, I think there is an example too).
Moving onto the questions about gut metabolism,and first pass effect, remember in this type of modelling you are not starting with PBPK (bottom up), but a top down model, but looking for the simplest model that describes your data well - it does not necessarily have direct correlations to physiological compartments or processes. As you start to make it more complicated and introduce more processes then probably you will have identifiability issues e.g. CV% increases markedly, You have 17 parameters in your model already so you may already be hitting that problem if you have issues with some parameters.
I’ve not worked out what is going on with your units but if I have some time later in the week I’ll take another look.
Simon
EDIT : I added an example of using the BQL option on your observations and also changed the metabolite to have a proportional error also, the fit looks a fair bit better. See the model comparer fot these summaries;
Name RetCode LogLik -2LL AIC BIC
2 comp Ext Par_Met_Prop Err 1 -2081.70 4163.41 4197.41 4290.4487
BQL 2 comp Ext Par_Met_Prop Err 3 -1822.08 3644.16 3678.16 3770.4571
Hi all,
I tried to fit the parent and metabolite simultaneously with the graphical mode discussed and got final estimates and the fit improved by adding a peripheral (second) compartment to metabolite also.
The parameters obtained from parent alone, metabolite alone and parent + Metabolite analysis are attached:
The volume of metabolite is not identifiable in the combined model. Can we assume Vmet similar to V1 parent.
Now i have the model how do i validate it?
I understand that the clearance parameters are apparent. Do i have to use Fm (fraction metabolized ) / Bioavailability to convert CLPar2Met to absolute value ?
Thanks in advance…
Regards,
Rc
Comparison.xlsx (10.8 KB)Parent_Met_SD_140416.phxproj (2.36 MB)
Raghav, it is very hard to read your results above, please attach the project file or at least a word or excel document containing these as actual tables.
Youn need to use the button ‘more reply options’ to add attachments.
Simon
Hi,
I tried to fit the parent and metabolite simultaneously with the graphical mode discussed and got final estimates and the fit improved by adding a peripheral (second) compartment to metabolite also.
The parameters obtained from parent alone, metabolite alone and parent + Metabolite analysis are attached:
The volume of metabolite is not identifiable in the combined model. Can we assume Vmet similar to V1 parent.
Now i have the model how do i validate it?
I understand that the clearance parameters are apparent. Do i have to use Fm (fraction metabolized ) / Bioavailability to convert CLPar2Met to absolute value ?
Thanks in advance…
Regards,
Rc
Comparison.xlsx (10.8 KB)Parent_Met_SD_140416.phxproj (2.36 MB)
