Jump to content


Photo
- - - - -

minimal PBPK in PML with IV PO dosing why is F calculated for IV?

PBPK IV PO dose F

  • Please log in to reply
8 replies to this topic

#1 William R. Wolowich

William R. Wolowich

    Member

  • Members
  • PipPip
  • 27 posts

Posted 08 January 2019 - 02:02 PM

I have built a minimal PBPK model based on that of Fiserova-Bergerova V. Inhalation Anesthesia using physiologically based pharmacokinetic models. Drug Metab Rev. 1992;24:531-557

 

The project is attached. My question, why is Phoenix reporting F for the IV dosed subjects? How do I get the software to recognize IV F = 1?

 

PS this data is not for sharing.

Attached Files



#2 dlweiner

dlweiner

    Member

  • Members
  • PipPip
  • 23 posts

Posted 08 January 2019 - 02:41 PM

William, there are likely multiple ways to do this, but I suggest you modify the Aa dosepoint statement as follows...

 

dosepoint(Aa, idosevar = AaDose, infdosevar = AaInfDose, infratevar = AaInfRate, bioavail=F)

 

then remove "F*" from the DEs for Aa and Agut, as dosepoint will then automatically adjust for F for oral dosing.

 

deriv(Aa = - (Aa * Ka))
deriv(Agut = (Qgut * C1)- (QHPV * Cgut) + (Aa * Ka))
 
Hope that helps,
Dan


#3 William R. Wolowich

William R. Wolowich

    Member

  • Members
  • PipPip
  • 27 posts

Posted 08 January 2019 - 06:12 PM

 

William, there are likely multiple ways to do this, but I suggest you modify the Aa dosepoint statement as follows...

 

dosepoint(Aa, idosevar = AaDose, infdosevar = AaInfDose, infratevar = AaInfRate, bioavail=F)

 

then remove "F*" from the DEs for Aa and Agut, as dosepoint will then automatically adjust for F for oral dosing.

 

deriv(Aa = - (Aa * Ka))
deriv(Agut = (Qgut * C1)- (QHPV * Cgut) + (Aa * Ka))
 
Hope that helps,
Dan

 

Yes thank-you for the prompt reply. 



#4 William R. Wolowich

William R. Wolowich

    Member

  • Members
  • PipPip
  • 27 posts

Posted 08 January 2019 - 08:35 PM

Dan

I made the changes you suggested, but I am still getting F reported for the intravenous subjects. The project is attached again.

Attached Files



#5 bwendt@certara.com

bwendt@certara.com

    Advanced Member

  • Administrators
  • 282 posts

Posted 28 January 2019 - 02:09 PM

William,

 

Since F is a structural parameter, NLME will try to find a population estimate and calculate a posthoc value for each subject irrespective of dose route. 
Note that the F values in posthoc-table for the subjects with IV are shrunk to the structural parameter value with zeroed eta: ilogit( tvF*exp(nF) = ilogit( 0.917678*exp(0)) = 0.714569
NLME failed to find nF for those subjects since it is impossible.
The attached project file shows you a method to distinguish the subjects by their respective dose routes. It defines a covariate 'PODoseFlag' and applies it to the definition of parameter F.
 
Bernd
 
Attached File  thc for certara.phxproj   2.06MB   420 downloads


#6 William R. Wolowich

William R. Wolowich

    Member

  • Members
  • PipPip
  • 27 posts

Posted 28 January 2019 - 07:00 PM

Bernd: As you probably noticed, I used up a lot of covariates in this model. How else can I get PML Phoenix to use weight and cardiac output of the individuals to multiply with a scalar (ie %CO * patients CO)  for a specific tissue without using covariates?

Bill



#7 bwendt@certara.com

bwendt@certara.com

    Advanced Member

  • Administrators
  • 282 posts

Posted 01 February 2019 - 12:29 PM

Bill, I do not see any issues to work with multiple covariates. The way you are using covariates in your model looks fine to me. 



#8 William R. Wolowich

William R. Wolowich

    Member

  • Members
  • PipPip
  • 27 posts

Posted 01 February 2019 - 01:47 PM

Bernd; You don't see a problem in this reduced version of the model, but when I add the 2 metabolites I triple the number of covariates and I think I run into the 32 covariate maximum, because PML seems to ignore the last added covariates. How is that 32 max counted? I think it is for every call on a covariate, ie for Vc1, Vc2 and Vc3 the wieght covar is called 3 times. It is easy to reach 32 when you add all of the tissue and flow parameters *3!

Bill

Can the covariate max be expanded?



#9 bwendt@certara.com

bwendt@certara.com

    Advanced Member

  • Administrators
  • 282 posts

Posted 04 February 2019 - 08:36 AM

Bill,

there is no maximum of 32 covariates. There might have been in a previous version of Phoenix, but the current version, 8.1., does not have a maximum there. The only restriction that we found for covariates  is *Maximum number of covariate categories or occasions = 40.

Please check the project file attached where we created a hypothetical examples in PHX8.1 with 45 covariates included in a single parameter, and found that PML correctly incorporates all the covariates inside. In addition, it runs without any issue, but extremely slow:

 

Attached File  SingleParamInvolving45Covariates.phxproj   1.29MB   389 downloads

 

Please note that there is a similar topic that was posted in 2013:

 

https://support.cert...pulation-model/.

 

It seems that, at that time, it has this maximum 32 restriction. To avoid this, Serge proposed using Scenario run mode to reduce the number of covariates, and then do a covariate search (see details inside this link).

 

Hope, this helps,

 

Bernd

 






0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users