Jump to content


Photo

Execution errors when conducting PopPK model with steady state

steady state Execution error PopPK model

  • Please log in to reply
4 replies to this topic

#1 Bellerin

Bellerin

    Newbie

  • Members
  • Pip
  • 3 posts

Posted 11 November 2020 - 12:35 PM

Dear all,

I have met some issues when conducting popPK models with steady state. In our study, oral drug A is given to subjects once daily for 2 weeks before PK sampling. Then, blood samples are collected at pre-dose and at 1, 2, 4, 6, 8 hours after drug intake. In theory, the drug concentrations should reach the steady state, which is also confirmed in previous studies. Thus, I tick the SS option in the Phoenix NLME 8.0. The plots showed that popPK models with steady state fitted the data better, especially for the concentrations at pre-dose. However, I met execution errors when doing Visual Predictive Check, warning some subjects are unable to achieve steady-state. I have no idea why it will happen and how to solve it. It will be much appreciated if someone can give me some suggestions.

 

I have attached my document and warning picture. Looking forward to your kind replies!

 

Attached Thumbnails

  • Execution error.png

Attached Files



#2 kniefort

kniefort

    Newbie

  • Members
  • Pip
  • 7 posts

Posted 11 November 2020 - 04:56 PM

I believe what's happening here is that there are not many points in your dataset that help to define the 2 cmpt parameter set (points between 10 and 24 hours).  During the estimation, some individuals parameter values (e.g., V2) are driven to values that equate to a very long t 1/2, hence the error that SS can't be achieved. 

 

For some reason, I'm not able to attach a file to my post right now, apologies. 

 

What I did was: 

  • Modified dataset to create a time variable that indicates SS dose record happens on Day 15 rather than Day 0 (e.g., add 336 hours to each of your time values). 
  • Removed the ETA terms from V2 and CL2, as individual estimates for V2 were being driven to very high values (via individual eta term) that would result in SS not being achieved at the time of sampling.

This resulted in a successful run and vpc.  

 

 

Hope this is helpful!

 

Keith

Attached Files


Edited by Simon Davis, 11 November 2020 - 09:45 PM.

  • Simon Davis and Bellerin like this

#3 Simon Davis

Simon Davis

    Advanced Member

  • Administrators
  • 1,134 posts

Posted 11 November 2020 - 09:46 PM

i made a smaller copy of Keith's file by removing some items and and attached it above as he had hit the size limit on the forum of 12Mb.



#4 Bellerin

Bellerin

    Newbie

  • Members
  • Pip
  • 3 posts

Posted 12 November 2020 - 04:02 AM

Dear Keith,

 

Many thanks for your reply. Your answer perfectly helps me understand the causes of this issue and the solutions to solve it, where I have been struggling for a long time. I will have a try right now and hope it also works well for other drugs in my dataset.

 

Kind regards,

 

Bellerin

 

I believe what's happening here is that there are not many points in your dataset that help to define the 2 cmpt parameter set (points between 10 and 24 hours).  During the estimation, some individuals parameter values (e.g., V2) are driven to values that equate to a very long t 1/2, hence the error that SS can't be achieved. 

 

For some reason, I'm not able to attach a file to my post right now, apologies. 

 

What I did was: 

  • Modified dataset to create a time variable that indicates SS dose record happens on Day 15 rather than Day 0 (e.g., add 336 hours to each of your time values). 
  • Removed the ETA terms from V2 and CL2, as individual estimates for V2 were being driven to very high values (via individual eta term) that would result in SS not being achieved at the time of sampling.

This resulted in a successful run and vpc.  

 

 

Hope this is helpful!

 

Keith



#5 Bellerin

Bellerin

    Newbie

  • Members
  • Pip
  • 3 posts

Posted 12 November 2020 - 04:04 AM

Dear Simon,

 

Also many thanks for your support to upload the file.

 

Kind regards,

 

 

i made a smaller copy of Keith's file by removing some items and and attached it above as he had hit the size limit on the forum of 12Mb.






0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users