Jump to content


Photo

POPPK model using single concentration data


  • Please log in to reply
9 replies to this topic

#1 vk_pharmacy

vk_pharmacy

    Newbie

  • Members
  • Pip
  • 7 posts

Posted 11 February 2019 - 07:57 PM

Dear All,

 

I have a dataset of a compound X. In this dataset steady-state concentration at three occasions was recorded. In each occasion, 1 sample was collected ( between 8 to 12 hrs).

So basically I have 3 ocassion steady-state data.

 

In literature, one compartment model with first-order absorption and two-compartment model with a transit compartment were reported. I am interested to know how can I define occasions in Phoenix NLME and what is the best way to start modeling with single concentration data.



#2 Simon Davis

Simon Davis

    Advanced Member

  • Administrators
  • 1,316 posts

Posted 11 February 2019 - 08:24 PM

how many subjects do you have?  I think it is going to be rather hard to have confidence in any fitting of three trough values when you have a 2 com EV model since you have more parameters than data points, which sound like they all are similar positions in the curve.  I'm not sure that you'll be able to observe an effect for OCCASION too...

 

However go ahead and set up your project and see how it goes, once you've got as far as you can , then you can post the project for others to comment upon.

 

 Simon.



#3 vk_pharmacy

vk_pharmacy

    Newbie

  • Members
  • Pip
  • 7 posts

Posted 11 February 2019 - 09:20 PM

Hi Simon,

 

I have 35 subjects. In literature previously one compartment with FO absorption is reported with ka fix. 



#4 Simon Davis

Simon Davis

    Advanced Member

  • Administrators
  • 1,316 posts

Posted 12 February 2019 - 06:02 AM

So you have about 100 data points total. I would probably start with the 1com model as I don't think you'll have any data points to describe an intial distribution phase.

 

What is your objective ? i.e. how to you hope to interperet and apply the results from this modelling exercise?

 

There is a worked example of OCC in the NLME user guide, it woudl probably be intersting for you to try that before you attempt to apply it to your own data - especially as I think you will find it difficult to observe an occaisional effect with this sort of data.  I would stick with coding this up with SS on input options and see how that looks.

 

  Simon.



#5 vk_pharmacy

vk_pharmacy

    Newbie

  • Members
  • Pip
  • 7 posts

Posted 14 February 2019 - 04:52 AM

Hi,

 

I have Time after the first dose and Time after dose column in the dataset. Which one I should use for the model building when data is on steady state. 

 

I tried TAFD column with OCC and without OCC and Ka fix to literature value. I am able to estimate Cl but V seems 4 time than the reported value. What should I do?



#6 Simon Davis

Simon Davis

    Advanced Member

  • Administrators
  • 1,316 posts

Posted 14 February 2019 - 08:29 AM

personally I think it is clearer to use TAFD. I don't think I can really work out what else might be occuring without seeing your project.

 

remember if it's truly 2 com but you can only fit 1 com then volumen might well appear higher, Simon.


Edited by Simon Davis, 14 February 2019 - 08:30 AM.


#7 vk_pharmacy

vk_pharmacy

    Newbie

  • Members
  • Pip
  • 7 posts

Posted 16 February 2019 - 02:03 AM

Hi Simon,

 

First of all sorry, I put into the wrong forum topic.

 

Here I have attached my model. Can you please help me out?

 

The reported Ka is 0.1 to 0.6 and Vd/F is 150 to 350 L and Cl /F is 5 to 10 L/h. I am able to get the Cl but not Vd.



#8 vk_pharmacy

vk_pharmacy

    Newbie

  • Members
  • Pip
  • 7 posts

Posted 16 February 2019 - 02:05 AM

[attachment=3339:NLME.phxproj]


Edited by Simon Davis, 26 August 2022 - 06:14 AM.

  • xushiqiang likes this

#9 Simon Davis

Simon Davis

    Advanced Member

  • Administrators
  • 1,316 posts

Posted 08 March 2019 - 02:21 PM

Hi sorry for the delat, for each subject you have 2 subject level structural parameters (V and CL) and 3 occasion level parameters (occasions for V). Thus, there are 5 parameters for each subject

Then NLME tries to fit the model using 3 points for each subject. If you look into eta shrinkage sheet, all values tends to be near 1. When ε-shrinkage is large, the individual predictions are of little value for assessing model adequacy because the individual predictions “shrink” back toward the observation, meaning that IPRED ≈ DV (observation).

 

By the way I don't think you can get observed earlier Volume of distribution with given data.

 

Trying this analytical solution for 1cmt fo absorption and fo elimination model (steady state) in R 

D = 600

V = 1401.54

ka = 0.30643

Cl = 12.7682

tau = 24

t = 22.08

k = Cl/V

C = D/V*ka/(ka-k)*(exp(-k*t)/(1-exp(-tau*k)) - exp(-ka*t)/(1-exp(-ka*tau)))

 

one can find the predicted value at some time point.

Another way is to create that simple model in Phoenix (about 10 clicks) and use Initial Estimates tab for quick estimations

 

Please also check again your data since Kel tends to be muc higher than Ka



#10 xushiqiang

xushiqiang

    Newbie

  • Members
  • Pip
  • 1 posts

Posted 25 August 2022 - 02:54 PM

I could not find the attachment.






0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users