Jump to content


Photo

I have ETAs; I want to simulate different dosing


  • Please log in to reply
34 replies to this topic

#21 sundar

sundar

    Advanced Member

  • Members
  • PipPipPip
  • 36 posts

Posted 22 April 2019 - 05:39 AM

Thanks for your reply. I am using Phoenix 8.1, now i can able to open the attachment. 



#22 sundar

sundar

    Advanced Member

  • Members
  • PipPipPip
  • 36 posts

Posted 24 April 2019 - 09:43 AM

Hi, 

 

I have a couple of questions with respect to different dose simulation,

 

1) For simulation after oral drug administration should we need to map ADDL in the main data sheet? If yes, how to map that. Since, in PHX only the option of bolus/infusions is available. 

 

2) Residual DV is to be considered as simulated results? Kindly confirm. 

 

Thanks

Sundar. M 



#23 Simon Davis

Simon Davis

    Advanced Member

  • Administrators
  • 1,119 posts

Posted 24 April 2019 - 11:53 AM

Hi Sundar, no you don't need to map ADDL to enable oral dosing, that is set in the model itself.  Note if you originally fitted IV data and then swithced to an extravascular model to simulate you would need to give an estimate for Ka. Do you know this with confidence?

 

ADDL is for you to specify  additional doses, without having to explicitly include every dose record in your source data set.  See page 114 of NLME user guide(under help menu).

 

And no, Residual DV is not the simulated results, look in the Simulation Table, be sure to request C or whatever your predicted variable was called in your model code, remember it is case sensitive.  Read the NLME guide, specifically Simulation run mode if you are using population mode.

Simulation is useful for situations where modelers have PK/PD parameter estimates from prior modeling
and want to explore the potential impact of modifying certain conditions, such as dosing regimens,
without having to collect more data. Simulations are based on the structural model and its parameter
estimates.

 

 Simon.



#24 sundar

sundar

    Advanced Member

  • Members
  • PipPipPip
  • 36 posts

Posted 24 April 2019 - 12:20 PM

Hi Simon, 

 

Thanks for your quick reply. I am simulating data from extra-vascular and dont have any IV data. Per your suggestion, if my understanding is correct - if simulation is purely based on extra-vascular data - no need to map  ADDL.

 

Regarding simulation table even i read the NLME guide, however i am not able to understand it clearly. 

 

Would you mind if i ask example for the same. 

 

Again thanks for your help.

 

Thanks

Sundar. M 



#25 Simon Davis

Simon Davis

    Advanced Member

  • Administrators
  • 1,119 posts

Posted 24 April 2019 - 01:19 PM

Please post you project containing your model and a copy of that object as yo uhave tried to set it up for siumulation. then we can advise you better.

What sort of regimen(s) do you want to simulate.

Simon.
PS regarding ADDL mapping, it depends on whether your simulation will have multiple doses and whether they are recorded each to their own line or you want to use ADDL to let the engine know another e.g. 12 ADDitionaL doses were taken

Edited by Simon Davis, 26 April 2019 - 07:22 AM.


#26 sundar

sundar

    Advanced Member

  • Members
  • PipPipPip
  • 36 posts

Posted 26 April 2019 - 11:16 AM

Attached File  PKPD simulation.phxproj   1014.1KB   120 downloadsHi Simon, 

 

I have uploaded my project file. 

 

I have a PKPD plasma concentration data (with baseline value) of 2mg and 4mg after oral administration. And using these data need to simulate PKPD profile for 6mg.

 

I have following questions regarding model development,

 

·       In the concentration vs time profile – baseline concentration to be considered or not during model development?

·       Both 2mg and 4mg data can be modeled simultaneously or have to model separately to have a final estimates for simulation

 

Thanks 

Sundar. M 



#27 sundar

sundar

    Advanced Member

  • Members
  • PipPipPip
  • 36 posts

Posted 30 April 2019 - 03:38 AM

Hi Simon, 

 

Hope you have received my project file. Kindly confirm. 

 

Thanks

Sundar. M 



#28 Simon Davis

Simon Davis

    Advanced Member

  • Administrators
  • 1,119 posts

Posted 02 May 2019 - 09:36 AM

I would recommend to model both doses at the same time, especially since you have PD data as this will be much more informative.

 

Regarding baseline concentrations, why are they there? is it an endogenous compound or because there have been previous doses? If it's already at SS then there is a dosing option under "Input options" to be support that.

 

I notice the ID as the same in the 2 and 4mg worksheets, are they the same individuals ? if not they should be assigned a unique ID as I have done in the attached project

 

Also your dose was incorrectly set with no column attributes of mg instead entering "2mg" at the dose event, this would not be read correctly. Also it's easier to work in the same mass units as your concentration

 

Are you sure your units are correct generally as you have a VERY small Volume.

 

I would suggest you check all your units and then fit your PK model, I attach a simple example of using SS with a 1 com additive model. Once you have that clear then copy that model and add in the PD data.

Attached Files


Edited by Simon Davis, 02 May 2019 - 09:55 AM.


#29 sundar

sundar

    Advanced Member

  • Members
  • PipPipPip
  • 36 posts

Posted 03 May 2019 - 02:02 PM

Hi

 

Baseline concentration are endogenous. If i map SS column during PK modeling its working fine. However, when i add PD not able to converge the model, i have tried almost all of the model & algorithm options. Hence, i did not map the SS option during PKPD modeling. 

 

Data from both 2mg and 4mg are from same individuals and different period.

 

I agree there was an error in data units, i have converted them now to maintain in mg units. 

 

I came across one more issue that, theta results are able to generate only following QREM algorithm, is there any particular reason for this?

 

While model refinement is required,  request you to guide me to set up the simulation steps.

 

Attached PD added model for your further guidance. 

 

 

 

Thanks

Sundar. M 

Attached Files



#30 Simon Davis

Simon Davis

    Advanced Member

  • Administrators
  • 1,119 posts

Posted 04 May 2019 - 05:53 AM

If the baseline concentration is endogenous and this is truly a naive dose then I think you need to build a placebo model first.   I just plotted your PD data and, in red, the change from baseline.  You are looking at small changes and I suspect some underlying diurnal variation.

 

with 2mg I suspect there is little real effect, and with 4mg there is some intial stimualtion it seems, possibly followed by a rebound, but it's hard to tell with out placebo data.  you need to gather that and then make a model for this before you can model this I suspect.

 

even without consdering this, why have a PD model set up with no base line?!

I added one and I can get a fit but I don't think it very useful.  And I am still not convinced by your tiny volume; what sort of compound is it, do you know the mechanism of action?

  I suspect you need to sit down with a PKPD consultant to plan future experiments and analyse them.  Eg you should probably freeze PK when first moving to PKPD whilst you work out inital estimates.

In the meantime try taking a look at Gabriellsson and Weiner's PKPD  book, PD15 example of modelling oscillating data. 

 

Simon

Attached Thumbnails

  • diurnal_PD.jpg

Attached Files



#31 sundar

sundar

    Advanced Member

  • Members
  • PipPipPip
  • 36 posts

Posted 07 May 2019 - 05:28 AM

Thanks for your suggestion. For some reason i cant disclose the drug's properties. However, i will look into your suggestion. 

 

Thanks

Sundar. M 



#32 sundar

sundar

    Advanced Member

  • Members
  • PipPipPip
  • 36 posts

Posted 16 May 2019 - 06:13 AM

Hi, 

 

I am getting error message ("__" duplicate of random effect) while mapping PKPD parameters eta values. 

 

Could you please explain how to sorted out or map eta values for simulation purpose. 

 

 

Thanks

Sundar. M 



#33 Simon Davis

Simon Davis

    Advanced Member

  • Administrators
  • 1,119 posts

Posted 16 May 2019 - 07:49 AM

Please post your project so we can see the mappings to help find the source of this error message.  Simon.



#34 sundar

sundar

    Advanced Member

  • Members
  • PipPipPip
  • 36 posts

Posted 16 May 2019 - 09:07 AM

Hi 

 

Please see uploaded file for your evaluation. 

 

Thanks

Sundar. M

 

 

 

Attached Files



#35 sundar

sundar

    Advanced Member

  • Members
  • PipPipPip
  • 36 posts

Posted 23 May 2019 - 06:30 AM

Hi

 

Have you had a chance to look at the uploaded file (Simulation_Test). 

 

Thanks

Sundar. M






0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users