Jump to content


Photo

Nonparametric superposition vs Phoenix model simulation

Nonparametric superposition

  • Please log in to reply
1 reply to this topic

#1 ballk

ballk

    Newbie

  • Members
  • Pip
  • 1 posts

Posted 15 October 2018 - 10:11 AM

Comparing steady-state Cmax after IV bolus (i.e. C0) obtained using nonparametric superposition versus Phoenix model simulation, I get the same Cmax,ss when simulating using a 1-compartment model, when accumulation is present. 

 

But when using a 2-compartment model, the two approaches give me different Cmax,ss values. I made sure to select an appropriate time range for lambda z estimation in the nonparametric superposition object, so I don't see why I should have any difference at all in the Cmax value, using the two different approaches. I know there are different methods and assumptions behind each approach, but in my specific case, I would think that selecting an appropriate lambda z should result in a similar level of accumulation for the two approaches.

 

Could you provide any ideas to explain this difference?

 

Please see attached Phoenix project file.

 

Thanks,

Kathryn

Attached Files



#2 Helmut Schütz

Helmut Schütz

    Advanced Member

  • Members
  • PipPipPip
  • 316 posts
  • LocationVienna, Austria

Posted 16 October 2018 - 02:11 PM

Hi Kathryn,

 

I’m not sure whether your selection of the time-range for the estimation of lambda-z is correct. In the 2comp-model (day 1) with a range of 1.5–1.99 you get a t½ of 4.1359 h but with only the last three time points (1.97–1.99) you would get 4.9573 h. That’s an indicator that in your time range elimination is “contaminated” by distribution. I didn’t dive deeper into your model but consider setting in up with the respective classical WNL-model (V/k not clearances) and have a look at the secondary parameters A&alpha and B&beta. The ratios A/alpha and A/beta gives you the fraction of AUCs driven by the distribution and elimination. If they are similar (say, 0.5 ± something) it will always be difficult to get a “clean” estimate of lambda-z without modeling (e.g., in NCA or NPS).

 


 Best regards,
Helmut

https://forum.bebac.at/






Also tagged with one or more of these keywords: Nonparametric superposition

0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users