PharmPK Discussion - Estimation of Emax and EC50 in WinNonlin 5.2

• On 4 Jun 2007 at 11:34:55, "Corti Natascia" (Natascia.Corti.aaa.usz.ch) sent the message
  

  
  Dear all
  
  We have been doing a two dose level steady state PK/PD Study with 10
  Patients where we measured drug plasmaconcentration at several time
  points after administering 30mg and 60mg of the substance. At the
  same time we determined plasma concentrations of two different
  hormones (PD) which are influenced by the drug. The plasma levels of
  hormone 1 correlated inversely with the plasamconcentration of the drug.
  I was asked to estimate Emax and EC50. I did this by modelling the
  drug concentration - response curve in WinNonlin Version 5.2.
  Because fitting of simple Emax model or sigmoid Emax model to all
  single measured values (drug plasmaconcentration against hormone
  plasma concentration) of all patients at any time point resulted in a
  far to high variance of data (CV% very high) I decided to use the
  following procedures to reduce variance :
  1) I calculated the patients mean of drug plasma concentration and
  hormone level at every pharmacokinetic time point and fitted the mean
  drug concentrations to mean hormone concentrations with the simple
  Emax and sigmoid Emax model.
  With following results:
  
  Simple Emax model:
  
  http://www.boomer.org/pkin/images/corti/ole0.png
  
  A= mean values of drug concentration at every blood sampling time point
  C= effect on hormone level expressed as % reduction compared to
  baseline values measurd without the drug.
  
  http://www.boomer.org/pkin/images/corti/ole1.png
  
  Sigmoid Emax model:
  
  http://www.boomer.org/pkin/images/corti/ole2.png
  
  A= mean values of drug concentration at every blood sampling time point
  C= effect on hormone level expressed as % reduction compared to
  baseline values measurd without the drug.
  
  http://www.boomer.org/pkin/images/corti/ole3.png
  http://www.boomer.org/pkin/images/corti/ole4.png
  
  2) Second procedure: I fitted data of every patient individually to a
  simple Emax or a sigmoid Emax model then took the best fit for every
  patient and finally calculated the mean of all Emax and EC50 values.
  
  Results:
  Patient name    Emax (%)        Emax_CV%        EC50 (ng/ml)    EC50_CV%
  AA              52.1    10.8    8.9     20.3
  BH              72.6    5.7     15.7    8.6
  EK              78.1    32.8    6.0     75.5
  GJ              71.6    12.2    6.0     26.1
  HG              95.1    10.3    10.1    27.5
  KS              100.8   8.7     10.9    17.4
  PM              71.7    9.6     13.2    17.7
  SD              55.7    11.1    9.0     39.3
  WA              72.8    11.1    15.1    12.7
  WD              65.5    18.4    27.6    46.5
  
  Mean    73.6    13.1    12.2    29.2
  Median  72.1    10.9    10.5    23.2
  SD      15.2    7.6     6.4     20.0
  
  Question:
  Are the two procedures I used to estimate Emax resp. EC50 appropriate?
  When publishing the data which method would you describe?
  
  Thank you for your comments!
  Natascia Corti
  natascia.corti.-at-.usz.ch
  
  --
  Natascia Corti, MD
  Department of clinical pharmacology and toxicology
  University Hospital of Zurich
  Raemistrasse 100
  CH-8091 Zurich
  Switzerland
  email. catascia.corti.at.usz.ch
  
  

  Back to the Top

Copyright 1995-2011 David W. A. Bourne (david@boomer.org)