NONMEM Users Network Archive

Hosted by Cognigen

Model Identifiability

From: Shuying Yang <shuying.y.yang>
Date: Thu, 4 Mar 2010 11:18:21 +0100

Dear NMusers

I have collected from several studies PK data from healthy subjects, via IV=
 =
and SubCut and patients via both routes and would like to model them all =
together. For all subjects, only one dose level via one administration was =
given. From exploration of raw data, it is observed that there might be som=
e=
 difference in bioavailability from SubCut between healthy subjects and =
patients. I tried the model using ADVAN5, the parameterisation is listed =
below.

My questions are:
Is the model with six THETAs and ETAs an identifiable model? NOTE: NONMEM=
 =
run with this model is successful, also the estimates for CL,V2,Q and V3 ar=
e=
 similar to results with IV data alone.
What is a better way to explore the potential difference in PK between =
healthy subjects and patients with SubCut?

Finally, I also tried to model K12 as K12=0, IF(ROUT.GT.1) =
K12=EXP(THETA(5)+ETA(5)), could anyone help explain why this is wrong? =
NONMEM gave the following error message:

0PRED EXIT CODE = 1
0INDIVIDUAL NO. 1 ID=0.10080000E+04 (WITHIN-INDIVIDUAL) DATA RE=
C=
 NO. 2
 THETA=
  1.10E+00 1.40E+00 5.00E-01 2.50E+00 -1.10E+00 5.00E-01
 NUMERICAL DIFFICULTIES OBTAINING THE SOLUTION.
 THE COEFFICIENT MATRIX IS ALGORITHMICALLY SINGULAR.

Thanks for the help.

Shuying Yang
Clinical Pharmacology Modelling and Simulation
GlaxoSmithKline

CMT = 1 for SubCut dose record, and CMT=2 for IV dose record
ROUT=1 for IV and ROUT=2 for SubCut


   CL = EXP(THETA(1)+ETA(1))
   V2 = EXP(THETA(2)+ETA(2))
   Q = EXP(THETA(3)+ETA(3))
   V3 = EXP(THETA(4)+ETA(4))

   K20 = CL/V2
   K23 = Q/V2
   K32 = Q/V3

   K12 = EXP(THETA(5)+ETA(5))

   RO = EXP(THETA(6)+ETA(6))
   F1=0
   IF(ROUT.GT.1) F1=RO/(1+RO)

$THETA
   1.0 ; CL
   1.2 ; V2
   0.5 ; Q
   1.5 ; V3

   -1.3 ; LOG(KA) /DAY
   1.3 ; LOGIT(F1)

$OMEGA 0.05 ;CL
$OMEGA 0.06 ;V2
$OMEGA 0.1 ; Q
$OMEGA 0.05 ;V3
$OMEGA 0.05 ; LOG KA
$OMEGA 0.05 ; LOGIT F1
$SIGMA
  0.02; PROPORTIONAL
  2000; ADDITIVE



-----------------------------------------------------------
This e-mail was sent by GlaxoSmithKline Services Unlimited
(registered in England and Wales No. 1047315), which is a
member of the GlaxoSmithKline group of companies. The
registered address of GlaxoSmithKline Services Unlimited
is 980 Great West Road, Brentford, Middlesex TW8 9GS.
-----------------------------------------------------------

Received on Thu Mar 04 2010 - 05:18:21 EST

The NONMEM Users Network is maintained by ICON plc. Requests to subscribe to the network should be sent to: nmusers-request@iconplc.com.

Once subscribed, you may contribute to the discussion by emailing: nmusers@globomaxnm.com.