From: Kristel Crommentuyn Apkcr@SLZ.NL
Subject: [NMusers] posthoc estimates
Date: Thu, May 20, 2004 5:37 am
Dear all,
Another problem with acquiring post hoc estimates.
The model describes the pharmacokinetics of an orally dosed compound that is
co-administered with an inhibitor of its metabolism. The model is described using
advan6, with
compartment 1 = depot for analyte
compartment 2 = central for analyte
compartment 3 = enzyme
compartment 4 = depot for enzyme inhibitor
compartment 5 = central for enzyme inhibitor
The contol stream that describes the model is shown below.
The pharmacokinetics of the enzyme inhibitor were described before with a separate
model. The posthoc estimates for Cl/F (RCL), V (RV), ka (RKA), lagtime (RALA) and
oral bioavailability (RF1) are added in the dataset (enclosed 2.csv). The dataset
consists of full pharmacokinetic curves and sparse samples for more than 100
subjects at different occasions. Since the model building is still in a starting
phase no interoccasion variability is estimated. Also, the influence of the enzyme
is left out of consideration.
Per occasion, 3 doserecords are present, 1 in the enzyme compartment (dose = 1,
given 672 hours before the drugs), 1 in the analyte compartment and 1 in the
inhibitor compartment.
As data check-out, we performed a posthoc evaluation (maxeval=0) and noticed that
posthocs estimates were present only at the doserecords, not at the DV records.
$PROBLEM RUN# 011
$INPUT C DROP DROP ID SEX AGE TAD TIME AMT II SS MDV EVID DV
CMT RCL RV RKA RALA RF1
$DATA 2.CSV IGNORE=C
$SUBROUTINE ADVAN6 TOL=5
$MODEL COMP=(DEPOT) COMP=(CENTRAL) COMP=(ENZYME)
COMP=(DEPOT) COMP=(CENTRAL)
$PK
TVK12=THETA(1)
K12=TVK12*EXP(ETA(1))
TVCL2=THETA(2)
CL2=TVCL2*EXP(ETA(2))
TVV2=THETA(3)
V2=TVV2*EXP(ETA(3))
K20=CL2/V2
S2=V2
KENZ=THETA(4)*EXP(ETA(4))
K45=RKA
CL5=RCL
V5=RV
ALAG4=RALA
F1=RF1
K50=CL5/V5
S5=V5
$DES
C1=A(1)
DADT(1)=-K12*A(1)
C2=A(2)/S2
DADT(2)=(K12*A(1))-(K20*A(2)*A(3))
C3=A(3)
DADT(3)=KENZ-KENZ*A(3)
C4=A(4)
DADT(4)=-RKA*A(4)
C5=A(5)/S5
DADT(5)=(RKA*A(4))-(K50*A(5))
$ERROR
Y=F+F*THETA(5)*ERR(1)+THETA(6)*ERR(2)
IPRED=F
W=SQRT(F*F*THETA(5)**2+THETA(6)**2)
IRES=DV-IPRED
IWRES=IRES/W
$THETA
(0.92) ;TVK12=TVKA
(5.3) ;TVCL2 IN (L/H)
(55.7) ;TVV2 IN (L)
(1 FIX) ;KENZ
(0.1) ;[P] PROP ERR
(0.25) ;[A] ADD ERR
$OMEGA
0.5 ;[P] IIV K12
0.2 ;[P] IIV CL
0.2 ;[P] IIV V
0.5 ;[P] IIV KENZ
$SIGMA
1 FIX ;PROP ERR
1 FIX ;ADD ERR
$ESTIMATION SIG=3 MAXEVAL=0 PRINT=5 NOABORT METHOD=0 POSTHOC
;$COVARIANCE PRINT=E
$TABLE ID TIME IPRED IWRES CL2 V2 K12 ETA(1) ETA(2) ETA(3)
SEX AGE NOPRINT ONEHEADER FILE=011.TAB
$TABLE ID TIME IPRED IWRES NOPRINT ONEHEADER FILE=SDTAB011
$TABLE ID CL2 V2 K12 KENZ ETA(1) ETA(2) ETA(3) NOPRINT ONEHEADER FILE=PATAB011
$TABLE ID SEX AGE NOPRINT ONEHEADER FILE=CATAB011
Can someone help us with this?
Best regards,
Kristel
___________________________________
Kristel Crommentuyn, pharmD
AIDS research pharmacist
Department of Pharmacy and Pharmacology
Slotervaart Hospital
Louwesweg 6
1066 EC Amsterdam
The Netherlands
apkcr@slz.nl
020-5124737
_________________________________
_______________________________________________________