From: Kimberly Kimshim <*pharmsuni*>

Date: Thu, 27 Feb 2014 18:03:34 +0900

Hello. This is Kim who needs to complete my thesis with regards to

population pharmacokinetics.

I collected drug A concentration data from outpatient records. Since it was

collected retrospectively, it was consisted of trough levels and had huge

variation in this dataset by visual checking.

Therefore, I fixed the Ka since I did not have much information to explain

absorption part. I estimated Ka and Vd only to find poor estimation of Vd

and then, I fixed Vd to reference value (without eta value) and added

covariates to both Ka and Vd to estimate covariate effects.

$PK

T1 = THETA(1)

T2 = THETA(2)

T3 = THETA(3)

TVCL=T1

TVV=T2

TVKA=T3

CL = TVCL *EXP(ETA(1))

V = TVV

KA = TVKA

S2=V/1000

K=CL/V

$ERROR

IPRED=F

W=IPRED+0.0001

IRES=IPRED-DV

IWRE = IRES/W

Y= IPRED + W*EPS(1)

$THETA

(28.8) ;

(454) ;

(4.5) FIX ;

$OMEGA

0.351

$SIGMA

0.275

I started with nineteen covariates and seventeen of them were related to Ka

and/or Vd in Full model. Final model is as follows;

TVCL=THETA1*EXP(THETA5*(SCRR-1.12))

TVV=THETA2*EXP(THETA4*(SCRR-1.12))*EXP(THETA6*(ALBU-3.82))

The final model looked okay and clinical correlation of each covariate to

Ka and Vd made senses to me. However, my advisor wants to fix the Vd but

consider body weight of each individuals. The problem is that only half of

the subjects had their body weights known. Base model showed huge~~~ Eta

value in clearance (2.0E+021%). Eta value in Vd was 68.5%.

$PK

T1 = THETA(1)

T2 = THETA(2)

T3 = THETA(3)

T4 = THETA(4)

TVCL=T1

TVV=T2*(BWTT/64.1+0.001)

TVKA=T3

CL = TVCL *EXP(ETA(1))

V = TVV *EXP(ETA(2))

KA = TVKA

S2=V/1000*(BWTT/64.1+0.001)

K=CL/V

$ERROR

IPRED = F

IRES = DV - IPRED

W = THETA(4)

WRES = IRES/W

Y = IPRED+W*EPS(1)

$THETA

(5.51) ;

(159) ;

(4.5) FIX ;

(3.36) ;

$OMEGA

120

0.327

$SIGMA

1 FIX

My question is which is the best way to establish base model for this

dataset (trough only)? Fixing Vd with covariate effect consideration (fixed

effect model?) Fixing Vd without estimation? Fixing Vd adaptive to body

weight?

Thank you for your reply in advance.

- Kim

--

***********************************************

Mi Kyong Kim Shim, Pharm.D., RPH,

Researcher, MS candidate in clinical pharmacy.

School of Pharmacy

Seoul National University

1 Gwanak-ro, Gwanak-gu, Seoul, South Korea 151-742

TEL. 82-2-2072-0335

FAX 82-2-766-9560

E-mail pharmsuni

Received on Thu Feb 27 2014 - 04:03:34 EST

Date: Thu, 27 Feb 2014 18:03:34 +0900

Hello. This is Kim who needs to complete my thesis with regards to

population pharmacokinetics.

I collected drug A concentration data from outpatient records. Since it was

collected retrospectively, it was consisted of trough levels and had huge

variation in this dataset by visual checking.

Therefore, I fixed the Ka since I did not have much information to explain

absorption part. I estimated Ka and Vd only to find poor estimation of Vd

and then, I fixed Vd to reference value (without eta value) and added

covariates to both Ka and Vd to estimate covariate effects.

$PK

T1 = THETA(1)

T2 = THETA(2)

T3 = THETA(3)

TVCL=T1

TVV=T2

TVKA=T3

CL = TVCL *EXP(ETA(1))

V = TVV

KA = TVKA

S2=V/1000

K=CL/V

$ERROR

IPRED=F

W=IPRED+0.0001

IRES=IPRED-DV

IWRE = IRES/W

Y= IPRED + W*EPS(1)

$THETA

(28.8) ;

(454) ;

(4.5) FIX ;

$OMEGA

0.351

$SIGMA

0.275

I started with nineteen covariates and seventeen of them were related to Ka

and/or Vd in Full model. Final model is as follows;

TVCL=THETA1*EXP(THETA5*(SCRR-1.12))

TVV=THETA2*EXP(THETA4*(SCRR-1.12))*EXP(THETA6*(ALBU-3.82))

The final model looked okay and clinical correlation of each covariate to

Ka and Vd made senses to me. However, my advisor wants to fix the Vd but

consider body weight of each individuals. The problem is that only half of

the subjects had their body weights known. Base model showed huge~~~ Eta

value in clearance (2.0E+021%). Eta value in Vd was 68.5%.

$PK

T1 = THETA(1)

T2 = THETA(2)

T3 = THETA(3)

T4 = THETA(4)

TVCL=T1

TVV=T2*(BWTT/64.1+0.001)

TVKA=T3

CL = TVCL *EXP(ETA(1))

V = TVV *EXP(ETA(2))

KA = TVKA

S2=V/1000*(BWTT/64.1+0.001)

K=CL/V

$ERROR

IPRED = F

IRES = DV - IPRED

W = THETA(4)

WRES = IRES/W

Y = IPRED+W*EPS(1)

$THETA

(5.51) ;

(159) ;

(4.5) FIX ;

(3.36) ;

$OMEGA

120

0.327

$SIGMA

1 FIX

My question is which is the best way to establish base model for this

dataset (trough only)? Fixing Vd with covariate effect consideration (fixed

effect model?) Fixing Vd without estimation? Fixing Vd adaptive to body

weight?

Thank you for your reply in advance.

- Kim

--

***********************************************

Mi Kyong Kim Shim, Pharm.D., RPH,

Researcher, MS candidate in clinical pharmacy.

School of Pharmacy

Seoul National University

1 Gwanak-ro, Gwanak-gu, Seoul, South Korea 151-742

TEL. 82-2-2072-0335

FAX 82-2-766-9560

E-mail pharmsuni

Received on Thu Feb 27 2014 - 04:03:34 EST