From: Pierre Maitre <maitre@cdg.ch>
Subject: [NMusers] Problem (bug ?) with RATE and F1 in $PK
Date: Thu, 01 Nov 2001 23:00:24 +0100

Dear Nonmem Users

I run into the following problem with NMTRAN

My model is a 2 cpt model with s/c and iv administration.
I use ADVAN4, TRANS4, and my control file for NMTRAN is as follow:

$PROB xxxx 2cpt sc and iv
$DATA ../data/xxxx_PK.nm NOREWIND
$INPUT ID TIME CMT DOSE=AMT RATE CP=DV MDV EVID
$SUBROUTINES ADVAN4 TRANS4
$PK
CALLFL=-2
;
CL = THETA(1)*EXP(ETA(1))
V2 = THETA(2)
Q = THETA(3)*EXP(ETA(2))
V3 = THETA(4)*EXP(ETA(3))
KA = THETA(5)*EXP(ETA(4))
F1 = THETA(6)
S2 = V2
$ERROR etc....

Drug input occurs either as an iv bolus (i.e. in cmpt 2 when using
advan4) or as a long continuous sub-cutaneous infusion (i.e. in
compartment 1 when using advan4).
Bioavailability of the sub-cutaneous depot is probably between 0.3 to
0.5

As long as F1 (the bioavailability for the depot compartment) is
constrained to be 1, there is no problem.
However, when F1 is smaller than 1, say 0.5, strange things happen: the
duration of infusion is reduced by a factor F1.

I found the reason in the NONMEM documentation:
The command:
nmhelp bioavailability
gives the following:

> If the amount is A, and the bioavailability fraction is F, only the
> fraction F of A actually appears in the dose compartment (either
> instantaneously at the time the dose enters the compartment - with a
> bolus dose, or over a period of time - with an infusion).

In other words, if the fraction F1 is 0.5, the amount A entering
compartment 1 is only A*0.5. However the RATE of administration (a known
value, in the data set, see $INPUT above)) is not modified by F1 and is
unchanged. Therefore, the duration of administrations, i.e. A/RATE is
reduced by a factor F1, leading to very strange fits.

Question: is this a bug in the NMTRAN code, or am I missing something?

Pierre Maitre
--
Dr Pierre-O. Maitre
Privat Docent
FMH Anesthésiologie
FMH Pharmacologie Clinique
Cabinet médical / A la Joy
CH-1273 Genolier Switzerland

 

*****

 

 

From: Leonid Gibiansky <lgibiansky@emmes.com>
Subject: Re: [NMusers] Problem (bug ?) with RATE and F1 in $PK
Date: Fri, 02 Nov 2001 08:43:36 -0500

I think this is a feature, not a bug. The motivation (at least how I
interpreted it when I came across this problem) is the following:
bioavailability is treated as some factor outside of the system. F1 affect
the input. You may assume that the tablet (or any other drug form) is
reduced (multiplied) by F1 at time of dosing. So available amount is
NEW_AMT=F1*AMT. This does not affect the rate. That is why attempts to
change rate by changing the bioavailability cannot be successful (and this
can be seen only by looking on the concentration or rate). Bioavailability
is used only once, for the dose record, to multiply amount by F1. To change
rate you probably need to assign duration D1 (computed as AMT/RATE). Then
any changes in F1 will lead to changes in RATE (NEW_ RATE = NEW_AMT/D1=
F1*AMT/ (AMT/RATE) =F1*RATE).

Leonid

 

 

*****

 

 

From: Pierre Maitre <maitre@cdg.ch>
Subject: Re: [NMusers] Problem (bug ?) with RATE and F1 in $PK
Date: Fri, 02 Nov 2001 17:16:00 +0100

Leonid Gibiansky a écrit :
>
> I think this is a feature, not a bug. The motivation (at least how I
> interpreted it when I came across this problem) is the following:
> bioavailability is treated as some factor outside of the system. F1 affect
> the input. You may assume that the tablet (or any other drug form) is
> reduced (multiplied) by F1 at time of dosing. So available amount is
> NEW_AMT=F1*AMT. This does not affect the rate.

This is precisely my point: F1 does not affect the data item RATE, and
it should:

When you administer a sub-cutaneous infusion, your NONMEM data set needs
a AMT data item, and a RATE data item. Both are known data values, they
are not estimated values. The duration of the infusion is computed
internally by NONMEM (or is it by PREDPP ?) as the ratio AMT/RATE. In
this case, the duration is a "computed" data , not an estimated
parameter (I know that you can *estimate* the duration with the
parameter D1, but this is not my intention here).

Introducing the bioavailability parameter F1, will - as you say -
multiply AMT by F1 at the time of dosing. The data set hasn't changed
(we still use the same data set) , but for NONMEM F1 acts as if AMT
were reduced by a factor F1. Now, because F1 does not also scale the
data item RATE, the duration of the infusion - as viewed by NONMEM - is
now reduced by a factor F1 (remember that duration is computed as
AMT/RATE)... and the fit is totally screwed up.

... it's hard to call this a "feature" :-)

In brief, I have a known amount of drug, a known rate of administration,
a known duration of administration, and I just would like to calculate
the bioavailability of a continuous sub-cutaneous infusion with ADVAN4.
If I could use a DURATION data item (instead of the RATE data item),
this would solve my problem, but there is no such animal in the NONMEM
zoo. Moreover, I do not intend to use the D1 parameter and estimate the
duration, because duration is a data that was precisely measured.

Has anyone had this problem, and how was it solved?

Thanks for your help

Pierre Maitre
--
Dr Pierre-O. Maitre
Privat Docent
FMH Anesthésiologie
FMH Pharmacologie Clinique
Cabinet médical / A la Joy
CH-1272 Genolier Switzerland

 

 

*****

 

 

From: "Bachman, William" <bachmanw@globomax.com>
Subject: [NMusers] FW: DURATION data item (instead of the RATE data item)
Date: Fri, 2 Nov 2001 13:23:01 -0500

Pierre Maitre writes:

>In brief, I have a known amount of drug, a known rate of administration,
>a known duration of administration, and I just would like to calculate
>the bioavailability of a continuous sub-cutaneous infusion with ADVAN4.
>If I could use a DURATION data item (instead of the RATE data item),
>this would solve my problem, but there is no such animal in the NONMEM
>zoo. Moreover, I do not intend to use the D1 parameter and estimate the
>duration, because duration is a data that was precisely measured.
>
>Has anyone had this problem, and how was it solved?

See PREDPP Guide page 26, for a discussion of one technique using a
modeled rate to get exactly what you want.

Another technique is to use a modeled duration and set it equal to the
duration data item (disguised as another type of data item).

Stuart Beal

 

 

*****

 

 

From: Leonid Gibiansky <lgibiansky@emmes.com>
Subject: Re: [NMusers] Problem (bug ?) with RATE and F1 in $PK
Date: Fri, 02 Nov 2001 13:31:01 -0500

I think you can find any animal or plant in the NONMEM zoo, and DURATION is
one of them. For example, you can pretend that you "model" duration and
instead take it from the data file column (say D1), computed in advance as
AMT/RATE. This should be pretty straightforward to realize.

Leonid

 

*****

 

 

From: Nick Holford <n.holford@auckland.ac.nz>
Subject: [NMusers] NONMEM Zoo
Date: Sat, 03 Nov 2001 08:06:52 +1300

Leonid Gibiansky wrote:
>
> I think you can find any animal or plant in the NONMEM zoo

An excellent point. It is useful to reflect when one feels that one has been bitten by a NONMEM bug that perhaps it is was a wise elephant bellowing in your ear to make you wake up and appreciate that a deeper issue needs to be considered.

On a historical note users should be thankful that one can now estimate F1 with a zero-order input. Early versions of NONMEM assumed that a zero-order input must mean a fully bioavailable parenteral input. This feature/bug bit me on the first PK problem I ever tried with NONMEM and put me off using it for some years.

Nick
--
Nick Holford, Divn Pharmacology & Clinical Pharmacology
University of Auckland, 85 Park Rd, Private Bag 92019, Auckland, New Zealand
email:n.holford@auckland.ac.nz tel:+64(9)373-7599x6730 fax:373-7556
http://www.phm.auckland.ac.nz/Staff/NHolford/nholford.htm

 

 

*****

 

 

From: Pierre Maitre <maitre@cdg.ch>
Subject: Re: [NMusers] Problem (bug ?) with RATE and F1 in $PK
Date: Mon, 05 Nov 2001 10:39:41 +0100

"Stuart Beal" a écrit :
>
> Another technique is to use a modeled duration and set it equal to the
> duration data item (disguised as another type of data item).
>
Leonid Gibiansky a écrit :
>
> I think you can find any animal or plant in the NONMEM zoo, and DURATION is
> one of them. For example, you can pretend that you "model" duration and
> instead take it from the data file column (say D1), computed in advance as
> AMT/RATE. This should be pretty straightforward to realize.

Good trick, it worked fine.
Thanks to Stuart and Leonid for their help

Pierre Maitre
--
Dr Pierre-O. Maitre
Privat Docent
FMH Anesthésiologie
FMH Pharmacologie Clinique
Cabinet médical / A la Joy
CH-1272 Genolier Switzerland

 

 

*****

 

 

From: "Piotrovskij, Vladimir [JanBe]" <VPIOTROV@janbe.jnj.com>
Subject: RE: [NMusers] Problem (bug ?) with RATE and F1 in $PK
Date: Mon, 5 Nov 2001 11:03:17 +0100

Pierre,

What you described is indeed a NONMEM feature, but not a bug. In your case of SC infusion you know just duration, not the rate (it depends on the bioavailability). Below is the way how to code your case. I reduced the model and dropped ETAs for simplicity; duration is 5; MODE=3D1 corresponds to SC and MODE=3D2 to IV:

$PROB sc and iv
$DATA nmd.ssc
$INPUT ID TIME CMT DOSE=3DAMT RATE CP=3DDV EVID MODE
$SUBROUTINES ADVAN2 TRANS2
$PK
CL =3D THETA(1)
V =3D THETA(2)
KA =3D THETA(3)
D1 =3D 5
F1 =3D THETA(4)
S2 =3D V
$ERROR =20
Y =3D F * (1 + ERR(1))
$THETA
(0 20)
(0 100)
(0 5)
(0 .5)
$OMEGA
.01
$SIM (34906)=20
$EST PRINT=3D10 MAX=3D9990=20
$TABLE ID TIME CMT DOSE RATE EVID MODE
FILE=3Dtab.ssc ONEHEADER NOPRINT

An example (single-subject) data set is as follows:

1 0 1 100 -2 0 4 1
1 1 2 0 0 1 0 1
1 2 2 0 0 1 0 1
1 3 2 0 0 1 0 1
1 4 2 0 0 1 0 1
1 5 2 0 0 1 0 1
1 6 2 0 0 1 0 1
1 7 2 0 0 1 0 1
1 8 2 0 0 1 0 1
1 9 2 0 0 1 0 1
1 10 2 0 0 1 0 1
1 0 2 100 0 0 4 2
1 1 2 0 0 1 0 2
1 2 2 0 0 1 0 2
1 3 2 0 0 1 0 2
1 4 2 0 0 1 0 2
1 5 2 0 0 1 0 2
1 6 2 0 0 1 0 2
1 7 2 0 0 1 0 2
1 8 2 0 0 1 0 2
1 9 2 0 0 1 0 2
1 10 2 0 0 1 0 2

Best regards,
Vladimir