- On 10 Jul 2006 at 16:09:01, "Ronette Gehring" (rgehring.-a-.vet.k-state.edu) sent the message

Back to the Top

The following message was posted to: PharmPK

Dear Listmembers,

I have a question regarding Pharsight's WinNonlin compiled PK models in

which K01 = K10 (e.g., Model #5 in the library). Since I have not seen

these models used much, I was wondering when it would be appropriate to

apply them. Would it be for data with sparse numbers of samples during

the initial absorption phase, for which it is otherwise difficult to

accurately estimate the value of K01?

Thanks

Ronette

Ronette Gehring BVSc, MMedVet (Pharm)

Assistant Professor

Department of Clinical Sciences (Agricultural Practices)

Kansas State University, College of Veterinary Medicine

I 107 Mosier Hall

Manhattan

KS 66506

Telephone: 785-532-2153

E-mail: rgehring.aaa.vet.ksu.edu - On 12 Jul 2006 at 16:14:05, helmut.schuetz.-a-.bebac.at sent the message

Back to the Top

[Sorry if some of the messages are out of order. I've had to retrieve =20

some of the PharmPK messages from my junk mail box. I'm using a =20

different computer and have to retrain the junk filter - db]

The following message was posted to: PharmPK

Dear Ronette!

If WinNonlin's Model #4 lead to instable estimates of k01

and k10 (i.e., values are close together and show high

variances), the model may be over-parametrized; in such

a case you may opt for Model #5.

IMHO there is no direct connection which rich or sparse sampling.

<>

best regards,

Helmut Sch=FCtz

--

Helmut Sch=FCtz

BEBAC

Consultancy Services for Bioequivalence and Bioavailability Studies

Neubaugasse 36/11

1070 Vienna/Austria

tel/fax +43 1 2311746

Web http://BEBAC.at

BE/BA Forum http://forum.bebac.at - On 17 Jul 2006 at 10:52:44, "Ronette Gehring" (rgehring.at.vet.k-state.edu) sent the message

Back to the Top

The following message was posted to: PharmPK

Dear Helmut,

Thanks for your reply. I have one further question:

What is then the meaning of K in this model? Is it a hybrid of K01

and K10? Can it be used in dosing calculations?

Regards

Ronette

Ronette Gehring BVSc, MMedVet (Pharm)

Diplomate of the American College of Veterinary Clinical Pharmacology

Assistant Professor

Department of Clinical Sciences (Agricultural Practices)

Kansas State University, College of Veterinary Medicine

I 107 Mosier Hall

Manhattan

KS 66506 - On 23 Jul 2006 at 12:32:49, =?ISO-8859-1?Q?Helmut_Sch=FCtz?= (helmut.schuetz.-at-.bebac.at) sent the message

Back to the Top

The following message was posted to: PharmPK

Dear Ronette!

>you wrote:

>What is then the meaning of K in this model? Is it a hybrid of K01

>and K10? Can it be used in dosing calculations?

>

<>As already said, you should apply WinNonlin's Model #5 only,

if you are running into problems with Model #4.

Problems arise from parameter indentifiability.

If the 'true' values k01 and k10 are close together, the

SSQ-surface may become 'flat', making it very difficult for

the minimization-algorithm to find the 'true' minimum.

It's also possible that two local minima exists (instead of

one global).

In such a case the final estimates are very sensitive to the

given side conditions (initial parameter values, model bounds,

estimates, convergence criteria, number of maximal iterations,

mimimization algorithm,...); you may notice 'instable'

estimates (values of k01 and k10 close together and highly

correlated, high variance).

In such a case you may opt for Model #5, which is a

'borderline' situation of a 'flip-flop' model.

In a 'flip-flop' model the apparent elimination is actually

the absorption phase (which is slower than the 'true'

elimination). Without i.v. data it's not possible to

distinguish between k01 and k10.

Maybe on of Harold Boxenbaum's article may help:

http://www.ualberta.ca/~csps/JPPS1(3)/H.Boxenbaum/flip-boxenbaum.pdf

<>Model #5 simply sets k01=k10 yielding at stable solution, having

only two parameters (volume of distribution and k):

C(t)=k*(Dose/V)*t*exp(-k*t)

And yes, you may use the estimated k in dosing calculations,

although personally I would do so only, if all subjects

of a reasonably large data set shows a failure of Model #4.

best regards,

Helmut

--

Helmut Schuetz

BEBAC

Consultancy Services for Bioequivalence and Bioavailability Studies

Neubaugasse 36/11

1070 Vienna/Austria

tel/fax +43 1 2311746

Web http://BEBAC.at

BE/BA Forum http://forum.bebac.at

http://www.goldmark.org/netrants/no-word/attach.html

Want to post a follow-up message on this topic? If this link does not work with your browser send a follow-up message to PharmPK@boomer.org with "WinNonLin k01 = k10 model" as the subject

PharmPK Discussion List Archive Index page

Copyright 1995-2010 David W. A. Bourne (david@boomer.org)