- On 9 Oct 2003 at 17:13:40, "Prabhu, Saileta" (sprabhu.aaa.sunesis.com) sent the message

Back to the Top

The following message was posted to: PharmPK

Hi everybody

I am wondering if one can model data that shows flipflop kinetics in

WinNonlin using one the library models. Also, is it possible to get a

plot of the "feathered" curve. Thanks.

-Saileta

Saileta Prabhu, Ph.D.

Scientist

ADME-PK

Sunesis Pharmaceuticals

341 Oyster Point Blvd

South San Francisco, CA 94080

Tel.: 650-266-3679

Fax: 650-266-3502 - On 13 Oct 2003 at 05:45:41, "Suresh.B.L." (suresh_bl1.-at-.rediffmail.com) sent the message

Back to the Top

Dear Saileta,

I have some idea about modelling data using winnonlin. Any one in the

group please correct me if i am wrong.

Flipflop kinetics: The diagnostic for this kind of data:

(i)If the terminal portion of the conc. time curve is not parallel for

Oral and IV data.

(ii) If Ka (Absorption rate constant) is less than Ke (Elimination rate

constant)

Calculate the initial estimates i.e., V/F, K01, K10, Lag time from your

conc. time profile. Initially by looking at K01 & K10 values you can

guess whether it undergoes flipflop kinetics; Then use model 4 - from

the winnonlin compiled models (I am assuming that your compound shows

one compartment behaviour). In the results look at your observed and

predicted profile, Std error, CV(%) and some other statistical

parameters like AIC, SS, WRSS and condition no. based on that we will

know whether the model fits the data well or not.

For the same data if you fit with model 5 and model 6 which is for

K01=K10 with and without lag time, the model fails. Hope this will help.

Thanks,

B.L.Suresh - On 13 Oct 2003 at 11:18:19, Roger Jelliffe (jelliffe.at.usc.edu) sent the message

Back to the Top

Dear Saileta:

Why do you want to use a basically obsolete method like

feathering? Why don't you consider using more up to date methods? Most

population modeling software will do a much better job on the data, and

nonparametric pop modeling such as NPAG currently appears to be the

best, as it is consistent, and not restricted by assumptions of normal

or lognormal parameter distributions. You can go to www.lapk.org, and

click on new advances in population modeling, and see Bob Leary's

careful simulation study about this.

Very best regards,

Roger Jelliffe

Roger W. Jelliffe, M.D. Professor of Medicine,

Division of Geriatric Medicine,

Laboratory of Applied Pharmacokinetics,

USC Keck School of Medicine

2250 Alcazar St, Los Angeles CA 90033, USA

Phone (323)442-1300, fax (323)442-1302, email= jelliffe.-a-.usc.edu

Our web site= http://www.lapk.org - On 14 Oct 2003 at 09:24:27, "J.H.Proost" (J.H.Proost.-a-.farm.rug.nl) sent the message

Back to the Top

The following message was posted to: PharmPK

Dear colleagues,

Dr. Suresh wrote:

> Flipflop kinetics: The diagnostic for this kind of data:

> (i)If the terminal portion of the conc. time curve is not parallel for

> Oral and IV data.

OK

> (ii) If Ka (Absorption rate constant) is less than Ke (Elimination

> rate constant)

If you have only oral (or any extravascular) data, there are always two

solutions for the model parameters. In case of a one-compartment model,

the values of k01 (or ka, absorption rate constant) and k10 (or ke,

elimination rate constant) can always be interchanged, resulting in

exactly the same plasma profile. The solutions where k01 > k10 is

usually regarded as the 'normal' solution, and the solution where k01 <

k10 as 'flip-flop'. Note that the value of V/F is different for both

solutions.

Which of the two solutions is obtained, depends primarily on the

software, and probably of the initial estimates of the parameters. E.g.

using initial estimates k01 > k10 will likely result in a solution

where k01 > k01, but this is not necessarily so. The major point here

is that there is no way to decide which solution reflects best the real

situation. Looking at the values of k01 and k10 does not solve the

problem! Since the 'flip-flop' may be true in the analysis of any

extravascular data, I would say that a fitting program should always

provide both solutions. If the user has additional information, he/she

may choose the most likely solution. The common practice of using the

solution where k01 > k10 is highly questionable.

If oral and iv data are available, simultaneous fitting solves the

problem completely. Actually there is no need to use the term

'flip-flop' since there is only one solution, and nothing is flipping

or flopping.

> Calculate the initial estimates i.e., V/F, K01, K10, Lag time from

> your conc. time profile. Initially by looking at K01 & K10 values you

> can guess whether it undergoes flipflop kinetics;

As stated above, looking at the values does not solve the problem.

> In the results look at your observed and predicted profile, Std error,

> CV(%) and some other statistical parameters like AIC, SS, WRSS and

> condition no.

As stated above, the plasma profile of both solutions is exactly the

same, and so are the statistical parameters.

Best regards,

Hans Proost

Johannes H. Proost

Dept. of Pharmacokinetics and Drug Delivery

University Centre for Pharmacy

Antonius Deusinglaan 1

9713 AV Groningen, The Netherlands

tel. 31-50 363 3292

fax 31-50 363 3247

Email: j.h.proost.aaa.farm.rug.nl

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 "Modeling flipflop in WinNonlin" as the subject

PharmPK Discussion List Archive Index page

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