- On 26 Apr 2003 at 02:19:03, "letasso" (letasso.at.bol.com.br) sent the message

Back to the Top

I have a problem to discuss some of my results when I

used the Scientist (Pharmacokinetic Program) to fit my

set data (when oral administration was used). A two

compartment open model was used to fit the data.

Statistic showed a MSC value over 2 with fator 1. The

fit was very good with a good correlation .

IndVars: T

DepVars: c

Params: a,b,d,alfa,beta,delta

C=(a*(exp(-alfa*t)))+(b*(exp(-beta*t)))-(d*(exp(-

delta*t)))

C is the same that A+B, C>A>B, gama>alfa>beta

After least Squares fit the values of A and C were very

large and almost the same, very different from the

middle values obtained.

C = A+B must be always true???

What can I do?

Any help that you could provide would be greatly

appreciated.

Best Regards,

Leandro Tasso

Laboratório de Farmacocinética

Programa de Pós Graduação em Ciências Farmacêuticas

Universidade Federal do Rio Grande do Sul

Av. Ipiranga, 2752, 90610-000

Porto Alegre-RS/Brasil

Fone: +55-51-3316-5215

e-mail: letasso.-a-.bol.com.br - On 28 Apr 2003 at 15:05:09, "Hans Proost" (j.h.proost.-a-.farm.rug.nl) sent the message

Back to the Top

The following message was posted to: PharmPK

Dear Leandro Tasso,

You wrote:

> IndVars: T

> DepVars: c

> Params: a,b,d,alfa,beta,delta

> C=(a*(exp(-alfa*t)))+(b*(exp(-beta*t)))-(d*(exp(-

> delta*t)))

> C is the same that A+B, C>A>B, gama>alfa>beta

> After least Squares fit the values of A and C were very

> large and almost the same, very different from the

> middle values obtained.

> C = A+B must be always true???

I trust that you mean 'd' instead of 'C' here since C refers to the

concentration? The minus sign in the equation suggests that 'delta' is

the

absorption rate constant (ka)?

In my experience it occurs rather often that ka and k (in case of

one-compartment) or alpha or beta (in case of two compartments) become

almost similar. As a results their coefficients a (or b) and d become

very

large, since this value is a composite value that can written as

F.Dose/V .

ka/(ka-k) (or a slightly different equation for two compartments).

I wrote a message to the PharmPK group several years ago, but i did not

get

a fully satisfactory response. From Monte Carlo simulation I learned

that

this happens rather often, also in cases where the true values for ka

and k

(or a or b) differ by a factor of, say, 2. Please note that

computational

problems may at least affect the problem. If ka and k are almost equal,

both

parameters become strongly correlated, and cannot be estimated

accurately.

But reparametrization does not solve the problem.

I still do not know why this happens so often. Any suggestion is

welcome!

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.-at-.farm.rug.nl - On 1 May 2003 at 01:24:35, "Christopher A. Thornton" (cthornton.at.micromath.com) sent the message

Back to the Top

The following message was posted to: PharmPK

Dear Dr. Tasso,

I read your post recent post regarding Scientist and forwarded the data

you provided to our technical support team. They have confirmed your

initial result (i.e. the convergence and solution you have inquired

about). After taking the following actions to adjust the initial

parameters as follows we they have obtained the following results which

appear to be more consistent:

If you edit rat3.par to reflect initial estimates corresponding to the

middle.par (A = 4.3106, B = 1.7903, D = 6.1294, ALFA = 0.65985, BETA =

0.070391, KA = 3.3189) and then perform a fit, you end up with fitted

estimates of A = 5.0608, B = 0.45005, D = 5.4789, ALFA = 0.26611, BETA =

4.4382e-17, KA = 6.5629. This seems more in line with what is accepted.

Of

course, the BETA value suggests that the concentration B is nearly

constant. But looking at middle.par, we see that BETA = 0.070391, so a

near zero value seems within tolerable experimental error.

It appears that your initial paramerter estimates converged on a

solution

that is physically unlikely. We will look at putting some additional

validation routines into our new version of Scientist (MicroMath

Scientist

3.0) scheduled for release this fall.

Thank you for your business.

Christopher A. Thornton

Manager, MicroMath Research

1710 South Brentwood Blvd.

Saint Louis, Missouri 63144

1.800.942.6284

www.micromath.com

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 "Pharmacokinetic (Scientist)" as the subject

PharmPK Discussion List Archive Index page

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