Back to the Top
Dear all,
I have a simple question I hope someone can help me with. When
investigating whether a compound has a linear PK, one generally looks
at AUC and/or concentrations vs. dose. The idea is to see e.g. if
doubling the dose will result in double exposure. If linear CL
prevails, a regression line will have a slope not different from 1 and
an intercept not different from 0.
It has come to my attention that some people use log AUC vs. log dose
to construct the graph, using a power model for regression. Is there
any special reason for this?
Toufigh Gordi
Associate Director of Clinical Pharmacology
CV Therapeutics Inc.
3172 Porter Dr.
Palo Alto, CA 94304
Back to the Top
The following message was posted to: PharmPK
Toufigh,
AUC is believed to be usually lognormally distributed rather than
normally. Also, the requirement for the intercept to be 0 is not a
necessary one, unless you are trying to establish linearity from the
dose of 0 to the doses in your studies.
Katya
Back to the Top
The following message was posted to: PharmPK
Toufigh,
The reference we use is
Kevin Gough, Michael Hutchison, Oliver Keene, Bill Byrom, Stuart Ellis,
Larry Lacey and John McKellar; Assessment of dose proportionality:
Report from the statisticians in the pharmaceutical
industry/Pharmacokinetics UK joint working party; Drug Information
Journal, 29, 1995, p1039-1048
Similar to BA/BE analysis, log transforming of the parameters is
"better" (not the statistically correct terminology).
Susan Shoaf
Sr. Pharmacokineticist
Otsuka Maryland Research Institute
Back to the Top
The following message was posted to: PharmPK
Dear Toufigh,
I think that the power model is adequate to assess the
linearity/nonlinearity as it gives a more realistic estimate of the
relationship. For example, a slope of 1.1 is probably linear, a slope of
2, is not.
Hope this helps,
Thanks,
Samia
Back to the Top
The following message was posted to: PharmPK
Katya,
If I recall correctly (I'm afraid David is too tough with his scissors
when posting replies so I dont know exactly what Toufigh wrote) the
question asked if there was a reason for using linear regression of log
dose vs log AUC in order to test for linearity.
This transform both sides (TBS) approach is fine if you believe that
the residual error in the AUC is log normally distributed but is this
assumption reasonable?
There are several sources of random effects in AUC. The first effect is
due to between and within subject variability. The second effect is due
to the measurement error in time and concentration and the use of the
trapezoidal rule to compute AUC.
I would think a better way to model the relationship between dose and
AUC is to use a mixed effects model e.g. with NONMEM that explicitly
recognizes these random effects and attempts to distinguish between
them.
Nick
--
Nick Holford, Dept Pharmacology & Clinical Pharmacology
University of Auckland, 85 Park Rd, Private Bag 92019, Auckland, New
Zealand
email:n.holford.-at-.auckland.ac.nz tel:+64(9)373-7599x86730 fax:373-7556
http://www.health.auckland.ac.nz/pharmacology/staff/nholford/
[Nick, you didn't see Toufigh's original message? - db]
Back to the Top
Can somebody critique the use of weighting to assess linearity,
especially when dealing with wide dosage range?
Sure
Back to the Top
The following message was posted to: PharmPK
Hi Toufigh,
have a look at
Shein-Chung Chow and Jen-pei Liu;
Design and Analysis of Bioavailability and Bioequivalence Studies.
Marcel Dekker, New York, Basel, pp.367-374 (2000)
Cave: p.371 contains a typo, it should read
1. 1 < L 2. 0.75 < L < U 1.0
_or_, NOT _and_
Best regards
Helmut
--
Helmut Schuetz
BEBAC
Consultancy Services for Bioequivalence and Bioavailability Studies
Neubaugasse 36/11
A-1070 Vienna/Austria
tel/fax +43 1 2311746
http://BEBAC.at Bioequivalence/Bioavailability Forum at
http://forum.bebac.at
http://www.goldmark.org/netrants/no-word/attach.html
PharmPK Discussion List Archive Index page
Copyright 1995-2010 David W. A. Bourne (david@boomer.org)