Back to the Top
In doing non-compartmental pharmacokinetics analysis, is it possible
that AUC0-inf will be less than AUC0-last or AUC0-all.
Regards,
Aqel W. Abu-Qare, Ph.D
Research Scientist.
Pharmacokinetics,Metabolism
and Bioanalytical.
NeoPharm, Inc.
1850 Lakeside Drive
Waukegan, IL 60085
Back to the Top
Aqel
> In doing non-compartmental pharmacokinetics analysis, is it possible
> that AUC0-inf will be less than AUC0-last or AUC0-all.
No. Assuming you are talking about a single dose study then since
AUC(0-inf) = AUC(0-last) + AUC(last-inf)
and AUC(last-inf) cannot be negative.
Regards
Steve
Steve Duffull
School of Pharmacy
University of Queensland
Brisbane QLD4072
Australia
Back to the Top
[Quite a few replies - db - see WDSOP message for my reply :-)]
From: "Aziz Karim"
Date: Thu, 27 Jun 2002 22:24:32 -0500
To: david.aaa.boomer.org
Subject: Re: PharmPK Re: AUC 0-inf
Status: U
Dear Aqel:
WinNonlin gives you the option of determining AUC (last), AUC (all) and AUC
(inf) As pointed out by Dr Duffel, AUC (last) can not be greater than
AUC (inf).
However there can be an artefact where by AUC (all) would end up
being greater than
AUC (inf).
We need to clearly define the AUC(last) and AUC (all) values:
1) AUC (last) would represent the AUC up to the last validated measurable
plasma concentration [C(last).
2) AUC up to the last point below the limit of quantitation [considered zero
in AUC calculation] represents AUC(all).
It is difficult describe in words the above distinction so I have attached a
figure which explains the differences between the two AUC values.
Please let me know if
you have difficulties in opening the attachment. If so I will send
you this figure by
fax or send you directly via your E-mail address.
[The attachment is not included - my attempt to resist the spread of
viruses - db]
AUC(last) by definition can not be greater than AUC(inf). However, as an
artefact, it is possible for AUC(all) to be larger than AUC(inf).
This occurs when
time span is large between the last validated plasma concentration
(Clast) in the
disposition phase and the next sample with concentration below the
limit of quantitation (see attached Figure for explanation).
As a note, I don't particularly like to use AUC(inf) for drugs with
short half lives (T1/2 = < 24 hours). The reason for this is one
single concentration [C(last)] in the beta phase has a high influence
on the AUC(inf) values since AUC(inf) = AUC(last) +
[C(last)/(0.693/T1/2)]. It is simply better to determine AUC(last)
with the most sensitive assay and taking blood samples for a time
period exceeding 4*T1/2.
AUC(inf) is more appropriate for drugs with long T1/2 (>48 hours)
provided that the ratio of [AUC(last)/AUC(inf) ]exceeds 0.80.
Regards,
Aziz Karim
---
From: daniel.martinez.at.beaufour-ipsen.com
Date: Fri, 28 Jun 2002 08:27:45 +0200
To: david.at.boomer.org
Subject: Re: PharmPK AUC 0-inf
Dear Abu-Qare,
I think that you have some problem when you do the Pharmacokinetics
analysis. I don't know any case where the AUCall is higher than the AUCinf.
If you think about the non-compartmental formulas of this parameters you
will see that:
AUCinf = AUClast + Clast / lz
so as minimum you can have an AUCinf = AUClast if Clast = 0.
I hope it helps.
Your Sincerely
Daniel Martínez
RIA Laboratory
Metabolism & Pharmacokinetics Service
Research & Development Department
IPSEN PHARMA, S.A.
Ctra. Laureŕ Miró 395
Sant Feliu de Llobregat, Barcelona, Spain
daniel.martinez.-at-.beaufour-ipsen.com
---
From: "Stephen Duffull"
Date: Fri, 28 Jun 2002 16:34:21 +1000
To: david.at.boomer.org
Subject: Re: PharmPK Re: AUC 0-inf
Aziz
So if I'm interpreting you correctly you have some conc value that is BLQ?
When you compute AUC(all) you include some estimate of the BQL conc value
but when you compute AUC(inf) you don't.
This is not really an artifact of the data - so much as an artifact of your
analysis. If you have
LOD > conc > BQL
then you can expect to get these results if you ignore this conc when doing
regression to compute AUC(inf).
Really, you should model your data (I realise that this does not help you).
Regards
Steve
Steve Duffull
School of Pharmacy
University of Queensland
Brisbane QLD4072
Australia
---
From: "Gregorio Encina"
Date: Fri, 28 Jun 2002 08:41:37 +0200
To: david.aaa.boomer.org
Subject: Re: PharmPK AUC 0-inf
It is impossible and illogical that AUC0-inf will be less than AUC0-last or
AUC0-all. But you must be carefull with some programs like WinNonlin,
because if you put in the worksheet sampling points of elimination phase
quantified LOQ as "0" , in some cases (fast elimination and wide sampling
time period), the AUC0-all obtained could be greater than AUC0-inf.
Gregorio Encina, Ph D.
Pharmacokinetics and Drug Metabolism Dpto
Laboratorios Dr. Esteve, S.A.
Avda Mare de Deu de Montserrat, 221
08041 Barcelona, Spain
gencina.-a-.esteve.es
---
From: "McBurney, Alan"
Date: Fri, 28 Jun 2002 08:14:52 +0100
To: david.-at-.boomer.org
Subject: Re: AUC0-inf
Sorry to disagree with an earlier reply on this but yes, it is possible to
obtain an AUCinf which is less than AUCt or AUCall. There has been previous
correspondence on this matter earlier this year.
The situation arises when the sampling period includes concentrations which
are BLQ towards the end of the sampling period. If these results are
entered as zero, then the trapezoidal rule can potentially calculate an area
between the last two sample points which is greater than the extrapolated
area.
Alan
HUNTINGDON LIFE.SCIENCES LIMITED
http://www.huntingdon.com
---
From: TOUBLANC_NATHALIE.-a-.Lilly.com
Date: Fri, 28 Jun 2002 14:06:43 +0100
To: david.-at-.boomer.org
Subject: Re: PharmPK Re: AUC 0-inf
Actually, when you do NCA, if your AUC0-inf is calculated using Clast
predicted and not Clast observed, and your Clast observed is higher
than your Clast predicted, you could end up having a higher
AUC(0-inf) (calculated using your Clast pred) than your AUClast using
your Clast obs.
Regards
Nathalie
Global PK/PD and Trial Simulation
Back to the Top
Dear Dr. Abu-Qare,
Could you please explain the possibility how AUC0-inf is less than
AUC0-last or AUC0-all in non-compartmental pharmacokinetics analysis.
Back to the Top
Imagine that you have a sampling scheme where the last data point is zero
(actually BLQ and treated as zero) and the second to the last point is not
zero. If the time between the last and second to last data points is large,
the area of the last trapezoid may be greater than the extrapolated area using
the second to the last point and beta.
Greater sampling density in the tail end of the curve would alleviate this
discrepancy.
Michael
PharmPK Discussion List Archive Index page
Copyright 1995-2010 David W. A. Bourne (david@boomer.org)