Back to the Top
Is is possible to have enterohepatic recirculation or first-pass
effect after an iv or ip bolus?
SS
Back to the Top
Satyendra,
First pass effect
I would be interested in hearing other list members' definitions of
when processes, such as metabolism switch, from being classified as
"first pass" to a more general systemic role (I am guessing that this
would have to be an operational definition related to how sampling
was performed). I understand that this can include passage through
the pulmonary vasculature so there can be first pass loss of
intravenously administered compounds if e.g. pulmonary metabolism is
appreciable. After intraperitoneal administration the dose is
usually considered to enter the portal circulation (e.g. Lukas et
al., 1971) and so there is the opportunity for first pass hepatic
metabolism when dosing by this route.
Enterohepatic recirculation
As long as the dose reaches the liver, EH circulation has potential
to occur, no matter what the route of administration.
All the very best,
Bernard
Bernard Murray, Ph.D.
Senior Research Investigator
Drug Metabolism, PCS, PPD, GPRD, Abbott Laboratories, Chicago, USA
Bernard.Murray.aaa.abbott.com
Back to the Top
SS,
Yes, if the parent drug is secreted in bile, or if a metabolite is
secreted and then converted back into parent in the intestinal lumen,
then it is available for reabsorption in the gut, hence recirculation.
It doesn't matter if it came from an iv or po dose. Under the right
circumstances, oral bioavailability can appear to be greater than 100%.
Note that you can also get entero-entero recirculation with compounds
that have long half lives and relatively low plasma protein binding.
This is because after the drug is gone from the lumen, but remains in
the blood, the concentration gradient is in the direction of
blood-intestinal mucosa-lumen, so the drug can be exsorbed back into the
lumen, and then reabsorbed if/when the concentration gradient becomes
higher in lumen. This has been reported for propranolol from iv doses.
Our GastroPlus software models these phenomena for oral and iv doses.
With sufficient data, for drugs that undergo enterohepatic circulation,
you can estimate the fraction of total clearance that is recirculated.
Walt Woltosz
Chairman & CEO
Simulations Plus, Inc. (AMEX: SLP)
1220 W. Avenue J
Lancaster, CA 93534-2902
U.S.A.
http://www.simulations-plus.com
Phone: (661) 723-7723
FAX: (661) 723-5524
E-mail: walt.-at-.simulations-plus.com
Back to the Top
Bernard Murray wrote:
"I would be interested in hearing other list members' definitions of
when processes, such as metabolism switch, from being classified as
"first pass" to a more general systemic role (I am guessing that this
would have to be an operational definition related to how sampling
was performed). . ."
IMHO, first pass extraction is the additional clearance resulting from
increased concentration in a clearing organ during absorption, over the
clearance that would have taken place anyway from the concentration due
to the systemic circulation.
For example, the liver receives blood from both portal vein and hepatic
artery. When there is no absorption, the plasma concentration in both is
the systemic plasma concentration - so ther is no net "new" drug passing
through. When there is absorption from the gut, the portal vein
concentration is higher, so that when they mix in the liver, the
concentration in liver is higher than it would have been without
absorption. The difference in the clearance is caused by the presence of
the additional "new" drug added to what was already circulating.
Because the clearance in an organ can be nonlinear, the most general
approach to calculating first pass extraction is, I believe, to
calculate the clearance at the higher (combined for the liver)
concentration, subtract the clearance that would have taken place at
just the systemic concentration, and integrate the difference over time
to get the accumulated first pass extraction.
I think the process would be similar for other clearing drugs in between
the site of absorption/administration and the systemic circulation.
I am also very interested in comments from others regarding this.
Walt Woltosz
Chairman & CEO
Simulations Plus, Inc. (AMEX: SLP)
1220 W. Avenue J
Lancaster, CA 93534-2902
U.S.A.
http://www.simulations-plus.com
Phone: (661) 723-7723
FAX: (661) 723-5524
E-mail: walt.at.simulations-plus.com
Back to the Top
Walt,
You wrote: "Under the right circumstances, oral bioavailability can
appear to be greater than 100%." I would be interested in learning when
this can happen.
Best regards,
Mats
--
Mats Karlsson, PhD
Professor of Pharmacometrics
Div. of Pharmacokinetics and Drug Therapy
Dept. of Pharmaceutical Biosciences
Faculty of Pharmacy
Uppsala University
Box 591
SE-751 24 Uppsala
Sweden
phone +46 18 471 4105
fax +46 18 471 4003
mats.karlsson.-at-.farmbio.uu.se
Back to the Top
Mats,
One example is chloramphenical (when given IM), some of the
chloramphenical is metabolized by muscle enzymes (don't remember which),
given a larger AUC with oral than IM. Clearly this is an exception,
don't
know if this qualifies as "the right circumstances", seem like a one
off.
Mark
Mark Sale M.D.
Global Director, Research Modeling and Simulation
GlaxoSmithKline
919-483-1808
Back to the Top
Mark,
Thanks for the example. However, I took Walt to mean absolute, rather
than relative, bioavailability. In that case we really need to look at
an IV dose as reference.
Mats
Back to the Top
Mats:
In theory, higher bioavailability following extravascular vs. "bolus"
vascular administration may occur in any case where CL increases with
concentration. Again, in theory, this may occur in cases where there is
saturable protein binding (e.g., naproxen, valproic acid), saturable
renal
reabsorption (e.g., riboflavin), and/or saturable "systemic protection"
(e.g., as in FcRn-mediated "protection" against the catabolism of IgG).
Joe
Joseph P. Balthasar, Ph.D.
Associate Professor
Department of Pharmaceutical Sciences
University at Buffalo
The State University of New York
457B Cooke Hall
Buffalo, New York 14260
Tel: 716-645-2842, x 270
Fax: 716-645-3693
Back to the Top
Joe,
Thanks. I appreciate that AUC (of total drug) can be larger following
oral than IV if you have saturable binding. However, that doesn't mean
bioavailability is above one. Whenever you have non-linear processes,
AUCs can't be directly used for bioavailability estimation (you have to
correct for average CL during the two occasions, see expression below).
In practise, I think model-based analysis is the way to estimate F in
such situations.
AUCpo * Div * CLpo
F = ------------------
AUCiv * Dpo * CLiv
Best regards,
Mats
--
Mats Karlsson, PhD
Professor of Pharmacometrics
Div. of Pharmacokinetics and Drug Therapy
Dept. of Pharmaceutical Biosciences
Faculty of Pharmacy
Uppsala University
Box 591
SE-751 24 Uppsala
Sweden
phone +46 18 471 4105
fax +46 18 471 4003
mats.karlsson.-a-.farmbio.uu.se
Back to the Top
Mats Karlsson said
"Thanks for the example. However, I took Walt to mean absolute, rather
than relative, bioavailability. In that case we really need to look at
an IV dose as reference."
My statement was:
"Under the right circumstances, oral bioavailability can appear to be
greater than 100%."
I actually meant to say fraction absorbed, but I'll stick with
bioavailability as well.
As Mats has noted, when enough drug is recirculated, the AUC/Dose for an
oral dose can exceed that for an iv dose, giving the appearance of
higher bioavailability for the oral dose. Other explanations have been
offered here as well.
Is this really absolute bioavailability? I'll leave that for others to
answer, but I can see an argument in favor of treating it as such from
an exposure standpoint - the subject is exposed to more drug for a
longer time, with all the therapeutic and adverse consequences of a
larger dose without recirculation. It may come down to what is a useful,
practical definition - why do we talk about bioavailability to begin
with?
One way "bioavailability" can be calculated via simulation is as the
integral of the amount of drug added to the systemic circulation
("central compartment") over time, divided by the dose. If some of it
comes out of the systemic circulation and gets added back, it's the same
as if it came from a bigger dose - as far as the central compartment
goes, a drug molecule is a drug molecule (without history). If this
amount is less than the dose, then F<1. If the amount exceeds the dose
because of recirculation, then it is >1.
Since (in simple systems, and assuming similar concentrations)
recirculation should occur under both iv and oral dosing, it could
become a matter of when the measurements were taken . . .
Suppose a drug has a biliary clearance fraction (the fraction of the
total clearance that is secreted in bile either as parent or as a
metabolite that gets converted back to parent in the intestinal lumen)
of 0.XX and that all clearance is by the liver. If the iv dose is given
at a time long before the next meal, then a significant amount of drug
can accumulate in the gall bladder (unless it's a rat experiment), while
some bypasses and continuously recirculates (all does in the rat). If
all the samples are taken before the next meal, then the drug stored in
the gall bladder would not contribute to AUC. If, on the other hand, the
oral data were taken over a longer period, which included one or more
meals, then the gall bladder emptying would contribute additional drug
for reabsorption, and the resulting AUC/Dose could exceed that for the
iv dose.
Walt Woltosz
Chairman & CEO
Simulations Plus, Inc. (AMEX: SLP)
1220 W. Avenue J
Lancaster, CA 93534-2902
U.S.A.
http://www.simulations-plus.com
Phone: (661) 723-7723
FAX: (661) 723-5524
E-mail: walt.-at-.simulations-plus.com
Back to the Top
Walt,
Thanks for many informative contributions to this listserv. On this one,
I don't entirely agree, even if some (or all) is semantics. I've given
some comments below.
Best regards,
Mats
--
Mats Karlsson, PhD
Professor of Pharmacometrics
Div. of Pharmacokinetics and Drug Therapy
Dept. of Pharmaceutical Biosciences
Faculty of Pharmacy
Uppsala University
Box 591
SE-751 24 Uppsala
Sweden
phone +46 18 471 4105
fax +46 18 471 4003
mats.karlsson.at.farmbio.uu.se
--
Walt Woltosz
31 August 2004 05:39
Subject: PharmPK Re: E/H recirc and FP effect
PharmPK - Discussions about Pharmacokinetics
Pharmacodynamics and related topics
Mats Karlsson said
"Thanks for the example. However, I took Walt to mean absolute, rather
than relative, bioavailability. In that case we really need to look at
an IV dose as reference."
My statement was:
"Under the right circumstances, oral bioavailability can appear to be
greater than 100%."
I actually meant to say fraction absorbed, but I'll stick with
bioavailability as well.
As Mats has noted, when enough drug is recirculated, the AUC/Dose for an
oral dose can exceed that for an iv dose, giving the appearance of
higher bioavailability for the oral dose. Other explanations have been
offered here as well.
M> Only when you have a nonlinear disposition and are ignorant about
that fact. With linear PK, I can't still see how recirculation would
make bioavailability seem to be larger than 1. But this is maybe where
you meant "fraction absorbed".
Is this really absolute bioavailability? I'll leave that for others to
answer, but I can see an argument in favor of treating it as such from
an exposure standpoint - the subject is exposed to more drug for a
longer time, with all the therapeutic and adverse consequences of a
larger dose without recirculation. It may come down to what is a useful,
practical definition - why do we talk about bioavailability to begin
with?
M> I prefer to separate bioavailability and elimination. Otherwise, for
example, for a drug with saturable metabolism, one could say that it has
different bioavailability depending on the infusion rate of a certain
dose.
One way "bioavailability" can be calculated via simulation is as the
integral of the amount of drug added to the systemic circulation
("central compartment") over time, divided by the dose. If some of it
comes out of the systemic circulation and gets added back, it's the same
as if it came from a bigger dose - as far as the central compartment
goes, a drug molecule is a drug molecule (without history). If this
amount is less than the dose, then F<1. If the amount exceeds the dose
because of recirculation, then it is >1.
M> With this argument any drug molecule that goes into a peripheral
compartment and comes back into the central compartment can be similarly
viewed! In my mind, EH is a distribution phenomena but with some
particular characteristics, the discreteness of the process with respect
to the gall bladder emptying is one, the ease by which it can turn from
a distribution to elimination process is another.
Since (in simple systems, and assuming similar concentrations)
recirculation should occur under both iv and oral dosing, it could
become a matter of when the measurements were taken . . .
Suppose a drug has a biliary clearance fraction (the fraction of the
total clearance that is secreted in bile either as parent or as a
metabolite that gets converted back to parent in the intestinal lumen)
of 0.XX and that all clearance is by the liver. If the iv dose is given
at a time long before the next meal, then a significant amount of drug
can accumulate in the gall bladder (unless it's a rat experiment), while
some bypasses and continuously recirculates (all does in the rat). If
all the samples are taken before the next meal, then the drug stored in
the gall bladder would not contribute to AUC. If, on the other hand, the
oral data were taken over a longer period, which included one or more
meals, then the gall bladder emptying would contribute additional drug
for reabsorption, and the resulting AUC/Dose could exceed that for the
iv dose.
M> A bad experimental design will generally causes problems in
interpretation.
Walt Woltosz
Back to the Top
Hi all,
I just wanted to make my comments on some issues:
Walt >" Under the right circumstances, oral bioavailability can appear
to be
greater than 100%."
I actually meant to say fraction absorbed, but I'll stick with
bioavailability as well.
As Mats has noted, when enough drug is recirculated, the AUC/Dose for an
oral dose can exceed that for an iv dose, giving the appearance of
higher bioavailability for the oral dose. Other explanations have been
offered here as well.
Mats> Only when you have a nonlinear disposition and are ignorant about
that fact. With linear PK, I can't still see how recirculation would
make bioavailability seem to be larger than 1. But this is maybe where
you meant "fraction absorbed".
Donatas> The fraction absorbed indeed can be more than 100% (if we
stick to the concept of absorption and if drug undergoes enterohepatic
recirculation).
But it is more complicated with absolute bioavailability, because you
compare IV with PO dose. IV dose also have enterohepatic recirculation
and if you have linear kinetics E/H recirculation will be similar to
PO dose. And absolute bioavailability will be < or equal 100%.
Different situation is nonlinear kinetics. If you have 10-15 time
higher PO dose than IV, if you have good absorption, if you have
saturable PPB, or saturable metabolism, and good renal cl , then in
principle drug concentration after PO administration can be much higher
in portal vein, then after IV administration, and you can have more
effective E/H recirculation after PO administration and absolute
bioavailability >1
..
Walt > Since (in simple systems, and assuming similar concentrations)
recirculation should occur under both iv and oral dosing, it could
become a matter of when the measurements were taken . . .
Suppose a drug has a biliary clearance fraction (the fraction of the
total clearance that is secreted in bile either as parent or as a
metabolite that gets converted back to parent in the intestinal lumen)
of 0.XX and that all clearance is by the liver. If the iv dose is given
at a time long before the next meal, then a significant amount of drug
can accumulate in the gall bladder (unless it's a rat experiment), while
some bypasses and continuously recirculates (all does in the rat). If
all the samples are taken before the next meal, then the drug stored in
the gall bladder would not contribute to AUC. If, on the other hand, the
oral data were taken over a longer period, which included one or more
meals, then the gall bladder emptying would contribute additional drug
for reabsorption, and the resulting AUC/Dose could exceed that for the
iv dose.
Mats> A bad experimental design will generally causes problems in
interpretation.
Donatas> In calculating F oral AUC(infinitive)/Dose is used.
Donatas Zmuidinavicius
Pharma Algorithms
Email donatas.aaa.ap-algorithms.com
[There shouldn't be any confusion if the right model is used. Trying to
use AUC as a measure of bioavailabilty or absorption for non-linear
models is wrong. Even if 'some' correction is made for changes in
clearance you lose information. A good model that is consistent across
doses/ROA can include terms for absorption (F and ka etc...). As Walt
has pointed out before F or ka may include even more detail. That is,
be replaced by more 'fundamental' parameters if you want or need to
build a bigger model. Just one example, F may be a split into
intestinal metabolism, reverse transport, liver first-pass etc. An
inconsistent value for F (or other parameter) means you don't have the
right model - db]
Back to the Top
Dear Group
The discussion is very interesting. However it would be important to
recall
that the assumption for using the relationship AUCiv and AUCev for
bioavailabity calculation is that clearance remains constant.
Fabiana Landoni
Prof. Dra Maria Fabiana Landoni, PhD
Catedra de Farmacologia.
Facultad de Ciencias Veterinarias
Universidad Nacional de La Plata
FAX +54-221-425-7980
Calle 60 y 118 cc 296
(1900) La Plata.
ARGENTINA
Back to the Top
Thanks for excellent comments by Mats, Donatas, Fabiana, and David!
Mats:
I have learned much from your comments, along with many other thought
leaders who post frequently to this listserv (David, I wonder if you
really appreciate the true value of this service to the global
pharmaceutical community!).
Although I would agree that good experimental design could eliminate
many instances of incorrect conclusions, I'm not sure we could always
blame poor experimental design for failure to include iv samples that
did not include a meal to trigger EHC, especially when a slow absorbing
oral data set covers a longer time period than the iv samples taken down
to BLQ. There are times, such as in a large study we are involved in
right now, where a controlled release dosage form has data taken for
much longer times than the iv doses (slow absorption, sometimes
controlled release, and long half-life).
Also, I would wholeheartedly agree that EHC and EEC (entero-entero
circulation) are effectively peripheral compartment effects - at least
in a mathematical sense. The difference to me is that our typical
multi-compartment models assume the compartments are within the system,
while drug molecules that leave the system and reenter look no different
than additional dosing. So how should it be treated? I guess it depends
on where you draw the lines around the "system".
I love G. Box's saying - "All models are wrong, some models are useful"
- so any model that lets you get your work done adequately may be OK to
use. I also have a saying "Adequate, not perfect" - to me, adequate
means satisfying all real requirements to get the job done properly -
not compromising on anything essential, but not unnecessarily refined to
the nth degree to make it more "perfect".
Mats, Fabiana, & David:
As you have both noted, the AUC/Dose ratio method of calculating
bioavailability is only valid for linear systems, but even with linear
systems, you can have the effects of "apparent" bioavailability > 1
because of EHC and EEC. I think it should only occur when the amount of
drug that is recirculated is different in the iv and po dose samples. As
I noted earlier, this could happen because of sampling times (which may
be unavoidable when po data extends far beyond any useful iv data), or
when differences in gall bladder secretion amount/rate occur between the
two routes.
Because in rat the enterohepatic circulation is continuous, for linear
systems the AUC ratio would not be expected to produce oral
bioavailability > 1, nor secondary peaks. If apparent differences in
fraction absorbed and bioavailability among species are noted, this
should be considered if the drug is known to be a likely candidate for
EHC or EEC (EEC is more likely with low plasma protein binding and long
half-life).
Walt
Walt Woltosz
Chairman & CEO
Simulations Plus, Inc. (AMEX: SLP)
1220 W. Avenue J
Lancaster, CA 93534-2902
U.S.A.
http://www.simulations-plus.com
Phone: (661) 723-7723
FAX: (661) 723-5524
E-mail: walt.at.simulations-plus.com
Back to the Top
Dear colleagues,
The discussion about bioavailability and fraction absorbed is
interesting,
but also confusing. The last message of Walt Woltosz was really
clarifying
and unifying the different opinions. I would like to make a few
comments to
this last message:
> There are times, such as in a large study we are involved in
> right now, where a controlled release dosage form has data taken for
> much longer times than the iv doses (slow absorption, sometimes
> controlled release, and long half-life).
It may be still a debate whether or not such a study design is
appropriate.
In cases where EHC or EEC is known to be likely to occur, sampling
after iv
dosing should be as long as after oral dosing.
> As you have both noted, the AUC/Dose ratio method of calculating
> bioavailability is only valid for linear systems, but even with linear
> systems, you can have the effects of "apparent" bioavailability > 1
> because of EHC and EEC. I think it should only occur when the amount
of
> drug that is recirculated is different in the iv and po dose samples.
As
> I noted earlier, this could happen because of sampling times (which
may
> be unavoidable when po data extends far beyond any useful iv data), or
> when differences in gall bladder secretion amount/rate occur between
the
> two routes.
The two possibilities are quite different in a methodological sense: (1)
Sampling times can be chosen freely, and I don't see why this is
unavoidable; what do you mean with 'useful iv data'? Any measurement is
useful, even if it is below LOD or LOQ. And in case of gall bladder
emptying, one might expect that plasma concentration rises, even if it
was
below LOD.
(2) Differences in gall bladder secretion between the two routes are
indeed
a possible reason for an 'apparent' bioavailability, but is it likely to
occur? Of course it may happen in an individual, but this should level
out
in a panel of subjects. But a real difference in gall bladder secretion
between the two routes seems rather unlikely to me, provided that the
experimental design with respect to time of administration, sampling,
and
meals is sound.
Before Walt's last message arrived, I was preparing some comments to the
earlier messages:
Walt wrote:
> Is this really absolute bioavailability? I'll leave that for others to
> answer, but I can see an argument in favor of treating it as such from
> an exposure standpoint - the subject is exposed to more drug for a
> longer time, with all the therapeutic and adverse consequences of a
> larger dose without recirculation. It may come down to what is a
useful,
> practical definition - why do we talk about bioavailability to begin
> with?
We talk about bioavailability because it is a major determinant of the
plasma concentration, both after single dose and after multiple dosing.
So,
there can be hardly any discussion about the importance and meaning of
bioavailability.
IMHO, Walt's reasoning is not correct. There is not more drug for a
longer
time. As was stated by Mats, the situation is comparable to that of a
peripheral compartment, although there is a time delay due to storage
in the
gall bladder. One should compare the AUC in case of EHC with the
situation
without EHC (similar clearance, different 'apparent' volume of
distribution;
enterohepatic recirculation does not affect AUC), and not with the
situation
where the drug disappears in the same way as during EHC but without
returning; in the latter case the clearance would be much higher, of
course.
> Donatas> The fraction absorbed indeed can be more than 100% (if we
> stick to the concept of absorption and if drug undergoes enterohepatic
> recirculation).
> But it is more complicated with absolute bioavailability, because you
> compare IV with PO dose. IV dose also have enterohepatic recirculation
> and if you have linear kinetics E/H recirculation will be similar to
> PO dose. And absolute bioavailability will be < or equal 100%.
I agree. But using the term 'fraction absorbed' for all drug molecules
entering the systemic circulation via the gut, divided by number of
molecules administered makes no sense to me. Why would we bother about
this
fraction? IMHO, only bioavailability is relevant. What does it matter
if a
molecule passes 10 times the gut? Of course, it may be interesting from
a
scientific point of view to know what happens exactly, but what can we
really learn from this information? On the other hand, bioavailability
is
extremely important, since it determines the (average) plasma
concentration
pattern. Please note that this discussion resembles the discussion with
respect to 'bioavailability at the site of drug effect', a few month
ago!
> David Bourne:
> [There shouldn't be any confusion if the right model is used. Trying
to
> use AUC as a measure of bioavailabilty or absorption for non-linear
> models is wrong. Even if 'some' correction is made for changes in
> clearance you lose information.
I agree. Most of the discussion was done wih the underlying assumption
of
linear kinetics. In case of nonlinear kinetics, one has to be very
careful
in interpreting AUC and bioavailability. The procedure suggested by Walt
seems rather curious, and I would not recommend it without a firm and
sound
theoretical basis.
Finally, I fully agree with Walt with respect to the value of this
service
to the global pharmaceutical community, and his compliments to David
Bourne!
Best regards,
Hans
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.-a-.farm.rug.nl
Back to the Top
Hans - thank you for your thoughtful comments.
Hans: "It may be still a debate whether or not such a study design is
appropriate. In cases where EHC or EEC is known to be likely to occur,
sampling after iv dosing should be as long as after oral dosing."
In most cases, I would agree. But in the study we are working on now,
some of the first nonzero data points for the oral doses are many hours
after any iv sample would be above LOD. It would be a waste of time and
money, and unnecessary risk for the subjects, to stick them many
additional times to obtain a sample that would not show anything (IMHO,
any time you take a sample has a small risk associated with it). I think
it is sometimes the case that significant EHC or EEC becomes known from
data that was already taken before it was known.
Hans: "The two possibilities are quite different in a methodological
sense: (1) Sampling times can be chosen freely, and I don't see why
this is unavoidable; what do you mean with 'useful iv data'? Any
measurement is useful, even if it is below LOD or LOQ. And in case of
gall bladder emptying, one might expect that plasma concentration rises,
even if it was below LOD."
I was referring to times when it is known beforehand (say, from earlier
iv studies) that continued measurements would be below LOD because there
is not enough drug left in the system to be detected. What purpose would
it serve to continue to sample, knowing that there would be insufficient
drug left to detect? Yet the oral doses could release slowly enough to
have detectable drug still in the system many hours later.
Hans: "(2) Differences in gall bladder secretion between the two routes
are indeed a possible reason for an 'apparent' bioavailability, but is
it likely to occur? Of course it may happen in an individual, but this
should level out in a panel of subjects. But a real difference in gall
bladder secretion between the two routes seems rather unlikely to me,
provided that the
experimental design with respect to time of administration, sampling,
and meals is sound."
I agree it is unlikely, but in this thread we are talking about
something that is unlikely - apparent bioavailability > 100%. So we are
examining the possibilities for such cases. I think proper experimental
design could eliminate the majority of these occurrences, but I'm not
sure it could eliminate all of them.
Hans: "We talk about bioavailability because it is a major determinant
of the plasma concentration, both after single dose and after multiple
dosing. So, there can be hardly any discussion about the importance and
meaning of bioavailability."
Exactly my point - I was not disparaging bioavailability at all, but
trying to think about how and why we use it, and therefore what its
meaning is/should be. If it is to be able to compare plasma
concentration-time between extravascular and iv administration, then it
would seem that when the extravascular plasma concentration-time (albeit
rarely) exhibits greater AUC, it may be useful to talk about
bioavailability > 100%. Theoretically, this could even happen with Fa
less than 100%, though probably not much less. Would it be useful to
talk about F>1.0 and Fa<1.0, in the sense that it would tell us
something useful about the drug? I don't know - it's a question I am
putting to the group.
Hans: "IMHO, Walt's reasoning is not correct. There is not more drug for
a longer time. As was stated by Mats, the situation is comparable to
that of a peripheral compartment, although there is a time delay due to
storage in the gall bladder. One should compare the AUC in case of EHC
with the situation without EHC (similar clearance, different 'apparent'
volume of distribution; enterohepatic recirculation does not affect
AUC), and not with the situation where the drug disappears in the same
way as during EHC but without returning; in the latter case the
clearance would be much higher, of course."
I'm interested in your statement that AUC is not affected by EHC (would
this also include EEC?). I'm guessing that you are coming from a
perspective of fitting data. When we fit models, must fit the data,
which has its AUC. So whether we fit with or without EHC, AUC does not
change (except with the goodness of the fit of each model).
But this is not the same as asking "If we have a drug that has certain
absorption and PK characteristics (Vc, CL, K12, K21, etc.), and twin
drug (perhaps a different enantiomer) with the same characteristics but
also happens to undergo EHC, will the AUC of the two be different?"
Lets consider a drug that is cleared only by the liver. We model EHC by
assigning a portion of liver clearance to EHC. Suppose we use 50%. Then
50% of the liver clearance is sent to the gall bladder - or I should say
toward the gall bladder. We allow this amount to be split into two parts
- one part is stored in the gall bladder compartment for later release,
and the remainder is sent directly to the duodenum compartment and made
available for reabsorption.
If we look at an extreme case, we could send all liver clearance to the
gall bladder with no bypass, and just let it stay there. The plasma
concentration-time curve would be identical to one that did not send any
drug to the gall bladder (the same amount was cleared from the system).
Now lets change things a little. Some time after dosing, say, 6 hours,
there is a meal given and the gall bladder empties into the duodenum
over some time period (perhaps 30 minutes). Now the drug that we
collected before has an opportunity to get back into the system. Will
AUC change? Yes, it will increase significantly. So does EHC change AUC?
I think it does if the question is interpreted as meaning if all else is
the same, does a molecule that undergoes EHC have a different AUC than
its near twin that does not.
Hans: " . . . using the term 'fraction absorbed' for all drug molecules
entering the systemic circulation via the gut, divided by number of
molecules administered makes no sense to me. Why would we bother about
this fraction?"
As I understand it, the term "fraction" is the ratio of amount of drug
absorbed divided by the dose. So if we want to ask "Was the amount of
drug that was absorbed greater than the dose?" then it might have
meaning. Perhaps only 90% of the original dose molecules were absorbed,
but perhaps 20% of that amount was secreted or exsorbed back into the
lumen and later (90%) reabsorbed. So effectively, we had 90 + .2*90 108% fraction absorbed. But is it fair or useful to think of it this
way? In the absence of circulation, how could we duplicate the effect?
Taking our twin molecule from above (alike in all respects except
recirculation), we could duplicate the recirculated Cp-time by dosing an
additional amount of the twin drug.
Again I ask, what is the reason we talk about fraction absorbed? If the
way we use it means that values > 100% are useful, then perhaps such
values should be allowed. If not, then we should have a way to deal with
those situations where it appears to be > 100%. If we go to a reference
some years after an experiment and see Fa = 90% or Fa=108%, which would
be more enlightening?
Hans: "Most of the discussion was done with the underlying assumption of
linear kinetics. In case of nonlinear kinetics, one has to be very
careful in interpreting AUC and bioavailability. The procedure suggested
by Walt seems rather curious, and I would not recommend it without a
firm and sound theoretical basis."
We had considerable internal discussion about this, as well as talking
about it with our external advisors. Perhaps this is another case of
defining terms. I have assumed that we want to say that bioavailability
refers to how much drug entering from an extravascular dose reaches the
systemic circulation.
Consider a simple case of an oral dose with nonlinear clearance only by
the liver. At first, there is no drug in the system, so drug that is
absorbed enters from the portal vein (which previously had no drug) into
the liver, and mixes with blood from the hepatic artery that as yet has
no drug. So for a short time, all of the clearance of drug in the liver
is first pass extraction, and the contribution to bioavailability is
what survives the first pass through the liver. This drug will circulate
and come back to the hepatic artery and portal vein after a couple of
minutes or so. Then the blood coming from the hepatic artery (and portal
vein) will contain some amount of drug that has already undergone first
pass extraction, so it was already bioavailable.
If absorption had stopped, all clearance at that point would be just
routine systemic clearance (not first pass). Both portal vein and
hepatic artery concentration would be the same - what we refer to as the
central compartment concentration (typically Cp, absent drug in
hematocrit). But if absorption is still happening, then the
concentration reaching the liver from the portal vein will contain both
drug that has already undergone first pass extraction as well as drug
that was just absorbed. At the liver, this will be mixed with drug from
the hepatic artery at Cp.
Now the question is, how much of this mixed drug is undergoing first
pass extraction, and how much is not. We propose that the drug that was
already in the system was already bioavailable. The just-absorbed drug
entering the liver for the first time is subject to first pass
extraction, so not all of it will be bioavailable. Absent absorption,
the nonlinear clearance based on Cp would have a certain value. When the
new drug is added by absorption, the clearance increases (if not
flow-limited). We propose that this difference in clearance is
effectively the first pass extraction of the newly absorbed drug, and
that this can be integrated over time to calculate a net first pass
extraction, and with that known, the bioavailability.
We would be interested in counter arguments to this approach, as to date
we have heard none (but I wish I had a dollar for every raised
eyebrow!).
Walt
Walt Woltosz
Chairman & CEO
Simulations Plus, Inc. (AMEX: SLP)
1220 W. Avenue J
Lancaster, CA 93534-2902
U.S.A.
http://www.simulations-plus.com
Phone: (661) 723-7723
FAX: (661) 723-5524
E-mail: walt.at.simulations-plus.com
Back to the Top
Walt,
I liked your second to last email as a nice conclusion to the
discussion. Particularly the "point of view" and that you talked about
"apparent" bioavalability being possibly being >100% (which I took was
due to experimental setup's sometimes failing to capture the terminal
portion of the curve). However, in the latest mail you again discuss as
if bioavailability truly could be >100% even with perfect design. Maybe
what you refer to here is when there is variability in CL between study
occasions, which again in a na=EFve analysis can appear as
>100% bioavailability (but certainly is not). Otherwise, I still fail to
see the situation when it can occur.
In your example of the "twin" drugs, you actually take away an
elimination pathway when you create your second drug, so it is not
surprising if AUC changes. If you add or take away an EHC without
changing elimination pathways, AUC will not change.
Lastly, you write " I'm guessing that you are coming from a
perspective of fitting data."
I'm pretty sure that Hans fits models, not data!
:-)
Best regards,
Mats
--
Mats Karlsson, PhD
Professor of Pharmacometrics
Div. of Pharmacokinetics and Drug Therapy
Dept. of Pharmaceutical Biosciences
Faculty of Pharmacy
Uppsala University
Box 591
SE-751 24 Uppsala
Sweden
phone +46 18 471 4105
fax +46 18 471 4003
mats.karlsson.aaa.farmbio.uu.se
Back to the Top
Mats,
Thank you for your comments:
Mats: "However, in the latest mail you again discuss as if
bioavailability truly could be >100% even with perfect design. Maybe
what you refer to here is when there is variability in CL between study
occasions, which again in a naive analysis can appear as >100%
bioavailability (but certainly is not). Otherwise, I still fail to see
the situation when it can occur."
What I was suggesting is that it may be useful to use the term
bioavailability to describe situations like the one I offered wherein
recirculation results in the same effect as an additional dose.
Molecules can be metabolized, secreted in bile, converted back into
parent, and reabsorbed as though another amount of drug had been dosed
orally. So in the case of linear kinetics, the AUC/Dose ratio between
oral and iv doses might be > 1. If you went to a reference years later
and saw F=90% or F=108% in my example, which would be more enlightening?
I think it comes down to how one interprets the definition of
bioavailability. Suppose, for example, that 80% of a dose is absorbed,
but 99% bioavailability is calculated using AUC/Dose ratios because of
EHC. Would you allow that 99% should be reported as the
bioavailability, even though only 80% was absorbed? If so, then what
chanes when the calculated bioavailability becomes 101%?
Mats: "In your example of the "twin" drugs, you actually take away an
elimination pathway when you create your second drug, so it is not
surprising if AUC changes. If you add or take away an EHC without
changing elimination pathways, AUC will not change."
The assumption was that total liver clearance was the same. The amount
going to the gall bladder would also be the same, but the twin drug's
metabolite would not be converted back to parent and reabsorbed. So just
as much would be secreted in bile, but none would be reabsorbed.
Mats: "Lastly, you write " I'm guessing that you are coming from a
perspective of fitting data." I'm pretty sure that Hans fits models, not
data! :-)"
Semantics, I think. We fit models to the data, so we here tend to say it
either way. When my wife goes to get a new dress, she says she is
getting fitted - but it's the dress that's changing, not her. Now it's
true, she looks like a model (IMHO), but now we're really getting into
word games! ; )
Walt Woltosz
Chairman & CEO
Simulations Plus, Inc. (AMEX: SLP)
1220 W. Avenue J
Lancaster, CA 93534-2902
U.S.A.
http://www.simulations-plus.com
Phone: (661) 723-7723
FAX: (661) 723-5524
E-mail: walt.-a-.simulations-plus.com
[Should I get into this one ;-) Walt's use of "...perspective of
fitting data" I believe implies "...perspective of fitting data with a
model". A common usage. The more universally incorrect statement would
be "fitting data to a model" which should be replaced with "fitting a
model to data" - db]
Back to the Top
Walt,
I liked your comment about our different view of the system as the root
to any differences in our way of describing pk processes. I am more
inclusive and I tend to see the intestinal contents as part of the body.
Therefore, molecules going into an EHC are not cleared by the liver,
merely participating in a distribution process. Thus, I think it is not
a question about being correct or not (anyone is entitled to make their
definitions), but which is the most useful way of describing the system.
I was taught PK from Rowland&Tozer's "Clinical Pharmacokinetics" and I
still adopt their definitions. I think they are useful because they,
without compromising physiology, focus on what is possible to obtain
from standard data and what is desirable to know to in order to provide
appropriate drug treatment. If you take an iv dose of a drug undergoing
EHC, with these definitions, you can use:
CL=Dose/AUC to estimate CL, and
AUC=Dose/CL to predict the systemic exposure
I don't know what you gain by instead defining EHC as a clearance plus
absorption process. Certainly CL cannot be as simply estimated as above
and fractions, according to the old definitions, seem to take on values
above one, which really does not make sense. Therefore, a new set of
parameter definitions, adopted to this view, seems be needed. [or else
we will go on with discussions like this forever]
So all the above is semantics. However, what I still don't get is when
you write "So in the case of linear kinetics, the AUC/Dose ratio between
oral and iv doses might be > 1." Assuming that the concentration you
measure is not in portal vein, I don't think EHC can result in this and
you haven't described how that may happen, apart from the example when a
sub-optimal design results in failure to capture the AUCs appropriately.
As a final semantic, sometimes I try to fit into my clothes (being no
model) and sometimes I make the clothes fit me. It can be useful to be
able to make the distinction.
Best regards,
Mats
--
Mats Karlsson, PhD
Professor of Pharmacometrics
Div. of Pharmacokinetics and Drug Therapy
Dept. of Pharmaceutical Biosciences
Faculty of Pharmacy
Uppsala University
Box 591
SE-751 24 Uppsala
Sweden
phone +46 18 471 4105
fax +46 18 471 4003
mats.karlsson.-at-.farmbio.uu.se
Back to the Top
Mats: "However, what I still don't get is when you write "So in the case
of linear kinetics, the AUC/Dose ratio between oral and iv doses might
be > 1." Assuming that the concentration you
measure is not in portal vein, I don't think EHC can result in this and
you haven't described how that may happen, apart from the example when a
sub-optimal design results in failure to capture the AUCs
appropriately."
It can happen under conditions of slow oral absorption - controlled
release is an extreme case. You can have oral plasma concentration-time
data that goes on long after the iv data is below LOD. Taking additional
iv data is fruitless and provides no information. But taking the oral
data at time points that are above LOD does provide information. For
long sampling times, EHC or EEC has longer to operate, so although study
design could not avoid it, the apparent bioavailability calculated based
on AUC shows a value greater than it would if the same drug had no EHC
or EEC. The question is, what does it mean? If it is calculated as
F=99%, it seems that no one worries about it (even if Fa is less). But
if it becomes > 100%, then the alarms are sounded.
I still think that a reference that reported F=108% when Fa=90% (as in
my earlier example) would be more useful than one that reported a lower
bioavailability with the same Fa - at least the 108% provides
information that some kind of recirculation is probable.
Mats: "As a final semantic, sometimes I try to fit into my clothes
(being no
model) and sometimes I make the clothes fit me. It can be useful to be
able to make the distinction."
Me, too . . . !! Isn't it odd that with all the advances in the fashion
industry, clothes still shrink as we get older? ; )
Best regards,
Walt
Walt Woltosz
Chairman & CEO
Simulations Plus, Inc. (AMEX: SLP)
1220 W. Avenue J
Lancaster, CA 93534-2902
U.S.A.
http://www.simulations-plus.com
Phone: (661) 723-7723
FAX: (661) 723-5524
E-mail: walt.aaa.simulations-plus.com
Back to the Top
Walt,
What you describe is a situation where the experimental conditions
(sampling times, assay limits) make AUCpo appear higher than AUCiv while
with perfect monitoring it would not have done so.
It is important to appreciate situations that may provide difficulties
in interpretation and this is one I haven't come across before. Thanks
for pointing it out.
Best regards,
Mats
--
Mats Karlsson, PhD
Professor of Pharmacometrics
Div. of Pharmacokinetics and Drug Therapy
Dept. of Pharmaceutical Biosciences
Faculty of Pharmacy
Uppsala University
Box 591
SE-751 24 Uppsala
Sweden
phone +46 18 471 4105
fax +46 18 471 4003
mats.karlsson.-a-.farmbio.uu.se
Back to the Top
Dear Mats,
I agree with your view in the discussion on EHC, bioavailability and
fraction absorbed.
> Lastly, you write " I'm guessing that you are coming from a
> perspective of fitting data."
> I'm pretty sure that Hans fits models, not data!
> :-)
Thank you for your confidence in my work!
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.-a-.farm.rug.nl
PharmPK Discussion List Archive Index page
Copyright 1995-2010 David W. A. Bourne (david@boomer.org)