Back to the Top
May I ask the forum for their views on the practicality of a possible
individual body weight correction factor in bioequivalence studies? I
am asking because we conduct bridging studies in Japanese and Caucasian
volunteers and the Japanese tend to weigh less. Companies try to get
round this by narrow selection criteria for body weight or Mass Index,
so as to limit the lower weights of Japanese and upper Caucasians which
can sometimes drastically affect recruitment. In a current study of 15
per group the group weights are still not matched very well despite the
inclusion restrictions, so would such a factor help to correct any
difference so engendered? How much individual Pk data would we need to
have a reliable estimate of such a factor? I guess that around 100
would be more than enough to at least compute a regression equation
between AUC for example and body weight, but would it have much
meaning? Could we manage with less?
With thanks
Andrew Sutton
Andrew Sutton, MBBS, MD(London), FFA
Medical Director
Guildford Clinical Pharmacology Ltd.
Hascombe Ward, Royal Surrey County Hospital,
Guildford, Surrey GU2 7YXX, UK
Tel: +44 (0)1483 406886
Back to the Top
The following message was posted to: PharmPK
Dear Andrew,
I assume by the time the type of bridging study you mention takes
place, a great deal is known about the effect of weight on the
particular compound of interest. This can be used to justify matching
or not matching the subjects for BW in each case. One could also dose
according to BW, although it might not be very practical. If there are
no non-linearities, one can estimate different parameters normalized
with regard to dose.
Toufigh Gordi
Back to the Top
The following message was posted to: PharmPK
Andrew Sutton wrote:
>
> May I ask the forum for their views on the practicality of a possible
> individual body weight correction factor in bioequivalence studies? I
> am asking because we conduct bridging studies in Japanese and Caucasian
> volunteers and the Japanese tend to weigh less. Companies try to get
> round this by narrow selection criteria for body weight or Mass Index,
> so as to limit the lower weights of Japanese and upper Caucasians which
> can sometimes drastically affect recruitment. In a current study of 15
> per group the group weights are still not matched very well despite the
> inclusion restrictions, so would such a factor help to correct any
> difference so engendered? How much individual Pk data would we need to
> have a reliable estimate of such a factor? I guess that around 100
> would be more than enough to at least compute a regression equation
> between AUC for example and body weight, but would it have much
> meaning? Could we manage with less?
IMHO it would be foolish to attempt to determine the relationship
between weight and AUC when the answer is know from extensive prior
knowledge and theory. CL is related to body weight via the allometric
relationship:
CL= CLstd*(WT/WTstd)**0.75
where CLstd is the CL in a person of standard weight (WTstd e.g. 70kg)
and CL is the clearance expected in a person of weight WT.
The relationship between WT and AUC will be the reciprocal of CL i.e.
AUC= 1/CL = 1/(CLstd*(WT/WTstd)**0.75)
Nick Holford, Dept Pharmacology & Clinical Pharmacology
University of Auckland, 85 Park Rd, Private Bag 92019, Auckland, New
Zealand
email:n.holford.aaa.auckland.ac.nz tel:+64(9)373-7599x86730 fax:373-7556
http://www.health.auckland.ac.nz/pharmacology/staff/nholford/
Back to the Top
The following message was posted to: PharmPK
Dear Andrew
In addition to replies by others you may try following. However
this needs to be written in the protocol before you can use this
approach
1. You can make a subgroup analysis and conclude BE based on the
same.
2. If just subgrouping is not working try normalising the given
dose by weight. Thus each volunteer will have a different dose in
mg/kg depending on his weight. At the same time you are still
comparing the same strength formulations.
Could please reply if any of these works?
with best regards
Dr.Prashant
Ph.D.
Back to the Top
The following message was posted to: PharmPK
Dear Toufig and Nick
Surprisingly, these Bridging studies are usually done quite early in
Phase 1
when little is known about the effect of body weight on the Pk profile.
That
is exactly why I wondering if there is any predictive mileage in using
the
correction factor I am dreaming of. I agree that dosing by body weight
is
not practical as the company's whole reason for doing the trial is to
compare Japanese and Caucasian bioavailability from a new capsule or
tablet
formulation
I appreciate the reminder of the equation Nick. Do you think it has any
application in this setting?
Andrew
Back to the Top
The following message was posted to: PharmPK
Dear Nick,
>> AUC= 1/CL = 1/(CLstd*(WT/WTstd)**0.75)
Is this also applicable for Cmax?
Another question:
Sometimes WT does not seem to affect AUC & Cmax so much,
i.e. WT**0.75, when PK are compared between the two populations
without narrowing the entry criteria for WT.
Does the WT normalization have any meaning in such case?
Regards,
Masaki Hiraoka
Back to the Top
The following message was posted to: PharmPK
Hello,
I'm joining in a little late in the discussion, but was wondering - is
not bioequivalence based on a ratio of test to reference in the same
individual? Would not weight be irrelevant in such a case?
Just asking.
Edmond Edwards, Ph.D.,
EDIT Research
Back to the Top
The following message was posted to: PharmPK
A simpler approach could be to normalize the PK parameter being tested
by
dose/kg or even by body surface area. Body weight differences are
commonly
encountered in studying gender differences in PK. Usually exposure in
females seems higher than males since the dose administered to both
groups
is the same and females are generally lighter than males. In these
cases,
the parameters like AUC and Cmax are normalized for body weight or body
surface area and then statistically tested. I have a couple of
references if
you are interested.
Clapton Dias, Ph.D.
Pharmacokineticist II,
Global Product Development Services,
PRA International
16400 College Boulevard,
Lenexa, Kansas-66219.
913-227-7257
Back to the Top
The following message was posted to: PharmPK
Andrew Sutton wrote:
....
> I appreciate the reminder of the equation Nick. Do you think it has
> any
> application in this setting?
> Andrew
Andrew,
You seem to have missed the point of my earlier reply. The relationship
between weight and clearance is already known based on sound biology,
mathematical theory and data ranging from elephants to mice (with
humans of various sizes in between). I strongly urge you to read the
literature on this topic before engaging in empirical experimentation.
Anderson BJ, van Lingen RA, Hansen TG, Lin YC, Holford NHG.
Acetaminophen developmental pharmacokinetics in premature neonates and
infants: a pooled population analysis. Anesthesiology
2002;96(6):1336-45.West GB, Brown JH, Enquist BJ. A general model for
the origin of allometric scaling laws in biology. Science
1997;276:122-26.
Peters R. The ecological implications of body size. Cambridge:
Cambridge University Press; 1983
West GB, Brown JH, Enquist BJ. The fourth dimension of life: fractal
geometry and allometric scaling of organisms. Science
1999;284(5420):1677-9.
http://www.anaesthetist.com/physiol/basics/scaling/Kleiber.htm
IMHO it is a waste of resources to attempt to rediscover these basic
facts of life. "Those who do not study history are destined to repeat
its mistakes"
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/
Back to the Top
The following message was posted to: PharmPK
"M.Hiraoka" wrote:
>
> PharmPK - Discussions about Pharmacokinetics
> Pharmacodynamics and related topics
>
> The following message was posted to: PharmPK
>
> Dear Nick,
>
>>> AUC= 1/CL = 1/(CLstd*(WT/WTstd)**0.75)
>
> Is this also applicable for Cmax?
Cmax is determined by several parameters e.g. CL, V, absorption
half-life. Allometric scalling models are different for CL and V.
If Cmax is largely determined by V (i.e. assume very rapid input) then:
Cmax~=Dose/V=Dose/(Vstd*WT/WTstd)
See my earlier posting to Andrew Sutton with literature references to
PK and allometry.
> Another question:
> Sometimes WT does not seem to affect AUC & Cmax so much,
> i.e. WT**0.75, when PK are compared between the two populations
> without narrowing the entry criteria for WT.
> Does the WT normalization have any meaning in such case?
IMHO WT *always* without exception is a determinant of CL and V. Your
experimental data may not confirm this because of confounding from
other covariates e.g. body composition, age, disease factors.
Nick
Nick Holford, Dept Pharmacology & Clinical Pharmacology
University of Auckland, 85 Park Rd, Private Bag 92019, Auckland, New
Zealand
email:n.holford.aaa.auckland.ac.nz tel:+64(9)373-7599x86730 fax:373-7556
http://www.health.auckland.ac.nz/pharmacology/staff/nholford/
Back to the Top
The following message was posted to: PharmPK
As far as BE is concern, using the cross over design where one patient
would get both drugs, the weight factor never enter the equation. Any
correcting would be frown upon as dressing up the data.
Back to the Top
The following message was posted to: PharmPK
Dear Vuong
The problem is that by definition your suggestion does not apply to
inter-racial comparisons.
Andrew
Back to the Top
The following message was posted to: PharmPK
Nick. I don't think I'm the one who has missed the point, but perhaps I
haven't expressed myself clearly. You seem to be agreeing that it is
logical
to use the knowledge you have so expertly described in the manner I am
hoping for, but I can see huge difficulties in defining the path of
justifying the application. You are just the kind of scientific expert
who
could help achieve this.
Specifically, could it help us to widen the weight ranges of included
volunteers when we compare a generally lighter Japanese population with
a
heavier Caucasian group because that would make it easier to find the
required number of Japanese volunteers yet still demonstrate
bioequivalence
when the data is analysed. I find that Japanese companies especially
want to
recruit very narrow weight range groups in an effort to guarantee
demonstrating bioequivalence. It's as if they think there are hundreds
of
possible Japanese volunteers in London just waiting to go into
trials....
By extension I suppose it could apply to other BE studies, but I am not
asking the question from a simple BE point of view.
Thanks
Andrew
Back to the Top
The following message was posted to: PharmPK
Hi Clapton,
Yes, I would be glad of the references if they help justify recruiting
wider
weight ranges in the two racial groups.
Thanks a lot.
Andrew
Back to the Top
The following message was posted to: PharmPK
Dear Prashant,
Unfortunately your suggestions don't apply because in bridging studies
we
are usually restricted to one group per population of Japanese and
Caucasians and one fixed dose precisely because we are asked to show
that
the specific formulation is bioequivalent. Some studies will use
different
dose levels but even they are not on a mg/kg basis, but in steps of the
formulation and again the key comparisons between the races are at the
same
step levels.
Sorry I cannot make a more constructive reply. I think it's a case of
"Watch
this forum space"
Best regards
Andrew
Back to the Top
The following message was posted to: PharmPK
Dear Dias:
Please could you mention the references?
Thanks,
Silvia
Back to the Top
The following message was posted to: PharmPK
Andrew,
Andrew Sutton wrote:
> Nick. I don't think I'm the one who has missed the point, but perhaps I
> haven't expressed myself clearly. You seem to be agreeing that it is
> logical
> to use the knowledge you have so expertly described in the manner I am
> hoping for, but I can see huge difficulties in defining the path of
> justifying the application. You are just the kind of scientific expert
> who
> could help achieve this.
I doubt it I could be helpful. My expert opinion would not only point
out that allometry has the answer but that pharmaceutical and
regulatory 'scientists' who did not agree with the proposed approach
need a brain upgrade. IMHO the use of allometric weight adjustment of
AUC has much stronger scientific grounds than than the uniformly
accepted log transformation of parameters for bioequivalence analysis.
Note that I think the log transformation is quite sensible -- among
other things because it does not assume negative AUC values are
possible like the normal distribution would -- but has little
experimental support in comparison to the weight of allometric evidence.
> Specifically, could it help us to widen the weight ranges of included
> volunteers when we compare a generally lighter Japanese population with
> a
> heavier Caucasian group because that would make it easier to find the
> required number of Japanese volunteers yet still demonstrate
> bioequivalence
> when the data is analysed. I find that Japanese companies especially
> want to
> recruit very narrow weight range groups in an effort to guarantee
> demonstrating bioequivalence. It's as if they think there are hundreds
> of
> possible Japanese volunteers in London just waiting to go into
> trials....
> By extension I suppose it could apply to other BE studies, but I am
> not
> asking the question from a simple BE point of view.
I see no reason not to widen the weight ranges for your BE studies. A
more critical and somewhat tricky selection issue would be to try to
get 'normal' body composition. The allometric scaling model is based on
a 'normal' body composition in relation to weight. I would not expect
allometry to correctly prediction size associated differences in
clearance if Sumo wrestlers (with highly abnormal body composition
based on visual appearances) were selected in order get some heavier
subjects!
It should also be recognized that the contribution of weight to between
subject variability in clearance (and thus in AUC) is probably
negligible in relation to seemingly random differences that are typical
of most drugs. If you have say a variability of 10% in weight and 30%
variability due to other random effects then the total variability will
be sqrt(10*10 + 30*30)=32%. If the variability in WT was doubled to 20%
the overall variability would only increase from 32% to 36%. This 4%
increase in expected variability is unlikely to be critical for
determination of bioequivalence.
--
Nick Holford, Dept Pharmacology & Clinical Pharmacology
University of Auckland, 85 Park Rd, Private Bag 92019, Auckland, New
Zealand
email:n.holford.-a-.auckland.ac.nz tel:+64(9)373-7599x86730 fax:373-7556
http://www.health.auckland.ac.nz/pharmacology/staff/nholford/
Back to the Top
The following message was posted to: PharmPK
Dear Andrew,
> I find that Japanese companies especially want to recruit
> very narrow weight range groups in an effort to guarantee
> demonstrating bioequivalence.
> It's as if they think there are hundreds of possible
> Japanese volunteers in London just waiting to go into
> trials....
I guess the purpose of your study is to show similarity in
PK of an agent between Japanese and Caucasians. And also I
guess the primary endpoint is the ratios of the AUC and
Cmax within the criteria for bioequivalence. Therefore you
are trying to reduce the differences/variances in BW, BMI, etc.
between the two groups of subjects.
Assuming these my guesses are correct and if it were my case,
the first thing I should consider is that the narrower the
subject demographics become the lower provability the subjects
represent the real populations. This means the extrapolation
from the study results to real populations would become
less reliable.
Then I should discuss with my clients about another primary
endpoint that sounds scientifically good and meet with the
purpose of the study.
Regards,
Masaki Hiraoka
Back to the Top
The following message was posted to: PharmPK
On 5/9/03 10:34 pm, "Nick Holford"wrote:
Andrew,
IMHO the use of allometric weight adjustment of AUC has much stronger
scientific grounds than than the uniformly accepted log transformation
of
parameters for bioequivalence analysis.
and
It should also be recognized that the contribution of weight to between
subject variability in clearance (and thus in AUC) is probably
negligible in
relation to seemingly random differences that are typical of most
drugs. If
you have say a variability of 10% in weight and 30% variability due to
other
random effects then the total variability will be sqrt(10*10 +
30*30)=32%.
If the variability in WT was doubled to 20% the overall variability
would
only increase from 32% to 36%. This 4% increase in expected variability
is
unlikely to be critical for determination of bioequivalence.
and
I see no reason not to widen the weight ranges for your BE studies. A
more critical and somewhat tricky selection issue would be to try to
get
'normal' body composition. The allometric scaling model is based on a
'normal' body composition in relation to weight. I would not expect
allometry to correctly prediction size associated differences in
clearance
if Sumo wrestlers (with highly abnormal body composition based on visual
appearances) were selected in order get some heavier subjects!
Nick,
Many thanks for your comments. I agree with you on all of them. On
body composition we attempt to standardise to some extent by making BMI
one
of the inclusion criteria. A typical range would be 20 to 25, which
would
exclude the Sumo wrestler and the anorexic... Your comments and the
calculation almost suggest that we could dispense with body weight
altogether and replace it by BMI alone.
Masaki Hiraoka also wrote>
.....the first thing I should consider is that the narrower the subject
demographics become the lower probability the subjects represent the
real
populations. This means the extrapolation from the study results to real
populations would become less reliable. Then I should discuss with my
clients about another primary endpoint that sounds scientifically good
and
meet with the purpose of the study.
Masaki
I again agree wholeheartedly and hope that BMI fulfils that role.
I would add that if anyone at the FDA, Koseicho or MHRA would like to
comment, that too could be very helpful, since fear of their reaction
is
used by company executives to justify the belt and braces approach we
get by
using both narrow body weight ranges and BMI.
If anyone in the forum should have some references on the relative
effects
of weight and BMI that too would be very helpful.
Incidentally we find in London that there are many more Japanese women
wanting to volunteer for clinical trials than men... so I am starting to
wonder just how big gender differences are and whether body composition
also
outweighs them. If BMI would overcome that to the same degree as body
weight
as Nick Holford has shown, their inclusion would certainly make the
study
population more representative of the end users..!
Many thanks
Andrew
Back to the Top
The following message was posted to: PharmPK
Dear Andrew Sutton,
I agreed that body weight variation will affect the volume of
distribution and so also many PK parameters in some way or the other.
But i donot see any regulatory authorty considering the application of
the body weight correction factors to conclude/document the
bioequivalence.Most of the BE studies are crossover in design so
anyways this variation ( body weight) is going to be reduced to large
extent.
In bioavailability studies, it makes sense to use it in early
development phases of new drugs ( Phase I) to validate the PK of new
investigational drug .Here one has to understand the difference between
the two ( Be and BA).
I hope this makes sense,
Regards
Pradeep Bhadauria
Ranbaxy Reserach Lboratories,
INDIA.
Back to the Top
The following message was posted to: PharmPK
Pradeep,
You are missing the primary basis for the study which is a comparison
between races. Although the objective is to prove BE, this is not a
crossover study, it is an effect of race study.
Clapton Dias, Ph.D.
Pharmacokineticist II,
Global Product Development Services,
PRA International
16400 College Boulevard,
Lenexa, Kansas-66219.
Ph: 913-227-7257
Fax:913-599-2753
Back to the Top
The following message was posted to: PharmPK
I do not understand the reason for the distinction of race as medicine
is always administered as mg/m2. Where race is a factor is in
metabolism and not in weight.
Back to the Top
The following message was posted to: PharmPK
Vuong Trieu wrote:
>
> I do not understand the reason for the distinction of race as medicine
> is always administered as mg/m2. Where race is a factor is in
> metabolism and not in weight.
>
I do not agree.
First of all very few medicines are dosed on a mg/m2 basis. It is only
for some specialist treatments (e.g. oncology) and sometimes for some
paediatric treatments that mg/m2 is used.
Secondly, there are very clear differences between races in their body
composition e.g. Asians tend to be lean while Polynesians tend to be
fat. An Asian and a Polynesian of the same weight may have quite
difference clearance and volume of distribution because 1) fat does not
eliminate drugs so the component of weight due to fat will change
weight but not clearance
2) fat may increase Vd for lipophilic drugs or may not affect Vd for
hydrophilic drugs so depending on the drug there may be an increase or
decrease in Vd depending on the %fat even if the weight is the same.
--
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/
Back to the Top
The following message was posted to: PharmPK
On 10/9/03 5:30 am, "b pradeep"wrote:
Hi Pradeep,
I see that Clapton Dias has already answered your point below and I
agree
with him that by definition these Bridging Studies cannot be crossover
designs. Therefore we do need to get around the problem that the
Japanese
we see in London at least, tend to be lighter in body weight than our
Caucasians. Such studies are required by the regulatory authorities in
Japan in particular to establish the similarity (or otherwise) of Pk
profiles in the 2 populations. This is largely due to the relatively
high
(~20%) incidence of so-called slow metabolisers in the Japanese
population.
Accordingly, if they will accept a body weight correction factor, or
better
still, to rely on BMI and drop BW, then life will be easier for those
of us
who conduct these trials.
Nick Holford has put forward a telling argument in favour of doing this
and
his Asian-Polynesian comparison is a good illustration, I think, of why
BMI
would be better than body weight.
Best regards
Andrew
Back to the Top
The following message was posted to: PharmPK
If I read correctly, you are trying to normalize body weight in hope of
getting rid of the slow metabolizer effect. If the effect of slow
metabolism is not due to something as simple as body fat, you will not
be able to do so. If it is body fat then you probably should look up
the materials on propofol and such. There you will see that women tend
to have higher clearance than men (and their weight is smaller too!).
If you can elaborate on the cause of slow metabolism maybe everyone will
be able to think this through better. What you are trying to do may not
be appropriate for the problem you have on hand. Just my 2 cents worth.
Back to the Top
The following message was posted to: PharmPK
Andrew,
I would propose to include weight as an independent variable in the
model. This will absorb the variability on the kinetics that is due to
weight, reduce the error term and yield a more precise BA ratio, even
when the two groups are perfectly weight-matched.
Since the determination of the BA ratio is commonly done on log scale,
the linear model can be defined as follows :
Log(PK parameter)= race + log(weight) + error
The additional convenience of logging both PK parameter and weight is
that
the estimate of log(weight) will be easily interpretable for AUC and
clearance. It is expected that if clearance will be fit to the data,
the estimate will be positive, whereas negative for AUC and Cmax. This
trick doesn't work on the original scale where you would need to fit two
models :
1) Clearance= race + weight
2) AUC or Cmax= race + 1/weight
It would however be a good idea to weight-match the individuals. You
will probably end up with a mean difference of at most 10kg.
I'm not aware of any (regulatory) objections against the above model but
I only observe that similar things are done in safety or efficacy
trials.
Regards,
Paul Meyvisch
Biostatistician
Back to the Top
The following message was posted to: PharmPK
Dear Paul,
I agree that in an ideal world the best procedure is to match the
weights
of the Japanese and Caucasian populations, but it is an impossible
counsel
of perfection to go as far as setting the recruitment criteria to a 10kg
range. It would take so long to find the subjects that it would be
impractical and very expensive.
Moreover, Nick Holford for example, has shown that body weight is a
minor
consideration compared with Body Mass Index ( height /weight squared)
and
that is because the major determinant of AUC is clearance. In his Asian
vs
Polynesian example if you relied only on body weight you could compare a
tall muscular Japanese with a much shorter but rotund South Sea
Islander of
the same overall weight but he would have a smaller clearance rate and a
tendency with lipophilic compounds in particular to have a very
different
distribution pattern.
Perhaps in a way I'm asking the forum if we can identify a logical
Outer
Limit of the weight criterion as opposed to the Inner limit you have
identified.
Cheers
Andrew
Back to the Top
The following message was posted to: PharmPK
Hi Vuong,
No it's not quite like that...In bridging studies the aim is to find a
comparable population in the two races so that when they are given the
same
formulation in very closely controlled conditions it is shown that they
have
similar Pk profiles. In that event the Japanese regulatory authorities
assume that it is safe to accept the safety data from all the trials
conducted in Caucasian populations. In other words, those trials do not
have to be repeated in Japan....with huge cost savings to the
pharmaceutical
company.
I would agree that it would be more logical to find a group of slow
metabolisers and compare them with the same race of normal
metabolisers, but
they are difficult to find so we would be back in the same recruitment
bind
as using too narrow weight ranges. I'm no expert on the definition of a
slow
metaboliser but as I expect you know, it is basically a genetic absence
of
certain metabolising enzymes. A good example is the inability of many
Japanese to eliminate alcohol at normal rates. In practice it does not
affect many drugs because most have multiple pathways of elimination..
I'm sure that others in the forum will give you more a detailed
definition.
On balance I don't the absence of a study in slow metabolisers
represents a
safety risk because they know who they are..from family history and
experience with alcohol mostly, so they know to take reduced doses.
Having
said that, perhaps there should be more specific warnings with drug that
have small therapeutic ratios
Best regards
Andrew
Back to the Top
The following message was posted to: PharmPK
Andrew,
Andrew Sutton wrote:
> Moreover, Nick Holford for example, has shown that body weight is a
> minor
> consideration compared with Body Mass Index ( height /weight squared)
> and that is because the major determinant of AUC is clearance.
Please note that the remarks I made about weight contributing only a
small part to the overall variability of AUC can be expected to apply
also to BMI. I did *not* say that "body weight is a minor consideration
compared with Body Mass Index"!
Allometric theory and experiment can predict the extent that clearance
will vary based on an "allometric normal weight". This ANW can be
thought of as the weight of a person with normal body composition.
Deviations of total body weight from ANW will contribute some error in
allometric predictions of clearance when body composition is not normal.
> In his Asian
> vs Polynesian example if you relied only on body weight you could
> compare a
> tall muscular Japanese with a much shorter but rotund South Sea
> Islander of the same overall weight but he would have a smaller
> clearance rate
> and a tendency with lipophilic compounds in particular to have a very
> different distribution pattern.
If you are considering people with similar weight but grossly different
body composition it is possible that using height as well as weight
might reduce the between subject variability in predicted clearance.
BMI is formula using weight and height aimed at trying to describe
obesity. Like IBW it is an empricial formula that has not been
developed (nor tested) as a predictor of ANW. Other measures of body
composition which reflect fat rather than mass e.g. height and sex,
skin fold thickness, bioelectric impedance, total body potassium, whole
body MRI scans, can be used to predict ANW. Of these methods perhaps
the commonest is the use of height and sex to predict ideal body
weight. IBW can then be used as an estimate of ANW. Note that IBW was
not developed for this purpose. I believe it was developed from
actuarial data used to predict life expectancy.
Bruce Green at the University of Queensland is closely involved in
experimental studies of these issues. Perhaps he can offer some
insights on methods for predicting ANW in obesity.
My enthusiasm for encouraging the use of allometry to predict clearance
is because of the solid theory and evidence that the allometric model
is a fundamental biological law that can be relied upon for
interpolation and extrapolation. This is in comparison with emprical
formulae such as BMI and IBW which are based on data without theory.
Allometry alone cannot predict all diffences between people (or
species) but it is a solid starting place. Other independent factors
such as body composition and differences in metabolic pathways (e.g.
CYP polymorphisms, between species metabolism differences) need to be
considered in addition to allometry. Trying to make a single formula
work based on one or two empirical factors is doomed to failure except
in individuals who are close to being average with respect to other
influential factors.
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/
Back to the Top
The following message was posted to: PharmPK
Andrew,
Andrew Sutton wrote:
....
> certain metabolising enzymes. A good example is the inability of many
> Japanese to eliminate alcohol at normal rates. In practice it does not
> affect many drugs because most have multiple pathways of elimination..
....
The explanation for differences in ethanol metabolism is more complex.
It is reasonably well established that there are polymorphisms for
enzymes involved in ethanol metabolism which lead to *faster* rates of
ethanol metabolism to acetaldehyde and (independently) *slower*
elimination of acetaldehyde. Acetaldehyde then causes adverse effects
such as facial flushing. See this link for a simple explanation:
http://chemcases.com/alcohol/alc-06.htm
and a more complex discussion:
http://www.biol.sc.edu/~elygen/Jenny%20Lake.htm
It is commonly asserted that ethanol elimination is slower in Asians
but as far as I know there has never been any study to demonstrate that
either the mixed order (Vmax, Km) or first order (CL) pathways for
ethanol are really any different so that the rate of blood ethanol
disappearance is much the same in all groups -- once weight and body
composition have been accounted for. If anyone has any hard data on
ethanol PK showing difference between races please let me know.
Nick
--
Nick Holford, Dept Pharmacology & Clinical Pharmacology
University of Auckland, 85 Park Rd, Private Bag 92019, Auckland, New
Zealand
email:n.holford.-a-.auckland.ac.nz tel:+64(9)373-7599x86730 fax:373-7556
http://www.health.auckland.ac.nz/pharmacology/staff/nholford/
Back to the Top
The following message was posted to: PharmPK
Andrew,
> Moreover, Nick Holford for example, has shown that body weight is a
> minor consideration compared with Body Mass Index ( height /weight
> squared)
> and that is because the major determinant of AUC is clearance.
This is not (at least from my reading) what Nick was getting at - or
indeed what has been found in empirical studies. BMI (WT/HT[m]
squared) is not a good descriptor of CL, although it does seem to be a
good descriptor of disease related morbidity in a population setting.
BMI does not account for the patients sex nor body composition and
therefore is not likely to be a good descriptor of CL.
Since there is some increase in lean mass with adipose mass it seems
likely that CL will increase slightly in an obese subject. However
this increase will generally not be in proportion with their total
weight, and hence CL per WT may appear to be lower.
Regards
Steve
Stephen Duffull
University of Queensland
Brisbane 4072
Australia
Back to the Top
Nick Holford
You wrote:
>> ”...Trying to make a single formula work based on one or two
>> empirical factors is doomed to failure except in individuals who are
>> close to being average with respect to other influential factors.
Nick...”>>
Nick:
I agree completely with your statement and just wonder if the usage
of a Physiologically Based Pharmacokinetic Model (PBPK) to interpret
the results obtained from mixed population would not solve the problem
of extrapolation between these two distinct subpopulations (and,
perhaps, even between individuals).
For decently characterized drug it does not take much more to
develop an adequate PBPK model. From its sensitivity analysis, it is
possible to determine these few critical parameters that affect mostly
the output. Then, one can stratify the data along the suspected
parameters (e.g., fraction of body fat, rate of metabolism, etc.), and
fit the model to the data, using those few actually measured parameters
(e.g., body weight) and predicting the unknown parameters, and finally,
optimizing the model separately for each subpopulation.
Most probably, as the result, one will get two sets of optimized
parameters that may be tested on selected individuals, set aside from
those distinct subpopulations for validation.
This methodology has been successfully applied in toxicology and
works well even for interspecies extrapolations.
Best wishes.
Janusz Z. Byczkowski, Ph.D.,D.Sc.,D.A.B.T.
Consultant
212 N. Central Ave.
Fairborn, OH 45324
voice (937)878-5531
secure fax (702)446-9127
e-mail januszb.-a-.AOL.com
homepage: http://members.aol.com/JanuszB/index.html
JZB Consulting web site: http://members.aol.com/JanuszB/consult.htm
PharmPK Discussion List Archive Index page
Copyright 1995-2010 David W. A. Bourne (david@boomer.org)