Back to the Top
Dear Group,
Should oral bioavailabilty be the limitation factor in
developing NCE for oral administration if it has
promising in-vivo activity? Normally it has been said
that NCE under development should possess promising
oral bioavailability profile in rodents and
non-rodents (for example if it is an analog then it
should exhibit 30%+ in rodents). But there were
exemptions in bioavailability profile of orally
administered drugs that are in market. Alendronate is
an example.
Alendronate, a biphosphonate derivative, used in the
treatment of osteoporosis was marketed as orally
administered drug despite of its poor oral
bioavailability in humans (<1%) because the bio phase
or receptor site in bone effectively takes up
Alendronate that is available in systemic circulation
and has half-life of 10 years.
I would highly appreciate if anyone from PHARMPK GROUP
explain me the criteria with respect to oral
bioavailability of NCEs in order to develop further as
oral drug. Can we compromise the bioavailability if it
has excellent in-vivo PD and half-life? Thanks in
advance.
Regards,
S Syed Mustafa,
Research Associate,
Drug Metabolism & Pharmacokinetics,
Discovery Research,
Dr Reddy's Laboratories Ltd,
Bollaram Road, Miyapur,
Hyderabad- 500049 INDIA
mustafas.aaa.drreddys.com
Back to the Top
Dear Syed Mustafa,
I suspect alendronate may represent an exception: current paradigms in
drug development don't accept usually candidate drugs with low
bioavailability!
Alendronate is poorly absorbed and bioavailable, however, when
alendronate was marketed, no other competitive therapies were available
in the some therapeutic field (calcitonin was a placebo-like
substance...). Additionally alendronate activity against bone mass
density decrease in osteoporosis was unexpected when alendronate was
first tested (I remember as probably it was developed during a
phosphate derivatives synthesis program for development of new washing
powders additives...)
It was really because no other therapies were (and probably are)
present that alendronate became a successful drug...not for ADME
properties!
Best regards
Dr. Stefano Porzio
Pharmacokinetic and Tox. Dept.
Inpharzam Ricerche SA - ZAMBON-GROUP
Taverne - Switzerland
Back to the Top
Hello Mustafa,
oral bioavailability is just one in the list of different parameters to
be considered when you design drug candidates for oral dosing. But this
is only ONE factor.
One should not drop a compound because of an oral bioavailability in
rodents less than 30% for example, especially if the drug is highly
efficient and non-toxic (ideal drug, isn't it ?). You sould also
consider the pharmacological activity and also as you write, the in
vivo PD and half-life.
Secondly you should also explore the reasons why your active is not
"bioavailable". Is it really a problem of F or a problem of absorption
? Formulation is a way to modify absorption parameters so that if your
compound issues are in the absorption process you will be able to
correct or modify this at least.
Hope this helps,
Frederic DOC
CEO and co-founder of ACRITER, drug discovery consulting
Back to the Top
Dear Mustafa,
First of all for NCE to be developed, we have to see the category of
drug and the same class of drugs. Then we have to compare the same class
of drugs in that category w.r.t to its pros and cons. Then check the
critical parameter which is failing in the clinic viz. activity, PK,
toxicity. Then as per the requirement in the clinic we have to
prioritize the compounds to be developed.
Obviously we know pharmacokinetics is the main area to be focused for
the drug development process in NCE research. I do agree that we will
see first bioavailability as main parameter of interest. But this is one
of the factor only. So other things include clearance, half-life and Vd.
As you rightly said, we can compromise on bioavailability if we have
potent in vivo activity and long half-life for the NCE.
But there are some more factors to be considered in this regard. First
of all it is better to check where the bioavailability is failing in
case of in vivo active molecules.
1. Check its physicochemical properties which can help in improving
absorption by making different salts of this compound if possible.
2. check different formulations by adding bioenhancers or different
suspending vehicles which may improve the compounds absorption further.
3. check the dose linearity and proportionality by giving different
doses. I feel most of the compounds may fail in this if we have very low
bioavailability. In these cases your compound may be potent but not
efficacious.
3. check its first pass-metabolism such that a chemist can try to
block the metabolic sites in his structure without loosing the activity.
4. check the possibility of alterations in bioavailability in other
species which you may try to predict by in vitro metabolic stability
studies in different species liver microsomes.
4. depending on the category and necessity of the compound one can
try to develop the NCE by changing the route of administration as last
step unless it has toxicity.
here I have exemplified only some of the methods of developing NCE if it
has low bioavailability and I appreciate if our group gives more light
on this topic.
Regards,
KANTHI KIRAN V.S. VARANASI, M.Pharm, MBA
Sr. Research Scientist,
Pharmacokinetics and Drug Metabolism Department,
Glenmark Pharmaceuticals LTD.
Glenmark Research Centre,
Plot No. A-607, T.T.C Industrial Area,
MIDC, Mahape,
NAVI MUMBAI-400709
INDIA.
Ph.: 91-22-55902491/92 ext 315
Fax: 91-22-55902318
Back to the Top
Hi,
I agree with other comments that the degree of bioavailability (F)
should not be the main basis for go or no go decision of a compound.
Many drugs can be potent enough that even if less than 5-10% of the dose
is available to the system, it is enough for their effect.
The major issue I see with a compound with low F is the inter-individual
and possibly even intra-individual variability of drug (plasma)
concentrations. Someone with an F of 7% has more than twice as much drug
in the body than another one with an F equal to 3%. If the candidate
compound has a narrow therapeutic window, this could effectively kill
the development work. On the other hand, if you are not worried about
the lack of efficacy or presence of side effects due to the variability,
i.e. you have the desired effect and no toxicity at those
concentrations, this could be a very good drug candidate.
Toufigh Gordi
Back to the Top
Dear Group,
I thank everyone who responded me. It has been a
common practice to rank order the NCEs based on the
systemic exposure or bioavailability during lead
selection process rather paying attention onto PK/PD
index of NCEs, which was the basis to bring about such
discussion in PHARMPK Group.
Also it would be appropriate to consider the use of
suitable animal models for PK (For example tumor
bearing mice for evaluating Pharmacokinetics of
oncology drugs),immune state and infection site
differences among various animal models and
differences in half-life when it is being extrapolated
to humans (In general half-life is 6-9 folds longer in
humans than rodents). I would highly appreciate
further comments on this issue.
Regards,
S Syed Mustafa,
Research Associate,
Drug Metabolism & Pharmacokinetics,
Discovery Research,
Dr Reddy's Laboratories Ltd,
Bollaram Road, Miyapur,
Hyderabad- 500049 INDIA
mustafas.aaa.drreddys.com
Back to the Top
Hi Mustafa
Lead selection is always based on three factors, potency or you may
call it as PD effect (foremost), good PK (to reduce the interindividual
varation and ease in further development) and toxicity (with minimal
side effects and toxicity is desired, again based on benfit to risk
ratio). I don't think lead selection (or rank ordering) is based on
bioavailability alone.
For compounds of same chemical class and having similar PD profile, you
can rank order based on bioavailability.
Coming to your second question, yeah, its always better to use the
appropriate PD model (tumor bearing mice for eg). but I guess for the
routine compounds, becuase of the cost involved you may use the animal
species that has better PK corrleation with the PD model. Its not
always necessary that you determine the PK of each and every molecule
in the PD model. Based on PK, you can shortlist candidates for further
development and final decision will be based on PK in appropriate
models such as disease induced or knockout animals.
Because of the tremendous cost involved in new drug discovery, these
practises are well appreciated in industry but final decision as
mentioned above will be a balace of potency, PK, toxicity and also CMC
issues.
Regards
Shruti
Back to the Top
Shruti Agrawal said:
"Lead selection is always based on three factors, potency or you may
call it as PD effect (foremost), good PK (to reduce the interindividual
varation and ease in further development) and toxicity (with minimal
side effects and toxicity is desired, again based on benfit to risk
ratio)."
I have to disagree just a little with your statement.
I think it is a mistake to say "foremost" for any single property.
Potency has traditionally been the ruling factor, but I believe that
over the years, that philosophy has resulted in enormous waste and
numerous clinical failures. As your further discussion states,
bioavailability and minimal side effects are also important. I think
they are equally important.
For example, which is better - a compound with an IC50 of 1 nmol and
bioavailability of 30%, or one with an IC50 of 2 nmol and
bioavailability of 90% (side effects being equal)? The latter is only
half as potent, yet the delivered effect is greater at the same dose. Or
suppose the bioavailabilities were equal, but the half-lives were 2
hours for the more potent molecule and 10 hours for the less potent
molecule? Or if the volume of distribution for the more potent molecule
was 10 L/kg but only 2 L/kg for the less potent?
Clearly, we should not choose the lowest IC50, nor the longest
half-life, nor the smallest volume of distribution, nor the highest
solubility, etc. We should choose the optimum balance (compromise) among
these factors.
In the aerospace industry in the 1970's, I jokingly put forth "Woltosz'
First Rule of Optimization", which states, "If any single property of a
system is optimum, then the entire system is not." I believe it applies
equally in the pharmaceutical industry.
An optimum design for a complex system (like a dosage form) is virtually
always a compromise among competing factors. In the case of drugs, those
factors include potency, toxicity, pharmacokinetics (CL, Vd),
deliverability (bioavailability), manufacturability, cost, and even
marketability (with respect to competition).
We need to take a systems approach to optimizing drug (and dosage form)
design, as other industries have learned with respect to their systems.
This means conducting projects in a way to gather the data needed to
assess the tradeoffs among the competing factors, and having the tools
that can assimilate these data and assess the interactions.
There is no other way to do this outside of good simulation and
modeling. Enormous amounts of data exist in spreadsheets, databases,
reports, etc. representing huge investments on the part of
pharmaceutical companies. Without the right tools, human minds can look
at such data for years and never appreciate the interactive effects that
exist.
Good simulation and modeling tools help you see what your data is trying
to tell you.
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
Dear Mustafa,
I do agree that a compound PK in diseased animal model will give you a
better idea of how it is behaving in pathological conditions and you can
try to correlate PK/PD.
But that is the second factor in general not the first and most factor
which you have to always check how a compound is handled by the normal
physiological system. It will give true PK parameters without the
influence of other factors. you knew even in clinical trials, we start
with PK in healthy volunteers rather than others. This may be different
in case of anti-cancer drugs because its side effects/toxicity in normal
volunteers.
But in general we have to see the common design of conducting studies to
get a proper meaningful data by doing in normal mice/rat/dog/monkey
models.
As our colleagues said it is better to get collect more data by doing
sequential design of studies for better decision making.
Regards,
KANTHI KIRAN V.S. VARANASI, M.Pharm, MBA
Sr. Research Scientist,
Drug Metabolism and Pharmacokinetics Department,
Glenmark Pharmaceuticals LTD.
Glenmark Research Centre,
Plot No. A-607, T.T.C Industrial Area,
MIDC, Mahape,
NAVI MUMBAI-400709
INDIA.
Ph.: 91-22-55902491/92 ext 315
Fax: 91-22-55902318
Back to the Top
Walt & Shruti,
You seem to be making a fundamental mistake in confusing potency (EC50,
sensitivity) with efficacy (Emax, intrinsic activity). Potency is more
or less irrelevant except for formulation issues in being able to
deliver a large mass of drug. It is efficacy that determines if a drug
how effective a drug will be.
As a corollary to "Woltosz' Rule" I suggest Holford's Rule -- there are
only 4 parameters of interest - CL, Vd, Emax and EC50. Of these CL and
Emax are the most important to understand. All the other stuff is fine
print.
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
Nick,
Thank you for your post. There is no confusion - I agree completely that
Emax and EC50 (and gamma if it is a sigmoid model) are the more relevant
properties - once they are known.
My point was that in preclinical work, before human (or even animal)
data are available, medicinal chemists are often enamored of IC50, at
the expense of almost everything else. When screening a library of
molecules for eventual use in human, we should be looking at a balance
among many parameters (many more than just CL, Vd, Emax, and EC50!).
Our technology today does not allow predicting everything perfectly, but
whatever can be reasonably predicted should be used at this stage to
select compounds to take forward (and eventually measure Cl, Vd, EC50,
Emax, and gamma).
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 Nick,
You wrote:
> As a corollary to "Woltosz' Rule" I suggest Holford's Rule --
> there are only 4 parameters of interest - CL, Vd, Emax and EC50.
> Of these CL and Emax are the most important to understand.
> All the other stuff is fine print.
It is nice to realize that 'the essentials of PK-PD' can be reduced to
a few
parameters! But I suggest to add F to the list. Several contributors to
this
thread have given arguments that F is not the most important parameter
to
decide whether or not a NCE is suitable for further development. One
might
say that a low F is simply a matter of increasing the dose. But in my
opinion, an 'essential' element of PK-PD is to get some insight in the
relationship between dose and effect. Others have argued that a low F
may be
associated with a high variability in F, which may cause problems in
clinical practice. Also, the question 'what happens with the fraction
'1-F'?', and lack of efficiency may be serious. Besides, I would not
recommend to replace CL and Vd by their 'apparent' (a terribly vague
term!)
counterparts CL/F and Vd/F, and thus taking into account F implicitly.
This
approach does not add insight.
Do you have convincing arguments to 'ignore' F in Holford's Rule?
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
Back to the Top
Hans and Nick,
Sorry, but I still cannot subscribe to such a simplistic approach. There
are too many other things to consider. We cannot select molecules based
only on CL, Vd, Emax, and EC50 (with or without F), or we'll ignore
factors that cause drugs to fail: toxicity, manufacturability,
marketability (competitive therapies), patient compliance (e.g., b.i.d.
or q.i.d), etc.
As I noted earlier, I am coming from a perspective of mostly preclinical
project decisions regarding which molecule(s) in a series to keep, and
which to kill early. I believe that was the context of the original post
for this thread.
It is clearly a systems optimization kind of problem, and it calls for
systems optimization thinking and tools. Otherwise, we will continue to
spend the vast majority of the industry's resources on failures.
I stand on Woltosz' First Rule of Optimization: "If any single property
of a system is optimum, then the entire system is not."
I should add a Second Rule of Optimization that is more general: "If any
subset of properties of a system are optimized, then the entire system
is not." In other words, if we adjust only some subset of parameters to
achieve some minimal cost function, but ignore other parameters, then we
again obtain a nonoptimum system. Of course, in real life, we almost
always deal with subsets, so we almost never achieve true optima. We
have to do the best we can with the parameters over which we exercise
some control.
Best regards,
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
Back to the Top
Walt Woltosz wrote:
> I stand on Woltosz' First Rule of Optimization: "If any single
> property of a system is optimum, then the entire system is not."
>
> I should add a Second Rule of Optimization that is more general: "If
> any subset of properties of a system are optimized, then the entire
> system is not."
I definitely echo those sentiments. In support of this notion I have
been encouraging our discovery chemists to read the commentary by Dennis
Smith and co-workers (Clinical Pharmacokinetics, 41:1005-1019 [2002])
in which they argue that "adequate", rather than "optimal" parameters
should be the realistic aim. Their focus is upon DMPK properties but
this could be extended to efficacy etc. Understanding what constitutes
"adequate" and the appropriate weightings to be used for "not optimal
but adequate" values needs to be done as early as possible. We need
to suppress the tendency to over-engineer any parameter simply because
it is easy to screen. Disobeying Walt's rules leads to an imbalance
that can only be rescued, if at all, with time and money (e.g time
to test new formulations or strange dosage forms). I agree that we
do not want to make tracts of chemspace forbidden a priori but we
need to stress that there is a financial cost in going out too far.
Knowledge of these guidelines also helps in deciding when to stop
(or pause) the search. Some teams just don't know when to stop.
Thanks to all for an interesting discussion.
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
Hans,
Hans Proost wrote:
>
> Dear Nick,
>
> You wrote:
> > As a corollary to "Woltosz' Rule" I suggest Holford's Rule --
> > there are only 4 parameters of interest - CL, Vd, Emax and EC50.
> > Of these CL and Emax are the most important to understand.
> > All the other stuff is fine print.
>
> It is nice to realize that 'the essentials of PK-PD' can be reduced to
> a few parameters! But I suggest to add F to the list.
....
> Do you have convincing arguments to 'ignore' F in Holford's Rule?
The only argument for knowing F is to know how predict equivalent doses
by different routes e.g. IV and oral. This is almost never an issue for
most drugs being developed today -- therefore I consider F to be a fine
print parameter.
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
Nick Holford wrote:
"The only argument for knowing F is to know how predict equivalent doses
by different routes e.g. IV and oral. This is almost never an issue for
most drugs being developed today -- therefore I consider F to be a fine
print parameter."
Predicting doses by different routes is almost never an issue. Wow!
F as a fine print parameter. Wow!
I must be working in a different industry! ; )
Respectfully, Nick, I know you're a very smart guy, so I think the only
explanation is a matter of perspective.
For a drugs used in clinical practice, perhaps your argument holds most
of the time. But for discovery and development work, these are clearly
issues that cause daily wringing of hands at research sites worldwide.
Especially so for drugs that undergo regionally dependent
absorption/bioavailability due to regionally dependent
properties/phenomena such as solubility(pH), first pass extraction (in
the gut wall), or (saturable) transporter effects. For such drugs, which
we see virtually every month at one customer site or another, F is not a
simple ratio for oral/iv doses, and in fact F can sometimes change
significantly with relatively small changes in formulation.
Once efficacy and safety have been established, you're in a different
game. Before then, the problem of selecting leads (the original question
of this thread) cannot be simplified to CL, Vd, EC50, Emax, gamma, and
F.
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
Nick
> The only argument for knowing F is to know how predict
> equivalent doses by different routes e.g. IV and oral. This
> is almost never an issue for most drugs being developed today
> -- therefore I consider F to be a fine print parameter.
I think that your comment is right from the viewpoint that you know F.
However if you don't know F in an individual but rather have a
population estimate which had some between-subject variability attached
to it then the value of F takes on a more important meaning. Say for
the purposes of simplicity that F = 1 - E. Then the ith patients value
of E (E_i) could be
E_i = E(1+ETA)
ETA ~ N(0, var(ETA))
Then the variance of E_i in the population would be given by
E^2*var(ETA)
This would therefore mean that the variability of F in the population
would be inversely related to its value. So knowing the population
average value of F would allow one to have some feeling for the
variability between individuals.
Hence, while F may be of limited importance its variance is not and if
its variance is of interest then the value of F is of interest too...
Steve
Stephen Duffull
School of Pharmacy
University of Queensland
Brisbane 4072
Australia
Tel +61 7 3365 8808
Fax +61 7 3365 1688
University Provider Number: 00025B
Email: sduffull.at.pharmacy.uq.edu.au
www: http://www.uq.edu.au/pharmacy/sduffull/duffull.htm
PFIM: http://www.uq.edu.au/pharmacy/sduffull/pfim.htm
MCMC PK example: http://www.uq.edu.au/pharmacy/sduffull/MCMC_eg.htm
Back to the Top
Attrition is primarily caused by lack of efficacy and/or toxicity.
Recently, there has been a rash of phase III attritions from failure to
eliminate undesirable candidates early on during the drug discovery
process. The role of PK/PD department is to eliminate as many
undesirable leads as possible to avoid costly mistakes down the line.
However, from experiences at Pfizer, Merck etc., there are simple rules
for NCEs which should be followed.
Lipinsky at Pfizer proposed a series of rules for NCEs. Understably
some of these can be discarded depending on your institution; keep in
mind, however, that there is no need to reinvent Pfizer from scratch
i.e. if one do not learn from history, history will repeat itself.
Poor absorption or permeation are more likely when there are:
More than 5 H-bond donors.
The MWT is over 500.
The CLog P is over 5 (or MLOGP is over 4.15).
The sum of N's and O's is over 10.
Substrates for transporters and natural products are exceptions.
Other rules include:
Rule of 3 for fragment screening
Less than 10 rotatable bonds for bioavailability
PSA < 60-70 A^2 for CNS activity
N+O <= to 5 for CNS activity
Log P - (N+O) is positive for CNS activity
Log D > 0 and < 3 for good permeability
Vuong Trieu PhD
Back to the Top
Steve Duffull wrote:
> Nick
>
> > The only argument for knowing F is to know how predict
> > equivalent doses by different routes e.g. IV and oral. This
> > is almost never an issue for most drugs being developed today
> > -- therefore I consider F to be a fine print parameter.
>
> I think that your comment is right from the viewpoint that you know F.
> However if you don't know F in an individual but rather have a
> population estimate which had some between-subject variability
attached
> to it then the value of F takes on a more important meaning. Say for
> the purposes of simplicity that F = 1 - E. Then the ith patients
value
> of E (E_i) could be
> E_i = E(1+ETA)
> ETA ~ N(0, var(ETA))
> Then the variance of E_i in the population would be given by
> E^2*var(ETA)
> This would therefore mean that the variability of F in the population
> would be inversely related to its value. So knowing the population
> average value of F would allow one to have some feeling for the
> variability between individuals.
>
> Hence, while F may be of limited importance its variance is not and if
> its variance is of interest then the value of F is of interest too...
I don't think the variance of F is likely to follow the model you
propose. When F is close to 1 (let's say > 0.9) then variability is
low. But after that it is very variable and more or less independent of
F as far as I can tell. F can be low because of high extraction (e.g.
morphine) or poor absorption (e.g. bisphosphonates). There are 2 very
different mechanisms contributing to F which means that F is variable
over a wide range.
Whatever model you choose for variance of F it doesn't add anything if
you already know CL/F and V/F (and their variability and covariance) in
clinical drug development.
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
Walt,
Walt Woltosz wrote:
>
> Nick Holford wrote:
>
> "The only argument for knowing F is to know how predict equivalent
doses
>
> by different routes e.g. IV and oral. This is almost never an issue
for
> most drugs being developed today -- therefore I consider F to be a
fine
> print parameter."
>
> Predicting doses by different routes is almost never an issue. Wow!
>
> F as a fine print parameter. Wow!
>
> I must be working in a different industry! ; )
>
> Respectfully, Nick, I know you're a very smart guy, so I think the
only
> explanation is a matter of perspective.
>
I agree. My perspective is on clinical drug development i.e. after you
have helped the discovery folks figue out if the NCE is brickdust.
After that we can forget about F.
If you want to convince me that F is important in clinical drug
development then you have to give me examples of real drugs where
knowledge of F made an important contribution to the label or a
development decision (assuming we have passed the pre-clinical "Is it
brickdust?" test).
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
Nick Holford wrote:
"If you want to convince me that F is important in clinical drug
development then you have to give me examples of real drugs where
knowledge of F made an important contribution to the label or a
development decision (assuming we have passed the pre-clinical "Is it
brickdust?" test)."
It was not I that advocated for F, although I know it is of major
concern to our customers in both discovery and development.
Because so many failures occur in clinical drug development that are not
due to efficacy, but in fact are due to deliverability and toxicity, I
must advocate for a systems approach that goes far beyond "CL, Vd, EC50,
and Emax".
IMHO, as long as the drug has not yet passed Phase III, these four
parameters are not sufficient. (Six parameters if gamma (exponent) and F
are added.)
Of course, I encourage performing the systems optimization at the
earliest possible stage, so that ideally, when we got to Phase II and
III, the 4-6 parameters would suffice. But I don't hold out much hope
that that will happen in my lifetime.
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
Nick,
> I don't think the variance of F is likely to follow the model
> you propose. When F is close to 1 (let's say > 0.9) then
> variability is low.
My proposed and obviously simplistic model was that variance was
associated with E (F=1-E) so as E gets bigger (F gets smaller) the
variance of E and F will increase.
> There
> are 2 very different mechanisms contributing to F which means
> that F is variable over a wide range.
Sure - but what is important is that the value of F is important as the
value of variance will most likely inversely depend on its value.
Steve
Stephen Duffull
School of Pharmacy
University of Queensland
Brisbane 4072
Australia
Tel +61 7 3365 8808
Fax +61 7 3365 1688
University Provider Number: 00025B
Email: sduffull.-at-.pharmacy.uq.edu.au
www: http://www.uq.edu.au/pharmacy/sduffull/duffull.htm
PFIM: http://www.uq.edu.au/pharmacy/sduffull/pfim.htm
MCMC PK example: http://www.uq.edu.au/pharmacy/sduffull/MCMC_eg.htm
Back to the Top
Vuong Trieu wrote:
"Lipinski at Pfizer proposed a series of rules for NCEs. Understandably
some of these can be discarded depending on your institution; keep in
mind, however, that there is no need to reinvent Pfizer from scratch
i.e. if one does not learn from history, history will repeat itself."
Lipinski's work has been of great value to the industry; however, after
a few more years of work (inspired by his original Rule of 5), we
believe there are better rules, as validated by the largest number of
drugs we can find with reported human fraction absorbed (115 compounds
in the Zhao, et al, data set, with transported molecules removed).
Our "J-Alert" score in QMPRPlus(TM) uses an entirely different set of
rules, and is significantly more discriminating (R^2 ~0.8 vs 0.2 for the
Lipinski Rule of 5 on the same data set):
(1) Moriguchi logP < -0.59
(2) S+ Peff (human effective permeability predicted by the Simulations
Plus artificial neural network ensemble model) < 0.864
(3) Sum of partial atomic charges on hydrogen bond acceptor oxygen atoms
< -2.406
(4) Number of quaternary amines > 0 OR number of sulfonium > 0 OR number
of diazo > 0
(NOTE: assign 2 points for rule #4)
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
Steve,
Steve Duffull wrote:
>
> Sure - but what is important is that the value of F is important as
the
> value of variance will most likely inversely depend on its value.
>
My point is that I don't see why we need either F or its variance to
make sensible drug development decisions or to write informative drug
labels (except dose conversions between routes). Can you explain why
this inverse relationship for the variance of F would be important for
drug development given we can describe the variability of CL/F and V/F
without any knowledge of F?
A real example would be helpful.
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
Nick,
I guess one example could be cyclosporine. The original formulation
Sandimmune gave rise to a large variability in exposure (AUC, Cav,ss).
The
work that led to the improved Neoral formulation was probably made much
simpler to motivate given the knowledge that bioavailability was low
(and
variable). Without knowledge about F and its variability, it would have
been
guesswork whether a new formulation could lower variability in exposure.
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
Dear Nick,
You wrote:
> The only argument for knowing F is to know how predict equivalent
doses
> by different routes e.g. IV and oral. This is almost never an issue
for
> most drugs being developed today -- therefore I consider F to be a
fine
> print parameter.
I agree, if you want to reduce the list from the point of view of NCE
selection for development. But from that point of view, I do not see the
argument why CL is of primary interest. Why is CL not a fine print
parameter, like F, V and EC50? This would leave only Emax as the primary
parameter, and then we are back in the good old days when PK-PD was
unknown
....
Besides, from your reply to Steve Duffull I understand that you
advocate the
use of CL/F as a key parameter? Such a mixed parameter does not seem
helpful
to me.
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
Back to the Top
On 3/11/04 12:26 am, "Nick Holford"wrote:
"A real example would be helpful"
Perhaps domperidone is a case in point. It has an average
bioavailability of
about 15% which means the range can easily be 5 or 6 fold...as from 5%
to
30% for example. I think this is an adequate explanation of the drug's
great
weakness from a clinician's point of view, ie: is its very unpredictable
efficacy in the face of potentially unpleasant adverse events such as
extrapyramidal phenomena. It is quite easy to ask a patient to double or
triple a dose but not so easy when the factor is 6 or more. In practice
the
delays in reaching an adequate dose for a rapid metaboliser also
undermine
confidence, especially when it's major market is dypepsia, the symptoms
of
which are inherently variable in the first place.
Andrew Sutton,
Guildford Clinical Pharmacology Ltd.
The Technology Centre, Occam Road
Guildford, Surrey, UK. GU2 7YG
Tel: +44 (0)1483 455375. Direct: 688303
Mobile +44 (0) 7770 820 145 (To 5pm EST)
URL: www.gcpl.co.uk
Back to the Top
Dear Colleagues:
There is one thread of comments in the current discussion on which PK
parameters are important and which are not, embodied perhaps most
clearly in something that Nick Holford wrote:
> ..... to make sensible drug development decisions or to write
> informative drug labels ....
>
Does this group believe that PK parameters are only or overwhelmingly
important for drug development? What about the optimization of a given
pharmacotherapy on an individualized basis? I did not see in all that
discuss any mention of which parameters may - or may not - be important
in determining how to treat patient X or patient Y at any particular
stage of his/her disease.
Some of you may have heard me argue that while blood-based measurements
may be useful for many drugs, in cancer and in a number of other
diseases the key determinants of response or toxicity are PK parameters
measurable only at the target site(s), and that is why we have been
developing noninvasive imaging methods to allow the measurement of such
parameters using noninvasive (imaging) methods, including nuclear
imaging and MRI/MRS.
--
Professor Walter Wolf, Ph.D.
President, Correlative Imaging Council, Society of Nuclear Medicine
Distinguished Professor of Pharmaceutical Sciences
Director, Pharmacokinetic Imaging Program
Department of Pharmaceutical Sciences, School of Pharmacy
University of Southern California
1985 Zonal Ave., Los Angeles, CA 90089-9121
E-Mail: wwolfw.-at-.usc.edu
Telephone: 323-442-1405
Fax: 323-442-9804
Back to the Top
Walter,
Thanks for emphasizing the need for dose optimization in individuals.
This thread was motivated by a discussion of what parameters are most
important during drug development. IMHO the label at the end of drug
development should contain information to help the prescriber and the
patient to optimize the dose. The need to obtain information in the
label to aid optimization is why I wrote:
> > ..... to make sensible drug development decisions or to write
> > informative drug labels ....
If drug development was target concentration oriented (Holford 1995)
then it would require only CL/F, V/F, Emax and EC50 as the key
parameters for development and a label that gave help for optimization.
I support this assertion with the following arguments:
The goal of therapy is to obtain a target effect (TE). Choice of TE
includes optimization of beneficial vs adverse effects. The target conc
can be predicted:
TC=TE*EC50/(Emax-TE)
Then the loading dose can be predicted:
LD=TC*V/F
and the maintenance dose rate:
MDR=TC*CL/F
Note that F is not needed explicitly. For most drugs F is unknown and
it only CL/F and V/F that are obtained.
The PD and PK models may need to be made more complex in individual
cases e.g. adding Hill coefficient to the PD model but all drug
development programs should attempt to identify at a minimum these 4
parameters.
The most important of the 4 are:
Emax because if you dont know this you will come on market with too big
a dose then lose $$ when it is found a lower dose is equieffective with
fewer AEs.
CL/F because this affects average exposure which in most cases is best
related to beneficial and adverse effects and predicts the recommended
average dose rate.
Note that interesting stuff like models for delayed effects are not
commonly needed for the label because the steady state effects are
mainly independent of any such delays.
The target concentration strategy aims to identify individual
differences in the key parameters, Emax, EC50, CL/F and V/F and
individualize the dose based on a change in Target Conc (maybe) and an
individual estimate of CL/F to predict the maintenance dose.
Holford NHG. The target concentration approach to clinical drug
development. Clinical Pharmacokinetics 1995;29(5):287-91.
--
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
Mats,
Mats Karlsson wrote:
>
> PharmPK - Discussions about Pharmacokinetics
> Pharmacodynamics and related topics
>
> Nick,
>
> I guess one example could be cyclosporine. The original formulation
> Sandimmune gave rise to a large variability in exposure (AUC, Cav,ss).
> The
> work that led to the improved Neoral formulation was probably made
much
> simpler to motivate given the knowledge that bioavailability was low
> (and
> variable). Without knowledge about F and its variability, it would
have
> been
> guesswork whether a new formulation could lower variability in
exposure.
Thanks for this example. But is a rather special case from Phase IV of
clinical drug development and I suspect much of the formulation
development was motivated by patent extension issues rather than
reducing variability of exposure.
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
Hans,
Thanks for your comments.
Hans Proost wrote:
> You wrote:
> > The only argument for knowing F is to know how predict equivalent
> doses
> > by different routes e.g. IV and oral. This is almost never an issue
> for
> > most drugs being developed today -- therefore I consider F to be a
> fine
> > print parameter.
>
> I agree, if you want to reduce the list from the point of view of NCE
> selection for development. But from that point of view, I do not see
the
> argument why CL is of primary interest. Why is CL not a fine print
> parameter, like F, V and EC50? This would leave only Emax as the
primary
> parameter, and then we are back in the good old days when PK-PD was
> unknown
> ...
I've tried to explain the need for CL/F in another reponse to this
thread.
> Besides, from your reply to Steve Duffull I understand that you
> advocate the
> use of CL/F as a key parameter? Such a mixed parameter does not seem
> helpful
> to me.
I am afraid we will have to agree to differ on the value of CL/F.
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
Andrew,
Andrew Sutton wrote:
>
> PharmPK - Discussions about Pharmacokinetics
> Pharmacodynamics and related topics
>
> On 3/11/04 12:26 am, "Nick Holford"
wrote:
> "A real example would be helpful"
>
> Perhaps domperidone is a case in point. It has an average
> bioavailability of
> about 15% which means the range can easily be 5 or 6 fold...as from 5%
> to
> 30% for example. I think this is an adequate explanation of the drug's
> great
> weakness from a clinician's point of view,
The explanation of variability could equally well be in terms of CL/F.
There is no need to know F.
> In practice the delays in reaching an adequate dose for a rapid
metaboliser also
> undermine confidence,
So it does seem there is a need to be able to identify rapid and slow
metabolizers. This is CL (or CL/F) not F.
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
Dear Nick,
Thank you for your response. After all, we do not differ that much. The
need
for CL/F is clear. But knowledge on CL/F does not seem helpful if EC50
(or
TC) is not known. So I would say: both or none.
Also with respect to the value of CL/F we do not really differ, I
think. I
agree that CL/F is quite useful to correlated dose to concentration.
But I
would prefer to know both CL and F. I agree that CL/F is definitely more
important than either CL or F.
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.at.farm.rug.nl
Back to the Top
Dear PharmPKers,
Interesting thread to follow. In his last reply to Walter, Nick wrote:
"Thanks for emphasizing the need for dose optimization in individuals.
[....]
If drug development was target concentration oriented (Holford 1995)
then it would require only CL/F, V/F, Emax and EC50 as the key
parameters for development and a label that gave help for optimization.
[....]
The most important of the 4 are:
Emax because if you dont know this you will come on market with too big
a dose then lose $$ when it is found a lower dose is equieffective with
fewer AEs.
CL/F because this affects average exposure which in most cases is best
related to beneficial and adverse effects and predicts the recommended
average dose rate."
While from Walter's original post:
"[...] I did not see in all that discuss any mention of which parameters
may - or may not - be important in determining how to treat patient X
or patient Y at any particular stage of his/her disease."
IMHO individual optmisation is only needed when the therapeutic margin
is low, i.e. the target concentration is close to concentrations evoking
serious adverse effects.
So for individual optimisation, we need to describe the peak
concentration,
and that can only be done if we have a description of relevant kinetics.
Which kinetics are relevant will depend on the drug itself. One needs
to
know how fast the absorption process is - eg k(a) - and whether a clear
distribution component is involved, so mono vs multicompartment
kinetics.
And of course a description of the peak requires the volume(s) of
distribution. If the clearance of a drug is (s)low in comparison with
the
dose and dosing interval one of course does not worry about peak values.
Furthermore one of course is not as much interested in parameters of the
average person as in those of an individual patient. So ideally one
would
describe factors that influence PK and PD parameters of a particular
drug,
for example age (V+Cl), body composition or BMI (V), disease state and
comedication (both the whole series Cl, V, EC50, Emax - ka?) - of course
in addition to normalisation to body weight or another allometric.
Best regards,
Jeroen
PharmPK Discussion List Archive Index page
Copyright 1995-2010 David W. A. Bourne (david@boomer.org)