Back to the Top
Thanks for the site. I have spent many hours reading the problems! I
now have a problem of my own.
How do we now look at the scaling of dose to first time in man for a
compound (not anti-cancer) that has a potential for serious side
effects. Is there a regulatory guideline for high risk toxic
compounds for extrapolating from animal PK data (two species in vitro
and in vivo plus ADME and protein binding) to Phase I? What are the
physiological variables that are presently regarded as giving a safe
allometric relationship for a high risk toxic compound? I would
normally use body weight and/or body surface area. I have heard
recently about brain weight (BW) and maximum life time potential (MLP)
but I guess getting the human measurements would be tricky.
Any thoughts and tips would be appreciated.
Back to the Top
If it is a highly toxic substance such as a traditional cancer
then you do not give it to volunteers but to patients, such as cancer
patients who have failed all other treatments. Then you enrol only ones
twos in the lowest dose because there is little point in continuing to
usual tens and twelves if two or three have already shown no adverse
The simplest method that I have found for calculating the 1st dose to
to start with the minimal effective concentration at the receptor or
if it is available and to multiply that by the compartment volume that
think is applicable. That gives you what I call the projected minimal
effective total dose (PMETD). If I do not know what compartment to use I
just use a total blood volume of 8 litres, knowing that this is a
conservative estimate since the drug is so likely to be distributed more
widely than that. Then you take a proportion of this PMETD for added
which depends upon the mechanism of action of the drug. Something as
as you mention might merit 25% while a non-pharmacologically active
or antibiotic might merit 100%.
As you should have a lower than effective dose in your first group you
need a classic ascending design. It's a great help if there is a
drug effect or target Cmax, which should be obvious from the
so as to detect the minimal effective dose in man. You can continue to
explore higher, more clinically useful doses provided that there are no
toxic effects of course. This can be done by doublings, although you
probably have only two or three before you reduced the size of
tradition in volunteers we are permitted 3 doses per subject which
can use within subject comparisons. I personally prefer them because
are more sensitive as detectors of dose response characteristics than
parallel designs...but that's a whole other debate!
I hope this is helpful.
Back to the Top
For a compound with serious toxicity it is not only a question of
have not yet come across any guidance in this respect. The only
know of to cover the step from non-clinical to the clinic is the ICH
of non- clinical studies guideline. This guideline is, however, not
with respect to PK requirements. In general, in this phase of
comes down to common sense and experience.
It will be important to predict the exposure levels at which these side
effects may occur in humans, especially in relation with the expected
effective levels. I tend to keep it simple in this phase, so my
recommendation is "use what you always use". However, in the assessment
the safe human use in Phase I, other factors are as important.
the case of serious toxicities the character of the toxicity will have
most impact on the way forward, like starting dose, dose escalation
and specific safety monitoring requirements. The relevant questions are
1 Is this serious toxicity likely to occur in humans? quite a lot of
toxicities are animal specific
2 Are there known differences in physiology between the animal species
the human with repsect to the tarhet organ?
3 Is this toxicity reversible in the non-clinical studies?
4 What does the tissue distribution study tell about the elimination
the target organ for the toxicity and for the persistence of the drug
its metabolites in the (rat) body?
5 Are the animal model used representative for the human with respect
metabolic profile? Since the only data you may have are from human
microsomes or hepatocytes, this is a question that can usually be
after phase I, but sometimes you already now early on that the value of
animal species will be very limited.
6 Is there a way to detect unexpected changes in exposure (AUC, Cmax)
early stage so that dosing can be stopped in time ?
7 Is there a way to detect the toxicity in an early stage that dosing
stopped before serious damage will be done to the person taking the
Hope this helps,
(by the way I am a toxicologist)
Back to the Top
Do you really want to develop a compound that has the potential for
severe toxicity in man?
If you really want, one suggestion might to start with the NOAEL (no
observed adverse event level) in the most sensitive animal species and
gender (if for instance the NOAEL the female rat is lower that that in
male rat use the former) then apply what is called an "uncertainty
factor" which is normally 60 to 100. Thus, the starting dose will be
1/60th or 1/100 of the NOAEL in the most sensitive animal
species/gender expressed as per kg of body weight. Other apply an
additional safety factor of 10 to select the staring dose. Doubling
dose increments are then usually applied.
Another more time consuming approach, is to use the physiologically
based pharmacodynamic/pharmacokinetic model (some of the big CRO
provide this as a service). The model integrates the in vitro
metabolism, in vivo preclinical pharmacokinetics, protein binding,
pharmacodynamics data in several animals species and man (in vitro
only) to extrapolate the rate and extent of exposure of the volunteers
in the FIM study. It takes into accounts for instance, intrinsic
clearance, organ blood flow, fraction unbound etc. etc. to extrapolate
the plasma concentration vs time curve in man. If you then have
toxicodynamics data you can then make a reasonable prediction. For this
you need a considerable amount of data and accurate prediction is not
guaranteed. Geoffrey Tucker at the University of Sheffield has
developed a software that it is supposed to do the job (sometimes does
the job sometimes doesn't).
Finally, if the compound is very toxic, you might think about
performing the FIM study using a slow intravenous infusion at very low
Even if this might appear crazy at first, it provides several
advantages with respect to the oral formulation.
First of all you can stop the infusion if serious AE's occurs. On the
contrary, once you have administered the oral formulation and an AE
occurs rapidly (before all the dose is actually absorbed) nothing can
be done to stop the absorption of the remaining portion of the dose.
Second it provides an immediate and direct indication of the exposure
and eliminate the risk of flip-flop kinetics which might prolong the
exposure (if biliary excretion and hepatic cycling can be reasonably
Third, the dosing can be accurately controlled down to the mg level
avoiding the risk of solubility/precipitation high local concentrations
Most of the Investigators with training in intensive care, would agree
with this approach.
I hope this help.
Stefano Persiani, PhD
Director, Drug Metabolism and Pharmacokinetics
Rotta Research Laboratorium SpA
Via Valosa di Sopra, 7-9
20052 Monza (MI)
PharmPK Discussion List Archive Index page
Copyright 1995-2010 David W. A. Bourne (firstname.lastname@example.org)