Back to the Top
[I may have sent this out earlier - sorry if it is a duplicate - db]
Colleagues:
"A failed experiment is never a complete loss, it can always serve as a bad
example"
Children will eat anything and sometimes they eat things that are poisonous,
including topical medications. Therefor we have been
trying to determine the (relative) bioavailability of methyl salicylate
(MeSal) from
ointments when ingested compared with that from oil of wintergreen. We
conducted a crossover study in which 4 volunteers ingested 5 g and 15 g of
15% MeSal ointment and 1 mL of oil of wintergreen on 3 separate occasions.
We collected plasma salicylate for 24 hours after each treatment. At the
time of planning the study, we somehow managed to forget (?!!! Dr. Levy
forgive me) that salicylate displays non-linear elimination at
concentrations
that would be reached in our study. Noncompartmental methods yield a B/A
of 150%,
obviously in error due to concentration dependent clearance. I have
searched the literature for
methods of determining the relative bioavailability of a nonlinear drug. I
attempted to use Martis and Levy's method that I adapted for relative
B/A. I attempted to use the Vd/F ratio from the WinNonlin fit to a
nonlinear model, ignoring the nonlinear protein binding and assuming that
the Vd would remain constant between treatments. That method and every
other I have tried has been frustrated by the inability to get a decent
estimate of Vmax and Km. It seems regardless what version of the Michaelis
Menten ( CL based or reparametized) differential equation I
use, Vmax and Km are highly correlated to each other and also to Vd and ka,
not to mention
extremely dependent on starting values.
I am fitting the 3 treatments simultaneously, to use the widest range of
plasma concentrations possible, as taught by a former mentor.
I will take David's suggestion and fix Km at a literature value, to try to
reduce variance in the Vmax and the other parameters. If that doesn't work
does anyone have any other tips on fitting non-linear oral absorption
models?
Bill Wolowich, PharmD
Central Ohio Poison Center
Children's Hospital and OSU
Back to the Top
Bill,
Several points. First, if you assayed for parent drug (methylsalicylate),
nonlinearity is NOT an issue. The kinetics of that compound are
linear. However,
if you assayed for TOTAL salicylate, then you have to deal with the
nonlinearity. Forgetting that point for a moment, why is an F value
greater than
100% an issue? You are comparing two oral forms not an iv form.
Invert the ratio
and now you have an F value less than 100%. It is a matter, in this
instance, of
what you designate the denominator to be. A recent paper in CPT from Neal
Benowitz's group (CPT 66, 128, 1999) indicates that the compound
undergoes total
first-pass metabolism on oral dosing. It is possible, if you are compairng the
same dosage form and adjusting for differences in doses that your larger dose
exceeded or saturated the first-pass and gave rise to greater methylsalicylate
concentrations.We have found measureable, but low, intact methylsalicylate
plasma concentrations upon dermal dosing (unpublished).
Hope these points are useful. By the way, how does one ingest an ointment
orally? Not easily, I suspect.
PharmPK Discussion List Archive Index page
Copyright 1995-2010 David W. A. Bourne (david@boomer.org)