Back to the Top
OK, so I know how to calculate CrCl, IDBW, Dosing Wt, T1/2, Ke, Vd, etc.
But where I, and others, need help is what gets done 1st, 2nd, 3rd, 4th,
......etc when working up real patients who require dosing computations. A
decision tree would be a big help or a a step by step approach to which
calculations to do and in what order.
Can anyone help us poor soles?
Thanx
Hank
Back to the Top
[A few replies - db]
Reply-To:
From: "Robert Aucoin"
To:
Subject: RE: PharmPK Kinetics help needed
Date: Thu, 10 Jun 1999 18:50:23 -0500
MIME-Version: 1.0
X-Priority: 3 (Normal)
Importance: Normal
The problem that I have found with working with "real patients" is that all
of the computations in the world will not predict a steady outcome. I have
tracked 62 critically ill children, ages 9 days to 18 months. While I did
get steady state levels with the target drugs after a while, initially it
was impossible to anticipate levels based on common pharmacokinetics.
ie: 18 month old female, spina bifida, in need of a nissan fundoplication
revision. Developed a lung abscess post, returned to surgery for excision.
36hrs post-op developed DIC, and ARDS. Placed on Vancomycin at 10mg/kg q
6hr. Pk/Tr drawn around 4th dose. Tr = 3.5; Pk = 12.6. Adjusted the dose
to 15mg/kg q6hr. Second Pk/Tr Tr =11.3; Pk=13.6. Checked with lab for draw
times. Had an ace nurse on duty who logged the "actual" time of both draws.
Her levels bounced all over the board for 4 days till she stabilized.
Short story:
In the Septic patient with ARDS, DIC and a "little bit" of
hepatomegelia,
kinetics is about a useful as teeth on a chicken.
I dose them depending on their age and "dry weight". Shoot for
something
resembling therapeutic and keep them from being in the mega zone.
Its not just antibiotics. A seizure kid with sepsis secondary to
aspiration pneumonia is very hard to dose on a number of anticonvulsants.
We often snow them with continuous Versed especially if they are vented.
I really admire all of you folks out there doing this work. I
understand
most of it, however when the rubber hits the road, a lot of what you
calculate is not usable. Keep up the good work. We need the studies to get
the drugs to market to help the people like the little tykes I work with.
Just passing through...... Robert
Peds Clinical Pharmacist 888-765-7428 (free)
The Children's Center at fax 225-765-7917
Our Lady of the Lake RMC
Baton Rouge, LA.
---
Date: Thu, 10 Jun 1999 21:49:10 -0400
From: Brennan
MIME-Version: 1.0
To: PharmPK.aaa.boomer.org
CC: Multiple recipients of PharmPK - Sent by
Subject: Re: PharmPK Kinetics help needed
Hank
The first thing to do is read the history and take a look at the patient.
Bob B
---
Date: Thu, 10 Jun 1999 20:24:14 -0300
From: "Peter W. Mullen"
MIME-Version: 1.0
To: PharmPK.aaa.boomer.org
Subject: Re: PharmPK Kinetics help needed
Am I the 'sole' person on this list to think that there's something "fishy"
about this request? There's a whole "school" of thought here? :-)
Regards,
Peter
Peter W. Mullen, Ph.D.
Kemic Bioresearch Laboratories
P.O. Box 878
Kentville
Nova Scotia, B4N 4H8
Canada Tel. (902) 678-8195 Fax (902) 678-2839
Back to the Top
[Two replies - db -- A few administrative commnets: As with all listserv
there are a number of bad addresses in the list. This doesn't mean much to
you but to me it means I get a bounced message for each message for each
bad address - consequently I sometime aggressively delete addresses that
send me something that looks like a bad address. Most of the time this is
appropriate, other times it may only be a temporary problem. If I delete
your address by mistake...SORRY...I didn't mean it...please resubscribe
from the PharmPK page at http://www.boomer.org/pkin (or send me a message).
With the new listserv software version it will now send a confirmation
message...simply replying to this message will confirm your subscription.
This should help cut down on the number of bad address. A reminder about
digests. If the number of messages is too many for your there is a digest
mode. Send 'set PharmPK digest' to macjordomo.at.boomer.org to receive the
messages in batches...now back to the replies]
From: "Bruce CHARLES"
Organization: School of Pharmacy
To: PharmPK.-at-.boomer.org
Date: Fri, 11 Jun 1999 14:08:13 +1000
MIME-Version: 1.0
Subject: Re: PharmPK Re: Kinetics help needed REPLY
X-Confirm-Reading-To: "Bruce CHARLES"
X-pmrqc: 1
Priority: normal
Why are you routinely monitoring vancomycin .....the cost benefits
are doubtful at best...??
However, whatever the drug of interest, I don't think PK has ever
claimed to be able to produce, simplified and comfortable solutions
to complex and unstable clinical situations such as you describe.
Certain underlying pathophysiological processes events are
causing this unpredictability (e.g. haemodynamic instability
affecting renal perfusion, changes in V due to tissue sequestration
of water in bad sepsis, etc ,etc, etc). Sometimes the only (slightly
painful) way is to acccumulate the data on a given drug of interest in
numbers approaching those you describe and try to identify and, if
you're lucky, quantify any trend(s).
If the dose of a drug in common use needs to be changed for
reasons of efficacy and/or toxicitiy then perhaps a population
approach in your very sick infants could be useful (linked to clinical
outcomes?), particularly as a foundation for Bayesian forecasting in
individuals.
You have a source of potentially rich data for doing this very thing.
I know this may not help in the short term for any given patient and
it may not end up giving the "right" answers all (or any) of the time,
but it may be better than throwing darts in Baton Rouge which is the
other alternative.
Cheers,
BC
Bruce CHARLES, PhD
Senior Lecturer
School of Pharmacy
The University of Queensland
Brisbane, QLD Australia 4072
TEL: +61 7 336 53194
FAX: +61 7 336 51688
Email: Bruce.aaa.pharmacy.uq.edu.au
---
From: "Thomas Senderovitz"
To:
Subject: Sv: PharmPK Kinetics help needed
Date: Fri, 11 Jun 1999 08:06:21 +0200
MIME-Version: 1.0
X-Priority: 3
Hank,
As i see it, this is not a simple question to answer. Are you asking how to
dose real patients by way of PK/PD computations? Are you a clinician?
If the drugs ind mind have very broad therapeutic intervals (TI), no TDM is
needed. If you can monitor the treatment clinically (i.e. good effect
monitoring), you also do not need TDM (therapeutic drug monitoring). So,
what you are probably after is some help in dosing drugs with very narrow
TI's, right? Then, first of all you need to know EXACTLY how well tese
intervals have been deermined. This is itself a very difficult task. For a
lot of drugs, these TI are old, and not based on good clinical studies.
Say, then, that you have good, valid intervals, there are several
possibilities. You might use software such as USC Pack or MW Pharm to
individualize your dosage regimens (Bayesian approach), but this requires
some knowledge of the softwae, the maths and PK. You could of course do
some basic calculations without knowledge of the population PK data, but I
think this is less useful. In other words: Could you be more specific in
your question? This way, you would probably get more specific answers.
Regards,
Thomas Senderovitz
Back to the Top
To expand on Thomas's comments:
>As i see it, this is not a simple question to answer. Are
you asking how to
>dose real patients by way of PK/PD computations? Are you a
clinician?
>If the drugs ind mind have very broad therapeutic intervals
(TI), no TDM is
>needed. If you can monitor the treatment clinically (i.e.
good effect
>monitoring), you also do not need TDM (therapeutic drug
monitoring).
Other areas when TDM (or specific implementations of TDM may
have difficulties) can be thought of with respect to
different aspects of variability in parameter values.
1) Between subject variability (also termed interindividual
variability). If this is very small compared to the
therapeutic index then TDM is not needed since in essence
one dose can be used for all. In contrast, if this type of
variability is very big then Bayesian type TDM procedures
become less helpful since the population data is of little
help in identifying an individual's response. Indeed as
between subject variability increases more observations/dose
will be required as the prior in the Bayesian approach
becomes less informative.
2) Between occasion variabilty (also termed interoccasion
variabiltiy). This is random variability in the values of
the parameters for an individual between dosing occasions.
If this is large compared to the therapeutic index then TDM
is of no value (indeed the drug would be very difficult to
dose). If there was no between occasion variability then
you would only need to estimate the parameter values once
for a given patient.
In addition the influence of large residual variability (eg
a bad assay, model misspecification etc etc) and systematic
changes in a patient's parameters values over the course of
treatment (eg the patient is getting better [hopefully])
will also make TDM more difficult to interpret.
These levels of variability do *not* mitigate against a
quantitative approach to dosing of drugs in specific patient
groups - indeed an indepth understanding of how various
levels of variability influence patient response is
paramount to good clinical practice.
Steve
=====================
Stephen Duffull
School of Pharmacy
University of Manchester
Manchester, M13 9PL, UK
Ph +44 161 275 2355
Fax +44 161 275 2396
PharmPK Discussion List Archive Index page
Copyright 1995-2010 David W. A. Bourne (david@boomer.org)