Back to the Top
The following message was posted to: PharmPK
I am a final year Pharmacy student at the Midwestern University Chicago
College of Pharmacy. I am doing my rotation at Apria Homehealth care. My
final project is on aminoglycoside once daily dosing. I was looking at the
protocols from different hospitals and also doing research on this topic. I
will really appreciate if you can guide me and send me the current
information on this topic. Some of the questions I need answer.
1. What is the suggested initial dose?
2. How to determined the maintenance dose?
3. How often should we monitor the levels?
4. What are the benefits of once daily dosing over conventional dosing method?
I will really appreciate if you can answer my question.
Thanks
Naheed Kazmi
Naheed26.-at-.aol.com
Back to the Top
[A few replies - db]
From: "Gendron, Richard T. (GS)"
Date: Thu, 6 Dec 2001 16:27:17 -0500
To: david.-a-.boomer.org
Subject: RE: PharmPK Re: Once daily aminoglycosides
Status: R
The following message was posted to: PharmPK
Dose is based on adjusted body weight if patient is obese:
Dosing weight = Ideal weight + 0.4*(total weight-ideal weight)
Gentamicin/tobramycin 5-7mg/kg ideal body weight every 24 hours if patient
can clear drug safely (usually OK w/ CrCl exceeding 60ml/min)
There are dosing nomograms available but we simply check a trough level with
the second dose to confirm adequate drug clearance (desire complete drug
clearance, but a trough up to 1 mcg/ml is probably OK). Recheck trough level
weekly or sooner depending on patient. Baseline and serial SCr to monitor
for renal toxicity. High frequency hearing/vestibular function tests advised
for long term therapy (greater than 3 weeks).
Theoretical benefits: Peak exceeding MIC by multiple factor
(concentration-dependent killing) with very low troughs to take advantage of
post antibiotic effect (possibly less toxic).
---
From: Robert Milne
Date: Fri, 7 Dec 2001 08:11:36 +1030
To: david.at.boomer.org
Subject: RE: PharmPK Re: Once daily aminoglycosides
Status: R
I would have thought that the obtaining of such information from the
literature, rather than seeking the assistance of others, would have been part
of the project.
Bob Milne,
School of Pharmacy,
University of South Australia,
Adelaide,
Australia
---
From: Brennan & Co
Date: Thu, 06 Dec 2001 18:18:18 -0500
To: david.-at-.boomer.org
Subject: Re: PharmPK Re: Once daily aminoglycosides
Status: R
The following message was posted to: PharmPK
I would strongly recommend the addition of analytical software to ODA
dosing. We
use a 5mg/kg dose given late in the evening so that a level can be drawn with
routine AM blood draws. This gives us fairly decent -kel information while
minimizing disruptions in routine and the errors associated with such
disruptions.
Analysis is done using the USC Pack software.
Bob
---
From: Norma Rohde
Date: Fri, 7 Dec 2001 11:24:59 +1100
To: david.at.boomer.org
Subject: RE: PharmPK Re: Once daily aminoglycosides
Status: R
The following message was posted to: PharmPK
Dear Naheed,
There is quite a bit of literature published on once daily dosing. If you
could manage to source a UK or European PI for gentamicin this may give you
the information you seek as I believe once daily dosing has been approved
there for some time.
Please note though, once daily dosing is not appropriate for all
aminoglycosides since they do not all have the same renal excretory
mechanism. Tobramycin could be nephrotoxic when given once daily. I
don't have information on the others.
Norma
Back to the Top
The following message was posted to: PharmPK
Dose is based on adjusted body weight if patient is obese:
Dosing weight = Ideal weight + 0.4*(total weight-ideal weight)
Gentamicin/tobramycin 5-7mg/kg ideal body weight every 24 hours if patient
can clear drug safely (usually OK w/ CrCl exceeding 60ml/min)
There are dosing nomograms available but we simply check a trough level with
the second dose to confirm adequate drug clearance (desire complete drug
clearance, but a trough up to 1 mcg/ml is probably OK). Recheck trough level
weekly or sooner depending on patient. Baseline and serial SCr to monitor
for renal toxicity. High frequency hearing/vestibular function tests advised
for long term therapy (greater than 3 weeks).
Theoretical benefits: Peak exceeding MIC by multiple factor
(concentration-dependent killing) with very low troughs to take advantage of
post antibiotic effect (possibly less toxic).
Back to the Top
Dear Richard:
What is really meant by ideal weight? I am interested to see,
over the years, how many of these formulae there are. While they have
a spurious appeal to common sense, very few of them are capable of
validation in any rigorous way. The insurance companies at least can
point to their ideal weights as being associated with greatest
longevity. Body mass index can be correlated with body fat, and
therefore with fat-free weight. The others - I have not seen any way
to really validate them.
Further, patients with very good renal function (young
healthy teenagers, or patients with CF) may well clear the drug too
fast, and may not be able to wait 24 hours between doses. It seems
that most of the q 24 h regimens are designed for middle aged
patients with CCr between 50 and 85 ml/min/1.73msq.
In all this, one usually wants a certain AUC above the MIC.
This is done by choosing, for each individual patient according to
his/her need for the drug, a desired peak, and a trough that is
acceptably low, and then by developing a dosage regimen and achieving
these goals as precisely as possible. Don't you want to see that the
selected target peak is actually achieved? Monitor it. Is the trough
properly low? Monitor it. Then, based on this information, make an
individualized model of the behavior of the drug in that patient.
Relate the AUC to the known or estimated MIC. This is the best way I
know to evaluate the patient's (and the organism's) clinical
sensitivity to the drug, and the best way to adjust the regimen to
hit the desired target goals, which actually may have changed, based
on your re-evaluation of the patient and his/her problems.
Moreover, why wait for the second dose to check the level?
Isn't it better to find out as soon as possible how the drug is
behaving in the patient? Monitor with the first dose, so you can make
the adjustment as soon as possible. There is no need at all to wait
for a steady state, for example.
Also, there is no reason to wait for distribution to be
complete after a dose. D-optimal theory has been well known for about
30 years now, and usually it is optimal for parameter estimation to
get a true peak sample, out of the opposite arm at the end of the IV
infusion, or at the time when the peak is computed to occur after an
oral dose, and then at a time when the concentrations have fallen to
about 1/3 of the peak. These can be easily computed, and are a
significant part of optimal experimental design in drug studies. They
can serve exactly the same role in clinical settings. Forget waiting
for a steady state or for distribution to take place. That approach
was needed when linear regression on the logs of the levels was the
only method for fitting data. We have moved beyond that years ago.
On our web site (www.lapk.org is the easy one to remember)
there are also materials under various teaching topics. You can click
on section 4, modeling diffusion into endocardial vegetations, the
post-antibiotic effect, and bacterial growth and kill, for example,
and get more discussion of these topics.
Very best regards, and best holiday wishes,
Roger Jelliffe
Roger W. Jelliffe, M.D. Professor of Medicine, USC
USC Laboratory of Applied Pharmacokinetics
2250 Alcazar St, Los Angeles CA 90033, USA
Phone (323)442-1300, fax (323)442-1302, email= jelliffe.at.hsc.usc.edu
Our web site= http://www.usc.edu/hsc/lab_apk
Back to the Top
[Two replies - more good support for TDM (therapeutic drug monitoring ;-) - db]
From: Brennan & Co
Date: Fri, 21 Dec 2001 00:09:46 -0500
To: david.-at-.boomer.org
Subject: Re: PharmPK Re:
The following message was posted to: PharmPK
Bravo Roger!
The need to apply pharmacokinetic tools to ODA situations just seems
to be common sense. Why guess when you can know? Use a nomogram or mg/kg
guideline to start, but follow up with real data for that patient.
Modeling software is available and easy to use - let's use it!
High bolus dosing of aminoglycosides makes sense on various levels. So why
do we feel reluctant to take it further? The protocols I've seen adjust
for decreased renal function, but what happens to the younger patient who
has the 1-2 hr half-life. Perhaps we should use 5-7mg/kg doses q12h in
these patients.
Bob
---
From: Paul Hutson
Date: Fri, 21 Dec 2001 11:09:12 -0600
To: david.-at-.boomer.org
Subject: Re: PharmPK Re: Once daily aminoglycosides
The following message was posted to: PharmPK
I agree with Roger about checking PK on the first dose, especially in
my neutropenic patients. Also as an aside, we just submitted a
retrospective evaluation of our PK experience with carboplatin. We
found very poor predictability of carboplatin clearance with renal
function using creatinine clearance estimates with and without
adjustment for "ideal" body weight. I think that DEXA scans would be
a better measure of body fat, but even better would be the use of
multiple frequency bioelectric spectroscopy; cheaper than DEXA scans,
too.
Merry Christmas to all.
Paul
Paul Hutson, Pharm.D.
Associate Professor (CHS)
UW School of Pharmacy
NOTE NEW ADDRESS effective 6/2001
777 Highland Avenue
Madison, WI 53705-2222
Tel: (608) 263-2496
FAX: (608) 265-5421
Pager: (608) 265-7000, #7856
Back to the Top
The following message was posted to: PharmPK
Once-Daily Aminoglycosides,
My approach to once-daily aminoglycoside dosing (or LDEI) in
"elderly" patients has been to adhere to the Hartford guidelines
of <0.5ug/ml (2ug/ml for amikacin) for >=4 hours, since Hartford(1)
has published data regarding safety/efficacy. A single trough taken
at least 20 hours after the first dose can be used with Bayesian
programs to adjust the interval, which is frequently 36-48 hours or
greater in the elderly after a 5-7mg/kg dose (20-28mg/kg for
amikacin). Follow-up troughs can be ordered as indicated by the
patient's renal function (serum creatinine), and can then be used to
further adjust the dose or dose interval. However, a Bayesian
approach seem more applicable in clinical situations involving
geriatric patients with fluctuating serum creatinine, as opposed to
the Hartford nomogram approach, since <0.5ug/ml for >=4hrs becomes
more problematic in these patients.
The body weight used is DBW (DBW=IBW + 0.4(ABW-IBW)) for patients
who are 40% above ideal body weight, but this weight can also be
applied to all patients, as is the usual approach with
aminoglyosides(1).
Less conservative trough guidelines (i.e. not <0.5ug/ml for
>=4 hours) used with once-daily aminoglycoside dosing has
resulted in an increased rate of nephrotoxiciy as published
in a clinical trial involving "geriatric" patients (2).
A Bayesian approach is available in the USC pack collection
of pharmacokinetic programs (Dr. Jelliffe)(3). However, other programs
are also available and can be written.
Mike Leibold, PharmD, RPh
m_leibold.-a-.hotmail.com
References
1) Nicolau, D.P. et al, Experience with a once-daily aminoglycoside
progam administered to 2,184 adult patients, Antimicrob Agents
Chemotherap 1995;39(3):650-655
2) Koo, J. et al, Comparison of once-daily versus pharmacokinetic
dosing of aminoglycosides in elderly pateints, The American Journal
of Med 1996;177-183
3) USC Pack Instruction Manual
Back to the Top
The following message was posted to: PharmPK
Prof Jeliffe makes a valid point about untested dosing schemes. I met the
"Ideal wt + 1/2(actual - ideal )wt formula about 40 yrs ago, proposed for
neuromuscular blocking drugs, by Prof. FF Foldes. As these drugs are
hydrophillic and have a small Vd, it sounded common sense. A few years ago I
tested this on recovery time data following vecuronium. Expressing the
dosage as dose/kg formula weight, I got an Rsq. of 0.32! Using body weight,
it was the same, 0.33! and a multiple regression using wt, age and sex it
was 0.28. So much for seemingly sensible formulae. Admittedly these blocking
drugs are rather unpredictable in their effect, but back to the drawing
board. The clinical problem in our hospital was to get the assays back quick
enough to be useful in designing regimens.
Tom Torda
Back to the Top
The following message was posted to: PharmPK
Once-Daily Aminoglyosides,
The actual criterion the Hartford group used when applying
DBW (DBW= IBW + 0.4(ABW-IBW))to once-daily dosing was >=20% over IBW
(1). However, DBW can be used for all dosing situations involving
aminoglycosides.
Mike Leibold, PharmD, RPH
m_leibold.at.hotmail.com
Reference
1) Nicolau, D.P. et al, Experience with a once-daily aminoglycoside
progam administered to 2,184 adult patients, Antimicrob Agents
Chemotherap 1995;39(3):650-655
Back to the Top
The following message was posted to: PharmPK
It's an interesting point you bring up. The application of q 24 hour
aminoglycoside dosing in children with good renal function doesn't
make logical sense when you consider their clearance. Yet we have
various physicians using these regimens with confidence. Would the
patient be better served by q 18 hour dosing? Possibly. Although this
may not be valid with the neonate.
I've never been comfortable with first dose levels. You make a good
point for their use, but what about the patient that is admitted and
only receives 2-3 doses of an aminoglycoside before it is
discontinued due to many possible factors (culture positive for
something else, different sensitivity, etc). In these situations it's
an unnecessary "stick" and laboratory cost. I think first dose levels
should be individualized based on the clinical situation.
Jerry Fahrni
Jerry L. Fahrni, Pharm.D.
Staff Pharmacist
Valley Childrens Hospital
Madera, CA 93638
Ph: (559) 353-5504
Fax: (559) 353-5515
JFahrni.-at-.Valleychildrens.org
Back to the Top
The following message was posted to: PharmPK
The dosing scheme of DBW = IBW + [0.4(TBW-IBW)] is not untested versus use
of IBW weight alone for inital aminoglycoside dosing. There are multiple
publications dating back to 1977 that compare the 2 initial dosing
schemes.
Larry Bauer
Back to the Top
The following message was posted to: PharmPK
Please! Do not bring up the spectre of 18hr dosing. In 20+ years of
practice, I have always managed to avoid this particular interval. From
an on the floor, day to day point of view, "odd" intervals cause
mistakes.
Perhaps others have happier experiences, but I have yet to see a
dosing interval based on 18hr, 36hr, etc last for more than 24hrs without
a problem. We aggressively try to convert dosing to routine intervals. It
may not be optimal, but it works much better.
Interestingly enough, I've always been able to find and alternative
regimen using standard dosing intervals. Not optimal, but effective.
Bob
[So appropriate intervals should be 24, 12, 8, 6, 4...hours? - out of
ignorance I must ask; can't dosing times be scheduled by computer for
any hour in a hospital setting? - db]
Back to the Top
The following message was posted to: PharmPK
From a practical point of view, administration of doses is done within a 24hr
framework. Yes, computers can schedule the doses on non traditional intervals,
but computers do not give the meds. Very busy, short staffed nurses give the
meds. Designing odd scheduled frequencies increases the risk of error.
Back to the Top
[Two more replies - db]
From:
Date: Mon, 24 Dec 2001 07:00:34 -0800 (PST)
To: david.aaa.boomer.org
Subject: Re: PharmPK Re: Once daily aminoglycosides
The following message was posted to: PharmPK
Hello David,
They certainly can but hospital staffing is a rate-limiting step.
Generally a hospital with once-daily dosing almost always schedules
the dose during the day or evening shift. As a result, night shift
never expects to have to administer the drug and won't even bother
looking for the dose on the medication administration record. Also,
if the trough levels need to be done late at night you can bet there
isn't enough lab staff on duty to accomodate this.
Even with the technology available to us drug administration is still
in the hands of humans and subject to their own limitations (or the
institution's budget). It truly is best to keep the dosing
conventional.
Happy Holidays!
Richard Molitor, R.Ph.
http://www.angelfire.com/wa/pharmacist
---
From: Roger Jelliffe
Date: Mon, 24 Dec 2001 10:04:16 -0800
To: david.-at-.boomer.org
Subject: Re: PharmPK Re: Once daily aminoglycosides
Dear Bob:
You raise an interesting point. For example, when creatinine
clearance gets to be about 15-25, it may well be that q 24 h is too
frequent and q 48 h not frequent enough. Many patients really need
such regimens. We used to have "smart" infusion apparatus that could
take instructions from the USC*PACK programs to infuse for a desired
period, to shut off for a desired period, and to repeat this as often
as desired, from about 1976 to 1982, before the PC came out. We even
had diagrams of the syringe or bag, showing how much should remain to
be infused at various times, so that one could easily check and
verify whether or not the infusion was going OK or not. . I think
some of this is still possible with some of the current equipment,
but I do not know. It really would be a great help to be able to do
this at times. In my experience, people, especially nurses, have a
big mental block with a dosage regimen that does not come at the same
time each day. This is just what a smart pump can to, just as with
your microwave or your p[alm pilot. All things are then possible,
easy, and easily verifiable.
Best wishes for the holidays and the New Year,
Roger Jelliffe
Roger W. Jelliffe, M.D. Professor of Medicine, USC
USC Laboratory of Applied Pharmacokinetics
2250 Alcazar St, Los Angeles CA 90033, USA
Phone (323)442-1300, fax (323)442-1302, email= jelliffe.at.hsc.usc.edu
Our web site= http://www.usc.edu/hsc/lab_apk
Back to the Top
The following message was posted to: PharmPK
Hello again (gee, two posts from me on the same thread, must be the
holidays...),
Dr. Jelliffe is certainly correct about the technology having existed
for a long time. I remember the old Intelliject pumps which were
originally designed for chemo administration but later the company
tried to expand their use into other areas such as antibiotic
administration (they were VERY expensive and bulky relative to the
competition). Currently pumps such as the Deltec CADDs et al. can
also accomplish "odd" dosing regimens but again, it's a cost and
maintenance issue. In this day and age of mangled care those old
infusion devices just keep getting older and the necessary support
staff keeps getting cut back. More research into how much faith we
can put into the "post-antibiotic effect" would sure be helpful.
One other note of interest is that some hospital pharmacy departments
have had to scale back the amount of clinical monitoring services
they can offer. As people move on the ability to sustain an optimal
program gets harder and harder.
Happy holidays to all,
Richard Molitor, R.Ph.
http://www.angelfire.com/wa/pharmacist
Seattle, WA USA
Back to the Top
[Four replies - db]
From: "Thao Nguyen"
To:
Subject: RE: PharmPK Re: Once daily aminoglycosides
Date: Tue, 25 Dec 2001 18:13:25 -0800
X-Priority: 3 (Normal)
Importance: Normal
I agree with Bob 98% but not completely. Once in a while, I have a
temptation to do q18h. It looks odd but the estimated levels appear better
than either q12h or q24h pharmacokinetically (and hopefully dynamically.)
Says, a 70-year old or above patient from a nursing home was admitted into a
hospital 2nd to severe COPD exacerbation and r/o pneumonia. The patient had
some degree of azotemia as usual, i.e.. BUN/Cr: 32/1.1, unstable VS, i.e..
low grade fever, tachypnea, tachycardia, and leukocytosis. The patient was
started on cefotaxime and gentamycin empirically by an ID and ask pharmacy
to f/u gent dosing. The hospital's computer system is superb (i.e.
Meditech), being able to generate an MAR with pre-calculated schedule time
(even for an odd schedule like q18h). Hospital policy requires double check
system. RN gives dose, records time, and sends the previous 24h MAR copy to
pharmacy for review. The dose was 2.5mg LD and 1.7mg/kg q12h using ABW since
the guy's wgt is about IBW +/- 10% to aim at goal peak of 8-10 for
pneumonia, and trough of < 2.0 but preferably 1.0-1.5). Levels came back at
the third dose as 5.9/1.8. Using 1-compartment model, extrapolated Cmax and
min were about 7.2 and 1.9 with t1/2 around 6hr (ke=0.113).
So, you have a few options: i) leave the level alone. The
clearance more
likely reflected the previous 24h period. Rationale: azotemia came down
(BUN/Cr: 24/0.8) after good IV rehydration. ii) adjust the dose based on
the levels. Rationale: old man, some degree of chronic renal insufficiency
after several courses of past dehydration in the NH). q12h/t1/2 ratio = 2
(too short to get goal peak without raising trough), q24h/t1/2 = 4 (too
long, concern of prolonged drug free interval and not-enough post-antibiotic
coverage from a peak of 7-10). q18h/t1/2 ratio = 3. Looks good to me.
Interval is long enough to reach target peak and trough. iii) Wait 1 more
day for sputum Cx result coming back for ABX re-evaluation (possible
discontinue gentamycin?)
But Bob is probably right: 'I've always been able to find an
alternative
regimen using standard dosing intervals. Not optimal, but effective.' So,
why bother with all of those above? Anyway, 'we treat patients, we don't
treat levels,' said one of my ID folks. The problem is when you do some
calculation, you want the numbers look good for the sake of calculation as
well as presentation. Second problem is it is the mean +/- 2 SD to make
your 95% confidence that your calculation is the true mean. Finally, the
reason for that 95% CI is you possibly have more than a few assumptions on
your calculation (one-compartment, stable renal function, stable volume of
distribution, etc.).
But I am proud to tell you this, Dr. Jelliffe and his colleagues at the
University of Southern California and around the world are very close to an
unprecedented pharmacokinetic wonder. Imagine you can estimate a dose of
many drugs which can give you an estimated levels from a much narrower range
by using stored database from real patients at your hospital. The computer
program will allow you to estimate an optimal dosage after taking into
account maximal therapeutic outcome and minimal toxicity concern. After 30
years of serving mankind, Dr. Jelliffe, the father of many great
pharmacokinetic formula, is doing us and our patients another big favor.
Congratulation and thanks, Dr. Jelliffe.
Thao Nguyen, Pharm.D., BCPS
Fellow in Antimicrobial Pharmacokinetics/Pharmacodynamics
School of Pharmacy
University of Southern California
1985 Zonal Ave
Los Angeles, CA 90089
Voice: 323-442-1334
Fax: 323-442-1395
Email: tdn.at.usc.edu
---
From: Roger Jelliffe
Date: Wed, 26 Dec 2001 13:23:03 -0800
To: david.-a-.boomer.org
Subject: Re: PharmPK Re: Once daily aminoglycosides
Dear All:
Another thing about aminoglycoside dosing. When using
D-optimal design, and simulation studies, it often turns out that a
good pair is at the peak with the first dose, and then at about 20-22
hours into the regimen, regardless of whether the dosage is q 8 h, q
12 h, or q 24 h, when creatinine clearance is at least 50
ml/min/12.73 msq. This puts the second sample about 2-3 hours before
the trough. Because of this, it may well be useful to center the
dosing schedule to come about 2-3 hours after the routine morning
blood drawing time. Then the AM second sample is close to D-optimal.
A combination of good math and practical common sense.
Very best wishes for the holidays and the New Year,
Roger Jelliffe
Roger W. Jelliffe, M.D. Professor of Medicine, USC
USC Laboratory of Applied Pharmacokinetics
2250 Alcazar St, Los Angeles CA 90033, USA
Phone (323)442-1300, fax (323)442-1302, email= jelliffe.-at-.hsc.usc.edu
Our web site= http://www.usc.edu/hsc/lab_apk
---
From: Roger Jelliffe
Date: Wed, 26 Dec 2001 14:31:09 -0800
To: david.-at-.boomer.org
Subject: Re: PharmPK Re: Once daily aminoglycosides
Dear Richard:
Thanks for your comments about faith in the post-antibiotic
effect. The Zhi model has been around for a while, and seems to be
not bad for a general "worst case" scenario in modeling the effect of
a drug on bacterial growth and kill.
We have combined this with a model of diffusion into a
simulated microorganism to mimic a post-antibiotic effect (saying
NOTHING about mechanism) with coefficients such that growth is
inhibited, for example, for 6 hours after the serum concentrations
fall below the MIC, to mimic a 6 hour post-antibiotic effect.
The combination of these 2 effects is used in the USC*PACK
programs to examine the possible overall effect of all this on the
growth and kill of organisms.
The model is NOT realistic in at least 2 aspects:
1. The organisms are always in their logarithmic growth phase.
This is not realistic, as Vinks, and probably others, point out
correctly that organisms run out of substrates and their growth
slows, and may even max out. Here, in contrast, they are always at
their worst, in their log phase of growth.
2. The organisms never get resistant in the model, while in
reality they do.
However, the model can be used as a useful "worst care" view
of what might be going on clinically. Yes, the organisms are in their
log phase at all times. In addition, you, as a clinician, can ask
yourself how resistant you think these organisms might get, and enter
that resistant value of the MIC . Then they are that resistant from
the very beginning.
Now, if a dosage regimen generates a serum concentration
profile that will result in a good kill in the model under stated
circumstances, there is a good chance that it should be successful
clinically. This is generally what we have seen. More detailed
information can be found on our web site, under teaching topics,
section #4, "modeling diffusion into endocardial vegetations, and the
post-antibiotic effect", and section 5, "modeling antibiotic growth
and kill".
I hope this is helpful. All the best wishes for the New Year,
Roger Jelliffe
Roger W. Jelliffe, M.D. Professor of Medicine, USC
USC Laboratory of Applied Pharmacokinetics
2250 Alcazar St, Los Angeles CA 90033, USA
Phone (323)442-1300, fax (323)442-1302, email= jelliffe.aaa.hsc.usc.edu
Our web site= http://www.usc.edu/hsc/lab_apk
---
From: "Carl Kirkpatrick"
Date: Thu, 27 Dec 2001 16:46:49 +1300
To: david.-a-.boomer.org
Subject: Once daily dosing
The following message was posted to: PharmPK
Hi All
I have a couple of questions/comments regarding this thread.
1) why would you want to go back to using a nomogram approach for
extended interval dosing, when these were proven to be less reliable
than dose individualisation methods (eg Sawchuck & Zaske and Bayesian
methods) at acheiving desired peaks and troughs for conventional
methods? If it is agreed that with conventional dosing that dose
individualisation (using the methods above) was "best practice", how
can it now be ethical to use nomograms with once daily dosing?
2) I have problems with nomograms and trying to identifying patients
with unusual or changing pharmacokinetics eg patients with poor or
deteriorating renal function between doses may be exposed to high
drug concentrations for long periods, potentially increasing the risk
of toxicity, as well as preventing reversal of adaptive resistance.
Also the Hartford method does not accommodate any ability to identify
those patients that may be under-dosed due to underlying
pathophysiological processes eg patients with burns, or cystic
fibrosis. Both these scenarios can be dealt with by using dose
individualisation methods.
3) What is the point of measuring a concentration at 20 hours? In
patients with normal renal function, this will be around or below the
limit of detection of the assay (0.3mg/L for gentamicin). Surely if
you are going to use bayesian analysis, the aim is to get a
concentration that provides the best information about Vd and CL with
minimal assay error. On this basis, a mid dose sample between 6 to 12
hours would seem far more useful if you were only going to take 1
sample. Whats wrong with taking two samples ie peak and mid dose to
provide information about Vd and CL? That way you can really dose
individualise!
4) Also, why not monitor after the first dose so that any future
therapy can be individualised to that patient from that point
onwards. If you wait until day 3, then you will have to wait another
24 hours for samples to come back, you have effectively been wasting
4 days of therapy. How can this be best practice?
5) Whats with the rush to give the next dose?? Are we forgetting
about adaptive resistance? Aminoglycosides are always administered
with another antibiotic with synergistic or additive effects, so the
patient is still effectively being "covered" by antibiotics.
Carl Kirkpatrick
Department of Clinical Pharmacology
Christchurch Hospital
New Zealand
Back to the Top
Dear Thao:
Many thanks for your very kind comments. First, about the
numbers, and the fact that some say that "we treat patients, not
levels". That is quite true. In doing this, we in our laboratory have
tried specifically to consider everything possible, to plan things in
advance so that the outcome is most likely to be optimal for each
individual patient. This involves weighing the specific risks and
benefits in a quantitative way, based on the principles of decision
analysis, for each patient. The "levels" chosen are usually what the
responsible clinician thinks are the best target goals to shoot for,
individualized for that particular patient, according to his/her
specific needs. That is why we like to set target level goals for
each individual patient.
If the patient's organism is so unkind as to have an MIC, or
the patient is so unaccommodating as to have a renal function that
might do better with a q 18 h or q 36 h dosing regimen, for example,
why should we then try to make the dose come at the same time each
day, knowing that it really is not optimal? I grant you it is
certainly easier, and much easier to keep track of, and all that also
needs to be taken specifically into account in planning the regimen
for each patient. I think you want the numbers to "be good" for each
patient, according to the patient's real needs, according to one's
best judgment about what the best target goals are for each
individual patient.
About your second comment, you are far too kind. We have
always tried to be able not only to take optimal care of the stable
patient, but also the highly unstable one. That is why we developed
our method of estimating creatinine clearance without a urine sample,
even when serum creatinine is changing rapidly from day to day, and
incorporated that into our USC*PACK software. That is why we have
used maximum aposteriori probability (MAP) Bayesian adaptive control
of dosage regimens, based on parametric population PK models, with
individual Bayesian posterior PK models, based on fitting to the
available serum data, the various doses given, and the patient's
changing renal function and changing body weight from dose to dose.
More recently, we have begun to develop a newer method of
Bayesian adaptive control based on a combination of nonparametric
(not noncompartmental) population PK models and "multiple model" (MM)
dosage design. The method is well known in the aerospace industry,
where it is widely used in flight control and missile guidance
systems. In this method, it is possible to estimate in advance the
weighted squared error with which any control input such as a dosage
regimen is predicted to hit a desired target goal. Based on this, one
can then compute the regimen which specifically minimizes that error.
Thus the new MM method develops dosage regimens designed to hit
target goals specifically with maximum predicted precision over the
planned period of therapy.
We are also incorporating a new Bayesian "interacting
multiple model" (IMM) approach (used in tracking evasive targets in
the aerospace industry) which can detect unsuspected changes in a
patient's Bayesian posterior PK parameter values during the period of
the data analysis. This has been a big help, we think, in tracking
the behavior of drugs (a patient's changes in aminoglycoside volume
of distribution, for example), as his/her clinical status changes. We
have seen this to be very helpful in at least one notable instance,
where the patient's Vd went from about .18 L/kg to .55, with a
surprise relapse. This took place about 10 years ago. At the time, we
got a very poor fit when using all the levels, and had to break up
the data into 3 phases, which more clearly reflected the patient's
different clinical phases. Later on, when the patient had
significantly improved, the Vd came back to about .2 L/kg. Such
changes, to our knowledge, were first described as taking place in
individual patients by Marcus Haug and Peter Slugg at the Cleveland
Clinic, who called it "Vd collapse" and correlated it with the fact
that the patient was probably going to get well.
The new "MM-USC*PACK" software is now in its beta phase of
testing, and should hopefully be ready for clinical use this spring
sometime. Give us an email if you would like to help us evaluate it.
More information on the MM and IMM approaches can be found on our web
site, under "teaching topics".
All the very best wishes for the New Year,
Roger Jelliffe
Roger W. Jelliffe, M.D. Professor of Medicine, USC
USC Laboratory of Applied Pharmacokinetics
2250 Alcazar St, Los Angeles CA 90033, USA
Phone (323)442-1300, fax (323)442-1302, email= jelliffe.-at-.hsc.usc.edu
Our web site= http://www.usc.edu/hsc/lab_apk
PharmPK Discussion List Archive Index page
Copyright 1995-2010 David W. A. Bourne (david@boomer.org)