Back to the Top
I am forwarding the attached message to this group. Please respond to
Lisa directly at pharmgirl.-a-.centuryinter.net.
My name is Lisa Cone and I am a hospital pharmacist at White River
Medical Center in Batesville, Arkansas. We have a once-daily dosing
protocol for aminoglycosides (mainly gentamicin) here. I have been
supplied with basic information supporting the use and efficacy of
this dosing schedule. However, none has been able to provide me with
the exact research that led to the specific way in which we dose the
patients. In short, the patient is given an initial dose based on
body size and creatinine clearance. Then a "random" gentamicin level
is drawn eight to ten hours post end of infusion. We have a nomogram
on which we locate time post infusion (y-axis) and gent level
(x-axis). The point of there intercept gives a dosing interval (q 24,
q 36, q 48 hrs.). The dose is supposed to remain constant, with only
the dosing interval changing. My request of you is this: Can you
direct me to a person, article, or other source of information which
will explain different modalities for dosing the "once-daily"
aminoglycoside regimen? I do not need to know the theory behind WHY
it works (I understand this already), just how to use it.
Your help is appreciated.
Lisa M. Cone, Pharm.D
Please use the following address to contact me. Thanks.
PharmGirl.aaa.centuryinter.net
--
Gopal Krishna, Ph.D.
Drug Metabolism and Pharmacokinetics, Schering-Plough
Email: Gopal.Krishna.-a-.spcorp.com, Phone: 908-298-6564 (work)
Editor and maintainer : Pharmacokinetics/Biopharmaceutics home page
http://griffin.vcu.edu/~gkrishna/PK/pk.html
Back to the Top
Dear PharmGirl
Clinical Trends in Pharmacy Practice Vol. 10 #1 1996 published by
Fujisawa USA Inc. has an article "Once Daily Aminoglycosides". The
telephone number is:
1-800-727-7003. The web site is: www.fujisawausa.com
Lippincott's HOSPITAL PHARMACY August 1995 Vol 30 #8 has an article:
Implementation of a once-daily aminoglycoside program in a large
community-teaching hospital The telephone numberr is: 1-800-777-2295
I hope this will help you.
Joan
Back to the Top
There are been many methods published for dosing once daily
aminoglycosides. The method that you are referring to is:
Nicolau, et al. Experience with a once-daily aminoglycoside
program administered to 2184 adult patients". The method uses a 7
mg/kg dose for all patients. Obese patients are dosed based on
adjusted body weight rather than total body weight. The dosing
interval is selected based on the estimated creatinine clearance.
Your can refer to the publication for specifics. This dosing
strategy is based on an MIC of 2 mcg/ml, a peak to MIC ratio of 10:1
and allowing the drug concentration to fall below 0.5 mcg/ml for at
least 4 hours during a dosing interval. The dosing works well most
of the time. There are two concerns:
1. Patients with fast elimination (t1/2 < 2 hours) will have
undetectable concentrations for >12 hours during each daily dosing
interval. However, this is usually not a problem since
aminoglycosides are combined with beta-lactam agents in most
patients. I would be concerned about this prolonged drug free period
in patients treated with aminoglycoside only outside the urinary
tract. However, I see few indications for using aminoglycoside
alone.
2. As renal function declines, some patients will have
measurable concentrations at the end of a dosing interval. Using a
dose of 7 mg/kg, peaks of 15-40 can be expected. If the trough is
still detectable at 24 hours, then the patient will have greater
exposure to aminoglycoside (AUC) than was used in the past when
dosing was adjusted to target peaks of 6-10 and troughs of 0.5-2.
There have been cases of severe toxicity with ODA aminoglycoside.
Some investigators continue to use 2 serum concentrations and
ensure that the AUC is not too high.
Begg, et al. A suggested approach to once-daily aminoglycoside
dosing. Br. J Clin Pharmacol. 1995;39: 605-9.
Barclay ML, et al. Experience of once-daily aminoglycoside dosing
using a target area under the concentration-time curve. Aust NZ J
Med. 1995;25:230-5.
This is always difficulty in patients that have borderline renal
function to allow once daily dosing. Prolonging the interval to 36
hours is logical, but many hospitals do not handle Q 36h dosing very
well. Going to 48 hours may be too big of a jump. Based on
available studies, once daily (pulse) dosing is well established in
patients with normal renal function and creatinine clearance >60-70
ml/min. Note that some patients with estimated creatinine clearance
of 65 ml/min for example may not pre predicted well, and may
eliminate gentamicin like a patient with predicted creatinine
clearance of 50 ml/min. Thus I feel that two levels may be
appropriate in patients with impaired renal function due to advanced
age or renal disease.
The targets for aminoglycoside ODA dosing were extrapolated from
studies performed 30 years ago with three times a day dosing. We are
not sure if the Cmax to MIC ratio should be 10, 15, or 20. Nearly
all studies involve concomitant beta-lactam agents. It's important
that except for neutropenic patients, and patients with Pseudomonas
bacteremia, combination therapy has not been shown to be better than
beta-lactam alone in a randomized clinical trial. There are some
exclusions for ODA including pregnancy patients, children (?),
patients with poor renal function (? prolonged interval with high
peak OK), and use of aminoglycoside for synergy (Enterococcus, S.
aureus, S. viridans endocarditis).
Below is my method for ODA aminoglycoside:
1. Start dosing with 7 mg/kg (adjusted body weight if obese)
2. Predict peak and trough levels using population PK parameters,
estimated creatinine clearance and weight just like we did
with traditional dosing.
3. Obtain single concentration at 8-12 h post dose. If not
detectable, make sure that patient is on concomitant
antibiotic effective against the suspected pathogen(s).
Compare the observed concentration with a predicted
concentration using population PK values. Make sure that
level make sense. If Cobs>Cpred, either CL or V is less than
predicted from population values. If CobsCL or V is greater than predicted. Use esperience to ensure
that the measured concentration seems reasonable.
4. Continue dosing as is or make adjustments based on pathogen MIC
if available (be more aggressive if MIC=4, less aggressive if
MIC=0.5), avoiding excessive exposure (AUC >100-120 mg*h/L)
Cmax:MIC target is 10 by default since there are no better
studies. Some patients may require more than one serum
concentrations if the initial level is considerably different
than predicted. Place most of your effort for patients that
are likely to be treated with aminoglycoside long term
duration of therapy is the strongest predictor of toxicity.
Back to the Top
Dear David Nix
A comment regarding your method...
> 3. Obtain single concentration at 8-12 h post dose. If not
> detectable, make sure that patient is on concomitant
> antibiotic effective against the suspected pathogen(s).
> Compare the observed concentration with a predicted
> concentration using population PK values. Make sure that
> level make sense. If Cobs>Cpred, either CL or V is less than
> predicted from population values. If Cobs> CL or V is greater than predicted. Use esperience to ensure
> that the measured concentration seems reasonable.
If you measure a single concentration at between 8 and 12 hours post dose
then differences between Cobs and Cpred are more likely to be due to a
variation in Cl rather than V. The influence of variability in V on the
concentration-time curve for a drug displaying 1-cpt PK is more marked (in
absolute terms) within the first half-life. After this Cl is the main
determinant. Given the relationship between Cl and AUC, a high Cobs
(relative to Cpred) will represent a higher than expected AUC (?increased
risk of toxicity). Interpreting Cobs when it is close to the limit of
detection may also be problematical (without the aid of Bayesian dose
individualisation methods).
> 4. Continue dosing as is or make adjustments based on pathogen MIC
> if available (be more aggressive if MIC=4, less aggressive if
> MIC=0.5), avoiding excessive exposure (AUC >100-120 mg*h/L)
> Cmax:MIC target is 10 by default since there are no better
> studies. Some patients may require more than one serum
> concentrations if the initial level is considerably different
> than predicted. Place most of your effort for patients that
> are likely to be treated with aminoglycoside long term s
Based on this method how do you account for patients who are not population
average, eg patients with burns, cystic fibrosis, in ICU ... ? These
patients may have a greater Vd than expected and hence not achieve the
Cmax:MIC ratio that is desirable.
I definitely agree about monitoring patients "closely" who will be
receiving Ag long term. This is an area that needs further work for ODA Ag.
I would be interested in your comments.
Regards
Steve
=======================Steve Duffull
Christchurch
New Zealand
Ph +64 3 3815280
Fax +64 3 3640902
Email sduffull.-at-.clear.net.nz
Back to the Top
> If you measure a single concentration at between 8 and 12 hours post dose
> then differences between Cobs and Cpred are more likely to be due to a
> variation in Cl rather than V. The influence of variability in V on the
> concentration-time curve for a drug displaying 1-cpt PK is more marked (in
> absolute terms) within the first half-life. After this Cl is the main
> determinant. Given the relationship between Cl and AUC, a high Cobs
> (relative to Cpred) will represent a higher than expected AUC (?increased
> risk of toxicity). Interpreting Cobs when it is close to the limit of
> detection may also be problematical (without the aid of Bayesian dose
> individualisation methods).
>
I totally agree with your comments. I am not happy with the
performance of a statistical Bayesian approach when only one sample
is available. I use an empirical Bayesian approach
(i.e. experience). First, I may modify the population parameter if
the patients is expected to have increased volume or clearance due to
underlying disease. The Cobs/Cpred ratio is reviewed within the
patient context. I agree that CL typically explains most of the
difference for samples taken at later time points (>2 t1/2 e.g.).
Based on the clinical evaluation, all of the error may be attributed
to CL or the error may be split between V and CL. Suppose
Cobs/Cpred=0.5, either CL or V or both are higher than the
population values. Would it make since for the volume to be higher
for the patient? If so, I may increase the volume by 25% then
recalculate the ke to make Cobs=Cpred. Hopefully, we can put enough
information into a computer someday so it can make such decisions.
I am concerned that younger practicioners may not get the experience
needed to may good decisions. Today, one only one sample, there is
rarely enough feedback to allow good estimates of V and CL. Without
good estimates, it is difficult to gain experience on V and CL is
affected by disease states and clinical condition.
> Based on this method how do you account for patients who are not population
> average, eg patients with burns, cystic fibrosis, in ICU ... ? These
> patients may have a greater Vd than expected and hence not achieve the
> Cmax:MIC ratio that is desirable.
I often will make empiric adjustments in the population parameters
for differenct clinical circumstances. The one concentration is then
used to confirm the adjustment. We often limit the use of
aminoglycosides in patients with unstable renal function, liver
dysfunction, or in patients at high risk for ARF - unless the
aminoglycoside has been proven to provide additional efficacy
(serious Pseudomonas infection, profound neutropenia, etc.) We also
may use the combination of beta-lactam/fluoroquinolone in such
patients.
>
> I definitely agree about monitoring patients "closely" who will be
> receiving Ag long term. This is an area that needs further work for ODA Ag.
>
ODA works well for patients with good renal function. I am not
totally confortable with using pulse dosing (extended interval) in
patients with impaired renal function (Ccr <60). We clearly need
specific, controlled studies in these patients.
Back to the Top
Another comment on ODA aminoglycosides:
Dear Steve and David:
It is good to hear from you, and to hear your views on ODA
aminoglycosides. I have the basic question, and that is, what are your
goals for the serum levels at the various times you get them? What SHOULD
these levels be? For example, what peak value do you expect when you start
with 7 mg/kg, and what sort of a level do you expect when you draw one at
8-12 hrs post-dose? Aren't you really setting those as your goals, and then
developing a regimen to hit these targets?
As you measure your level, or levels, how do you distinguish between a
patient with a different Vd from predicted, or one with a different
clearance or Kel? It may well make a difference in the kill you get from
the regimen, depending on the MIC, for example. It is true that you can do
Bayesian fitting with only 1 level and see what you get, and it helps, but
I have usually had the experience that a single level usually does not
distinguish well between Vd and either clearance or K. But I don't have any
big disagreement there, just the general impression that any experimental
design is richer when you have at least 1 level for each parameter you are
seeking to estimate, whether it is in a research study or for a study of a
patient to optimize his care.
In general, it seems to me that what we really want to do is to select
some targets such as peak and 8-12 hr levels, and then plan, monitor, and
adjust our regimen until we do an OK job of hitting these targets. This
then puts us in a much more flexible and optimal position, especially with
regard to renal function, than does a more rote ODA approach.
Also, some patients are quite different from others - ICU patients from
general nedical patients, for example. And it looks as though each patient
makes this transition from one to the other, as did the patient we saw
together in 1991, who unexpectedly relapsed and went into septic shock,
whose Vd of his central compartment went from about .18 to about .55 L/kg,
and then came back toward the smaller value when he got better again.
In general, I have liked to set target goals - a peak and an 8-12 hr
target, for example, based on the seriousness of the patient's need for the
drug and the risk of adverse reactions it appears justified to accept in
order to get the benefits of the drug, and then to hit these targets as
precisely as possible. One doesn't want the patient to run any more risk
than is justifiable, but also we want the patient to get maximum benefit
(as much drug or AUC or whatever) within that constraint. This is why I
always like to set target goals, and then to individualize the regimen to
hit the targets as precisely as possible. One of the strengths of
nonparametric population modeling is that it leads naturally into the
"multiple model" method of designing dosage regimens specifically to hit
the selected targets with the greatest possible precision. This is not
possible with current methods that use only a single value for each PK
parameter.
I agree heartily that you want to know or estimate the MIC of the
bug and
to select the target goals with this in mind.
When Cobs is near the limit of detection, then we get into the real
difference between toxicology, where you have no other info except from the
assay as to whether there is drug present or not, and TDM, where you know
the time after the dose, and you therefore know there is drug there, and
your only question is how much. We have talked recently in PharmPK about
this, and it seems to me that for TDM, there really is no lower detection
limit, as you already knoe the drug is there. It is a totaly different
question we are asking in TDM (How much - we know it is there) from
toxicology (is there anything there or not, and how do we tell this from a
blank?).
Hope this adds something - I just wanted to get back to target
goals, the
patient's renal function, and not just ODA. I think this is important.
Best regards to all,
Roger Jelliffe
At 09:46 AM 10/30/97 -0600, you wrote:
>PharmPK - Discussions about Pharmacokinetics
> Pharmacodynamics and related topics
>
>Dear David Nix
>
>A comment regarding your method...
>
>> Below is my method for ODA aminoglycoside:
>>
>> 1. Start dosing with 7 mg/kg (adjusted body weight if obese)
>> 2. Predict peak and trough levels using population PK parameters,
>> estimated creatinine clearance and weight just like we did
>> with traditional dosing.
>> 3. Obtain single concentration at 8-12 h post dose. If not
>> detectable, make sure that patient is on concomitant
>> antibiotic effective against the suspected pathogen(s).
>> Compare the observed concentration with a predicted
>> concentration using population PK values. Make sure that
>> level make sense. If Cobs>Cpred, either CL or V is less than
>> predicted from population values. If Cobs>> CL or V is greater than predicted. Use esperience to ensure
>> that the measured concentration seems reasonable.
>
>If you measure a single concentration at between 8 and 12 hours post dose
>then differences between Cobs and Cpred are more likely to be due to a
>variation in Cl rather than V. The influence of variability in V on the
>concentration-time curve for a drug displaying 1-cpt PK is more marked (in
>absolute terms) within the first half-life. After this Cl is the main
>determinant. Given the relationship between Cl and AUC, a high Cobs
>(relative to Cpred) will represent a higher than expected AUC (?increased
>risk of toxicity). Interpreting Cobs when it is close to the limit of
>detection may also be problematical (without the aid of Bayesian dose
>individualisation methods).
>
>> 4. Continue dosing as is or make adjustments based on pathogen MIC
>> if available (be more aggressive if MIC=4, less aggressive if
>> MIC=0.5), avoiding excessive exposure (AUC >100-120 mg*h/L)
>> Cmax:MIC target is 10 by default since there are no better
>> studies. Some patients may require more than one serum
>> concentrations if the initial level is considerably different
>> than predicted. Place most of your effort for patients that
>> are likely to be treated with aminoglycoside long term s
>
>Based on this method how do you account for patients who are not population
>average, eg patients with burns, cystic fibrosis, in ICU ... ? These
>patients may have a greater Vd than expected and hence not achieve the
>Cmax:MIC ratio that is desirable.
>
>I definitely agree about monitoring patients "closely" who will be
>receiving Ag long term. This is an area that needs further work for ODA Ag.
>
>I would be interested in your comments.
>
>Regards
>
>Steve
>=======================>Steve Duffull
>Christchurch
>New Zealand
>Ph +64 3 3815280
>Fax +64 3 3640902
>Email sduffull.-a-.clear.net.nz
************************************************
Roger W. Jelliffe, M.D.
USC Lab of Applied Pharmacokinetics
CSC 134-B, 2250 Alcazar St, Los Angeles CA 90033
Phone (213)342-1300, Fax (213)342-1302
email=jelliffe.at.hsc.usc.edu
************************************************
Take a look at our Web page for announcements of
new software and upcoming workshops and events!!
It is http://www.usc.edu/hsc/lab_apk/
************************************************
Back to the Top
Hi Roger
> As you measure your level, or levels, how do you distinguish between a
> patient with a different Vd from predicted, or one with a different
> clearance or Kel?
We measure two concentrations. Hence both Vd and Cl can be estimated. We
usually use Bayesian techniques for this. Even a single conc can give some
info about either Vd or Cl (depending when the blood sample was taken).
> Also, some patients are quite different from others - ICU patients from
> general nedical patients, for example. And it looks as though each patient
> makes this transition from one to the other, as did the patient we saw
> together in 1991, who unexpectedly relapsed and went into septic shock,
> whose Vd of his central compartment went from about .18 to about .55 L/kg,
> and then came back toward the smaller value when he got better again.
We have experienced this problem frequently (ie rapidly changing values of
Vd and Cl without any changes in covariates). Hence in very sick patients
we monitor more regularly (every 1-2 days), versus normally every 3 days.
Regards
Steve
=======================Steve Duffull
Christchurch
New Zealand
Ph +64 3 3815280
Fax +64 3 3640902
Email sduffull.aaa.clear.net.nz
Back to the Top
Dear Steve:
Thanks for your note. Glad you usually measure 2 levels. Yes, some
patients can change their parameter values, including Vd, quite rapidly
with changes in their clinical state.
Very best regards to all,
Roger J
**********************
Roger W. Jelliffe, M.D.
USC Lab of Applied Pharmacokinetics
CSC 134-B, 2250 Alcazar St, Los Angeles CA 90033
Phone (213)342-1300, Fax (213)342-1302
email=jelliffe.-a-.hsc.usc.edu
**********************
Take a look at our Web page for announcements of
new software and upcoming workshops and events!!
It is http://www.usc.edu/hsc/lab_apk/
*********************
Back to the Top
Dear Anaizi
Re: ODA
> It is difficult to understand the sort of scientific evidence t hat must
> be provided to support the pulse dosing of aminoglycosides. Are we to
> make a distinction between evidence derived from clinical trials and
> evidence derived from basic science (laboratory / vitro / animal)
> research? What constitutes a scientific basis?
I agree that to date we have scientific in-vitro evidence that
aminoglycosides (Ag) display dose dependent kill patterns, we also know
that they have a
post-antibiotic effect, they display adaptive resistance and they have
toxicity patterns that support low trough concentrations as safer. Data
from clinical trials, and the various meta-analyses support some of these
claims most of the time (especially reduced toxicity).
The question is not whether to once daily dose, but how to once daily dose?
What we don't know is what is the best method of dosing Ag. The claim for
a lack of scientific basis is simply that it has not yet been determined
how ODA Ag should be monitored? What targets should be employed? How
should Ags be dosed (q24h or q48h)? etc. The AUC method seems sensible in
that the highest peak concentration is achieved while constraining the
exposure to an acceptable level. The "acceptable level" has also not been
defined and is currently determined as the level of exposure that was
previously accepted as appropriate when Ag's were dosed conventionally (ie
between 70 and 100 mg/L.h).
> at the present time there is no evidence to support the need for
> AUC-based dosing. In fact he suggests that at some point we should look
> into " .. whether the AUC is important or not ...".
Currently it is not known whether the AUC is the "best" method. But it is
a method that can be used to ensure that patients are not exposed to
excessive amounts of Ag. In my experience high AUC values (> 120 mg/L.h)
for long periods of time do result in toxicity. A recent paper in Clinical
Nephrology discusses the use of a single dose of 10 mg/kg of gentamicin in
the treatment of pyelonephritis followed by ciprofloxacin PO. Perhaps a
*single* pulse (ie "hit and run") is optimal?
What seems to be needed is a prospective study that addresses the effect of
different administration regimens (eg conventional vs q24h vs a48h?) while
keeping the
total daily "exposure" the same. Hence the true validity, and advantages,
of changing administration regimens (alone) can be assessed.
Kind regards
Steve
=======================Steve Duffull
Christchurch
New Zealand
Ph +64 3 3815280
Fax +64 3 3640902
Email sduffull.aaa.clear.net.nz
PharmPK Discussion List Archive Index page
Copyright 1995-2010 David W. A. Bourne (david@boomer.org)