Back to the Top
The following message was posted to: PharmPK
Dear Wolowich,
You wrote:
> This reminds me of the debacle with the MDRD
> equation...
Could you be more specific about the 'debacle with the
MDRD equation'? Is there something wrong with the MDRD
equation?
Best regards,
Hans Proost
Johannes H. Proost
Dept. of Pharmacokinetics and Drug Delivery
University Centre for Pharmacy
Antonius Deusinglaan 1
9713 AV Groningen, The Netherlands
tel. 31-50 363 3292
fax 31-50 363 3247
Email: j.h.proost.-a-.rug.nl
Back to the Top
The following message was posted to: PharmPK
Dr. Proost:
The MDRD equation is another tool to estimate GFR, no more no less.
The debacle that I refer to is the problem with the interpretation of
the results. The MDRD in its original form produces an estimated GFR
in units of mL/min/1.73m**2.
In clincial practice adjustment of drug dosing for renal
insufficiency has for many years used the Cockcroft and Gault
formula, which provides an estimate of creatinine clearance in mL/
min. The problem originates when naive clinicians assume that mL/min/
1.73 m**2 is identical to mL/min even when the patient in question is
nowhere near 1.73m**2. I have personally seen cases of improper doses
of contrast agent being given because of the misinterpreted MDRD result.
for more information see: Pharmacotherapy 2005;25:1283-84 and
Pharmacotherapy 2006;26:1041-46
W.R.Wolowich
Back to the Top
The following message was posted to: PharmPK
Dear Dr. Wolowich,
Thank you for your explanation about 'the debacle with the MDRD
equation'.
The use of two different units for creatinine clearance, i.e. ml/min and
ml/min/1.73m** is indeed confusing. In my opinion both forms have their
merits: for example, 'ml/min' should be used for dose adjustments, and
'ml/min/1.73m**' gives a better insight in the degree of renal
function of
the patient. This is not specific for the MDRD equation, but also to
Cockcroft&Gault and to the equations by Dr. Jelliffe.
Best regards,
Hans Proost
Johannes H. Proost
Dept. of Pharmacokinetics and Drug Delivery
University Centre for Pharmacy
Antonius Deusinglaan 1
9713 AV Groningen, The Netherlands
tel. 31-50 363 3292
fax 31-50 363 3247
Email: j.h.proost.-a-.rug.nl
Back to the Top
The following message was posted to: PharmPK
Hans,
I dont understand why you propose to use ml/min for dose adjustments
and ml/min/1.73m** for renal function. The ml/min/1.73m** metric is
the traditional method of obtaining a size normalized value but it
uses the discredited surface area theory of allometric scaling. A
more consistent approach which is compatible with allometric theory
and observation is illustrated below.
I prefer to convert metrics such as predicted Clcr to a nominal renal
function (RF) and then use the RF to predict the renal component of
clearance (CLr). RF will have a value of 1 in a person with 'normal'
renal function. The RF value can be used directly by clinicians to
think of renal function on 'per cent' type of scale (RF*100). The CLr
can be used as appropriate for dose adjustment.
e.g. Cockcroft and Gault formula for CLcr in a 70 kg man with a
particular Age and Scr:
CLcr(Age,Scr) = (140-Age)/(Scr mg/dL)/72*70
Notice particularly that the RF value does not require knowledge of
the patient's actual weight. The RF is normalized relative to a
person of a standard weight (70 kg in my example). The RF value will
be 1 for a patient who weighs 35 kg and has an actual CLcr of 50 ml/
min or a patient who weighs 140 kg with an actual CLcr of 200 ml/min.
The RF can be computed using other metrics (e.g. MDRD or Jelliffe
prediction of CLcr). It only requires you to choose a suitable
'normal' value for the standard 70 kg human.
RF = CLcr(Age,Scr)/CLcrSTD (normalize RF to a standard CLcr of 100
ml/min [CLcrSTD] in a 70 kg human)
The calculation of non-renal clearance (CLnr) and renal clearance
(CLr) is based on a population estimate for a standard human (70kg).
Then the size adjusted total CL (CL) is predicted from the sum of
these size standardardized values using an identical allometric scale
factor for both renal and non-renal clearance.
CL = (CLnrSTD + CLrSTD*RF)*(WT/70)**0.75
Nick
--
Nick Holford, Dept Pharmacology & Clinical Pharmacology
University of Auckland, 85 Park Rd, Private Bag 92019, Auckland, New
Zealand
email:n.holford.-a-.auckland.ac.nz tel:+64(9)373-7599x86730 fax:373-7556
http://www.health.auckland.ac.nz/pharmacology/staff/nholford/
Back to the Top
FYI, It is important to realize that renal dose adjustments in the
clinical setting should be made based on the FDA-approved drug label
and PK studies conducted during drug development. The most common
approach is an algorithm based on the Cockcroft-Gault estimation of
creatinine clearance, which is usually reported in mL/min and rarely
mL/min/1.73m2 units. THe MDRD equation provides an estimate of GFR
(eGFR), which was developed and valided based on iothalamate
clearance in non-diabetic CKD patients, and its use for drug dosing
should be avoided until further studies evaluating relationship
between MDRD eGFR and drug PK are conducted.
Tom
Back to the Top
The following message was posted to: PharmPK
Dear Nick,
Thank you for your response. I'm not an expert in renal function, and I
don't know if using RF instead of 'ml/min/1.73m** for renal function is
better. I leave that to the clinicians. I hope you don't mind that I
focus
on dosing, since that's what I'm interested in.
> Notice particularly that the RF value does not require knowledge of
> the patient's actual weight. The RF is normalized relative to a
> person of a standard weight (70 kg in my example). The RF value will
> be 1 for a patient who weighs 35 kg and has an actual CLcr of 50 ml/
> min or a patient who weighs 140 kg with an actual CLcr of 200 ml/min.
I don't understand this. What is your exact definition of RF? Rowland
&Tozer
(3rd Ed) define RF as the creatinine clearance of a patient divided
by the
creatinine clearance of the typical (standard) patient (55 y, 70 kg).
How
can you obtain the creatinine clearance without the patient's body
weight
(or body surface area in other formulas)? In the three patients the
RF = 1,
and their common property seems to be that the ratio of creatinine
clearance
and BW is 100/70 = 50/35 = 200/140 ml/min/kg. Is this what you mean?
This
implies that your definition is different from that of Rowland&Tozer.
> The calculation of non-renal clearance (CLnr) and renal clearance
> (CLr) is based on a population estimate for a standard human (70kg).
> Then the size adjusted total CL (CL) is predicted from the sum of
> these size standardardized values using an identical allometric scale
> factor for both renal and non-renal clearance.
>
> CL = (CLnrSTD + CLrSTD*RF)*(WT/70)**0.75
I do not see the rationale of the allometric conversion of the renal
clearance if one knows the creatinine clearance of the patient. I would
propose to use the following equation:
CL = (CLnrSTD)*(WT/70)**0.75 + CLrSTD * CLcr / CLcrSTD
where CLcr and CLcrSTD is the creatinine clearance of the patient and
the
standard patient, respectively, both in ml/min. In this equation the
first
term at the right side gives the allometrically scaled nonrenal
clearance
(similar in your approach). The second term converts the renal
clearance of
the standard patient to that of the patient, assuming that the renal
clearance of the drug is proportional to the creatinine clearance (in
ml/min).
Best regards,
Hans Proost
Johannes H. Proost
Dept. of Pharmacokinetics and Drug Delivery
University Centre for Pharmacy
Antonius Deusinglaan 1
9713 AV Groningen, The Netherlands
tel. 31-50 363 3292
fax 31-50 363 3247
Email: j.h.proost.-at-.rug.nl
Back to the Top
The following message was posted to: PharmPK
Hans,
See comments below...
Hans Proost wrote:
> Nick Holford wrote:
> > Notice particularly that the RF value does not require
knowledge of
> > the patient's actual weight. The RF is normalized relative to a
> > person of a standard weight (70 kg in my example). The RF value
will
> > be 1 for a patient who weighs 35 kg and has an actual CLcr of
50 ml/
> > min or a patient who weighs 140 kg with an actual CLcr of 200
ml/min.
>
> I don't understand this. What is your exact definition of RF? Rowland
> &Tozer (3rd Ed) define RF as the creatinine clearance of a patient
divided
> by the creatinine clearance of the typical (standard) patient (55
y, 70 kg).
I define RF as predicted CLcr for an individual based on age and
serum creatinine (140-age)/Scr*70/72 i.e. as if they weighed 70 kg)
divided by a standard value for a 70 kg person (e.g. 100 ml/min/
70kg). Rowland and Tozer define a RF relative to someone else rather
than a patient of similar size. Its seems a bit silly to say that a
child with perfectly normal kidneys has RF less than a normal adult
simply because they are smaller. I prefer to use the concept of RF as
a reflection of renal function not something confounded with body size.
> How can you obtain the creatinine clearance without the patient's
body
> weight (or body surface area in other formulas)?
See above for how to calculate a standardized creatinine clearance.
The idea is no different from the traditional but theoretically
invalid method of standardising on predicted BSA.
> In the three patients the RF = 1, and their common property seems
to be that the ratio of creatinine
> clearance and BW is 100/70 = 50/35 = 200/140 ml/min/kg. Is this
what you mean?
> This implies that your definition is different from that of
Rowland&Tozer.
You are correct. I dont agree with their size dependent definition of
RF. I choose to calculate the creatinine clearance for a patient of a
given age and serum creatinine (and sex) as if they weighed 70 kg.
Scaling the estimate for weight is then done in a second stage using
allometric concepts (see below).
> > The calculation of non-renal clearance (CLnr) and renal clearance
> > (CLr) is based on a population estimate for a standard human
(70kg).
> > Then the size adjusted total CL (CL) is predicted from the sum of
> > these size standardardized values using an identical allometric
scale
> > factor for both renal and non-renal clearance.
> >
> > CL = (CLnrSTD + CLrSTD*RF)*(WT/70)**0.75
>
> I do not see the rationale of the allometric conversion of the renal
> clearance if one knows the creatinine clearance of the patient. I
would
> propose to use the following equation:
> CL = (CLnrSTD)*(WT/70)**0.75 + CLrSTD * CLcr / CLcrSTD
>
> where CLcr and CLcrSTD is the creatinine clearance of the patient and
> the standard patient, respectively, both in ml/min. In this
equation the
> first term at the right side gives the allometrically scaled nonrenal
> clearance (similar in your approach). The second term converts
the renal
> clearance of the standard patient to that of the patient, assuming
that the renal
> clearance of the drug is proportional to the creatinine clearance (in
> ml/min).
>
When you say you 'know' the creatinine clearance you mean you are
making a guess for people of different sizes based on an assumption
that creatinine production is linearly proportional to weight. It is
important to appreciate that the usual formulae for predicting
creatinine clearance are in fact formulae for predicting creatinine
production rate (CPR). The measured serum creatinine is then used
with this prediction to obtain an estimate of creatinine clearance.
It is my guess that CPR, just like many other functional properties
of the body, might be expected to scale non-linearly with weight e.g.
like clearance to the 3/4 power of weight.
I liek to think about what determines CPR without making the
assumption that CPR is linearly proportional to weight. This can be
done by predicting the CPR based on age for the men studied by
Cockcroft and Gault (they point out almost parenthetically that the
average weight of the group they studied was around 70 kg). So I use
their predictions for an age adjusted value for a standard 70 kg
person. They I put my faith in allometry to predict CPR in people of
different weights.
The differences in our approaches comes from what assumptions we
prefer to make about how CPR scales with weight. I have to admit to
not being able to cite any direct evidence for my opinion that CPR
will scale as 3/4 power or weight. If you can offer empirical
evidence that your linear assumption has been tested against the 3/4
power assumption and shown to be better then I would be quite happy
to change my mind.
Nick
--
Nick Holford, Dept Pharmacology & Clinical Pharmacology
University of Auckland, 85 Park Rd, Private Bag 92019, Auckland, New
Zealand
email:n.holford.aaa.auckland.ac.nz tel:+64(9)373-7599x86730 fax:373-7556
http://www.health.auckland.ac.nz/pharmacology/staff/nholford/
Back to the Top
The following message was posted to: PharmPK
Dear Nick,
Thank you for your thoughtful comments.
> I define RF as predicted CLcr for an individual based on age and
> serum creatinine (140-age)/Scr*70/72 i.e. as if they weighed 70 kg)
> divided by a standard value for a 70 kg person (e.g. 100 ml/min/
> 70kg).
I'm not sure that I appreciate this approach (I try to formulate
carefully
since I know you have good reasons). The phrase 'as if they weighed
70 kg'
sounds strange to me. This implies a normalization to 70 kg, but
later in
your message you argue that an individualization (the reversed
process of
normalization) from a standard weight of 70 kg should not be based on a
simple kg-conversion.
Please note that for children the Cockcroft&Gault equation should not be
used. Other formulae proposed for children, e.g the Schwartz equation
CLcr 0.55 * height(cm) / Scr, does not allow this conversion.
The Cockcroft&Gault equation is also likely to overestimate the renal
function of obese patients, since it is unlikely that the creatinine
production rate is proportional to actual body weight. It has been
proposed
that Ideal Body Weight or Lean Body Mass, as estimated from several
formulae
(Devine, Chennavasin, James), may give better estimates of creatinine
clearance in obese patients.
> Rowland and Tozer define a RF relative to someone else rather
> than a patient of similar size. Its seems a bit silly to say that a
> child with perfectly normal kidneys has RF less than a normal adult
> simply because they are smaller. I prefer to use the concept of RF as
> a reflection of renal function not something confounded with body
size.
OK, but using the same term for different things is confusing.
> When you say you 'know' the creatinine clearance you mean you are
> making a guess for people of different sizes based on an assumption
> that creatinine production is linearly proportional to weight. It is
> important to appreciate that the usual formulae for predicting
> creatinine clearance are in fact formulae for predicting creatinine
> production rate (CPR). The measured serum creatinine is then used
> with this prediction to obtain an estimate of creatinine clearance.
> It is my guess that CPR, just like many other functional properties
> of the body, might be expected to scale non-linearly with weight e.g.
> like clearance to the 3/4 power of weight.
OK, this sounds good. In fact, you suggest that the Cockcroft&Gault
equation
is not appropriate. Why not changing the Cockcroft&Gault equation
with your
suggested 3/4 power of weight, and then applying the dosing equation
proposed in my earlier message. This would keep the dosing approach
straightforward.
Best regards,
Hans Proost
Johannes H. Proost
Dept. of Pharmacokinetics and Drug Delivery
University Centre for Pharmacy
Antonius Deusinglaan 1
9713 AV Groningen, The Netherlands
tel. 31-50 363 3292
fax 31-50 363 3247
Email: j.h.proost.aaa.rug.nl
Back to the Top
The following message was posted to: PharmPK
Hans,
Thanks for your thoughts. I know this can be a tricky area to
appreciate because algebraically many of the expressions appear to be
the same. But for me its like the preference for expressing PK in
clearance and volume rather than rate constants.
First of all 'normalization' does seem similar to 'as if they weighed
70 kg'. If you look carefully at the Cockcroft and Gault paper you
will find they report creatinine production rate as a function of
age. Their parameters are not scaled for weight. They note that the
average weight was around 70 kg. So the C&G formula is primarily
predicting an age adjusted CPR for a population with an average
weight of 70 kg.
If you accept this then it seems reasonable to use their parameters
to predict the CPR for someone 'as if they weighed 70 kg'. After
doing that one can choose to scale to another weight using some other
method. I prefer allometry because of its very strong experimental
and theoretical basis compared with the naive per kg scaling approaches.
I know it is traditional to say that one should not use the C&G
equation for children. But that does not mean that one should not
consider using the principle of the C&G method i.e. predict an age
specific CPR then scale for body size. I have been doing this in
collaboration with Brian Anderson and we have developed estimates of
how CPR appears to increase with age in children (as it should
because they are growing and muscle bulk increases with age). We use
allometry of course to adjust for differences in body size.
Finally, I agree its important to recognize that abnormal body
composition requires an additional adjustment in order to get a
suitable weight for allometric purposes e.g. predicted normal weight
(Duffull SB, Dooley MJ, Green B, Poole SG, Kirkpatrick CM. A standard
weight descriptor for dose adjustment in the obese patient. Clin
Pharmacokinet. 2004;43(15):1167-78.)
Note that the approach I prefer is to separate and treat each factor
independently as far as possible i.e. size independent of abnormal
composition, renal function independent of size and then try to deal
with age related phenomena after accounting for size and renal
function. I think that if one properly understands how each of these
elements is formed then they can be combined more appropriately for
application to such things as dose prediction.
Nick
--
Nick Holford, Dept Pharmacology & Clinical Pharmacology
University of Auckland, 85 Park Rd, Private Bag 92019, Auckland, New
Zealand
email:n.holford.aaa.auckland.ac.nz tel:+64(9)373-7599x86730 fax:373-7556
http://www.health.auckland.ac.nz/pharmacology/staff/nholford/
Back to the Top
The following message was posted to: PharmPK
Dear Nick,
Thank you for your explanations. I fully agree with your approach to
separate and treat each factor independently. Actually this is also the
guiding principle in the approach proposed in my first message.
However, you
correctly pointed to the fact that the Cockcroft&Gault equation (and
probably all the others as well) are not based on sound allometric
principles with respect to the estimate of creatinine production rate
(CPR).
As stated in my previous message, I would say that it is more logical to
modify the Cockcroft&Gault equation, but that implies 'revolution' in
the
field of nephrology. It is likely that you choose the 'peaceful' way
to take
this into account in the dose adjustment, and not in the calculation of
creatinine clearance.
Best regards,
Hans Proost
Johannes H. Proost
Dept. of Pharmacokinetics and Drug Delivery
University Centre for Pharmacy
Antonius Deusinglaan 1
9713 AV Groningen, The Netherlands
tel. 31-50 363 3292
fax 31-50 363 3247
Email: j.h.proost.-at-.rug.nl
PharmPK Discussion List Archive Index page
Copyright 1995-2010 David W. A. Bourne (david@boomer.org)