Hi,Back to the Top
I'm wondering what are the characteristics one looks for in picking an
intenal standard for quantitation of a novel drug by LC-MS (other than
unique retention time)? In the literature, people seem to pick a drug
related to their compound of interest.
Thanks,
Noelle Williams
Noelle,Back to the Top
Actually, unlike with other chromatographic techniques,
your LC-MS internal standard doesn't have to possess a
different retention time compared to your analyte (and if
you use a heavy isotope internal standard you are actually
very unlikely to see separation). This doesn't matter as
the MS does all the hard work (I'll leave out cases where
there could be ion suppression and other artifacts caused
by co-eluting compounds). I try and find an internal standard
that behaves in a generally similar manner to the analyte.
For the MS side of things I hope for a similar MW (so it is in
the same general mass range as the analyte), has similar
analysis conditions (same ionisation mode). For the LC I want
something that will elute under the same conditions as the
analyte (e.g. avoid highly polar compounds that might get lost
in the solvent front). For the extraction conditions I want
something that will extract reporoducibly with the analyte
(and is chemically stable, photostable etc.). The easiest way
to achieve all of these criteria (or a good starting point) is
a a structural analogue of the analyte of interest. Suitable
behaviour during extraction is almost always the most difficult
target to hit.
I hope that this helps.
All the very best,
Bernard
Bernard Murray, Ph.D.
Senior Research Investigator
Drug Metabolism, PCS, PPD, GPRD, Abbott Laboratories, Chicago, USA
Bernard.Murray.-at-.abbott.com
Hi Noelle:Back to the Top
Generally one would go for a stable isotope (H-2 or C-13) of the
compound of interest to use as an IS for LC-MS analysis. Hope this
helps.
Parnali Chatterjee
Dear Noelle,Back to the Top
Adding to what Bernard has said, the requirements of an internal
standard
for use in LC-MS method depends on your LC part of it. If you have an LC
method that separates the internal standard from test compound, you can
very
well go ahead and select an internal standard much the same way you do
for
LC method (i.e.., similar extraction properties, stable etc). Ofcourse,
it
should ionise in the MS conditions you use for your test compound. But,
if
your LC method does not separate your test compound and internal
standard
(as in most cases) an additional criteria is that your internal standard
should have a different mass than that of your test compound.
More care should be exercised when analysing PK or metabolism samples
using
LC-MS method where you can have metabolites. In such situations the
possibilities are that:
1. metabolite(s) may have same mass as that of the parent compound
2. metabolite(s) may have same mass as that of the internal standard or
3. both
In such situations you can still use your existing internal standard
and MS
conditions if you can achieve their separation by altering the LC
method.
However, with LC-MS-MS you can use internal standard with same mass as
that
of parent compound if they have different fragmentation patterns and
minimum
criteria in this case would be that of your LC part.
Hope this helps.
Kasiram.
Victorian College of Pharmacy
Monash University
Deuterated analogs are the best usually (they have almost the sameBack to the Top
retention
time on most LC systems) but aren't usually used until the clinical
stage
because of cost and there are good alternatives. Analogs work very
well.
If you can get something with one or two more or less halogens (F or
Cl),
that works about as well as a deuterated compound in my experience.
You can
use something that is different by some alkyl moiety (e.g. perhaps
there is
a dimethyl phenyl analog vs. phenyl). We always try to be mindful of
the
choice being potentially the same MW as a metabolite (e.g. a Cl for F
substitution is the same as a mono-oxygenated metabolite). Having said
that, even if it has the same MW, the fragment ion most likely won't be
the
same so it is OK if you have MS/MS and chromatography reduces the odds
of
problems because of separation in time. If the halogen modification is
further away from ionizable groups, this is helpful as the electron
withdrawing elements may affect the chromatography retention time
relative
to parent. For LC/MS, generally you like to have the retention time
be
close if possible to parent. Helps minimize problems in variable
response
with time of the MS detector. One example that is a hot topic these
past
years is if you have ion suppression problems which happen with
formulations
high in PEG since many compounds are so lipophilic. There are a series
of
polymer peaks that may coelute with your peak of interest (but you'll
never
see in the MRM channel) and since formulation concentrations are so much
larger than drug or IS, your MS signal may drop in study samples and
not be
mimicked by your standards or QCs as there is competition for protons
in the
ion source. You can look at your IS response for study samples vs.
standards and QCs and if there is a large drop in study samples, the
formulation may be causing problems. That's why it is useful to have
the RT
be the same as the analyte as you are covering it also for ion
suppression
problems. The same suppression problems can occur from endogenous
plasma
components and you can do the infusion experiment to check for it as
described by R. King a few years back. Sometimes it is easier to
develop a
L/L or SPE cleanup to eliminate the PEG problems.
Dear Noelle,Back to the Top
Personally, I don't use internal standard in LC/MS. I am working in the
MS-MS mode so that in general, the sample preparation is very simple. I
also
use samples / QC / unknown samples organized in bracketed sequence. This
enables to develop a method father and avoid any problem with different
matrix effects on the internal standard and on the drug of interest.
Internal standard mode is promoted by bio analysts who were working with
LC/UV detection and complicated sample preparation. Generally, it is not
recommended by Mass Spec experts. I use to work with M-Scan, a company
with
very strong expertise in MS and LC/MS-MS and located in the UK,
Switzerland,
Germany, US, Korea: they work the same way as I.
Best regards
Bernard Ludwig, PhD, CChem M.R.S.C.
ADDEX Pharmaceuticals Ltd
+41 22 884 15 82 (direct)
+41 22 884 15 55
bernard.ludwig.-a-.addexpharma.com
Dear Noelle,Back to the Top
As has already been mentioned here, internal standard in lc/ms, apart
from its usual function, can and should be used to deal with the ion
suppression problems. In this sense, the choice of the IS will be
dictated by the ability to force its co-elution with the analyte. The
good news is that it will normally shorten your run time.
However, one cannot assume that the extent of ion suppression for the
IS will be the same as for the analyte simply because they co-elute -
you have to verify that there is no over- or undercompensation.
Andrew Volosov
UCB Research
I agree with the contributors who note that there are more requirementsBack to the Top
for a good chromatographic internal standard that is to be used with
mass spec detection/quantification than for one that is to be used with
uv detection/quantification, and that a stable isotope analog is usually
superior to a structural analog. We recently encountered a case in
which a structural analog internal standard that matched the analyte
chromatographic properties was still marginal because it did not
completely match its mass spec properties. (We'll be presenting this
example as a poster at the Eighth International Symposium on Hyphenated
Techniques in Chromatography and Hyphenated Chromatographic Analyzers,
Bruges Belgium, in early February: Hua-Fen Liu, Cheng-I Mau, Thomas
Tarnowski; "Stable Isotope Analog vs Related Structural Analog as an
LC/MS/MS Internal Standard: Differences in Instrument Parameters That
Affect Compound Behavior and Assay Performance")
It is important to have an adequate internal standard for methods in
support of clinical studies, but also in support of GLP tox studies that
precede and enable clinical studies.
Note that a stable isotope internal standard needs to differ from its
analyte by several mass units (e.g., a monodeuterated version is almost
never good enough).
Thomas L. Tarnowski, Ph. D.
Project Team Leader
Dept. Head (Dep), Drug Metabolism and Pharmacokinetics
Roche Palo Alto
Back to the Top
I can't agree with you,
The only way to avoid/minimize the "matrix effects" is to improve your
extraction method. The use of stable labelled internal standard (2H,
13C)
helps you to minimize this problem, but if you don't use IS when you
quantitate from biological samples, your results may be subject to large
errors.
These errors may come not only from the extraction step but also from
the
analytical system (volume injected, matrix effects........). The fact
that
you use QC samples does not assure that your results are correct
because the
matrix of the unknown samples will never be the same than yours.
On the other hand, the shorter the chromatographic time, the higher the
posibilities of suffering matrix effects. The use of tandem MS does not
assure the absence of these effects because they take place in the ion
source during the ion evaporation step and not in the analyzer.
I have never heard any MS expert claiming not to use internal standards
in
quantitative LC-MS or LC-MS/MS; even the new achievements in LC-MALDI
for
quantitation purposes, must be carried out with an internal standard to
correct for differences in absolute ion signal due to the presence of
the
matrices (the MALDI "matrix" and the biological "matrix") and/or
ionization
efficiency due to the laser shot impact area in the MALDI spot. The
first
time I heard this was in a lecture by Prof. Karas (one of the
developers of
MALDI-MS).
So if you want to have a robust and reliable method (from my point of
view)
you need a clean extract, a suitable chromatographic separation and a
good
internal standard (stable labelled if possible).
Hope this helps
Jos\0xC8 Antonio Allu\0xC8
Mass Spectrometry Laboratory
Metabolism & Pharmacokinetics Service
Research & Development Department
IPSEN-PHARMA S.A. Laboratories
Ctra. Laure\0x2021 Mir\0xDB 395
Sant Feliu de Llobregat, Barcelona, Spain
Telf.: 936858100
Back to the Top
Dear all,
I do not agree with the statement that an internal standard should be
used
in LC/MS-MS in order to generate more reliable data. We and our
colleagues
from M-Scan (who are expert in mass spec) practically never use internal
standard. We prefer to utilize bracketed standard/QC/unknown samples
sequences and work in the external mod. Especially when using such
techniques as column switching that minimize the sample preparation
step.
Also be aware that an internal standard only improves the accuracy and
precision of the analysis when the CV of the peak measurement of the
i.std.
is better than the CV of the peak area or height of the product to be
assayed. Otherwise, the precision is worth when using internal standard
calibration than when using external standard calibration and carefully
manipulation at the bunch. There is a paper of Haeffelfinger et al. I
believe in J. Chromatography sometime in the early 80's, which is
dealing
with this matter. I think most analysts use internal standard in order
not
to be worried with careful pipeting etc. Especially in CRO's that often
exhibit a high turn over of their technicians and scientist, which makes
difficult to have true senior analysts (i.e. with at least 10 years
experience in bio analytics at the bunch)
Best wishes
Bernard Ludwig, PhD, CChem M.R.S.C.
ADDEX Pharmaceuticals Ltd
+41 22 884 15 82 (direct)
+41 22 884 15 55
bernard.ludwig.aaa.addexpharma.com
Back to the Top
Dear M. Allu\0xC8,
I cannot agree with your view.
Matrix effect is something we studied (as request by the US FDA in its
Guideline dated May 2001) in each LC-MS-MS method we are developing.
When
using an i.std. you have to study the matrix effect for the drug of
concern
and the i.std.
Why should the matrix effect be different form unknown to QC samples
when
used at the same concentrations and in the same matrix (e.g. pool blank
plasma)? Why should we accept that an i.std. could eliminate the matrix
effect? Making the same error or un precision two times will not
increase
the precision / accuracy of the final result.
Matrix effect is not alchemy. This is due to the large quantity of
products
entering the MS source at the same time. The matrix effect is related
to the
instrument you use. It is also varying with the product concentration.
Thus,
it is generally difficult to generate calibration lines on a large
range of
concentration. It is better to have short range calibration and dilute
the
unknown / QC.
We are using fast HPLC connected to MS-MS (LC run < 2.5 min) and ultra
fast
LC (10 seconds) with column switching. This enables to decrease the
matrix
effect because a lower quantity of the plasma or urine endogeneous
products
are entering into the ionization source. Under these conditions, we
already
observed than adding an i.std. as a tri-D compound did not improve nor
the
accuracy or the precision of the method.
Best wishes,
Bernard Ludwig, PhD, CChem M.R.S.C.
ADDEX Pharmaceuticals Ltd
+41 22 884 15 82 (direct)
+41 22 884 15 55
bernard.ludwig.-a-.addexpharma.com
Back to the Top
Dear all
In deed it was great discussion from all of you whether to use Internal
standard or not for quantification purposes. Many views expressed by
the experts were quite interesting and laudable. Further to this i
would like to add few comments on the topic.
1.After all the data being generated will be used for further decision
making. So the accuracy needs to be ensured at the outset for all the
experimentation.
and
2. Regulatory acceptance needs to be ensured.
If the two criteria are met one through with the job.
There are two issues, which everybody has pointed out. One is
extraction accuracy for the biological samples. And other is
instrunmentation accuracy.
The first step we do by diffrent extraction methods. say LL, SPE or PP
methods. All these methods give us good extraction. But the accuracy
and precision will be maitained by using proper internal standard.Since
there may be many number of step involved, many scientists do process
samples, error creeps in every step of sample preparaion. A thorough
IS will minimize the error but not truly eliminate it.The IS can be a
stable Isotope or an anlogue.Procesing samples without IS will always
give us the different values, if you repeat the same batch. This is not
the case with the IS added sample. THe repetition will give us the
accepatble values. So in my point I would like to go for IS maintaing
accuracy in each step which will ensure the final value.
Second point of my comment is Instrumentation accuracy. As per the
literature and also different views expressed in this forum i would
like add more to this.The assumption is LCMSMS instruments are more
sensitive and minimal sample preparation is required.This is always not
true. We need to have a cleaner sample for better accuracy and we need
to work for sensitivity.
Normally the instrunment response and repeatblity will be checked by
injecting suitable standards.Then processed samples will be injected.
As per as the fact sheet goes instrunment response for a particular
compound will vary over a period of time.Also the response may vary for
the analyte.How to ensure the instrument response stable over a period
of time? Some times we need to inject as many as 400 samples in
batch.This is not under our control. Under controlled environmental
conditions also the change in the response was observed. If you have
the IS which is analogue or stable isotope will maximize the
accuracy..In this situations ading IS may be better altenative to
control instrumentation accuracy. THe Quality control samples will give
additional validation of the batch run.Here again two things will creep
in whether to use or not to use the IS. or to use Stable isotope or
anlogue.
As per the technical aspects goes ionization efficency and trasmission
of ions with same velocity and its fragmentation will vary according to
the conditions of MS and chromatography.Having a IS with same
ionization charcteristics and chromatography will minimize the
variablity in response and ensure the stability of signal over a period
of time. Taking about matrix effects we need to ensure during method
development. Changing to different source some times avoid the
problem.Also slight modification of chromatographic conditions do
ensure elimination matrix effects. After all, however the clean sample
is, possibility of coeluting compounds do present which will interfere
with the analysis.
any suggestion and comments on the topic are welcome.
Regards
Akula T. Bapuji Ph.D
Group Leader-Bioanalytical
Ranabxy Research labs
Gurgaon.
Back to the Top
Dear Bernard,
I have never said that the use of an IS would eliminate the possible
matrix
effects, and of course the absence of these effects must also be
checked for
the IS as well. On the other hand, the matrix you use for your QC and
calibration samples (or the one you use to check for matrix effects), in
most of the cases will never be the same than that of your real samples
(not
all the rat plasma pools are identical...............). So this is why I
think that the use of an stable labelled IS is a main advantage if it is
available for you. The use of IS is not due to correct for incorrect
pipeting or sample handling; it is much more than that biassed point of
view.
As you said, these effects come from the large quantity of compounds
entering the ion source, so once again, the cleaner the extract the
lower
the matrix effects.
In our lab we also have a large experience in quantitative LC-MS; from
magnetic sectors to quadrupoles, from electron impact ionisation to
electrospray, FAB, chemical ionization, APCI, Particle Beam, etc. We
also
use fast LC in some situations, with fast sample preparation methods (on
line SPE), and our experience tells us that most of the times the
combination of two sample preparation techniques (i.e. protein
precipitation
+ on line SPE, off line + on line SPE......) drastically reduces the
matrix
effects, and that the use of and IS is a very helpful tool.
Best wishes.
Jos\0xC8 Antonio Allu\0xC8
Mass Spectrometry Laboratory
Metabolism & Pharmacokinetics Service
Research & Development Department
IPSEN-PHARMA S.A. Laboratories
Ctra. Laure\0x2021 Mir\0xDB 395
Sant Feliu de Llobregat, Barcelona, Spain
Telf.: 936858100
Back to the Top
To answer to the discussion and arguments raised by doctor Ludwig, I
thing that it is necessary to explain why internal standards are
used...
If you use internal standard to increase the reliability of your datas
after one or multiple sample preparations steps such as extractions or
precipitation, I think that IS use is really necessary, because you
will obtain different extractions rates between the matrix tested and
the pooled plasma. In everyday clinical practice and therapeutic
monitoring, the skilled analysts know that the matrix effect will be
specific of the patient state, what he has eaten, the other drugs
plasma concentration at the sampling time etc... So in that case you
cannot work without IS because pooled plasmas will never mimic the
interaction between your patient's plasma and the drug your are trying
to quantitate. The IS choice has to be as close as the drug you are
trying to analyte. I do agree with others in this forum, claiming for
stable isotopes as IS.
If you have no preparation steps (and I do think that is Dr Ludwig's
case according to the informations he provided), IS use is less
necessary because the matrix effect can effectively been greatly
reduced by dilution (if you don't have to determine very low
concentrations of your drug)
In conclusion the approach is different according to the analysis
purposes and methodology... as it is always in analytical chemistry.
I hope these precision will help
F. Lagarce
Dear Bernard,Back to the Top
I feel a bit more detailed clarification is in order here.
First and foremost, answering your question why matrix effect may be
different in unknown samples and in QCs or standards prepared from
blank plasma. The answer is very simple - because the unknown samples
were not "prepared" from blank plasma! These unknown samples may (and
more often than not will) be very different, because they contain
excipients from the formulation (PEG-400, Tween, etc.). Furthermore, in
the PK studies, the unknown samples will not only be different from the
QCs/standards, but also from each other, because the excipients undergo
elimination just like the drug of interest. Very often we see in our
studies that samples from early sampling times (especially IV samples)
produce huge matrix suppression, while samples taken at later times
give almost no suppression at all.
Secondly, as I have already mentioned in my previous reply, we do not
assume that the internal standard eliminates matrix effect - rather, it
can be used as a convenient tool to compensate for suppression. Do not
forget that strictly speaking, you do not know whether you have ion
suppression until and unless you actually measure it. Once you have
measured it, you can decide what is the best way to deal with it.
And finally, in this case two wrongs do make a right - in other words,
internal standard suppressed to the same extent as the analyte will
give a correct result.
Best regards,
Andrew Volosov
UCB Research
Back to the Top
Dear Andrew,
Thank you for your valuable and interesting comments
I am not so convince by the excipient issue. I rather like to generate
data
and make comparison. In a study where I assayed down to 2 - 3pg/mL, I am
simply not convinced on the need of i.std. Simply by calculating the
same
analytical data batch in the i.std. and external mod! In this later
case,
pharmacokinetisists and my colleagues preferred to stay with the i.std.
because it is generally well accepted that it is more accurate and
precise.
The facts did not confirmed this (in this later case). I have never seen
reports clearly demonstrating the needs of i.std. excepted as I wrote
previously this publication of Haefelfinger et al. in appro. 1984. I
rather
has the impression that people use to utilize an i.std. and then they
don't
question themselves anymore. Equipment has changed during these last
year.
Sciex 3000 for example, suffered from memory effect. This was improved
by re
designing the apparatus (Prof Hopfgartner, Uni Geneva). The last
LC/MS-MS
from Finigan (Quantum) exhibits very low memory effect. Thus, using
external
calculation in these two might well lead to results as good as with
i.st.d
when not better. Because memory effect could also play a role in this
story
too.
You should be aware that some companies have not the money to pay for
manufacturing isotope compound. Sometime, I wonder whether making
analytical
things complicated is not a good way for the pharm industry to protect
themselves from generics...
I would appreciate to discuss this further with you and other. I am
trying
to plan some experiments, I hope with my colleagues from M-Scan, in
order
to bring experimental results from various sources in the debate.
Best wishes.
Bernard Ludwig, PhD, CChem M.R.S.C.
ADDEX Pharmaceuticals Ltd
+41 22 884 15 82 (direct)
+41 22 884 15 55
bernard.ludwig.at.addexpharma.com
Back to the Top
Dear Bernard,
thanks at lot, you speak from my heart.
In our experience, an i.st. in some instances really make things worse.
This conclusion is not new, but many people just spread a rumour
instead of following common sense, early references and - yes -
experimental results.
One of the main points in method validation is 'suitability for the
intended use'; so if we are able to develop and validate an analytical
method without an internal standard, fine.
two references:
P. HAEFELFINGER;
Limits of the Internal Standard Technique in Chromatography
J. Chromatogr. 218, 73-81 (1981)
KARNES, T.H., SHIU, G. and V.P. SHAH;
Validation of Bioanalytical Methods
Pharm. Res. 8, 4, 421-426 (1991)
(P.S. this was *after* the 1990 Washington Conference on Validation!)
Regards
Helmut
Helmut Sch\0xB8tz Biokinet GmbH / Dept Biostatistics
Neubaugasse 36/11 Nattergasse 4
A-1070 Vienna/Austria A-1170 Vienna/Austria
tel/fax +43 1 9713935 tel +43 1 4856969 62
no cell phone ;-) fax +43 1 4856969 90
Dear Helmut,Back to the Top
You said the final word. I think people seem to worry too much about a
simple issue. Internal standard is in deed a good tool, when the
problem is
complicated. It is best to live without one if it is Ok with your
problem in
hand.
Vinayak Nadiger
Back to the Top
thank you very much Helmut
We agrre with you. Actually, I proposed to some of those writing in this
forum to participate to a common study aiming to check the validity of
the
i.std. in LC/MS-MS with several type of drug substance. Nobody accepted
it
up to now. With M-Scan 5Dr Luc Alain Savoy, Geneva\0x221E we are discussiong
to
start such a study . We intend to publsih the results. If you agree you
could join us. In any case, we shall keep you inform with the results.
The puiblication of Haefefinger, I know it as I was working with Roche
Basel
for 18 years. The one of Karnes et al. I don't know it. I will look for
it
and read it carefully for sure.
Best regards
PharmPK Discussion List Archive Index page
Copyright 1995-2010 David W. A. Bourne (david@boomer.org)