|
|
|
QUESTIONS
& ANSWERS |
| If you have any questions
regarding method validation, instrument performance
and calibration, method development and pharmaceutical
analysis, please forward them to cvg@cvg.ca. We will
try to answer your questions or direct your questions
to the experts who may be able to answer them. |
|
 |
I am looking for information regarding the theoretical values of caffeine UV maxima and minimum. I am trying to incorporate those values in my HPLC VW Detector SOP and I like to have something official about those values. I have looked into Merck Index, Handbook of Chemistry and Physics but none of those mention those values. Several vendors have provided information but I got values that differ somehow from each other. For example, one vendor uses 272 nm and another 273 nm, For the minimum, one vendor uses 244 nm another uses 245 nm. |
 |
The lamda maximum and minimum of a
caffeine solution depend on the solvent used and the spectral bandwidth the spectrophotometer
or detector. Typically, the spectral bandwidth of
UV detectors is quite broad. For a caffeine solution
in water measured using a spectral bandwidth of 6
nm, the lamda maximum occurs at 273 and 205 nm and
minimum occurs at 244 nm. These values are taken using
a calibrated high resolution spectrophotometer. I
hope this help. |
 |
|
I would like to validate
two incubators (25
oC/60% RH and 40
oC/75% RH). We
are a pharmaceutical plant, so we have to meet FDA
regulations. I would like to have these validations
performed by a specialized firm, since we are not
equipped to perform them. Can you help me with this
matter? |
|
The following companies do this type of work:
Validation
Technologies
Pharmaceutical
Technologies |
|
|
Can you please let me know, what we
mean by Quantity (Q)?
Where or how is it derived and
what are the advantages and disadvantages of using
Quantity (Q) in dissolution testing ? |
|
I am not sure about the origin of "Q-value"
in dissolution test. However, USP defines the criteria
for stage 1, 2 and 3 dissolution testing using Q-value,
therefore all US submissions (e.g., NDA, ANDA) to
FDA should have the dissolution specifications based
on Q-values. Stage 1,2 and 3 testings are only used
by USP and therefore they are the only one who use
Q-values. For submissions to other countries (e.g.,
EU), there is no need to use Q-values.
For example, the dissolution specifications for
the same drug product will be:
US submission: meet USP requirements, where Q=75%
in 30 minutes. Other submissions: not less than
80% dissolved in 30 minutes. They are the same specification
limit, as stage 1 USP testing requires all units
to be not less than Q+5% (i.e., 80%).
Keep in mind that Q=75% means all units will be
not less than 80% (not 75%) dissolved. |
|
|
We are about to transfer some validated analytical
methods to our own laboratory. Am I correct in thinking
ruggedness and precision are the only tests required
to certify us to use these methods. By ruggedness
I mean 2 analysts test 6 samples from each of 2
different manufacturing lots on 2 different instruments
on 2 different days?
Would the %RSD of the grand mean have to be less
than 2.0% (main active assay) Would this spec apply
to within and between laboratories. Is two different
manufacturing lots enough? Is this approach overkill
or not enough? |
|
I am assuming the method you are transferring
is the potency method. Based on the info you mentioned,
your proposal would be sufficient. The specification
should be applicable to both labs. The acceptance
criteria for the method transfer will be based on
the variability of your method and should be within
the normal acceptable variability.
However, for impurity methods, you will need to ensure
sensitivity of the instrument in the new lab vs. your
lab. |
|
|
When doing the precision test on the assay of
main actives, I made six sample preparations from
the same sample, I then calculated the % RSD of
six replicate results. The criteria is % RSD less
than 2.0%. How to perform this test on the tablet
dissolution, what is the criteria in this case?
For the ruggedness test (deviation from dissolution
method conditions), how many factors that I have
to take into consideration, and what are the accepted
% RSDs (compared to the unmodified conditions)?
If the dissolution assay is to be carried out
by a second analyst (on the same HPLC and dissolution
unit) what is the accepted % RSD? These questions
are raised because the individual tablet weights
are different, and the content uniformity test has
a high accepted % RSD (for example in USP, % RSD
is not more than 6.0%)
|
|
I will perform 6 replicate of tablet dissolution
and analyze the results (by HPLC or whatever you
choose). RSD should be less than 5%. This precision
includes the variation caused by dissolution and
by HPLC analysis.
I prefer to call it "robustness" instead
of "ruggedness". Technically speaking,
"ruggedness" is "day to day variation,
that means ICH intermediate precision". The
number of factors are up to you. In most cases,
I will use less than 6 factors. Some of the factors
that I will investigate are " degas method,
rpm, sinker type, dissolution medium composition).
In general, I expect to see the "mean"
dissolution results from all conditions to be within
5% of each other.
In order to avoid tablet-to-tablet variation to
affect your results, you can compare the "mean"
of one analyst versus the "mean" of another
analyst. The mean results should be less than 5%
of each other. In addition, you can use some "normalization"
technique to account for tablet to tablet variation.
I can't go into too much details here.
|
|
|
I was wondering if you could help
me to find information regarding the requirements
on robustness of an HPLC method. I looked in the ICH
guidelines and I found the document: "Validation
of analytical procedures: methodology". They
mention that robustness has to be evaluated but there
is no limit or requirement to meet. |
|
The following companies do this type of work:
Validation
Technologies &
Pharmaceutical
Technologies |
|
|
I would like to validate
two incubators (25 oC/60% RH and 40 oC/75% RH). We
are a pharmaceutical plant, so we have to meet FDA
regulations. I would like to have these validations
performed by a specialized firm, since we are not
equipped to perform them. Can you help me with this
matter? |
|
The following companies do this type of work:
Validation
Technologies &
Pharmaceutical
Technologies |
|
|
Can you please let me know, what we
mean by Quantity (Q)? Where or how is it derived and
what are the advantages and disadvantages of using
Quantity (Q) in dissolution testing ? |
|
I am not sure about the origin of "Q-value"
in dissolution test. However, USP defines the criteria
for stage 1, 2 and 3 dissolution testing using Q-value,
therefore all US submissions (e.g., NDA, ANDA) to
FDA should have the dissolution specifications based
on Q-values. Stage 1,2 and 3 testings are only used
by USP and therefore they are the only one who use
Q-values. For submissions to other countries (e.g.,
EU), there is no need to use Q-values.
For example, the dissolution specifications for
the same drug product will be:
US submission: meet USP requirements, where Q=75%
in 30 minutes. Other submissions: not less than
80% dissolved in 30 minutes. They are the same specification
limit, as stage 1 USP testing requires all units
to be not less than Q+5% (i.e., 80%).
Keep in mind that Q=75% means all units will be
not less than 80% (not 75%) dissolved. |
|
|
We are about to transfer some validated analytical
methods to our own laboratory. Am I correct in thinking
ruggedness and precision are the only tests required
to certify us to use these methods. By ruggedness
I mean 2 analysts test 6 samples from each of 2
different manufacturing lots on 2 different instruments
on 2 different days?
Would the %RSD of the grand mean have to be less
than 2.0% (main active assay) Would this spec apply
to within and between laboratories. Is two different
manufacturing lots enough? Is this approach overkill
or not enough? |
|
I am assuming the method you are transferring
is the potency method. Based on the info you mentioned,
your proposal would be sufficient. The specification
should be applicable to both labs. The acceptance
criteria for the method transfer will be based on
the variability of your method and should be within
the normal acceptable variability.
However, for impurity methods, you will need to
ensure sensitivity of the instrument in the new
lab vs. your lab. |
|
|
When doing the precision test on the assay of
main actives, I made six sample preparations from
the same sample, I then calculated the % RSD of
six replicate results. The criteria is % RSD less
than 2.0%. How to perform this test on the tablet
dissolution, what is the criteria in this case?
For the ruggedness test (deviation from dissolution
method conditions), how many factors that I have
to take into consideration, and what are the accepted
% RSDs (compared to the unmodified conditions)?
If the dissolution assay is to be carried out
by a second analyst (on the same HPLC and dissolution
unit) what is the accepted % RSD? These questions
are raised because the individual tablet weights
are different, and the content uniformity test has
a high accepted % RSD (for example in USP, % RSD
is not more than 6.0%) |
|
I am assuming the method you are transferring
is the potency method. Based on the info you mentioned,
your proposal would be sufficient. The specification
should be applicable to both labs. The acceptance
criteria for the method transfer will be based on
the variability of your method and should be within
the normal acceptable variability.
However, for impurity methods, you will need to
ensure sensitivity of the instrument in the new
lab vs. your lab. |
|
|
I was wondering if you could help
me to find information regarding the requirements
on robustness of an HPLC method. I looked in the ICH
guidelines and I found the document: "Validation
of analytical procedures: methodology". They
mention that robustness has to be evaluated but there
is no limit or requirement to meet. |
|
The current ICH guideline on method
validation (Q2B) only gives a brief outline on the
expectations for method robustness. In general, the
following suggestions may be considered.
Focus on critical parameters only: Since it is
impossible to investigate all parameters in the
method, I only focus on those critical parameters
(i.e., parameters that have significant impact on
the method performance). We can usually identify
the critical parameters during method development.
For example, if you know that % organic will cause
significant change in peak shape, this may be one
of the parameters in the method validation.
Variation range in method validation:
The range should reflect the expected day-to-day
variation, so the range should be quite tight. For
example, if % organic is one of the parameters,
I will suggest to use +/- 2% as the range as this
reflects the expected range when the procedure is
performed properly. I have seen method robustness
problems just because of using a too-wide variation
range in method validation.
Investigate sample preparation/ HPLC separation
independently
Method robustness should be investigated separately
for sample preparation vs. HPLC chromatography.
Sample preparation: look for effect of varying sample
preparation on the results (e.g., sample solvent
composition, sonication time....) HPLC conditions:
look for similar effects by varying HPLC conditions
(e.g., detection wavelength, mobile phase composition,
column temperature...)
Try to use experimental design to minimize work
Since the workload will increase exponentially
when you try to change multiple parameters at the
same time, I like use traditional experimental design
to minimize the workload (e.g., fractional factorial
design...)
Tie to method procedure
If you demonstrate that the method is very sensitive
to a particular parameter, then you need to specify
a very tight range in method procedure (e.g., %
organic should be between 48% to 52%). This range
will ensure the method performance is reproducible
on a day-to-day basis. |
|
|
How often should the calibration weights
be recalibrated? Our balances are used in laboratory
and processing situations. |
|
We calibrate the standard weights for every year,
in order to have control over their accuracy as
indicated by GMP's. And they also can be calibrated
under request if there is any suspicion of change
of weight or for any sign of deterioration. This
will be my recommendation since auditors were very
comfortable with this policy. |
|
|
Please share some experiences on any
lab glassware or dishwasher cleaning validation. The
topic seems to be one of the frequently asked questions
by the regulatory agencies. |
|
For glassware cleaning validation,
it may be useful to assume the worst case scenarios.
1. Select the worst case compound in the laboratory.
That is the compound with the lowest solubility
in most solvent types, and in most aqueous conditions.
2. Select the worst case concentrations. There
are two 2 different situations:
a. When someone prepares a high concentration of
this compound (e.g., for impurity assay), clean
the glassware in a normal manner. Then use the same
glassware to prepare a "blank" and detect
if any active present in the blank.
b. Use low volume pipettes (e.g., 2 mL, 5 mL) to
pipette the high concentration solutions. Clean
the pipettes and similarly determine the carry over
in blank. Pipettes tend to be the most difficult
to clean.
3. Acceptance criteria
The worst case is that from 2a above, if there is
any carry over (from insufficient cleaning of glassware)
and the same glassware is used to prepare a low
concentration of the same compound (e.g., potency).
Relatively speaking it will cause the highest bias
in the potency assay.
In 2a. The carry over in the blank should be less
than 1% of potency if the same glassware is used
for potency determination.
In 2b. The carry over in the blank should be less
than the uncertainty in the pipette volume for a
class A pipette.
Reminder: It is always necessary to "develop"
the normal cleaning process.
Different detergent, wash cycle, pre-rinse, soaking
of pipettes..... will affect the cleaning effectiveness
significantly. Therefore before we validate the
process, we have to fine tune the procedure and
ensure we'll meet the criteria. |
|
|
I enjoy reading your insightful articles in "Laboratory
Focus" and also appreciate your initiative
to bring together calibration and validation specialists
so we can learn and share information. I have two
questions re: validation requirements for analytical
and general methods that I am hoping you might be
able to clarify.
1) Do general methods (eg, pH measurement of a
product solution, osmolality determination, etc)
listed in USP require validation for regulatory
compliance when applied specifically to product
testing?
2) I have several test methods for various finished
products which require validation (eg., "Formaldehyde
Determination by the Regular Nash Method" for
Product A, Product B, etc.). Many of these product-specific
test methods are based on a (non-compendia) general
test method (eg., "Formaldehyde Determination
by the Regular Nash Method"). Question: Is
a full validation required/necessary for each product
for compliance? Or can one full validation be performed
for one product and a partial validation (eg., just
specificity and, perhaps, robustness) for the rest?
The assumption in the latter approach is that the
fully validated method can be referenced for the
missing but potentially transferable validation
elements since the testing conditions were practically
identical.
|
|
1) Yes, general USP methods do require some form
of method validation. It is because USP usually
does not include enough details, and therefore chances
are scientists need to supplement with some product
specific details (or some local practices) to make
the procedures work. One typical example is the
Karl Fischer titration method. Even though it is
primarily a general procedure, I have seen much
time that product specific details have to be added
to the procedure. For example, appropriate sample
size, number of replicate, sample handling procedure,
proper titration medium, data acceptance criteria,
system suitability criteria. Having said that, I
don't mean to have very comprehensive method validation
for some simple procedures (e.g., pH). For something
like pH measurement, we just need to demonstrate
the accuracy (i.e., absence of matrix interference)
and precision.
2) It is perfectly fine to use method validation
data from other methods. Provided that these data
are not affected by sample matrix (i.e., will not
change from product to product).
For example, in some cases we demonstrate the linearity
of the method by plotting response vs. concentration
of active with no excipient. Therefore, the same
linearity results can be shared among several products
of the same active.
|
|
|
I am inquiring about what happens to the uncertainty
of measurement, when one measurement device is used
to calibrate another measurement device.
For example, if a reference thermometer known to
have an error of measurement of say +/- one degree
C, is used to calibrate a second thermometer. In
this example, the second thermometer would now have
an uncertainty of +/-4deg C (1deg C times 4 = 4
deg. C).
|
|
Typically, there are two factors to
be considered in doing calibration, one factor is
the uncertainty ratio and the other factor is the
traceability. For the Uncertainty Ratio (UR), it is
recommended to be 1:4, meaning that the standard is
four times better than the instrument to be calibrated.
For example, if the process limits are +/- 4 deg C,
an instrument with an uncertainty of +/-1 deg C should
be used and calibrate the instrument with a traceable
standard with an uncertainty of +/- 0.25 Deg C. However,
sometimes there are limitations (technological or
budget) where the UR ratio of 1:2 may be justifiable
on a case by case basis. An UR equals or larger than
1 is not acceptable. In addition, any chosen standard
must be traceable to a recognized agency (eg. NIST,
ASTM) and re-certified periodically. |
|
|
I would like to begin by saying that
I enjoyed the IQ/OQ/PV course very much and I hope
to attend more courses involving this subject area
in the near future. My question at this time is if
you could recommend to me any articles that deal with
or discuss the validation of computer software. |
|
There are three books on computer
system and software validation that may be useful
to you.
Validation
And Qualification In Analytical Laboratories, Ludwig
Huber, Interpharm Press.
Validation Automated Manufacturing And Laboratory
Applications
Guy Wingate, Interpharm Press.
The Survive And Thrive Guide to Computer Validation, Teri Stokes, Interpharm Press. |
|
 |
|
|
