Transcript A Compliant Stability Testing Program

PREPARING FOR A METHOD
VALIDATION
Steven S. Kuwahara, Ph.D.
GXP BioTechnology, LLC
PMB 506, 1669-2 Hollenbeck Avenue
Sunnyvale, CA 94087-5042
Tel. & FAX: (408) 530-9338
e-Mail: [email protected]
Website: www.gxpbiotech.org
IVTMethVal1205
1
GENERAL I.
• In general, when a validation is planned, there
should be enough information available that the
workers will have reason to believe that the
validation will be successful.
• The method validation study should not be used to
“discover” the assay parameters such as accuracy,
precision, and linearity.
• The validation study should confirm proper
performance under normal conditions of use or
suitability for the intended use under normal
conditions, including extremes.
IVTMethVal1205
2
GENERAL II.
• Quite often the method development or adaptation
takes place in a R & D laboratory or under other
conditions where dedicated equipment and
personnel are used.
• This may result in the “discovery” of new factors
when the method is transferred to the user
laboratory, but this should be unexpected.
• No matter how well the assay performed in the
development laboratory, if it does not perform
well in the user laboratory, it is useless.
IVTMethVal1205
3
SYSTEM SUITABILITY vs VALIDATION
• System suitability testing is not the same as validation.
Therefore while the development and setting of system
suitability criteria may be a part of an assay development
and validation program, it cannot be considered to be the
equivalent of an assay validation.
• A full validation study will consider many factors in addition
to the criteria included in system suitability testing, and must
involve testing of the actual sample, not system suit. samples.
• Some organizations regard an assay for which system
suitability criteria have been developed and emplaced as
being “qualified but not validated.”
• The same is true for assays that have not been tested with the
specific sample.
IVTMethVal1205
4
SYSTEM SUITABILITY vs VALIDATION
• 15. METHODS VALIDATION
• “System suitability data alone is insufficient for
and does not constitute method validation.”
• GUIDE 1. GUIDE TO INSPECTIONS OF
PHARMACEUTICAL QUALITY CONTROL
LABORATORIES
• Note: This document is reference material for
investigators and other FDA personnel.
IVTMethVal1205
5
WHAT ARE YOU VALIDATING?
• If the test method will be used for an analyte in
different products (e.g. you are testing for an
excipient used in different products) or for an
analyte in different matrices ( e.g. an active
substance in samples from intermediate stages of
manufacture) then you must use all of these
products in your validation study.
• This will expand your study, but it’s better to do it
all now than to go at the validation in a piecemeal
fashion.
IVTMethVal1205
6
Schering-Plough Products, LLC. 2001
San Juan District Office, FDA
• According to the firm, the AE method for the new
product, Nasonex Unscented Nasal Spray (MAA)
is considered validated based on the AE method
validation for KTL. This validation was completed
on 3/96 according to protocol dated 12/85. On
12/98 the firm approved a new SOP for the AE
method validation. Even when this new SOP
includes the use of negative and positive controls
as well as the gram stain of test organisms during
the AE method validation, the firm has not
revalidated the existing AE method for KTL.
IVTMethVal1205
7
WHEN TO VALIDATE? I.
• The best time is at the end of the method
development process or at the point of transfer to
the user laboratory.
– At this point the reagents and equipment are still
available, the procedure is fresh in people’s memory,
and the method developer should be available.
• Validate at this point, do not put things off until
“late phase III.”
– Work piles up and then a whole bunch of validations
must be done under time pressure.
– Poor planning or a change in schedules can create
major problems, then work is put off and forgotten.
– Putting off work to save money only works if the
product fails.
IVTMethVal1205
8
WHEN TO VALIDATE? II.
• If you wait, people forget, equipment may be
modified.
– This will result in a need for re-training, re-calibrations,
re-conditioning.
– Time will be wasted, and additional resources
consumed.
• If the validation study reveals problems, and you
are in late Phase III, you may not have the time to
go back and fix the assay.
– There may be intense pressure to complete the
validation and ignore the problems, but FDA may see it
in their labs.
IVTMethVal1205
9
WHO VALIDATES? I.
• The end users should do the validation, since the
validation should reflect the actual working
conditions.
– This means that real samples, with real sample
collection and preparation should be used.
– Real analysts working under real conditions.
– If a senior analyst is taken aside and only works on the
method validation to the exclusion of other work, this is
an artificial situation.
– The real situation may be a junior analyst running the
test while doing other tests.
IVTMethVal1205
10
WHO VALIDATES? II.
• Personnel are very important. In system
suitability testing, the analyst should be
considered to be a part of the system.
• If the laboratory has a “validation specialist” who
does only validation work to the exclusion of other
work, this is an artificial situation.
• If there are five analysts, any one of whom could
be called upon to run the test, all five of them
should participate in the analyst-to-analyst part of
the repeatability and intermediate precision
studies.
IVTMethVal1205
11
WHERE IS THE VALIDATION DONE? I.
• The validation should be done in the working
laboratory using the equipment and facilities that
will be normally used.
• If special equipment will need to be transferred to
the QC lab, then do that before the validation.
• If equipment will be shared among different
assays, that should be factored into the
intermediate precision study.
IVTMethVal1205
12
WHERE IS THE VALIDATION DONE? II.
• If an instrument will be parked under the air
conditioner vent and next to a drying oven, then it
should be there during the validation study.
• If a shared HPLC will mean that a column will
need to be changed and re-conditioned before the
assay can be run, that event should be included in
the validation study.
• If the lab operates 24/7 the shifts and days should
be included.
IVTMethVal1205
13
Case Study. Part 1. Initial Conditions
Taken from: Dolan, J.W., LCGC (2000) 18: 1136.
• Assay development and validation done by
Methods Development Laboratory.
• Procedure transferred to QC Laboratory for
routine use.
• Isocratic LC with UV detection at 214 nm.
• Impurity peaks at 5, 6, and 15 min.
• Peaks at 5 & 6 min have USP tailing factor <2.
Peak at 15 min is marginal with tailing factor of
2.1,
IVTMethVal1205
14
Case Study. Part 2. Initial Conditions
• Method transfer protocol consisted mainly of
having QC repeat the separations obtained by
Methods Development.
• Results of Method Transfer Study showed that QC
obtained the same results and transfer was
considered to be successful.
• A few days after the transfer study, manufacturing
of the product increased and QC work load
increased.
IVTMethVal1205
15
Case Study. Problem
• The chromatograms now show irregular peaks.
• In some cases, the peak shapes fail requirements
for the tailing factors.
• In addition to the irregular shapes, there is some
indication of a rising baseline during some of the
runs.
• The results are clearly OOS, and the irregular
peak shapes suggest that the impurities are not
pure. (Remember this was an impurity assay.)
IVTMethVal1205
16
Case Study. Initial Investigation. Part 1
• When peak distortions such as tailing occur for all
peaks, one possibility is a blocked frit or bubble at
the column head.
• A quick solution was to change the column.
– When this was done, the problem appeared to go away,
but quickly came back. Flushing the column head,
manually, had no effect.
– Therefore the column was probably not the source of
the problem.
IVTMethVal1205
17
Case Study. Initial Investigation. Part 2
• To check the irregularity of the peak shape, an
autosampler precision check was performed by
repeatedly injecting the same sample under
standard run conditions.
• The result was a relative standard deviation of less
than 1% when checking the peak areas, despite
their irregular shape.
• This was well within specifications.
IVTMethVal1205
18
Case Study. Initial Investigation. Part 3
• The autosampler was cleaned and reassembled to
see if the problem was being caused by carryover
or cross contamination or a bad sampler.
– There was no change.
• There was some suggestion that the baseline noise
was occurring in a regular pattern, and this led to
several suggestions and checks.
IVTMethVal1205
19
Case Study. Initial Investigation. Part 4
• A bubble or leak in the pump should reduce the
flow rate, but there was no evidence of a
significant change in the retention times of the
peaks.
• The possibility of bubbles in the detector was
checked by extra degassing of the solvent.
• The detector lamp was replaced.
• Nothing changed, suggesting that there were no
bubbles in the detector and the lamp was not
failing.
IVTMethVal1205
20
Case Study. Further Investigation. Part 1
• The baseline should be constant in an isocratic
run, so the rising baseline suggests that the
samples contain a contaminating material that
“bleeds” off the column.
• In an isocratic separation, a contaminated mobile
phase will result in a shift of the baseline, but it
should not rise throughout the separation.
IVTMethVal1205
21
Case Study. Further Investigation. Part 2
• Interviews with the Methods Development analyst
revealed little except that it was customary to
make two runs in the morning and two in the
afternoon. During lunch and overnight, the pump
was left running at a lower rate “to keep the
column conditioned.”
• This procedure was eliminated after the method
transfer study because the HPLC was needed for
other work and “to save on the cost of solvents.”
IVTMethVal1205
22
Case Study. Further Investigation. Part 3
• It was decided to check the baseline change by
letting the column continue to run past the normal
20 min run time. In addition, the flow rate was
increased.
• The run was allowed to go for 180 min. It was
found that the baseline continued to rise in an
irregular manner over the whole 180 min with
several changes in its slope. In addition there were
two, very broad peaks centered around 70 and 150
min. They were superimposed on the rising
baseline.
IVTMethVal1205
23
Case Study. Conclusion
• In isocratic runs, it is known that peak widths
increase as the retention time increases.
Consequently, the presence of the broad peaks
suggests that the samples contain two impurities
that “bleed” through several cycles of separation.
• During Method Development, the extended
“washing” of the columns eluted the
contaminants.
• The introduction of a “flushing” step with a strong
solvent at the end of the 20 min run eliminated the
problem.
IVTMethVal1205
24
Case Study. Corrective Action and Follow-Up.
• The flushing step required “pushing” a strong
solvent through the column for 5 min after the 20
min separation run.
– As this changed the column conditioning and
the programming, it was felt that a revalidation
was needed, but this was not done because of a
bigger problem. (See Below)
IVTMethVal1205
25
Case Study. Follow-Up. 1.
• Part of the problem here arose from the fact that
the validation and method transfer studies were
not conducted “under actual conditions of use.”
(21 CFR 211.194(a)(2) The suitability of all testing
methods used shall be verified under actual
conditions of use.)
IVTMethVal1205
26
Case Study. Follow-Up. 2.
• The “bigger” problem here, of course, is the fact
that two, new, impurities were found. It was now
necessary to identify these substances, decide on
their importance, and redesign the
chromatographic procedure to include them in the
routine analytical run.
• A new validation was needed.
IVTMethVal1205
27
21 CFR 312.23(a)(7)(iv)(a)
• (iv) Reflecting the distinctions described in this paragraph
(a)(7), and based on the phase(s) to be studied, the
submission is required to contain the following:
• (a) Drug substance. A description of the drug substance,
including its physical, chemical, or biological
characteristics; the name and address of its manufacturer;
the general method of preparation of the drug substance;
the acceptable limits and analytical methods used to assure
the identity, strength, quality, and purity of the drug
substance; and information sufficient to support stability
of the drug substance during the toxicological studies and
the planned clinical studies. Reference to the current
edition of the United States Pharmacopeia--National
Formulary may satisfy relevant requirements in this
paragraph.
IVTMethVal1205
28
21 CFR 312.23(a)(7)(iv)(b)
(b) Drug product. A list of all components, which may include
reasonable alternatives for inactive compounds, used in the
manufacture of the investigational drug product, including
both those components intended to appear in the drug
product and those which may not appear but which are
used in the manufacturing process, and, where applicable,
the quantitative composition of the investigational drug
product, including any reasonable variations that may be
expected during the investigational stage; the name and
address of the drug product manufacturer; a brief general
description of the manufacturing and packaging procedure
as appropriate for the product; the acceptable limits and
analytical methods used to assure the identity, strength,
quality, and purity of the drug product; and information
sufficient to assure the product's stability during the
planned clinical studies.
IVTMethVal1205
29
WIL Research Laboratories, Inc. 1999
GLP and Bioequivalence Investigations Branch
Division of Scientific Investigations, Office of Medical Policy,
CDER
• Your response indicates that none of the FDA 483 observations
suggest non-compliance with the GLP regulations. The response
also indicates that, as a contract laboratory, WIL Research cannot
be held responsible for sponsor conducted activities (analyses of
dosing formulations) which are not under WIL Research’s direct
control.
• This response does not satisfy the requirements as set forth in
Section 58.31(d) which requires that testing facility management
shall assure that test and control articles or mixtures have been
appropriately tested. Specifically, when analytical tests for
homogeneity, concentration, and stability of dosing formulations
are performed by another laboratory, it is the testing facility’s
responsibility to assure that such tests are performed,
IVTMethVal1205
30
Sec. 58.31(d) Testing facility management.
• For each nonclinical laboratory study, testing
facility management shall:
• (d) Assure that test and control articles or
mixtures have been appropriately tested for
identity, strength, purity, stability, and uniformity,
as applicable.
IVTMethVal1205
31
Analytical Procedures and Methods Validation. CMC
Guideline: August, 2000
• C. Stability-Indicating Assay: A stability-indicating
assay is a validated quantitative analytical
procedure that can detect the changes with time
in the pertinent properties of the drug substance
and drug product. A stability-indicating assay
accurately measures the active ingredients,
without interference from degradation products,
process impurities, excipients, or other potential
impurities. . . . Assay analytical procedures for
stability studies should be stability-indicating,
unless scientifically justified.
IVTMethVal1205
32
483 Observation:
“There is no data to show that the method used for XX
stability testing has been validated as stability-indicating
with respect to acid and base hydrolysis and photolysis;
there was inadequate data for oxidation and thermal
degradation.”
IVTMethVal1205
33
483s on Impurities/Degradants
“Analysts can use “suppress unknowns” to omit reporting
of unknown peaks in HPLC run”
“Unidentified HPLC peaks found during stability
testing of the validation lots were not identified or
evaluated.”
“Systematic identification of degradation products has
not been performed.”
IVTMethVal1205
34
DEVELOPMENT REPORT I.
• You should not attempt to validate an
undeveloped or poorly developed test.
• What tests were considered? Is this a compendial
test? Is this the right test?
• What is known about the test?
– Problems, interferences, cross reactions, chemistry.
• What are the literature references behind the test?
• Are preliminary estimates available for validation
parameters?
IVTMethVal1205
35
DEVELOPMENT REPORT II.
• The development report should have enough
information that you feel sure that the validation
will be successful.
• The stability of reagents and intermediate
solutions and the “stopping points” should be
identified for the test.
• The standards and controls should have been
defined, SOPs written and material transferred to
the validation workers.
IVTMethVal1205
36
FACILITY I.
• Is your laboratory ready?
• Do you have space for the work, the personnel, the
instrument? Is the space clear or will it need to be
shared?
• This is sort of an IQ for the lab. Can you install
your validation study without disrupting things?
• The work will need to be done within the context
of existing work. You cannot shove it off into a
corner and then expand it later.
IVTMethVal1205
37
FACILITY II.
• What about data storage?
• If you do the validation, where are you going to
store the records of the work?
– Do you at least have a fireproof file cabinet?
– Is there space in the document storage facility?
• Are there any special records such as
electrophoresis gels, animal or cell culture tissues,
computerized data that are recorded on discs.
– How will you store these?
– Do you have a back up location to store copies?
IVTMethVal1205
38
EQUIPMENT I.
• Do you know if the needed equipment and utensils
are ready and suitable for use. (how about a DQ?)
• All of the important instruments should have been
qualified (IQ and OQ at least).
• Do you have calibration and maintenance SOPs
available for the instruments?
• At the point where you start the validation, the
equipment should have been recently calibrated
and required maintenance should have been done.
IVTMethVal1205
39
EQUIPMENT II.
• If the normal calibration and maintenance cycles
will call for this work to be done again during the
validation study, do it.
– Do not hold off on calibration and maintenance just
because you are in the middle of a validation study.
• Do you have spare parts and other material that
may be needed for calibration and maintenance
work?
• Do not overlook small items. Pipettors may need
calibration and maintenance too.
IVTMethVal1205
40
EQUIPMENT III.
• If computers and computerized equipment are
used, they should be compliant with 21 CFR 11 at
least to the extent required by the guidelines.
• How will you validate any special software,
especially if it is used to work up final data?
• How will you archive the computer records?
• Do you have computer techs or other assistance
available if needed?
• What about consumables such as paper, discs,
tapes, flashdrives, etc.?
IVTMethVal1205
41
IT IS A CAPITAL MISTAKE
TO THEORIZE BEFORE ONE HAS
DATA
• Sir Arthur Conan Doyle, aka Sherlock Holmes
IVTMethVal1205
42
Test Method Procedure (SOP) I.
• If there is no SOP for the test, how do you know
what you are validating?
– Note that the validation study is only valid for the test
as described in the SOP.
• Is the SOP clearly written? Does it describe
reagent preparations? Is the procedure for data
work up clear?
• The SOP should clearly specify the supplies
needed, the length of time the test will require.
• The SOP MUST describe safety precautions and
potential hazards.
IVTMethVal1205
43
Test Method Procedure (SOP) II.
• Reagent stability should be clearly stated.
• Stopping points in the assay should be clearly
identified.
– If the analyst needs to take a break, where can the
procedure be stopped without affecting the test results?
– Where are the steps when a break in the activity
sequence cannot be tolerated or where stopping the test
will affect the outcome?
• The SOP should contain references so that
analysts can refer to the original literature, if
necessary.
IVTMethVal1205
44
STANDARDS
• If a compendial or other well-recognized standard
exists, your working standard for the validation
should have been calibrated against it.
• There should be an SOP describing the
preparation of the working standard even if it is
nothing more than a solution of a purchased
preparation. The solvent and purchase
specifications should be stated.
• If it had to be “prepared” by synthesis or special
purification this should be described.
• The stability of the standard in its “working form”
should be known and documented.
IVTMethVal1205
45
Tri-State Analytical Laboratory LLC 2003
FDA New Orleans District, Nashville Branch Office
• Analytical results were reported to [redacted]
stating that a sample met specifications when
either out-of-specifications (OOS) results were
obtained on the sample analysis or on the quality
control samples used to determine the validity of
the analytical results. These OOS results were not
investigated/documented properly to assure
results reported to [redacted] were accurate and
valid.
• Use of reference standards and reagent solutions
for extended periods of time without data in the
analytical records supporting time of use.
IVTMethVal1205
46
CONTROLS I.
• The control should resemble the actual test sample
to the degree that it can act as a suitable surrogate
for the actual samples.
• Several controls may be needed if samples can
come in several forms, e.g. the analyte may be
dissolved in different solvents or the sample
matrix may vary.
• Negative controls should be the sample matrix
without the analyte.
• A positive control should be a real sample.
– Positive controls prepared by adding a standard to the
negative control are not as good.
IVTMethVal1205
47
CONTROLS II.
• Positive and negative controls should be used.
• Do not use the solvent blank for the “zero” standard
as the negative control, unless it is real.
• Controls can be prepared by using “practice” runs
from your development work. Later, positive controls
can be obtained from failed lots or validation lots.
• Ideally, you can prepare standard curves by diluting
your positive control with the negative control. At
least try to have “high/low” controls by diluting the
positive control with the negative.
• System suitability samples can be prepared by spiking
controls. You want a difficult sample.
IVTMethVal1205
48
TRAINING I.
• Everyone who will participate in the validation
must be properly trained to run the assay.
– This includes the supervisor. If the supervisor does not
understand what is being supervised, how can
supervision be done?
– Analyst to analyst variation should not be due to
different levels of training. Analyst skills will change
and mature over time.
– If a reproducibility (lab to lab variation) study will be
done, it is important to show that all analysts were
equally trained and qualified.
IVTMethVal1205
49
TRAINING II.
• The training should be completed BEFORE the
validation study is launched.
• Do not use the validation study as your training
platform. Otherwise training activities will create
artificial situations that are not related to the
validation or normal testing.
• Specialist training (statistics, instrument
maintenance and calibration) should be completed
before the validation is attempted, so that the
specialists can have proper input to the validation
protocol.
IVTMethVal1205
50
TRAINING III.
• The training must be documented and proof of
training supplied.
• Simple signatures to the effect that analysts have
read protocols or that they have been observed to
perform the test correctly are not sufficient.
• Proof should be provided in the form of test
results obtained with standards and controls. This
should include calculations and use of
spreadsheets. System suitability or precision and
accuracy specifications can be employed.
IVTMethVal1205
51
TRAINING IV.
• Statistics is an important body of knowledge for
performing method validations and good training
in statistics should be provided for all QC
analysts. Simple basic statistics is sufficient.
• Specialist training in DOE or fractional factorial
designed experiments is needed for at least a few
members of the method validation team.
• Intermediate precision and robustness are best
studied using fractional factorial designs. The
DOE studies will minimize the total number of
tests runs that are needed.
IVTMethVal1205
52
Germiphene Corporation 2005
Division of Manufacturing and Product Quality
Center for Drug Evaluation and Research
• Several observations cited a lack of following
procedures even though there was documentation
that your employees were trained in the
procedure. For example, when asked about the
procedure for determining the status of equipment
prior to use, the employee could only say that the
analysts were trained to check the calibration
prior to use. However, the piece of equipment in
question was in use and out of calibration for 1.1
months
IVTMethVal1205
53
RAW MATERIAL I.
• If you have special raw material needs, you will
need to control these items just as you would with
manufacturing raw material.
– For instance, if you have a nucleic acid primer or a
monoclonal antibody that is custom made for your
assay, you will need to have purchasing specs. that
define the properties needed. If you need ACS grade do
you have it? Will USP grade substitute?
– You will need raw material testing SOPs to verify that
the material delivered is what you ordered and what
you need.
IVTMethVal1205
54
RAW MATERIAL II.
• If you do not have test procedures for these, how
did you qualify the vendor? Reputation is not
enough!
– Did you audit the vendor?
– How about confirming the vendor’s certificate of
analysis? NEVER TRUST a CoA from a new source.
• The raw material qualification should be done
before you use the test material in the validation
study.
– What if the validation goes bad because the reagent
does not perform properly?
IVTMethVal1205
55
REAGENTS I.
• In addition to qualifying your raw material, you
should have SOPs describing how solutions and
other test reagents are made and, if necessary,
checked for proper concentration or potency.
• If the procedure is simple, like making 0.85%
(w/v) NaCl or 0.1 N HCl, then you could write it
into the test method SOP.
• If a reagent requires extensive preparation or
where careful checking needs to be done, an SOP
will be needed.
IVTMethVal1205
56
REAGENTS II.
• Consider a monoclonal antibody that needs to be
isolated from culture supernatant, derivatized
with a fluorescent tag or an enzyme, and have its
binding constant determined. More than one SOP
may be needed.
• If culture media are needed, how are you going to
qualify the media in your laboratory? The
manufacturer may obtain good results, but what
about your laboratory?
IVTMethVal1205
57
REAGENTS III.
• What is the stability of these reagents? Even
solutions of inorganic salts will evaporate over
time and with use.
– You need to specify an expiration date for all prepared
solutions. A real stability study may be required.
• Are there special storage requirements?
– Is the solution photosensitive?
– Does it need refrigeration?
– If it needs refrigeration, can it work if added cold? Does
it need to be brought to room temperature? If so, how
long before it expires at room temperature?
IVTMethVal1205
58
PLANNING I.
• Have the costs for the validation study been
factored into the budget?
• What facilities and equipment will be needed and
when? How can these needs be woven into the
normal laboratory activities?
• What people are needed and when? How will these
needs be woven into normal laboratory activities?
• Can the needs really be superimposed on normal
laboratory operations? Are additional personnel
and facilities needed? Is the budget sufficient for
all the activities?
IVTMethVal1205
59
PLANNING II.
• Do the supervisors understand their
responsibilities and have plans for dealing with the
needs in their individual areas of responsibility?
• Remember that simply imposing a study on a lab
by decree and expecting them to deal with the
added workload without a plan and supporting
structure, is a sure recipe for disaster and is a
clear symptom of bad management.
• BY FAILING TO PLAN, YOU ARE PLANNING TO FAIL!
– Deming?
IVTMethVal1205
60
Validation Protocol I.
• A method validation should be a PROSPECTIVE
study, so a protocol should be written BEFORE
the work is started.
• The protocol needs to state what is being validated
– The test method (SOP # and effective date), the types of
material to be studied (different matrices or
concentrations).
• What elements will constitute the validation?
– Based on ICH Q2A or other works, what are you
checking? Accuracy, linearity, LOD or LOQ?
IVTMethVal1205
61
Validation Protocol II.
• The sequence of work to be done and by whom.
– What do you study first? Who does this study?
– Who works up the data?
– How is the work to be done?
• How is the information to be recorded and how is
it to be stored?
• Who is responsible for supervising the work,
preparing the protocol, and preparing the final
report? Responsibilities must be assigned
otherwise things fall into cracks.
IVTMethVal1205
62
Validation Protocol III.
• What are the criteria that must be met for a
successful validation?
• A timetable for activities should be proposed with
intermediate check points for verifying that the
work is progressing well or for clearing away
problems that may develop.
• The final protocol approval should be done by an
individual or group who verify that the validation
study is ready for execution.
IVTMethVal1205
63
SPECIAL NEEDS I.
• All tests should be validated, but not all tests
follow the same general pattern for execution.
Microbiological tests, dissolution tests, content
uniformity and endotoxin tests are often cited as
examples. With content uniformity and dissolution
tests, there are actually two stages of testing. The
second stage, where the analyte is being measured
can follow the usual procedures, but validating the
sample acquisition or preparation may be a
problem.
IVTMethVal1205
64
SPECIAL NEEDS II.
• With any test, the accuracy and reproducibility
must be established over various test conditions.
• For accuracy, it is important to establish the fact
that the test really does measure the factor itself in
the magnitude that is present. For instance, in
dissolution testing, the assay may do a really good
job of measuring the analyte, but that is not the
question; it is the rate of appearance of the analyte
in the solution that is critical. It is very important
to know what is being measured.
IVTMethVal1205
65
SPECIAL NEEDS III.
• Reproducibility is important because a highly
variable test will cast doubt on the accuracy of the
test result. If the test is highly variable, the results
will be virtually irreproducible. In which case it
will be useless for manufacturing or dosing
control.
• There is always one reproducibility check that can
be done. If the test is working at all, it should be
possible to choose some concentration of analyte
and some test condition that should give a result
within the normal response range.
IVTMethVal1205
66
SPECIAL NEEDS IV.
• This condition can be created and reproduced
about six times and the test could be run, even
without a standard curve, as long as there is some
measurable response. This can be done with any
test, including whole animal or cell culture tests.
• This check, which gives repeatability (or within
run precision) should show the minimum possible
variation. If the variation here is large, then it
shows the presence of uncontrolled factors in the
test, and these must be found and controlled
during the method development process.
IVTMethVal1205
67
SPECIAL NEEDS V.
• There are some who run animal or tissue culture
based tests who have been known to accept
potency test results ranging from 5% to 160% of
the expected values.
• They take the fatalistic attitude that “that’s the
way it is with biological tests.”
• The data are so variable that it is not possible to
assign sensible values for doses or therapeutic
windows.
• Often an investigation will show that virtually no
effort was expended in trying to reduce the
variability in the assay.
IVTMethVal1205
68
SPECIAL NEEDS VI.
• High variability can only be accepted when a
development laboratory shows that it made
serious attempts to reduce the variability of the
test. Often the problem arises because nonanalysts are placed in the position of developing
tests.
• If the previous repeatability test shows a low
variability, then the high test variability is due to
factors normally covered in the intermediate
precision portion of the validation study.
IVTMethVal1205
69
The Dogbert Conjecture
• “If your controlled tests have never found
psychic powers, how do you know the tests
work for that sort of thing?”
• “Isn’t that like using a metal detector to find
out if there are unicorns in your sock drawer?”
• - Dogbert (1/20/98).
IVTMethVal1205
70