Out of Specification Results (OOS) A One Day Workshop
description
Transcript of Out of Specification Results (OOS) A One Day Workshop
Out of Specification Results
(OOS)A One Day Workshop
Presented by: Karen S. GinsburyPCI Pharmaceutical Consulting Israel Ltd.
For IFFMarch 2007
PCI Pharmaceutical Consulting Israel Ltd 2/82
Purpose of Workshop
Understand Current Industry Practice asit relates to all aspects of handling:– Out of Specification– Unusual– Out of Trend Results
When is retesting legitimateWhen does testing have to stopWhen do the authorities need to be notified
PCI Pharmaceutical Consulting Israel Ltd 3/82
Workshop Structure
09:00 – 10:30 10:30 – 11:00 11:00 – 12:30
12:30 – 13:30 13:30 – 15:00
15:00 – 15:15 15:15 – 16:30
Background and Overview: Barr and AbleCoffee BreakThe Making of the FDA GuideCase study: what NOT to doLaboratory investigation and checklistLunch BreakFDA Guide Continued:Extended investigation, retesting protocolQualitative and quantitative testsCase studyTea BreakCase Study and workshop wrap-upReporting results; reporting to regulators
PCI Pharmaceutical Consulting Israel Ltd 4/82
PART 1 – BARR and ABLE
LitigationUSA regulatory environment around
late 1980’s through 1993Current regulatory environment in USAble - 2005
PCI Pharmaceutical Consulting Israel Ltd 5/82
Some Definitions
What is an OOS Result?
What is a test result
Definition of reportable value
“A reportable value is the end result of the complete measurement method as documented. It is the value compared with thespecification, the values collected when the term replicates is used, the values used for official reports, and the values used for anystatistical calculation or analysis.”
Torbeck (February 1999, Pharmaceutical Technology)
PCI Pharmaceutical Consulting Israel Ltd 6/82
Before Barr – Current Practice
Prior to 1993 and the court decision – it wasCOMMON practice to retestonceor in exceptionally good companies twiceand to release the batch if the retest result waswithin the specification
Companies had not really thought about thepractice
But then…nor had the regulators
PCI Pharmaceutical Consulting Israel Ltd 7/82
The Barr Court Case and Judge Wolin
From the New York TimesFebruary 6, 1993, Saturday(AP); Financial Desk
COMPANY NEWS; Judge Rules On Barr LabsA generic drug manufacturer must recall batches ofsome of its medicines and stop distributingothers until the company completes studies of itsmanufacturing process, a Federal judge ruled onThursday. But United States District Judge Alfred M. Wolin refused a request by Federal pharmaceuticalregulators to order a complete shutdown
PCI Pharmaceutical Consulting Israel Ltd 8/82
Barr and OOS
Faced with potential closure, the companytook FDA to court
The judge went into great details as to themeaning and implications of OOS results
The outcome: FDA draft guidance: 1998FDA final guidance: 2006
PCI Pharmaceutical Consulting Israel Ltd 9/82
Barr: What happened in court
The judge heard experts onbehalf of FDA and Barr regardingthe practice of retesting
FDA wanted retesting to be bannedunder all circumstances
After a long hearing at which five industry experts, anFDA investigator, and several company employeestestified, Judge Alfred M. Wolin, U.S. District Judge forthe District of New Jersey, issued a 79-page opinion
PCI Pharmaceutical Consulting Israel Ltd 10/82
The Barr Court Case 1993
Reported problems include– misplaced records– test data recorded on scrap paper– failure to control manufacturing steps such as
those governing products' physical properties– release of products not meeting their
specifications– inadequate investigation of failed products
PCI Pharmaceutical Consulting Israel Ltd 11/82
Barr: “Testing into Compliance”
Barr had numerous failuresPerformed retests with
– no investigations– no regard for process and product history
Tested until results met specificationsThen irrespective of previous OOS results
for the batch, released productreporting only the passing resultsreporting only the passing results
Q: How do you report passing OOS’s on COA?
PCI Pharmaceutical Consulting Israel Ltd 12/82
Reading the Judgment
Reading the Barr Court Judgment is likereading FDA’s draft guidance and pretty similar to thefinal guidance
Judge Wolin preferred to use the term"out-of-specification" (OOS) laboratory results ratherthan the term "product failure" which was morecommon to (preferred by?) FDA's investigators
Ruled that an OOS result identified as laboratory error by a failure investigation or an outlier test, orovercome by retesting is not a product failurenot a product failureBUT
Limited situations where laboratory error could be used
PCI Pharmaceutical Consulting Israel Ltd 13/82
Guide to Inspection: QC Labs
Issued July 1993 (must have been working onit while the court case was ongoing)
Addresses OOS results and instructsinspectors to be alert
“Evaluate the company's system to investigate laboratory test failures. These investigationsrepresent a key issue in deciding whether aproduct may be released or rejected and formthe basis for retesting, and resampling “
(Most of the information is now in FDA’s guide)
PCI Pharmaceutical Consulting Israel Ltd 14/82
Guide to Inspection: QC Labs
OOS results fall into three categories:– laboratory error– non-process related or operator error– process related or manufacturing process error
Evaluate the company's retesting SOP for compliance withscientifically sound and appropriate procedures
A very important ruling in one recent court decision sets forth aprocedure to govern the retesting program
This district court ruling provides an excellent guide to use inevaluating some aspects of a pharmaceutical laboratory, butshould not be considered as law, regulation or bindinglegal precedent
The court ruled that a firm should have a predetermined testingprocedure and should consider a point where testing ends andproduct is evaluated. If results are not satisfactory, product isrejected.
PCI Pharmaceutical Consulting Israel Ltd 15/82
After Barr – Current Industry Practice 2007
ALL pharmaceutical companies in developedcountries and many in less developedcountries have SOPs for handlingOut of Specification results
There are still many investigational findingsconcerning out of specification results
There are still numerous issues, particularlywith transparency:What do you report on the COA?
Able laboratories suspended activity in 2005because of Out of Specification results
PCI Pharmaceutical Consulting Israel Ltd 16/82
When it all goes wrong….Able 2005
The Quality Unit failed to: Review computer audit trails in the Waters Empower
Data Acquisition System Provide adequate training to analytical chemistsThese practices led to: The QU releasing batches failing in-process, failing in-process,
finished product and stability specificationsfinished product and stability specifications Submission of erroneous data in Annual Reports
and Prior Approval Supplements Ceasing manufacture, distribution and recall of all
products as of 13 May 2005 and withdrawal of atleast 5 ANDAs
PCI Pharmaceutical Consulting Israel Ltd 17/82
Resample, Re-injection, Reprocessing
18
Time for a Break
PCI Pharmaceutical Consulting Israel Ltd 19/82
From Able Laboratories 483
Notebooks and binders lacked data from all testing conducted in the QC Laboratory.
Records did not include all data such as OOSresults, chromatograms, sample weights, andprocessing methods.
OOS results were substituted by passing results by Analysts and Supervisors.
The substitution of data was performed by cutting and pasting of chromatograms, substituting vials, changing sample weights and changingprocessing methods
PCI Pharmaceutical Consulting Israel Ltd 20/82
That isn’t relevant…it’s fraud
Think about the following scenariobefore saying it couldn’t happen:– You are just starting up your HPLC system in
the morning– You inject 5 replicates of standard– The first four give perfect responses– The fifth is way off / makes no sense,
obviously incorrectWhat do you do?
PCI Pharmaceutical Consulting Israel Ltd 21/82
It couldn’t happen in my company
A split second wrong judgment can appear asfraud to the inspector
You inject one more standard injection and itgives a perfect response
Clearly you were right and the original “5th”injection was dirty glassware or mis-injected,or, or, or……
PCI Pharmaceutical Consulting Israel Ltd 22/82
It couldn’t happen in my company
But you don’t want to show 4 good injections, one lousy and another good one
So you cut out the bad one and documentonly the good ones in the notebook
Sounds “horrendous?” It has been done, probably in some of your
laboratories! Can you be sure none of youranalysts ever did this?How can you be so sure?
PCI Pharmaceutical Consulting Israel Ltd 23/82
GMP and OOS Results
Do you know where the term OOS appears in
– EU GMP regulations it didn’tas of June 2006 it
does
– US GMPs it doesn’t
– Q7A GMP for APIs it DOESbecause Q7A was written after 1993
PCI Pharmaceutical Consulting Israel Ltd 24/82
Q7A on OOS
PCI Pharmaceutical Consulting Israel Ltd 25/82
Ever heard this….
“The Lab don’t know how
to test”
“Give it to Pete…
he knows how to do
that test…
Dave always gets badresults!”
PCI Pharmaceutical Consulting Israel Ltd 26/82
OOS Case Study #1
Complex biochemical assay: 6 replicates Inherent variability allows for wider than usual
specification of 80 – 120%Results: 45, 50, 46, 52, 65, 69%
What would you think if it happened to you?The lab technician has 20 years seniorityNo other technician is familiar with the test
PCI Pharmaceutical Consulting Israel Ltd 27/82
OOS Case Study #1 - continued
“Must be my mistake!”WRONGAnalyst retested (not in accordance with SOP)Results: 72, 69, 81, 80, 82, 81%
NOW what would you think?
PCI Pharmaceutical Consulting Israel Ltd 28/82
OOS Case Study #1 - continued
“Results: 72, 69, 81, 80, 82, 81%
72 and 69% must be “OUTLIERS”
Average 81, 80, 82 and 81 and result passesBatch can be releasedReport only this set of results
PCI Pharmaceutical Consulting Israel Ltd 29/82
OOS Case Study #1 - continued
The outcome…..Product complaints from patients and doctors:
Product is sub-potentLitigationProduct recall Investigation reveals weighing error in
production
PCI Pharmaceutical Consulting Israel Ltd 30/82
What is the purpose of testing?
To find out the value of a specific parameterTo assess if that parameter meets the
pre-determined specification for each lotSo as to make a sound scientific judgment
regarding product release or rejectionWhat happens in your company when there is
an OOS result? IDEALLY the analyst doesn’t know the
specification because of “BIAS.”
PCI Pharmaceutical Consulting Israel Ltd 31/82
Is it possible to ensure a correct result?
Not at a 100% certainty levelJust as it is not possible to prevent an incorrect
result (at the 100% certainty level)What is a correct result?A result that is identical to the true result….
BUTWhen testing, we NEVER know the true result
PCI Pharmaceutical Consulting Israel Ltd 32/82
Consequences of an Incorrect Test Result?
FALSE NEGATIVE
Product declared fit for use when not
Side Effects Death
FALSE POSITIVE
Product declared unfit for use when fitRejection Financial Loss
PCI Pharmaceutical Consulting Israel Ltd 33/82
Is it possible to ensure a correct result?
It is possible to ensure a result as close aspossible to the true result
Using a quality assurance program in thelaboratory
Validate methodsQualify analystsFollow methods as writtenQualify, calibrate and maintain equipmentReport deviations / malfunctions
PCI Pharmaceutical Consulting Israel Ltd 34/82
Types of Tests
Quantitative: assays and some limits testsQualitative: e.g. sterility test, appearanceChemicalPhysicalMicrobiological
PCI Pharmaceutical Consulting Israel Ltd 35/82
Chemical Testing
Inherently reliablePrecision is usually considerably better than
for microbiolgical and biochemical testingOutlier testing is forbidden by the FDA guide
for chemical testing(usually have less replicates anyway)
Don’t forget case study #1 – How NOT tohandle OOS results
PCI Pharmaceutical Consulting Israel Ltd 36/82
Physical Tests
e.g.1 Black spots in powdere.g.2 Fill volume
are particularly problematic, since results arealmost certainly correct.Is there any place for retesting?In e.g. 1 Probably notIn e.g. 2 Maybe
Is there any place for resampling?
PCI Pharmaceutical Consulting Israel Ltd 37/82
FDA OOS Guidance, October 2006
Draft from September 1998 for comments Was confusing and open to misinterpretation
particularly by persons not familiar with theindustry
Final guidance is still confusing but a lot better thanthe draft
The draft was referenced, as was the Barr court case during inspections, so the final guidance will definitelybe used
Should be considered in preparing an OOS SOP
PCI Pharmaceutical Consulting Israel Ltd 38/82
FDA Guide on OOS Results
Finalized October 2006Contains “nonbinding recommendations”Requires that an investigation is performed
ANYTIME that an OOS is obtainedWants to include ALL suspect resultsWants an SOP stating number of retests
UPFRONT
PCI Pharmaceutical Consulting Israel Ltd 39/82
Table of Contents
IntroductionBackground Identifying and Assessing OOS Results –
Phase I: Laboratory Investigation– Responsibility of Analyst– Responsibilities of Laboratory Supervisor
PCI Pharmaceutical Consulting Israel Ltd 40/82
Table of Contents (continued)
Investigating OOS Results –Phase II: Full Scale OOS Investigation– Review of Production– Additional Laboratory Testing– Reporting Testing Results
Concluding the Investigation– Interpretation of Investigation Results– Cautions– Field Alert Reports
PCI Pharmaceutical Consulting Israel Ltd 41/82
Out of Specification Results
IF you don’t know the specification youshould never have an OOS, but we don’tlive in an ideal world!
HOW can you have an OOS if:– the method is validated– analysts are qualified– equipment is calibrated and
well-maintained– notebooks have full record of testing
PCI Pharmaceutical Consulting Israel Ltd 42/82
How to Identify OOS Results
Compare result with specificationBe sure that specification is to required degree
of accuracy and round the result to this value
E.g. spec 95 - 105 94.9 passesspec 95.0 - 105.0 94.9 fails
Message for your R&D Department:think carefully before setting specifications!!!
PCI Pharmaceutical Consulting Israel Ltd 43/82
How to Identify OOT Results
Out of Trend or unusual results are generallyresults that:– MEET the product specification– ARE outside the control limits of the process,
where control charts are usedor
– ARE different to results usually obtained(e.g. spec: 95.0 – 110.0usual results: 98.5 – 101.0OOT result: 96.4
PCI Pharmaceutical Consulting Israel Ltd 44/82
Initial Assessment of OOS Result
Bear in mind prior:– Product history– Process history– Test history– Reliability of equipment– Reliability of the analyst– Precision of the test (validation)
PCI Pharmaceutical Consulting Israel Ltd 45/82
Out of Specification Results
Failure of a control is NOT an OOS eg:– Standard shows impurity peak and assays at
incorrect retention time– Standard shows assay of 80%– System suitability drift– In this case all test results SHOULD BE INVALIDATED
and the test REPEATED using an additional aliquot ofthe same preparation if possible, or an additionalaliquot from the same sample (What if product degrades and this is not possible? – need to defineduplicate sampling procedure so that always havespare material)
– Data collected is retained on file ANYWAY– What happens if results are OOS? Do you need to err
on the side of caution e.g. possible homogeneity issue?
PCI Pharmaceutical Consulting Israel Ltd 46/82
OOS – Laboratory Investigation
IF you don’t question an “in-spec” result, whydo you question an OOS?
Must have a Retesting Policy:– Laboratory investigation
• equipment -e.g. glassware, calibration, maintenance
• question large differences between replicates
• calculations
• Product history etc. etc. etc. (to be continued)
• 4M’s: man, machine, methods, materials
47
Time for Lunch
PCI Pharmaceutical Consulting Israel Ltd 48/82
FDA Guide: Introduction
Applies to chemistry – based testing[KSG: primarily HPLC and GC]
Includes in-process testing except (footnote), where the purpose is to prevent process drift[KSG: discuss – could still have OOS!]
Principles apply to CONTRACT firmsTimely, unbiased investigation [KSG: OOS Policy is major SOP approved by
senior management]
PCI Pharmaceutical Consulting Israel Ltd 49/82
FDA: Laboratory Investigation
The source of OOS should be identified as:– Aberration of measurement process– Aberration of manufacturing process
[KSG: sampling process?]Even if the batch is rejected, an investigation is
required:– Other batches involved?– Other products involved?
Need written investigation, conclusions, follow-up
PCI Pharmaceutical Consulting Israel Ltd 50/82
FDA: Laboratory Investigation
Assess accuracy of lab data:– Before test preparations are discarded
(composite / homogeneous source of aliquot tested)
– Hypothesis testing using same test preparationfor laboratory error or instrument malfunction
– IF no meaningful errors were made [found?],conduct a full scale OOS investigation
– Contract lab conveys data and findings to you
PCI Pharmaceutical Consulting Israel Ltd 51/82
Invalidate resultPerform new test on same sample
Convincing evidenceof
Laboratory Error
Correct if possibleor
Reject Batch
Production Error
ResampleDouble sample
Revise sampling procedures
Sampling Error
QA decisionregarding
batch disposition
All withinspecifications
Reject Batch
One resultOOS
Retest ???further aliquots
from original sample
InconclusiveNo evidence of production error
QA Investigation
Inconclusive
Laboratory InvestigationReport to QA (copy in batch file?)
OOS result
OOS Flow Chart
PCI Pharmaceutical Consulting Israel Ltd 52/82
Responsibility of Analyst
To follow test procedure as writtenTo be alert to errors and STOP test BEFORE
obtaining the result if error is suspected,recording what happened e.g. spill
Analyst responsible for ensuring thatinstruments meet performance specificationsand are properly calibrated [KSG: maintained?]
Once an OOS result is obtained to review allrecords relative to the test to identify possiblelaboratory error
PCI Pharmaceutical Consulting Israel Ltd 53/82
Supervisory Role
Supervisors / team leaders / laboratory head:– Should be experienced analysts– Frequently audit while tests ARE BEING
performedin order to be able to objectively investigateOOS results
PCI Pharmaceutical Consulting Israel Ltd 54/82
Responsibility of Supervisor
Objective assessment withoutpreconceived assumptions as to cause of OOS
Immediate assessment may include:– Re-examination of:
• actual solutions• Test units• Glassware
used in the original measurements and preparations
This could provide more credibility for laboratory error hypotheses
PCI Pharmaceutical Consulting Israel Ltd 55/82
Responsibility of Supervisor
To INVESTIGATE:– Review notebook / worksheet with analyst :
• Was method was followed: with a copy of themethod in your hand, have the analyst describeexactly how they performed each step:confirm that the method was understood & followed
• Review raw data: Perform calculations againincluding checking dilution schemes
• Unauthorised changes to automated calculations• Examine reagents, (reference) standards, solutions• Examine glassware• Performance of instruments
PCI Pharmaceutical Consulting Israel Ltd 56/82
Hypothesis Testing
If you re-create an OOS conclusions are stronger Might retest recent sample to confirm / refute
equipment malfunction: retro consequences? Examine retained samples e.g.
– Re-inject solutions where a transient equipmentmalfunction is suspected [air bubbles]Hard to prove, but re-injection of samepreparation can provide strong evidence
– Release rate testing of slow release tab / cap:retest of original unit may show it was damaged duringtesting
– Additional extraction to show that problem ismethod related – then revise method
PCI Pharmaceutical Consulting Israel Ltd 57/82
Hypothesis Testing, Case Study #2
LAL test on radioactive product (short half-life) is positive
Product history – no previous failureOther products tested in same series passed Initial laboratory investigation: no evidence
of lab errorDilution used: 1:70
specification allows up to 1:140 dilution
PCI Pharmaceutical Consulting Israel Ltd 58/82
Hypothesis Testing, Case Study #2
– Perform retest in parallel with productioninvestigation because of short half-life
– Production investigation included samplingglassware and equipment for LAL residuesNO positive results
– Repeat test at 1:70 and 1:140 dilutionsResults were in spec i.e. no positive LAL
– New LAL reagent prepared and same sampletested with old and new reagent:
• Old: failed• New: passed
PCI Pharmaceutical Consulting Israel Ltd 59/82
Out of Specification Results
If the laboratory investigation is conclusiveconclusive– Document findings– INVALIDATE original test– Perform NEW test on same sample– Report original result with investigation as well as new
result in batch record for QA review prior to release– COA carries new result only; some companies use an
asterisk and indicate that there was an OOS If the laboratory investigation is NOT NOT conclusive
inform QA (or customer for contract lab)
PCI Pharmaceutical Consulting Israel Ltd 60/82
Out of Trend Results
If the laboratory investigation is conclusive orinconclusive, consult with QA
In most cases, DO NOT perform any additionaltesting or sampling
Make product disposition judgment based on:– Original result– Product history (e.g. stability data – statistical analyses
of particular use here)– Batch history (e.g. review indicates that there were no
processing errors / there were errors)– Other investigational findings
PCI Pharmaceutical Consulting Israel Ltd 61/82
And now some work for you.....
Specification is 90.0 – 110.0% for an oralsuspensionThree batches are tested simultaneously, eachsample in duplicateTwo batches show (average) 98.2% and 99.7%respectivelyThird batch gives one result of 88.2% and asecond result of 89.3%
PREPARE A CHECKLIST OF QUESTIONSTO REVIEW AS PART OF THE LABORATORY INVESTIGATION
PCI Pharmaceutical Consulting Israel Ltd 62/82
FDA: Full-Scale OOS Investigation
Use a pre-defined procedure Production/ process review and / or additional
laboratory work Identify root cause and implement CAPA QA / QCU responsibility, includes CMOs if used Documented in the batch record Involves all aspects of manufacture, quality
control and sampling Describes corrective actions and endpoint Is performed PRIOR to ANY retesting
PCI Pharmaceutical Consulting Israel Ltd 63/82
FDA: Full-Scale OOS Investigation
If cause of OOS is identified, batch is rejected In this case need CAPA on process / product May not identify cause and may need additional
lab testing:– Retest additional portion of original sample– Resample
PCI Pharmaceutical Consulting Israel Ltd 64/82
FDA: Retesting and Resampling
Use another analyst? Where possibleThe maximum number of retests should be
specified in advance in an SOPMay, on rare occasions, deviate from SOP but
with documented rationale and protocolThe number may vary depending upon the
variability of the test method and NOTdepending on the results obtained
Resampling raises questions as to samplingprocedure validity
PCI Pharmaceutical Consulting Israel Ltd 65/82
Interpretation and Results Reporting
Report all results, passing and suspect andconsider for batch release decision
Averaging– Has appropriate and inappropriate uses:
• Appropriate: optical rotation test: several discretemeasurements averaged for result; microbiology;HPLC average of 2 or 3 peak responses fromreplicate injections of the same preparation = onetest and one result as described in the test methodHave acceptance criteria for deviation between replicates
This is different to analysis of different portions ofa single batch
PCI Pharmaceutical Consulting Israel Ltd 66/82
Interpretation and Results Reporting
Averaging– Inappropriate uses:
• May conceal variations in different portions of thebatch or samplee.g. for powder blend uniformity or dosage formuniformity of content.
• Averaging the results of the original OOS andadditional retest or resample results is inappropriate
• All individual results must be provided andconsidered by the QCU / QA for release
67
Time for a Break
PCI Pharmaceutical Consulting Israel Ltd 68/82
Outlier Tests
Possible use of the outlier test should bedetermined in advance
Should be written in an SOP for datainterpretation
Should include the specific test to be usedShould specify the minimum number of results
required to obtain a statistically significantassessment from the test
For validated chemical tests its use is suspect
PCI Pharmaceutical Consulting Israel Ltd 69/82
FDA: Concluding the Investigation
Evaluate results and determine batch qualityRelease decision by QCU / QAAn initial OOS does not necessarily mean
that the batch fails and must be rejected.Where the suspect result is invalidated, the
result should not be used to evaluate thequality of the batch or lot
For inconclusive results – give fullconsideration to the OOS result
PCI Pharmaceutical Consulting Israel Ltd 70/82
FDA: Concluding the Investigation
Example given shows seven retest resultswhich gives the only indication in the guideregarding numbers of retests
The example given is also extreme:89.5% OOS99.0, 98.9, 99.0, 99.1, 98.8, 99.1, 99.0%
Consider method precision and validation datain making release / reject decision
PCI Pharmaceutical Consulting Israel Ltd 71/82
FDA: Cautions
Where a series of assay results (to producea single reportable result) have some individual results OOS and some in spec andall within the known method variability,passing results no more likely to representthe true value than the OOS result.
In this case the company must err on the side of caution and reject the batch
A result that is low but in specification shouldalso be a cause of concern
PCI Pharmaceutical Consulting Israel Ltd 72/82
FDA on Field Alert Reports (FARs)
For products with approved (a)/NDAs,regulations require a FAR within 3 workingdays concerning any failure of a distributedbatch to meet any of the specificationsestablished in an application
Unless the OOS result is found to be invalidwithin 3 days, an initial FAR should besubmitted
A follow-up FAR should be submitted whenthe investigation is completed
PCI Pharmaceutical Consulting Israel Ltd 73/82
OOS Case Study #3
BN 200602 gave an assay result for 3 month stability study: 89.2% (long term / RT)
Limits: 90.0 – 110.0%Test performed at a contract laboratory using
an internal instrument controlBN 200701 tested at the same time gave a
result of 95.4% (i.e. in spec. – release test)
DO YOU INFORM THE REGULATORS?
PCI Pharmaceutical Consulting Israel Ltd 74/82
OOS Case Study #3 – cont/
The samples have to be diluted duringpreparation and as far as records showedthere was no evidence of laboratory error
Calculations were satisfactoryNo evidence that reagents were outdated or
faulty techniqueControl sample showed a 3.5% difference at
beginning and end of run: usually around 1%;NMT 5% allowed by method
PCI Pharmaceutical Consulting Israel Ltd 75/82
OOS Case Study #3 – cont/
Batch number 200701; in-process result: 98.2%Batch number 200602; time zero result: 92.8%
i.e. an apparent difference of 3.5% betweenresults in both cases
Hypothesis formulated:problem is with the control sample / equipment and this can be tested using both batchesi.e. in-spec and OOS
PCI Pharmaceutical Consulting Israel Ltd 76/82
OOS Case Study #3 – cont/
Retesting Protocol (before start of retesting):– BN 200602: OOS on inverted stability sample
at 3 monthsPerform retest at three dilutions in duplicateon:
• Same sample = 6 results vs 2 original• Upright sample = 6 additional results
– BN 200701: In spec. but 3.5% lower than IPCPerform retest at three dilutions in duplicate
• Same sample = 6 results vs 2 original
PCI Pharmaceutical Consulting Israel Ltd 77/82
OOS Case Study #3 – cont/
Outcome:
– BN 200602: All retest results in specification
average result: 92.2%– BN 200701: All retest results in specification
average result: 97.9%– Control sample: difference of 1% beginning and
end of run
PCI Pharmaceutical Consulting Israel Ltd 78/82
OOS Case Study #3 – Conclusion
Original OOS result for BN 200602 is invalidOriginal OOT result for BN 200701 is invalid ?CAPA required regarding external laboratory
as follows:– Tighten allowed limits for control sample
(NMT 5% is too high)– Revalidate method? Investigation showed
validation last done 15 years ago and a newinstrument had been introduced since then!
– Closely follow BN 200602 at additional stability stations
PCI Pharmaceutical Consulting Israel Ltd 79/82
OOS Logs and Trending Data
Should keep a record of OOS resultsAnalyse periodically according to:
– Analyst– Instrument– Product– Test method
Take appropriate actions where repeatproblems appear
PCI Pharmaceutical Consulting Israel Ltd 80/82
And in Conclusion....
Always report OOS or OOT result toSupervisor
Have a pre-approved SOP that provides aflow-chart for handling OOS
Conduct detailed and genuine investigationDocument investigational findings clearly and
conciselyTransparency regarding reporting on COAAt some point...notification of regulators
PCI Pharmaceutical Consulting Israel Ltd 81/82
In conclusion
All test results are subject to some element of doubt Use controls and validated, approved test methods Use replicates where inherent variation exists Use qualified, calibrated and well maintained
equipment Perform trend analyses and report deviations NEVER continue a suspect test At some point, testing ends and a product disposition
decision must be made
PCI Pharmaceutical Consulting Israel Ltd 82/82
Thank you for your attention
Any questions?
Find me at [email protected]