Validation Good Practice, Long But Perfect. Recommended !

ECA Validation Good Practice Guide
Table of Contents
1.0 Introduction
2.0 A new view on Validation: The move to process

understanding
3.0 Examples: How to do
3.1 Risk-based Approach to Process Validation

3.2 Risk-based Qualification
3.3 Statistics in Process Validation and Continued
Process Verification
3.4 Process Validation and Statistical Trending in
Biopharmaceutical Manufacturing
4.0 Special Case: Legacy Processes
5.0 Statistical Background / Manufacturing Excellence
5.1 Statistical Process Control (SPC)
Annexes
Annex 1 Literature
Annex 2 Survey on the Annex 15 Revision
Annex 3 Interpretation of the FDA Guidance on Process
Validation
Annex 4 Interpretation of the EMA Guideline on Process
Validation
Annex 5 Interpretation of the Annex 15 Revision
1
© European Compliance Academy

2
1.0 Introduction
The topic Qualification & Validation has come to the centre of
attention of industry and authorities. Although the concept
had already existed for many years, a discussion has started
about a “modern approach” to qualify equipment and to vali-
date processes.
Especially in the field of process validation the change is evi-

dent. Both EU and FDA regulators have questioned the famous
“3 batch approach”. There is now a common agreement that
process understanding is the prerequisite for process valida-
tion. However, the concrete steps are not clear yet. FDA pub-
lished a new Guidance on process validation 2011 and the
revised Annex 15 is valid since October, 1st 2015.
As there are different validation approaches possible in the EU

and no practical examples are explained in the US Guidance,
concrete steps are not clear. The ECA has therefore set up an
Expert working group to develop a Good Practice Guide. With
the Annex 15 revision there was a need to implement its new
requirements in this Good Practice Guide as a 2nd edition. This
document is intended to provide support to both regulators
and industry. On one hand the Guide should contain the main
elements of a new approach (“what to do”) but on the other
hand it should also serve as supporting guide for the imple-
mentation (“how to do”).
Members of the ECA Expert Working Group:

Richard Bonner, ECA Chairman
Dr Jean-Denis Mallet, Head of ECA´s Validation Group
Dr Thomas Schneppe, Bayer Pharma AG
Gert Moelgaard, NNE Pharmaplan
Dr Renate Schenk-Gröninger, Boehringer Ingelheim Pharma
GmbH & Co. KG
Sven Pommeranz, Concept Heidelberg GmbH

3

4
2.0 A new view on Validation:

The move to process understanding
Sven Pommeranz, Concept Heidelberg, Heidelberg
2.1 Validation in practice from the past to today
The issue of validation with a focus on sterilisation processes

found broader public interest following the FDA draft1 of an
“upgrade” of CGMP regulations in 1976. In the early eighties,
the FDA also attributed its validation focus on other processes
instead of only sterilisation2. The FDA defined the line of ap-
proach for the companies it inspected with the first publication
of the draft of the FDA Guideline on Process Validation in
1983. Initial interpretations by the industry3,4 followed in the
USA. In Europe, the FIP Guideline represented the starting
point for this topic5 also following in Europe in the early eight-
ies. Publication of the EU Guideline to Good Manufacturing
Practice6 late in the eighties and its enforcement from 1992
rendered the activities of validation and qualification relevant
throughout the entire EU.
Whereas several pharmaceutical companies in the early nine-

ties had a lack of knowledge or even ignorance concerning
validation and especially qualification operations, knowledge
had increased extensively by the end of the century. This was
due to the increased experience gained by regulatory inspec-
tions in the respective companies as well as some professional
articles and additional guidelines with an interpretative charac-
ter such as the ISPE Baseline Commissioning and Qualifica-
tion7. This development was also promoted by pragmatic tem-
plates8,9.
Many suppliers of equipment, engineering and consulting to

5
the pharmaceutical industry followed suit – in some cases with
a delay of several years – and offered qualification documents
and services when selling equipment or delivering facility pro-
jects. These qualification documents had different levels of
quality and therefore 3rd party service-providers were increas-
ingly engaged into validation and qualification operations10.
Not least as a reaction to the so-called Barr ruling11, the FDA
put the focus of its investigative inspections during the nine-
ties on validation. This could clearly be seen in the analysis of
warning letters12. As a consequence, validation and qualifica-
tion operations often became almost self-serving in the sense
that they were both managed and executed by service provid-
er companies. Gradually commissioning and qualification activ-
ities increased significantly to avoid the risk of not complying
with the regulations or the regulatory inspection observations.
It became a common assertion that “This is the way the FDA
wants it” that often stifled some criticism of these increasingly
expensive and time-consuming qualification activities.
The original idea that quality management should contribute

to drug efficacy, safety and quality via – among others - the
qualification and validation operations was often forgotten.
Many companies qualified and validated nearly everything. As
an example, seminars on validation could go deeply into the
best qualification strategy for magnetic stirrers or the number
of signatures on a validation document without considering
the actual process performance.
But basically process validation should demonstrate that the

design, implementation, procedures and processes used for
manufacturing drugs are fit for their intended use and ensures
the corresponding quality of the pharmaceutical product. Pro-
cess validation should examine the conditions and especially
the critical parameters that can influence the quality of the
6
product and the robustness of the process. Process validation
should not compensate for a missing or insufficient process
development or process optimisation. It is the task of process
validation to prove the suitability of the process design and not
to develop or improve the design.
In short, process validation was often mainly a documentation

exercise rather than a process and product quality demonstra-
tion.
The common approach to process validation has been the

‘three validation batch’ approach, where three product batches
were produced on the basis of fixed and mostly average pa-
rameters without realising the risks and limits of the process13.
In some cases, validation was regarded as optimisation – with
the resulting process failures during routine operation. In or-
der to avoid the very restrictive regulatory change manage-
ment in Europe14,15,16 as well as in the USA17,18,19,20 further
technical developments or developments concerning process
optimisation became a rare exception in industry practices.
A few voices challenging the common industry approach was

raised in some conferences and articles in the nineties. As ear-
ly as 1993, Jim Akers asked “Validation – a cult-like Activity?”21
and he had hoped to lead validation back to the essential pro-
cesses. Now the benefit-cost effect of validation and therefore
qualification activities as well were discussed: “Validation –
How Much Can the World Afford? Are We Getting Value for
Money”22 and “Can Validation Improve the Bottom Line?”23,
the latter with a special reference to computer validation, were
two exemplary articles on this topic. Increasingly, the cost
topic became public, too. The author of an article24 in 1998
said that the validation costs are equal to 10% of the acquisi-
tion costs of equipment. According to this estimates, maintain-
7
ing the validated status in running production equals 1% of
annual acquisition costs. Depending on the facility qualifica-
tion, costs of 3 – 30% based on the order value are referred
to.25
With the introduction of the concept of Good Engineering

Practice combined with an impact assessment, the ISPE Base-
line “Qualification and Commissioning” marked a new way of
thinking, although the PIC-document PH 1/96 (today PIC/S PI
006) mentions explicitly engineering activities for IQ and OQ
and also a view on risks. But in the pharmaceutical industry IQ
and OQ activities were QA driven at that time. This change
was also due to the EU GMP Annex 15 on Qualification and
Validation which encouraged a risk assessment in order “to
determine the scope and extent of validation” and to the FDA’s
risk-based approach of the “cGMP for the 21st Century” initia-
tive. In lectures by renowned FDA officials (such as the 2005
lecture at ECA´s 2nd European GMP conference by David Hor-
owitz, then head of the Office of Compliance), the significance
of the typical three validation runs was questioned.
The pharmaceutical industry also demanded a change of

thinking: “Validation is not a religious experience or FDA wants
McDonalds and we give them Cordon Bleu”, was the essence
of one article 26, for example. With special reference to qualifi-
cation, two industry guidelines were dedicated to the topic of
“excessive” qualification activities27,28. ISPE’s White Paper is
primarily future-oriented and has become part of a new ISPE
Guide on a science- and risk based verification as an alterna-
tive to the Commissioning and Qualification Baseline Guide
from 2003. The new approach is based on an ASTM standard
from 2007 nowadays called E2500-12 (“Standard Guide for
Specification, Design, and Verification of Pharmaceutical and
Biopharmaceutical Manufacturing Systems and Equipment”)
8
that also has already been integrated in GAMP529. The tenor of
both documents is to focus qualification activities based on an
appropriate risk analysis, whereby the focus is mainly on the
critical aspects of the manufacturing system and the process.
Very detailed explained is that in an additional ISPE Guide
(“Science and Risk-Based Approach for the Delivery of Facili-
ties, Systems, and Equipment” from 2011, see below which
gives an interpretation of ASTM E 2500.
As part of the cGMP for the 21st Century Initiative, FDA also
started a Process Analytical Technology (PAT) Initiative30.
Shortly afterwards, the European EMA followed suit31 and it
became part of a fairly harmonised approach where focus on
“process understanding and control” became the main topic.
This new development was complemented by the Compliance
Policy Guide 7132c.08, which no longer mentioned a defined
number of validation runs und started to demand continuous
validation. Then, in November 2008, the FDA published a
long-awaited Draft Guidance on Process Validation constituting
what has become the current thinking of FDA and continuing
the path pursued by the Compliance Guide right into the final
version of FDA Process Validation Guidance from 2011, which
is now the regulatory expectation. The focus on product and
process understanding now has become the centre of interest
with a strong emphasis on scientific tools, including statistics.
The FDA now defines validation as a Life Cycle Process in 3

stages:
1. Process Design
2. Process Qualification
3. Continued Process Verification
In parallel to this, the International Conference on Harmoniza-
tion (ICH) , where the main drivers are Europe, USA and Ja-
9
pan had developed three new guidelines in the new millenni-
um: ICH Q8 (Product Development), ICH Q932 (Quality Risk
Management) and ICH Q1033 (Pharmaceutical Quality System,
now part III of the EU GMP Guideline). This three ICH guide-
lines were also one of the reasons for the annex 15 revision.
From a qualification and validation perspective this has be-

come a “light on the horizon”. With the combination of these
three guidelines, followed by the ICH Q1134 , there is a com-
mon understanding in Europe, USA and Japan of product and
process understanding as the new FDA Process Validation
Guidance (and EMA1 initiatives as well, see below) has made
even more clear. It aims towards breaking up the gap often
seen between the GxP systems in development and produc-
tion. Now a continuous improvement process is expected. New
technologies can be approached more offensively and the fear
of an inspection risk from insufficient qualification is replaced
by a much stronger emphasis on demonstrating product and
process understanding and -control.
This is part of the new, harmonised approach and also in Eu-

rope the EMA is turning the focus towards process under-
standing. Under the umbrella of its PAT initiative EMA pub-
lished in 2006 a Q&A Paper that says that enhanced process
understanding and monitoring could replace 3 PQ-Batches as
“continuous validation” 35. Also EMA now thinks in terms of a
validation life cycle (process design, process qualification, con-
tinued process verification) based on ICH Q8, Q9 and Q10
principles36. In 2010 EMA announced a Concept Paper to re-
vise their Note for Guidance on Process Validation and to im-
1
EMA is the abbreviation for European Medicines Agency. The Agency is responsible
for the scientific evaluation of medicines developed by pharmaceutical companies for
use in the European Union The former Name was EMEA, so some links lead to docu-
ments announced under the name EMEA
10
plement highlights of ICH Q8, Q9 and Q1037. A Draft Guideline
has been published in 201238. The final version39 is valid since
2014 (see also Annex 4). EMA has planned to update the An-
nex 15 of EU GMP Guideline regarding ICH Q8, Q9 and Q1040.
And since October 1st 2015 a revised Annex 1541 is valid, see
also Annex 5 in this Guide.
In parallel to this, the process validation in ISO 900142 and

1348543 validation is increasingly demanded by the medical
devices industry, too. A separate Process Validation Guidance44
of the Global Harmonisation Task Force (GHTF), nowadays
available via IMDRF45 offers help with implementation, espe-
cially for the medical devices industry. As the FDA’s new Guid-
ance points out in a footnote, this Guidance is also worth
reading by drug manufacturers since it puts an enhanced fo-
cus on statistics.
In summary, Fig. 1 shows an overview of the validation history

until 2015.

11
Fig. 1 Validation History, Dr Michael Hiob, Ministry of Employment,

Social Affairs and Health of Schleswig-Holstein, Kiel, Germany
2.2 Qualification/Validation is the focus of authorities
What is the regulatory authority’s point of view concerning

qualification and validation? This question can be answered
relatively simply for the FDA-regulated sector by analysing the
warning letters. Process validation always ranked top regard-
ing Warning Letter analysis.
In Europe it is more difficult to obtain data due to the more

complex system of national health care authorities. There are
no current statistics available, But in 2007, EMA published a
ten-year analysis (1995-2005) of GMP deficiencies in the con-
text of inspections of centralised marketing authorisations46.
12
With reference to all GMP deficiencies, process validation
ranks at 9 and equipment validation at 12. If the deficiencies
are listed according to their criticality however, process valida-
tion already ranks 5th in terms of critical deficiencies and
equipment validation comes in 9th. Another striking point is
represented by the large number of deficiencies concerning
the design of premises and equipment. This is an indication
that in Europe design qualification was asked in inspections
und might be a topic in future inspections.
With an incidence of 10.7%, process validation ranks 3rd of

the critical API GMP deficiencies among UK manufacturers
according to the results of inspections during the years 2004 –
2009 published by the British regulatory authority MHRA47.
In the inspection results48 of the Austrian regulatory authority

AGES PharmMed for the year 2007, qualification ranks 2nd
(29%) among the five most common deficiencies in manufac-
turing companies and process validation ranks 3rd (14%) on
the deficiency list while deficiencies concerning instructions on
maintenance, calibration and equipment cleaning and mainte-
nance come in 4th (13%) in this analysis.
2.3 Qualification: Current Industry Practice
In what way could a qualification be carried out today which is

compliant to GMP and nevertheless efficient? Matthias Klein,
CSL-Behring, has made some proposals49. He recognises that
the key to success lies in the efficient use of already existing
documents and processes (including change control and cali-
bration systems) in pharmaceutical companies. Furthermore,
he advises a structured process with control of time, resources
and costs. Expenditure minimisation is also possible by con-
sistently applying Good Engineering Practice according to the
13
ISPE Baseline Guide on Commissioning and Qualification in
terms of an integrated qualification concept. This signifies the
preferably extensive use of technical implementation docu-
ments also as test documents. Initial approaches already took
place in this sector within the validation of computerised sys-
tems according to GAMP at the turn of the millennium.50 In
this context, especially documents concerning the factory ac-
ceptance test (FAT) are mentioned or functional testing and
acceptance inspections at the contractor’s site by the contrac-
tor and the contractee, the so-called site acceptance test
(SAT); see Fig. 2.
Fig. 2 Including GEP documents in qualification, M. Klein, CSL-

Behring GmbH, Marburg, Germany
In this context it is important that FAT and especially SAT are

carried out in compliance with GMP (trained personnel, docu-
mentation in a timely manner, existence of acceptance crite-
14
ECA Valid
dation Goo
od Practice
e Guide
ria) and that they are furthermorre documen nted in comp pliance
with GMP ((permanentt records, integrity of d dates etc.). The
acceptance e of SAT and FAT, if it is documen nted well annd was
performed according GMP,
G was already
a acceepted by Euuropean
inspectors since yearss. This is exactly what tthe new Annnex 15
now descriibes. Furthe ermore, thiss process sh
hould at least be
described iin the Validaation Maste
er Plan.
Another ke ey to successs is the inte

egrated quaalification between
engineering and qualification by mixed team ms consisting of
engineers a ers, in orderr to avoid d uplication of
and qualifie o work,
(see Fig. 3). This prop posal was already mad e in 1998 in n an
article as sstreamliningg measure to reduce tim me and costs51.
Fig. 3: Integ
grated qualiffication conc
cept, M. Kleiin, CSL-Behrring
GmbH, Marburg, Germa any

15
The most important step towards an efficient and pragmatic
qualification is the appropriate use of risk analysis as part of a
risk management process. Although first articles on the use of
FMEA as an example of risk analysis in the context of valida-
tion52,53 had already been published during the mid-nineties,
only EU GMP Annex 15 made this topic relevant for the au-
thorities and has with its revision expand risk analysis re-
quirements. Here, too, the initial uncertainty as to the meth-
ods and extent of the risk analysis gave way to the meanwhile
practically established procedures in the companies. To this
effect, ICH Q9 with Annex 1 as a tool collection contributed
immensely. If in the beginning of the use of risk analysis strict
adherence to the methods was widespread, ICH Q9 demon-
strated the variety of possible methods right up to the informal
risk analysis. The ISPE-Baseline Guide on "Commissioning and
Qualification" with its direct/indirect-impact approach was also
of help. Predefined and structured risk analysis in the form of
guidelines or SOPs for standard equipment further reduces the
qualification effort. Here many qualification parameters always
remain the same (such as distribution and penetration of heat
in the context of OQ and PQ during sterilisation), see figure 4.

16
ECA Valid
dation Goo
od Practice
e Guide
Fig. 4: Spec
cified document structure for drawin
ng up qualific
cation
documents,, M. Klein, CL
LS-Behring GmbH,
G Marb urg, Germanny
Based on tthe product requiremen nts the risk analysis is often

carried outt in several stages toda ay. A generaal risk analy ysis is
the basis fo or DQ; the overall facillity is divideed into partss and
the GMP re elevant partts are identified (e.g. b by an impacct as-
sessment a according to
o ISPE). The en the com mponents rellevant
for qualificcation are id
dentified andd the exten nt of qualificcation is
determined d by a detail risk analysis. The riskk analysis should
also be furrther considered in the course of q qualification
n and
beyond rou utine operattion; see Fig
g. 5.

17
ECA Valid
dation Goo
od Practice
e Guide
Fig. 5: Risk analysis in the

t life cycle
e approach o
of validation,, modi-
fied accordiing ZLG Aidee memoire Qualification/ /Validation
In the pastt there was extensive uncertainty

u concerning
g design
54
qualificatio
on . The PIC/S docume ent did not elaborate the
t DQ
in a more d detailed ma anner, eitheer. In the orriginal FDA guide-
g
line, this te
erm does no ot exist at all.
a The ISPE E-Baseline was
w the
first to sho
ow the adva antages of a DQ (called d Enhanced Design
Review as a distinction from med dical devicess). In Annex x 15,
DQ is direcctly incorporrated into the life cyclee of the quaalifica-
tion. A “sleender DQ” can
c be another measurre towards
“streamline e validation”. At least in the past, the comparison
between user requirem ment speciffication and functional design
specificatioon was often seen as something sstatic and ch hanges
have afterw wards been n maintained d circumsta ntially. An integra-
tive approa ach makes it i easier. Th
he DQ includ des the who ole pro-
cess from d design to th
he beginning of qualificcation. It is not a
18
separate document but it is based on the existing “living” pro-
ject documentation (such as user requirement specifications
and functional design specifications). And the Annex 15 revi-
sion concrete DQ as stage 2. The qualification starts regarding
Annex 15 revision with User Requirement Specifications
(and/or Functional Design Specifications).
During the qualification project, change control can be han-

dled almost as pragmatically. A simplified change control pro-
cedure with document traceability facilitates the activities
compared to the standard change control during routine oper-
ation. Approvals in the project change control are then „only”
effected among the project team. This is taken one step fur-
ther by the ASTM document 2500 which introduces a com-
bined effort of Design Review and Risk Management in order
to not only identify risk but also mitigate risk aspects by
changes to the design in an iterative process during a project
(see below).
Rationales for qualification tests are increasingly questioned in

the context of inspections and audits where the integrative
concept and an appropriate traceability matrix can ensure that
only those tests that come from the risk analysis are carried
out. With the traceability matrix, the direct reference of a test
coming from the risk analysis is guaranteed from the qualifica-
tion protocol to the qualification report; see Fig. 6.

19
ECA Valid
dation Goo
od Practice
e Guide
Fig.6: Extra
act from a Trraceability Matrix, Dr Micchael Hiob, Ministry
M
of Employm ment, Social Affairs
A and Health
H of Sch
hleswig-Hols stein,
Kiel, Germaany
The topic rre-qualificattion is now a own chap pter in the Annex

15 revision
n. A periodiccal evaluatioon of premiises, system ms,
equipmentt and utilitiees concerninng the quali fication stattus is
required. AAlso small changes sho ould be asseessed. Requ ualifica-
tion cycles should be based on a rationale. FFurthermore e, with
the introduuction of PQQR in the EUU-GMP Guid deline proceess-
related datta and the qualification
q n status of eequipment emerge
e
which enab bles a statement conce erning contiinuous suita ability55.
In the US tthe Annual Product Rev view (21 CFFR 211.18056) re-
quires at le
east annually an evalua ation of thee quality standards
of each druug product.
The new FFDA Guidancce “Process Validation”” consequen ntly no

longer contains the teerm revalidaation. In thee sense of the Life
Cycle Apprroach the toopic is conta
ained in stagge 3 Contin nued
Process Ve
erification. The
T target iss to show in n this third stage
20
that the process remains in a validated status even during
routine production. For this purpose, a system is requested
that detects unplanned process deviations. Shifts have to be
assessed accordingly so that the process does not get out of
control. Here a direct reference is made to the Annual Product
Review pursuant to 21 CFR 211.180) in order to support this
on-going programme. And great importance is attached to
statistics. Also the revised Annex 15 requires now continued
process verification stage with regard to statistics, in the EU
called ongoing process verification. And consequently the term
revalidation is also deleted in the Annex 15 revision.
And the future?

Two documents, a White Paper by the ISPE and an ASTM
(American Society of Testing and Materials) Standard (E2500)
offer some proposals. The ISPE White Paper (which set the
direction of the ASTM E2500 standard) targets a "value- added
qualification”. In principle, this directs the focus of a pharma-
ceutical company towards the user requirement specifications
(URS) and the Process Qualification (PQ); see Fig. 7.

21
ECA Valid
dation Goo
od Practice
e Guide
Fig. 7: Simp
plified V-Mod
del with inte
egrative DQ, according to
o M.
Klein, CLS-B
Behring Gmb bH, Marburg, Germany
Thus IQ annd OQ are of

o less impo ortance. Inteerestingly, this
t
perception was alreaddy hinted att by an FDA A representa ative in
1994:”The extent of IQ and OQ is up to the pharmaceu utical
company tto decide. Itt is usually a relatively low priority
y in FDA
inspectionss”57.
The ISPE WWhite Paperr summarise es the „Prin

nciples for th
he 21st
Century Quualification””58 in ten po
oints that is an excellen
nt sum-
mary of the
e principless outlined inn the ASTM E2500 stan ndard
and the ISPE Guide on n Science and Risk Bassed Verification:

22
1. Focus on product quality: the product quality is to be
guaranteed by qualification activities. The user require-
ments specifications are mentioned expressly as most im-
portant element.
2. "Requirements": "User requirements" should be based on
the process. They are confirmed as being satisfied in the
PQ. This means that IQ and OQ are subordinate in im-
portance.
3. "Risk assessments": Process development and experi-
mental design are the key elements in order to identify
critical functions and parameters.
4. Only critical process parameters will be used as the basis
for qualification.
5. All activities must contribute value to the process ("We
won't do anything just for the sake of regulatory compli-
ance").
6. "Risk based asset delivery": Qualification activities should
be co-ordinated with the facility’s complexity (no use of
“cook books”). The GAMP classification is mentioned ex-
pressly as an example.
7. "Value-added documents": The necessity for documenta-
tion is mentioned expressly, but only data which serves a
useful purpose should be collected. It should not be col-
lected to merely fulfil "some imagined regulatory expecta-
tions".
8. Use of supplier documentation: If possible supplier docu-
mentation (test plans etc.) should be used if it has been
assessed accordingly.
9. Tests: As a rule tests should only be carried out once. But
it is said expressly that PQ tests may require additional ac-
tivities. It may be necessary to repeat some tests which
have been carried out in an earlier stage of development.

23
10. Fostering innovation: At this point, the flexibility of the
qualification programs is called for in order to be able to
implement new trends.
All this would require great organisational changes in the

companies (the authors consider this to be the greatest obsta-
cle for success), which often have relatively rigid qualification
and validation concepts. Quality Management wouldn’t any
more inspect each individual IOQ protocol. Instead, it would
focus more on processes and risk analyses. The Quality Man-
agement Department would be the umbrella to oversee and
approve risk analyses, qualification plans and qualification re-
ports. On the other hand, it would also be necessary to
strengthen the quality systems of the special departments
(e.g. engineering, automation/IT etc.) and production. Risk for
the product quality would be determined in the context of a
risk disclosure on the basis of process understanding. The con-
trol of those risks would then be qualified in a comprehensive
PQ.
Ultimately, the ASTM E2500 Standard Guide propagates an

approach which goes from requirement definitions through
specification and design and verification to acceptance and
release. This process takes place under the umbrella of the
Good Engineering Practice and on the base of risk manage-
ment, change management and design review. Input to the
process is product knowledge, process knowledge, regulatory
requirements and company quality requirements. The ex-
pected output of this process is operation & continuous im-
provement. The model is similar to the quality management
process of ISO 9001-08.

24
Interestingly, ASTM E2500 compares the status of the final
Standard with regulatory guidelines. It mentions the US Food
& Drug Administration (FDA) which participated in the devel-
opment of the document. There are no comments on the
Guide from European authorities though. However, it is still a
question how European authorities regards a consistent appli-
cation of ASTM E2500 Guide. An ISPE Guide “Science and
Risk-Based Approach for the Delivery of Facilities, Systems,
and Equipment” from 2011 gives in an Appendix hints how the
“traditional” IQ,OQ,PQ concept fits in the verification model of
ASTM E2500.
Moreover, the concept is also considered as usable for com-

puter validation. This incorporation has already been carried
out to a great extent by GAMP 5. Here the use of the tradi-
tional terms IQ/OQ/PQ is purely optional. This means GAMP 5
departs from the strict delineation of phases in the V-model59
that has always been part of the GAMP approach. It will be
interesting to see how the European authorities assess the
implementation. The revised Annex 11 of the EU GMP Guide-
line gives some hints.
With the new Process Validation Guidance by FDA the topic of

qualification also changes for the FDA. Now qualification activi-
ties are part (“Design of a Facility and Qualification of Utilities
and Equipment”) of Stage 2 of the process validation life cycle
of process qualification. The terms DQ, IQ and OQ are no
longer used in the document, and the FDA has made public
that this is on purpose because they never were a formal FDA
requirement as it was used by industry. This is consistent with
the ASTM E2500 Standard Guide and GAMP 5. Both docu-
ments refrain from using these terms and substitute them with
“verification”. In the Process Validation Guidance the FDA also
uses the term verification with regard to the activities concern-
25
ing installation and function testing. The intent is to make it
clear that the industry practices of DQ, IQ and OQ etc. is not a
regulatory requirement and to encourage a stronger applica-
tion of Good Engineering Practices (GEP). Also the Annex 15
revision goes in the GEP direction and has implemented FAT
and SAT. But the Annex 15 revision still mentions qualification
stages as DQ, IQ, OQ, PQ although they are not mandatory.
The FDA Process Validation Guidance only offers rather gen-

eral descriptions of the content of protocols in the Guidance
called plans for qualification activities but very detailed out-
lines the Process Performance Qualification (PPQ) protocol and
report. The PPQ batches need to be carried out by the produc-
tion employees under normal conditions – they are manufac-
tured under commercial manufacturing conditions and may
ultimately be released as normal commercial batches, depend-
ing on a successful overall Process Qualification result.
This also means that the worst-case scenario within the

framework of validation runs is no longer a special issue. For
the FDA the PPQ combines the qualified premises and facilities
as well as the trained personnel with the commercial manufac-
turing process. The FDA considers PPQ as an important mile-
stone in the product life cycle: “A successful PPQ will confirm
the process design and demonstrate that the commercial
manufacturing process performs as expected”. Hence the
completion of PQ before commercialisation is mandatory.
The FDA sees PPQ in the sense of the PIC/S-document PI 006.

It has equated PQ and process validation since 1996. In the
revised Annex 15, PQ and Process Validation can combined
together. This goes in the same direction as in the FDA guid-
ance. But the Annex 15 revision gives the option of three vali-
dation approaches
26
x An traditional approach (where the “magic 3 runs” are

still mentioned)
x Continuous process verification as described in ICH Q 8
and
x An Hybrid approach as a mixture of the both above
mentioned approaches.
In the US Process Validation Guidance only one approach is

mentioned and the validation life cycle is more emphasized
than in the EU. This difference to the EU Annex 15 revision
and might cause some confusion.

27

28
2.4 Process Validation: A new Approach
2.4.1 Summary of the new FDA Process Validation

Guide
It is of no surprise that the new guidance does not mention a

fixed number of validation runs to prove the process validity.
This became clear in the Compliance Guide on Validation pub-
lished in 2004. The new guidance relies on the 3-stage life
cycle model. During Stage 1 and 2 manufacturers gain a lot of
data. With this data it can be demonstrate that the process is
capable (including conditions that pose a high risk of process
failure). The new key word is "process understanding". New
definitions for process validation and performance qualification
show the strong connection to "scientifically sound". Strong
emphasis is also placed on the use of statistical methods.
Based on a risk based approach, statistics should give

a scientific sound rationale, when a process is valid. It is based
on the same type of methods as used in e.g. six sigma, DoE,
process capability indexes (Cpk) and statistical process control.
Clearly IQ and OQ are not mentioned anymore, but the term
qualification is still used. However, when describing how, the
Process Validation guidance points to verification activities as a
basis for PPQ. The PPQ is now the key element of the process
validation life cycle and is meant to be carried out under nor-
mal commercial conditions. Within the framework of continued
process verification, both the topic of trending and mainte-
nance are rated highly.
By the way, Prospective Validation and Concurrent Validation

are no longer mentioned. You can only find some statements
on Concurrent Release, which addresses special situations
where batches could be distributed before an appropriate PQ
has been completed and concluded (e.g. orphan drugs). Also
both revalidation and retrospective validation are not men-

29
tioned any longer. Old processes should be evaluated towards
development experiences, qualification work and of course
routine production experience. So (re)validation of old pro-
cesses start with stage 3. A lot of ASTM-Guides are referenced
as state of the art. For details, see also Annex 3 in this Guide.
2.4.2 Summary of EMA’s Process Validation Guidance

EMAs Process Validation Guidance is a marketing authorisation
document, a fact which is clearly addressed already in the
title. It is valid only for medicinal products but not for legacy
products. Its applicability for active ingredients and biological
products is mentioned analogously as possible. The introduc-
tion of a validation life cycle and the integration of a continu-
ous process verification (CPV) are completely new although
this approach is already known from ICH Q8. The "traditional
approach" remains accepted and the magical number three is
mentioned for the validation runs. This aligns with the infor-
mation from the Annex 15. In this point a discrepancy remains
to the FDA Process Validation Guidance which doesn't contain
a number for the validation runs any more. Just as in the re-
vised Annex 15 the hybrid approach remains a little bit "nebu-
lous" in the final document. The aim to integrate modern ele-
ments from ICH Q8, Q10 (and Q11) into the document is
clearly noticeable. But there are strikingly few concrete refer-
ences to ICH Q9. The more detailed definition of the bracket-
ing approach in the glossary is laudable. Bracketing is known
in the context of process validation also from the USA. But a
closer overlap with the FDA Guidance would have been desir-
able, nevertheless. The FDA Guidance addresses also APIs and
biological substances and the process validation lifecycle runs
like a red thread through the complete FDA document. And
the FDA Guidance contains GMP aspects. The FDA Guidance
refers explicitly also to legacy products which are to be inte-
grated in the lifecycle in step 3. There is another big differ-

30
ence. In contrast to the FDA Guidance the revised document
does not contain a strong emphasis on statistical methods.
(see also Annex 4).
2.4.3 Summary of revised Annex 15

The revision is very comprehensive. The influences of the
guidelines ICH Q8, 9 and 10 can be seen clearly, even in the
glossary. The alignment with the EMA Guideline on Process
Validation, revised in 2014, is also striking. Now the topic de-
sign space (ICH Q8) is included in the part on process valida-
tion. Now a lot of risk assessments (ICH Q9) are mandatory.
And the lifecycle approach and the topic process capabilities
(ICH Q10) are included, too. Deviation management has as-
sumed new importance. Third party services are allowed ex-
plicitly if the supplier has been qualified accordingly. This can
also be seen as an adaptation to reality. It is positive that the
preliminary release for the next step (such as qualification)
was mentioned for example in the case of deviations if there is
a documented assessment that there is no significant impact
on the next activity. It was not implemented a clear separation
between qualification (with regard to facilities and equipment)
and validation (with regard to processes).
New concepts appear without being defined in the glossary,

such as FAT and SAT or a functional specification. Retrospec-
tive validation and the term (periodic) revalidation were
dropped completely. But the effectiveness of the manual
cleaning processes should be confirmed at a justified frequen-
cy. Is this not also a sort of periodic revalidation?

31
The inclusion of user requirements as separate step and the
mention of FAT and SAT - even if merely as could requirement
- have made qualification more extensive. FAT and SAT are
typical elements of Good Engineering Practice (GEP). Here, the
link between GMP and GEP is missing in the document. Can
other GEP elements now also be applied (without problems) in
the GMP environment? Will even more GEP elements become
mandatory in the future? The greater flexibility as concerns
the qualification steps IQ/OQ which now can be carried out
together, is a positive element. But this has already been done
frequently by industry. There are no in-depth references to
alternatives for the qualification, such as ASTM E2500 alt-
hough the main qualification steps are merely a could option.
Unfortunately, information on the qualification of legacy facili-
ties were dropped completely. It happens again and again that
non-GMP environments are becoming GMP environments. May
these environments still qualify their old equipment and if so,
how? The reference to transport verification, packaging valida-
tion, validation of utilities and validation of analytical methods
comes as a surprise. Other regulations are more concrete
(such as ICH Q2 (R1) as concerns the validation of analytical
methods). The transport verification would probably have bet-
ter been placed in the context of regulations on Good Distribu-
tion Practice. Until now Annex 15 has been a general guidance
to the topic validation/qualification, why has there been this
specification? Should not the logical consequence be that vali-
dation of the sterilisation process or media fills for example
must also be included?
Now, there are three different approaches for process valida-

tion, a (modern) continuous verification approach, a traditional
approach that is still based on the classical 3 validation runs
and a mixture between both, the hybrid approach. But process
robustness has to be established in any case. Despite the ac-
32
cordingly listed restrictions and the necessary justifications as
concerns the decision for the number 3, the statement that
generally at least three consecutive runs are regarded as ac-
ceptable is a clear difference to the FDA Guideline on Process
Validation. This Guideline doesn't define a number any more.
It is positive, that a bracketing approach may be used in justi-
fied cases. The possibility of a hybrid approach remains
somewhat unclear. This term is not defined in the glossary.
But it is mentioned that a change of the validation strategy
(from the traditional approach to a verification) is possible with
an increasing level of knowledge and understanding of the
process. Unfortunately, the possibility of a review in the course
of re-qualifications is not explicitly mentioned any more. But
could the required evaluation concerning re-qualification not
still be carried out as review, where applicable?
The chapter Change Control now refers very strongly to regu-

latory aspects. The requirement for an efficiency control after
implementation of a change is new. It has already been re-
quired in chapter 1 of the EU GMP Guidelines Part I. Due to
the introduction of the feed-back loop the term change man-
agement would actually have been better for this chapter
(see also Annex 5 in this Guide).
2.4.4 Conclusion
The future of qualification seems to lie in an even more inte-

grated qualification involving engineers and qualifiers, in a
structured project on the basis of risk analysis and involving as
many GEP documents as possible. Surely the use of software
tools will be much more widespread than today. The testing
of URS precepts in the PQ as proposed in the ISPE White Pa-
per will be the basis for the main qualification tests. It will be
exciting if and in what way the new model of “verification”
33
from the ASTM Standard E2500 will be able to replace the
classical qualification stages? Also the revised Annex 15 gives
more flexibility in qualification.
The future of validation will be exciting. It seems to be clear

that the focus lies on process understanding. Statistics will be
helpful to support these aspects. You can see this in the revi-
sion of chapter 1 of the EU GMP Guide. It will be interesting to
see how industry reacts to the FDA Guidance and the devel-
opments in the EU. Especially the discontinuation of the "mag-
ic 3" - even though long anticipated - will probably (have to?)
lead to new rationales in order to prove a validation. In Europe
the “traditional” approach will still be available as stated in
EMA’s Guideline on Process Validation and revised Annex 15.
This traditional approach is not mentioned in the FDA Guid-
ance anymore. Will this still be a gap between Europe and
USA? Also in ICH Q 7 (EU GMP Guideline Part II) “GMP for
APIs” the three batches approach is still mentioned. How
about legacy products (old products)? The FDA says it starts
with stage 3 “Continued Process Verification”. And in Europe?
EMA’s Process Validation Guidance covers not legacy prod-
ucts, because it’s a document tending to the Marketing Au-
thorisation. In the revised Annex 15 is no hint on legacy prod-
ucts although the link to PQR indicates that legacy products
are also in the focus of ongoing process verification. EMA´s
Process Validation Guidance addresses drug products, not
APIs. But also in both documents (EMA Guidance and revised
Annex 15) is a “Validation Life Cycle” mentioned. The EMA
focus is very intensively on “continuous process verification”
now, more than in the FDA Guidance. There is also a “mix-up”
between traditional and enhanced approach feasible (Hybrid-
approach), regarding EMA Guidance. The same is with the
revised Annex 15. At least the EMA Guidance mentions not to
define new requirements.The EMA Guidance and the revised
34
Annex 15 could also be used for APIs, but are not mandatory
for APIs.
Also the new definition of PQ in the FDA Guideline possibly

leads to irritations. Up till now, PQ was often seen as being
primarily related to equipment. The perspective changed with
the new expression PPQ as part of PQ (now known as Process
Performance Qualification).

35

36
3.0 Examples: How to do

3.1 Risk-based Approach to Process Validation
Dick Bonner
The publication of the updated annex 15 in March 2015, which

became effective on 1st October 2015, takes into account
changes to other sections of the EudraLex, Volume 4, Part I,
relationship to Part II, Annex 11, ICH Q8, Q9, Q10 and Q11,
QWP guidance on process validation, and changes in manufac-
turing technology.
The traditional approach to process validation is no longer a

3-batch, one off validation with re-validation required at pre-
determined time intervals.
Annex 15 states:
“A quality risk management approach should be applied
throughout the lifecycle of a medicinal product. As part of a
quality risk management system, decisions on the scope and
extent of qualification and validation should be based on a
justified and documented risk assessment of the facilities,
equipment, utilities and processes. Retrospective validation is
no longer considered an acceptable approach. Data supporting
qualification and/or validation studies which were obtained
from sources outside of the manufacturers own programmes
may be used provided that this approach has been justified
and that there is adequate assurance that controls were in
place throughout the acquisition of such data. “
Validation should take into account not only the life cycle of
the product and the process, but also the qualification of the
facilities, equipment and utilities as well.

37
Validation may be thought of in three distinct phases:
1. Process Design is the critical step where the commer-

cial process is defined during this stage based on
knowledge gained through development and scale-up
activities.
“For products developed by a quality by design ap-

proach, where it has been scientifically established dur-
ing development that the established control strategy
provides a high degree of assurance of product quality,
then continuous process verification can be used as an
alternative to traditional process validation. In the tra-
ditional approach, a number of batches of the finished
product are manufactured under routine conditions to
confirm reproducibility. The number of batches manu-
factured and the number of samples taken should be
based on quality risk management principles, allow the
normal range of variation and trends to be established
and provide sufficient data for evaluation. Each manu-
facturer must determine and justify the number of
batches necessary to demonstrate a high level of as-
surance that the process is capable of consistently de-
livering quality product. It is generally considered ac-
ceptable that a minimum of three consecutive batches
manufactured under routine conditions could constitute
a validation of the process. An alternative number of
batches may be justified taking into account whether
standard methods of manufacture are used and
whether similar products or processes are already used
at the site. An initial validation exercise with three
batches may need to be supplemented with further da-
ta obtained from subsequent batches as part of an on-
going process verification exercise.

38
A hybrid of the traditional approach and continuous
process verification could be used where there is a
substantial amount of product and process knowledge
and understanding which has been gained from manu-
facturing experience and
historical batch data.”
2. Process Qualification is only confirmation of operation

within the pre-agreed limits established.
3. Process Verification is the vehicle for continual im-

provement using statistical analysis to monitor the pro-
cess.
There may be no re-validation required unless something

changes because the emphasis is based on showing the pro-
cess is capable and in-control using statistical analysis. Ongo-
ing process verification is required which can be documented
and assessed on an ongoing basis and collated using the
Product Quality Review, usually annually.
The emphasis should be that validation is never completed but

is an ongoing process that requires a comprehensive under-
standing of the process originating from R&D to identify all
sources of variation and to review all the Critical Quality At-
tributes and Critical Control Points, identify what risk these
present, determine if the risk is acceptable and if not devise
means of mitigating, reducing or eliminating the risk until a
process is achieved that remains consistently in a state of vali-
dation.
There is a recognition that more knowledge is gained during

commercialization which may result in changes to the process,
equipment or materials to ensure continuous improvements
whilst maintaining control and capability.

39
ECA Valid
dation Goo
od Practice
e Guide
This 2015 release of annex
a 15 dooes not cha nge the conncepts
outlined in the ICH Q8 8, Q9 and Q10
Q documeents. A risk based
approach tto achievingg and mainttaining a va lidated proccess can
be achieveed using theese tools. Th
his can be g
graphically repre-
r
sented in tthe following
g diagram:
Hence the propose prrocess unit operations

o aand variable
e pa-
rameters n need to be identified annd studied aat the development
stage. Thiss leads to th
he identifica
ation of all ssources of variabil-
v
ity for each
h unit operaation and innput sourcess and allowss initial
acceptablee ranges to be calculate ed.
Selected process steps and variables can theen be used in rep-

resentativee models to identify critical operattional param
meters
and operatting ranges.. This in turrn leads to mmechanisms to
limit, contrrol or mitiga
ate the varia
ability based
d on the ex
xperi-
mental datta to arrive at an accep ptable risk wwhere somee input
variability ccan be accoommodated within the process but still
result in a consistent acceptable
a output.
When suffiicient inform mation is av
vailable at th
he developm
ment
40
phase to produce a consistent process this can be the starting
point to initiate a technology transfer to the production divi-
sion so a comprehensive analysis of the process at a produc-
tion scale can be carried out. This would normally entail statis-
tically based additional sampling to give added assurance that
the process can operate at this commercial scale (or at least at
>10% of commercial scale as long as all operating parameters
are the same as at full scale.) There should be monitoring of
all Critical Control Points and product characteristics and full
investigations carried out if any problems do occur so root
causes can be identified and corrective actions taken to re-
solve the issues.
Using Q8 and Q9 principles a design space can be created

after identifying the Critical Quality Attributes and Critical Pro-
cess Parameters that have been demonstrated to provide as-
surance of quality. Working within this design space is not
considered as a change, while any move out of the design
space is considered to be a change which may require regula-
tory approval. At the time of submission the design space is
proposed by the applicant and will be part of the normal regu-
latory approval process.

41
ECA Valid
dation Goo
od Practice
e Guide
In the abovve example e, if the para

ameters weere submitteed as
part of the
e registration
n process thhen any chaange that occurs
o
within the shaded are ea would not need a reegulatory submis-
sion allowing greater regulatory flexibility
f to
o continuoussly im-
prove the pprocess.
The key qu
uestion is ho
ow do you decide
d whicch Quality attrib-
a
utes and P
Process Paraameters are
e really criticcal?
To do this you need too identify each unit op

peration and d exam-
actor associated with th
ine each fa hat unit opeeration to ascertain
which facto
ors can crea
ate the mosst risk (to th
he process and/or
a

42
ECA Valid
dation Goo
od Practice
e Guide
the patientt) if they areen’t adequaately controolled .
There are many techn niques to asssist in deteermining risk
k. In
this instancce I will use
e Failure Moode Effects Analysis (FM MEA) to
assess eacch part of an n operation and this wiill arrive at a rela-
tive risk sccore. The higgher score equates to a higher rissk. An
example off this is sho own for a grranulation o operation in the
following ta ables:
Where
S=Severityy on a scale
e of 1 to 10 (most seveere)
P=Probability on a scaale of 1 to 7 (most pro
obable)
D=Detectaability on a scale
s of 5 to
o 1 (most p
probably dettected)
The next sstep would be

b to take the
t highest risk areas which
w
are shown by the high her RPN numbers, whicch in this ca ase
would be tthe use of fa
ailed materials and equuipment con ntami-
nation and to repeat the
t FMEA ex xercise afteer putting in
n place
some correective action
ns as shownn in the nexxt table:
43
ECA Valid
dation Goo
od Practice
e Guide
This approoach can be applied to all the unit operations from

the start off the API prroduction to o the final DDrug Producct for-
mulation and packagin ng stages. The
T acceptaability of the e RPN
needs to bbe pre-deterrmined , butt the scaless used for th he S,P
and D can also be cha anged to en nsure a good d distributio
on of
RPNs so thhere are a limited numb ber of “Highh risk” areass to
work on firrst. If that doesn’t
d resu
ult in a proccess that is con-
trolled and
d capable th hen the nextt lower RPN N area is wo orked on
and so on until the prrocess rema ains in an accceptable sttate.
Any changes to the fa acilities, equ
uipment (inccluding com mputer-

44
ised systems), utilities, process or materials used which may
affect the quality of the product must be documented using
the change-control system. The impact this may have on the
validation of the product, and hence the possible risk, must be
formally documented using a recognised set of risk assess-
ment tools. Statistical tools should be used, where appropri-
ate, to support any conclusions with regard to the variability
and capability of the process.
Process Validation can be summarised as requiring all Critical

Control Points and Critical Quality Attributes to be identified.
Process Validation should involve sound design / development
and Risk Assessment which remains a regulatory requirement
for all drugs. Reproducibility at full-scale is an essential part of
Process Validation and Commercial phase improvements are
critical. The Product Review should be used to continuously
assess and improve the state of validation and the risk to the
patient should always be assessed.
It is up to each company to ensure that all risks are identified,

documented and justified using the appropriate scientific data.

45

46
3.2 Next generation of Qualification base on the
new US and EU guidances
(Gert Moelgaard)
From a qualification point of view the FDA Process Validation

Guide from 2011 and the EU GMP Annex 15 on Qualification
and Validation seems quite different and difficult to comply
with at the same time. FDA’s guide hardly mentions qualifica-
tion and the new Annex 15 goes even more in details than its
predecessor.
But the philosophy behind is almost the same and it is very

possible to comply with both at the same time. Behind the
scenes they both build on the same approach, called the Risk-
based approach, or more correctly: the Science- and Risk-
based approach. As do several supporting industry guidelines
such as the GAMP 5 guideline for computerised systems (Good
Automated Manufacturing Practice) and recent ISPE guidelines
on science- and risk-based verification.
However, this is easier to write than to do. In fact many com-

panies are struggling with updating their existing procedures
and practices to the new approach without making it much
more complicated - which was not the intention with any of
the guidelines. In contrary…
Qualification and Validation

What FDA’s Process Validation Guide and Annex 15 has in
common is the principles from the ICH Q guidelines for the
science- and risk based paradigm, especially the science ap-
proach from ICH Q8 on Pharmaceutical Development, the risk
management principles from ICH Q9 on Quality Risk Manage-
ment and the quality management principles from ICH Q10 on
the Pharmaceutical Quality System.
For the overall process validation approach - and including the

qualification approach - it means a high focus on the aspects
47
of manufacturing systems that potentially has a high impact
on the product quality with the eyes of the patient. Not the
only focus, but a high focus. So in qualification activities there
should be supporting quality risk management activities that
helps direct level of effort and documentation towards the
areas with the highest risk, both in order to mitigate or elimi-
nate the risk where possible and in order to ensure the sci-
ence- and risk-based focus on the manufacturing system
througout its life-cycle.
FDA’s Process Validation Guide only mentions the qualification

activities very briefly. As mentioned elsewhere in this guide-
line, it is a life-cycle approach with three stages: 1. Process
Design, 2. Process Qualification and 3. Continued Process Veri-
fication. In the section C about 2. Process Qualification there
are two parts, a technical part and a process performance
part. Section C1 about “Design of a Facility and Qualification of
Utilities and Equipment” use the term qualification for “activi-
ties undertaken to demonstrate that utilities and equipment
are suitable for their intended use and perform properly”. It
also states that “These activities necessarily precede manufac-
turing products at the commercial scale”, in other word they
are prerequisite for the process performance qualification. It
does not define how it should be done and it clearly does not
prescribe a specific approach such as DQ-IQ-OQ-PQ. More
about the terminology later but it is important that it does not
prescribe a standard method for qualification activities.
The Annex 15 of 2015 is much more specific, but not as spe-

cific as many companies interpret it. In fact it adds activities to
the former Annex 15, that only had DQ-IQ-OQ-PQ, because
the new Annex 15 has URS (and/or FDS)-DQ-FAT-SAT-IQ-OQ-
PQ and therefore many see it as more prescriptive than its
predecessor. But Annex 15 is actually not prescribing a specific
method. It is just exemplified more than FDA’s Process Valida-
tion guide and for some companies, this is a help. For others it
is a trap.
48
The European Approach
The confusion comes from that Annex 15 does not require that
its method has to be followed. it is simply a list of elements in
qualification activities that has become common and popular.
As it states: “Qualification activities should consider all stages
from initial development of the user requirements specification
through to the end of use of the equipment, facility, utility or
system. The main stages and some suggested criteria (alt-
hough this depends on individual project circumstances and
may be different) which could be included in each stage are
indicated below”. And then follows the descriptions of the
documents and activities like URS, DQ, FAT, SAT etc. Even
with the comment that “OQ normally follows IQ but depending
on the complexity of the equipment, it may be performed as a
combined Installation/Operation Qualification (IOQ)”, as has
become common practice in many companies.
So the full sequence of activities is not a requirement. In my

opinion they are mostly added to illustrate that there are many
sources of the quality records that goes into qualification and
some are outside the classical IQ-OQ-PQ. With good suppliers
of equipment that practices Good Engineering Practices etc.
many of the quality records of qualification are generated dur-
ing pre-qualification activities such as FAT and SAT. And if
supplier-generated quality records documents quality aspects
of the manufacturing system, they count as well. In fact, they
do not have to be repeated during subsequent testing, if a
proper configuration management and change management
system is in place. So the new Annex 15 activities for qualifica-
tion was not added to enhance the scope of qualification with
new activities but to enable pharmaceutical companies to lev-
erage quality records that are generated in earlier stages of
the manufacturing system life cycle than IQ-OQ-PQ. They are
added to give flexiblity, not to add complexity to qualification.
However, it is important to note the addition of the User Re-

quirement Specification to Annex 15 because that is becoming
49
an expected reference document. It has been part of industry
guidance such as GAMP from more than a decade, but now it
is one of the elements in Annex 15 and I think it will be a reg-
ulatory expectation in the future for all critical manufacturing
systems.
The FDA Approach

As mentioned, FDA’s Process Validation Guidance from 2011
has a totally different approach to qualification - at first sight.
It is stated that
“Qualification of utilities and equipment generally includes the

following activities:
● Selecting utilities and equipment construction materi-
als, operating principles, and performance characteristics
based on whether they are appropriate for their specific uses.
● Verifying that utility systems and equipment are built

and installed in compliance with the design specifications
(e.g., built as designed with proper materials, capacity, and
functions, and properly connected and calibrated).
● Verifying that utility systems and equipment operate in

accordance with the process requirements in all anticipated
operating ranges. This should include challenging the equip-
ment or system functions while under load comparable to that
expected during routine production. It should also include the
performance of interventions, stoppage, and start-up as is
expected during routine production. Operating ranges should
be shown capable of being held as long as would be necessary
during routine production.”
In other words, the FDA approach to qualification has more
focus on the performance of utility systems and equipment
than on the exact way the qualification is done. It mentions
the core elements of DQ, IQ and OQ but not the activities
themselves and not whether the documentation is established
in FAT, SAT, IQ, OQ etc. So in the FDA approach to qualifica-
50
tion the flexibility on how the quality records are established is
made more explicit and clear than in the European Annex 15.
Similarities between EU and US qualification approaches

There are many similarities between the two approaches. Nei-
ther the EU Annex 15 nor the FDA PV Guidance leaves any
doubt: Qualification should be planned and executed under
responsibility of the quality function of a company.
FDA states it as “The project plan should also include the

firm’s requirements for the evaluation of changes. Qualification
activities should be documented and summarized in a report
with conclusions that address criteria in the plan. The quality
control unit must review and approve the qualification plan
and report (§ 211.22). “.
Similarly EU Annex 15 states that “Qualification/validation per-

sonnel should report as defined in the pharmaceutical quality
system although this may not necessarily be to a quality man-
agement or a quality assurance function. However, there
should be appropriate quality oversight over the whole valida-
tion life cycle.”
They both rely on the science- and risk-based approach of ICH

Q8, Q9 and Q10 as mentioned above. In fact, this new para-
digm is the main reason why the both were updated and it
leaves not only new regulatory expectations but also new reg-
ulatory flexibilities that many companies have not been able to
capitalize on - at least yet.
FDA’s Process Validation guidance states that “Qualification of

utilities and equipment can be covered under individual plans
or as part of an overall project plan. The plan should consider
the requirements of use and can incorporate risk management
to prioritize certain activities and to identify a level of effort in
both the performance and documentation of qualification ac-
tivities. “
51
The EU Annex 15 is more direct on the quality risk manage-
ment aspect of qualification:
“A quality risk management approach should be applied
throughout the lifecycle of a
medicinal product. As part of a quality risk management sys-
tem, decisions on the scope
and extent of qualification and validation should be based on a
justified and documented
risk assessment of the facilities, equipment, utilities and pro-
cesses. Retrospective
validation is no longer considered an acceptable approach”
Another similarity is that both stress the opportunity of using

documentation from outside the qualification and validation
program.
EU Annex 15 states that “Data supporting qualification and/or

validation studies which were obtained from sources outside of
the manufacturers own programmes may be used provided
that this approach has been justified and that there is ade-
quate assurance that controls were in place throughout the
acquisition of such data”
FDA’s Process Validation Guidance does not have a similar,

clear statement on the potential role of outside documenta-
tion, but it does have an indirect reference. A footnote in
FDA’s guide refers to the industry standard ASTM E2500
“Standard Guide for Specification, Design, and Verification of
Pharmaceutical and Biopharmaceutical Manufacturing Systems
and Equipment”, which suggests the following approach to the
use of vendor documentation:
“Vendor documentation, including test documents may be

used as part of the verification documentation, providing the
regulated company has assessed the vendor, and has evi-
dence of:

52
x An acceptable vendor quality system,
x Vendor technical capability, and
x Vendor application of GEP such that information ob-
tained from the vendor will be accurate and suitable to
meet the purpose of verification.
x If inadequacies are found in the vendor quality system,
technical capability, or application of GEP, then the
regulated company may choose to mitigate potential
risks by applying specific, targeted, additional verifica-
tion checks or other controls rather than repeating
vendor activities and replicating vendor documentation.
x The decision and justification to use vendor documen-
tation, to support the verification of critical aspects of
the manufacturing element, should be based on the in-
tended use of the manufacturing system, and should
be documented and approved by subject matter ex-
perts including the quality unit”
These are only some of the similarities between the FDA and
the EU approaches to process validation but it may be worth-
while to reflect on why the similar philosophies to validation
and qualification has been implemented so different over the
years, especially since some of these differences are causing
daily pain in many international pharmaceutical companies
that must manage double compliance with both the US and EU
approaches at the same time.
Qualification in the history of process validation.

Ever since the FDA published its original process validation
guide, the 1987 „Guide to General Principles of Process Valida-
tion“ it has been a requirement that there has to be ‘some-
thing’ that demonstrates ‘fitness for use’ in the design and
implementation of a pharmaceutical manufacturing system.
FDA’s old1987 guide is only shortly outlining the requirements
this way: “Installation qualification studies establish confidence
that the process equipment and ancillary systems are capable
of consistently operating within established limits and toler-
53
ances. After process equipment is designed or selected, it
should be evaluated and tested to verify that it is capable of
operating satisfactorily within the operating limits required by
the process”.
Without going into details on the terminology changes from

this early and broad definition of Installation Qualification, it
suffice to say that the requirements for IQ/OQ/PQ etc. in the
original 1987 guidance were brief and broad compared to the
large efforts and cost that hav become industry standard qual-
ification activities.
The very big amount of qualification work and documentation

was never the intention, neither from the regulatory agencies
that made the regulations nor from professional organisations
such as ISPE, PDA and others that made the guidance to sup-
port it. But it grew over time into industry practices that re-
quired more and more formality and rigor in change manage-
ment, deviation management, sequence of actions, QA ap-
proval signatures etc. In many companies it has become a big
part of every investment project. Very often both the Quality
Unit and the Engineering/Technical Unit realise that it adds
very little value compared to the effort and money invested,
but many think it is a burden that comes from “requirement
from the FDA”.
When the ISPE Engineering Baseline Guide Volume 5 for

Commissioning and Qualification (C&Q) was launched in 2001,
it gained broad acceptance in pharmaceutical companies
around the world and has been widely applied as the refer-
ence for a more streamlined approach compared to the older
concepts of validation that were used quite differently in vari-
ous companies.
The C&Q Baseline Guide introduced a few key concepts that

was significant improvements compared to traditional qualifi-
cation. Since then many companies have established Good
54
Engineering Practices (GEP) and increased the leverage of
commissioning tests into Installation and Operational Qualifi-
cation (IQ, OQ or combined IOQ). The concept of impact as-
sessments has also been very effective in identifying the man-
ufacturing systems that have direct or indirect impact on
product quality – a simple, first approach to risk management.
This was the reason for industry and FDA to initiate develop-
ment of the ASTM E2500 “Standard Guide for Specification,
Design, and Verification of Pharmaceutical and Biopharmaceu-
tical Manufacturing Systems and Equipment”. It was a reaction
on several international investment projects that were delayed
and had big budget overruns due to complext C&Q and Com-
puter Validation industry practices. The ASTM E2500 standard
guide was made as a follow-up to the new quality manage-
ment paradigm of ICH Q8, 9 and 10 as well as the FDA PAT
initiative, the Quality by Design activities (QbD) and other FDA
guides supporting a science- and risk-based approach to
pharmaceutical manufacturing.
The idea is in short to replace „one size fits all“ C&Q practices
with a more flexible approach build on the same risk manag-
ment practices as ICH’s Q9 guideline for Risk Management. So
the ASTM E2500 standard outlines a science- and risk based
approach to the verification or qualification of manufacturing
systems It builds on the ICH Q9 concept that “The level of
effort, formality and documentation of the quality risk man-
agement process should be commensurate with the level of
risk and be based on scientific knowledge”.
ASTM E2500 uses the term ‚Verification’ rather than ‘Qualifica-

tion’ in order to facilitate that the C&Q activities adapt the
„level of effort, formality and documentation“ to the risks in
the manufacturing system.
The ASTM E2500 has both a four step-wise approach and a
number of supporting activities that follow through all phases
of a verification project.
55
The four steps are:
1. Requirements definition
2. Specification and design
3. Verification
4. Acceptance and release
The supporting activities are:

· Good Engineering Practices
· Risk Management
· Design Review
· Change Management
Beside of this there is are a number of basic principles that are
related to the points mentioned above such as
· Risk-based Approach
· Science-based Approach
· Critical Aspects of Manufacturing Systems
· Quality by Design
· Good Engineering Practice
· Subject Matter Experts
· Use of Vendor Documentation
· Continuous Process Improvement
Good Engineering Practices are fundamental to the whole
E2500 concept as well as to the traditional C&Q concept.
These are the technical standards, the core GMP require-
ments, the best practices in industry as well as the skills and
training of the professionals involved. In short it is about a
quality management system to ensure Good Engineering Prac-
tices. Some pharmaceutical companies have success in imple-
menting it as a quality agreement between the Quality Unit
and the Technical Unit/Engineering Dept. and with involved
suppliers.
Quality management by auditing, surveillance etc. applies and

‘trust must be earned’ in the sense that the overall quality
responsibility in a pharmaceutical organisation relies within the
quality unit. Tasks may be delegated, but will always be under
the final responsibility of the Quality Unit. Thus an Engineering
56
Unit, as an example, will have to earn the trust of the quality
unit in order to take over more responsibility as ‘Good Engi-
neering Practices’.
Another fundamental principle is the involvemtment of Subject

Matter Experts (SMEs) in certain areas, rather than using peo-
ple from the Quality organisation for all reviews, approvals,
signatures, etc. The SMEs enables a more lean and result-
oriented approach such as using technical specifications as
reference for commissioning and qualification activities.
There are several good examples of companies that have tak-

en a science- and risk based approach on board in their quali-
fication activities. Some of them have even dropped their for-
mer internal commissioning and qualification procedures with
a set of verification procedures following the ASTM E2500
Standard.
The companies that change to new science- and risk-based

approaches to qualification uses the ASTM E2500 Verification
Standard or other risk-based approaches to ensure that the
main focus is on the Critical Aspects of the manufacturing sys-
tem based on a good understanding of the product, processes
and equipment in scope and a documented risk management
process.
One way to ensure success in a transformation from traditional

qualification practices to a science and risk-based approach
like the ASTM E2500 verification approach is by combining a
cross-organisational involvement of functions such as Manu-
facturing, Engineering, Automation, Quality, and Process De-
velopment together with a strong sponsorship from senior
management.
It is a challenge but necessary to introduce the cross-

functional team that leads the verification program to product
and process information, as it relates to regulatory require-
57
ments, product quality and patient safety, because this
knowledge, including the manufacturing history, should be
used as the basis for making science- and risk-based decisions
that ensure that the manufacturing systems are designed and
verified to be fit for use. This includes understanding critical
quality attributes, critical process parameters, GMP require-
ments and other manufacturing- and quality-related critical
aspects.
It is also a challenge to train the team in the application of

Risk Management tools and methods, including Quality Risk
Management that should be applied during the whole specifi-
cation, design and verification process. In some companies
this has led to a very complex approach to the new EU GMP
Annex 15 and the FDA guide, which was never the intention of
the regulators.
Many pharmaceutical companies struggle to establish and en-

sure the use of Good Engineering Practice (GEP) principles to
support all verification test activities and to identify and en-
gage suitably qualified Subject Matter Experts (SMEs) during
the planning, review and execution of the verification activi-
ties.. They have difficulties in implementing the concept that
verification activities and corresponding formality of documen-
tation should be commensurate to risk to patient safety and
product quality. A reason may be that many companies have
been using a ‘one-size-fits-all’ approach to C&Q so the ability
to distinguish and focus on critical aspects in the manufactur-
ing system requires new and additional training. However,
companies that have already trained their staff in Quality Risk
Management based thinking, see the benefit of leveraging the
approach throughout their business activities.
Here is an example of a practical approach from companies

that have implemented a transformation to a science- and
risk-based to science- and risk-based C&Q that complies with
both EU Annex 15 and the FDA Process Validation Guide;
58
1. Ensuring the collection all relevant product and process
knowledge, including relevant knowledge from the
manufacturing history of similar processes and equip-
ment and relevant requirements from regulatory guide-
lines and the company’s own pharmaceutical quality
system.
2. Development of an initial User Requirement Specifica-

tion (URS) document based on the above knowledge as
well as the intended use of the manufacturing system.
It should be shorter and more focused than traditional
User Requirement Specifications to enable a clear focus
on what is critical. Practical experiences shows that this
step is the most critical: if companies don’t get the new
URS concept right, the whole benefit of the verification
approach is at risk.
3. Development of the core design and specification doc-

uments that are the basis for the implementation and
verification. In the ASTM E2500 approach this devel-
opment is typically iterative, i.e. it is repeated several
times in a combination of risk management and design
review meetings with all appropriate parties involved.
The focus is not only on design and verification but al-
so on practical risk management with the purpose of
risk mitigation by design changes, operational require-
ments etc. and is repeated several times during the
project execution.
4. Developing a specific Quality Verification activity plan

that outlines the verifications activities based on the
design, specification and the risk management activi-
ties, including the vendor capabilities and the product
and process risk areas. The verification plan should
clearly outline roles, responsibilities and risk manage-
ment focus on all involved parties, including Subject
Matter Experts, Quality Unit, external suppliers etc. It
59
may include other risks such as EH&S, business risk
etc.
5. Preparation of the specific verification activities, includ-

ing possibly vendor audits, FAT/SAT/IQ/OQ/... test
plans and other quality measures that serves to sup-
port appropriate quality management focus from a risk-
management perspective
6. Execution of test and inspections covering all activities

that will serve to document the verification, including
FAT, SAT and other inspection and test activities that
will serve to document the ‘fitness for use’ of the man-
ufacturing system. The actual verification testing is
based on the criticality and risk as determined during
the previous activities.
7. Overall evaluation and conclusion regarding acceptance

and release of the manufacturing system for subse-
quent validation life cycle activities, including process
performance qualification, sterilisation validation and
other validation activities.
8. Release of the manufacturing system and continuation

into the Process Qualification or Process Performance
Qualification activities.
One has to be careful about the wording because EU Annex 15

and FDA’s Process Validation guide uses different terminology.
EU Annex 15 uses the traditional URS/DQ/FAT/SAT/IQ/OQ/PQ
approach and terminology for qualification and after these
follows Process Validation and Ongoing Process Validation.
FDA’s approach is a life-cycle approach that does not use the

wording from the URS/DQ/FAT/SAT/IQ/OQ/PQ approach at all.
It distinguishes between stage 1 Process Design, stage 2 Pro-
cess Qualification and stage 3 Continued Process Qualificaiton.

60
Stage 2 corresponds roughly the european approach. Stage 3
is similar to the new European concept of Ongoing Process
Qualification.
The potential confusion in companies qualification and valida-

tion procedures can be omitted by using the US term Process
Performance Qualification instead of the European term Pro-
cess Validation. And to use the EU term Ongoing Process Veri-
fication rather than the US term Continued Process Verifica-
tion.
The confusion risk is even higher when the term Continuous

Process Verification is introduced, as it is in the EU GMP Annex
15 and the EU Process Validation Guide. Here it is used in a
similar meaning as the US term Process Analytical Technology
(PAT) and the confusion can be prevented by using the PAT
terminology and not the Continuous Process Verification ter-
minology.
In the real-life implementation of the verification approach it is

important that all participants are well trained in the risk man-
agement principles as well as the mindset in the science- and
risk-based approach to quality risk management etc. Another
critical success factor is the understanding and implementation
of Good Engineering Practices that must support the verifica-
tion approach. The consequences of insufficient understanding
of the quality risk, management principles and the Good Engi-
neering Practices among the Subject Matter Experts in the
program, can jeopardize the success of the verification pro-
gram. But with the new EU GMP Annex 15 and the new FDA
Process Validation there is an additional risk that people does
not understand the new terminology as well as the new con-
cepts.

61
Ensuring a strong involvement of the quality unit during plan-
ning of the verification and risk management activities during
the project is very critical to ensure the overall understanding
and that the manufacturing system is indeed fit for intended
use in all critical aspects.

62
3.3 Statistics in Process Validation and Continued
Process Verification (some thoughts on "how to do")
Dr. Thomas Schneppe
Disclaimer and Explanations on this document
¾ These "thoughts" are my personnel view and may need

correction, improvement and/or additions… and sorry for my
"English" - corrections highly welcome. As it is unclear up to
now what will be the further way how to deal with this draft I
just called it "some thoughts" for to be as neutral as possible.
The way how to continue with this document will depend very
much on the comments, corrections and additions from the
experts who received it for commenting.
¾ In the statistics arena there are often different options to go

ahead. The selection of tests, confidence levels and evaluation
strategies (as e.g. defined in statistical software) may lead to
different potential solutions per case. In many industries dif-
ferent statistical standards are available. They reflect a wide
range state of the art, which gets additional interpretation and
differentiation with the creation of individual company rules. In
some areas of statistics there are different definitions and
rules established in different parts of the world, e.g. different
approaches exist in establishment of warning limits, confi-
dence levels and related rules and limits and terminology.
¾ This document is intended to give some hints and to point

out aspects and positions that might be needed to be clarified
within a company to come to a clear approach on how to
manage statistics between auditors and audited functions,
authorities and companies, customers and suppliers to come
to a common understanding of e.g. procedures and ac-
ceptance limits.

63
¾ Statistical methods are in use to support many pharmaceu-
tical and GMP-related activities in research, development and
production. They can be used in a more reactive mode (e.g.
with reviews, Key Performance Indicator Reporting, Product
Quality Review, Annual Product Review) or in a more proactive
mode e.g., establishment and continuous verification of pro-
cess capability indices, Statistical Control (SPC)-Charts, CAPA
investigations. This document discusses mainly aspects of the
implementation of process capability indices and SPC.
¾ In a later stage of drafting we may add some more specific

overview on statistical tests and, process models and calcula-
tion schemes. The pros and cons to do so should be thorough-
ly discussed to avoid creating a too complex document that
leads more to "hesitation instead of motivation". Furthermore
it may be very challenging to go in such level of detail as there
are various potential approaches described in literature - at
least for me it was not possible to identify one "golden path".
3.3.1 Introduction, Aim and Scope
FDA 2011 Guidance for Industry on Process Validation:

Process validation is defined as the collection and
evaluation of data from the process design stage
throughout production, which establishes scientific
evidence, that a process is capable of consistently de-
livering quality products.
By the beginning of 2011 the FDA published a “Guidance for

Industry on Process Validation” which gave strong emphasis
on understanding the life cycle of process validation by foster-
ing the use of statistical methods to establish and continuously
verify the capability of manufacturing processes.
Other authorities (e.g. EMA), institutions (e.g. WHO) and as-

64
sociations (e.g. EFPIA) confirmed this approach at least as an
additional option beside the established traditional "3 batch
approach".
Therefore one key challenge is now to align the new approach

with the existing "3 batch validation" concept as used actually
as the most common approach to demonstrate validated pro-
cesses.
As pharmaceutical manufacturing is highly regulated and actu-

ally a confirmatory rather than a proactive approach is com-
mon in process validation this document shall serve as a basis
to define common understanding between manufactures and
authorities on how to establish capable and valid, not just val-
idated, processes.
This document is intended to give some hints, answer some

frequently asked questions and provide some proposals, expe-
riences and examples on potential approaches. It cannot re-
place the statistical expertise that is necessary in the compa-
nies and functions to implement reliable and sustainable capa-
bility indices and statistical control mechanisms like SPC-
charts.
The use of statistical methods supports either initial validation

activities as confirmation and/or improvement of existing
manufacturing process and is interfaced with other GMP key
processes like CAPA, Product Quality Review-Mgmt., Continu-
ous Improvement, Operational Excellence Initiatives and
Technology Transfer.

65
3.3.2 Key Messages
¾ The manufacturer has to select and implement adequate

statistical tools and evaluation procedures depending on the
process and intention (Improvement, steering, verification…).
No authority dictates which tools to use.
¾ Actually there are still differences between established sta-

tistical rule sets, e.g. in the automotive industry area. For ex-
ample the US AIAG Manuals act with a different standard level
for significance levels and related failure risks than the Ger-
man automotive industry (VDA, Verband der Automobilindus-
trie) standards. Some definitions are different in the process
capability area. Nevertheless the general principles are similar
and in real life the statistical standards, philosophies and ap-
proaches are harmonizing more and more.
¾ In some cases the statistics and the GMP requirements

need specific alignment, which shall be demonstrated with one
example:
- With statistical activities outliers are identified and elimi-

nated according to certain rules e.g. to allow establishment
of meaningful limits in statistical control charts. The source
of outliers should be identified and eliminated.
- The GMP rules expect thorough investigations of outliers

which may lead to formal procedures like CAPA. In the
field of Quality Control outliers may undergo very sophisti-
cated Out of Specification investigations.
- On the other hand the use of statistical methods supports

timely investigation of observations in the environment of
e.g. OOS evaluation or CAPA activities to identify if the ob-
servation is founded in the process.
¾ Establishment and verification of capable processes (evalu-

66
ation of allocation and variety of parameters and attributes)
does not contradict or conflict with the traditional 3 batch vali-
dation approach (verification of specification adequacy and
conformity of variables and parameters). It is just a different
view on the same data set and can confirm each other:
- The traditional validation approach confirms that critical

process parameters and critical quality attributes are able
to be held in a given range (within the specification) and
lead to a product in conformity with the quality require-
ments.
- If this can be supported with adequate process capability

indices and continuously verified with SPC this is very val-
uable additional information as it can be proven not only
that the process is in specification but furthermore now,
how much it is in specification and stable (Cp, CpK).
- Therefore it is no contradiction to reflect the traditional

and the capability-driven approach on process validation in
one validation report (see Table 3.3.2.1).

67
Aligment of validation approaches The new and traditional approach rely on
same set of data; the validation report
Example: Blistering Process may reflect and underpin both:
¾ Parameters/Attributes „in specification“
Validation Planning: ¾ Process Capability (Cp/CpK) is demon-
Risk Assessment or strated and sustained e.g. via SPC-charts
e.g. „Quality by Design“
to identify relevant para- traditional approach:
meters/attributes, e.g.: 3 batches
- Time TI Validation Reporting:
- Pressure PR Evaluates TI TE PR
- Temperature TE (challenge) data.
new Evaluation of data set new

FDA that reflect 3 batches FDA
GfI GfI
Stage 1
QC:
Stage 2 Tight
blister
Data of batches 1-n… Process Capability
Design
Intensive IPC samp- TI x x x x x x x x x x x x x … (e.g. Cp, CpK) for TI, TE,
Space ling/testing of the TE x x x x x x x x x x x x x … Batch release to be based
critical parameters PR x x x x x x x x x x x x x … on significant no. of data.
Pressure PR Stage 3
Routine Life Cycle
Proof of capability leads
to reduced routine IPCs
Continuous proof of capability e.g.
via SPC charts replaces revalidation
Table 3.3.2.1: How to align the 3 batch evaluation" and "continu-

ous verification"?
¾ Use of SPC and a proven acceptable process capability may

allow reduction of some QC testing activities which shall be
illustrated with the following example:
- The key process parameters for sealing of blisters (see

Table 2.1) are sealing time, sealing pressure and sealing
temperature.
- If the capability (Cpk) of all 3 parameters is proven and

continuously monitored (e.g. via control charts) the inten-
sity of QC testing for tight blisters may be reduced.
- In case of deficiently sealed blisters the root cause investi-

gation can exclude the process and may focus e.g. on the
materials quality.
¾ SPC charts can be established for qualitative data (e.g.
68
complies / does not comply) and for quantitative data as well.
As the quantitative SPC charts are more meaningful it is rec-
ommended to prioritize these parameters when implementing
SPC.
¾ Instead of SPC charts other and less statistically oriented

control chart systems may be implemented e.g. the Pre-
control System which is less powerful in terms of interpreta-
tion and significance but which is easy to understand and im-
plement on the shop floor level. Furthermore Pre Control
works well with small batch sizes as these are sometimes the
outcome of "lean six sigma" projects.
3.3.3 General Process Flow
Table 3.3.2.2 reflect an example for a potential process flow to

implement process capability indices and statistical process
control charts. Some aspects are pointed out more in detail in
the following comments (cmts)

69
ECA Valid
dation Goo
od Practice
e Guide
Table 3.3.2..2 General process flow
CMT 1 and CMT 2: Statistical

S Process Conttrol is well estab-
e
e of the art in many industries witth a specificc termi-
lished state
nology whiich differs in
n some case es from thee GMP techn nology
he same purpose. For eexample is a
but serves basically th
measuring system ana alysis requirred in orderr to ensure mean-
a to be evalu
ingful data uated?
This princip ple is implemented in the

t pharmaaceutical ind dustry
already viaa calibrationn, qualification of laboraatory equip pment,
System Suitability Tessts and valid dation of annalytical methods,
covering asspects like e.g.
e accuraccy (bias), reepeatability, repro-
ducibility, rrecovery, roobustness, selectivity,
s llinearity, sta
ability,
precision, a and limits of
o detection//quantitatio on.

70
Statistical data and capability indices (machine capability
Cm/Cmk) can be supportive during the qualification of equip-
ment as well. The process orientation and close interlink of
equipment qualification and process validation can be shown
in using the relevant process parameter data for evaluation of
the machine capability. Continuation of the data collection can
be done until sufficient data are available for the evaluation of
the preliminary process capability and after further continua-
tion of data collection and the evaluation of the long term ca-
pability is possible. One frequently asked question is how
many data points are needed. Orientation may be taken from
other industries where capability evaluations are well estab-
lished, e.g. the VDA 4.1 document (Verband der Automobilin-
dustrie, Germany) mentioned 50 data points for calculation of
machine capability, 125 data points for the preliminary process
capability and data points from 5 production days for the long-
term capability. These orientation points may need "transfor-
mation" into the pharmaceutical environment.
CMT 3: The preliminary evaluation is important to get orienta-

tion on the stability of the process and to get data on spread
and location of the sample data. The preliminary evaluation in
many industries is done with approximately 125 data points
(e.g. 25 samples with n=5). In case less data points have to
be used (e.g. due to destructive testing need) the capability
indicators target values may be raised. For the evaluation of
process parameters in pharmaceutical manufacturing process-
es usually the sample size is n=1. Individual approaches have
to be defined in the validation planning to come to adequate
sampling procedures to gain enough meaningful and sufficient
data to allow adequate preliminary process capability evalua-
tions.
CMT 4: In general process improvement should have first

priority. In the meantime the process should be under close
surveillance (Pp/PpK, SPC charts, Pre-control, additional test-
ing…). For validation purposes a stable process has to be ex-
71
pected from stage 1.
CMT 5, CMT 6 and CMT 7: The identification of the best

fitting process model (acc. to ISO 21747-2006) is a key step
as the process capability calculation schemes and the SPC-
charts are oriented to specific process models.
CMT 8, CMT 9: In case an unstable / not yet stable process is

identified it may be acceptable nevertheless to establish pro-
cess indices (Pp / PpK) and SPC charts to get some better con-
trol on the process. Sometimes the expectations on the target
indicator are higher for Pp/PpK than for Cp/CpK to have an
equivalent for the non stability e.g., instead of a CpK of 1,33 a
PpK of 1,66 may be expected.
Parallel to this it should be tried to stabilize the process and to

establish capability indices (Cp / CpK) and SPC charts with
adjusted limits for the stabilized process. In case it is not pos-
sible to stabilize a process adequate additional control
measures may have to be established. For validation purposes
this cannot be a permanent solution, the process has to go
back to stage 1 for re-design.
CMT 10: The IM/R-chart is a control chart that works with

samples sizes of n=1 which often appear in the field of pro-
cess parameters. Within given intervals temperatures, pres-
sures… are documented as 1 data point per time. The other
control charts deal with samples sizes of n>1 per sample
(common standard is 5 per sample).
CMT 11: The continuous sample verification (stage 3 acc. to

FDA 2011 Guidance for Industry on Process Validation) reflects
the steady state and sustainable verified the process capabil-
ity. The review data may serve as KPI and support PQRs,
APRs, and CAPA investigations.

72
ECA Valid
dation Goo
od Practice
e Guide
3.3.4 Spe
ecifics on identificattion of Con
ntrol Chartts
Table 3.3.22.3 demonsstrates in prrinciple the ssame proceess flow

like Table 3
3.3.2.2 but with more focus on th he specifics of the
manufacturing processs and its chhallenges an nd options.
Table 3.3.2..3 Process Fllow within the organizattion
Strategies and pathwa ays for iden

ntification off adequate control
charts are defined in commercial
c ly available statistic so
oftware.
No specificc user knowwledge on sttatistical tessting is need
ded to
receive infoormation onn the adequuate type off control cha art.
These path hways can be
b adapted to companyy specific needs
and rules. This allows the establishment of ccompany sp pecific
statistics approaches as
a a kind off "blackbox"".
Table 3.3.2 2.4 describe
es one of many
m approaaches to deccide on
SPC chartss. Pathways as describe ed in the exxample in ta able
3.3.2.4 aree normally part
p of comm mercial stattistics softw
ware.

73
Table 3.3.2.4: Example for a pathway to identify adequate SPC-

Charts
3.3.5 Implementation Support
One major aspect to be solved while implementing process

capability evaluations and/or statistical process control is to
describe the company approach. E.g. how do you align these
activities and resulting data and conclusions with the existing
"3 batch validation" approach. Traditional validation strategies
and capability evaluations and/or SPC activities rely basically
on the same set of attributes, parameters and data which
supports the alignment of the approaches (see Table 3.3.2.5).

74
Table 3.3.2.5: Challenges and Steps in the Implementation Phase
3.3.6 Literature
Aide Memoire 07121105, Inspektion von Qualifizierung und

Validierung…, ZLG 2010
FDA Guidance for Industry on Process Validation, 2011
ISO 8258-1991 Shewhart Control Charts
ISO 21747-2006 Statistical methods - Process performance

and capability statistics for measured quality characteristics
TEQ Six Sigma Blackbelt Training Documentation 2009

75
3.3.7 Abbreviations and Terms and their meaning as
applied in this documents
Cp Process Potential Index for spread (stable

processes)
CpK Process Capability Index for spread and
position (stable processes)
I/MR ChartOne type of SPC-Charts for n=1:
Individuals and Moving Range Chart
Pp Process Potential Index for spread
(unstable or not yet stable process)
PpK Process Performance Index for spread and
position (unstable or not yet stable process)
Run 7 data points in a row above or below mean
SPC Statistical Process Control (unstable or not
yet stable process)
Trend 7 data points in a row raising or falling
X-bar/s-Chart Statistical Control chart for n>1 and normally

distributed data

76
3.4 Process Validation and Statistical Trending in
Biopharmaceutical Manufacturing
Dr Renate Schenk-Gröninger
1. Introduction in Biopharmaceutical Processes

In the Pharmaceutical Manufacturing World Biophar-
maceutical Production Processes play an extra role as
they mainly cover the API (=active pharmaceutical in-
gredient) manufacturing using the growth and produc-
tivity of microorganisms or cell cultures. After purifica-
tion the API is formulated with excipients and then
usually aseptically filled into vials, syringes cartridges
etc.
Biopharmaceutical processes used to produce proteins,

peptides or other biochemical molecules as active
pharmaceutical ingredients are time-consuming and
complex. The majority of the process is covered by the
fermentation process, followed by the purification of
the API. The API manufacturing process is a liquid pro-
cess which may take several days or weeks.
After formulation of the API with final excipients the

solution is dispensed into single or multiple-dose units.
This filling process is less time-consuming and takes
only a few hours or days.
Cell culture processes may start with a working cell

bank vial of e.g. 1ml in the inoculum and are then
scaled up over several cycles to a final production bio-
reactor volume of e.g. several thousand liters. The
scale up and production process may take e.g. 4-6
weeks for a fed batch production or even longer for a
continuous perfusion fermentation.
During this time intensive control of the cell culture is

required and physical/physicochemical process controls
77
are monitored continuously while analytical control is at
least performed daily. The growth of the cell culture
may be controlled by the measurement of e.g. cell
density, viability, product titer and metabolites. Physical
parameters like temperature, stirring as well as physi-
cochemical parameters like O2 and pH may be continu-
ously measured using probes.
For microbial processes e.g. temperature, turbidity and

pH are tested. At the end of the fermentation process
cells are harvested and removed and the supernatant
is further processed.
Microbes for example are pelleted and cracked to ex-

tract the product containing inclusion bodies.
During downstream processing the API is purified using

various chromatographic and filtration techniques over
several process steps. The purification process is moni-
tored using a variety of analytical methods (e.g. protein
concentration, purity checks, HPLC methods, Bioburden
and Endotoxin testing) and physicochemical parame-
ters (conductivity, UV spectra, pressure control etc.).
Process time varies from a few days to a few weeks in-
cluding the final formulation process of the API.
During the filling and e.g. lyophilisation process which

takes only a few hours or days several parameters are
controlled to assure the aseptic homogeneous manu-
facturing of single or multiple dose units as application
forms. The parameters may include e.g. filling weight,
stopper positions and other machinability parameters
(filling speed, pressure etc.).
Drug substance and drug product is controlled for pa-

rameters like bioactivity, binding activity, protein con-
centration, impurities, purity, identity, physicochemical
78
characteristics, protein structure and posttranslational
modifications, isoelectric pattern, sterility etc.
2. Process development, reevaluation of commer-

cial processes and definition of parameters
New processes
The Product life cycle is divided into several stages,
e.g. determination of the genetic information, genetic
engineering, clone identification and clone stability, cell
bank development, fermentation development, purifica-
tion and formulation development and development of
the final product in its application form. Various scales
are applied within the product life cycle to produce ma-
terial for several clinical phases.
During process development for each process step pro-

cess parameters are defined as
- critical parameters having impact on product
quality attributes
- Key parameters having impact on process
performance
- non-key parameters having no impact on product
quality attributes
The parameters may be categorized as

- input
- output and
- performance parameters.
Product and process characterization studies are per-

formed supported by DoE analysis to identify the criti-
cality of the parameters.
For the product quality critical quality attributes are de-

fined and usually result in final release specifications.

79
The parameters are monitored during process devel-
opment at various scales (small scale to large scale)
and data is gathered for further data analysis.
Approved Commercial processes (“Old” processes)

are mostly validated based on the former 3 validation
batches approach. Revalidation may be necessary after
reevaluation or due to process changes.
These “old” processes consist of several process steps

too where a number of parameters are controlled.
These parameters are
- input or
- output
- or performance parameters too.
They are of
- technical
- physical
- biochemical
- biological origin.
Due to the historical background they are usually not

categorized into a critical, key and non-key classifica-
tion. Based on historical experience with the process,
scientific knowledge, development data and the change
history, the parameters may be categorized in a risk
assessment (e.g. FMEA) on the manufacturing process.
Release specification and quality attributes (product

quality, safety, potency and efficacy) have to be taken
into account in this risk assessment as well. The risk
assessment of the parameters is then to be combined
with a statistical evaluation (plotting of historical data
in a Shewart control chart against existing process lim-
its and specification, calculation of the mean and +/- 3-
SD-ranges around the mean, process capability calcula-
80
tion with regard to existing limits, min-max calculation
and data distribution analysis) of the historical data
from routine manufacturing and release and historical
development data.
The scientific relevance and the impact on product

quality and process performance further helps to cate-
gorize these parameters and their limits.
Due to this categorization a process validation and con-

tinuous process performance qualification program is to
be designed.
3. Parameters and control

Product quality in Biopharmaceutical processes is ini-
tially defined by the organism (animal or human cell
line or microbe) used in the fermentation process and
the conditions under which these organisms grow and
produce the API. During API purification various effects
(filtration and separation techniques, pH, conductivity
etc.) may influence product quality too.
Later on the final product quality is influenced by the

pharmaceutical manufacturing process design and pa-
rameters, e.g. temperature and shear forces during the
final sterile filtration step, freeze and thaw conditions
during lyophilisation.
Results of these parameters are either numerical or al-

phanumerical.
They can be categorized as

x Quantitative
x Discrete (Endotoxine results)
x Continuous (Protein concentration results)
x Qualitative results.

81
They may in some cases be normally and in the majori-
ty non-normally distributed. Especially semi-
quantitative results and results from limit tests are
usually non-normally distributed.
Some parameters are of technical/physical origin and

thus easy to control during the process. In most cases
these parameters in biopharmaceutical processes are
input parameters and can be continuously monitored
online.
Other parameters are output parameters of the process

and are not easy to control. Many of them are usually
analytical parameters and are only generated in distinct
time intervals.
Some analytical output parameters may be defined as
critical quality attributes and final release specifica-
tions.
Analytical methods to measure these parameters need

to be validated. Analytical instruments and manufactur-
ing equipment need to be qualified prior to use as well
as software programs.
4. Process Performance Validation Approach

The historical approach of Process Validation covered
the “3 Batches validation approach”. Three consecutive
batches in the final commercial scale format were
manufactured and data was gathered for a variety of
parameters. The results of the parameters were re-
ported and compared to predefined validation ac-
ceptance criteria. The success of the Validation Proce-
dure was demonstrated by meeting these criteria com-
pletely for the 3 batches only. If the process parameter
ranges and release specifications were met, no further
evaluation regarding the validated status of the process
was needed for routine manufacturing.
82
The ICH Q8 Guideline, the US Process Validation
Guideline and the Annex 15, although not mandatory
for APIs, now require a different approach.
Biopharmaceutical processes are scaled up during pro-

cess development to result in the final commercial pro-
cess format and scale. During development a limited
number of batches are produced in medium and final
scale. Scale up has to be demonstrated to generate
comparable product quality during development. Espe-
cially the micro-heterogeneity (Oligopeptide structure,
posttranslational modification pattern) of the complex
protein molecules needs to be comparable.
Since comparability of the complex biomolecules is

demonstrated and scale effects can be disregarded,
development data may be used as an additional data
basis for statistical evaluation of the process perfor-
mance. Here as well data may be plotted in control
charts, mean values and +/- 2SD or 3-SD-ranges
around the mean may be calculated, a min-max calcu-
lation and a data distribution analysis as well as a pro-
cess capability analysis may be performed with regard
to existing limits. A new definition of limits may be
necessary.
Additionally, small scale data from characterization ex-

periments may support the statistical evaluation men-
tioned above and gain in a deeper process understand-
ing.
Based on the number of development, characterization

and clinical pre-validation batches, the number of pro-
cess performance validation batches may be defined
individually (e.g. statistical number n >5 to 10 for all
batches).

83
A justification for the number of batches to be validat-
ed is supportive for the validation strategy.
If development data is not sufficient, the “3 batches

approach” may be used further or may be extended by
additional batches.
The Process Performance validation program for the

defined number of batches shall cover all critical and
key parameters and may be supported by non-key pa-
rameters. Product quality attributes need to be evalu-
ated as well.
The biopharmaceutical API manufacturing process is a

liquid (continuous) batch manufacturing process.
As homogeneity of the liquid is once demonstrated by

mixing studies using the manufacturing equipment and
mixing times are predefined, only single representative
samples/measures may be taken per process step and
parameter. Repeated sampling/measurement to
demonstrate homogeneity of the liquid is then not re-
quired.
In this case single results per batch per process pa-

rameter and process step are generated. Statistical
evaluation (control charting, process capability calcula-
tion, analysis of variance) can only be performed over
the predefined number of process performance valida-
tion batches. The development and clinical batches da-
ta may support this statistical evaluation.
5. Trending program and choice of parameters

Based on a review and evaluation after completion of
the Process Performance Validation the parameters for
a monitoring/trending program are defined. Usually all
84
critical and key parameters are part of the trending
program, non-key parameters may be omitted.
For the whole process single process steps and rele-

vant parameters are defined for the trending program.
Further the critical product quality attributes are de-
fined as release specifications and may be part of the
trending program.
In biotechnological manufacturing processes several

parameters are monitored over various process steps
over many days.
Cell culture e.g. is controlled continuously online for

temperature, stirring etc.. Analytical parameters (Os-
molality, Glucose concentration, Product titer etc.) are
controlled for example daily or a few times per day.
Technical input parameters which are easy to control
may be omitted in a trending program.
Output parameters, for example cell density may be

relevant for the “within batch extension” of the cell cul-
ture process and are less easy to control. For transfer
steps to the next bioreactor cell density is usually a key
parameter and is then part of the trending program.
A trending procedure has to be defined and may cover
statistical rules (e.g. setting of +/- 2SD or 3-SD-ranges
around the mean, Nelson rules, Cpk value calculations).
Data may be plotted in SPC charts which present the
data of various batches per product for one parameter
and one process step.
The trending procedure may occur periodically over the

year to get an overview on the long-term performance
of the biotechnological process. The actual trending pe-
riod may be compared to historical trending periods,
e.g. by visually comparison or in comparison of Cpk
85
values. Trending results may be discussed within the
PQR and/or APR.
The precision of the data trended is at least the same

as defined for the specifications/parameter ranges. A
more precise trending of data is preferable whereas the
scientifically feasible precision of the measurement or
analytical method has to be considered.
Trends may be identified using statistical techniques.

One way may be to set control limits for the parame-
ters based on a +/- 2SD or 3-SD-ranges around the
mean, which may serve as warning or action limits, but
also other ranges may be applied.
In connection with control limits Nelson rules may be

applied. Nelson rules are a method in process control
of determining if some measured variable is out of con-
trol (unpredictable versus consistent).
http://en.wikipedia.org/wiki/Nelson_rules).
They allow detection of unusual patterns in the data
stream.
Another way is to apply process capability analyses. An

overall Cpk may be calculated for a longer trending pe-
riod. The Cpk value is defined by the distance of the
process mean to the nearest specification limit(s) (ei-
ther the upper limit USL or the lower limit LSL or both)
normalized by 3 times the standard deviation SD.

86
It is estimated by
USL x
Cpk when only USL is given
3SD
x LSL
Cpk when only LSL is given
3SD
Minimum USL x , x LSL

Cpk when LSL and USL
3SD
are given
A Cpk greater then 1.33 is generally regarded as suffi-

cient for a capable process. A Cpk value between 1 and
1.33 indicates borderline capability. For a Cpk value
less then 1 the processes capability is not sufficient.
On smaller time scales a moving Cpk (mCpk) may be

suitable in order to detect out-of-trend situations. The
moving Cpk is a hybrid procedure incorporating the
moving window technique in connection with process
capability analysis. A mCpk control chart is set up by
moving a window of fixed span (e.g. 5 to 10 values)
over the data stream and plotting the Cpk values suc-
cessively. Especially for small spans the variation of the
mCpk will be high and will signal too often that the
process is inadequate, even when it is not. Therefore, a
lower threshold for the mCpk of 0.5 to 0.6 (compared
to 1 or 1.33 in long term analyses) has proven practical
to decide whether the process is capable or not.

87
Shewart
Shewart
Control
Control
Chart
Chartwith
upper
with and
lower
upper
control
and
limits
lower
Control
Control
chart
chart
present-
present-
ing
ing
moving
moving
Cpk
Cpk
Figure 1: Shewart Control Chart showing e.g. the product con-

centration for a number of lots produced over time. The center
line (black dotted line) represents the mean of all results over
time in addition to the upper and lower control limits (red dot-
ted lines). The control limits are calculated using a ±3 SD from
the center line. Furthermore the upper and lower specifica-
tions are marked as red broken lines. The process is not cen-
tered, but slightly shifted to the upper specification. The con-
trol chart presents all data from historical lots, the red vertical
lines present the trending period to be discussed. The lower
part of the chart presents the moving Cpk chart in parallel to
the Shewart control chart. For this data set the moving Cpk
calculations demonstrate no violations which means that no
results have a Cpk < 0,5.

88
Figure 2: Shewart Control Chart showing e.g. the impurity

concentration for a number of lots produced over time. The
center line (black dotted line) represents the mean of all re-
sults over time in addition to the upper control limit (red dot-
ted line). The control limit is calculated using a +3 SD range
from the center line. The lower control limit is not presented
as it is calculated to be < 0. Furthermore the upper specifica-
tion is marked as red broken line; a lower specification is de-
fined as “0”. The process is not centered, but shifted to the
lower specification which is expected for an impurity. The con-
trol chart presents all data from historical lots; the red vertical
lines present the trending period to be discussed. The lower
part of the chart presents the moving Cpk chart in parallel to
the Shewart control chart. For this data set the moving Cpk
calculations demonstrate no violations which means that no
results have a Cpk < 0,5.

89
Figure 3: Shewart Control Chart showing e.g. a product quality

parameter for a number of lots produced over time. The upper
and lower specifications are marked as red broken lines. Con-
trol limits (±3 SD from the center line) are not presented as
they are calculated to be wider than the specifications. The
results of this parameter are reported in 4 conditions only due
to the precision of the assay and the rounding. The process is
not centered, but slightly shifted to the lower specification.
The control chart presents all data, the red vertical lines pre-
sent the trending period to be discussed. The lower part of the
chart presents the moving Cpk chart in parallel to the Shewart
control chart which shows two violations.
The application of statistical rules in trending procedures shall

be investigated thoroughly and on a scientifically sound basis
as plenty of violations may occur resulting from the variety of
parameters with different precisions and by applying multiple
rules in connection with large amounts of data. For example a
set of 30-50 historical results each of 2-5 product parameters
are plotted in a control chart and Nelson rules 1-4 are applied.
Then the number of Nelson violations is registered per rule
and the violations are investigated with regard to known pro-
90
cess and method deviations.
Preferably statistical rules shall be applied which indicate the

relevant violations and highlight process abnormalities, drifts
and changes in product quality.
For complex, time-consuming biotechnological processes usu-

ally a lot of parameters are analyzed or controlled for the pro-
duction of one batch.
It is important to set up trending program based on a set of

parameters which is relevant for product quality and focus on
the consistency of the manufacturing process over time.

91

92
4.0 Special Case: Legacy Processes

Jean Denis Mallet
Introduction
There are few doubts that when the American FDA published
in 2003 the draft report on “Pharmaceutical cGMPs for the
21st Century; a risk-based approach” [a] that document trig-
gered for some real changes in the way our industry validates
its processes. However, consideration could be given that it
may not be because the FDA advocated so intensively for it
that the new paradigm finally arose, but more probably be-
cause the related supporting manufacturing and the analytical
technologies were or unavailable or not really mature in the
pharmaceutical world before the end of the 20th Century.
Thus, it can be noted that the practical implementation of a

number a Pharmaceutical Quality Tools, final product testing
tools and more and more generally process controlling tools
now used in the Process Analytical Technology (PAT) approach
have taken a quite long time to emerge and to be accepted as
the next standards. For example, numerous Pharmacopoeial
monographs described the use of Thin Layer Chromatography
(TLC) until the time High Pressure Liquid Chromatography
(HPLC) has been itself challenged by the next generation of
that technique, Ultra Pressure Liquid Chromatography (UPLC) !
This should remind us that, conceptually and historically (and

this 2002 paper relies heavily on historical aspects of technol-
ogy and technological changes) the pharmaceutical industry is
conservative and tends to consider the equation “ legacy =
quality ”. Still today we, in the pharmaceutical industry, are
prone to think: “ Why running the risk to change something in
our established procedures, processes and dossiers as they
are already filed in the Competent authorities drawers ” ?

93
A brief history of validation
We all remember that the very first formalized document to
the (current) good manufacturing practices had been issued in
the United States in June 1963 (FR 63-6336). From this point
the concept of documenting all operations related to the man-
ufacture, testing, release and distribution of pharmaceuticals
spread all over the world. However it took several years to
reach most of the industrialized countries probably because
every authority had then the will of developing a “proprietary”
GMP standard rather than to adopt the existing one(s).
However we (especially in Europe) have to acknowledge that

the origin of several of our good manufacturing practices
guidances is undoubtedly rooted in the transatlantic cGMPs.
This, especially, for the word “validation” which appeared with
the amended version of the FDA’s cGMP principles published in
the Federal Register dated 29 September 1978. There, the
validation concept was clearly associated with the computer-
ized data (211.68), the supplier’s test results (211.84), the
manufacturing processes (211.110) including especially the
sterilization processes (211.113) and last with the analytical
methods (211.165 & 211.166). Unfortunately no definition of
the words “to validate” or “validation” was given in the Part
210 (Glossary of terms). Therefore a very interesting paper [b]
based on a conference delivered in June 1978 by Bernard T.
Loftus, a Director in the FDA, gives then some light on what
was expected by the agency at the time.
Acknowledging that FDA did not define the term “validation”,

he referred first to the then Compliance Program which men-
tioned that “A validated manufactured process is one which
has been proved to do what it purports or is represented to
do”’, and he added that, according to the FDA 1978 thinking
that “The proof of validation is obtained through the collection
and evaluation of data preferably beginning from the process
development phase and continuing through into the produc-
tion phase. Validation necessarily includes process qualification
94
(the qualification of material, equipment, systems, buildings,
and personnel) but it also includes the control of the entire
process for repeated batches or runs. […]”. No number of
batches was mentioned by B. Loftus. But we certainly neglect-
ed one important part of this sentence: “and continuing
through into the production phase”. For sure, we will re-
discuss it at the light of the latest US(FDA) and European
guidances at the end of this article.
Because of the need to define and formalize the way the

pharmaceutical industry had to validate its processes, FDA-
CDER and FDA-CDRH jointly developed in 1983 a “Guideline
on general principles of process validation that was finally pub-
lished in May 1987. This document contained the following
definition for (process) validation: “Process validation is estab-
lishing documented evidence which provides a high degree of
assurance that a specific process will consistently produce a
product meeting its pre-determined specifications and quality
characteristics“. It also contained a number of other definitions
that may have not all clearly understood or used within the
pharmaceutical industry: “installation qualification”, “process
performance qualification”, “product performance qualifica-
tion”, “prospective validation”, “retrospective validation”, “vali-
dation protocol” and the last but not the least “worst case”.
However, at least from the opinion of the author, the imple-

mentation of this guideline was handicapped with the fact that
the concept of “qualification” (of tools and equipment) and the
concept of “validation” (of processes) while normally expected
to be easily assimilated by the industry were frequently mixed-
up even sometimes in certain Corporate policies.
We will see hereunder that definitions are one of the most

important aspects when designing a validation plan.
The 20th Century (legacy) approach to validation

An important date in the FDA history on good manufacturing
95
practices enforcement is the February 1993 decision of the
New Jersey Court elaborated by Judge A. Wolin in the now
famous USA vs Barr Laboratories. During the late 80s, repeat-
ed failing FDA inspections in the manufacturing sites operated
by Barr Laboratories led to this Court’s Ruling. Amongst twelve
clear rules, Judge Wolin ruled that (prospective and concur-
rent) process validation should be based on at least three
(successful) consecutive batches. Therefore, since 1993, this
“magic” number (three) appeared in most validation protocols
established in the pharmaceutical industry while some claimed
that there is (was) no statistical value with this number. Also,
the Judge Wolin mentioned then, that for retrospective valida-
tion, a company had to compile data from 20 to 30 batches.
At this stage we have to cross back the Atlantic Ocean to have

a look on the then available good manufacturing practices
guidances existing in Europe. Most of the European authorities
adhered to the concept of validation and mentioned it in their
guidances during the 80s. The famous PIC guidance PH 5/89
which was then used by the European Commission to prepare
the first EU-GMP guide contained both a definition for the vali-
dation of processes and some simple requirements rendering
process validation mandatory in European Member States.
The EU-GMP guide was completed in 2001 (the year before
FDA issued its 21st Century Initiative) with Annex 15 “Qualifi-
cation and validation”. This document considered (in § 25)
that 3 consecutive batches/runs would be generally acceptable
for prospective validation. The revised Annex 15 has been
quite recently published in March 2015 and, while, stating that
the number of batches should normally be based on a risk
approach, ”it is generally considered acceptable that a mini-
mum of three consecutive batches manufactured under rou-
tine conditions could constitute a validation of the process”.
Therefore, the current generation of pharmacists and pharma-
ceutical engineers (mostly educated during the 20th Century)
has been with this magic number: 3 consecutive batches are
needed (and are sufficient) for the validation of a process.
96
And, up to now –meaning up to the publication of the new set
of validation guidances [c], [d], no one validation officer would
have really tried to establish validation protocols that would
have been called for a different scheme ! This would have
noted even if, on October 25th, 2005, the European Commis-
sion published a new version of the Chapter 1 of the European
GMP guide introducing (§ 1.5) the Concept of Product Quality
Review where (and very few professionals actually noted it)
the new requirement was clearly calling for the compilation of
elements of a… continuous verification of the validated state
of their pharmaceutical products.
A brief analysis of the 2011 guidance

Since 2003, when the FDA published its “Pharmaceutical
cGMPs for the 21st Century : a risk-based approach” [a] draft
report, it was clear that a number of ideas and their related
supporting documents needed to be revisited. Consequently,
this happened in January 2011 for the 1987 “Guideline on
General Principles of Process Validation” with the publication
of a new guidance for industry incorporating the content of
the recent ICH guidelines Q8, Q9 and Q10.
The ICH Q8 guideline is probably the one that was the more
impacting the old legacy concept of validation. Indeed, the Q8
philosophy is that quality, safety and efficacy of a medicine are
designed into the product while quality cannot be assured by
simple in-process and finished product testing. Consequently,
FDA redefined the approach to process validation as a contin-
uous collection of data taking place all over the lifecycle of the
product and of its manufacturing process. Revised FDA expec-
tation is that the validation of a given pair (product/process)
would be developed in three stages: 1. Process Design, 2.
Process Qualification, and 3. Continued Process Verification. A
strong emphasis is also given that a real successful validation
will be only possible if the owner of the pair (the manufactur-
er) has a real product knowledge and process understanding
and of the different possible sources of variation.
97
In order to achieve a science-based process validation the FDA
guidance recommends the companies to integrate a multidis-
ciplinary team approach (see also Q9), to write and to conduct
protocols according to sound scientific principles and to assess
its degree of control over the product attributes.
According to the new 2011 guidance the three steps towards

the process validation “status” are:
Process Design, embedding a) the building and capturing pro-

cess knowledge and understanding, and b) the establishment
of a strategy for process control ;
Process Qualification, embedding a) the qualification of the

facilities, utilities and equipment used in the process, b) the
issue of a process performance qualification programme or
“PPQ” of all items pertaining to the manufacturing process
supported by an intense sampling and testing programme, c)
a written protocol covering all the inputs and outputs generat-
ed during the process, and d) the PPQ approved report ;
A never-ending “Continued Process Verification” stage embed-

ding an ongoing programme to obtain a continual assurance
through data collection and compilation that the manufactur-
ing process and the product(s) remain within their attributes
during commercial manufacture.
One can draw the following diagram representing this “new”

approach to process validation that sometimes deliberately
borrows vocabulary to the EU-GMP guidances:
When reading that diagram it is quite easy to identify the con-
tinuity of FDA’s logic through the decades. We can see that
the white core of the new guidance corresponds to the classi-
cal qualification approach into DQ (selecting of utilities and
equipment), IQ (verification that utilities and systems are built
and installed in compliance with their specifications), and OQ
& PQ (verification that systems and equipment operate within
98
their anticipated range, if needed through challenging loads
comparable to the product itself). This step has to be com-
pleted with the Process Performance Qualification which, actu-
ally, is quite similar to the European Process Validation
The March 2015 revised “Annex 15” to EU-GMP

One should simply read the reasons for change as mentioned
on the cover page of this revised European Annex 15 : “since
[it was first] published in 2001 the [international] manufactur-
ing and regulatory environment has changed significantly and
an update is required to reflect this changed environment.
[Especially ) this revision takes into account changes to other
sections of the EudraLex, Volume 4, Part I, relationship to Part
II, Annex 11, ICH Q8, Q9, Q10 and Q11, QWP guidance on
process validation, and changes in manufacturing technology”.
In practice what really changed in the European approach can

be found in paragraphs 1.7 (emphasis on quality risk man-
agement approach), 3.2 (introduction of the term URS: User
Requirements Specifications), 3.4-3.7 (introduction of the FAT
& SAT steps), 5.2 (stating that the Annex 15 should be used in
conjunction with the EMA recommendations on Process Valida-
tion), and 5.3. This paragraph 5.3 - combined with 5.23 - is
the real key because is now states that, depending on how the
manufacturing processes have been developed (or against a
traditional approach or against a continuous verification ap-
proach), the validation strategy can (should) be, respectively,
or based on a traditional process validation approach like the
one described in the first 2001 edition of Annex 15 or based
on a continuous process verification. Another point is that ret-
rospective validation is no longer an acceptable approach.
Which is more than logic if we consider the introduction of the
Product Quality Review in 2005. Also, and quite logically, the
new Annex states (5.26) that a hybrid of the traditional ap-
proach and of the continuous process verification could be
used where there is a substantial amount of product and pro-
cess knowledge and understanding
99
Last, the new Annex 15 introduces the concepts of verification

of transportation and … of validation of packaging.
A debate on legacy (traditional) validation approach versus the
21st Century validation scheme
As usual when a new guidance is set up the trend is to oppose

it to the previous existing requirements. However, there is
actually little difference between the 1987 FDA document and
the 2011 FDA document. First, their title is identical, second
they are “compatible” in that, schematically, the second stage
of the new 2011 approach corresponds to the 1987 core.
When reading the 1987 document one can found the “PPQ”
wording and we can see that the concepts are similar.
Nevertheless, the 21st Century guidance is more developed

than the 20th Century guidance: in the older, there was ap-
parently no absolute requirement for the process design stage
and for the continual process verification. This is quite under-
standable for the design considerations, because the ICH Q8
recommendations only reached version 4.3 in November 2004.
It is not so true for the “Continued Process Verification” be-

cause the requirement to evaluate annually the manufacturing
and testing data from the records and reports pertaining to
each pair (product/process) dates back to… 1978 !
Conclusion
While this chapter has been revised since the first ECA edition
in 2012, its conclusion is not really changed. Practically, the
modification of the conclusion consists into the wording used.
The FDA guidance calls for Continued Process Verification

while the European guidances (EMA, 2014 and EC 2015) re-
quire that a Continuous Process Verification is made. From
the author’s point of view : a) there is no real philosophical
difference in between the two sides of the Atlantic Ocean and
100
both regulatory authorities expect the process validation no
longer being a one-off exercise (as it is also written since 2003
in the World Health Organization good manufacturing practice
guide referenced TRS 908 and; b) there is no need, for legacy
(traditional) products to dismantle the existing processes and
products validation systems.
However, as soon as some new knowledge is gained from the

current manufacturing operations or from the re-validation
runs performed, they should be added into or hybridised with
the concerned process-dedicated knowledge management
system. Little by little this will probably give the manufacturers
the possibility of formalizing a “reverse design” approach.
To conclude it should be underlined that nothing was really

invented but only formalized in the 2011 guidance when one
reads the Bernard T. Loftus 1978 paper again:
“The proof of validation is obtained through the collection and

evaluation of data preferably beginning from the process de-
velopment phase and continuing through into the production
phase. Validation necessarily includes process qualification
(the qualification of materials, equipment, systems, buildings,
and personnel) but it also includes the control of the entire
process for repeated batches or runs. […]”.
Executive summary
Since the late 70s “validation” is one of the most powerful
GMP word that can be used by regulators and in inspectors’ or
auditors’ reports. Or, much more precisely the two words “not
validated” that trigger for a long and probably litigious discus-
sion case in between industry and agencies.
Why this? Probably because being involved in the protection of

life and protection of health our industry has no ethical right to
worsen any patient’s situation. Then a “validated” product (i.e.
a product obtained through a validated process) is considered
101
as the minimum quality that the pharmaceutical industry has
to offer. People who use pharmaceutical products have to be
assured that they will be cured or find relief without any ad-
verse effects arising from the manufacturing processes. Here,
validation is used to demonstrate that our industry has worked
well in producing effective products and any failure during a
validation can really be considered as a critical issue.
The new validation paradigm about on-going verification is

now mandatory for new products and more than encouraged
for legacy (traditional) products on both sides of the Atlantic
Ocean by the FDA and the EMA. However, because not all
processes may be converted to the new QbD approach, the
traditional process validation as described in the recent EMA
guideline should remain acceptable during the coming years or
even decades and the dossier will include: “the description of
the manufacturing process, the tests to be performed and
acceptance criteria, a description of the additional controls in
place and the data to be collected”. One may say “as usual”…
References
[a] “Pharmaceutical cGMPs for the 21st Century. A Risk-Based

Approach, FDA, August 2002
[b] Validation and Stability, Bernard T Loftus, PDA Journal

1978, Vol 32, N°6
[c] Guidance for Industry Process Validation: General Princi-

ples and Practices, FDA, January 2011
[d] Guideline on process validation for finished products in-

formation and data to be provided in regulatory submissions
CHMP & CVMP, EMA, dated 27 February 2014
[e] Annex 15 (Qualification and Validation) to the EU Guideline

for Good Manufacturing Processes, dated 30 March 2015
102
5.0 Statistical Background/

Manufacturing Excellence
5.1 Statistical Process Control (SPC)
The following is an excerpt from a translation of the Aide
memoire “Inspection of qualification and validation in pharma-
ceutical manufacture and quality control” from the Central
Authority of the German Federal Laender for Health Protection
Regarding Medicinal Products and Medical Devices (Zen-
tralstelle der Länder für Gesundheitsschutz bei Arzneimitteln
und Medizinprodukten). The translation was prepared by Con-
cept Heidelberg on behalf of the ECA Foundation.
Control charts serve for presenting variation. Therefore, they

are a tool to control processes. Variations in processes can
have random or systematic causes. It is possible to prepare
control charts for quantitative (measurable variables) and
qualitative characteristics (attributive properties).
Figure 1: Control chart

103
Random influences cause a natural variation on the control
chart. Usually, they cannot be omitted; for instance, system-
related fluctuations in temperature, dosing precision.
Systematic influences lead to a slow shift of variation. Exam-

ples are faults due to equipment malfunction, wear or opera-
tion errors. In the case of systematic disturbing influences,
corrective action must be taken as concerns the process has to
be taken.
Control charts aim at recognising systematic disturbing influ-
ences and the success of the measures taken.
A control chart is a graph of the course of original values or of

the statistical parameters calculated from this data.
In practice, the following types of control charts have become

established:
▪ mean chart
▪ original value chart
▪ standard deviation chart
▪ median chart
To prepare a control chart, samples of the size n of consecu-

tively produced products are extracted at regular intervals
from the ongoing process. They are then examined with refer-
ence to certain quality-related parameters. The results are
entered directly above the horizontal time axis.
The set point defines the desired position of the process. Apart
from the graph on the chronological course of variation of a
process, all causes and measures influencing the variation are
noted (so that correlations between the process and its sur-
rounding circumstances can be understood retrospectively).
These records and evaluation of the process data give an indi-
cation of the variation and its influences.

104
The control chart contains action limits and alert limits. Basi-
cally, action and alert limits can be defined in two ways:
1. process control chart

2. acceptance control chart
In the case of process control charts, action and alert limits

are not defined on the basis of pre-defined specifications, but
extrapolated from the knowledge of already known or previ-
ously performed processes (and adjusted periodically to the
extent new data gives rise to do so).
Position and/or variation values of quality-related process pa-

rameters (such as temperature, pressures and torques) are
taken as a basis for parameters.
In the case of acceptance control charts, action and alert limits

are defined on the basis of pre-defined tolerance limits (e.g.
specifications of intermediate or final products such as water
content, tablet hardness, impurity level).
Usually, action limits are set tighter than tolerance limits in

order to ensure timely intervention, so that the product is in
conformity with the specifications throughout the entire pro-
cess.
Apart from limit violations, other deviations in the curve can

also be observed which require intervention in the process:
x data trends (continuous ascent or descent of a set of pa-
rameters)
x so-called "runs" (a set of parameters lies exclusively above
or below the line of symmetry of the control chart)
x abnormal distribution (e.g. periodic course) of a set of pa-
rameters
Figure 2 Trend
105
Figure 3: Pattern
A "pattern" is a not randomly shaped curve, such as the pe-

riodic "oscillation" around a pre-defined mean. It can be at-
tributed to temperature fluctuations, for example, causing
parts manufactured to be bigger one time and smaller the
next.

106
Figure 4: Run
If seven measuring points are above or below the pre-defined

mean, a new actual mean was practically created. This is
called a "run". It can indicate, for example, that the punch of
a tablet press has been damaged and now produces tablets
which are bigger or smaller.

107

108
Annex 1: Literature
1
FDA, “FDAs Proposed Revisions in Drug GMPs”, Federal Register , Vol. 41, No. 31, Feb. 13, 1976,
pp. 6878-6894
2
Chapman, K. G., A suggested validation lexicon, Pharmacweutical Technology, August 1983
3
Loftus, B. T. and Nash, R. A. Pharmaceutical Process Validation, Marcel Dekker, Inc. New York –
Basel 1984
4
Carleton, F. J. and Agalloco, J. P. ‘Validation of Aseptic Pharmaceutical Processes, Marcel Dekker,
Inc. New York – Basle 1986
5
FIP-Richtlinien für die gute Validierungspraxis (FIP Guideline on Validation Practices) Phar-
mazeutische Industrie, 1980, 42 (10), 982 – 984 , in German language
6
http://ec.europa.eu/health/documents/eudralex/vol-4/index_en.htm
7
ISPE Baseline Pharmaceutical Engineering Guides, Vol. 5: Commissioning and Qualification Guide,
(2001)
8
http://www.ispe.org/jett/jett-sample-documents
9
Schneppe, T. und Müller, Rainer, H., Qualitätsmanagement und Validierung in der phar-
mazeutischen Praxis (Quality Management and Validation in Pharmaceutical Practice), 2. erw.
Auflage 2003, Editio Cantor Verlag, Aulendorf, in German language
10
Gardner, P.B., Outsourcing Validation Results in Significant Cost Saving, Pharmaceutical Engi-
neering, p. 24-28, Sept/Oct. 1998
11
Gold Sheet, Vol. 27, No. 2, February, 1993
12
Marrer – GMP Inspektionen (GMP Inspections), Pharm. Ind, 62, No. 7 (2000), in German lan-
guage
13
Dr. M. Hiob, "Ministry of Employment, Social Affairs and Health of Schleswig Holstein, Kiel,
Germany, personally communication
14
http://ec.europa.eu/health/files/eudralex/vol-1/reg_2003_1084/reg_2003_1084_en.pdf
15
16
17
21 Code of Federal Regulations Part 314.70 - Supplements and other changes to an approved
application
18
http://www.fda.gov/downloads/Drugs/.../Guidances/UCM217043.pdf
19
http://www.fda.gov/downloads/drugs/guidancecomplianceregulatoryinformation/guidances/ucm07
7097.pdf
20
http://www.fda.gov/downloads/drugs/guidancecomplianceregulatoryinformation/guidances/ucm12
2871.pdf
21
Akers, J. Simplifying and Improving Process Validation, PDA Journal of Pharmaceutical Science &
Technology, Vol. 47, No. 6 (1993)
22
Anisfeld, H. Michael, Validation – How Much Can the World Afford? Are we Getting Value for
Money, PDA Journal of Pharmaceutical Science & Technology, Vol. 48, No. 1 (1994)
23
Selby, D., Can Validation Improve the Bottom Line?, Pharmaceutical Engineering, Vol. 19, No 6
(1999)
24
Powell-Evens, “Streamlining Validation”, Pharmaceutical Technology Europe, Vol. 10, no. 12, pp.
48-52, 1998,
25
Stinecker et. al, Anlagenqualifizierung um jeden Preis? (Equipment qualification at any costs?),
Pharm. Ind, 61, No. 1, 66 – 68, (1999), in German language
26
Crosson, Campell, Warren, Journal of Validation Technology 08/03
27
White Paper on Risk-Based Qualification for the 21st Century, ISPE, 9 March 2005. 9. Branning,
R., et al
28
ASTM, Specification, Design, and Verification of Pharmaceutical and Biopharmaceutical Manufac-
turing Systems and Equipment” (E2500-07, today E2500-12)
29
ISPE, GAMP 5 A Risk-Based Approach to Compliant GxP Computerized Systems (2008)
30
http://www.fda.gov/downloads/Drugs/Guidances/ucm070305.pdf
31
http://www.ema.europa.eu/docs/en_GB/document_library/Other/2009/10/WC500004890.pdf
...
109
32
http://www.ich.org/fileadmin/Public_Web_Site/ICH_Products/Guidelines/Quality/Q9/Step4/Q9_Gui
deline.pdf
33
http://www.ich.org/fileadmin/Public_Web_Site/ICH_Products/Guidelines/Quality/Q10/Step4/Q10_G
uideline.pdf
34
http://www.ich.org/fileadmin/Public_Web_Site/ICH_Products/Guidelines/Quality/Q11/Q11_Step_4.
pdf
35
http://www.ema.europa.eu/ema/index.jsp?curl=pages/regulation/general/pat_qa.jsp&mid=WC0b0
1ac058006e0f2
36
http://www.ema.europa.eu/docs/en_GB/document_library/Scientific_guideline/2009/09/WC50000
2835.pdf
37
http://www.gmp-compliance.org/enews_01951_EMA-revises-its-Process-Validation-
Guidelines.html
38
5399.pdf
39
136.pdf
40
GMP-News 05 December 2012, http://www.gmp-
compliance.org/enews_03426_EMA%20publishes%20Concept%20Paper%20on%20the%20Revisi
on%20of%20Annex%2015.html
41
http://ec.europa.eu/health/files/eudralex/vol-4/2015-10_annex15.pdf
42
Beuth Verlag, DIN EN ISO 9001:2008, Quality management systems – Requirements (2008-12)
43
Beuth Verlag, DIN EN ISO 13485:2012, Medical Devices - Quality management systems -
Requirements for regulatory purposes (2012-11)
44
GHTF, Quality Management Systems – Process Validation Guidance, GHTF/SG3/N99-10:2004
(Edition2): http://www.imdrf.org/docs/ghtf/final/sg3/technical-docs/ghtf-sg3-n99-10-2004-qms-
process-guidance-04010.pdf
45
http://www.imdrf.org/
46
GMP-News dated 21 February 2007 (http://www.gmp-compliance.org/eca_news_891.html), ECA
47
Holloway, Ian, Memoscript: Regulatory GMP Inspections – The views of an MHRA inspector,
How to pass EU and FDA Inspections & GMP Compliance Inspections, ECA Education Course, 25-
27 May 2011, Praque
48
Unterkofler, B., Memoscript: Die AGES – Erfahrungen mit Inspektionen (The AGES – Experiences
with Inspections), Die neue AMBO, Concept-Heidelberg, 25 September 2008, Vienna, in German
language
49
Klein, M., Memoscript: Risikomanagement in der Qualifizierung (Risk Management in Qualifica-
tion), 3 rd . German Risk Management Conference, Concept-Heidelberg, 10/11 June 2008, Ham-
burg, in German language
50
Matthew, G. Roberge, Factory Acceptance Testing (FAT) of Pharmaceutical Equipment, Pharma-
ceutical Engineering, p. 9-16, Nov./Dec. 2000
51
James, Phil, Integrated Validation: A Way of Streamlining Projects, to Reduce Project Validation
Time Cost, Pharmaceutical Engineering, p. 72-82, Jan/Feb 1998
52
R. G. Kieffer, Validation, Risk Benefit Analysis, PDA Journal of Pharmaceutical Science & Tech-
nology, Vol 49, Nr. 5 (1995)
53
R. G. Kieffer, S. Bureau, A. Borgmann, Application of FMEA in the Pharmaceutical Industry,
Pharma Technology Europe, Sept., 1997
54
Powell-Evans, K., Design Qualification : Mists, Mirrors and Witchcraft, Pharmaceutical Technolo-
gy Europe, 1999
...
110
55
Pommeranz, S. und Hiob, M., Qualifizierung/Validierung nach Annex 15 des EG-GMP-Leitfadens
– Teil 2, (Qualification/Validation according Annex 15 of the EC GMP Guide – Part 2), S. 26-50,
Pharma-Technologie-Journal Nr. 1095, GMP-/FDA-gerechte Validierung, 2nd Edition 2010, Editio
Cantor Verlag in German language
56
21 Code of Federal Regulations Part 211 Current Good Manufacturing Practice for Finished
Pharmaceuticals
57
Agalloco, J. PDA Journal of Pharmaceutical Science & Technology, Vol 49, Nr. 4 (1995)
58
GMP-News of 25 November 2009 (http://www.gmp-
compliance.org/eca_news_1798_6350,6264,6230,6351,6288.html), ECA
59
Loch, Memoscript: Computer Validation in the Medical Device Field, ECA Course Medical Devices
Validation Manager, 17/18 January 2012, Heidelberg
Last access: November 2, 2015

111
112
Annex 2:
Validation - Revision of Annex 15: How does the
Industry see the Changes?
Results of an ECA Survey
With the publication of a draft on the revision of Annex 15 at

the beginning of February 2014, changes in the area of
validation/qualification were to be expected. Even then, the
ECA Academy started an industry survey. Now the final
document is published and the changes will become valid on 1
October 2015. But how does the industry see the new
requirements?
A total of 53 people took part in the survey, the majority of

which answered all the questions. Almost 50% (49.0%) of the
participants in the survey work for companies with more than
500 employees while 40.9% work for companies with more
than 101-500 employees.
A total of 61.2% of the participants are employed in the

pharmaceutical industry. 16.3% of the them work in the API
industry (chemical API and biotechnologically produced API),
for further distribution, see Figure 1.
113
ECA Valid
dation Goo
od Practice
e Guide
Figure 1 Wh
hat products
s do you man
nufacture in your company?
25% of the e participan

nts that speccified their rrole in the
company w work in the "quality" seector. 21% w work in valiidation
and the rem maining parrticipants arre from a raange of sectors
(development processs, various manager
m funnctions, etc.).That
is all as farr as demogrraphic data is concerneed.
It is ratherr surprising that the subject of acttive substan

nces was
addressed in the final revision. And in this reegard, the first
f
question w was: How do o you assess the statem ment that Annex
A
15 can opttionally be used
u upplement iin the area of
as a su
active subsstances with hout additio
onal requireements in th
he EU
Guidelines Part II? 1/3 3 (34%) of the particippants consid
der
Annex 15 tto be a posiitive supplement to Parrt II. For 1//3, it is
not clear and just und der 1/4 (22%%) find thatt Annex 15 should
also be com mpletely ap pplicable to active substtances. 10%
% do
114
not consider Annex 15 necessary in the area of active

substances.
The second question dealt with the subject of conditional

approval: Will you use the option of conditional approval to
the next stage in case of deviations with a documented
assessment?
More than half of the participants (54.2%) will use this option,
just as they did in the past. Almost 1/3 (29.8%) will accept
this approach only in "exceptional circumstances". 14.6% will
use it even more frequently than in the past. Only 2.1% will
not allow this approach.
With regard to question 3 ("How do you assess the fact that

FAT and SAT are only "could" options?") 68.8% regard it as a
good idea that the tests are only optional. Yet almost 1/5
(18.8%) consider these tests mandatory. For 10.4% it is not
clear what it is about. 2.1% do not consider the tests to be
necessary.
50% of the participants assessed the option of combining

qualification stages (IOQ, OPQ) in question 4 as a good idea
and will also do so. For just under 1/3 (29.2%) it is already
common practice. 12.5% would not like to use combinations
and for 8.3% the approach is not clear.
As part of revising Annex 15, the definition of PQ has also

changed. 68.2% answered question 5 (Do you understand the
new PQ definition?) in the affirmative, for 1/5 (20.8%) the
definition is not clear and 10.4% also answered no. This
means that almost 1/3 of the participants are not sure about
the new PQ definition.
The answers to question 6 on how the participants assess the

fact that a minimum of 3 validation batches can show the
115
ECA Valid
dation Goo
od Practice
e Guide
validation o
of a processs were relattively evenlyy distribute
ed.
39.6% wou uld like furtther explana
ations. 37.5
5% will conttinue to
"always" mmake 3 valid dation batchhes. 22.9% believe tha at the
requiremen nt may lead d to confusio
on (see Figuure 2).
Figure 2: Diistribution of the answerrs on 3 valid

dation batche
es
47.9% ansswered yes to question n 7 (Do you understand d what

the revision means byy hybrid appproach for pprocess
validation??). 52.2% in
n total answ
wered "not cclear" or "no
o".
More than half of the participantss assessed removing th he

routine validation (que estion 8) ass understan
ndable as it is
replaced byy the ongoing process verification n. But 36.6%% would
still like to regularly re
evalidate ba
ased on a ri sk assessment,
and 6.4% even without a risk analysis. Just as many
116
participants will only do this in a sterile area (e.g. at Medial

Fills).
In relation to question 8, we asked question 9 to find out if the

participants expect the ongoing process verification to involve
additional work. 60.4% answered yes. For 1/4 the question
was not clear and 14.6% do not believe it will generate
additional work. We asked the 14.6% to give reasons for their
assessment. In summary, it can be said this group considered
this requirement already currently fulfilled (e.g. through APR,
current monitoring).
The answers to question 10 were also interesting (Do you

think that the requirements of packaging validation will require
more work?). While 35.4% answered yes, it is (still) not clear
to just as many (35.4%) and 29.2% answered no. We also
asked those who answered yes to substantiate their answer.
In summary, it can be said that the answers reveal that little
(or less) was validated in this area in the past.
69.4% answered question 11 "How do you handle the fact

that only PDE is stated as the acceptance criterion for the
cleaning validation (even for old products)?" with "not clear
yet". 22.4% will start pharmacological investigations to obtain
PDE calculation data. 8.2% want to keep the old limit values
since they do not have data to calculate PDE values.
Question 12 relates to question 11 and also deals with the

subject of cleaning validation (How will you deal with the fact
that the number of validation batches for cleaning validation is
no longer specified?)
69.4% answered that they will determine the number of

cleaning validation runs based on a risk assessment. 16.3%
will continue to do 3 runs and implementation is currently still
unclear to 14.3%.
117
ECA Valid
dation Goo
od Practice
e Guide
When askeed question 13, to wha at extent do

oes it conforrm with
the FDA prrocess validation guidance, 68.9%% answered 75-
100%. 30%% answered d 50-75%, 4.4%
4 said < 50%, 6.7%
answered 100% confo ormity.
The answe er to the lasst question (no.

( 14) on how the re
evision
is generallyy assessed is very interesting:
The verdictt is surprisin

ngly positive (very goo
od: 6%, goo od:
52%). 28% % are satisfied. Only 14 4% would liike improveements
made and we specifically asked theset peoplee what
improveme ents they would
w suggesst. More preecise inform
mation
was requessted and a greater com mparison wiith the FDA process
validation g
guidance, particularly
p with
w referen nce to the terms
t
PQ and PPQ. One com mment was even that tthe document
raises moree questionss than it ansswers.
Figure 3: Ge
eneral asses
ssment of the
e revision
118
Conclusion
The participants have an overall positive opinion of the
revision of Annex 15. In total 58% consider the document to
be very good or good. 28% said the revision is "satisfactory".
Some of the new options are already being used by the
industry (e.g. temporary release, combinations of qualification
stages). Easing the requirement for FAT/SAT tests ("could")
compared to the draft was well received. 68.8% find it good.
The new PQ definition is now surprisingly clear (68.2%), on
the other hand, almost 1/3 of the participants are still
unsatisfied with the definition. Just as with the survey about
the draft, there is still considerable uncertainty regarding the 3
validation batches. Here the industry would like more specific
information or to carry on as previously ("always produce 3
batches"). The hybrid approach is also only clear to just under
50% of participants. With the request for ongoing process
verification, the majority of the industry (60.4%) believes this
will result in additional work, while 50% also believe
discontinuing the routine revalidation will result in additional
work. The analysis of the industry on packaging validation is
also varied. Around 1/3 believe this will result in additional
work, but almost as many think the opposite and also 1/3
cannot currently say whether any additional work will arise.
The industry is mostly sceptical about the PDE concept in
connection with the cleaning validation. For 69.4% the
concept is not yet clear and just under 1/5 would like to start
pharmacological investigations to calculate PDE values. The
industry is dealing very pragmatically with the discontinuation
of the three batches during cleaning validation. 69.4% would
like to determine the runs based on a risk assessment. The
majority of participants believe it conforms with the FDA
process validation guidance to a relatively great extent.
119
120
Annex 3
Interpretation of the FDA Guidance on Process
Validation
Part 1: First Draft

On a total of 20 pages and subdivided into 7 chapters, the
FDA described their current thinking (2008) with regard to
process validation. The chapters are subdivided into:
I Introduction
II Background
III Statutory and Regulatory Requirements for Process
Validation
IV Recommendations
V Concurrent Release for Performance Qualification
Batches
VI Documentation
VII Analytical Methodology
I Introduction
The introduction points out expressly that process validation is
connected to a product life cycle. Thus, process validation
includes development towards routine production and routine
production itself. Furthermore, the guidance document intends
to promote modern manufacturing principles, process
improvement, innovations and sound science. It also refers to
ICH Q8, 9 and 10 in connection with the life cycle concept.
The introduction then lists the products subject to the

guideline: "human drugs", "veterinary drugs", "biological and
biotechnology products", "finished products", "combination
products" (drug and medical device), but also "Active
Pharmaceutical Ingredients" (with reference to ICH Q7a).
As a countermove, exclusions are listed as well, e. g. medical
devices and dietary supplements. It is also pointed out that
121
the guideline does not give any information on the

documentation necessary to apply for a marketing
authorisation. The guideline does not apply to the validation of
"automated process control systems".
II Background
The history briefly deals with the FDA Guideline on Process
Validation of 1987, the basic principles of which have been
taken up again in the new draft. Interestingly, the text
expressly mentions the GHTF Guideline on Process Validation
relevant to medical devices as being likewise useful for drug
manufacturing. The current guidance document is based on
experience values gathered since 1987 and also on the FDA
Initiative cGMPs for the 21st century - a risk-based approach.
Here, the text contains another reference to modern
manufacturing techniques and to "risk management" and
"quality management tools" and "concepts" - however, without
going into detail.
What is new is the definition of process validation as

"the collection and evaluation of data, from the process design
stage through production, which establishes scientific evidence
that a process is capable of consistently delivering quality
products". Thus, process validation is now split up into 3
stages:
x Stage 1 "Process Design" (The commercial process is

based on experiences gained from development and scale-
up)
x Stage 2 "Process Qualification" (During this stage, the

reproducible, commercial scale is confirmed on the basis of
process design)
122
x Stage 3 "Continued Process Verification" (This stage is

meant to show that the process is in a state of control
during routine production)
The text states expressly that in practice these 3 stages might

overlap. With emphasis, it urges manufacturers to prove with
a high degree of assurance that the product can be
manufactured according to the quality attributes before a
batch is placed on the market. For this purpose, data from
laboratory-, scale-up and industrial scale are meant to be
used. The data are explicitly meant to cover conditions
involving a great risk of process variation.
For this reason, the manufacturer should
x understand the process variations

x detect these process variations and assess their extent
x understand the influence on the process and the product
x and control such variations depending on the risk they
represent
Again, the text points out expressly that qualification activities

lacking the basis of a sound process understanding will not
lead to an accordingly qualitatively safe product. The chapter
closes by pointing out that the process must be maintained
during routine operation. This includes materials, equipment,
the environment, personnel and changes in the manufacturing
procedures.
III Statutory and Regulatory Requirements for

Process Validation
In this chapter, the FDA refers to paragraphs in 21 Code of
Federal Regulation 210/211 that also have to be applied with
regard to validation (and qualification of equipment):
123
211.100(a), 211.110(a)(b), 211.160(b)(3), 211.165 (a)(c)(d),

211.180(e), 211.42, 211.63, 211,68.
IV Recommendations
This is the central chapter of the guidance. At the beginning,
Good Project Management and good archiving are pointed out
as effective and efficient means for the product life cycle. A
team approach to process validation is mentioned as well, with
a statistician listed as possible team member. Furthermore,
the text reminds of the fact that the full support of senior
management is necessary. All studies conducted within the
framework of process validation should be documented
accordingly and conducted on the basis of sound scientific
principles.
Then the recommended activities in the 3 stages are dealt
with emphatically.
Stage 1 - Process Design

a) Building and Capturing Process Knowledge and
Understanding: In this stage 1, the manufacturing process is
meant to be defined, which will then be reflected in the
manufacturing and testing documentation. Also in view of
Q10, it is expressly pointed out that earlier development
stages do not have to be conducted under cGMP. However,
here, too, the basis should be sound scientific methods and
principles, including Good Documentation Practice. It is
considered to be no regulatory expectation that the process be
developed and tested until it fails, but the combination of
conditions involving a high process risk should be known. In
order to achieve this level of process understanding, among
other things the implementation of design of experiments in
connection with risk analysis tools is recommended. However,
other methods, like classical laboratory tests, are also
considered as acceptable. What is considered to be essential is
the adequate documentation of the process understanding
based on rationales, above all in view of the life cycle.
124
b) Establishing a Strategy for Process Control: Process

knowledge and understanding are considered to be the basis
for process control. Apart from in-process controls, the text
mentions the possible use of Process Analytical Technologies
(PAT) and quotes the corresponding PAT guideline.
Stage 2 - Process Qualification

It is meant to prove that the process design is suitable for
reproducibly manufacturing commercial batches. This stage
has 2 elements: On the one hand the qualification activities
regarding premises and equipment, on the other hand
performance qualification (PQ). Strictly speaking, this stage
encompasses those activities that are currently summarised
under process validation: On the basis of qualified equipment,
it is then demonstrated that the process can create a product
in conformity with the specifications.
The text deals with the constituents of the qualification

activities that are the prerequisites for PQ. Without mentioning
the terms DQ, IQ, OQ, these activities are described as
constituents of the qualification. The necessary documentation
on qualification, too, is dealt with. The description of
qualification activities is possible as individual plans or as part
of a project plan. The possibility to integrate risk management
in order to determine priorities and scope of performance and
documentation is only mentioned as a "can" option. The
contents of the plan should be:
1. Description of the tests

2. Acceptance criteria
3. A schedule
4. Responsibilities
5. Information on the documentation and release of the
qualification results
6. Data on the change control procedure
125
The results are meant to be summarised in a report making

reference to the acceptance criteria in the plan. Both the plan
and the report are meant to be reviewed and released by the
quality control unit.
A specific subchapter is dedicated to the performance

qualification approach, the second element of stage 2 in
process validation. The PQ combines the qualified premises
and equipment as well as the trained personnel with the
commercial manufacturing process. What could in principle be
regarded as a definition of PQ is the document's statement: "A
successful PQ will confirm the process design and demonstrate
that the commercial manufacturing process performs as
expected". The PQ is considered to be an important milestone
within the product life cycle. Its completion is a prerequisite
for marketing, and the decision for marketing the product
should be based on data gained from the commercial
manufacturing scale, if necessary supported by laboratory and
scale-up studies.
The guidance points out expressly that a sound science should

serve as approach to PQ. This goes as far as the FDA requiring
in the guidance: "We strongly recommend firms employ
objective measures (e.g. statistical metrics), wherever feasible
and meaningful to achieve adequate assurance". Insofar,
according to the guidance, the scope of PQ regarding
sampling, additional tests is more comprehensive than in
normal production. Finally, the text points to very specific
aspects of biotechnological manufacture and to PAT
implementations within the framework of PQ, which take
account of a different PQ approach.
The PQ plan (here called "protocol") is meant to discuss the

following points:
126
1. The manufacturing conditions including the parameters to

be set, the process limits and the use of raw material
2. Data that are collected and how they are evaluated
3. Tests to be performed (IPC, release, characterisation) and
acceptance criteria for each significant process step
4. A detailed sampling plan (where, how much, how often)
based on a statistical basis (within a batch and between
batches) equalling a risk analysis
5. Criteria providing a rationale for the question whether the
process constantly produces a qualitative product. Among
them are a description of the statistical methods used for
data analysis (referring to variabilities not only within a
batch, but also between batches) and the description of
the handling of deviations
6. If need be, qualification aspects of the premises and
equipment, training certificates and a verification of the
materials used (raw materials and primary packaging
materials)
7. Validation status of analytical methods used for
measurements in process, for in-process tests and tests on
the finished product
8. Review and release by the corresponding departments and
the quality unit
An individual item regulates the performance of the PQ tests

and reporting: Not new, but still mentioned in the text is the
requirement not to carry out the activities until the protocols
have been released. Changes to the protocols should be
evaluated correspondingly and released by the departments
concerned and by the quality unit. The PQ batches are meant
to be performed by production staff under normal conditions.
This puts and end to discussions about "worst case" conditions
within the framework of the validation runs.
127
Required components of the PQ report:

1. Discussion of the results with cross references to the PQ
protocol
2. A summary and evaluation of the data according to the
specifications in the protocol
3. Evaluation of all unexpected observations and additional
data compared to the protocol
4. Summary and discussion of all "manufacturing non-
conformances", like e.g. deviations
5. Detailed description of corrective actions or changes with
regard to procedures and controls that have already been
laid down
6. Clear indication of a result if the data show that the
process is in compliance with the specifications in the
protocol and that the process is in a sufficient state of
control
7. All necessary reviews and releases by the departments and
the quality unit
Stage 3 - Continued Process Verification

During this stage 3, the objective is to keep up the validated
state of the process also in routine production. For this the
manufacturer is required to establish a system detecting
unplanned process variations. Shifts are meant to be
evaluated accordingly so that the process does not get out of
control. There is a direct reference to 21 CFR 211.180(e) in
order to support this ongoing programme. The data must be
statistically trended, and the analysis be done by a trained
person. The text recommends explicitly that a statistician or at
least an employee trained in statistical techniques works out
the sampling plans and carries out the evaluation of the data
with regard to process stability and process capability. The
evaluation and the trending should be done according to
SOPs. These evaluations are meant to be reviewed by the
quality unit in order to detect changes in the process (alert
128
limits) at an early stage and to be able to implement process

improvements. Also with regard to unexpected process
changes that can also occur in a well developed process, the
guidance recommends "that the manufacturer use
quantitative, statistical methods whenever feasible" in order to
identify and characterise them and investigate the root cause.
Here, too, variations within a batch and between batches are
addressed explicitly. At the beginning of routine production,
the guidance recommends the same scope of monitoring
activities and samples as in the process qualification stage
until enough data have been collected to allow a - statistically
secured - adjustment of this scope.
Data from complaints, OOS results, deviations etc. can also

give hints regarding process variability. Employees in the
production line and in quality assurance should be encouraged
to give feedback on the process performance. Operator errors
should also be tracked in order to check if training measures
are appropriate. The text explicitly recommends regular
meetings between quality assurance and production in order
to evaluate the above data and to discuss possible trends and
drifts with the corresponding correction and follow-up
measures.
The above results can then contribute to process

improvements. However, the guidance points out that changes
may only be implemented in a structured way and with the
final approval by the quality assurance. Additional measures
regarding process design (stage 1) and process qualification
activities (stage 2) might become necessary. Another topic
that is rated very important by the guidance is that of
maintenance (including calibration). The corresponding
maintenance and calibration cycles should then be performed
based on the gathered data.
129
V Concurrent Release of Performance Qualification

Batches
In rare cases concurrent release is also possible, i.e. a release
of the product before the entire PQ protocol has been
completed. Possible cases include orphan drugs (limited
demand for the product) and radiopharmaceuticals. Yet, the
guidance recommends the manufacturer to contact the FDA
before implementing concurrent release. Batches marketed
within the framework of concurrent release should be traced
back very closely in order to be able to take immediate action
in case of complaints (root cause). Furthermore, each batch
manufactured under concurrent release should immediately be
included in a stability programme.
VI Documentation
Documentation is considered very important during each stage
of the process validation life cycle. On the basis of the process
design (stage 1), the guidance recommends to draw up
process flow charts for the full-scale process. Apart from that,
the text refers to 21 CFR 211.22 and 211.100.
VII Analytical Methodology

Analytical methods do not have to be validated in the (early)
development stages, but they have to be scientifically sound.
However, analytical methods for clinical phases 2 and 3 and 4
are subject to cGMP.
Conclusion:
It does not really surprise that the new guidance does not
mention a fixed number of validation runs proving process
validity. This became clear in the Compliance Guide on
Validation published in 2004. The new guidance relies on a 3-
stage life cycle model. The new catchword is "process
understanding". New definitions for process validation and
performance qualification show the strong connection to
"scientific sound". Strong emphasis is also placed on the use
130
of statistical methods. "Modern" methods from the world of six

sigma, like DoE, process capability indexes (Cpk) and
statistical process control are mentioned directly or indirectly.
What does surprise is the non-mention of the qualification
stages DQ, IQ, OQ, even though the activities that are usually
subsumed under these terms are expressly addressed in the
guidance as a basis for PQ. The PQ is now the key element of
the process validation life cycle and is meant to be carried out
under normal conditions. Thus, worst-case considerations with
industrial scale batch sizes are excluded. Within the framework
of continued process verification, apart from the topic of
trending, maintenance is rated highly. Both revalidation and
retrospective validation are not mentioned any longer. It will
be interesting to see how industry reacts to this draft.
Especially the discontinuation of the "magic 3" - even though
long anticipated - will probably (have to?) lead to new
rationales in order to prove a validation. The new definition of
PQ might lead to irritations. Up till now, PQ was often seen as
being primarily related to equipment. It remains to be seen in
how far statistics will find their way into process validation.
131
Part 2: Comparison Draft to Final
On 25 January 2011 the FDA published the final version of a

new Process Validation Guidance. In the following you can
read an analysis of the changes regarding to the draft from
November 2008 (in general minor changes are not mentioned
explicitly but more significant changes are printed in italic):
Table of Content
x New structure concerning changes
x New subchapter II A Process Validation and Drug Quality,
x New subchapter II B Approach to Process Validation
x New structure of chapter Recommendations in A (General
Considerations), B (Stage 1), C (Stage 2), D. (Stage 3)
x New Glossary
x References mentioned
I. INTRODUCTION
x Emphasis to use ICH Q 8, 9 und 10
x Explicitly mentioned: term commercial manufacturing (old
term was commercial production) with definition explained
in a footnote
II. BACKGROUND
B. Approach to Process Validation
x Slight different definition of process validation and stage 1

and stage 2:
Process Validation: The collection and evaluation of data, from

the process design stage through commercial production,
which establishes scientific evidence that a process is capable
of consistently delivering quality product.
132
Stage 1 – Process Design: The commercial manufacturing

process is defined during this stage based on knowledge
gained through development and scale-up activities.
Stage 2 – Process Qualification: During this stage, the process

design is evaluated to determine if the process is capable of
reproducible commercial manufacturing.
x Occurrence of multiple stages mentioned

x Statement: the approach is to control of the manufacturing
process that results in products with the desired quality
attributes.
x Statement: focusing exclusively on qualification efforts
...may not lead to adequate assurance of quality.
x New clause regarding process improvements:
Manufacturers should use ongoing programs to collect and
analyse product and process data to evaluate the state of
control of the process. These programs may identify
process or product problems or opportunities for process
improvements that can be evaluated and implemented
through some of the activities described in Stages 1 and 2.
x New clause regarding process improvements for legacy
products: Manufacturers of legacy products can take
advantage of the knowledge gained from the original
process development and qualification work as well as
manufacturing experience to continually improve their
processes. Implementation of the recommendations in this
guidance for legacy products and processes would likely
begin with the activities described in Stage 3.
133
III. STATUTORY AND REGULATORY REQUIREMENTS

FOR PROCESS VALIDATION
x New sentence: The CGMP regulations require that
manufacturing processes be designed and controlled to
assure that in-process materials and the finished
product meet predetermined quality requirements and
do so consistently and reliably.
x Concretisation regarding CFR requirements for
sampling in process control for validation.
x Additional minor changes regarding the wording
IV. RECOMMENDATIONS
A. General Considerations for Process Validation
x New bullet point regarding the terms attribute(s) and
parameter(s).The emphasis is on a "risk based
approach". Variations of terminology usage is individual
and acceptable but should be communicated to the
Agency: The terms attribute(s) (e.g., quality, product,
component) and parameter(s) (e.g., process, opera-
ting, and equipment) are not categorized with respect
to their criticality in this guidance. With a lifecycle
approach to process validation that employs risk based
decision making throughout that lifecycle, the
perception of criticality as a continuum rather than a
binary state is more useful. All attributes and
parameters should be evaluated in terms of their roles
in the process and impact on the product or in-process
material, and re-evaluated as new information
becomes available. The degree of control over those
attributes or parameters should be commensurate with
their risk to the process and process output. In other
words, a higher degree of control is appropriate for
attributes or parameters that pose a higher risk. The
Agency recognizes that terminology usage can vary
and expects that each manufacturer will communicate
134
the meaning and intent of its terminology and

categorization to the Agency.
x New bullet point regarding single source products and
complicated manufacturing processes with the
message, that batch homogeneity is the goal of
validation: Many products are single-source or involve
complicated manufacturing processes. Homogeneity
within a batch and consistency between batches are
goals of process validation activities. Validation offers
assurance that a process is reasonably protected
against sources of variability that could affect
production output, cause supply problems, and
negatively affect public health.
B. Stage 1 - Process Design

x Slight change regarding Process design: Process
design is the activity of defining the commercial
manufacturing process that will be reflected in planned
master production and control records
1. Building and Capturing Process Knowledge and

Understanding
x Clarification of viral and impurity clearance studies:

Although often performed at small-scale laboratories,
most viral inactivation and impurity clearance studies
cannot be considered early process design
experiments. Viral and impurity clearance studies
intended to evaluate and estimate product quality at
commercial scale should have a level of quality unit
oversight that will ensure that the studies follow sound
scientific methods and principles and the conclusions
are supported by the data.
135
2. Establishing a Strategy for Process Control
x Two new, very interesting sentences: Decisions

regarding the type and extent of process controls can
be aided by earlier risk assessments, then enhanced
and improved as process experience is gained. FDA
expects controls to include both examination of
material quality and equipment monitoring.
x Footnotes on ASTM Guides regarding PAT
C. Stage 2 - Process Qualification

x New term Process Performance Qualification (PPQ)
1. Design of Facility and Qualification of Utilities and

Equipment
x Term qualification mentioned: Here, the term

qualification refers to activities undertaken to
demonstrate that utilities and equipment are suitable
for their intended use and perform properly.
x Qualification plan: supplement of the item 4
responsibilities: The plan should identify the following
items:
2. The responsibilities of relevant departments and the

quality unit
3. Process Performance Qualification (new title of this

subchapter)
x Generally: term PPQ substitutes old term PQ

x Amendment of the PPQ (former PQ) approach: The
approach to PPQ should be based on sound science
and the manufacturer’s overall level of product and
process understanding and demonstrable control.
136
x Enhanced sampling and monitoring activities required:

The increased level of scrutiny, testing, and sampling
should continue through the process verification stage
as appropriate, to establish levels and frequency of
routine sampling and monitoring for the particular
product and process. Considerations for the duration of
the heightened sampling and monitoring period could
include, but are not limited to, volume of production,
process complexity, level of process understanding,
and experience with similar products and processes.
x Process Performance Qualification Protocol: last bullet
point contains new term and is supplemented due to
risk-based approach and process ability: Criteria and
process performance indicators that allow for a science
and risk-based decision about the ability of the process
to consistently produce quality products.
x New Footnote about FDA Guidance on handling OOS-
results regarding deviation
D. Stage 3 - Continued Process Verification

x New term "undesired process variability" (deletion of
the term process drift) and more focus on "CAPA-like"
actions: Adherence to the CGMP requirements,
specifically, the collection and evaluation of information
and data about the performance of the process, will
allow detection of undesired process variability.
Evaluating the performance of the process identifies
problems and determines whether action must be
taken to correct, anticipate, and prevent problems so
that the process remains in control
x Quality attributes should be appropriately controlled
x Footnote about some references, e. g. ASTM-Guides
regarding Capability Indices, ASTM-Guide regarding
Verification
137
x Change of the term (process) drift to unintended

process variability and undesirable process variation.
x Supplement of the sentence: We recommend that the
manufacturer use quantitative, statistical methods
whenever appropriate and feasible.
x Remark on 211.180(e) for the continued process
verification program
x Changes in the monitoring/sampling chapter: We
recommend continued monitoring and sampling of
process parameters and quality attributes at the level
established during the process qualification stage until
sufficient data are available to generate significant
variability estimates. These estimates can provide the
basis for establishing levels and frequency of routine
sampling and monitoring for the particular product and
process.
x Change concerning process changes: Depending on
how the proposed change might affect product quality,
additional process design and process qualification
activities could be warranted (with regard to a footnote
mentioning 21 CFR 314.70 and 601.12 as an example)
x Concretisation regarding re-qualification: equipment
and facility qualification data should be assessed
periodically...
V CONCURRENT RELEASE OF PPQ BATCHES

Generally: term PPQ substitutes old term PQ
x Focus on high degree of process assurance for

distribution: In most cases, the PPQ study needs to be
completed successfully and a high degree of assurance
in the process achieved before commercial distribution
of a product.
x Supplement regarding concurrent release: In special
situations, the PPQ protocol can be designed to release
138
a PPQ batch for distribution before complete execution

of the protocol steps and activities, i.e., concurrent
release.
x Clarification of products which might be concurrent
released: Concurrent release might be appropriate for
processes used infrequently for various reasons, such
as to manufacture drugs for which there is limited
demand (e.g., orphan drugs, minor use and minor
species veterinary drugs) or which have short half lives
(e.g., radiopharmaceuticals, including positron
emission tomography drugs).
x New subchapter explaining the necessary
of completing PPQ: Conclusions about a commercial
manufacturing process can only be made after the PPQ
protocol is fully executed and the data are fully
evaluated. If Stage 2 qualification is not successful
(i.e., does not demonstrate that the process as
designed is capable of reproducible performance at
commercial scale), then additional design studies and
qualification may be necessary. The new product and
process understanding obtained from the unsuccessful
qualification study(ies) can have negative implications
if any lot was already distributed. Full execution of
Stages 1 and 2 of process validation is intended to
preclude or minimize that outcome.
x New subchapter explaining circumstances and rationale
for concurrent release: Circumstances and rationale for
concurrent release should be fully described in the PPQ
protocol. Even when process performance assessment
based on the PPQ protocol is still outstanding, any lot
released concurrently must comply with all cGMPs,
regulatory approval requirements, and PPQ protocol lot
release criteria. Lot release under a PPQ protocol is
based upon meeting confidence levels appropriate for
each quality attribute of the drug.
139
x Supplements concerning failures in PPQ and

concerning stability programme: Concurrently released
lots must also be assessed in light of any negative PPQ
study finding or conclusions and appropriate corrective
action must be taken (§§ 211.100(a), 211.180(e), and
211.192). We recommend that each batch in a
concurrent release program be evaluated for inclusion
in the stability program.
VI DOCUMENTATION
x New supplement regarding the validation life cycle: The
degree and type of documentation required by cGMP
vary during the validation lifecycle.
VII ANALYTICAL METHODOLOGY

x New sentence: Validated analytical methods are not
necessarily required during product- and process-
development activities or when used in characterization
studies.
x Clarification of the use of new analytical methods and
regarding clinical studies: New analytical technology
and modifications to existing technology are continually
being developed and can be used to characterize the
process or the product. Use of these methods is
particularly appropriate when they reduce risk by
providing greater understanding or control of product
quality. However, analytical methods supporting
commercial batch release must follow cGMPs in parts
210 and 211. Clinical supply production should follow
the cGMPs appropriate for the particular phase of the
clinical studies.
GLOSSARY
x Chapter completely new
140
REFERENCES
x Updating of FDA references and supplement of ASTM
Guides.
Summary
The final version does include neither the term prospective,

nor the terms concurrent (only concurrent release)
or retrospective validation. Process validation is a life cycle
approach with 3 stages. The number of validation batches is
not mentioned. Based on a risk based approach
statistics should give a scientific sound rationale, when a
process is valid. Old processes should be evaluated using
development experiences, qualification work and of course
routine production experience. So (re)validation of old
processes start with stage 3. IQ and OQ are not mentioned,
but the term qualification is still mentioned. Terms like critical
quality attribute and critical process parameter are not
included. These terms are covered automatically in validation,
if you use ICH Q8, Q9 and Q10. You don’t find the expression
worst case. During Stage 1 and 2 manufacturers gain a lot of
data. With these data they can demonstrate that the process
is capable (including conditions that pose a high risk of
process failure). A lot of ASTM-Guides are referenced as state
of the art.
This new FDA Guidance is actually the most modern GMP

Guidance on Process Validation and will probably have a signal
effect for other authorities.
141
142
Annex 4:
The new EMA Guideline on Process Validation –
A detailed Analysis
It didn't really come as a surprise that EMA published its

revised Guideline on Process Validation on 27 February 2014
after the publication of the draft on the revision of Annex 15 at
the beginning of February. EMA had announced the revision in
a concept paper already a long time ago. The aim was to
include modern aspects of GMP such as ICH Q8-11, PAT, QbD,
RTRT. Another goal mentioned was a harmonization with the
FDA Guidance on Process Validation and later also with the
Annex 15 revision
The following is a detailed analysis concerning also the original

draft of the revision from March 2012.
The Note for Guidance on Process Validation which in the

beginning only comprised seven pages now has more than
doubled its content (15 pages). Even the original draft for the
revision only consisted of 11 pages. It is striking that the title
was changed into "Guideline on process validation for finished
products- information and data to be provided in regulatory
submissions". This states already clearly that the document
concerns authorisational affairs.
After the publication of the draft of Annex 15 at the beginning

of February 2014 EMA followed suit with the revision of its
Guideline on Process Validation. The final document was
published on 27 February 2014. EMA had announced the
revision in a concept paper already a long time ago. As aim of
the revision the inclusion of modern aspects of GMP was
mentioned.
143
Just like the draft, the document is divided into eight

numbered chapters, an executive summary, definitions,
references, an Annex I (Process validation scheme) and an
Annex II concerning the topic standard/non standard
processes. Annex II is new as compared to the draft. The sub-
chapter "Design space verification" in the chapter Process
validation also is new.
In comparison with the draft there are only few changes in the
content of the executive summary. The executive summary
still explains that the draft serves for bringing the guideline
into line with the ICH Guidelines Q8, Q9 and Q 10 and that it
is possible to use continuous process verifications (CPV). The
indication that the document does not introduce new
requirements on medicinal products already authorised also
remained unchanged.
1. Introduction/Background
The first chapter also contains only very few changes as
compared to the draft. New elements are the reference to ICH
Q7 in the term "quality attribute" and the reference that all the
critical elements in the manufacturing process should be
covered and included in the dossier for regulatory submission.
Apart from that, the possibility of a CPV based on the
knowledge from product and process development and the
corresponding process and / or previous manufacturing
experience is addressed just as before. CPV may be applied in
addition to the "traditional approach" as described in the
current guideline or independently in the sense of an
"enhanced approach". The draft mentions the possibility to
use in-line, on-line or at-line monitoring to evaluate process
performance. It is stated that the combination of the contents
of EMA's guideline "Note for guidance on development
pharmaceutics" (CPMP/QWP/155/96) and ICH Q8(R2) with the
current document covers all the critical elements in the
manufacturing process. A separate reference is made to
144
veterinary medicinal products to which ICH Q8(R2) does not

apply. But the principles detailed in the draft may, in principle,
also be applied to veterinary medicinal products. Hence, there
is nothing new compared to the draft.
The indication that process validation is no one-off event and

that a lifecycle approach (product and process development,
validation of the commercial manufacturing process and
maintenance of the process in a state of control during routine
commercial production) continues to be valid unchanged.
2. Scope
Chapter 2 now points out more frequently that the document
refers to the commercial dosage form of chemical medicinal
products. Just as in the draft it is indicated, furthermore, that
the general principles also apply to active substances.
However, information on the validation of non-sterile active
substances is not required in the dossier. Now, reference is
made to ICH Q11 concerning further details on active
substances. It has remained unchanged in the document that
the principles are also applicable to biological products and
that these should, however, be considered on a case-by-case
basis in view of the complex nature and inherent variability of
such products. The part that the document provides guidance
on the information to be considered for dossier submission
and as such is mainly aimed at industry and assessors and
that the information may, however, also be useful for
inspectors was deleted. The expectation is new that the
information / data requested in this guideline be present in the
dossier at the time of regulatory submission. In conclusion,
reference is now made to the product lifecycle and the data
required in this document is assigned to the second stage
(validation of the manufacturing process). Regarding the first
stage (process design) reference is made to the "Note for
Guidance on pharmaceutical Development" (ICH
145
Q8R2/EMEA/CVMP/315/98) and concerning the third stage

(on-going process verification) to GMP (Annex 15).
3. Legal basis
This chapter refers to individual sections of Directives
2001/83/EC and 2001/82/EC - without any changes as
compared to the draft.
4. General Considerations
This chapter mentions once again that validation is generally
required before the product is placed on the market. This
corresponds exactly to the draft. But now there is a reference
to Annex 15 if in exceptional circumstances concurrent
validation may be accepted. The requirement that validation
should cover all manufactured strengths and all manufacturing
sites used for production of the marketed product has
remained unchanged. The reference made to the bracketing
approach (still called matrix-approach in the draft) and to its
use in the case of different strengths, batch sizes and pack
sizes is new. But each site has to be taken into consideration.
This list is comparable to point 4.4 in the revision draft of
Annex 15. This is followed by the indication that validation
should be carried out in accordance with GMP and that data
should be held at the manufacturing location and made
available for inspection if not required in the dossier. This part
was taken over from chapter 5.1 of the draft into this chapter.
Later on there is the – new - indication that the traditional way
can be performed regardless of the approach to the
development taken. Then continuous process verification and
process improvements which could enable a CPV are
addressed.
146
Chapter 5
Process validation) is – unaltered - subdivided into four sub-
chapters (“Traditional process validation”, “Continuous process
verification”(CPV), “Hybrid approach” and - new – “Design
space verification”). For this the sub-chapter “Continued
process verification during the lifecycle” was dropped.
Compared to the draft, the statement was introduced in sub-

chapter 5.1 (“Traditional process validation”) that a traditional
process validation is normally performed when the
pharmaceutical development and/or process development is
concluded, after scale-up to production scale and prior to
marketing of the finished product. The document still contains
the statement that some process validation studies may be
conducted on pilot scale batches if the process has not yet
been scaled up to production scale. Here, the (old) 10% or
100,000 units rule is mentioned once more for solid dosage
forms. For the data to be conclusive, a justified approach is
required in the case of smaller scales. This was also required
in the draft. The reference to other solid dosage forms was
added. For these dosage forms the pilot batch size should be
justified taking into account the risk to the patient of failure of
the dosage form. If it is not useful to conduct full validation
studies, the process validation scheme outlined in Annex I of
this guideline should be completed at production scale. Now, a
bracketing approach is also possible. The authority that could
limit the scale-up of the batch after the authorisation is not
mentioned anymore. Then the draft points out where
justifications for the chosen process validation scheme should
be presented in the authorisation documentation. The
indication from the draft that process validation should focus
on the control strategy which includes critical process
parameters and the indication concerning the capability of the
process are no longer contained in this document.
147
There were no changes, however, as concerns the

requirement that in certain cases it is considered necessary to
provide production scale validation data, for example when the
product is a biological / biotech product or in the case of a
non-standard method of manufacture (now with a reference to
chapter 8 and Annex II). Non-standard sterilisation methods
or aseptic processing are not mentioned any longer. Should it
be considered necessary to provide validation data in the
dossier the number of consecutive batches should be based on
the variability of the process, the complexity of the process /
product, process knowledge gained during development,
supportive data at commercial scale during technology transfer
and the overall experience of the manufacturer. As compared
to the draft this is a clear extension of the factors to be
included for the definition of the validation batches. Data on a
minimum of 3 validation runs should be submitted unless
otherwise justified. Data on 1 or 2 validation batches may
suffice where these are supported by pilot scale batches and
by a justification as highlighted above. This corresponds to the
draft.
The chapter ends with the requirement that the studies should
address the critical steps of manufacture, by conducting
additional testing as necessary. This was already required in
the draft. The original indications concerning the validation in
the case of the implementation of a design space contained in
the draft have not been taken over for this document.
In chapter 5.2 "Continuous process verification" CPV is defined

as an alternative approach to traditional process validation in
which manufacturing process performance is continuously
monitored and evaluated (cited from ICH Q8). This is no
variation compared with the draft. It has been added that CPV
can be used in addition to, or instead of, traditional process
validation.
148
The following paragraph has remained unchanged: It is a

science and risk-based real-time approach to verify and
demonstrate that a process that operates within the
predefined specified parameters produces material which
meets all its Critical Quality Attributes (CQAs) and control
strategy requirements. Again, the focus is put explicitly on
extensive in-line or at-line controls and on monitoring process
performance and product quality in a timely manner. Relevant
process quality attributes of the starting materials should be
collected. This should include the verification of attributes,
parameters and end points, and assessment of CQA and
Critical Process Parameter (CPP) trends. Process analytical
technology applications such as NIR spectroscopy (with
application examples) and multivariate statistical process
control (MSPC) can be viewed as enablers.
Just as in the draft, sufficient knowledge and understanding of

the process are pre-requisites for the continuous process
verification. Scope and extent of CPV are influenced by a
number of factors including:
x development and manufacturing knowledge from

similar products and/or processes;
x the extent of process understanding gained from
development studies and commercial manufacturing
experience;
x the complexity of the product and/or manufacturing
process;
x the level of process automation and analytical
technologies used;
x new is the reference to legacy products: with reference
to the product lifecycle, process robustness and
manufacturing history since point of commercialization
as appropriate.
149
The indications from the draft concerning the design space

and that the process should be verified at commercial scale
prior to marketing are deleted here.
The next paragraph has remained almost unchanged

compared to the draft: A discussion on the appropriateness
and feasibility of the continuous process verification strategy
should be included in the development section of the dossier
and should be supported with data from at least laboratory or
pilot scale batches. A description of the CPV strategy including
the process parameters and material attributes that will be
monitored, as well as the analytical methods that will be
employed, should be included (reference to Annex 1), with
cross reference to the validation section of the dossier. Actual
CPV data generated at commercial scale should be held at the
site for inspection. The applicant should define the stage at
which the process is considered to be under control (the
underlined part is the new wording, originally the wording was
put as follows: "process validated") and the validation exercise
completed prior to release of the product to the market, and
the basis on which that decision will be made. This rationale
should include a justification for the number of batches used
based on the complexity and expected variability of the
process and existing manufacturing experience.
The following paragraph has also remained nearly unchanged

as compared to the draft: Continuous process verification can
be introduced at any time in the lifecycle of the product i. e.
for the initial commercial production (the wording: "to design
process validation protocols for the initial commercial
production" is not used anymore), to re-validate as part of
process changes or to support continual improvement
throughout the lifecycle.
150
Continuous process verification depends ("strongly" has been

omitted) on compliance with GMP, if necessary complemented
by pharmaceutical quality systems (PQS) as described in ICH
Q10. GMP matters and PQS should not be included in the
submission. They are assessed and handled in the course of
GMP inspections. Apart from the omission of the word
"strongly" nothing has changed as compared to the draft.
Some small changes were carried out in sub-chapter 5.3

concerning the hybrid approach. As before, either the
traditional process validation or the continuous process
verification approach may be used for different steps within
the manufacturing process. Now, it should be clear in the
dossier which approach to validation has been taken for which
steps in the manufacturing process. Surprisingly, a justification
for using the hybrid approach is no longer required in the
dossier. The validation requirements in terms of batch size and
number of batches that depend on the extent to which
continuous process verification has been used have remained
unchanged. The requirement in the case of non-standard
processes has been changed (reference to chapter 8). If
continuous process verification does not address the critical
unit operation(s) in the case of the use of non-standard
manufacturing methods, the process validation requirements
highlighted in section 5.1 (traditional approach) should be
applied unless otherwise justified.
Sub-chapter 5.4. ("Design space verification") is completely

new. To start with, the term "design space verification" is
explained. Changes in the NOR (normal operating range)
might represent higher risks not previously identified during
initial establishment of the design space. Depending on how
the design space was originally established and how the
process was validated, there will be situations where it will be
necessary to confirm the suitability of the design space and to
verify that all product quality attributes are still being met in
151
the new area of operation within the design space. This

confirmation is called "design space verification".
If the parameters investigated during development of the
design space have not been shown to be scale independent
and the process has been validated using traditional process
validation, design space verification would be required on the
basis of a verification protocol and the results should be
provided in the dossier. If CPV has been utilised, the validity of
the design space could take place throughout the product
lifecycle. In this case a "design space verification strategy"
should be part of the "continuous process verification
strategy".
Depending on the change itself and the extent of movement

within the design space (i.e. distance from validated target
parameters/NOR or new area of design space with higher or
unknown risk) protocols for verification may include controls of
quality attributes (QA’s) and process parameters (PP’s) not
included in the routine control system. Examples mentioned
are monitoring or testing of QA’s and PP’s that are scale
dependant and when applicable, equipment dependant. It is
stated explicitly that it is not necessary to verify entire areas of
the Design Space or the limits of the process ("edge of
failure"). A stepwise approach taking into consideration the
need to adjust the NOR within the approved design space
during product lifecycle is acceptable.
The explanations concerning continuous process verifications

during the lifecycle that were previously contained in this sub-
chapter are now deleted completely.
152
Chapter 6
Contrary to the previous sub-chapter there are relatively few
changes in the chapter "Scale-up" compared to the draft. Due
to the fact that the wording of the draft is worded almost
identically with the actual Note for Guidance there are
relatively few changes the "authoriser" has to deal with.
Compared to the actual Note for Guidance those parameters
listed in the process validation scheme (Annex I) will not
generally need to be re-validated once further scale-up is
proposed post-authorisation if the process has been tested to
be scale independent or if - and that is new as compared to
the draft - continuous process verification is employed.
Chapter 7
Chapter 7 concerning changes after authorisation ("Post
approval change control") doesn't contain many changes
compared to the draft, either (and relatively few compared to
the actual Note for Guidance). The document contains the
new requirement that the consistency of the approved control
strategy must also be observed. The draft still contains the
requirement to ensure consistency with the approved
specifications, this part was deleted. Furthermore, this chapter
explicitly lays down that change control is part of GMP and is
not normally specified in the dossier. Reference is still made to
the European Commission guidance on Type I and Type II
variations on the basis of the regulations 1234/2008/EC and
now also 712/2012/EC.
Chapter 8
Chapter 8 (“Standard vs. non-standard methods for
manufacture”) is much shorter than in the draft, but it is still
only relevant for processes which have not been validated
using continuous process verification. Now many contents of
the draft were taken over in Annex II of the document. A new
requirement is the inclusion of full production-scale data in the
dossier for non-standard products or processes which were
153
validated using traditional process validation. Concerning this,

a reference is made to chapter 5.1 ("Traditional process
validation"). Now, the exceptions an applicant may use to
justify that the product process can be considered standard for
a particular manufacturer / or a particular site are described in
more concrete terms. Thereby, the risk to the patient of failure
of the product or process should be taken into consideration.
Such justifications should be assessed on a case by case basis,
but the information provided for each manufacturing site
should include:
x experience with the same or essentially similar product
or process (number of products authorised / marketed
in the EU/EEA and number of batches (including
information on scale) manufactured); The names/
marketing authorisation numbers in the relevant
EU/EEA member state should be provided.
x amount of knowledge gained during the development
of the product (number and scale of batches
manufactured at each manufacturing site involved);
x history of GMP compliance of the manufacturing site
for that type of process. The applicants should clearly
state in the application whether they consider the
manufacturing process to be standard or non-standard.
This decision should be justified.
For further information reference is made to Annex II of the

document.
The following definitions are new, compared to the draft:

x At-line
x Bracketing approach (relatively detailed)
x Enhanced approach
x In-line
x Ongoing process verification
x On-line
x Traditional approach
154
The following definitions have been deleted:

x Continued Process Verification
x High impact models
There are more references which now also contain EU and

EMA documents.
Annex I
In Annex I (Process validation scheme) the contents
concerning the traditional process validation have remained
unchanged compared to the draft. But the requirements
contained in the sub-chapter CPV have changed. Where CPV is
used, a process validation scheme should be submitted by the
applicant. This scheme should outline the monitoring on
production scale. The data should be available for verification
post authorisation by the supervisory authority. The process
validation scheme should include the following information
concerning the monitoring:
x details of on-line / in-line / at-line monitoring including

parameters tested, number of samples, size of samples
and frequency of monitoring;
x details of analytical methods (with references to the
dossier);
x acceptance criteria;
x information/ data including, as appropriate, information
on statistical models or tools used to determine
whether the continuous verification data supports the
ability of the process and controls to produce
reproducible product at a commercial scale;
x if a design space has been developed, how the
proposed monitoring will contribute to design space
verification.
155
Annex II
Annex II (“Standard / non-standard processes”) contains
many elements that were already contained in Chapter 8 of
the draft. Insofar the final document now takes the original
course and has integrated large parts of the actual Annex II
(CPMP/QWP/2054/03, EMEA/CVMP/395/03) in the actual Note
for Guidance on Process Validation. It is new that all biological
products are considered to be non-standard. The non-
standard processes (non-standard methods of sterilisation,
aseptic processing, lyophilisation, micro-encapsulation, certain
mixing processes, coating processes) listed as examples in the
draft are not contained in the text any more. But in the
detailed requirements concerning specialised pharmaceutical
dosage forms, new technologies and non-standard methods of
sterilisation some of these points re-appear.
Conclusion
There haven't been great changes to the draft released in
2012. But the chapter "Design space verification" is completely
new; many other parts have been updated to the state of the
art. The chapter on continued process verification has been
deleted. This part is now contained in the revised Annex 15 as
"ongoing process verification" and there is a corresponding
reference in the document. Especially welcome is the fact that
"high impact models" are not mentioned as possibility for a
part of the CPV any more. In the draft further information on
the meaning of these models was missing. Compared to the
draft, information concerning standard / non-standard
processes is again contained as independent Annex II, just as
was the case in the Note for Guidance applicable so far.
And what are the changes compared to the “old” Note for
Guidance on Process Validation?
Compared to the “old” Note for Guidance the revision remains

rather difficult to read and general in its final version. It is a
156
marketing authorisation document, a fact which is clearly

addressed already in the title. It is valid only for medicinal
products but not for legacy products. Its applicability for active
ingredients and biological products is mentioned analogously
as possible. The introduction of a validation life cycle and the
integration of a continuous process verification (CPV) are
completely new although this approach is already known from
ICH Q8. The "traditional approach" remains accepted and the
magical number three is mentioned for the validation runs.
This aligns with the information from the Annex 15 draft. In
this point a discrepancy remains to the FDA Process Validation
Guidance which doesn't contain a number for the validation
runs any more. The aim to integrate modern elements from
ICH Q8, Q10 (and Q11) into the document is clearly
noticeable. But there are strikingly few concrete references to
ICH Q9. The more detailed definition of the bracketing
approach in the glossary is laudable. Bracketing is known in
the context of process validation also from the USA. But a
closer overlap with the FDA Guidance would have been
desirable, nevertheless. The FDA Guidance addresses also
APIs and biological substances and the process validation
lifecycle runs like a red thread through the complete FDA
document. And the FDA Guidance contains GMP aspects. The
FDA Guidance refers explicitly also to legacy products which
are to be integrated in the lifecycle in step 3. There is another
big difference. In contrast to the FDA Guidance the revised
document does not contain a strong emphasis on statistical
methods.
And what is interesting for the GMP world? On one hand

relatively little on the other one quite a lot. How can this be
explained? Direct references to Annex 15 can be found
concerning " process verification" and "concurrent validation".
This is relatively little if the complete document is taken into
consideration. But it is required to carry out validation in
general, continuous process verification and change control
157
according to GMP. Since these are the essential parts of the

document, (almost) the complete document has to be seen
from a GMP perspective.
It is good that a comparison was made to the (at that time)

draft of Annex 15 before the document was finalised. This
might be the reason for the relatively long period between the
publication of the draft in March 2012 and the publication of
the finalised document in February 2014.
The EMA-Guideline on Process Validation is effective since
August 2014.
158
Annex 5:
Detailed analysis of the new Annex 15
The draft of the revision of Annex 15 was published in

February 2014. Some of the changes were significant
compared to the old version of Annex 15. The draft was
published as final document and came into operation on
1st October 2015.
What are the changes?
In the following you will read a comparison between the "old"

(valid until 30 September 2015) Annex 15 and the revised
version. Significant changes between the draft and the final
version will also be mentioned.
Compared to the draft it was a surprise that the revised Annex

15 may also be used as a supplementary optional guidance for
active substances, but without introduction of additional
requirements of Part II EU GMP Guide.
In general, there is a greater distinction between the terms

qualification and validation. Utilities now are part of the
qualification activities.
A general chapter as well as subsections to the chapters

"Qualification" and "Process Validation" were added. The
topics verification of transportation, validation of packaging,
qualification of utilities and validation of test methods are also
new. The chapter "Revalidation" was replaced by the chapter
"Re-qualification".
159
Principle
A life cycle of the product and process that should accompany
qualification and validation is already mentioned in the part
"Principle". As a sort of extension it is now pointed out that
changes should also assess the influence on the control
strategy. Furthermore, it is indicated that computerised
systems should be validated according to the requirements of
Annex 11. The relevant concepts presented in ICH Q8-Q11
should also be taken into account.
A completely new, general part ("General") was added. It

states that decisions on the scope and extent of validation and
qualification should be based on a justified and documented
risk assessment as part of a quality risk management
approach. The principles in ICH Q8-11 formerly mentioned are
dropped as compared to the draft. Now they are only
mentioned in the part "Principle". The last part of this chapter
states that data obtained from sources outside of the
manufacturers own programmes may be used provided that
this approach has been justified and that there is adequate
assurance that controls were in place throughout the
acquisition of such data. New, as compared to the draft, is the
explicit statement that retrospective validation is no longer
considered an acceptable approach.
Organising and Planning for Qualification and

Validation
The first chapter originally named "Planning for Validation"
now is called „Organising and Planning for Qualification and
Validation“. It is stressed that the life cycle of facilities,
equipment, utilities, process and product should be taken into
consideration and that the personnel should be trained
suitably and should follow approved procedures. Validation
personnel should report as defined in the internal quality
system although this may not necessarily be to a quality
management or a quality assurance function. However, there
160
should be appropriate quality oversight over the whole

validation life cycle.
Next, the validation master plan (VMP) is addressed. The

requirements haven't changed much. (Unfortunately), the
explicit requirement that the VMP should be brief, concise and
clear was deleted. The range of the VMP was extended to
qualification. To the organisational structure already required
in the "old" Annex 15 roles and responsibilities for qualification
and validation activities were added. This also is new as
compared to the draft. The requirement that the ("current"
has been deleted as compared to the draft) validation status
of the facilities, equipment, systems, processes on site should
be summarised was added. Furthermore, the revised Annex 15
requires the description of a deviation management in the
validation master plan, the developing of acceptance criteria
(in the draft "handling") and the description of a qualification
and validation strategy, including re-qualification. As compared
to the draft the "ongoing validation strategy" was deleted as
well as the requirement to describe a revalidation (substituted
by re-qualification). The assessment of the resources required
which was still mentioned in the draft is not required any
more. This is also true for the confirmation in the validation
master plan that the materials used are of the required quality
and that suppliers are qualified to the appropriate level.
Newly added were the last two requirements in this chapter:

x In the context of the quality risk management
approach the risk assessments should be repeated in
the light of increased knowledge and understanding
from any changes. It is explicitly stated that the ways
in which risk assessments are used should be clearly
documented.
x Checks to ensure the integrity of all data obtained.
161
Documentation
The second chapter "Documentation" (there is again a
reference to the VMP) begins with demanding the use of
"good documentation practices" in order to support knowledge
management throughout the product lifecycle. All validation
documents should be approved and authorised as defined in
the quality system. The inter-relationship between documents
in complex validation projects should be clearly defined.
Validation protocols should be prepared which define the
critical systems, attributes and parameters and the associated
acceptance criteria. The revised Annex 15 explicitly states the
possibility to combine together qualification documents (such
as IQ and OQ). Where validation protocols are supplied by a
third party the user should confirm their suitability and
compliance with internal procedures before approval. The
possibility that vendor protocols may be supplemented by
additional documentation/test protocols is new as compared to
the draft. Any significant changes to the approved protocol
(such as acceptance criteria, operating parameters) should be
documented as a deviation and be scientifically justified.
Results which fail to meet the acceptance criteria should also
be recorded as a deviation and be fully investigated according
to internal procedures. Any implications for the validation
should be discussed in the report. The content of the
requirements concerning the report did not change. Cross-
references to the protocol are not required any more but the
results obtained should be summarised against the acceptance
criteria. Review and conclusions should be part of the report.
The final document showed flexibility concerning the formal

release for the next stage in the qualification and validation
process that can either be part of the validation report
approval or as a separate summary document. Completely
new is the conditional approval to proceed to the next
qualification stage that can be given in the case of deviations
162
for example where there is a documented assessment that

there is no significant impact on the next activity.
Qualification
The main stages of qualification and some suggested criteria
are indicated in the third chapter "Qualification Stages for
Equipment, Facilities, Utilities and Systems" as a "could"
option. Is this a reference to alternative procedures? The
contents of the qualification steps themselves (IQ, OQ, PQ)
remained "should" requirements. In the draft they still have
been "could" requirements. The definitions for the single
qualification steps DQ, IQ and OQ remained unchanged, only
the definition for PQ was adjusted. Already at the beginning of
the chapter it is indicated that qualification should consider all
stages from the development of the user requirements
specification (URS) through to the end of use. The
specification for new equipment, facilities, utilities or systems
should be defined in a user requirements specification and/or
a functional specification. Objective is to build in quality at this
stage and to mitigate any GMP risks. The URS should be a
point of reference throughout the validation life cycle.
Now, DQ is the second step in the qualification. There have

been few changes as concerns the content apart from the fact
that the requirements of the user requirements specification
should be verified during the design qualification.
New parts in the final document are the factory acceptance

test (FAT) and the site acceptance test (SAT). A FAT could
eventually be carried out for equipment, especially if
incorporating novel or complex technology. This is a clear
defusing as compared to the draft, that still required a FAT as
"should". Where appropriate and justified, documentation
review and some tests could be performed at the FAT or other
stages without the need to repeat on site at IQ/OQ if it can be
shown that the functionality is not affected by the transport
163
and installation. According to the final document a FAT may

be supplemented by the execution of a SAT. Prior to
installation, equipment should be confirmed to comply with
the URS/ functional specification at the vendor site, if
applicable.
The requirements concerning the IQ have remained quasi

unchanged. Now the verification of the correct installation of
components and instrumentation against the engineering
drawings and specifications is explicitly required. In the
revised Annex 15 verification of the correct installation against
pre-defined criteria is now required in an own sub-point.
Depending on the complexity of the equipment, The OQ may

now be performed as a combined Installation/Operation
Qualification (IOQ). The formal release after the OQ is not
required any more. The OQ test should ensure the system is
operating as designed. It is not mentioned any more that the
calibration should be reviewed at this stage.
For PQ the possibility is stated explicitly that it may (in some

cases) be appropriate to perform it in conjunction with OQ or
Process Validation. New is the use of worst case batch sizes.
Now, the frequency of sampling should be justified. It is new
that the extent of PQ tests depends on the results from
development.
Re-qualification
Chapter 4 "Re-qualification" is new. The state of qualification
should be evaluated at regular intervals and the period should
be justified. Furthermore, the possibility of small changes over
time should be assessed. The requirement in the draft that
where manual processes are used (such as for cleaning), the
continued effectiveness of the process should be confirmed at
a justified frequency is not contained in this chapter in the
final document. Compared to the still valid Annex 15 the
164
possibility of a review within the re-qualification is not stated

any more.
The sub-chapter on the qualification of established (in-use)

facilities, systems and equipment was dropped completely.
Chapter 5: Process Validation

It is implicit in this annex that a robust product development
process is in place to enable successful process validation. It is
explicitly referred to the current EMA Guideline on Process
Validation which will also have to be taken into consideration.
In so far some parallels can be found between both
documents. It is also stated that the GMP requirements go
beyond the validation requirements for the documents
concerning regulatory submissions. Reference is also made to
the lifecycle approach. The traditional approach or a
continuous verification approach or a hybrid approach is
mentioned as option. Interestingly enough it is indicated that
irrespective of the approach used, processes must be shown
to be robust and ensure consistent product quality before any
product is released to the market. Validation for new products
should cover all intended marketed strengths and sites of
manufacture. For products which are transferred from one site
to another or within the same site, and where there is existing
product knowledge a bracketing approach could be used
concerning the number of batches, strengths, batch sizes and
pack sizes/ container types, if justified. In the case of new
products, condition for bracketing is extensive process
knowledge from the development stage in conjunction with an
appropriate ongoing verification programme.
The identification of critical process parameters and critical

quality attributes should be based on a risk assessment. In the
case of a site transfer the conditions of the Marketing
Authorisation should still be met. Batches manufactured for
process validation deviating from the size of commercial scale
165
batches should be justified. The draft still mentions a

continuous manufacturing process as example for a deviating
batch size. This is not contained in the final document any
more. It is explicitly expected that production personnel are
involved in the validation activities (product understanding)
but personnel from development or site transfer may also be
involved. The suppliers of critical starting and packaging
materials should be qualified prior to the manufacture of
validation batches. Otherwise there should be a written
justification based on the application of quality risk
management principles. Process knowledge and mathematical
models should be available in the case of design spaces.
Where validation batches are released to the market, this
should be pre-defined. The products should fully comply with
GMP and with the marketing authorisation or clinical trial
authorisation (reference to Annex 13 EU GMP Guide).
There are very strict rules for concurrent validation which may
be carried out only after a justified and strict risk-benefit-
assessment. This approach must be documented in the
validation master plan and be approved by authorised
personnel. There should be sufficient data to support the
conclusion that the process is uniform and can meet the
defined acceptance criteria. This should be documented and
the data should be available to the Qualified Person prior to
release of the batch.
When using the traditional approach (manufacture of

validation batches under routine conditions), the number of
batches and the number of samples taken should be based on
quality risk management principles to allow the normal range
of variation and trends to be established and to provide
sufficient data for evaluation. Each manufacturer must
determine and justify the number of batches necessary to
demonstrate that the process is capable of consistently
delivering quality products.
166
Without prejudice to the above it is generally considered

acceptable that a minimum of three consecutive batches
manufactured could constitute a validation of the process,
states the text in the revised version. But it is mentioned
explicitly that an alternative number of batches may be
justified taking into account whether standard methods of
manufacture are used and whether similar products or
processes are already used. If this way is chosen it should be
justified. An initial validation exercise with three batches may
need to be supplemented with further data obtained from
subsequent batches as part of an on-going process verification
exercise. The validation protocol should define the critical
process parameters (CPP), critical quality attributes (CQA) and
the associated acceptance criteria which should be based on
development data or documented process knowledge. The
following requirements were added to the validation protocol:
• Summary of the CQAs

• CPPs (with limits)
The protocol should now also contain a summary of non-

critical attributes and parameters which will be investigated
during the validation activity. This is a difference to the FDA
Process Validation Guidance. There should be given reasons
for the inclusion of these attributes and parameters in the
validation. The method validation of the relevant analyses
should be stated as well as the criteria for the process for
release, if applicable. There should exist a rationale for the
sampling plan. And furthermore, reasons should be given why
each in-process control is selected during the validation. But a
time table is not required any more.
There are special sub-chapters for "continuous process

verification" and "ongoing process verification". The condition
for continuous process verification is a quality by design
approach, where it has been scientifically established during
167
development that the established control strategy provides a

high degree of assurance of product quality. The continuous
process verification should naturally be specified in writing and
contain a science based control strategy. Process Analytical
Technology and multivariate statistical process control may be
used as tools. Each manufacturer must determine and justify
the number of batches necessary to demonstrate that the
process is capable of consistently delivering quality products.
The same wording is used as for the traditional approach. A
hybrid approach using the traditional approach and continuous
process verification can also be used. Where there is a
substantial amount of product and process knowledge and
understanding which has been gained from manufacturing
experience and historical batch data, continuous process
verification may also be used for any validation activities after
changes or during ongoing process verification. This applies
even if the product was initially validated using a traditional
approach.
In the context of the ongoing process verification product

quality should be monitored to be able to demonstrate that a
state of control is maintained throughout the product lifecycle
with the relevant process trends evaluated. This concerns all
possible validation approaches (traditional, continuous or
hybrid). The extent and frequency of ongoing process
verification should be reviewed periodically and modified if
appropriate, according to the level of process understanding.
Ongoing process verification should be conducted under an
approved protocol or equivalent documents - this is new as
compared to the draft - and a corresponding report should be
prepared. Statistical tools should be used, where appropriate,
to support any conclusions with regard to process variability
and capability. The ongoing process verification should also be
reflected in the PQR. Incremental changes over time should
also be considered and the need for any additional actions,
168
e.g. enhanced sampling, should be assessed. The topic

(routine) revalidation has been dropped completely.
Chapter on the verification of transportation

The transport of medicinal products, investigational medicinal
products, bulk products and samples should take place in
accordance with the conditions defined in the Marketing
Authorisation. It is mentioned explicitly, that the verification of
transportation (not validation as in the draft) may be
challenging due to the variable factors involved.
Transportation routes should be clearly defined, however.
Seasonal variations should also be considered for transports.
The original requirement in the draft that seasonal variations
should only be considered for transport across continents has
been dropped. A risk assessment should be performed to
consider the impact of variables in the transportation process
other than those conditions which are continuously controlled
or monitored, e.g. delays during transportation, failure of
monitoring devices, topping up liquid nitrogen, product
susceptibility and any other relevant factors. Continuous
monitoring of any critical environmental conditions to which
the product may be subjected should be performed.
Chapter on the validation of packaging

Contrary to the draft it is not explicitly stated any more that
primary packaging processes should undergo validation.
Variation in equipment processing parameters of packaging
equipment may have a significant impact on the integrity and
correct functioning of the pack, therefore primary and
secondary packaging equipment for finished and bulk products
should be qualified. Qualification of the equipment used for
primary packing should be carried out at the minimum and
maximum operating ranges defined for the critical process
parameters such as temperature, machine speed and sealing
pressure or for any other factors.
169
Chapter on the qualification of utilities

The quality of steam, water, air, other gases etc. should be
confirmed following installation using the qualification steps
described in the chapter qualification. The requirement in the
draft that coolants should also be qualified has been dropped.
The period and extent of qualification should also reflect any
seasonal variations, if applicable, and the intended use of the
utility. A risk assessment should be carried out where there
may be direct contact with the product (e.g. HVAC systems) or
in order to mitigate the risk of failure due to indirect contact
(such as through heat exchangers).
Chapter on the validation of test methods

All analytical test methods used in qualification, validation or
cleaning exercises should be validated with an appropriate
detection and quantification limit, where necessary, as defined
in Chapter 6 of the EU GMP Guidelines Part I. Where microbial
testing of product is carried out, the method should be
validated to confirm that the product does not influence the
recovery of microorganisms. This is also valid for microbial
testing of surfaces in clean rooms. They should be validated to
confirm that sanitising agents do not influence the recovery of
microorganisms.
Chapter on cleaning validation

Chapter 9 "Cleaning Validation" now has more than twice the
number of sub-chapters. There were also changes as
compared to the draft. Now grouping together of different
equipment is mentioned as possible, if the relevant grouping is
justified. It is new (also in comparison to the draft) that
simulating agents may be used with appropriate scientific
justification. The acceptance criteria "visibly clean" as single
acceptance criterion is described as not acceptable. Now it is
recognised that a cleaning validation programme may take
some time to complete and validation with verification after
each batch may be required for some products, e.g.
170
investigational medicinal products. There should be sufficient

data from the verification to support a conclusion that the
equipment is clean. Validation should consider the level of
automation in the cleaning process. Automatic processes
should be validated as concerns the specified normal operating
conditions. Detailed requirements from the draft for manual
cleaning, such as the identification of variable factors or worst
case approaches have been generalised to all cleaning
processes. It is new as compared to the draft that where
manual cleaning is performed, the effectiveness of the manual
process should be confirmed at a justified frequency. Limits for
the carry over of product residues should be based on a
toxicological evaluation. The permitted daily exposure value
should be documented in a risk assessment which includes the
relevant references. The removal of any cleaning agents used
should also be confirmed. And acceptance criteria should
consider the cumulative effect of multiple equipment in the
process equipment train. New (also in comparison the draft) is
the exemplary reference to (TOC) and conductivity if it is not
feasible to test for specific product residues. Dirty and clean-
hold times should be defined as part of the cleaning validation.
Where campaign manufacture is carried out, the ease of
cleaning between batches and the maximum length of a
campaign (in both time and number of batches) should be the
basis for cleaning validation exercises. The potential for
microbial and, or if relevant, endotoxin contamination, should
be assessed during validation. The use of worst case products
should be based on a scientific rationale and be assessed
again if new products are introduced. Criteria for determining
the worst case may include solubility, toxicity, (and new
compared to the draft) cleanability and potency of the
product. Cleaning validation protocols should detail the
locations to be sampled and the rationale for the selection.
Furthermore, acceptance criteria should be defined. As
techniques for sampling, swabbing and/or rinsing or other
means depending on the sampling location are stated. The
171
swab material should not influence the result. The following

requirement in the draft was dropped: If rinse methods are
used, the sampling should be performed during the final rinse
in the cleaning procedure. Recovery rates should be defined.
Interestingly, the number of validation runs should be based
on a risk assessment. The wording in the draft that for
investigational medicinal products or products which are only
manufactured infrequently, cleaning verification may be used
instead of cleaning validation has already been mentioned
above in this section. Where a cleaning process is ineffective
or is not appropriate for some equipment dedicated equipment
should be used for each product (in the draft the reference to
chapters 3 and 5 GMP Guideline Part I is missing).
Change Control
The chapter Change Control grew from originally two to seven
sub-chapters. Change processes are an important part of
knowledge management and should be handled within the
pharmaceutical quality system during the life cycle. Changes in
the product range, the batch size or the design space now are
listed as examples for activities requiring change control.
Reference is made to a possible need for any regulatory
actions concerning the Marketing Authorisation in the case of
a changed design space. Quality risk management should be
used to evaluate the potential impact of changes on product
quality, pharmaceutical quality systems, documentation,
validation, regulatory status, calibration, maintenance and on
other systems. This should also help to plan for any necessary
process validation, verification or re-qualification efforts.
Changes should be authorised and approved by the
responsible persons or relevant functional personnel in
accordance with the pharmaceutical quality system. Following
implementation an evaluation of the effectiveness of change
should be carried out. This has already been required after the
revision of chapter 1 of the EU GMP Guidelines Part I (valid as
of 13 January 2013). More importance than in the draft is
172
attached to supporting data (e.g. copies of documents) to

confirm that the impact of the change was demonstrated prior
to final approval.
Glossary
The following are new definitions in the glossary: continuous
process verification (with reference to ICH Q8), control
strategy (with reference to ICH Q10), critical process
parameter (with reference to ICH Q8), critical quality attribute
(with reference to ICH Q8), design space (with reference to
ICH Q8), knowledge management (with reference to ICH
Q10), life cycle, ongoing process verification (it is stated
explicitly that this term also is known as continued process
verification), product realisation (with reference to ICH Q10),
quality by design (with reference to ICH Q8), quality risk
management (with reference to ICH Q9), state of control,
traditional approach, user requirements specification (not
included in the draft). In the definition of PQ the connection
between facilities, systems and equipment is not explicitly
mentioned any more. It is striking that the definition for
process validation has remained unchanged in spite of the
introduction of the validation lifecycle and that the lifecycle
insofar is not part of the definition.
Conclusion
The revision is very comprehensive. It is questionable how the
optional application oft he revised Annex 15 can be carried out
regarding active substances without introducing additional
requirements. The FDA Process Validation Guidance includes
active substances. The influences of the guidelines ICH Q8, 9
and 10 can be seen clearly, even in the glossary. The
alignment with the EMA Guideline on Process Validation,
revised in 2014, is also striking. Now the topic design space
(ICH Q8) is included in the part on process validation. Now a
lot of risk assessments (ICH Q9) are mandatory. And the
lifecycle approach and the topic process capabilities (ICH Q10)
173
are included, too. Deviation management has assumed new

importance. Third party services are allowed explicitly if the
supplier has been qualified accordingly. This can also be seen
as an adaptation to reality. It is positive that the preliminary
release for the next step (such as qualification) was mentioned
for example in the case of deviations if there is a documented
assessment that there is no significant impact on the next
activity. It was not implemented a clear separation between
qualification (with regard to facilities and equipment) and
validation (with regard to processes).
New concepts appear without being defined in the glossary,

such as FAT and SAT or a functional specification.
Retrospective validation and the term (routine) revalidation
were dropped completely. But the effectiveness of the manual
cleaning processes should be confirmed at a justified
frequency. Is this not also a sort of routine revalidation?
The inclusion of user requirements as separate step and the

mention of FAT and SAT - even if merely as could requirement
- have made qualification more extensive. FAT and SAT are
typical elements of Good Engineering Practice (GEP). Here, the
link between GMP and GEP is missing in the document. Can
other GEP elements now also be applied (without problems) in
the GMP environment? Will even more GEP elements become
mandatory in the future? The greater flexibility as concerns
the qualification steps IQ/OQ which now can be carried out
together, is a positive element. But this has already been done
frequently by industry. There are no in-depth references to
alternatives for the qualification, such as ASTM E2500
although the main qualification steps are merely a could
option. The most PQ tests are done with connected facilities,
systems or equipment. Why the change of the definition
(connection not mentioned anymore)? Unfortunately,
information on the qualification of legacy facilities, systems
and equipment were dropped completely. It happens again
174
and again that non-GMP environments are becoming GMP

environments. May these environments still qualify their old
equipment and if so, how? The reference to transport
verification, packaging validation, validation of utilities and
validation of analytical methods comes as a surprise. Other
regulations are more concrete (such as ICH Q2 (R1) as
concerns the validation of analytical methods). The transport
verification would probably have better been placed in the
context of regulations on Good Distribution Practice. The „old“
Annex 15 has been a general guidance to the topic validation/
qualification, why has there been this specification? Should not
the logical consequence be that validation of the sterilisation
process or media fills for example must also be included?
Now, there are three different approaches for process

validation, a (modern) continuous verification approach, a
traditional approach that is still based on the classical 3
validation runs and a mixture between both, the hybrid
approach. But process robustness has to be established in any
case. Despite the accordingly listed restrictions and the
necessary justifications as concerns the decision for the
number 3, the statement that generally at least three
consecutive runs are regarded as acceptable is a clear
difference to the FDA Guideline on Process Validation. This
Guideline doesn't define a number any more. It is positive,
that a bracketing approach may be used in justified cases. The
possibility of a hybrid approach remains somewhat unclear.
This term is not defined in the glossary. But it is mentioned
that a change of the validation strategy (from the traditional
approach to a verification) is possible with an increasing level
of knowledge and understanding of the process.
Unfortunately, the possibility of a review in the course of re-
qualifications is not explicitly mentioned any more. But could
the required evaluation concerning re-qualification not still be
carried out as review, where applicable?
175
As concerns the cleaning validation the criterion "visibly clean"

as single acceptance criterion is indicated as not acceptable
any more. Now, grouping together of different equipment in
the course of the cleaning validation is expressly stated as
possibility, if justified. In this case the final version is an
adaptation to general practice. Nothing but a toxicological
justification (PDE) is targeted as acceptance criterion. This
innovation has been expected but it might cause insecurities.
The statement that the decision on the number on cleaning
validation runs has to be taken on the basis of a risk
assessment is very interesting. It was refrained expressly from
mentioning the magical number three - in contrast to the
process validation. It is good that now a cleaning verification is
also possible for products which are only manufactured
infrequently. The requirement for dirty and clean hold times
constitutes an adaptation to the state-of-the-art.
The chapter Change Control now refers very strongly to

regulatory aspects. The requirement for an efficiency control
after implementation of a change is new. It has already been
required in chapter 1 of the EU GMP Guidelines Part I. Due to
the introduction of the feedback loop the term change
management would actually have been better for this chapter.
All in all there is an abundance of new requirements, which in

part, however, only reproduce the state- of-the-art. Many of
the new requirements such as user requirement specifications
for the qualification or hold times in the cleaning validation
have been used like this by industry for years. The PDE
concept is a complete novelty in the cleaning validation. Due
to the (necessary) inclusion of ICH Q8-11 and the lifecycle
approach the annex has become more holistic, but
unfortunately also more ambiguous. Although it is mentioned
explicitly, the lifecycle model is not as stringently incorporated
in the complete document as in the FDA Guideline on Process
Validation. The same is true for the very intensive emphasis
176
on statistics in the FDA Guideline which cannot be found to the

same extent in the new Annex 15. Insofar a more strict
coordination with the FDA Guideline on Process Validation
(also for example as concerns the topic PPQ) would have been
desirable.
177

Validation Good Practice, Long But Perfect. Recommended !

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Validation Good Practice, Long But Perfect. Recommended !

Uploaded by

Copyright:

Available Formats

ECA Validation Good Practice Guide

2.0 A new view on Validation: The move to process

3.0 Examples: How to do

3.1 Risk-based Approach to Process Validation

4.0 Special Case: Legacy Processes

5.0 Statistical Background / Manufacturing Excellence

5.1 Statistical Process Control (SPC)

© European Compliance Academy

Especially in the field of process validation the change is evi-

As there are different validation approaches possible in the EU

Members of the ECA Expert Working Group:

© European Compliance Academy

© European Compliance Academy

2.0 A new view on Validation:

2.1 Validation in practice from the past to today

The issue of validation with a focus on sterilisation processes

Whereas several pharmaceutical companies in the early nine-

Many suppliers of equipment, engineering and consulting to

The original idea that quality management should contribute

But basically process validation should demonstrate that the

In short, process validation was often mainly a documentation

The common approach to process validation has been the

A few voices challenging the common industry approach was

With the introduction of the concept of Good Engineering

The pharmaceutical industry also demanded a change of

The FDA now defines validation as a Life Cycle Process in 3

From a qualification and validation perspective this has be-

This is part of the new, harmonised approach and also in Eu-

In parallel to this, the process validation in ISO 900142 and

In summary, Fig. 1 shows an overview of the validation history

© European Compliance Academy

Fig. 1 Validation History, Dr Michael Hiob, Ministry of Employment,

2.2 Qualification/Validation is the focus of authorities

What is the regulatory authority’s point of view concerning

In Europe it is more difficult to obtain data due to the more

With an incidence of 10.7%, process validation ranks 3rd of

In the inspection results48 of the Austrian regulatory authority

2.3 Qualification: Current Industry Practice

In what way could a qualification be carried out today which is

Fig. 2 Including GEP documents in qualification, M. Klein, CSL-

In this context it is important that FAT and especially SAT are

Another ke ey to successs is the inte

© European Compliance Academy

© European Compliance Academy

Based on tthe product requiremen nts the risk analysis is often

© European Compliance Academy

Fig. 5: Risk analysis in the

In the pastt there was extensive uncertainty

During the qualification project, change control can be han-

Rationales for qualification tests are increasingly questioned in

© European Compliance Academy

The topic rre-qualificattion is now a own chap pter in the Annex

The new FFDA Guidancce “Process Validation”” consequen ntly no

And the future?

© European Compliance Academy

Thus IQ annd OQ are of

The ISPE WWhite Paperr summarise es the „Prin

© European Compliance Academy

© European Compliance Academy

All this would require great organisational changes in the

Ultimately, the ASTM E2500 Standard Guide propagates an

© European Compliance Academy

Moreover, the concept is also considered as usable for com-