The Basis for Acceptance

Criteria in Process Validation

Dawn Tavalsky
 Disclaimers

• The material and perspectives offered in this presentation are

solely those of the presenter, and do not reflect the views or
opinions of Sanofi.

• The presenter does not receive any compensation, grants, equity

interest, or funding from any group except her employer.

• This presentation includes intellectual property that may belong to

the presenter, to Sanofi and/or to third parties which is protected
under the copyright, trademark and other laws of the United States
of America and other countries. All rights are reserved.
 Some Common Types of Validation of
Processes

Mixing Validation
Hold Time Validation for Raw Materials
In-Process Hold Time Validation
Validation of a Process Step (Typically called Process
Validation)
Consistency Validation
Comparability Validation
Cleaning Validation
There are Lots of
Sterilization Validation Types of Validation of
Extractable/Leachable Validation Processes
Container Closure Validation
Shipping Validation

3
 The Focus of This Presentation

For this presentation – we will focus on Validation of a

Process Step (Typically called Process Validation) –
although most all of the topics we will discuss are
directly applicable to other types of validation
presented on the previous slide

4
 References

FDA, May 1987. General principles of process validation

Note for Guidance on Process Validation – 2001 - EMEA
PDA technical report No 42 on Process Validation (2005)
Compliance Policy Guide 7132c.08 – section 490 100 on
Process Validation – FDA – 2004
Guidance for Industry – FDA - 2011

5
Regulations?

MCA EC GMPs

CND Inspectors FDA: 21CFR EMEA EC GMPs

SPAC Chinese GMPs
FDA FDA: 21CFR

CPAC Japanese GMPs

MGM WHO GMPs

TGA AUS GMPs

Legend: EMEA - EUROPEAN MEDICINES

FDA = FOOD & DRUG
ADMINISTRATION EVALUATION AGENCY
Regulatory body Regulations

6
 Key Terms

There are some Key Terms that we need to define and

understand before we go any further…….

7
 Input Variables

Operational process parameter (input variable)

An input variable or condition of the manufacturing process that
can be directly controlled in the process. Typically, these
parameters are physical or chemical (e.g. temperature, process
time, column flow rate, column wash volume, reagent
concentration or buffer pH (synonym: process parameter) (PDA
technical report no 42).

8
 Types of Input Variables

Critical operational process parameter

An input process parameter that should be controlled within a
meaningful, narrow operating range to ensure that drug substance
quality attributes meet their specification. Although, parameters
with wide operating ranges may also impact product quality, they
are generally easily controlled and not as likely to result in
excursions that impact quality and are therefore low risk. (synonym
: critical process parameter - CPP). (PDA technical report no 42)

Key operational parameter

An input process parameter that should be carefully controlled
within a narrow range and is essential for process performance
(PDA technical report no 42)

Controlled operational parameter

An input process parameter that has been demonstrated to be
easily controlled or has a wide acceptable limit (PDA technical
report no 42).

9
 Output Variables

Performance parameter
An output variable or outcome that cannot be directly
controlled but is an indicator that the process performed as
expected (synonym: performance attribute) (PDA technical
report no 42).

OUTPUTS

10
 Types of Output Variables

Critical quality attributes

Process measures or outputs that indicate product potency, yield, and/or
purity. They are evaluated to determine whether or not a statistically
significant change occurs in any specific process. Critical quality attributes
become the basis for routine product release testing. A process can have
Intermediate CQA’s and Final CQA’s.

Process Monitoring
Process measures or outputs that are not directly linked to product
potency, yield, and/or purity but do indicate process performance and
repeatability. They are evaluated to determine whether or not a statistically
significant change occurs in any specific process. Some are monitored due
to buisiness (rather than quality) concerns

Process Mapping
Process trending of outputs that do not yet have enough collective data to
statistically indicate the outputs impact or optimal range. Results of
process mapping can feed to new CQA’s or to new monitoring points

11
 A Few Additional Terms

Acceptance criteria
Numerical limits, ranges or other suitable measures for
acceptance of test results. Exceeding the acceptable range for
a critical parameter during validation may result in
questionable product quality that would require initiation of an
investigation and most likely would cause the process
validation to be deemed a failure. Exceeding the operational
range should be documented and explained in the validation
report and evaluated for validation study impact. (PDA
technical report no 42).

12
 Warnings about Acceptance Criteria

Acceptance criteria
Note that the FDA has rejected some process validation lots
(i.e., authored 483’s) due to firms not meeting their acceptance
criteria and then writing protocol deviations to try to justify why
the study was still successful

You must not mix development with validation by trying to

“adjust” your acceptance criteria after execution has started

You must choose your “acceptance criteria” carefully!!

13
 A Few Additional Terms

CMC
Chemistry, Manufacturing, and Controls (CMC) constitutes that
part of pharmaceutical development that deals with the nature
of the drug substance and drug product, the manner in which
both are made, and the manner by which the manufacturing
process is shown to be in control.

The CMC is a complete section of guidance documents

provided by the FDA on what is required for the chemistry,
manufacturing, and controls portion of the submission to the
agency

14
FDA Webpage

15
 The Cliff Notes Process
Validation

Process validation
The documented evidence that the process,
operated within established critical process parameters (CPPs)

can perform effectively and reproducibly to produce an

intermediate or final drug substance
meeting its predefined critical quality attributes (CQA’s)

OUTPUTS

(PDA technical report no 42)

16
 Example: Process Inputs and Outputs

INPUTS OUTPUT
(Process Parameters) (Product Requirements)
I : Insertion Depth RF : Removal Force

F : Insertion Force Bottle

Capping
ID : Inside Diameter Cap Process
OD : Outside Diameter Bottle

17
 Process Validation:
Inputs and Acceptance Criteria

There are several approaches that have been utilized

with respect to process inputs and acceptance criteria
1. All defined inputs are acceptance criteria (critical, key, and
control)– this is not recommended and may be overly
conservative
2. All CPP’s are acceptance criteria while all Key and Control
parameters are “expected conditions.” Expected conditions
may or may not be included in the process validation
protocol.
3. All inputs are “expected conditions” which become locked
as the validated parameters after the validation. Processing
of acceptable future lots with respect to input variables
becomes dependent upon the values encountered during
process validation

18
 Process Validation:
Outputs and Acceptance Criteria

Typically for outputs there is one approach:

All Critical Quality Attributes (CQAs) are acceptance criteria.
All process monitored values are “expected conditions” and
may or may not be included in the process validation
protocol.
Process mapping has no acceptance criteria or expected
conditions associated with it and is not included in the
process validation protocol.

19
 The Order of Things……

Let’s take a few minutes to discuss how we prepare for

process validation….including how we classify the
parameters and choose the acceptance criteria…..

20
 In a Perfect World……

All of the below validations would be complete

BEFORE you do process validation….
Mixing Validation
Hold Time Validation for Raw Materials
Cleaning Validation
Sterilization Validation
Extractable/Leachable Validation
Container Closure Validation
Quality Management of Raw Materials
Qualification of all Equipment
IQ, OQ, PQ

21
 In the Real World……We have ICHQ9 and the
Ability to Utilize Risk Assessments

All of the below must be complete BEFORE you do

process validation….
Mixing parameters must be recorded to ensure they fall
within “to be validated” parameters
Hold Time parameters for Raw Materials must be recorded to
ensure they fall within “to be validated” parameters
Equipment must be verified to be clean
Sterilization Validation (no way to do verification)
Extractable/Leachable risk assessment
Container Closure risk assessment
Qualification of all Equipment
IQ, OQ, PQ
Raw Material Quality Management Data must be recorded to
ensure they fall within “quality management” parameters

22
Changed Paradigm

Pharmaceutical Development (Q8)

Changed Past: Data transfer / Variable output
Paradigm Present: Knowledge transfer / Science based /
Consistent output

Quality Risk Management (Q9)

Past: Used, however poorly defined
Present: Opportunity to use structured
process thinking

Pharmaceutical Quality Systems

(Q10)
Past: GMP checklist
Q9

Future: Quality Systems across product

life cycle

23
 Overall Reduction in Risk!!

Process
Understanding Process Process
Understanding Understanding
CMCCMC
regulatory
regulatory
Oversight CMCCMCregulatory
(Submission)
oversight regulatory
Oversight
(Submission)
oversight CMC
regulatory
oversight
cGMP cGMP
regulatory
regulatory
oversight Q8 cGMPcGMP
regulatory
regulatory
oversight
Q10 cGMP
regulatory
(Inspection)
oversight
& (Inspections)
oversight
& oversight

Company’s
Quality system
Q9 Company’s Q9 Company’s
Quality system
Quality system
Post
Continuous
Approval
approval
Change Improvement
change
(PAC) PAC
PAC toto
Continuous
Continuous
Risk
Risk Improvement Risk
Improvement
(perceived
(P/R)& real) Risk

24
 Process Development

Before you start Process Validation you must do

Process Development – it is from Process
Development that you determine what is critical to your
process

25
Objectives of the Process Development
Phase
Optimization
Understanding of process
Process control
Process reliability/flexibility

26
 Some Types of Process Development

Characterization studies
Typically conducted on a process during development and
industrialization prior to process validation study that evaluates
the process to increase process knowledge and examines
proposed operational ranges and their individual and/or combined
impact on product quality (synonyms: evaluation studies, process
justification studies, engineering studies, development studies,
robustness studies). (PDA technical report no 42)
Process robustness studies
Demonstration that the process, operated within the whole range
of the established critical process parameters limits, can produce
an intermediate or API (drug substance) meeting its predefined
specifications and quality attributes. Process Robustness can be
demonstrated through Experimental Design methodology and/or
using worst case operating conditions.

27
 Some Types of Process Development
(continued)
Clearance Studies
Demonstrate that a process or unit operation removes or
inactivates process or product related impurities. Some
examples:
Process related: Solvents, Enzymes
Product related: Cells, Nucleic Acids, Protein
Note this is removal of impurities, not contaminants (your
cleaning validation and sterilization validation cover
contaminates)

28
 Six Sigma Validation - IQ-OQ-PQ
IQ OQ PQ
(1) Risk Assessment (4) Develop (5) Validate (10) Verify Capability
Measurable Outputs Measurement System of Outputs
LSL USL
FMEA LSL USL
Item Function Failure Mode Causes Effects O S D RPN
(2) Lid Prevents Spills Lid falls off due to looseness Lid diameter Burn 1 5 3 15
too large

Body diameter Burn 3 5 3 45

to small

Surfaces too slick Burn 2 5 2 20

Leaks from drinking hole Hole too large Stain 2 2 1 4

Hole has too Stain 3 2 2 12

direct of path

Spill when removing lid Lid diameter Burn 1 5 3 15

too small

Body diameter Burn 3 5 3 45

to large

Material catches Burn 3 5 5 75

15 20 25 30 35
Seal Strength
(6) Determine and
Characterize Key Inputs (7) Determine Variable
(2) Mistake Proof (11) Process Performance
Relationships
70
K = 20 Testing
T 60 200
1.02
5.6 OC Curve
PRESS O 1.1 1.0

Temperature Hot Bar (of)

R 50 2.0

Q K = -20 1.7
40
U
E 175
1.4 Pa 0.5
30
1.4
36.0 37.0 38.0
1.7
L 0.0
2.0 0 2 4 6 8 10

150 Percent Defective

0.5 0.75
(8) Finalize Control Plan
1.0

Dwell Time (sec)

(3) EquipmentChallenge Control Plan (12) Product Performance

Testing Item Function Failure Mode Causes Effects Control
(9) Determine and Test Testing
(2) Lid Prevents Spills Lid falls off due to looseness Lid diameter
too large
Burn Go-no go gauge
Worst-Case Conditions 1.0
Life Testing
Body diameter Burn Go-no go gauge
to small 0.8

Failure Rate
Surfaces too slick Burn Sampling Plan

Leaks from drinking hole Hole too large Stain Pin gauge 0.6
Hole has too Stain
direct of path 0.4
Spill when removing lid Lid diameter Burn (above)
too small
0.2
Body diameter Burn (above)
to large

29
Material catches Burn Sampling Plan
0.0
0 1 2 3 4 5 6 7 8 9 10

Years

© Copyright 2000, Taylor Enterprises Inc.

 Design for Six Sigma - IDOV
Identify Design Optimize Validate
(1) Define (3) Identify Potential (7) Candidate Inputs (8) Key Inputs (12) Verify Capability
Customer Failure Modes 70 K = 20 of Outputs
T
Requirements FMEA
Pump
O
60
LS L USL

Item Function Failure Mode Causes Effects O S D RPN

R : Piston Radius (in.) F : Flow Rate (ml/min) R 50

(2) Lid Prevents Spills Lid falls off due to looseness Lid diameter Burn 1 5 3 15 L : Stroke Length (in.)
Devel o
Co
pedBy:Way
pyrigh t 20
00
ne Taylor
INPUT
, TaylorEnter p

/
r is es Inc.

Q K = -20
too large
40
Body diameter Burn 3 5 3 45
OUTPUT U
VOC to small

Surfaces too slick Burn 2 5 2 20

S : Motor Speed (rpm) SYSTEM
E 30
Leaks from drinking hole Hole too large Stain 2 2 1 4
B : Valve Backflow (ml)
Hole has too Stain 3 2 2 12
36.0 37.0 38.0
15 20 25 30 35
direct of path

Spill when removing lid Lid diameter

too small
Burn 1 5 3 15
L
Body diameter
to large
Burn 3 5 3 45 S e a l S t r e n g th
(9) Characterize Behaviors of Key Inputs
Material catches Burn 3 5 5 75

(13) Perform Other

(4) Mistake Proof 10.2 Upper Control Limit Testing
OC Curve
1.0
10.0
PRESS
Radius
9.8 Lower Control Limit 0.9 1.0 1.1
Pa 0.5
Radius
9.6

(2) Establish
1 5 10 15 20 (11) Set Targets, 0.0
0 2 4 6 8 10

Measures and
Subgroup
Tolerances and Percent Defective
1.0
Life Testing
Standards Controls 0.8

(10) Discover Variable

Failure Rate
(5) Develop Measurable 0.6

Outputs Relationships 0.4

Requirements 0.2

Definition Seal Strength Std. Dev.

0.0
0 1 2 3 4 5 6 7 8 9 10

Years
Category Requirement Verification Segment Importance 1.02
200 5.6
Function Cup not spill when Roll cup on side for All customers 1
tipped over briefly 5 seconds

Cup not spill when Body diameter All customers 1 1.1

Temperature Hot Bar (o f)

(14) Implement
dropped to small
X3 X6 X9 X2 X8 X5 X7 X1 X10 X4
Cup keep coffee hot Coffee initially heated All customers 2

2.0
to 150oF should stay
above 130o F for at
least 15 minutes

(6) Validate Controls

Cup keep cold drinks cold Water initially cooled All customers 2
to 50oF should stay
below 60oF for at
least 15 minutes
1.7
Fit Cups are stackable Cups should easily Customers 3
stack with the base buying

Measurement System
of one cup fitting inside multiple cups
the lid of another

1.4
175
With lids removed

Seal Strength
Control Plan
cups should fit inside
each other and lids
should easily stack

Cup fits into cup holders Should fit all major Most customers

1.4
provided in cars models of Ford, GM,
Chrysler, Toyota,

Maintenance Cup can be easily

BMW

No hard to clean All customers

Item Function Failure Mode Causes Effects Control
washed crevices or surfaces (2) Lid Prevents Spills Lid falls off due to looseness Lid diameter Burn Go-no go gauge
Cup is dishwasher Should be no warping All customers 1.7 too large
safe when placed in
200o F water for 60 Body diameter Burn Go-no go gauge
minutes to small
Environment Recyclable Made of type A-C
plastic material
All customers
2.0 Surfaces too slick Burn Sampling Plan

Test for traces of lead, All customers

Leaks from drinking hole Hole too large Stain Pin gauge
No dangerous heavy metals, or
substances cancergenic Hole has too Stain
substances direct of path

Ascetics Pleasing color Variety of colors All customers 150 Spill when removing lid Lid diameter Burn (above)
including bright and too small
neutral colors
0.5 0.75 1.0

30
Pleasing shape Get customer All customers Body diameter Burn (above)
input of different to large

Dwell Time (sec)

shapes
Material catches Burn Sampling Plan

Dwell Time
© Copyright 2000, Taylor Enterprises Inc.
 Product and Process Quality Knowledge: Science-Risk Based cGMP’s

Quality by Design GMP/CMC FOCUS

Process Design 1st Design qualification
Principles

MECHANISTIC
UNDERSTANDING
Focused; Critical
Process Control
CAUSAL LINKS Points (PAT)
PREDICT PERFORMANCE

DECISIONS BASED ON
UNIVARIATE APPROACH
Extensive;
Every
Step
DATA DERIVED FROM (CURRENT)
TRIAL-N-ERROR EXPERIMENTATION
31
 Average - 3 Std. Dev. Average Average + 3 Std. Dev.

99.73%

Frequently, histograms take on a bell-shaped appearance that is referred to as the

normal curve as shown below. For the normal curve, 99.73% of the units fall within  3
standard deviations of the average.

32
Statistical Methods and Tools for Validation and Quality Systems

STABLE PROCESS
Total
Variation

e
m
Ti

33
Statistical Methods and Tools for Validation and Quality Systems

UNSTABLE PROCESS
Total
Variation

e
m
Ti

34
Goal: A Capable Process

USL
LSL
You cannot Validate
something that is not in
Control!!

? USL
LSL

35
 Goal: Process Characterization

Proof through documented testing that the following

is understood for each critical parameter...

Optimum Setting

SOP Setting Range

Proven Acceptable Range (PAR)

Product Works at these Process Settings

Product Fails
Acceptable Product in Red Process Setting
“The Cliff”

36
Weighing the Alternatives

Low Sensitivity High Sensitivity

Run at nominal Target PQ runs include worst case

settings. runs.
Process control needed Process control critical to
PQ mainly for cost, not quality. good product.
Not worried about DOE & confirmation runs
interactions much.
Interactions critical
Some latitude on raw
Raw materials critical
materials.
PQ may need to be run at
Run PQ at SOP limits
37 PAR limits
 Lots of Tools to Choose From -

7 tools (histograms, trend plots, Pareto, fishbone, etc…)

Variance Component Analysis (VCA)
Measurement Systems Analysis (MSA)
Risk Analysis (FMEA, etc)
Sampling Plans (for attribute, continuous data)
Descriptive Statistics (average, standard deviation, etc)
Confidence Intervals (CI)
Regression (Predicting outcomes)
Design of Experiments (DOE)
Response Surface Methodology (RSM)
Capability Analysis (Cp, Cpk, statistical tolerance intervals)
Reliability (MTTR, MTBF, etc)
Statistical Process Control (SPC)

38
 Analyze: What Is Critical?

 Check Sheets Defect Type Nov. Nov. Nov. Nov. Nov. Total
18 19 20 21 22 Freq.
 Pareto’s, 1st level, 2nd level, … III I III II
Rust 9
 Cause & Effect Diagrams IIIII II IIII III IIII II III
Dents 24
 Scatter Plots
IIII II III III I I IIII
 Risk Analysis (FMEA) Burs 20

 Process Flow Diagrams Scratches

I III I
5
IIII III IIII IIII III IIII IIII IIII IIII
Dings 50
Material IIII II II

effect 300

250

200
Method Machines
150

100
240
50

ns

in
ps
es

s
s

ck

to
he
io

hi
ch
220

ra
ra
ct

tc

C
at

lo
fe

C
ra
cr

co
r

Sc
pe
S

is
Im
ti c

D
ns
as

Le
d
ol
Pl

M
5.0 6.0

39
 Measure: Establish A Baseline

 Start With Run Charts of Individual data, add Control Limits

 Start With Histograms, add Spec Limits
Capable ?
(conforms to spec’s)

Yes No

USL USL
UCL
Yes
LCL
UCL
LSL LSL

LCL
In Control ? LSL USL
(stable process)

USL USL
UCL UCL
No
LCL LCL
LSL LSL

40
 Analyze

 Factors in your process are Time, Temperature and Force. You have a high and
low specification for each.
 How do you prove that the product produced is OK for each combination of
settings?
 Do you run the high and low of each holding the other two fixed at nominal?
 What are worst case conditions
 Do you run them 3 times?

Design of Experiments (DOE)

Ti Te Fo Diam
1 – – – 17
2 + – – 26
3 – + – 25
4 + + – 85
Temp.

5 – – + 19
6 + – + 24
7 – + + 23
8 + + + 84

Force
41
Analyze

Response Surface Methodology C

B
A

42
 Process Development Reports

Somewhere you need to summarize the results of the process

developmental studies, optimization, scale-up, ect.
Must have something to justify the process parameters used
when performing the performance qualification studies (PQ).
This may be a Process Development Report, a Technology
Transfer Report, or something similar.

43
 Process Development Reports

The developmental report will take

information from places like laboratory
notebooks and pilot batch records and
relate this information to what finally
appears in the master batch
production record.
It is the process described in the
device master record that is the
“validated” process.

44
Process Development Reports

Will include:
History of process development. Who did what, when, how, what
were the results (Good and Bad).
Identification of the important process variables and explanation of
how they affect product.
Summary of development batches including all the operating
parameters used. (A matrix is good for this)
Specifications for the process including upper and lower control
limits. Clear identification of the nominal values. Explanation of how
these were found.
Preliminary Process capability statistics (if applicable)

45
Acceptance Criteria Graphic

46
 In Summary

Fundamental to process validation is process

knowledge and full understanding of the technology
and equipment to be used in the process.

47
Timeline

48
 Process Validation Acceptance Criteria Requirements

Process validation is based on the following elements :

Primary elements with the demonstration of
Process consistency

Clearance of contaminants

Ability to meet predefined acceptance criteria for performance

parameters.

Additional studies such as

stability of process solutions or

chromatography and reusable filter membrane lifetime

49
 Process Validation Options

The different options for process validation :

Prospective validation
For new manufacturing processes
Concurrent validation
For early phase (I/II) clinical trials due to adequate lot-by lot
characterization.
When a retrospective validation/retrospective analysis has also been
performed
When process monitoring (such as PAT) allows a high level of process
control.
Rétrospective validation
Retrospective validation alone is not adequate for new or changed
processes, but can be used to validate well established and well
characterized processes.

50
 Process Validation Options
Laboratory or Industrial scale ?
Process validation studies may consist of a combination of full
scale and scale-down studies, depending on the
appropriateness of the model being used and the nature of the
study.
In cases where scaled-down studies are used, the
appropriateness and the “representativity” of the model must
be demonstrated.

51
 Process Validation vs Consistency Lot Production

Validation of process consistency

Final completion of the validation program requires a
demonstration of process consistency on a minimum of 3
consecutives batches.
Demonstration of process consistency is a confirmation that
the manufacturing facility, equipment and procedures are
capable of consistently manufacturing a drug substance that
meets specifications and quality attributes.
In some cases, process validation and process consistency are
combined in the same 3 lots

52
Validation of process consistency

If a failure occurs on a validation batch,

a full investigation must be conducted to determine the root cause and
apply the appropriate corrective and preventative actions.
The root cause of the failure has to be determined :
If the origin of the failure is not linked with the manufacturing process
then the concerned batch will be discussed in the validation exercise,
and this batch need to be replaced to meet validation requirements.
When the failure is identified as process related, the entire study should
be repeated following the implementation of corrective actions to obtain
three consecutives batches

53
Additional validation studies

Viral clearance
Studies are normally performed by spiking the process stream of a unit
operation in a small scale model of the step. The removal or
inactivation of viruses is then demonstrated by testing the processed
intermediate for the virus after the unit operation has been performed
The timing of viral clearance, number and types of virus depend on
several factors, among them the source of the cell line, knowledge of
virus in the cell line, the nature of additive and media that potentially
contain viruses, the disease that will be studied and the potential of
process steps to inactivate virus
Impurity clearance
Removal of impurities that have the potential to adversely affect
product safety or intended biological activity to sufficiently low levels
should be consistently demonstrated.

54
 Process validation requirements

Additional validation studies

Process intermediate stability (hold time)
This can be accomplished directly at manufacturing scale or in
combination with lab scale validation studies.
Process solution stability
The most straightforward way to validate the stability of these solutions
is to conduct hold time studies during the period which the solutions are
held over the maximum duration in full scale tanks. The validation
protocol should address bioburden, endotoxin and chemical stability.
Drug substance fill, freeze, thaw and storage
Sterilizing filtration
Thawing (process validation should encompass the freezing and
subsequent thawing steps)
Filling processes (product homogeneity all along the filling process,
process capability, absence of particulates, sterility)
Stability studies 55
Additional validation studies
Mixing studies
Mixing steps that are critical process steps must be validated.
Validations studies based on both product homogeneity and product
shear stress, through the following parameters :
Stirring – Volume – Temperature – Mixing time
Resins and reusable filter (UF cassettes) life time validation
The ability of chromatographic resins and reusable membrane to
function consistently over their intended life time (i.e. life cycle) must be
demonstrated.
Small scale studies are typically performed to establish a
recommended life time , that is confirmed at full scale studies via
concurrent validation.
The study performed in the course of the manufacturing include
measurements such as HEPT, water flux, … and demonstration of
impurities clearance.
Storage times and conditions required to maintain performance
characteristics of the resin or membrane for the intended life time
should also be demonstrated
56
Post validation activities

Change Control
On going process monitoring
Carried out through in process controls at specific steps
Test performed at critical decision making steps usually with action
limits and acceptance criteria
Additional statistical analyses of operational parameters
Process Revalidation
Process re validation is change based and not time based, and is
required when significant changes to a process or equipment have
been made.

57
Validation
Change control
In case of changes, consider the change and its impact on
the process validation
Examples of changes (requiring revalidation):
manufacturing process (e.g. mixing times, drying
temperatures)
equipment (e.g. addition of automatic detection
systems)
production area and support system changes
transfer of processes to another site
unexpected changes (e.g. those observed during
self-inspection or during routine analysis of process
trend data)

58
Validation

Prospective validation
SUMMARY

59
Validation

Prospective validation
SUMMARY

Concurrent validation

60
Validation

Prospective validation
SUMMARY

Concurrent validation

Retrospective validation

61
Validation

Prospective validation
SUMMARY

Concurrent validation

Revalidation

Retrospective validation

62
Validation

Prospective validation
SUMMARY

Concurrent validation

Revalidation

Retrospective validation

Change control

63
 Trends,
1 Master
Planning
5 Annual
Reviews
Validation
Plan 2
Process EIQ/functional OQ
Monitoring

Equip.Specification
Validation
Final Report

The Equip IQ Protocol

(checklist)
Result/Data
Validation OQ Protocol for
functional qual.
Process Life Cycle
Validation
Protocol
Result/Data
Process EIQ Final
Specification Report

4 Multi-vari
3
Process Process
Analysis/
Process Specification- Cpk DOE
Optimization (high
Validation PAR level OQ)

64
 Example Process – Baking a Chocolate Cake

Lets assume that you want to license your process for

producing a chocolate cake, and, so that our
discussion is representative of the pharmaceutical
world, that you have already sent batches of your
chocolate cake to clinicals, which went well, so now
you have to make the exact same chocolate cake that
you made for clinicals repeatably to send to
commercial AND you have to validate the process.

65
 Example Process – Baking a Chocolate Cake

Quality Managed Raw Materials

Flour (commodity #1 from vendor A)
Sugar (commodity #2 from vendor A)
Baking Soda (commodity #3 from vendor A)
Eggs (commodity #8 from vendor B)
Milk (commodity #12 from vendor B)
Butter (commodity #15 from vendor B)
Cocoa (commodity #11 from vendor C)
Water for Ingestion (commodity #23 made onsite)

66
 Example Process – Baking a Chocolate Cake

Qualified Equipment
Bowls #100 to 150
Mixers #132 and 133
Oven #300
Cake Pans #425 to 450
Calibrated Stainless Steel Measuring Cups part# 67 to 71
Calibrated Stainless Steel Measuring Spoons part# 75 to 80
Teflon Spatula part# 78
Cleaning Validation is Complete
Mixing Validation is Complete
Hold Time Validation is Complete
All Raw Materials are Quality Managed
EM Validation on grade C area

67
 Example Process – Baking a Chocolate Cake

Output Parameters of the Final Product

R&D has determined that the following output parameters for
the final product are the key indicators of quality

Firmness
Cocoa Concentration
Sugar Concentration
Homogeneity
Moistness of final product

68
The Process

69
The Process

70
The Process

71
The Process

72
The Process

73
The Process

Critical Quality Attributes

Firmness
Homogeneity
Cocoa Concentration
Sugar Concentration
Moistness

74
Baking a Chocolate Cake – Flow Diagram

¾ cups Butter
2 cups cocoa
2tsb baking

1 cup milk
3.5 cups

1.5 cups
powder

2 eggs

Sugar
Flour
Powder Liquid
Bowl# Bowl#

Bowl # Bowl #

Add Raw Materials to Bowl Add Raw Materials to Bowl

Mix with a spatula until visually blended Mix with mixer# on med speed for 5 minutes

2 cups WFI
(minimum of 15 rotations) Store at Room Temperature
Store at Room Temperature

Bowl Oven

Bowl #

Pour cake mix into greased

baking pan (greased with 1 tbs
crisco and coated with flower)
Pre-heat oven to 375F
Place baking pan into oven for 45
Combine Contents of two bowls
minutes
into Liquid Bowl
Check with stick after 45 min,
Begin to blend with spatula
continue baking in intervals of 2
Add WFI 1 cup at a time, using
min until stick comes out clean
spatula to blend while adding
Allow cake to cool on rack for 2
Mix on Med speed for 5 min
hours

75
 Determine the Input Parameters

Make a list of all Input parameters.

Note – there are many input parameters that we do not
control (e.g., elevation of the baking facility, time of day that
cake is manufactured, humidity inside the oven, etc). These
parameters can be identified and listed – but they would be
deemed to be non-critical, non-controlled parameters for the
process as I have developed it. By listing them, you identify
potential process improvement/process issues that may
need to be addressed in the future.

76
 Determine the Input Parameters

ANY input parameter taken to the EXTREMES will

impact the CQA
We need to work within our Design Space defined by the
FDA as:
“The multidimensional combination and interaction of input
variables (e.g., material attributes) and process parameters
that have been demonstrated to provide assurance of
quality. Working within the design space is not considered
as a change. Movement out of the design space is
considered to be a change and would normally initiate a
regulatory postapproval change process. Design space is
proposed by the applicant and is subject to regulatory
assessment and approval.”

77
Define our Design Space

78
 Define our Design Space

The room is a grade C controlled room

The temp and humidity of the room are controlled by the
BAS (building automation system) – therefore for input
variables like “temp during mixing” we can’t go to the point
of freezing or boiling…..the room would be controlled to a
much narrower design space than that
The Equipment is qualified
The Process is defined in the batch record
(e.g., order in which you add ingredients, which equipment
you use when, etc)
Design of Experiments (DOE) and other statistical tools
can be used to define the Design Space based upon the
development data that you have

79
The Input Parameters

Direct Impact on Indirect Impact on No Impact on CQA?

CQA? CQA? Controlled
Input Parameter Critical Parameter Key Parameter Parameter
Volume of Cocoa charged Cocoa Conc
Volume of Flour charged Conc and Firmness
Volume of baking soda charged Firmness
Volume of sugar charged Sugar Conc
Volume of butter charged Conc
Volume of milk charged Conc
Volume of water charged Conc
Volume of eggs charged Firmness
Temp during mixing of solids X
Humidity during mixing of solids X
# of times stirred X
Hold Time of powder mixture X
Temp during mixing of liquids X
Humidity during mixing of liquids X
Mixer Speed Homogeneity
Mixing time Homogeneity
Hold Time of liquid mixture X
Temp during mixing of combined X
Humidity during mixing of combined X
# of times stirred X
Mixer Speed Homogeneity
Mixing time Homogeneity
Hold Time of combined X
Volume of Crisco X
Volume of flour to grease pan X
Oven Temp Moisture
Total Time baked Moisture
Total Time cooled X

80
 So what are my Acceptance Criteria?

If I follow the model that only the CPP’s and the CQA’s
are acceptance criteria – for this example I would have
the following:
CPPS
Volume of Baking Soda
Volume of Eggs
Oven Temp
Total Baking Time
CQAs
Moisture
Conc of Cocoa
Conc of Sugar
Firmness
Homogeneity

81
 What do I file in my Submission

Be VERY careful not to file any more than you HAVE to.
Your flow diagram should only have values for the CPPs
Your flow diagram should show the key and control
parameters – but only in name, no values
Be careful in validation reports that are filed with the agency
Lesson learned with batch records…..you CAN NOT change
your batch record after you file and before you get your
approval

82
Example - Tableting

You are given a tablet

manufacturing flow chart
to study

List the critical steps that are

required to be validated

List the critical equipment

required to be qualified

Identify the variables and

construct a table of inputs
and outputs

83
Determining Critical Quality Attributes of the
Final Product

Output - Particle size distribution of the active(s)

Output – Final moisture content

Output - Granule active content and homogeneity

84
Drying of Active

Inputs for step:

Air Temperature
Dry Time
Amount of Active Charged

Outputs for step:

% moisture

85
Determining Which Inputs Have Direct Impact
on the Outputs

Output - Particle size distribution of the active(s)

Input - Blending time for the powder
Input - Granulating time and speed; amount of
granulating fluid and binder concentration

Output – final moisture content

Input - Drying time
Input – Room Humidity

Output - Granule active content and homogeneity

Input - Blending time of external phase
Input – Amount of active charged
Input – Amount of binder charged

86
Example: Determining critical control points

87
 The Mixing Becomes Critical

Inputs:
Mix Speed
Mix Time
Volume in mixer

88
Tablet compression input variables

Fill volume

Compression force

Turntable speed

Dwell time

Granule size and feed

Ejection force, lubrication

89
 Tablet Compression Output Parameters

l Mass
l Hardness
l Moisture
l Friability
l Disintegration
l Dissolution
l Thickness

90
Questions?

91