PDA Visual Inspection of Injectable Products

Follow Pharma Broadcast on LinkedIn for more
Visual Inspection of Injectable

Products:
More than Sorting Good
from Bad …
John G. Shabushnig, Ph.D.

Insight Pharma Consulting, LLC
[email protected]
December 2021
An Introduction to Visual Inspection
© 2021 John G. Shabushnig
Why We Inspect

2
Why Inspect?
• Patient Risk
– Physiological Implications
– Chemical and Microbiological Implications
• Compendial Requirements
• Regulatory Requirements
• Process Knowledge and Continuous Process
Improvement

3
Particulate Size Ranges
<100 nm 100 - 1,000 nm 1 - 100 µm >100 µm
Nanometer Sub-micron Sub-visible Visible
• SEC (Size Exclusion • Light Obscuration • Manual / Human

Chromatography) • Microscopy • Semi-Automated
• FFF (Field Flow • Flow Microscopy • Automated
Fractionation) • Coulter Counter
• SDS-Page Gels
• AUC (Analytical Ultra-
Centrifugation)
Narhi, et al. J Pharm Sci, 2012
4
Particulate Matter Definitions
• Extrinsic (from outside the process,

uncontrolled)
– Environmental Contaminants
• insect parts, hair, fibers, paint, rust
• Intrinsic (from within the process, unplanned)
– Processing Equipment, Primary Package
• qualified product contact materials (e.g. stainless steel,
glass, rubber, silicone oil)
• Inherent (part of the formulation, controlled
and expected)
RISK
– Protein agglomerates
5
FDA Sterile Injectable Drug Recall
Notices 2017-2021
Visible Particles
Lack of Sterility Assurance
35% Labeling
Container
Other*
15%
33% 3%
* Incl. incorrect potency
14% or dose, discoloration,
impurities/degradation
products and storage
temp excursions.
Data obtained from the FDA Recall and Safety Alerts Archive,
https://www.fda.gov/Safety/Recalls/default.htm
6
US FDA 483 Themes
• Must establish a maximum allowable reject rate.

• Must control reinspection of product, including
when appropriate, inspection conditions and
number of reinspections permitted.
• Inspectors must be trained and training
documented.
• Inspectors must be periodically recertified.
• Training and certification conditions must align
with routine 100% inspection conditions.
• Address inspection fatigue during qualification.

7
US FDA 483 Themes
• Must conduct thorough investigations. Identify

particulate matter when performing investigations.
• Must use statistically sound sampling plan(s) for
AQL inspection.

8
Pharmacopeial Requirements
USP <790> EP 2.9.20 JP 6.06

Illumination 2,000-3,750 2,000-3,750 2,000-3,750 lux
Intensity (lux) (8,000-10,000)*
Inspection 10 sec 10 sec 10 sec
Time (sec)
Background Black/White Black/White Black/White
Acceptance “essentially free “clear and “free of readily
Criteria from visible practically detectable
particulates” particle-free” foreign insoluble
ANSI/ASQ Z1.4 matter”
AQL=0.65%
* Illumination intensity for plastic containers

9
United States Pharmacopeia USP 43
• USP <790> Visible Particulates in Injections
– Inspection conditions defined
• Harmonized with EP
• 2,000-3,750 lux
• Black and white backgrounds
• No magnification
• 5 sec viewing against each background
• Swirl and/or invert sample
– Applies to Extrinsic and Intrinsic particles
– Inherent particles addressed in individual
monographs or approved regulatory filings

10
USP <790> Acceptance Criteria
• At Time of Batch Release
– 100% inspection followed by acceptance sampling
– ANSI/ASQ Z1.4 or ISO 2859
– AQL= 0.65%, UQL= 2.3-3.3% typical
– Alternate and equivalent plans acceptable
• For Product in Distribution
– n = 20, a = 0
– AQL= 0.26%, UQL= 10.9%

11
USP <790> Acceptance Criteria
• Supplemental Inspection
– Where the nature of the contents or the container–
closure system permits only limited capability for
inspection of the total contents, the 100% inspection of
a batch shall be supplemented with the inspection of
constituted (e.g., dried) or withdrawn (e.g., dark amber
container, suspensions, highly colored liquids) contents
of a sample of containers from the batch. The
destructive nature of these tests requires the use of a
sample smaller than those traditionally used for non-
destructive acceptance sampling after 100% inspection.

12
USP <1790>
• <1790> Visual Inspection of Injections
– Information Chapter
– Key elements of an inspection process
• Patient Risk
• Elements of a good inspection process
• Lifecycle / Continuous Improvement
• Visible Defect Types
– Extrinsic, Intrinsic and Inherent
• Inspection Technologies
– Published in USP 40 1st Supplement
• Official Aug 2017

13
EC GMP Annex 1 / WHO Annex 6
• Finishing of Sterile Products
EC 124 / WHO 11.3. Filled containers of parenteral
products should be inspected individually for extraneous
contamination or other defects. When inspection is done
visually, it should be done under suitable and controlled
conditions of illumination and background. Operators
doing the inspection should pass regular eye-sight checks,
with spectacles if worn, and be allowed frequent breaks
from inspection. …

14
EC GMP Annex 1 / WHO Annex 6
• Finishing of Sterile Products
EC 124 / WHO 11.3. … Where other methods of
inspection are used, the process should be validated and
the performance of the equipment checked at intervals.
Results should be recorded.

15
Proposed Revisions to Annex 1
Manufacture of Sterile Medicinal Products

Published Dec 20, 2017
Additional comments requested Feb 18, 2020
8. Finishing of Sterile Products
8.21 Sealing/Container Integrity
8.29 General Visual Inspection
8.30 Manual Inspection
8.31 Automated Inspection
8.32 Trending

16
European Pharmacopeia
• EP 2.9.20 Particulate Contamination: Visible
Particles
Particulate contamination consists of mobile undissolved
substances other than gas bubbles, unintentionally
present in liquid preparations.
The test is intended to provide a simple procedure for the
visual assessment of the quality of liquid preparations, if
applicable after reconstitution, as regards visible particles.

17
• Apparatus
– Vertical matte black panel
– Vertical non-glare white panel next to black panel
– Lamp holder … with … shaded, white light source and …
diffuser (… two 13W fluorescent tubes, each 525 mm
(20.7 in) in length, or an … LED light source). …
illumination at the viewing point is … between 2000
and 3750 lux, although higher values may be required
for for coloured glass and plastic containers and for
coloured or turbid preparations.

18
• Changes …
– Official Jan 2020
– Applicable to reconstituted solutions
– Increased light levels for turbid or colored solutions, as
well as plastic or colored glass containers
– Transfer to clear containers for evaluation when
needed
– Addition of LED as acceptable light source

19
Inspection Methods and
Technologies

20
Manual Inspection

21
Critical Inspection Parameters
• Lighting
– Illumination Intensity
– Uniform, Flicker-free
• Fluorescent, Incandescent, LED
– Tyndall (dark-field)
• Background
– Black / White
• Presentation and Manipulation
– Swirl and/or invert
• Pace
– 10 sec / container reference
22
Human Inspection Performance
100
Probability of Detection (%)
90
Borchert
80 Knapp
70 Ryan
60 Androver
Borchert
50
Melchore
40
30
20
10
0
0 50 100 150 200 250
Particle Diameter (µm)
From Shabushnig, Melchore, Geiger, Chrai and Gerger, PDA Annual Meeting 1995

23
Inspection Performance
• Human Inspection
– Visual acuity
– Fatigue
– Flexibility
• Probabilistic
– Especially true for particulate matter due to
continuously changing presentation.

24
Semi-Automated Inspection
• Machine Material Handling

– Transport, Spin/Rotation, Traying
• Consistent lighting and presentation
– Manual or Machine Rejection
• Human Inspection
– Quality Decision

25
• Semi-Automated
– Similar to manual
– May have poor sensitivity for heavy particles
• Particles stop moving before inspection
– Improved ergonomics
– Improved throughput

26
• Machine Inspection
– Adjustable Sensitivity
• Increased sensitivity must be balanced with increased false
rejection rates.
• Often product dependent
– viscosity, surface tension, container design and variability
• Improved Reproducibility
• Improved Throughput
• High Initial Investment

27
What technique is used for inspection for /
of …
2014 2008 2003 1996
Particles
Manual 49% 33% 46% 33%
Semi-Automated 17% 24% 19% 20%
Automated 33% 43% 35% 42%
Container/Closure
Manual 54% 36% 63% 48%
Semi-Automated 18% 26% 15% 42%
Automated 28% 39% 20% 5%

© 2021 John G. Shabushnig 28
Laboratory: Manual
Inspection Exercise

29
Equipment

Inspection Station Set-up
Switch Measure
on both light
Mark position at
lamps. intensity.
~3000 Lux

Inspection Timing / Pace
• Just listen to the metronome while inspecting

to get the right timing.
• You will here a click for each second and a tone
when you should start and stop the inspection
at each background.

Inspection Technique
• Pick up vial and look for heavy particles on bottom.

• Invert and gently swirl. Do not introduce air bubbles.
• Look for moving particles against black background.
• Repeat against white background.

Defect Classification

34
Defect Classifications
Critical defects are those which make the

product unfit for use. This defect may pose a
risk to patient safety.
Major defects are those which may impair

functionality, processing, or handling that
may lead to a loss of performance.
Minor defects are those which represent a

general lowering of perceived quality (i.e.
appearance) but do not limit the function of
the product or make it unsafe.
35
Critical Defects
• Patient Safety
– Potential hazard to customer
– Vial integrity may be compromised
• Regulatory Conformance
– Does not conform to specifications (e.g. fails sterility,
specific lot release assays, integrity of container
compromised)

36
Critical Defects
• Customer Impact
– Will render product totally unfit for use
• Customer Relations
– Potential loss of customer / market
– Will result in losses greater than the value of the
product (e.g. field alert, product recall, litigation)

37
Major Defects
• Patient Safety
– Will not cause permanent injury
– Conforms to registered specifications

38
Major Defects
• Customer Impact
- Will make the product difficult to use
- Customer complaint or rejection
- Defect noticed by customer
- Potential loss of customer, resulting in reduced
market share

39
Minor Defects
• Patient Safety
– Will not cause personal injury
– Conforms to registered specifications

40
Minor Defects
• Customer Impact
– No functional impact
– Appearance or neatness of product
questionable
– May or may not be noticed by customer
– Low risk of customer rejection
– May be customer /market dependant
– Unlikely to lose customer

41
Typical Drug Product Defects
• High / Low Fill (Gross)
• Particulates / Foreign Material
• Haze / Discolored Solution
• Meltback (lyo)
• Collapsed Cake (lyo)
• Non-homogeneous Suspension (suspension)
• Product does not re-suspend (suspension)

42
Typical Vial Defects
• Crack
• Check
• Bubble/Blister (molded)
• Deformed/Misshaped
• Tooling Marks
• Stone/Inclusion
• Scratch/Airline
• Dirt on Exterior Surface

43
Typical Ampoule Defects
• Crack
• Check
• Deformed/Misshaped
• Stone/Inclusion
• Scratch/Airline
• Dirt on Exterior Surface
• Misshaped Tip
• Incomplete Seal / Pinhole
• Burned Product in Tip
• Wrong Color Band or Color Dot
44
Typical Syringe Defects
• Crack
• Check
• Deformed / Misshaped
• Stone / Inclusion
• Scratch / Airline
• Foreign Material on Exterior Surface
• Missing / Bent Needle
• Missing / Misplaced / Damaged Needle Shield
• Missing / Misplaced Plug
• Liquid Between Ribs in Plug
45
Typical Syringe Defects (cont.)
• Missing / Broken Plunger
• Cracked / Broken Flange

46
Typical Stopper / Plug Defects
• Leaking Container
• Missing or Misplaced Stopper
• Wrong Stopper Type
• “Splash-up”/Dried Product in Seal Area
• Foreign Material on Stopper

47
Typical Seal / Cap Defects
• Missing Cap
• Missing Flip-off Button
• Wrong Cap / Wrong Color Flip-off Button
• Loose Cap
• Incomplete Seal
• “Skirted” Seal
• Torn Seal
• Dented Seal

48
Atypical Defects
• Process to handle atypical defects can be
predefined
• Often requires identification through lab analysis
• Work closely with Quality to determine root
cause and implement corrective actions
• Do not over-react to these types of defects
• Think twice before modifying the inspection
process for these types of defects
• Should track to provide process feedback

49
Inspector Selection and
Qualification

50
Manual Inspection
• Objective of the Manual Inspection Process

– To carefully and consistently observe and remove visual
defects from a controlled process.

Selection Criteria - Vision Testing
• Visual Acuity Testing

– Must pass exam with 20/20
corrected near vision
– Tested annually
– Correction adjusted as needed
• Color Impairment Testing

Training
• Manual Inspection Training Process

– Enrollment into Inspector Training Curriculum
– Attend Injectable Process Overview Training
– Attend Injectable Dosage Form Training
– Tour Inspection and Production areas
– Begin On-the-Job-Training (OJT) process
• Demo phase--->Practice phase--->Proficiency
– Inspection specific SOP training

Training
• Manual Inspection Training Process
– Defect identification/categorization
• Defect Reference Manual
• Defect Samples
– Demo of inspection procedure by Trainer
– Practice in non-production training environment (>2000
vials)
– Inspection of a 'seeded' tray

Training
• Trainee moves to the production environment
– Inspection line pacing
– All vials are verified by trainer prior to release

Training
• Inspector accomplishes ‘proficiency’ on this product
• Inspector ready to inspect on their own in
production environment
• Repeat process for other product
sizes/families/production lines

Performance Maintenance &
Monitoring
• Tray Audit
– Evaluation for missed defects in inspected tray
– On-line immediate feedback after inspection
– A customized database is maintained
– Profile individuals, shift, or unit results
– The inspector's product trays are audited at a rate of 1
full and 3 partial trays each month making sure that
each product is audited annually

Performance Maintenance &
Monitoring
• Procedure Audit
– Each inspector's inspection procedure is blindly audited
to be sure that they are performing the correct
inspection steps
– Confirm compliance to SOP
– Immediate feedback to inspector
– Each inspector is audited at a rate of 2 audits/week
making sure that each product type is audited annually

Inspector Accountability
• Inspector is responsible for keeping current training
on all assigned SOPs
• Inspector will stop and ask for assistance if
something unusual is observed
• Annual vision checks
• Best practices review

Breaks and Rotation
• Various types of breaks help keep the inspector alert
and focused
– Shift & Lunch breaks (2x/shift)
– 5 minute ‘eye breaks’ (each hour)
– Micro-breaks during the inspection process
• Work Center rotations (each hour)
– Each Inspector rotates to a different production line each
hour.

Ergonomic Awareness
• Each inspector is trained in Ergonomics and is

accountable for properly setting up the inspection
station
– Adjustments are made to:
• Chair height
• Light position
• Reach to the product

Stretching
• Each inspector is required to participate in a
focused stretching program 4x/shift
• Reduces fatigue on key areas:
– Hands/Wrists
– Shoulders/Neck
• Reduction in lost time

Acceptance Sampling

63
Sampling vs. 100% Inspection
• Sampling preferred when:
– Test is destructive
– Test cost is high
– Lot size is very large
• 100% Inspection preferred:
– To remove low numbers of randomly distributed
defects
– When risk of a defective unit is high

64
Acceptance Sampling Terms
• Acceptance Sampling
– Inspection of a sample from a lot to decide
whether to accept the lot.
• Lot Size (N)
– The total number of items to be considered
together for acceptance or rejection.
• Sample Size (n)
– The number of items selected at random from
the lot for testing or inspection.

65
• Acceptable Quality Limit (AQL)
– The defect level that will be routinely accepted by
the sampling plan. 95% of the time, lots of this
quality will be accepted. Defines the producer’s
risk.
• Unacceptable Quality Limit (UQL) or Lot Total
Percent Defective (LTPD)
– The defect level that will be routinely rejected by
the sampling plan. 90% of the time, lots of this
quality will be rejected. Defines the customer’s
(patient’s) risk.

66
• Operating Characteristic (OC) Curve
– A plot of the probability of accepting a lot (y-axis)
versus the lot percent defective (x-axis). This curve is
descriptive of the protection provided by a given
sampling plan.

67
Operating Characteristic Curve
1.0
AQL (95%)
Probability of Acceptance
0.8
Single
N = 50,000
0.6
n = 315
0.4 a=6
AQL = 1.1%
0.2 UQL = 3.3%
UQL (10%)
0.0
0 1 2 3 4 5
Lot Percent Defective

68
• Sampling Plan
– Defined by the sample size (n) and the accept
number (a) for a given lot size (N). Performance
is shown by the operating characteristic curve
and values for AQL and LTPD.
• Accept (a)/Reject (r) Number
– When the sample contains defectives <= the
accept for the sampling plan, the lot should be
accepted. When the sample contains defectives
>= to the reject number it should be rejected.

69
• Single Sampling Plan
– Sampling inspection in which the decision to
accept or reject a lot is based on the inspection
of a single sample.
• Double Sampling Plan
– Sampling inspection in which the inspection of
the first sample leads to a decision to accept the
lot, reject it or take a second sample. The
inspection of the second sample, when required,
then leads to a decision to accept or reject the
lot.

70
Effect of Sample Size (n)
1.0
n=8
a=1
Probability of Acceptance
0.8
n=20 Single
a=1
0.6 n=50 N = 50,000
a=2
AQL = 1.1%
n=125
0.4 a=3
n=315
a=6
0.2 n=500
n=1,250 a=9
0.0 a=19
0 1 2 3 4 5
Lot Percent Defective

71
Standard Sampling Plans
• Mil Std 105e “Sampling Procedures and Tables for
Inspection by Attributes”
– Discontinued in 1991
• ANSI/ASQ Z1.4-2003 (R2013) “Sampling
Procedures and Tables for Inspection by Attributes”
– ISO 2859-1:1999
– JIS Z 9015-1:2006
• Dodge-Romig
– Average Outgoing Quality Limit (AOQL)
• Mil Std 1916
– Zero accept plans
72
What AQL value (in %) do you use for acceptance
sampling of these defect categories?
30
48% use patient risk to set AQL values. Critical
25
Major
Responses
20
Minor
15
10 Cosmetic
AQL (%)

73
ANSI Z1.4 Sampling Table

74
ANSI Z1.4 Sampling Table

75
Sample Collection
• Random
• Stratified
– Subgroups
• Periodic
– Simple procedure
– Easy to automate

76
How should a sampling plan be used?
• As a Safety Net
– 100% inspection is not 100% effective
– To detect new defect types with automated
systems
– To detect a process shift
• Manufacturing or Inspection process

77
What should you do when you exceed
the Accept Number?
• Investigate (Always)
– Inspection Process
– Manufacturing Process
• Reinspect (Sometimes)
– When it is appropriate (based on investigation)
• Reject Lot (Sometimes)
– When reinspection is not successful

78
Visual Inspection
Strategies

79
Normal Inspection
• 100% inspection of filled, unlabelled product

• Remove defects using a pre-defined process
• Determine if the Total Defect Level (established
through analysis of historical performance) has
been exceeded
• Critical ~ 0.1%
• Major ~ 3%
• Minor ~ 5%
• Classify defects
• Perform AQL/UQL inspection of accepted units

80
Normal Inspection
• Determine if lot meets acceptance sampling plan

criteria
• Critical ~ 0.065% (median)
• Major ~ 0.65% (median)
• Minor ~ 4% (median)
• If no limits are exceeded, the inspection process is
complete
• If lot exceeds limits, an investigation is initiated

81
Re-Inspection
• Consider re-inspection when defect limits are
exceeded and after investigation
• Consider “focused inspection” based on
information gained from investigation
• Special situations
– e.g. critical defect found post-inspection
• SOP should guide re-inspection process
– Investigation, inspection conditions, limit number of
times re-inspection may be performed

82
Sequential Inspection
• A two-step process, often a hybrid of an
automated inspection (for some specific
attributes, i.e. particulate) and a Manual
Inspection for the remaining attributes of the
product (container, closure, etc.)
• Entire lot inspected by each inspection process or
step
• Also used in conjunction with vial integrity / leak
detection systems
• Uses each method to it’s strengths

83
2-Stage Inspection
• First-stage typically machine inspection.
• Implemented to address high false-reject rate for
a known cause (e.g. air bubbles, container
variation).
• Units accepted by machine are sampled and
assessed against established AQL limits.

84
2-Stage Inspection
• Uncertain units from machine are manually re-
evaluated (second-stage).
• Inspection method may focus on specific attributes
based on machine sorting capabilities and sensitivity
may be increased by allowing additional inspection
time
• Units accepted from second-stage inspection are
sampled and assessed against established AQL
limits.
• The use of tightened limits (larger sample size) is
recommended.

85
Empty Container Inspection
• Inspection process of product container before filling
– Used when drug product is expensive
– Used with special containers / packaging systems
– Used when requested by customer
– Can be used in conjunction with Incoming Quality process
to verify glass quality levels

86
Key Points
• Procedures must be established and
approved for all of these strategies BEFORE
using them.
• Terminology used to describe units within
the inspection process is important.
– Difficult to defend returning a “rejected” unit to
the lot.

87
Visual Inspection Survey

88
Survey Format and Participation
• Objective:
– Document current industry practice for visual inspection of
injectable products.
• On-line survey with multiple choice responses
• 77 questions with blinded responses
• Open to PDA members and non-members
• Response requested by site, so may have multiple
entries for the same company
• 186 Participants

Please keep in mind …
• The same population (PDA Members) was
sampled for each survey, but the specific
companies and manufacturing sites that
responded each year are different. This limits to
some degree the identification of trends.
• The survey documents current industry practice,
but does not indicate if these are good or bad
practices.

Summary and Conclusions
• Good geographic representation in plant location
with NA (48%), EU (29%) and Japan + A/P (19%).
• Good geographic representation of markets
supplied with NA (82%), EU (72%), Japan (58%),
A/P (54%) and SA (52%).
• Good representation of small (<1M units/year) to
large (>100M units/year) manufacturing sites.

• The majority (77%) of surveyed products inspected
are for human use and include a significant amount
(54%) of biological/biotech products.
• The majority of surveyed products inspected are
aqueous solutions (84%) or lyophilized powders
(59%).
• These products are mostly packaged in tubing (70%)
and molded (55%) glass vials, with a significant
number in glass syringes (40%) and ampoules (29%).

• Manual inspection continues to be the most used
method for both particles (46%) and
container/closure (50%).
• Continued interest in using automated inspection
with 50% of firms having plans to implement
systems in the next two years. Similar results
observed in previous surveys.
• Automated systems are validated with production
defects (83%) to be equivalent to manual
inspection (51%).

• Most firms (73%) control manual inspection time
and do not use magnification or polarized light.
• The median inspection time was 6-10 sec per
container which agrees with the current EP and
USP inspection conditions.
• Illumination intensity is typically 2,000-4,000 lux
(60%) which agrees with the current EP and USP
inspection conditions with some (28%) using
higher values.

• Inspection continues to be performed most often
(79%) off-line, but a significant amount (58%) is
also performed in-line with packaging.
• Training (94%), a test of visual acuity (91%) and
inspection performance (90%) are part of the
typical inspector qualification process.
• Annual requalification (79%) continues to be the
typical time interval used for human inspectors.

• Test sets with 100-300 units (50%) with a defect rate
of 5-10% (35%) are used most often for inspector
qualification.
• Inspectors are given a 5 minute (50%) break every 60
minutes / 1 hour (47%) or every 30 minutes (33%).
• Most firms use the same inspection conditions for
different regions (78%), veterinary products (77%) or
clinical supplies (86%).

• The typical total reject rate is 1-2% for aqueous
solutions and <1% for lyophilized powders.
• Differences in typical rejects rates are likely due to
detection ability rather than underlying quality.
• Particles and specifically lint/fibers continue to be
the most common defects observed.

• After 100% inspection, lots are routinely (92%)
audited most often (71%) by QA per equivalent
standards ANSI/ASQ Z1.4, ISO 2859 or JIS Z9015.
• The median values for AQL’s used with these plans
are 0.065% for Critical, 0.65% for Major and 2.5%
for Minor.
• There was a shift in the median AQL value used
for Critical defects from 0.10% to 0.065% and for
Minor defects from 4.0% to 2.5% between 2008
and 2014.

• More firms (56%) classify glass particles as Critical
versus Major (37%)
• There has been a shift to a Critical classification
for particles likely due to regulatory pressure but
this is not consistent with the new USP <790>.
• Firms have established alert/action limits based
on 100% inspection results (89%) and investigate
(88%) and/or reinspect (69%) when these limits
are exceeded.

Regulatory Experience and Future
Expectations
• 44% have been challenged by a regulatory
inspector on their inspection method or
acceptance criteria in the last two years.
– Various reasons
• 38% expect changes in customer expectations in
the next five years.
– Tighter particle limits
• 79% expect changes in regulatory expectations in
the next five years.
– Tighter particle limits

Difficult to Inspect
Products

101
Definition
• A Difficult to Inspect Parenteral (DIP) Product is

one:
“Where the nature of the contents or the container–
closure system permits only limited capability for
inspection of the total contents …” from USP <790>
With “… a diminished capability for 100% visual
inspection of particulate matter due to the container
system, the formulation, or both..” from TR 79

102
Common DIP Examples
• Product or Formulation
• opaque and deeply colored solutions
• lyophilized cakes, powders
• concentrated suspensions, and emulsions
• Package
• amber glass vials and syringes
• plastic syringes
• blow-fill-seal packaging
• medical devices.

103
PDA TR 79
• PDA TR 79 Particulate Matter Control

in Difficult to Inspect Parenterals
– 1.0 Introduction
– 2.0 Glossary
– 3.0 Overview
– 4.0 Expectations for DIP Inspection
– 5.0 Inspection Approaches for DIP
Products/Containers/Devices Released
– 6.0 Defect Prevention: A Lifecycle Approach 2018
– 7.0 Conclusion
– 8.0 References
– 9.0 Appendices
104
Common Destructive Test Methods
• Method 1: Reconstitution
– Suitable for solid powders and lyo products
– Dissolve product in original container with terminally
filtered diluent/solvent
– Perform compendial manual visual inspection.
– The use of non-coring needles to add diluent/solvent is
highly recommended to avoid false positives from
stopper fragments.

105
• Method 2: Filtration
– Dissolve product in filtered diluent/solvent and pass
through appropriate filter membrane (0.45 or 5 µm).
– Count particles >100 µm retained on filter using
microscope.
– Similar to USP <788> Method 2.
– Provides basic particle identification information as well
as count.
– Clean equipment and working environment are
required to avoid false positives.

106
• Method 3: Clarification
– Suitable for suspensions and emulsions
– Add solvent, acid or base to dissolve suspended solids,
break emulsion or dilute solution.
– Alternatively, the clarified solution may be filtered and
examined as in Method 2.

107
• Method 4: Transfer/Dilution
– For use with darkly colored containers (amber/red) or
solutions.
– Product may be transferred to a clear container.
– Darkly colored solutions may de diluted in the original
containers (if sufficient capacity) or transferred to a
larger container for dilution.
– Increased light intensity may be beneficial for visual
inspection (5,000-10,000 lux).
– Transferring product from the original container always
increased the risk of contamination and a false positive.
108
• Method 5: Sieve/Mesh
– Suitable for suspension products
– Like Method 2 but uses a coarser sieve or mesh
(5-30 µm) to retain particles >100 µm .
– Count particles >100 µm retained on sieve using
microscope.
– Provides basic particle identification information
as well as count.

109
• Method 6: Panning
– Suitable for suspension products with a broad
particle size range that cannot be tested using
Method 5 Sieve/Mesh.
– Transfer product to a verified clean petri dish in a
clean environment. Gently swirl to distribute
drug particles and reveal any intrinsic or extrinsic
particles.

110
• Method 7: Rinse/Flush + Filtration
– Suitable for implanted medical devices, empty
contains, infusion tubing and associated
components.
– Component(s) is rinsed with terminally filtered
water or solvent which is collected in a verified
clean container.
– Collected water or solvent is filtered and the filter
examined as in Method 2.

111
Supplemental Test Sample Size
• USP <790>
– Recommends ANSI/ASQ Z1.4 S plans
• USP <1790>
– Recommends the ANSI/ASQ Z1.4 S-3 or S-4 plans.
• TR 79
– Recommends ANSI/ASQ Z1.4 S plans
• These sampling plans generally result in a
sample size of 20 (accept 0/fail ≥1) for most
commercial lot sizes. (Confirm using tables
in the standard.)
112
Automated Inspection

113
Imaging Sensors / Camera Types
• Matrix
Pixel
• Line Scan

114
Image Processing Tools
• Windows
• Edge Detection
• Pixel Count
• Thresholding
• Image Subtraction
• Texture Analysis

115
Front and Back Lighting
Front or Bright-Field
Useful to record surface detail.
Back Lighting
Useful with dark colored

containers and solutions.
Shadow only, no surface detail.
116
Dark-Field Illumination
the “Tyndall Effect”
black Side illumination

background
black
background Bottom illumination
Useful to highlight particles

and cracks.

117
Material Handling
• Transport and Alignment

• Starwheels
• Pucks
• Rotation/Spin
• Grippers
• Allows bottom view
Bosch AIM 8000

118
Artificial Intelligence (AI) and Machine
Learning (ML)
• AI/ML used for inspection recipe development
– Faster
– Better discrimination of good and bad units
• Traditional validation once recipe developed and
locked
– IQ/OQ/PQ
– Challenge with known defects
• Supervised and Unsupervised learning
• PDA PtC on AI/ML applied to AVI in devlopment

119
Key Points
• Automated inspection is generally more
reproducible and offers greater throughput
when compared to manual inspection.
• Automated inspection best for high volume
products.
• Not all products are suitable for automated
inspection.
• Assess performance with samples before
committing to purchase.

120
Validation Methods

121
Validation vs. Qualification
• Equipment must be validated

– IQ/OQ/PQ
• Human inspectors are trained and qualified
– NOT validated

122
Human Performance Baseline
• Establish human inspection performance
with defect test set.
• Can use to make direct comparison to
compendial inspection method.
– USP and EP

123
Test Set Composition
• Defects to include in Test Set
– Multiple examples of anticipated defects
– Weighted toward critical defects
– Defect examples are qualified by multiple
inspection by qualified inspectors. Reject zone
defects (with POD ≥ 70%) selected for inclusion.
• Typical Test Set Size
– 500 to 1000 units

124
Test Set Composition
• Production Defects vs. Standards
– “Real” defects necessary to validate production
performance
– Standard spheres useful to establish baseline,
compare methods and settings and monitor
routine performance
• Defect rate in test set
– For human studies, 10% or less is preferred to
avoid Hawthorne Effect (positive reinforcement)
– Not relevant for machine studies

125
Human Qualification
• Inspector Selection
– Visual Acuity (near-vision)
– Color Perception
• Initial Training
– Defect Examples
• Initial Qualification
• Periodic Requalification

126
Knapp Method
• Multiple inspections of test set to determine
reject probabilities of individual units
• Sort results into the following ranges:
– Accept Zone: P = 0.0 to 0.3
– Gray Zone: P = 0.3 to 0.7
– Reject Zone: P = 0.7 to1.0
• Calculate Reject Zone Efficiency (RZE)
• Calculate Accept Zone Loss (AZL)

127
RZE
RZE = RZR/RZN
Where:
RZR = Reject Zone Rejects, the number of vials
rejected in the Reject Zone
RZN = Reject Zone Number, the total number of vials
in the Reject Zone

128
Knapp Method
• Calculate comparable terms for the Gray and
Accept Zones.
• Accept Zone Loss (AZL) is a measure of the false
reject rate.
• The RZE for an alternative method should be the
same or better than the reference method.
• Gray Zone vials may be “sacrificed” to achieve
higher RZE.

129
Knapp Method
• Remember, as originally published, this method
was designed to assess inspection for particles
only.
• To apply the method to the full range of visible
defects normally addressed, it is necessary to
categorize defects by risk and determine an RZE
for each risk category.
• As published, defect test sets have a ~30%
defect rate. This is very high and likely to bias
the results. A defect rate of 10% or less is
recommended.

130
Equipment Validation (Manual and
Semi-Auto Inspection)
• Equipment Specifications
• Installation Qualification (IQ)
– Utilities
• Operational Qualification (OQ)
– Light Intensity
– Inspection Rate (Semi-Auto)
– Rotation (Semi-Auto)
– Rejection (Semi-Auto)
• Process Qualification (PQ)
– Operator Training and Qualification

131
Specification / Requirements
• User Requirements and Specifications (URS)

– Good validation starts with clear documentation of
the performance expectations for the new
equipment.

132
Factory Acceptance Test (FAT)
• Equipment performance should be

confirmed before acceptance for shipment.
– Check against URS
– Inspection performance should be tested with
samples defects.
– The false reject rate should also be determined.

133
Installation Qualification (IQ)
• Installation Qualification (IQ) should document
receipt and installation of equipment
– Model and serial number
– Features / operating ranges
– Version numbers of software / firmware
– Verify utility connection(s)
– Calibration
– Spare parts
– Change parts

134
Operational Qualification (OQ)
• Operational Qualification (OQ) should document
proper function of equipment component
systems
– Emergency stop(s)
– Eject system(s)
– Man/Machine Interface (MMI)
• Reports
– Other features?
• e.g., lamp failure detection
– Establish appropriate operating ranges

135
Operational Qualification (OQ)
– Detection probability for each defect type
• Compare with human baseline
• Establish reference for routine continuous performance
verification.

136
Validation Criteria
• 100% validate automated inspection equipment.
• Validation Criteria:
– Equivalent to manual: 51%
– Better than manual: 28%
– Other, Not compared to manual: 21%
From 2014 PDA Survey of Visual Inspection Practices

137
Performance/Process Qualification
• Performance / Process Qualification (PQ)
confirms expected performance with full
production lots
– Method 1
• Inspect three production lots by both manual and
automated methods
• Compare defect detection rates
• Determine false reject rate
– Method 2
• Inspect three production lots by automated method
and use a tightened sampling plan to assess
performance
• Determine false reject rate
138
Routine Performance Verification
• Typically run before each batch
• Small test set to challenge each sensor/camera
station
• Gross examples to assure rejection, Go/No Go
test
• Does not challenge sensitivity, but rather
camera alignment, functionality and proper
operation of reject system.

139
Inspection Myths

140
Inspection Myth #1
• 100% inspection means detection and

elimination of all visible defects (e.g.
particulate matter, cracks, etc.)
– Inspection is a probabilistic process.
– Detection probability is dependant on inspection
conditions and defect characteristics.
– Particles <200 um generally have a detection
probability <100%.

141
Inspection Myth #2
• Human manual inspection is a “validatable”
process.
– Human inspectors are not “validatable”.
– Qualified human inspectors can provide reliable
performance.
• Defined selection and training criteria
• Controlled inspection conditions
– Lighting, Background, Duration
– SOP’s

142
Inspection Myth #3
• Magnification always improves human manual
inspection performance.
– Inspectors will move head position to minimize eye-
strain during extended inspection, reducing apparent
magnification.
– Controlled studies have not found increased
detection of particulates or container defects with 3x
magnification.

143
Inspection Myth #4
• If you use a sampling plan with an AQL of 0.1%

and do not exceed the accept number in your
sample, the defect rate in your batch will not
exceed 0.1%.
– AQL is the Acceptable Quality Level and is the defect
rate where the rejection probability is 5%. 95% of
batches with this defect rate will be accepted. This is
a measure of the risk of rejecting good batches.
– The UQL is the Unacceptable Quality Level and is the
defect rate where the rejection probability is 90% for
the batch.

144
Conclusions

145
Conclusions
• Current industry performance is generally at

or beyond the limits of medical risk.
• Compendial guidance is ambiguous, but
getting better.
• “Zero defects” is a valuable goal, not a
practical limit for particulate matter.
• Need to develop practical limits based on risk
assessment and process capability measures.

146
References and
Acknowledgements

147
Papers
• Good Practices in Visual Inspection
– Drury, C.G., Watson, J, Federal Aviation Administration, Flight
Standards Service (2002), www.faa.gov
• Visual Inspection: A Review of the Literature
– See, J.E., Sandia National Laboratories Report SAND2012-8590.
(2012) www.sandia.gov
• Rare Items Often Missed in Visual Searches
– Wolfe, J., Horowitz, T. and Kenner, N., Nature, 435, 439-440 (2005)

148
Papers
• Generalized Methodology for Evaluation of Parenteral
Inspection Procedures
– JZ Knapp and HR Kushner, J. Parent. Sci & Techn. 34 (1), pgs. 14-61
(1980)
• Implementation and Automation of a Particle Detection
System for Parenteral Products
– JZ Knapp and HR Kushner, J. Parent. Sci & Techn. 34 (5), pgs. 369-
393 (1980)
• PDA Survey – 2014 Visual Inspection
– JG Shabushnig, Parenteral Drug Association, October 2015

149
Papers
• Intravenous Fluids: A Solution Containing Such Particles
Must Not Be Used
– JM Garvan and BW Gunner, Med J Austr. 2, pgs. 140-145 (1963)
• The Harmful Effects of Particles in Intravenous Fluids
– JM Garvan and BW Gunner, Med. J. Austr. 2, pgs. 1-6 (1964)
• Particles in Intravenous Solutions; A Review
– WH Thomas and YK Lee, New Zealand Med. J. 80, pgs. 170-178
(1974)
• Foreign Particle Embolism in Drug Addicts: Respiratory
Pathophysiology
– FG Douglas, et al, Annals Int. Med. 75, pgs 865-872 (1971)

150
Papers
• Potentially Hazardous Effects of Introducing Particulate
Matter into the Vascular System of Man and Animals
– AM Jonas, Proceedings of the Safety of Large Volume Parenteral
Solution Symposium, Washington D.C., July 28-29, 1966
• Industry Perspective on the Medical Risk of Visible
Particles in Injectable Drug Products
– Bukofzer, S., Ayres, J., Chavez, A., Devera, M., Miller, J., Ross, D.,
Shabushnig, J., Vargo, S., Watson, H., and Watson, R., PDA J Pharm
Sci and Technol 69, 123-139 (2015)

151
Papers
• Visible Particulates in Injections – A History and a Proposal
to Revise USP General Chapter Injections <1>
– RE Madsen, RT Cherris, JG Shabushnig and DG Hunt, Pharmacopeial
Forum, 35(5) pgs. 1383-1387, Sept-Oct 2009.
• Particulate Matter in Injectable Drug Products
– Stephen E. Langille, PDA J Pharm Sci and Technol, 67 (3) pgs. 186-
200 (2013)
• Considerations for Design and Use of Container Challenge
Sets for Qualification and Validation of Visible Particulate
Inspection
– James A. Melchore and Dan Berdovich, PDA J Pharm Sci and
Technol, 66 (3) pgs. 273-284 (2012)

152
Books and Journals
• Visual Inspection and Particulate Control
– D. Scott Aldrich, Roy T. Cherris and John G. Shabushnig, DHI Press
©2016, PDA Bookstore
• Control of Particulate Matter Contamination in Healthcare
Manufacturing
– Thomas A. Barber, CRC Press ©1999
• Pharmaceutical Particulate Matter; Analysis and Control
– Thomas A. Barber, Interpharm Press ©1993
• Particulate Matter; Sources and Resources for Healthcare
Manufacturers
– Michael J. Groves, Interpharm Press ©1993

153
Books and Journals
• Liquid & Surface-Borne Particle Measurement Handbook
– Julius Z. Knapp, et. al., Marcel Dekker ©1997
• Illuminating Engineering Society of North America (IESNA)
Lighting Handbook
– Ed. Mark S. Rea, 9th Edition, ©2000
• Guide to Acceptance Sampling
– Wayne A. Taylor, Taylor Enterprises, Lake Villa, IL, ©1992

154
Journals
• PDA Journal of Pharmaceutical Science and Technology
• PDA Technical Report No. 43 (Revised 2013):
Identification and Classification of Nonconformities in
Molded and Tubular Glass Containers for Pharmaceutical
Manufacturing: Covering Ampoules, Bottles, Cartridges,
Syringes and Vials (2013)
• PDA Technical Report No. 76: Identification and
Classification of Visible Nonconformities in Elastomeric
Components and Aluminum Seals for Parenteral
Packaging (2016)
• PDA Technical Report No. 79: Particulate Matter Control
in Difficult to Inspect Parenterals (2018)
155
Regulatory & Compendial
• US Pharmacopoeia (USP)
– <771> Ophthalmic Products – Quality Tests
– <787> Subvisible Particulate Matter in Therapeutic Protein
Injections
– <788> Particulate Matter in Injections
– <789> Particulate Matter in Ophthalmic Solutions
– <790> Visible Particulates in Injections
– <1787> Measurement of Subvisible Particulate Matter in
Therapeutic Protein Injections
– <1788> Methods for the Determination of Particulate Matter in
Injections and Ophthalmic Solutions
– <1790> Visual Inspection of Injections

156
• European Pharmacopeia / Pharm Europa (EP)
– 2.9.19 Particulate Contamination: Sub-Visible Particles
– 2.9.20 Particulate Contamination: Visible Particles
– 5.17.2 Recommendations on testing of particulate contamination:
visible particles (DRAFT)
• Japanese Pharmacopoeia (JP)

– 6.06 Foreign Insoluble Matter Test
– 6.07 Insoluble Particulate Matter Test for Injections

157
• US Code of Federal Regulations (CFR) 211 Food and Drugs
Subpart B – Organization and Personnel
– 211.25 Personnel qualifications
Subpart C – Buildings and Facilities
– 211.42 Design and construction features
– 211.56 Sanitation
Subpart D -Equipment
– 211.63 Equipment design, size and location
– 211.65 Equipment construction
– 211.67 Equipment cleaning and maintenance
– 211.68 Automatic, mechanical, and electronic equipment

158
Subpart E - Control of Component and Drug Product Containers and
Closures
– 211.80– General requirements
– 211.84 Testing and approval or rejection of components, drug
product containers, and closures
– 211.94 Drug product containers and closures
Subpart F – Production and Process Controls
– 211.100 Written procedures: deviations
– 211.110 Sampling and testing of in-process materials and drug
products
Subpart I – Laboratory Controls
– 211.160 Laboratory controls – general requirements
– 211.165 Testing and release for distribution

159
Subpart J – Records and Reports
– 211.188 Batch production and control records
– 211.192 Production record review
– 211.194 Laboratory records
– 211.198 Complaint files
Subchapter F - Biologics
– 600.10 Personnel
– 600.11 Physical establishment, equipment, animals, and care

160
• EC Guide to Good Manufacturing Practice – Annex 1
Manufacture of Sterile Medicinal Products
• British Pharmacopeia (BP)
• Chinese Pharmacopeia (ChP)
• Japanese Guidance for Industry: Sterile Drug Products
Produced by Aseptic Processing
• German Pharmaceutical Codex (DAC)
• WHO International Pharmacopoeia
• FDA Warning Letters and 483 Observations
– FDA website
– GMP Trends

161
Conferences and Meetings
• PDA Visual Inspection of Parenterals Interest Group
• PDA Visual Inspection Forums

162
Equipment Vendors
• Antares Vision
– Brescia, Italy www.antaresvision.com
• Brevetti C.E.A., S.p.A.
– Sovizzo, Italy www.brevetti-cea.com
• Bonfiglioli Engineering, S.r.l.
– Vigarano Pieve, Italy www.bonfiglioliengineering.com
• Dabrico, Inc.
– Kankakee, IL www.dabrico.com
• Syntegon (formerly Eisai, Bosch)
– Tokyo, Japan www.eisai-mc.co.jp/english
• InnoScan K/S (Stevenato Group)
– Braband, Denmark www.innoscan.dk

163
Equipment Vendors
• Optrel (Stevenato Group)
– Padova, Italy www.optrelinspection.com
• Phoenix Imaging
– Livonia, MI www.phoeniximaging.com
• Seidenader Maschinenbau, GmbH (Korber)
– Munich, Germany www.seidenader.de
• Unchained Labs (Rap.ID Particle Systems)
– Pleasanton, CA www.unchainedlabs.com
• Victor International Marketing, Inc.
– Morristown, NJ, www.victorinternational.com
• Wilco AG
– Wohlen, Switzerland www.wilco.com

164
Standards Vendors
Standard Particles:
• Duke Scientific Corp.
– Palo Alto, CA www.dukescientific.com
• Mo-Sci Corp.
– Rolla, MO www.mo-sci.com
• National Institute of Standards (NIST)
– Gaithersburg, MD www.nist.gov
• Poly Sciences, Inc.
– Warrington, PA www.polysciences.com

165
Standards Vendors
Finished Standard Containers:

• Material Analytischer Service (M.A.S.)
– Freiburg, Germany www.ma-service.de
• Micro Measurement Laboratories, Inc.
– Wheeling, IL www.mmlabs.com
• RJ Lee Group
– Monroeville, PA www.rjlg.com
• SoloHill Engineering, Inc
– Ann Arbor, MI www.particlestandards.com

166
Acknowledgments
• PDA Task Force

– Julius Z. Knapp – R&D Associates
– Roy T. Cherris – Bridge Associates International
– Russell E. Madsen – The Williamsburg Group, LLC
• Pfizer Inc
– Stephen J. Borchert (retired)
– D. Scott Aldrich (retired)
• MIBIC, GmbH
– Markus Lankers

167

PDA Visual Inspection of Injectable Products

Uploaded by

Copyright:

Available Formats

PDA Visual Inspection of Injectable Products

Uploaded by

Document Information

Original Description:

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

PDA Visual Inspection of Injectable Products

Uploaded by

Copyright:

Available Formats

Follow Pharma Broadcast on LinkedIn for more

Visual Inspection of Injectable

John G. Shabushnig, Ph.D.

An Introduction to Visual Inspection

An Introduction to Visual Inspection

<100 nm 100 - 1,000 nm 1 - 100 µm >100 µm

Nanometer Sub-micron Sub-visible Visible

• SEC (Size Exclusion • Light Obscuration • Manual / Human

• Extrinsic (from outside the process,

• Must establish a maximum allowable reject rate.

An Introduction to Visual Inspection

• Must conduct thorough investigations. Identify

An Introduction to Visual Inspection

USP <790> EP 2.9.20 JP 6.06

* Illumination intensity for plastic containers

An Introduction to Visual Inspection

An Introduction to Visual Inspection

An Introduction to Visual Inspection

An Introduction to Visual Inspection

An Introduction to Visual Inspection

An Introduction to Visual Inspection

An Introduction to Visual Inspection

Manufacture of Sterile Medicinal Products

An Introduction to Visual Inspection

An Introduction to Visual Inspection

An Introduction to Visual Inspection

An Introduction to Visual Inspection

An Introduction to Visual Inspection

An Introduction to Visual Inspection

An Introduction to Visual Inspection

An Introduction to Visual Inspection

• Machine Material Handling

An Introduction to Visual Inspection

An Introduction to Visual Inspection

An Introduction to Visual Inspection

An Introduction to Visual Inspection

An Introduction to Visual Inspection

An Introduction to Visual Inspection

An Introduction to Visual Inspection

• Just listen to the metronome while inspecting

An Introduction to Visual Inspection

• Pick up vial and look for heavy particles on bottom.

An Introduction to Visual Inspection

An Introduction to Visual Inspection

Critical defects are those which make the

Major defects are those which may impair

Minor defects are those which represent a

An Introduction to Visual Inspection

An Introduction to Visual Inspection

An Introduction to Visual Inspection

An Introduction to Visual Inspection

An Introduction to Visual Inspection

An Introduction to Visual Inspection

An Introduction to Visual Inspection

An Introduction to Visual Inspection

An Introduction to Visual Inspection

An Introduction to Visual Inspection

An Introduction to Visual Inspection

An Introduction to Visual Inspection