See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/225064007

Considerations for Design and Use of Container Challenge Sets for

Qualification and Validation of Visible Particulate Inspection

Article · May 2012

DOI: 10.5731/pdajpst.2012.00862 · Source: PubMed

CITATIONS READS
6 2,546

2 authors:

James A Melchore Jr. Dan Berdovich

Melchore Consulting LLC Micro Measurement Labs
10 PUBLICATIONS   36 CITATIONS    3 PUBLICATIONS   6 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Challenge sets used for the validation of inspection methods View project

Creating Placebo/Mimic solutions to simulate drug products for standards View project

All content following this page was uploaded by James A Melchore Jr. on 07 October 2015.

The user has requested enhancement of the downloaded file.

 Downloaded from journal.pda.org on May 30, 2012

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 Tech 2012, 66 273-284

Access the most recent version at doi:10.5731/pdajpst.2012.00862
 Downloaded from journal.pda.org on May 30, 2012

Considerations for Design and Use of Container Challenge

Sets for Qualification and Validation of Visible Particulate
Inspection
JAMES A. MELCHORE1,* and DAN BERDOVICH2
1
Melchore Consulting and 2Micro Measurement Laboratories ©PDA, Inc. 2012

ABSTRACT: The major compendia require sterile injectable and ophthalmic drugs, to be prepared in a manner that
is designed to exclude particulate matter. This requirement is satisfied by testing for subvisual particles in the
laboratory and 100% inspection of all containers for the presence of visible particles. Inspection for visible particles
is performed in the operations area using one of three methods. Manual inspection is based on human visual acuity,
the ability of the inspector to discern between conforming and nonconforming containers, and the ability to remove
nonconforming units. Semi-automated inspection is a variation of manual inspection, in which a roller conveyor
handles and presents the containers to the human inspector. Fully automated inspection systems perform handling,
inspection, and rejection of defective containers. All inspection methods must meet the compendial requirement for
sterile drug product to be “essentially free” of visible particulates.
Given the random occurrence of particles within the batch, visual detection of a particle in an individual container is
probabilistic. The probability of detection for a specific particle is affected by many variables that include product
attributes, container size and shape, particle composition and size, and inspection capability. The challenge set is a useful
tool to assess the particle detection in a product, and it may also be used to evaluate detection of container/closure defects.
While the importance of a well-designed challenge set is not always recognized or understood, it serves as the cornerstone
for qualification and/or validation of all inspection methods. This article is intended to provide useful information for the
design, composition, and use of container challenge sets for particulate inspection studies.

KEYWORDS: Challenge sets, Container/closure defects, Inspection validation, Kits, NIST spheres, Particle inspec-
tion, Particulate matter

LAY ABSTRACT: Regulations require drug products intended for injection or ophthalmic use to be sterile and free of
particles that could harm the patient. This requirement is meet by 100% inspection of every drug container in the lot
and the removal of any defective unit before it is released for patient use. Great progress has been made through the
creation of a harmonized method for the detection of small particles in drug product and universal recognition of
container defects.
Differing opinions concerning the conduct of large particle inspection have hindered the creation of a harmonized method.
The absence of a standard method has created confusion that must be resolved for acceptance of drug products in the global
marketplace. While the importance of a well-designed test set for qualification of these methods is overlooked or
misunderstood, it can serve as the cornerstone for qualification of all inspection methods. This article is intended to provide
valuable information for test sets used qualify inspection systems. The proper design and use of these test sets will provide
clarity for inspection qualification, which can be applied to the inspection of commercial product.

Introduction tential harm to the patient; however, the circumstances

surrounding their detection and rejection are very dif-
Inspection of product in the operations area is primar- ferent. Most container/closure defects are easier to
ily focused on container/closure defects and particu- detect than particulate matter in product and guidance
late matter in product. Both type defects present po- provided by the Parenteral Drug Association (PDA)
and the U.S. Food and Drug Administration (FDA) has
enabled companies to create clearly defined accep-
* Correspondence author: james.melchore@me.com
tance criteria and disposition for each type of contain-
doi: 10.5731/pdajpst.2012.00862
er/closure defect (5–7).

Vol. 66, No. 3, May–June 2012 273

 Downloaded from journal.pda.org on May 30, 2012

The major compendia have harmonized the testing Automated Inspection: There is increased use of
methodology and acceptance criteria for subvisible fully automated inspection systems that perform both
particles (1– 4); however, the absence of a harmonized container/closure and particulate inspection in a single
guidance for “visible particles” has led to confusion in pass. Data from numerous studies has proven that
the global industry. The wording may vary, but the automated inspection systems are more sensitive and
major compendia share the vague expectation for in- consistent in detecting rejects than manual inspection.
jectable and ophthalmic products to be “essentially Automated inspection settings are dependent on the
free” of visible particulate matter. This expectation is physical properties of the product, container, fill level,
based on human visual acuity, which is subjective and and the particle type and size. It is essential that the
can be affected by many variables. In the absence of a challenge set has the same attributes as the commer-
harmonized guidance, this review article is based on cial product and that it contains a sufficient number of
published studies and experience gained in the design defective units to provide an accurate assessment of
and use of challenge sets for particulate matter studies. process capability.

Manual Inspection: The pioneering work of Knapp The Role of Challenge Sets for Inspection
and Kushner in 1980 introduced an inspection meth- Qualification/Validation
odology that minimized manual inspection variability
and provided a statistical means to measure inspection Challenge sets have many uses. They can be used to
performance (8). Controlled manual inspection contin- train inspectors, qualify and requalify inspectors, and
ues to be a practical means to detect visible particulate establish a manual performance baseline. The manual
matter. The USP, the Code of Federal Regulations,
baseline can be used to validate alternate inspection
and the FDA provide minimal information regarding
methods, evaluate changes to the inspection process,
the conduct of particulate inspection. Section 2.9.20 in
or perform preproduction functionality testing of an
the European Pharmacopeia (EP) provides the most
automated inspection system. In our experience, the
definitive guidance for manual inspection methodol-
size and scope of a challenge set may expand with
ogy, which is consistent with many documented stud-
experience, and useful information that is gained can
ies. In the absence of a harmonized guidance, the EP
lead to process improvements. Challenge sets provide
provides a sound basis for qualification of manual
a measurable link between the probability of detection
inspection. Although human visual inspection may be
(POD) for various particle types and sizes. All chal-
affected by many variables, published data indicates
lenge sets contain a subset of defect-free containers
that most of these variables can be controlled to pro-
and a subset of containers having one known defect.
vide reproducible data (8 –19).
The challenge set, in combination with predetermined
Semi-Automated Inspection: Many companies use acceptance criteria and a sound inspection strategy,
semi-automated systems, installed in-line with filling enables qualification/validation of inspection methods
or packaging operations to assist with container/clo- designed to differentiate between defect-free versus
sure inspection. Filling and packaging lines typically defective containers.
run at speeds ⱖ100 units/min and inspection must be
qualified using representative defects and the same The principles for the design and use of a challenge set
conditions intended for the inspection of commercial apply to manual qualification and validation of auto-
product. The inability to qualify in-line particulate mated systems for particle detection or container/clo-
inspection at excessive line speeds, is the reason why sure defect detection. During inspection of commer-
of the injectable drug manufacturers that participated cial product, container/closure defect and particulate
in the latest PDA survey, only 24% perform particu- inspection may be performed in a one- or two-step
late inspection in-line (20). In-line inspection for vis- process. Inspection for container/closure defects in-
ible particles under these conditions compares unfa- volves slow rotation of the container to examine all
vorably to manual inspection performed in an surfaces, while particulate detection requires a greater
inspection booth, having a median inspection time of rotation speed to put particles in motion. Maintaining
5– 6 seconds against each color background (20). separate data and calculations for each type of defect
Semi-automated inspection systems may be qualified is essential to gain an understanding of both aspects of
for particulate inspection, when installed off-line, with inspection. Cherris supported this position at the 2011
additional inspection time. PDA Visual Inspection Forum (9).

274 PDA Journal of Pharmaceutical Science and Technology

 Downloaded from journal.pda.org on May 30, 2012

Considerations in the Design and Use of a the importance of a stable formulation and how pro-
Challenge Set tein aggregates ⱕ50 ␮m may lead to visible particle
detection (27).
Shelf Life: It takes a considerable amount of time and
resources to manufacture a quality challenge set. A subvisual particle analysis of a biopharmaceutical or
Borchert and other investigators quickly realized the complex bulk formulations can provide valuable in-
value of having a stable challenge set (10, 11, 13, 14, formation regarding the enumeration and characteriza-
22, 27). Challenge sets intended for inspector qualifi- tion of particles inherent to the formulation (33). This
cation/requalification and establishing a manual base- is especially useful for products that have opalescence,
line must be stable and have reproducible results for turbidity, suspensions, and smaller particles that may
long-term use. Aqueous challenge sets that are filled join to form larger (aggregated) particles. Additional
aseptically with a stable solution and compatible con- testing for subvisual particles in the product, after it is
tainer/closure typically have a useful shelf life of 3 to filled and sealed, may provide insight into external
5 years. particle contribution and potential interaction between
the formulation and container/closure. Having this
Challenge Sets for Manual Inspector Training and knowledge before beginning visual inspection studies
Qualification: A challenge set used to qualify/ can save a great deal of time and failed studies. The
requalify manual inspectors or to establish an in-house scope of this article does not permit going into further
process capability must be stable and capable of with- detail regarding biopharmaceuticals or other difficult-
standing repeated handling over time. Many investi- to-inspect products; however, the reference article by
gators have recognized the importance of using a Singh et al. provides detailed information regarding
well-designed challenge set for inspector qualification the best practices for formulation and manufacturing
and to establish a statistically sound manual baseline of biotech drugs (26). Expertise with orthogonal par-
for validation of automated systems (8, 9, 11, 13, 14, ticle investigational methods has grown at laboratories
21–23). Challenge sets intended for manual inspection with new instrumentation, and future studies should
studies must contain particles representative of those provide more detailed information on the inspection of
found in commercial product (e.g., glass, rubber, steel, biopharmaceutical products.
etc.) that are within the range of visual acuity. The
particles may be sourced from actual manufacturing Container Configuration: Container configuration
rejects or may be simulated particles manufactured in and the material of construction can affect the ability
a laboratory. A stable challenge set can also be used to to detect particles. Amber glass and opaque plastic
evaluate process drift or measure the effect(s) of a containers require increased light intensity and/or in-
change to the inspection process (21). creased camera gain to detect particles. When in-
creased light intensity or increased inspection duration
Product Characteristics: “Know thy product” seems is used to inspect photosensitive products, the same
overly simple, but product-specific attributes such as inspection method intended for commercial inspection
viscosity, specific gravity, surface tension, tendency to should be used to inspect stability samples to ensure
form air bubbles, head space, terminally sterilized that product stability is not adversely affected by
products containing schlieren lines (stratified layers of enhanced inspection lighting. By constructing a small
different density), and other factors can directly affect challenge set with each of the container sizes and fill
inspection methodology and the composition of the volumes, with specifically designated particle sizes
challenge set. Drug product attributes can significantly and types, the differences in POD and other inspection
affect inspection methodology, and they must be un- variables can be better compared with statistical con-
derstood before designing a challenge set (23–30). fidence. Syringes present challenges due to their
Studies by Rathore, Sing, Shnek, Gidth, and Deng small-bore size, which provides less room to form a
provide a great deal of detailed information concern- vortex during spinning.
ing the characteristics of biopharmaceutical products
and the idiosyncrasies of inspecting these products Studies by Shnek et al. experienced problems with
(26 –30). Rathore et al. discussed in detail how the standard spheres sticking to the interior surfaces of
viscosity, density, and surface tension of biopharma- prefilled syringes filled with biologic products (28).
ceutical formulation could affect particle detection These studies indicate that physical properties of the
with automated inspection (26). Singh et al. stressed particle can affect detection in syringes. The studies

Vol. 66, No. 3, May–June 2012 275

 Downloaded from journal.pda.org on May 30, 2012

Figure 1

Probability of detecting particle sizes in small volume clear glass containers (12).

also included the seeding of syringes with one, versus detected 0 –3% of the time with multiple inspections.
multiple, spheres (28). Gidh et al. approached the The inclusion of a 50 ␮m particle adds little value for
problem of inspecting viscous liquids in prefilled sy- routine qualification, but it may provide confidence in
ringes by suspending particles in varied formulations studies designed to evaluate the lower range of visual
using NaCl, polyethylene glycol (PEG), water for in- detection.
jection (WFI), and 5% polysorbate to increase particle
suspension and detection (29). Deng et al. took a In the search for a quantitative approach to validation,
similar approach by using a placebo solution that much attention has been given to particle size as the
mimicked the properties of the drug product. Their criterion that contributes to particle detection. Other
findings concluded that particle detection is more dif- particle attributes including shape, composition, re-
ficult in syringes filled with viscous solutions, which flectivity, color, contrast, density, and particle behav-
tend to generate air bubbles during spinning (30). ior during spinning and inspection must also be con-
Unique (undetectable) methods have been introduced sidered. Glass shards will tend to twinkle due to the
to seal the needle path to prevent particle loss. edges of the particle, while this is not seen with a
NIST (National Institute of Standards and Technol-
How Many Containers Belong in the Challenge ogy) glass sphere. Container type, size, shape, and fill
Set?: Knapp recommended the use of a challenge set volume affect the apparent size of a particle and must
consisting of 250 containers in his initial validation be considered when selecting particle sizes for testing.
procedure (8). The absolute number of containers in The studies referenced in this article indicate that one
the challenge set is not as critical as having a sufficient particle size range for manual inspection may not be
number of containers that represent the range of par- suitable for all product/container combinations. Parti-
ticle types and sizes required to provide a statistical cle sizes and types to be seeded into challenge set con-
assessment of process capability. The design of exper- tainers are determined by the end-use of the set, and they
iment (DOE) and end-uses of the challenge set should should be specified in the DOE. The particle sizes used in
determine the number of required containers in the set. Figure 1 serve as a practical guideline for studies con-
ducted with small volume clear glass containers (12). For
Particle Size and Type: What particle sizes and types certain containers, this may not be evident. A limited,
should be seeded into a challenge set container? There small inspection study, using a few challenge contain-
is no universally accepted requirement for specific ers with small sizes, may provide useful information
particle sizes to be included in inspection studies. for the overall particle size range in the DOE.
When considering the overall range of particle sizes,
50 ␮m is accepted as the threshold for a human Dunham conducted manual inspection studies utilizing
inspector having 20/20 visual acuity. This value is a light intensity of ⬎500 foot-candles (⬎5382 lux),
often misunderstood, as a 50 ␮m particle may only be combined with a 15 s inspection duration, to determine

276 PDA Journal of Pharmaceutical Science and Technology

 Downloaded from journal.pda.org on May 30, 2012

Figure 2

Probability of detecting particles in a 250 mL bottle (31).

the POD for NIST-traceable spheres in 20 mL glass each container in the set. Knapp’s original validation
vials, 250 mL glass bottles, 50 mL flexible containers procedure recommended at least five (5) inspections of
(infusion bags), and 250 mL flexible containers (31). each container by multiple inspectors for a total of 70
Detection of a 163 ␮m particle was comparable to inspections to provide data with a 95% confidence
earlier studies reported by Shabushnig and Melchore, limit. Due to the probabilistic nature of particle detec-
using light intensities ranging from 225 to 375 foot- tion, statistical confidence in the POD increases with
candles (12). In contrast, Budd reported a 93% POD the number of inspections, and investigators refer-
for detecting a comparable sized particle when using a enced in this article have utilized ⬎10 inspections to
light intensity of 5500 lux (36). These studies serve as calculate the POD. At some point practicality limits
a useful guideline for conducting SVP inspection, with the number of manual inspections that can be per-
the understanding that there will be site-to-site varia- formed, and a risk assessment, conducted by a multi-
tion dependent on operational differences. Dunham’s
disciplinary team, provides a sound basis to determine
data also indicates there is little to be gained by
the number of inspections to be performed.
increasing particle size once a given particle size is
detected with approximately 90% reproducibility in 20
Statistical Subsets: The Knapp validation procedure
mL vials.
divides the challenge set into three subsets based upon
the POD, which is calculated by a minimum of 20
The data in Figure 2 taken from Dunham’s studies
inspections (9). The containers are placed into one of
conducted with a 250 mL bottle indicate that the same
sized particle detected in the 20 mL vial had a signif- three subgroups based on their POD:
icantly lower POD when seeded in a 250 mL bottle.
● The Accept subset consists of 170 containers that
The larger inspection window and increased depth of
field of the 250 mL bottle affects the apparent particle have a rejection probability ranging from 0 ⬍ 0.3.
size to the manual inspector, which is evidenced by a It represents ⬇68% of the challenge set.
decreased POD. Automated inspection machines are
● The Grey Zone subset consists of 40 containers
designed to inspect multiple sized containers, and
container size has less impact on automated systems. that have a rejection probability ranging from
0.3 ⬍ 0.7. It represents ⬇16% of the challenge set.
POD Calculation: The challenge set must be in- Grey zone containers are not true rejects, but are
spected multiple times to develop a statistical POD for containers that are sensitive to subtle changes in

Vol. 66, No. 3, May–June 2012 277

 Downloaded from journal.pda.org on May 30, 2012

the inspection process. They provide security that tive measures should be taken to prevent particles
the reject containers are detected. from adhering to interior surfaces. Other investigators
referenced in this article stress the importance of seed-
● The Reject subset consists of 40 containers that ing each container with only one particle in order to
have a rejection probability ranging from 0.7 ⱕ determine if a specific particle type and size was
1.0. It represents ⬇16% of the challenge set. detected (11, 14, 26, 36).

With effective particulate control procedures in place, Containers may be seeded with intrinsic particles
inspectors routinely experience rejection rates ranging found in the manufacturing process, extrinsic particles
from 1 to 2% during inspection of commercial drug that enter from the environment, or simulated particles
lots. Great care must be taken when designing chal- prepared in a laboratory that react in the same manner
lenge sets to avoid positive reinforcement (sensitiza- as natural particles during spinning and manual in-
tion) of human inspectors, as challenge sets typically spection. A variety of suitable materials have been
have a greater percentage of defective containers than identified for container seeding (20 –25):
with commercial lots. Knapp reported positive rein-
forcement when the Reject subset exceeded 25% of ● Fibers: fibers in challenge sets should include fi-
the challenge set (8). Many studies reported in this bers from materials that are found in the produc-
article have taken a more conservative approach and tion process. They can be either natural or syn-
kept the reject subset between 20 –25% of the total thetic. Commonly found materials are cotton,
containers in the challenge set, for studies involving rayon, nylon, polyester, filters (polypropylene,
human visual inspectors (11, 12, 14, 15, 19, 20, 22). Teflon威 fluoropolymer resin, Teflon威 fluorinated
Contrary to urban myth, automated inspection systems ethylene propylene), wipers, garments, and others.
are not subject to sensitization. The proportion of
rejects in the challenge set or the order in which the ● Glass shards or glass lamella of the specific prod-
containers are inspected does not influence the results uct/configuration.
with an automated inspection system because these
systems are not subjective. ● Rubber fragments from stoppers; diaphragm
pumps or “O” rings.
Preproduction Functionality Test Automated Sys-
tems: Typically, companies have a procedure to verify ● Stainless steel shards from manufacturing or filling
and document automated system settings prior to in- equipment.
spection of commercial product. Many companies also
perform a preproduction inspection of a challenge set Challenge Sets Made from NIST-Traceable Spheres
that contains sufficient acceptable and reject contain- versus Real Particles
ers to ensure that each station of the automated system
is functional. This type of challenge set may consist of Challenge sets may also be prepared by seeding con-
a portion or the entire validation challenge set, and it tainers with NIST traceable spheres in lieu of actual
provides added security of machine function (20). particles. Standard spheres are manufactured from
glass, polystyrene latex, stainless steel, or various
Seeding Containers: Independent studies performed resins of uniform shape and size. Standard spheres
by Borchert, Shnek, and Gidh compared the effect on provide a quantitative measurement of size due to their
POD when containers were seeded with one versus uniform shape and consistency and have limited vari-
multiple NIST spheres (10, 27, 28). Their independent ability compared to real particles. Unlike real parti-
findings confirm that seeding containers with more cles, they do not break down with repeated spinning
than one particle will result in an artificially increased and are suitable for extensive use during factory ac-
POD. This is due to a multiplying effect of two smaller ceptance testing, site acceptance testing, and support-
particles in proximity being detected as one large ive engineering studies. Various sized spheres enable
particle, and the identity of the particle that was de- the plotting of a calibration curve, which provides a
tected will be unknown. The practice of adding more direct correlation between POD and specific particle
than one sphere was started to compensate for spheres sizes (15, 34 –37). The calibration curve is useful for
that went undetected due to particles adhering to the comparing inspection methods, inspector groups, or
interior syringe surfaces. In such cases, other correc- evaluating changes to the inspection process. In addi-

278 PDA Journal of Pharmaceutical Science and Technology

 Downloaded from journal.pda.org on May 30, 2012

Figure 3

Detection of glass shards versus glass spheres (32).

tion, spheres are suitable for long-term usage because One of these studies revealed a common problem with
they do not breakdown with repeated spinning like syringe inspection. Two of 10 syringes, all seeded
some natural particles. with a standard sphere, went undetected in 50 inspec-
tions, while the other eight syringes in the same group
McCormick and Bariexca evaluated particle detection were detected in all 50 inspections. It was determined
with vials and syringes seeded with Type I USP boro- that the sphere was hidden in the syringe tip cap on
silicate glass shards or NIST-traceable spheres of var- these two syringes (a new technique is now used to
ied sizes (31). The data in Figure 3 indicates that glass plug the needle path). It is important to inspect each
shards were more readily detected than comparably syringe after particle seeding to ensure that the sphere
sized standard spheres when inspected on the Eisai or particle remains visible during inspection. Once this
Automatic Inspection Machine (AIM). The dynamics problem was realized, creative methods have been
of putting a sphere into motion are very different than used to prevent particles from going undetected.
with natural or simulated particles that have edges.
Also, standard spheres may sink quickly after spin- Matrix Grouping for Qualification/Validation of
ning, which provides less detection time. Multiple Products

Although standard spheres have many desirable char- The creation of an individual challenge set to qualify
acteristics, they have minimal utility for training and or validate the inspection of every drug product is time
testing manual inspectors. The physical appearance of consuming, costly, and may be unnecessary. The ma-
standard spheres does not prepare manual inspectors trix approach places product/container combinations
for detection of real particulate matter found in com- into groups that have comparable physical properties.
mercial product. Also, the dynamics of putting a The matrix approach can reduce the number of chal-
sphere into motion and its behavior after spinning are lenge sets required for validation of multiple products.
very different than with irregularly shaped particles. One representative drug product is used to validate
These differences affect both manual and automated each drug product/container group; however, this ap-
inspection. Optimized settings for automated inspec- proach is only possible when the representative drug
tion of commercial product can only be achieved with product has the same physical properties as the other
real particulate matter found in product (17). drug products in this class. The key requirements for

Vol. 66, No. 3, May–June 2012 279

 Downloaded from journal.pda.org on May 30, 2012

products placed in the same group are that a suspended the challenge container. If the container is to be par-
particle must react in the same manner during spin- ticle-free it is immediately sealed after filling. Filter-
ning, braking, and inspection. An example of a prod- ing is a desirable step in preparing challenge sets filled
uct grouping would consist of three aqueous products with biopharmaceutical products. However, filtering
that have the same physical appearance, viscosity, some large-molecule formulations can remove smaller
container, fill volume, and inspection characteristics. particles but may cause aggregation or result in parti-
If all attributes are comparable, one drug can be used cles that may interact to form larger particles. Another
to validate the inspection of the drug product group in alternative for making challenge sets with biopharma-
the matrix. ceuticals is to use a placebo solution. The key to
success is to select a placebo that looks like and
Before considering the details of any proposed behaves in the same manner during spinning and in-
challenge set, the purpose of the challenge set must spection as the actual product.
be described and the specifications for the contain-
ers must be determined. This is best accomplished by Reject containers are seeded with one particle or
a risk analysis of the key issues and concerns (risks) sphere of a known composition and size before being
for each segment of the inspection process. The out- sealed under aseptic conditions. After sealing, the
come of the risk analysis should lead to design qual- presence of the particle in the container is verified and
ification or a user requirement specification to define the seeded containers are randomly distributed
the uses of the challenge set, as well as the particle throughout the challenge set. All containers receive a
sizes and types to be used. When all these consider-
permanent unique number, with the container contents
ations have been identified, defined, and evaluated, the
recorded in an inventory list. The advantage of using
particle types, sizes, and characteristics can be speci-
this type challenge set is that the composition and size
fied for inclusion into the challenge set.
of each particle in the containers is known, verified,
and documented.
Preparation of Specific Challenge Sets
Manufacturing a precision challenge set is time-con-
Certified Standard Challenge Set: This type of
suming and tedious, and these precautions are required
challenge set is designed for frequent handling over
to create a challenge set suitable for inspector quali-
long time periods and is expected to have a stable
fication over a number of years. Once received, proper
shelf life from 3 to 5 years. It provides the greatest
storage of the challenge set is required for maximum
amount of information regarding the inspection pro-
shelf life.
cess because each particle size and type is known,
which enables a more exact comparison of particle
size and POD. Certified standards are suitable for Characterized Challenge Set: Containers may be
qualification/requalification of manual inspectors, filled with diluent or bulk product in the laboratory,
establishing a manual performance baseline or val- or with the contents of containers taken from the
idation of automated systems for inspection of aque- filling line. As with the certified standard set, only
ous products. containers free of visible particles are selected for
further work. Because the solution was not filtered
A compatible container/closure is selected and all prior to filling, the subvisible enumeration and sizes
components are washed and sterilized prior to use. of the particles are unknown. Freedom of visible
Sterilized components are meticulously rinsed with particles must be verified by multiple inspections to
WFI under a high-efficiency particulate air (HEPA) determine the POD for each container. Only con-
filter hood, using aseptic technique. The rinsed con- tainers free of visible particles can be aseptically
tainers are filled with WFI (with or without preserva- seeded with a particle of known composition and
tive), a proprietary formulation having bacteriostatic size in the laboratory. As with the certified standard
properties, or a stable bulk drug product. The liquid is set, knowing the particle type and size enables a
filtered prior to filling the containers to minimize the direct correlation between particle size and POD,
chance of any subvisible particle ⬎25 ␮m from enter- but the effect of subvisible particles on POD is
ing the container, as subvisible particles may attach to unknown. The characterized challenge set is useful
each other or be in close proximity at the time of for nonbiologics or formulations known to have few
inspection and interfere with the intended contents of subvisible particles.

280 PDA Journal of Pharmaceutical Science and Technology

 Downloaded from journal.pda.org on May 30, 2012

Challenge Sets Using Production Line Containers placebo solution offers a safe alternative for labora-
(Knapp Boot Strap Method): This challenge set uti- tory personnel preparing containers with toxic or po-
lizes containers filled with drug product taken directly tent drugs. It is critical that the placebo formulation
from the filling line/inspection area to serve as the has the same physical properties of the finished prod-
basis for a challenge set. The incorporation of con- uct and that particles suspended in the placebo solu-
tainers taken directly from the inspection booth or tion must react in the same manner as the actual drug
filling line can not be directly placed into a challenge product, from the time the placebo solution is manu-
set unless they are inspected multiple times to provide factured until the inspection study is completed. As
a POD. The use of containers removed from the pro- with the Knapp Boot Strap Method, containers taken
duction line becomes less of an issue as the number of directly from the filling line and inspection booth must
inspections is increased, as there is greater confidence be carefully examined under optimized conditions be-
in the POD. This method has had mixed success with fore they are placed into a challenge set.
some products due to small particles in the formula-
tion, which have the potential to come into contact General Comments Regarding Challenge Sets:
with each other or be in close proximity during in- Each time this challenge set is used, the current test
spection. The POD is based on multiple inspections, data should be compared with historical data in order
without knowing the particle composition or size. to determine if the current classification of a container
Qualification of individual inspectors compared to the is consistent with historical data. During this review,
inspector pool is possible with this type of challenge any container that has a significant change in classifi-
set, as the POD of the individual inspector is compared cation should be closely examined, characterized, or
to the POD of the inspector pool. The manual inspec- removed from the set. As noted by Leversee and
tion baseline may also be used as a basis to validate an Shabushnig, challenge sets should be inspected peri-
automated inspection system. The disadvantages of odically for microbial growth, change in product ap-
this approach include not being able to correlate POD pearance, or damage to the container/closure (20). A
directly to particle size. In addition, the composition written standard operating procedure, specifying a re-
of some particles may be unknown and the potential quired container examination review period, with doc-
effect of subvisible particles on the POD will be umented findings, provides a means to ensure that the
unknown. challenge set has consistent performance.

A great deal of time can be saved if an experienced When constructing a challenge set, it is prudent to
technician or inspector examines the containers taken have additional containers prepared for each subset to
from the production line and examines each container replace broken or unusable containers through re-
carefully before creating a challenge set to be in- peated use. This is particularly important for rare or
spected multiple times by the inspector pool. Using specific in-house manufacturing particles that occur
increased light intensity, longer inspection duration, less frequently. Challenge sets must be stored in a
and a useful device such as the Eisai APK, which secure and controlled environment.
spins, brakes, and enables inspection of the container
with enhanced lighting. These added steps can facili- Containers having any unidentified particles should be
tate improved selection of containers for the challenge sent to a qualified laboratory for analysis. Understand-
set. ing the composition of each particle in the challenge
set provides a better understanding of the inspection
Challenge Sets Containing Toxic or Potent Com- process and can be used to determine the root cause of
pounds: Custom-filling each visual inspection stan- particle generation.
dard in the laboratory with a drug product can only be
performed when the product is not harmful to humans, Summary
as the steps involved in filling and seeding several
containers results in prolonged exposure for the mi- Inspection of filled drug product containers serves as a
croscopist. Drug exposure can be minimized by using monitor of particulate contribution from upstream
drug-filled containers taken from the production line manufacturing processes. Inspection does not add
as described under uncharacterized challenge sets or quality to the product, but it does provide a checkpoint
by the use of a placebo solution that mimics the where nonconforming containers can be detected and
physical properties of the drug product. The use of a rejected from the lot. As the industry strives toward

Vol. 66, No. 3, May–June 2012 281

 Downloaded from journal.pda.org on May 30, 2012

manufacturing excellence, technological improve- ticle detection is affected by many variables. In the
ments such as barrier technology will reduce the par- past, visible inspection and subvisible particle compli-
ticulate load in product. The latest PDA survey indi- ance have been considered separately, but recent stud-
cates that many manufacturing operations have ies concerning particles in biotech and other products
achieved particulate rejection rates between 1 to 2% indicate that both regions (all particle sizes) must be
and that exceptional operations having particulate re- considered collectively. The present acceptance crite-
ject rates ⱕ1% (20). ria can be further defined for each product by estab-
lishing in-house performance criteria through the use
Although this article is concerned with challenge sets, of challenge sets and can bring different products and
issues with the absence of a harmonized methodology methods closer to harmonization. The authors hope
for visual inspection need to be evaluated. A harmo- that this article has clarified questions concerning
nized methodology for subvisual particulates exists,
challenge sets and that we have emphasized the im-
but there is great disparity concerning methodology
portant role that they play in the inspection process.
and acceptance criteria for “visible particles.” The
major compendia is based on the detection of large
Conflict of Interest Declaration
particles using visual inspection. However, none of the
compendia have a requirement for inspector visual
The authors declare that they have no competing in-
acuity. Data published by Knapp indicates that the
terests.
difference in particle size detection between 20/20 and
20/30 visual acuity is the ability to see a 58.2 ␮m
particle versus a 87.3 ␮m particle (35). Requirement References
for 20/20 visual acuity is essential for visual inspec-
tion (natural or corrected). PDA surveys verify that 1. USP (USP) Particulate Matter in Injections, 具788典;
most companies have adapted this requirement (20). Particulate Matter in Ophthalmic Solutions 具789典;
The authors recommend that 20/20 visual acuity be- Methods for the Determination of Particulate Mat-
come a universal requirement. We base this recom- ter in Injections and Ophthalmic Solutions,
mendation on numerous studies, industry practice, and 具1788典, USP 34-NF 29, 2011.
optical data concerning contrast thresholds of the hu-
man eye that were published as far back by Blackwell 2. European Pharmacopeia (EP): Particulate Con-
in 1946 (38). tamination: Sub-visible Particles; Method 2.9.20,
6.0, 1, 302, 2011.
Differing opinions concerning light source, light in-
tensity, and methodology are likely to continue. In
3. British Pharmacopeia (BP). Appendix XIII B.
addition, the variability of products in the market
Particulate Contamination: Visible Particles; Ph.
requires different inspection techniques, and this has
Eur. Method 2.9.20, I, A322, 2011.
hindered the creation of a harmonized inspection
method. The guidance provided by the EP has the
4. Japanese Pharmacopoeia (JP). Insoluble Particu-
essential elements of visual inspection, and the light
late Matter Test for Injections: The Ministry of
intensity specified in the EP has been used in most of
the referenced studies. The only missing elements in Health, Labor and Welfare Ministerial Notifica-
this guidance are a specified time for inspection and tion No. 285, 15, 6.07, 110, 2007.
technique. If Knapp’s recommendation for a 5 s in-
spection in front of each color background is added, 5. Guidance for Industry. Container Closure Systems
the EP guidance could serve as a reasonable interim for Packaging, Human Drugs and Biologics,
document until there is a harmonized method, and this Chemistry, Manufacturing and Controls Docu-
might reduce differences in product acceptance in the mentation, FDA, May 1999.
global marketplace.
6. Technical Report No. 43. Identification and Clas-
Regulatory agencies and industry want to see a qual- sification of Nonconformities in Molded and Tu-
itative number replace the vague acceptance criteria of bular Glass Containers for Pharmaceutical Manu-
“essentially free” from visible inspection. However, facturing. PDA J. Pharm. Sci. Technol. 2007, 61,
the studies referenced in this article indicate that par- 1–31.

282 PDA Journal of Pharmaceutical Science and Technology

 Downloaded from journal.pda.org on May 30, 2012

7. Technical Report No. 27. Pharmaceutical Package 17. Melchore, J. A. Prerequisites for optimized per-
Integrity Identification and Classification of Non- formance of the Eisai 1088W Automated Inspec-
conformities in Molded and Tubular Glass Con- tion System. PDA J. Pharm. Sci. Technol. 2010,
tainers for Pharmaceutical Manufacturing. PDA 64 (6), 574 –580.
J. Pharm. Sci. Technol. 1998, 52, 1– 48.
18. Knapp, J. Z.; Abramson, L. R. Automated par-
8. Knapp, J. Z.; Kushner, H. K. Generalized meth- ticulate inspection systems: strategies and im-
odology for evaluation of parenteral inspection plications. J. Parent. Sci. Technol. 1990, 44,
procedures. J. Parent. Drug Assoc. 1980, 34 (1), 74 –107.
14 – 61.
19. Borchert, S. J.; Maxwell, R. J.; Davidson, R. L.;
9. Cherris, R. Visual Inspection Lifecycle; Particu- Aldrich, D. S. Standard particulate sets for visual
late and Container/Closure Defects. Paper pre- inspection systems: their preparation, evaluation
sented at the PDA Visual Inspection Forum, and applications. J. Parent. Sci. Technol. 1986, 40
Bethesda, MD, 2011. (6), 265.

10. Borchert, S. J.; Abe, A.; Aldrich, S.; Fox, L. E.; 20. Leversee, R. L.; Shabushnig, J. G. A Survey of
Freeman, J. G.; White, R. W.; Particulate matter Industry Practice for the Visual Inspection of In-
in parenteral products: a review. J. Parent. Sci. jectable Products (PDA Preliminary Report). Pa-
Technol. 1986, 40 (5), 212–241. per presented at the PDA Visual Inspection Fo-
rum, Berlin, Germany, 2008.
11. Melchore, J. A. Sound practices for consistent
human visual inspection. AAPS PharmSciTech 21. Leversee, R. L., Shabushnig, J. G. The Use of
2011, 12 (1), 215–221. Standard Particles in the Qualification, Valida-
tion, and Routine Control of Visual Inspection
12. Shabushnig, J. G.; Melchore, J. G.; Geiger, M.; Operations. Paper presented at the PDA Visual
Chrai, S.; Gerger, M. E. A Proposed Working Inspection Forum, Bethesda, MD, 2009.
Standard for the Validation of Particulate In-
spection in Sterile Solutions. Paper presented at 22. Berdovich, D. Visual Inspection Standards. Paper
the PDA Annual Meeting, Philadelphia, PA, presented at the PDA Visual Inspection Forum,
1994. Bethesda, MD, 2007.

13. Melchore, J. A. Establishing a Statistically Sound 23. Berdovich, D. Visual Inspection Standards. Paper
Manual Inspection Baseline. Paper presented at presented at the PDA Visual Inspection Forum,
the PDA Visual Inspection Forum, Bethesda, MD, Bethesda, MD, 2005.
2009.
24. Aldrich, S., Particulate Matter—Visual Detection
14. Melchore, J. A.; Berdovich, D. Design and Use of to Identification. Paper presented at the PDA Vi-
Challenge Sets for Inspection Studies. Paper pre- sual Inspection Forum, Bethesda, MD, 2009.
sented at the PDA Visual Inspection Forum,
Bethesda, MD, 2011. 25. Cherris, R. T. Visible Particle Inspection, Micros-
copy, and Characterization—Tools Supporting
15. Knapp, J. Z. The Bridge between Visible Particle Visual Inspection Acceptance Criteria. Paper pre-
Data and the Accuracy and Repeatability Require- sented at the PDA Visual Inspection Forum,
ments of PAT, One Safe Step at a Time. Paper Bethesda, MD, 2005.
presented at the PDA Visual Inspection Forum,
Bethesda, MD, 2005. 26. Rathore, N.; Chen, C.; Gonzalez, O.; Wenchang,
J. Challenges and strategies for implementing au-
16. Knapp, J. Z. The scientific basis for visible parti- tomated visual inspection for biopharmaceuticals.
cle inspection. PDA J. Pharm. Sci. Technol. 1999, Pharmaceutical Technology, Injectable Drug De-
53 (6), 291–302. livery Supplement 2009, 33, S25–S29.

Vol. 66, No. 3, May–June 2012 283

 Downloaded from journal.pda.org on May 30, 2012

27. Singh, S. K., Rathore, N.; McAuley, A., Rathore, sented at the PDA Visual Inspection Forum, Ber-
A. S. Best practices for formulation and manufac- lin, Germany, 2010.
turing of biotech drug products. BioPharm Int.
2009, 22, 32– 48. 33. Grieb, E. Scrutinizing the Sub-Visible, Pharm.
Technol. 2010, 35, 44 – 48.
28. Shnek, D, Carrion-Martinez, M., Gigh, A., Se-
garra, E., Bruno, M, Aquino, A., Paranandi, M. 34. Berdovich, D. USP 具788典, USP 具789典 Issues in
Comparing Manual and Automated Inspection of Measuring Particles. Paper presented at the USP
Seeded Syringe Defects. Paper presented at the Workshop on Particles, Rockville, MD, Dec.
PDA Visual Inspection Forum, Bethesda, MD, 8 –10, 2010.
2007.
35. Knapp, J. Z. Overview of Accurate NIST Sized
29. Gidh, A., Carrion-Martinez, M., Cai, X., Dimensions of Visible Particle Contamination
Paranandi, M., Shnek, D. Particle Standard Sets Data in Sealed Containers of Injectable Products:
for Automated Syringe Inspection. Paper pre- An Essential Validation and Product Acceptance
sented at the PDA Visual Inspection Forum, Tool. Paper presented at the PDA Visual Inspec-
Bethesda MD, 2007. tion Forum, Bethesda, MD, 2007.

30. Deng, S.; Shnek, D., Carrion-Martinez, M., Dom- 36. Budd, G. W. Referee Level NIST Traceable Par-
penciel, R., Baez, L., Sullivan, J., Sing, N., Ruhl, ticle Contaminating Particle Standards in Sealed
S. Evaluation of Automated Visual Inspection Containers Used for the Creation of a Visible
(AVI) Technologies for Biologic Products. Paper Calibration Curve. Paper presented at the PDA
presented at the PDA Annual Meeting, Las Vegas, Visible Inspection Forum, Bethesda, MD, 2009.
NV, 2009.
37. Budd, G. W. Optimizing the Correlation of Auto-
31. Dunham, A. Visible Particulate Matter Detection mated Inspection Data Obtained from Electronic
in LVIs. Paper Presented at the PDA Annual Sensors to (NIST Traceable) Referee Level Parti-
Meeting, Colorado Springs, CO, April 2008. cle Standards. Paper presented at the PDA Visual
Inspection Forum, Bethesda, MD, 2009.
32. McCormick, S.; Bariexca, S. Glass Sphere vs.
Glass Shard Particle Detection Equivalency Using 38. Blackwell, H. R. Contrast thresholds of the human
Automatic Static Division Inspection. Paper pre- eye. J. Opt. Soc. Am. 1946, 36 (11), 624 – 641.

284 PDA Journal of Pharmaceutical Science and Technology

 Downloaded from journal.pda.org on May 30, 2012

An Authorized User of the electronic PDA Journal of Pharmaceutical Science and

Technology (the PDA Journal) is a PDA Member in good standing. Authorized Users are
permitted to do the following:

·Search and view the content of the PDA Journal

·Download a single article for the individual use of an Authorized User
·Assemble and distribute links that point to the PDA Journal
·Print individual articles from the PDA Journal for the individual use of an Authorized User
·Make a reasonable number of photocopies of a printed article for the individual use of an
Authorized User or for the use by or distribution to other Authorized Users

Authorized Users are not permitted to do the following:

·Except as mentioned above, allow anyone other than an Authorized User to use or access the
PDA Journal
· Display or otherwise make any information from the PDA Journal available to anyone other
than an Authorized User
·Post articles from the PDA Journal on Web sites, either available on the Internet or an Intranet,
or in any form of online publications
·Transmit electronically, via e-mail or any other file transfer protocols, any portion of the PDA
Journal
·Create a searchable archive of any portion of the PDA Journal
·Use robots or intelligent agents to access, search and/or systematically download any portion
of the PDA Journal
·Sell, re-sell, rent, lease, license, sublicense, assign or otherwise transfer the use of the PDA
Journal or its content
·Use or copy the PDA Journal for document delivery, fee-for-service use, or bulk reproduction or
distribution of materials in any form, or any substantially similar commercial purpose
·Alter, modify, repackage or adapt any portion of the PDA Journal
·Make any edits or derivative works with respect to any portion of the PDA Journal including any
text or graphics
·Delete or remove in any form or format, including on a printed article or photocopy, any
copyright information or notice contained in the PDA Journal

View publication stats