a biopharmaceutical industry perspective on the control of ... · commentary a biopharmaceutical...

10.5731/pdajpst.2015.006189Access the most recent version at doi: 392-40870, 2016 PDA J Pharm Sci and Tech

Serge Mathonet, Hanns-Christian Mahler, Stefan T. Esswein, et al. ProductsVisible Particles in Biotechnology-Derived Injectable Drug A Biopharmaceutical Industry Perspective on the Control of

on July 18, 2016journal.pda.orgDownloaded from on July 18, 2016journal.pda.orgDownloaded from

http://journal.pda.org/lookup/doi/10.5731/pdajpst.2015.006189

http://journal.pda.org/


COMMENTARY

A Biopharmaceutical Industry Perspective on the Control ofVisible Particles in Biotechnology-Derived Injectable DrugProductsSERGE MATHONET1,*, HANNS-CHRISTIAN MAHLER2, STEFAN T. ESSWEIN3, MARYAM MAZAHERI4,PATRICIA W. CASH4, KLAUS WUCHNER5, GEORG KALLMEYER6, TAPAN K. DAS7, CHRISTOF FINKLER8,and ANDREW LENNARD9

1Global Regulatory Affairs—Biologics CMC, Sanofi R&D, 91385 Chilly-Mazarin, France; 2Drug Product Services,Lonza AG, 4002 Basel, Switzerland; 3NBE Analytical R&D, Abbvie Deutschland GmbH & Co. KG, 67061Ludwigshafen, Germany; 4Analytical Biotechnology, MedImmune, Gaithersburg, MD; 5PDMS Analytical Development,Janssen R&D, 8205 Schaffhausen, Switzerland; 6Quality Combination Products, Roche, 68305 Mannheim, Germany;7Biologics Development, Bristol Myers Squibb, Hopewell, NJ; 8Analytical Development & Quality Control, PharmaTechnical Development Biologics EU, Roche, 4002 Basel, Switzerland; and 9Regional Regulatory Affairs CMC,Amgen Ltd, Uxbridge, UK ©PDA, Inc. 2016

ABSTRACT: Regulatory monographs in Europe and the United States require drug products for parenteral administrationto be “practically free” or “essentially free” of visible particles, respectively. Both terms have been used interchange-ably and acknowledge the probabilistic nature of visual particle inspection. The probability of seeing a particle in adrug product container varies according to the size and nature of the particles as well as container and inspectionconditions. Therefore, the term “without visible particles” can be highly misleading in the context of what ispractically achievable. This may lead to differences in understanding between industry practitioners and regulatoryagencies. Is this term intended to mean “zero particles”, or is there any intention to distinguish between particle typesuch as “zero extraneous visible particles” or “zero proteinaceous particles”? Furthermore, how can “zero” particlesas a criterion for release testing be reconciled with “practically free from particles” as stated in the definition and alow, justified level of proteinaceous particles after production?The purpose of this position paper is to review best practices in the industry in terms of visual inspection process andassociated operator training, quality control sampling, testing, and setting acceptance criteria corresponding to“practically free of visible particles” and providing considerations when visible proteinaceous particles are deemedunavoidable. It also provides a brief overview of visible particle characterization and gives perspectives on patientsafety. This position paper applies to biotechnology-derived drug products including monoclonal antibodies inlate-phase development to licensed products.

KEYWORDS: 100% inspection, Lyophilized, Quality control sampling, QC sampling, Particle identification,proteinaceous particles, visible particles.

LAY ABSTRACT: In the 2011 monoclonal antibody monograph revision, European Pharmacopoeia experts acknowledgedthat protein products may also contain proteinaceous particles at release or that protein particles may form during storage.Indeed, industry experience has demonstrated that therapeutic proteins such as monoclonal antibodies can exhibit apropensity for self-association leading to the formation of aggregates that range in size from nanometres (oligomers) tomicrons (subvisible and visible particles). As a result, the requirement for drug product appearance for monoclonalantibodies was changed from “without visible particles” to “without visible particles unless otherwise authorised orjustified”. In our view, “practically free from particles” should be considered a suitable acceptance criterion for injectablebiotechnology and small-molecule products, as long as appropriately defined. Furthermore, we argue that visual inspection is asuitable quality control release test and that “practically free from particles” is a suitable specification when adequately described.

* Corresponding Author: e-mail: [email protected]; telephone: !33 6 73 23 26 85.

doi: 10.5731/pdajpst.2015.006189

392 PDA Journal of Pharmaceutical Science and Technology

on July 18, 2016journal.pda.orgDownloaded from


1. Problem Statement

The European Pharmacopoeia (EP) and United StatesPharmacopeia (USP) monographs for parenteral prep-arations require drug products for parenteral adminis-tration to be “practically free” or “essentially free” ofvisible particles, respectively. Both terms have beenused interchangeably. These definitions acknowledgethe probabilistic nature of visual particle inspection.The EP monograph 2031, Monoclonal Antibodies forHuman Use, as revised in 2011 set the followingexpectations and requirements for visible particles inmonoclonal antibody drug products, as noted below inexcerpts from different sections in the monograph:

Definition: “Examined under suitable conditions ofvisibility, they are practically free from particles.”

Production: “As part of the in-process control eachcontainer (vial, syringe or ampoule) is inspected afterfilling to eliminate containers that contain visible par-ticles. During development of the product it must bedemonstrated that either the process will not generatevisible proteinaceous particles in the final lot or suchparticles are reduced to a low level as justified andauthorised.”

Tests—Appearance: “They are without visible parti-cles, unless otherwise justified and authorised.”

The question remains unresolved for a clear meaningof the term “without visible particles”; this may leadto differences in understanding between industry andagencies. Is this term intended to mean “zero parti-cles”, or is there any intention to distinguish betweenparticle type such as “zero extraneous visible parti-cles” or “zero proteinaceous particles”? Furthermore,how can “zero” particles as a criterion for releasetesting be reconciled with “practically free from par-ticles” as stated in the definition and with a low,justified level of particles after production? Is zeroparticles a consistent, realistic outcome?

PHARMEUROPA Vol. 23, No. 3, July 2011 providedexplanatory notes and summary of changes to themonograph 2031 to be published in Supplement 7.3.

The merit of the 2011 monoclonal antibody mono-graph revision was that EP experts acknowledged thatprotein products may also contain proteinaceous par-ticles at release or that protein particles may formduring storage. Indeed, industry experience has dem-

onstrated that therapeutic proteins such as monoclonalantibodies can exhibit a propensity for self-associa-tion, leading to the formation of aggregates that rangein size from nanometres (oligomers) to microns (sub-visible and visible particles). As a result, the require-ment for drug product appearance for monoclonalantibodies was changed from “without visible parti-cles” to “without visible particles unless otherwiseauthorised or justified”.

It was stated in the explanatory notes that “withoutvisible particles” was intentionally kept to give clearguidance that the presence of visible particles is un-wanted and the appropriate formulation studies shouldbe performed during development to minimize visibleparticle formation; practically free could not be apass/fail criteria in a test, and that visual inspection isnot a quality control (QC) test, even though performedat the end of the production.

As a result of the probabilistic nature (1) of detectingparticles by visual inspection method, “without visibleparticles”—meaning zero visible particles—is an un-realistic requirement for QC release/shelf life testingof any parenteral product and especially those of bio-technological origin. Even with significant formula-tion and container development, supported by long-term stability studies and stress stability studies (e.g.,agitation), the probability of a visible particle beingpresent cannot be completely eliminated. Interest-ingly, a requirement of without (zero) visible particlesis not aligned with the requirement for small-moleculeparenterals, which states “essentially/practically free”of visible particles.

In our view, “practically free from particles” should beconsidered a suitable acceptance criterion for inject-able biotechnology and small-molecule products, aslong as appropriately defined. Furthermore, we arguethat visual inspection is a suitable QC release test andthat “practically free from particles” is a suitable spec-ification when adequately described.

The purpose of this position paper is to review bestpractices in the industry in terms of visual inspectionprocess and associated operator training, QC sam-pling, testing and setting acceptance criteria corre-sponding to “practically free of visible particles”, andproviding considerations when visible proteinaceousparticles are deemed unavoidable. It also provides abrief overview of visible particle characterization, andperspectives on patient safety.

393Vol. 70, No. 4, July–August 2016



This position paper applies to biotechnology-deriveddrug products including monoclonal antibodies in late-phase development to licensed products.

2. Visual Inspection at End of Drug ProductManufacturing—100% Inspection Followed bySampling Inspection

The presence of particles in parenteral finished drugproducts is dependent on the manufacturing processand manufacturing environment (design, qualification,validation, execution) as well as post-production han-dling, storage conditions, transportation, and handlingby end users. This includes the choice and processingof primary packaging components, and also the designand stability of the formulation, particularly for bio-technology products. All products intended for paren-teral administration must be visually inspected forvarious (critical, major, and minor) defects, includingthe presence of visible particles as required by thepharmacopoeia and current good manufacturing prac-tice (cGMP).

The drug product manufacturing processes and asso-ciated controls are designed to yield, for example,single- or multiple-dose vials, syringes, cartridges, orampoules containing solution or lyophilised solid of aformulated drug product that are, to the extent possi-ble, free from visible particles.

As a last unit step (mandatory 100% inspection) dur-ing the post-filling operation, every unit of filled con-tainers is visually inspected for critical, major, andminor defects including the presence of visible parti-cles. For products with specification of “practicallyfree from visible particles”, any filled unit found witha critical or major defect is rejected. Visible particlesare classified as at least a major defect; refer to Sectionto 3.3 for exceptions.

Visual inspection is conducted according to USP chap-ters "1#, "790#, and European Pharmacopeia2.9.20 using manual, semi-automated, or automatedequipment. Of note, inspection conditions defined inEP/USP monographs are at least 2000 –3750 lux, blackand white backgrounds, and at least 5 s viewingagainst black and white backgrounds.

Detection of visible particles is a statistical processand, in routine practice, the detection limit depends onthe individual operator, the inspection system, theinspection time, properties of the particle such as

morphology and number, properties of the solutionsuch as viscosity rheology, opalescence, and refractiveindex of the solutions, as well as other factors includ-ing the primary container (e.g., container size andshape, fill volume).

There is no single size cut-off for a particle beingvisible to the human eye. The detection limit forparticles in a product is proportionate to the probabil-ity of detection such that the detectable size rangevaries from less than 50 $m, at near 1% probability ofdetection, to 200 $m, for #95% probability of detec-tion under experimental conditions. As described, theprobability of detecting a particle is influenced byseveral factors. For example, detection of 50 $mspherical polystyrene particles is at an approximately1% probability of detection (2), whereas a 100 $mparticle in 1 mL solution, in a 1 mL glass syringe, canbe detected at about 70% detection probability. Sizecut-off may be higher based on the nature of particlesand other factors. Detection limits may be specified incompany-specific procedures.

Also, companies may deviate from pharmacopoeialmethods for manual visual inspection to improve thesensitivity or ergonomics of the process, such as theuse of aids, for example, anti-glare glasses, observa-tion times, or swirling procedures; variation in lightintensity is already a flexibility provided in pharma-copoeia. These method operational parameters mayhave a pronounced effect on reject rates.

Slightly opaque or coloured container materials, forexample, opaque plastic syringes, typically require useof a higher light intensity (e.g., 8000 –10,000 lux)during 100% inspection and subsequent sampling test-ing to successfully detect visible particles. Highlycoloured containers may require transfer to a clean,transparent container for visual inspection.

Visual inspection operators are trained (Appendix 1)using several runs of defect sets that may includevarious particle types, for example, product contactmaterials (stainless steel, aluminum, glass, rubber),particle size standards and fibres, hair, and so forth.The specific design, and also the stability, of these testkits need to be described in company-specific standardoperating procedures (SOPs) and may be tailored tosuit a specific process and product.

An operator must be able to detect a predefined pro-portion of defects in the training set (not exceeding a




maximum amount of false positives) in order to bequalified (Appendix 1).

During routine 100% inspection, all units with de-tected defects are segregated, classified, and counted.All units with confirmed particle defects are discarded;refer to Section 3.3 for exceptions.

For products “practically free of visible particles”, anyunit with an observed visible particle defect irrespec-tive of its nature is discarded at 100% inspection, andthis applies to liquid and lyophilisate. Products spec-ified as “may contain proteinaceous particles”, if jus-tified, can accept protein particle at 100% inspectionwith appropriate training and certification through useof product-specific training sets and guidance materi-als, for example, photographs/video, actual examplesof the proteinaceous particles for direct comparison.

One hundred percent inspection may be manual, au-tomated, or a combination. When automated inspec-tion is used, ejected units can be manually inspected toverify the defect (e.g., presence of particulate materialfrom other particle phenomenon such as micro-bub-bles), and the detected defects are classified accordingto company procedures. Ejected units confirmed forthe presence of a defect such as visible particles wouldbe rejected; refer to Section 3.3 for exceptions.

Biopharmaceutical manufacturers are expected to haveassessed the capability of their commercial fill/finishprocess for each product (and strength) at each man-ufacturing facility. This assessment often includes100% inspection results, and sponsors often use thiscapability to set action limits, or a maximum allowablereject rate, for the visible particle defect rate in com-mercial products. If the maximum defect rate is ex-ceeded then an investigation is initiated. For clinicalproducts, a defect rate criterion may not be viablegiven the limited manufacturing history. For manufac-turing trending, bracketing and matrixing approachescan be used, for example, performing trending only forthe highest and lowest dose strength (not every singledose strength) or, for example, clustering of compa-rable product families.

The 100% inspection is unable to assure detection ofall visible particles due to the direct relationship be-tween the probability of visual detection and particlesize (as well as other parameters). Therefore addi-tional assurance is provided by a second inspectionbased on statistical sampling—for example, accept-

able quality limit (AQL), sampling based on pre-defined sampling plans, such as ANSI Z1.4, ISO2859-1 or better or equivalent. This second inspectionis performed by different inspectors to those perform-ing 100% inspection.

The additional visual inspection test based on a sam-pling plan is most commonly referred to as AQL(acceptable quality limit) testing. It might also bereferred to as an acceptance sampling plan (ASP) insome companies.

The AQL chosen for use with sampling plans can bebased on a survey of current inspection practices con-ducted by the Parenteral Drug Association (PDA) (3).This survey of both European and US-based manufac-turers established that the industry median maximumAQL for the major (Level 2) defect category is 0.65%.

AQL maximum of 0.65% has been adopted in USP"790# and is typically used as the criterion for theAQL test, unless otherwise justified (4).

Circumstances dictate when switching to a higherlevel of testing is required, which could mean anincreased sample size and lower maximum AQL for atightened sampling plan. A tighter sampling plan maybe triggered by detection of an atypical particle type,or increased particle detection frequency, suggestingprocedural issues that may require corrective and pre-ventive actions (CAPAs) and may reflect a systemfailure that presents greater risk to patients.

ANSI Z1.4 and ISO 2859-1 use AQL tables to deter-mine the appropriate sample size; however, othermethods may also be used that result in an acceptableAQL for visual inspection. Methods also exist thathave reduced dependence on batch size, for example,use of operating characteristics curves as described byKnapp and Budd in 2005 or Taylor in 1992 (5, 6).

As an illustrative example of AQL testing: Applying aLevel II ANSI/ASQ Z1.4 plan for batch sizes of10,001 to 35,000, the batch is considered to be “prac-tically free of particles” if no more than five (5) unitsare observed to contain one or more visible particles ina sample of 315 units examined under the conditionsdescribed above. Detection of six (6) units with visibleparticle(s) would fail the test. Visual inspection LevelII testing may be considered suitable for detection ofvisible particles when product development did notreveal concerns of visible particle generation. Addi-




tional investigation of particles detected during AQLtesting would not be required unless they are unex-pected, atypical particles that suggest system failure(e.g., insect parts), in which case it is recommended totrigger further investigation.

As an alternative to directly using the AQL tablesprovided in ANSI/ASQ Z1.4, operating characteristic(OC) curves may be generated for the proposed sam-pling plans with parameters adjusted to meet the de-sired AQL and unacceptable quality limit (UQL). TheOC curve then determines the sample size and accept/reject numbers. The impact of lot size on a samplingplan is minimised by fixing both AQL and UQL andensuring that the sample size (n) is less than 0.1 % N,where N is the lot size.

Batches that fail AQL criteria may be considered for100% re-inspection dependent upon the outcome of aninvestigation and on specific company procedures.

Should 100% inspection be repeated, another AQL-based inspection would need to be passed and may usetighter AQL criteria such as increased sample size.The tiers of retesting should be limited and describedin company procedures (usually not more than twocycles of re-inspection). Ultimately, repeated test fail-ure may result in batch rejection.

When 100% inspection action limits show specialcause variation (6, 7) in the process, additional AQLsamples may be required to provide additional assur-ance of quality.

When the batch size is small (e.g., less than 501 units)then reduced, special sampling plans may need to beconsidered and justified by the sponsor. The SpecialSampling Plans described in ANSI Z1.4/ISO 2859-1may also be considered.

For product with a history for formation of protein-aceous particles, units containing visible protein-aceous particles may be accepted during the inspectionprovided that these expected protein (product-related)particles can be differentiated from foreign matter bya qualified procedure. This could include specifictraining and use of guidance materials for visual iden-tification of known, expected particle types includingexamples of product particles from that manufacturingline, in connection with specific analytical testing.

The visual inspection process is depicted in Figure 1for a liquid product (liquid-filled units with integral

container closure) to be classified as practically freefrom visible particles.

Two key concepts—that is, presence of particles qual-ified as “occasional random occurrence (first concept)or representative of a system failure (second con-cept)—are introduced in this flow chart and furtherelaborated in Appendix 2:

● A batch may be passed with occasional, randomoccurrence of particles if these are within the AQLacceptance criterion.

● Any critical visible defect issues with regards tonature of particles and representing a system fail-ure (insect wing, rust, paint, hairs, etc.) will lead toan investigation/re-inspection.

If the investigation indicates the potential for sterilitybreach during aseptic processing, this type of systemfailure may lead to batch rejection.

To summarize this section, 100% inspection (whetherdone manually, semi-automated, or automated) doesnot result in 100% of units without any visible parti-cles. Although zero defects is the desired goal andshould drive continuous process improvement, it is nota workable acceptance criterion for visible particlesbecause of current capabilities in the industry forpackaging components, processing, and facility. Inaddition, protein particles may be an inherent qualityattribute for protein biologic products particularly athigh protein concentrations, despite efforts to developan optimized formulation and container system.

Even when using state-of-art technology and ap-proaches, an absolute statement that all units will be“free from visible particles” cannot be made. It is forthis reason that the USP and EP in their generalmonographs on sterile injectable products use thephrases “essentially” or “practically” free of visibleparticles.

The “practically free from visible particles” require-ment is intended to specify a particle-free productwhile considering the limitations of the visible par-ticle inspection process, thus acknowledging that alldrug product units cannot be “without” particles.Based on the above concepts, our suggestion is toimprove clarity and wording of the monoclonal an-tibody monograph in the EP by reflecting this con-cept and, for example, adopting language to reflect




that during QC, “drug product units are practicallyfree of visible particles, unless otherwise justifiedand authorised”.

3. QC Sample Testing

This section discusses best practice approaches forthe QC assessment of visible particles for productbatch release and stability testing of parenteralsproducts containing biotechnology-derived activepharmaceutical ingredients, including monoclonalantibodies. This includes liquid products, colouredor opaque containers, and drug/device combinationproducts such as prefilled syringes (PFSs), pens,autoinjectors, and delivery pumps. Sampling planrecommendations are provided for biologics when“practically free from particles” is specified. Theseprinciples may be applied to parenteral drug prod-ucts in general. Considerations are also provided for

the control of products that may contain protein-aceous product particles. Lyophilised products arediscussed in Section 4.

3.1. Products “Practically Free From Particles”

When the product is specified as “practically free fromparticles”, batch testing may be achieved using differ-ent strategies depending on the manufacturing in-pro-cess and QC batch release control procedures in place.Therefore, different approaches to QC control of vis-ible particles are considered equally applicable.

3.1.1. Post-100% Inspection, AQL Testing May Re-place End-Product Testing: For clear, liquid dosageforms in clear containers using non-destructive visualinspection, end-product release testing may take theacceptance sampling plan (AQL) result as sufficientand appropriate to confirm the specification of “prac-

Figure 1

Schematics of typical visual inspection process—liquid-filled units.




tically free from particles”. For products that requiredestructive testing for visual inspection, refer to Sec-tions 3.4 and 4. The approach of moving the specifiedcontrol of visible particles upstream to the manufac-turing in-process AQL test requires that the AQLinspectors are trained for the detection of visibleparticles, to the same level as would QC end-prod-uct release testing inspectors (Appendix 1) for par-ticle detection endpoints. Depending on the time-frame between manufacture and the AQL testing,use of post-100% inspection AQL testing to supportcompliance of a batch to its visible particle speci-fication may also be considered as a real-time re-lease testing (RTRT) alternative to end-product QCrelease testing. Conformance of RTRT to cGMP canbe confirmed at the manufacturing site inspectionand should permit relief from EU import testingwhen manufacture is performed in a non–EU mem-ber state without a Mutual Recognition Agreementin place.

3.1.2. Batch Release End-Product QC Testing:When a drug product has successfully passed 100%inspection and AQL testing during manufacture, aproduct is typically released on the basis of end-product testing in a QC environment.

As described in Section 3.1.1, the AQL result alonemay be justified as suitable for batch release with nofurther end-product release testing for visible parti-cles. Alternatively, if a successful AQL result is notbeing used for batch release, then another confirma-tory visual inspection may be performed using a pre-defined sample size, for the purpose of QC release(Section 3.1.3).

3.1.3. Sampling Sizes: A survey among the biophar-maceutical manufacturers contributing to the elabora-tion of this position paper suggested that in manycases, 10 to 20 samples have been used to inspect forQC batch release when end-product testing was re-quired.

3.1.4. Reuse of Visually Inspected Units: Samplesused for visual inspection may be reused for other QCtests (e.g., purity, protein content) if desired and jus-tified.

Re-use applies to both non-destructive or destructivetesting when the container is opened and content ma-nipulated by, for example, reconstitution or transfer toa clean container (see also Section 4. Control of Vis-

ible Particles in Lyophilized Products). However, sta-bility of these samples during visual inspection needsto be assessed, and it needs to be ensured that condi-tions the sample is exposed to during analysis, forexample, temperature cycling and sample handlingand visual testing assessment, is not affecting subse-quent analyses and results.

3.2. Use of In-Line Filters during or Prior toAdministration

In general, the presence of visible particles may bemitigated, for example, by use of a filter during intra-venous or even subcutaneous administration. This isalso acknowledged in the respective EP and USPmonographs. However, the compatibility of a productsolution and effectiveness of an in-line filter to reduceparticles should be qualified by the sponsor. Further-more, the use of in-line filters cannot be used to justifyaccepting units with any type of visible particles dur-ing 100% manufacturing inspection and AQL testing.In-line filters may be used with products that areexpected to form protein particle after adequate justi-fication (see Section 3.3 for a discussion on proteinparticles).

The end-product release specifications would, in thesecases, possibly change from “practically free of visibleparticles” to an acceptance criterion reflecting thepresence of (protein) particles once sufficient manu-facturing and stability knowledge is gained. EU andUS product information would then be consistent withthe proposed acceptance criteria (see Section 6).

3.3. Products that may contain Proteinaceous Particles

Biologics, including monoclonal antibodies, can havean inherent molecular property to self-associate, oraggregate and form proteinaceous particles despiteformulation, manufacturing process, and containerclosure development to minimise visible protein par-ticles. This propensity is a basic thermodynamic prop-erty of the molecule that cannot be totally overcome.Such particles may form over time, often exist inequilibrium, and may or may not be reversible as aresult of non-covalent interactions. Many biologictherapeutic agents require several milligrams per kil-logram of dose for efficacy, resulting in tens to hun-dreds of milligrams protein to be administered. For asubcutaneous injection this requires formulation of ahigh-protein concentration of product with appropriateformulation to achieve acceptable physical properties




such as viscosity, and to provide a stable product withminimal aggregate and particulate content (high-mo-lecular-mass species, sub-visible particles, and visibleparticles). Increasing protein concentration can pro-mote molecular interactions and hence protein particleformation.

The Notes to the EP monograph for monoclonal anti-bodies for human use (EP 2031) recognises that finalproduct “may contain proteinaceous visible particlesthat are intrinsic to the product”. If the sponsor cansuitably justify that formulation development has beensufficiently performed to “minimise the presence ofsuch visible proteinaceous particles”, then rare occur-rence of formation of such particles may be consideredan acceptable quality attribute after risk assessmentconcluding no impact to safety and efficacy.

If suitably justified (e.g., by risk assessment that in-cludes evaluation of possible patient risks, such asimmunogenicity, embolism, and other adverse ef-fects), well characterised proteinaceous particles at adefined level and frequency may be considered accept-able in the drug product, even without use of an in-linefilter (Section 5 and Appendix 2).

There can be various approaches for 100% inspection,AQL testing, and end-product batch release. For ex-ample, the level and frequency of visible protein-aceous particles is compared to the specification anddocumentation that describes the expected or allowedcharacteristics of particles for that product. This pro-cedure would allow the presence of units with proteinparticles that are similar to those expected and withina defined frequency. Any unit with a non-protein par-ticle would continue to be discarded. Should the de-fined frequency of particles be exceeded, then thenon-conformance should trigger investigation leadingto potential batch rejection. Specific training, usingproduct-specific test panels, including actual productmanufacturing line samples, and photographs/videosand so forth, may be used to qualify and validate anappropriate procedure to ensure that certified inspec-tors can differentiate proteinaceous from non-proteinparticles based on a different morphology and behav-iour, in connection with adequate analytical technol-ogy.

End-product QC release testing of a protein with ex-pected inherent protein particles may use the AQLtesting results after successful 100% inspection with-out requiring an additional, confirmatory QC test.

When a product has a specification that allows for thepresence of proteinaceous particles, a means of quan-tification to detect changes that may occur becomesmore challenging, and the instructions should be out-lined in company procedures.

To support the overall control strategy for visibleparticles and justification for the presence of low-levelprotein particles at QC release, it is recommended thatthe product’s development and manufacturing historyis adequately documented, including the potential oc-currence of visible particles during manufacturing,stability, and transportation-related handling as wellas any clinical findings and adverse events. This his-tory may derive from separate documents according tocompany practice or constitute the product pharma-ceutical development report.

The particle history would vary according to theparticle types detected. In documenting the particlehistory, it may be useful to include information onparticle detection trends, for example, frequencyacross a lot, and number of particles per container ifpossible; description and characterization (e.g., buoy-ancy, composition) data on detected particles and re-cords of actions taken; risk management plan for thepresence of any “expected” (previously detected forthe product) and “unexpected” (novel, not previouslydetected) visible particles; and support and justifica-tion for the visible particle specification. It is acknowl-edged that particle history might be documented dif-ferently depending on company-specificpharmaceutical and quality systems requirements.

3.4. Considerations for Opaque or Coloured Productsor Primary Packaging Material and Products withRestricted View

Products that are slightly opaque and products inslightly coloured or opaque primary containers mayuse higher light intensity for visual inspection forvisible particles as permitted in EP 2.9.20 and USP"790# (e.g., 8000 –10,000 lux). For AQL testing,products in coloured or opaque primary containersmay be transferred to clean, transparent containers.Also, for products that are highly opaque or coloured,dilution may be required. Extra care is required, suchas use of particle-free water and aseptic sample han-dling in a controlled environment, when transferringor diluting the product to avoid introduction of for-eign, extraneous particles. Similar sampling consider-




ations as presented above for liquid drug productshould apply.

It may not be readily feasible to perform visual in-spection of PFSs once assembled into a pen, autoin-jector, infusion pump, or patch device if the prefilledprimary container of drug cannot be easily removed,for example, the device assembly components lock thePFS inside. In this case, it is sufficient to inspect thePFS, cartridge, or other primary container type prior toassembly to demonstrate that the assembled product is“essentially free from visible particles” provided thatthe assembly process steps are shown not to introduceparticles into the drug product container closure sys-tem.

The inspection window in a device might be inade-quate for visual inspection and particle detection andtherefore a method qualification for visual inspectionmay not be possible. Destructive disassembly is oftennot recommended because the procedure risks affect-ing the container closure integrity and creating othercosmetic defects such as scratching, but it may beperformed if the procedure is suitably qualified. If theproduct may be easily disassembled for inspection andanalysis, without a possibility to affect container clo-sure integrity and the procedure is scalable, then thecontainer should be inspected at AQL testing, as de-scribed for a typical liquid product. Disassemblyshould not be required at 100% inspection. Similarprocedures are recommended for other assembleddrug/device combination products.

3.5. Visible Particle Testing on Stability

Ideally, formulation, product, and process develop-ment studies should result in “practically free fromparticles” product that remains consistent with thisacceptance criterion throughout the claimed shelf-life.

However, although undesired, particles may form overtime with slow kinetics, for example, proteinaceousparticles or particles resulting from interaction withprimary packaging (8 –11) and/or excipients and dueto shaking during transport. Also, the appearance ofprotein particles is typically not uniform across allcontainers in a batch as a function of time, and eachcontainer with protein particles will look slightly dif-ferent from another. Additionally, the probabilisticnature of visible particle detection means that it is notpossible to entirely discount the detection of pre-existing particles in stability testing when new units of

product are being inspected at each time point of thestability protocol.

The detection of pre-existing particles in stabilityshould be avoided because stability testing shouldfocus on changes in the product over time and notdiscovery of particles that would have been present attime zero (an exception is a product with expected,inherent proteinaceous matter at release). Therefore, apre-screen of stability samples may be performed toeliminate units with unexpected visible particles attime zero. If a product is justified to have protein-aceous particles at release, then that would constitute“expected” particles and be permitted in stability ac-cording to the stability specification.

In addition to the tendency of protein product toaggregate, other time-induced changes can occur thatrelate to the formulation and container closure mate-rials of construction. One such phenomenon is theformation of visible glass lamellae resembling flake-like particles. When detected during product develop-ment, these visible particles should be avoided to theextent possible by formulation, process, and productdevelopment, including potentially considering chang-ing the primary packaging washing/sterilization pro-cess, formulation and/or container, as these particleswould indicate a system failure.

The QC release visual inspection result can be used forthe time zero stability result. When the AQL resultused for release is not suitable for stability testing,then additional units are required for the time zerovisual inspection. The number of units tested at timezero and additional time points may depend on theproperties of the finished product and should be jus-tified in company procedures. Sampling sizes for QCstability testing differ across companies.

A survey among the biopharmaceutical manufacturerscontributing to the elaboration of this position papersuggested that in many cases, 10 to 20 samples havebeen inspected during QC stability testing at each timepoint for both liquids and lyophilisates.

A product with no history of particle formation mayjustify stability sampling with a lower number of unitsthan a product with a known tendency to form pro-teinaceous particles.

When the visual inspection test is not destructive, thesame samples may also be used for subsequent stabil-




ity time points tested. This must be supported bystudies that assure product would not be at increasedrisk of particle formation over time as a result ofsample handling stress during repeated inspection andthat samples can be clearly traced to avoid mix-up ofsamples over the course of the stability programme.Defective samples may be replaced to maintain thesample size. Alternatively, separate units can be in-spected for each time point, taking into considerationthat unexpected, extraneous, visible particles may befound, as 100% and AQL inspection operations areprobabilistic and single units with rare and randomlyoccurring particles may occur.

Stability samples used for visual inspection may bereused for other batch control tests or re-inspected atfurther time points, if it is demonstrated that the visualinspection does not affect other QC test parameters.

Semi-quantitative methods such as visual comparisonagainst barium sulphate precipitate (12) or other stan-dards and instrument-based methods are in develop-ment to evaluate the levels of inherent particles in aproduct. If these can be validated (13), such methodscould possibly be introduced into release/stabilityproduct testing.

Given sufficient product understanding and history toassure stability over the proposed shelf-life, it may bejustifiable to reduce frequency of testing for visibleparticles for annual batches placed into the productstability protocol.

4. Control of Visible Particles in LyophilizedProducts

Lyophilised products present unique challenges for thecontrol of visible particles because the cake will ob-scure detection of most visible particles. Therefore,100% inspection, though still required and of value forgeneral container and lyophilised cake appearance, haslimited capability for detection of particles. Therefore,reconstitution of the product is to date recommendedduring QC for inspection for visible particles. Becausethe reconstitution is destructive to the sample, it can-not be performed for 100% inspection, nor is it feasi-ble to use AQL sampling plans to the extent possiblefor a liquid product. A different approach is requiredfor visible particle control for lyophilised product to aliquid product. At minimum a predetermined samplesize of lyophilisate in stoppered, capped vials shouldbe reconstituted, during end-product QC testing, as

justified by the sponser. The sampling plan may bebased on the current industry standard of 10 to 20units, which is supported by the history of marketedlyophilised products, or by using ANSI Z1.4(ISO2859-1) Special Sampling Plans. Equivalent orbetter sampling plans may be used.

Note that post-lyophilisation sample handling, for ex-ample, reconstitution, can itself result in false-positivedetection of particles. The risk of introducing particlesof any source during reconstitution is not pertinent tothe control of particle during manufacture, productbatch release, or stability testing and should be aconcern of product handling and investigated as partof method development, container closure qualifica-tion, and compatibility with the product. Thus, recon-stitution and handling in QC should take into consid-eration the instructions for use also provided for thereconstitution by end users (e.g., handling by healthcare providers).

Visual inspection for lyophilisates in a stability pro-gramme again requires reconstitution of a predeter-mined sample size at each time point, as describedin Section 3.5.

Samples of reconstituted lyophilisate, after visual in-spection, can be also used for other QC tests, with thesame considerations applying as for liquid products.

5. Particle Identification and Characterization

The presence of visible particle(s) in a drug productcontainer may drive an investigation that often in-cludes further characterization to determine particleattributes, such as size, morphology, and composition,in an effort to find root cause. An investigation wouldbe triggered when, for example, atypical particles aredetected, unexpected growth of inherent protein par-ticles over time is observed, or if acceptance criteria ofunits with visible particles are exceeded. Particle char-acterization and identification typically involve re-search methods that are qualitative in nature, and thesemethods are not recommended as routine tests, such asassays for batch release.

The origin of the visible particles (14) can includeenvironmental, process, packaging components, theprotein drug itself, and any combinations of these. Inthe cases of atypical findings, the knowledge of theparticle origin may support the risk management andcontrol strategy and drive CAPA.




Non-conformance investigations as wejll as advancedparticle characterization as part of product and processdevelopment may require the use of a combination ofvarious analytical techniques. A multi-pronged ap-proach may be necessary because a single analyticaltechnique may not be able to reveal the key attributesof a particle. A first step is to determine the size,number, and aspect of the particle(s) in the unopenedprimary container, to the extent possible. Focusedvisual inspection under enhanced visualization condi-tions (such as polarized light and/or magnification,stereo- or inverted-light microscopy, advanced imageanalysis) can be used as additional analytical tools fornon-destructive analyses. Particle attributes from non-destructive analyses such as size, number, shape (e.g.,fiber, sphere), color, type (e.g., metallic, glass) andsettling behavior may support particle identification.

However in some rare instances, further analyticalcharacterization may be necessary to gain more reli-able information on composition and origin of theparticles. These tests are destructive in nature (i.e.,requiring opening of the container) and may ofteninvolve isolation of the particle(s) for further charac-terization. Therefore, careful sample handling is veryimportant to maintain integrity of the particle(s)planned for identification and ensuring no introductionof environmental particles during handling and anal-ysis. In some cases, the dosage form may contain asingle particle or a very few particles. Furthermore, itis quite common that only a single container (withatypical particle) is available for detailed characteriza-tion. Required sample volume for such extended char-acterization tests is also an important consideration,especially in early stages of development. Therefore, apredetermined strategy should be considered to ensureadequate data can be collected using limited numberof containers and/or limited volume of sample.

For example, the tests that are non-destructive in na-ture (not requiring container opening) should be con-ducted first (see flow chart in Figure 2). These includevisual inspection as well as use of enhanced visual-ization conditions such as magnification, lighting, con-trast, and optical microscopy. Subsequent tests forfurther characterization will require opening of thecontainer. Physico-chemical characterization of thenature and composition of particles can be accom-plished following the filtration of the particles onappropriate membrane or other sampling techniques.The sample handling step is critical and needs to becarefully controlled because it may alter the particle

by, for example, oxidation or breakage or dissolutionof fragile particles, and result in additional particleingress (false positives) or in a loss of the particle ofinterest. Tests with the isolated particle(s) may includeoptical microscopy for detailed morphological assess-ment, Fourier transform infrared (FTIR) or Ramanmicroscopy for compositional analysis, and other tech-niques such as Time-Of-Flight Secondary Ion MassSpectroscopy (TOF-SIMS) that may be applicable forcertain cases. In this stage of the analyses, isolatedparticles are not subjected to significant treatmentsother than isolating/filtering. If further characteriza-tion is needed and sufficient material is available, thenas the last step in the characterisation process, isolatedparticle(s) can be biochemically treated to enable testssuch as Matrix Assisted Laser Desorption Ionization(MALDI), sodium dodecyl sulfate capillary electro-phoresis (CE-SDS), Sodium dodecyl sulfate-poly-acrylamide gel electrophoresis (SDS-PAGE), peptidemapping, and the like.

Visual inspection as well as enhanced visualizationtests provide a good clue of the type of particle basedon shape/morphology. For example, protein particlesare often fluffy, amorphous, and translucent in nature,and may not reflect well under polarized light inspec-tion conditions. Protein particles also can be fibrillous(fiber-like). On the other hand, glass and metal parti-cles are often shiny and/or crystalline in nature.

Extended characterization tests for composition andorigin of particles include FTIR microscopy and Ra-

Figure 2

Flow chart for usage of containers for particlecharacterization.




man microscopy. FTIR/Raman microscopic tests aresuitable for measuring morphology of visible particlesas well as compositional analysis through spectralfingerprinting. Several types of particle compositionsincluding proteinaceous, inorganic, and organic can bestudied by FTIR/Raman microscopy. These tests mayalso be suitable to study subvisible particles in low-micron size range (approximately #20 $m). In Ramanmicroscopy assay, it needs to be considered that laser-induced damage of the sample can likely occur. Lightmicroscopy tests using polarized light or fluorescencemicroscopy (with and without staining or labeling ofparticles) may provide additional insight into the na-ture and characteristics of the particles.

Additional supporting tests may include elementalanalyses such as by scanning electron microscopy(SEM) coupled with energy-dispersive (EDX) orwavelength-dispersive X-ray detectors. However,SEM-EDX elemental data may not be confirmatorybecause particles of different origins may have similarelemental composition. Atomic force microscopy(AFM) and transmission electron microscopy (TEM)imaging have also been described to provide support-ing information about the particle morphology and finestructure at the molecular level (e.g., by cryo- ornegative staining TEM). Again, careful sample han-dling and understanding method capabilities are im-portant considerations for these tests.

The analysis of the size distribution (size and numberof particles at each size) is particularly important forproteinaceous particles, which cover a wide rangefrom visible precipitates to soluble sub-micron aggre-gates (15–17). During product development, it is im-portant to study a wide size range of aggregates/particles to understand a possible mechanistic link ofsmall aggregates and subvisible/submicron particles tothe formation of visible particles. Light-obscuration,flow microscopy, electrical sensing zone, and flowcytometry techniques are suitable to detect and countsmall amounts of particles in a size range from ap-proximately 1 to 100 $m. These analytical techniquesprovide comparable information of size distributiontrends, but particle counts may vary by orders ofmagnitude. Techniques to measure particles in thesub-micron size range include several emerging tech-niques such as field flow fractionation, nanoparticletracking analysis, and resonant mass measurements.This is an active area of research, and robustness ofthese techniques are yet to be established.

In summary, establishing the nature or origin of visi-ble particles is a complex process that requires thor-ough analytical due diligence. A toolset of variousorthogonal analytical techniques provides the best op-portunity to characterize particles, particularly theproteinaceous particles. The identification of visibleparticles should be considered for individual investi-gations and not for routine or QC testing.

6. Patient Safety

Particles remain a critical quality attribute, as “essen-tially/practically free of visible particles” is a require-ment for parenteral dosage forms (unless, e.g., usedwith a filter during administration or when particlesare a designed part of the formulation, e.g., insulinprotein crystals). Various clinical empirical observa-tions have been connected to particles in parenteralproducts, although there are no systematic studies thathave correlated specific types and levels of particles tospecific adverse effects. Literature in the field ismostly related to drug abuse or anecdotal evidence. Aconnection of (sub-visible or visible) proteinaceousparticles and immunogenicity has been hypothesized,although there is no direct clinical evidence for thishypothesis. Clinical relevance of particles has beenreviewed in recent literature (18, 19).

In general, small number of particles in parenteralproducts likely do not lead to any adverse impact onsafety. This is corroborated by a review of the ProductInformation for monoclonal antibody products ap-proved through the centralized procedure in the Euro-pean Union and through Biological License Applica-tion in the Unites States of America (see Table I): Inthe table above, particle presence is referred to as“may”, emphasizing the unpredictable and probabilis-tic nature of these particles.

Where unavoidable, an acceptance limit for protein-aceous visible particles should be based on exposureand adverse event data from non-clinical and/or clin-ical studies, assuming the product is not intended to beadministered to the patient through an in-line filter asprescribed in the Product Information.

7. Conclusions

Visible particle assessments remain a challenge in thedevelopment of parenteral products, especially of bio-technology products. The biopharmaceutical industryis striving towards improving the quality and cleanli-




TABLE IExamples of EU/US Product Information for Biotechnology Products Potentially Containing Protein Particles

Generic Name (Routeof Administration) Product Name

ProductInformation

EU or US Section Text

Anakinra injection(subcutaneous)

Kineret US 2.4 Administration There may be trace amounts of small,translucent-to-white amorphous particles ofprotein in the solution. If the number oftranslucent-to-white amorphous particles in agiven syringe appears excessive, do not use thissyringe.

11 Description The solution may contain trace amounts of small,translucent-to-white amorphous proteinaceousparticles.

Etanercept injection(subcutaneous)

Enbrel EU 6.6 Special precautionsfor disposal andother handling

The solution should be clear to slightly opalescent,colourless or pale yellow and may contain smalltranslucent or white particles of protein.

US 2.4 Preparation ofEnbrel

Parenteral drug products should be inspectedvisually for particulate matter and discolorationprior to administration.

There may be small white particles of protein inthe solution. This is not unusual forproteinaceous solutions.

The solution should not be used if discolored orcloudy, or if foreign particulate matter is present.

Denosumab injection(subcutaneous)

Prolia EU 6.6 Special precautionsfor disposal andother handling

The solution may contain trace amounts oftranslucent to white proteinaceous particles.Do not inject the solution if it is cloudy ordiscoloured.

US 2.2 Preparation andAdministration

Prolia is a clear, colorless to pale yellow solutionthat may contain trace amounts of translucentto white proteinaceous particles.

Golimumab injection(intravenous)

Simponi ARIA US 2.3 ImportantAdministrationInstructions

The solution may develop a few fine translucentparticles, as golimumab is a protein. Do not useif opaque particles, discoloration, or other foreignparticles are present.

Ipilimumab injection(intravenous)

Yervoy injection US 2.3 Preparation andAdministration

Discard vial if solution is cloudy, there ispronounced discoloration (solution may have apale yellow color), or there is foreign particulatematter other than translucent-to-white,amorphous particles.

Infliximab for injection(intravenous)

Remicade US 2.11 GeneralConsiderations andInstructions forPreparation andAdministration

The solution should be colorless to light yellow andopalescent, and the solution may develop a fewtranslucent particles, as infliximab is a protein.

Nivolumab injection(intravenous)

Opdivo US 2.3 Preparation andAdministration

Visually inspect drug product solution forparticulate matter and discoloration prior toadministration. Opdivo is a clear to opalescent,colorless to pale-yellow solution. Discard the vialif the solution is cloudy, is discolored, or containsextraneous particulate matter other than a fewtranslucent-to-white, proteinaceous particles.

11 Description Opdivo is a sterile, preservative-free, non-pyrogenic,clear to opalescent, colorless to pale yellow liquidthat may contain light (few) particles.

Panitumumab injection(intravenous)

Vectibix US 2.3 Preparation andAdministration

Although Vectibix should be colorless, the solutionmay contain a small amount of visibletranslucent-to-white, amorphous,proteinaceous, panitumumab particulates(which will be removed by filtration).

Pembrolizumab forinjection(intravenous)

Keytruda US 2.3 Preparation andAdministration

Visually inspect the reconstituted solution forparticulate matter and discoloration prior toadministration. Reconstituted Keytruda is a clearto slightly opalescent, colorless to slightly yellowsolution. Discard reconstituted vial if extraneousparticulate matter other than translucent towhite proteinaceous particles is observed.

Ustekinumab solutionfor injection(subcutaneous)

Stelara EU 6.6 Special precautionsfor disposal andother handling

The solution is clear to slightly opalescent,colourless to light yellow and may contain a fewsmall translucent or white particles of protein.This appearance is not unusual for proteinaceoussolutions. The medicinal product should not beused if the solution is discoloured or cloudy, or ifforeign particulate matter is present.




ness of products, including addressing the upstreamsources of particles. Although the nature of visualinspection is probabilistic and the presence of tracelevels of particles in parenteral products cannot befully excluded, a holistic approach to minimizing thepresence of visible particles in parenteral protein drugproducts as for other injectable drug products is pro-posed. However a requirement of zero (or without)visible particles is overly stringent and practically notattainable. Particle controls are and should be one ofthe main formulation and process design criteria ap-plied by the biopharmaceutical industry. This contin-uous improvement objective being acknowledgedacross the industry, the current industry positionreflected and justified at length in this position paperis to consider “practically free from visible particlesunless otherwise justified and authorised” as a stan-dard requirement for QC release of biotechnology-derived drug products including monoclonal anti-bodies subject to compliance with the EuropeanPharmacopoeia.

Acknowledgments

The paper was written in collaboration with otherexperts from European biopharmaceutical enterprises(EBE) Visible Particle topic group and from the EBEBiological Manufacturing Group member companiesthat contributed and supported the preparation of thisdocument: Atanas Koulov, Analytical Development &Quality Control, Pharma Technical Development Bi-ologics EU, Roche; Sharon Adderley, ManufacturingScience and Technology, Analytical Science, Pfizer;and Rober W. Kozak, Regulatory Affairs, BayerHealtcare LLC.

Conflict of Interest Declaration

The authors declare that they have no competing in-terests.

References

1. Knapp, J.; Kushner, H. Generalized Methodologyfor Evaluation of Parenteral Inspection Proce-dures. J. Parenter. Drug Assoc. 1980, 34 (1),14 – 61.

2. Melchore, J. A. Sound Practices for ConsistentHuman Visual Inspection. AAPS PharmSciTech2011, 12 (1), 215–221.

3. Leversee R.L.; Shabushnig J.G. A Survey of In-dustry Practice for the Visual Inspection of Inject-able Products (Preliminary Report). PDA VisualInspection Forum, October 2008.

4. There may be clinical circumstances where tighterAQL values (limits) may be appropriate for high-risk patients and for other routes of administrationbased on an evaluation of patient risk. Bukofzer,S.; Ayers, J.; Chavez, A.; Devera, M.; Miller, J.;Ross, D.; Shabushnig, J.; Vargo, S.; Watson, H.;Watson, R. Industry Perspective on the MedicalRisk of Visible Particles in Injectable Drug Prod-ucts. PDA J. Pharm. Sci. Technol. 2015, 69 (1),123–139.

5. Knapp, J. Z.; Budd, G. W. Part II: Visible Inspec-tion Data and the Sampling Inspection. PDA An-nual Conference, April 4-6 2005, Chicago, IL.

6. Taylor, W. A. Guide to Acceptance Sampling;Taylor Enterprises, Lake Villa, Il 1992.

7. Gershon, M. Statistical Process Control for thePharmaceutical Industry. J. Parenter. Sci. Tech-nol. 1991, 45(1), 41–50.

8. Guadagnion, E.; Zuccato, D. Delamination Pro-pensity of Pharmaceutical Glass Containers byAccelerated Testing with Different ExtractionMedia. PDA J. Pharm. Sci. Technol. 2012, 66 (2),116 –125.

9. Jiang, G.; Goss, M.; Li, G.; Jing, W.; Shen, H.;Fujimori, K.; Le, L.; Wong, L.; Wen, Z. Q.;Nashed-Samuel, Y.; Riker, K.; Germansderfer,A.; Tsang, P.; Ricci, M. Novel Mechanism ofGlass Delamination in Type 1A BorosilicateVials Containing Frozen Protein Formulations.PDA J. Pharm. Sci. Technol. 2013, 67 (4),323–335.

10. Lacocca, R. G.; Toltl, N.; Allgeier, M.; Bustard,B.; Dong, X.; Foubert, M.; Hofer, J.; Peoples, S.;Shelbourn, T. Factors Affecting the Chemical Du-rability of Glass Used in the Pharmaceutical In-dustry. AAPS PharmSciTech 2010, 11 (3), 1340 –1349.

11. Sloey, C.; Gleason, C.; Phillips J. Determining theDelamination Propensity of Pharmaceutical Glass




Vials Using a Direct Stress Method. PDAJ. Pharm. Sci. Technol. 2013, 67 (1), 35– 42.

12. Cash, P.; Narwal, R.; Levitskaya, S. V.; Krause,S.; Murphy, D.; Mazaheri, M. Semi-QuantitativeAnalysis of Inherent Visible Particles for Biop-harmaceutical Products. PDA J. Pharm. Sci. Tech-nol. 2016, doi: 10.5731/pdajpst.2015.

13. Saddic, G. Challenges In Developing a Semi-Quantitative Visible Particulate Method for Rou-tine Commercial Testing. 1st International Sym-posium on Higher Order Structure of ProteinTherapeutics, Rockville, MD, September 27,2011.

14. United States Pharmacopeia. "790# Visible Par-ticulates in Injections. United States Pharmaco-peia–National Formulary; USP 37 NF 32, 2014.

15. Wuchner, K.; Buchler, J.; Spycher, R.; Dal-monte, P.; Volkin, D. B. Development of aMicroflow Digital Imaging Assay to Character-ize Protein Particulates during Storage of aHigh Concentration IgG1 Monoclonal AntibodyFormulation. J. Pharm. Sci. 2010, 99 (8), 3343–3361.

16. Das, T. K. Protein Particulate Detection Issues inBiotherapeutics Development-Current Status.AAPS PharmSciTech 2012, 13 (2), 732–746.

17. Singh, S. K.; Afonina, N.; Awwad, M.; Bech-told-Peters, K.; Blue, J. T.; Chou, D.; Crom-well, M.; Krause, H. J.; Mahler, H. C.; Meyer,B. K.; Narhi, L.; Nesta, D. P.; Spitznagel, T. AnIndustry Perspective on the Monitoring of Sub-visible Particles as a Quality Attribute for Pro-tein Therapeutics. J. Pharm. Sci. 2010, 99 (8),3302–3302.

18. Doesegger, L.; Mahler, H. C.; Szczesny, P.; Rock-stroh, H.; Kallmeyer, G.; Langenkamp, A.; Herr-mann, J.; Famulare, J. The Potential Clinical Rel-evance of Visible Particles in Parenteral Drugs.J. Pharm. Sci. 2012, 101 (8), 2635–2644.

19. Bukofzer, S.; Ayres, J.; Chavez, A.; Devera, M.;Miller, J.; Ross, D.; Shabushnig, J.; Vargo, S.;Watson, H.; Watson, R. Industry Perspective onthe Medical Risk of Visible Particles in Injectable

Drug Products. PDA J. Pharm. Sci. Technol.2015, 69 (1), 123–139.

20. Melchore J. A.; Berdovich, D. Considerations forDesign and Use of Container Challenge Sets forQualification and Validation of Visible ParticulateInspection. PDA J. Pharm. Sci. Technol. 2012, 66(3), 273–284.

21. FDA Guidance for Industry. Immunogenicity as-sessment for therapeutic protein product. August2014. Accessed February 3, 2016; http://www.fda.gov/downloads/drugs/guidancecomplianceregulatoryinformation/guidances/ucm338856.pdf.

Appendix 1: Inspector Certification

The visual appearance test is subjective and the resultis probabilistic since the instrument is the analyst’seye. Effective inspector selection, training, and mon-itoring are therefore necessary for a reliable and con-sistent visual inspection program. The training forinspectors performing manual visual inspection of un-labelled drug product containers requires certification,which would include assessment of visual acuity andtechnical expertise that includes the ability to detectparticles in test panels.

The same principle of training and monitoring shouldbe provided for all personnel performing visual in-spection for visible particles during manufacture aspart of 100% manual inspection, AQL testing, QCrelease and stability testing, QA reserves/retention,and product complaints.

Inspector Selection and Qualification

1. Selection Criteria for Visual Inspector

a. Each trainee must have an eye exam completed,for example, including vision acuity test andcolor blindness test.

b. Trainees should pass eye examination beforebeing qualified for visual inspection training.This documentation should be archived for fu-ture reference.

2. Training Process for Inspector

a. Introduction to general visual inspectionmethod.



http://www.fda.gov/downloads/drugs/guidancecomplianceregulatoryinformation/guidances/ucm338856.pdf




b. Introduction to visual inspection station withblack and white background, according to PhEur2.9.20.

c. If a defect library is available, a demonstrationof typical defects should be performed. Defectset criteria and its stability, should be capturedaccording to predefined SOP.

d. Demonstration of visual inspection method bytrainer.

e. Demonstration of sample handling and swirlingmethod, according to predefined SOP.

f. Introduction to defect categories and associatedAQL levels as per SOP.

g. Proficiency runs using appropriate test panels(qualification defect test sets) for the processand product, as many as required.

Training sets (20) should include a range of defecttypes, including particles that would be relevant todefects possibly found during the manufacturing pro-cess. These test panels should use containers both withand without defects. It is recommended to includerepresentative pictures or videos (with and withoutmagnification) of the particle types as part of trainingmaterial.

Training sets should be maintained using definedcriteria. All sets should have expiration or retestdate and may be reassessed, per SOP, for continueduse. Note: Qualification conditions need to mimicthe actual inspection conditions, for example, usingthe same rate of inspection for units and using thesame equipment.

3. Maintenance and Monitoring of Visual Inspector

a. Regular, for example, annual, procedure reviewand assessment; compliance to SOP and abilityto detect defects.

b. Eye exams must be performed regularly, forexample, annually.

i. Inspector must be re-qualified at regular inter-vals, for example, annually, to be considered aqualified visual inspector.

Appendix 2: Risk Assessment of Visible Particles

Generally risk assessments evaluate the likelihood(occurrence) an event will happen (e.g., visible parti-cle), the impact (severity) of the event on productsafety and efficacy, and the ability to detect either theevent or the impact.

The risk assessment should have input from all rele-vant disciplines—for example, product quality, CMCproduct development (process, formulation, analyti-cal), regulatory, manufacturing, clinical, clinical phar-macology, drug safety, drug metabolism (pharmaco-kinetic/pharmacodynamics/immunogenicity), andpharmacovigilance.

The origin of the visible particle(s) may be environ-mental, related to the process, packaging components,drug product formulation or proteinaceous makeupand/or likely mixtures thereof. Categorization of par-ticles into inherent, intrinsic, and extrinsic particles isdefined in USP "790# monograph.

As mentioned in Section 5, in cases of atypical find-ings indicative of system failure, the knowledge of theparticle origin/source may support the risk manage-ment and control strategy and drives CAPA. Systemfailure is defined as opposite to “randomly and rarelyoccurring”. For example, the formation of glass lamel-lae resembling flake-like particles over time wouldsignal a system failure (Section 3.6). Insect parts thatmight be detected during AQL testing at end of man-ufacturing would also be considered as particles re-flective of a system failure. By contrast, as discussedin Sections 2 and 3, rare occurrence of particles orpresence of proteinaceous particles is a concept thatdefines what might be considered as expected or nor-mal based on the product history.

Based on batch history, and trend analysis of com-plaints, the likelihood of occurrence of proteinaceousparticles can be determined as, for example, extremelyrare, remote, occasional, likely, probable, almost cer-tain, or even unknown, allowing differentiation ofproducts that would fulfill requirements for being cat-egorized as “product practically free from particles”from products that may contain protein particles.

For products “practically free from particles”, someisolated (occasional random) occurrence of visibleparticles (including protein particles) may occur, andbe identified during QC testing at release or during




stability studies or in complaints from patients/healthcare professionals and would trigger an investigationto identify the root cause, characterize the particles inorder to assess possible safety and efficacy impact topatients taking into consideration the route of admin-istration of the product, patient status, and so forth.

From a medical risk perspective, extraneous particlesintroduced by intravenous or intravitreal route hasbeen a greater concern compared to intramuscular orsubcutaneous routes. Proteinaceous particles may be agreater concern for subcuntanous administration (21)if a link of particles and immunogenicity is estab-lished. Particles injected into closed spaces (e.g., in-trathecal, intraarticular) may not be eliminated ordrained and also represent a higher potential for harm.Consequently the nature, size, and capability of theparticles to dissolve or disaggregate might have to beassessed.

For particle characterization including protein parti-cles, the following parameters may be considered:

● Potential dissolution of the particles by swirlingand exposure to room or even physiologic temper-ature.

● Solubility, buoyancy properties of particles.

● Characterization (size, number, volume, morphol-ogy) (see Section 5).

● Quantity of particle (weight).

In the patient safety/efficacy assessment, the followingparameters may be considered:

● Use of in-line filter for administration to patients.

● Adverse events: immunogenicity and impact ofpotential Anti-drug Antibodies (ADA), thrombo-embolic event (less of a risk of capillary occlusionwith protein particle as they are likely flexible/deformable), intraocular inflammation, and soforth.

● Daily intake—acute versus chronic therapy andexposure duration.

● Population treated, patient general state, patientimmune status.

● Potential induction of pathological change.

● Amount, size distribution of particles per individ-ual container.

● Dissolution and type of particles.

● Possibility to migrate from the injection site tobloodstream.

● Loss of potency.

● Alteration of the pharmacodynamic effect.

● Biodistribution.

● Amino acid sequence similarity to endogenousproteins, permitting antibody cross-reactivity.

Ultimately the overall risk assessment can take intoconsiderations the relevant parameters mentionedabove to define a risk level associated with the parti-cle(s) and support the decision for product disposition,and for preventive or corrective strategies. The overallrisk level could be categorized as high, moderate, low,very low, or negligible.




Authorized User or for the use by or distribution to other Authorized Users·Make a reasonable number of photocopies of a printed article for the individual use of an·Print individual articles from the PDA Journal for the individual use of an Authorized User ·Assemble and distribute links that point to the PDA Journal·Download a single article for the individual use of an Authorized User·Search and view the content of the PDA Journal permitted to do the following:Technology (the PDA Journal) is a PDA Member in good standing. Authorized Users are An Authorized User of the electronic PDA Journal of Pharmaceutical Science and

copyright information or notice contained in the PDA Journal·Delete or remove in any form or format, including on a printed article or photocopy, anytext or graphics·Make any edits or derivative works with respect to any portion of the PDA Journal including any·Alter, modify, repackage or adapt any portion of the PDA Journaldistribution of materials in any form, or any substantially similar commercial purpose·Use or copy the PDA Journal for document delivery, fee-for-service use, or bulk reproduction orJournal or its content·Sell, re-sell, rent, lease, license, sublicense, assign or otherwise transfer the use of the PDAof the PDA Journal ·Use robots or intelligent agents to access, search and/or systematically download any portion·Create a searchable archive of any portion of the PDA JournalJournal·Transmit electronically, via e-mail or any other file transfer protocols, any portion of the PDAor in any form of online publications·Post articles from the PDA Journal on Web sites, either available on the Internet or an Intranet,than an Authorized User· Display or otherwise make any information from the PDA Journal available to anyone otherPDA Journal·Except as mentioned above, allow anyone other than an Authorized User to use or access the Authorized Users are not permitted to do the following:



a biopharmaceutical industry perspective on the control of ... · commentary a biopharmaceutical...

Documents