Conducting Clinical Risk Assessments for Visible Particulate Matter in Parenteral Preparations

Background

Injectable product inspection standards date to the early 20th century. Historically, the presence of unintended visible particulate matter has been regarded as a safety-impacting quality defect, resulting in re-inspection, rejection, or recall for product batches. The rationale that visible defects were inherently hazardous arose from an innate, visceral perception deeply rooted in pharmaceutical culture. It is important to realize, however, that our innate perceptions and presumptions are often misguided when it involves biologic and physical sciences (10). As a result, the notion that visible particulate matter posed a unique patient harm likely drove superfluous batch-directed actions despite an acceptable margin of safety.

As process and inspection capabilities improved, ensuing standards promoted an idealistic expectation of a zero-defect quality paradigm for foreign matter and particles (11). What has ultimately evolved from this are visual inspection processes and batch disposition requirements predicated on the ability to visibly detect particles—generally fixed to the limits of human visual acuity—so any particle detected above an established detection threshold for that product would be—should be—rejected. As automated inspection capabilities improve and the detectable particle size decreases, clinical input will become an essential factor in establishing acceptance criteria. Otherwise, if acceptance criteria are based solely upon inspection process capability significant resources might be expended—and otherwise wholly acceptable product discarded—absent any meaningful impact to the product's benefit–risk profile. Such actions may serve a substantial negative impact to the availability of injectable pharmaceutical medications or lead to justifying the distribution of suboptimal product fearing a potential drug shortage (12).

The absolute lower limit of unaided human visual detection is generally accepted at around 50 microns. Pooled-study inspection results under ideal conditions and prepared particle standards provide the following findings (13):

• the probability of detection of a seeded, 100 μm particle standard is approximately 40%;

• the threshold for routine, reliable detection (≥70% probability of detection) of individual visible particles is often near 150 μm in diameter; and

• routine, reliable detection typically exceeds 95% for particles that are 200 μm and larger.

It is essential to acknowledge that the likelihood of visual detection of macroscopic particulates is a probabilistic event and detection by a manual, semi-automatic or automated visual inspection is a cumulative function of the particles' visible attributes such as particle size, shape, color, density, and reflectivity and is not an unequivocal certainty. However, in spite of this fact, there has been a history of considerable trepidation concerning the treatment of visible particles, especially those falling within the ‘region of probabilistic detection’ (around 50–200 microns in the ideal conditions) and subsequently identified in a released or approved product batch. Until the adoption of USP <790> (14), this equivocation was compounded by the fact that the compendia have not provided definitive guidance concerning units containing particles below the level of routine, reliable visible detection might be managed. Thus, particles falling within the 0–99% detection window [i.e., the Knapp Zone (15)] might not be detected during the 100% inspection or the statistical acceptable quality limit (AQL) sampling plan inspection, but found subsequent to batch release. The resulting outcome might include batch recall—for a finding with less than a 50% probability of initial detection—ignoring the fact that the probability of detection and the probability of non-detection are two inseparably linked covariants.

The relevant particle-related U.S. Pharmacopeia requirements are excerpted below: Foreign and Particulate Matter USP <1>: Articles intended for parenteral administration should be prepared in a manner designed to exclude particulate matter as defined in Subvisible Particulate Matter in Therapeutic Protein Injections <787> (16), Particulate Matter in Injections <788> (9), or Particulate Matter in Ophthalmic Solutions <789> (17), as well as excluding other foreign matter as appropriate for the dosage form. Each final container of all parenteral preparations should be inspected to the fullest extent possible for the presence of observable foreign and particulate matter (hereafter termed visible particulates) in its contents. The inspection process should be designed and qualified to ensure that every lot of all parenteral preparations is essentially free from visible particulates, as defined in Visible Particulates in Injections <790>. Qualification of the inspection process should be performed with reference to particulates in the visible range and those particulates that might emanate from the manufacturing or filling process. Every container in which the contents show evidence of visible particulates must be rejected. The inspection for visible particulates may take place during examination for other defects such as cracked or defective containers or seals, or when characterizing the appearance of a lyophilized product.

While the qualifying language “essentially free from visible particulates” found in USP <1> acknowledges that neither parenteral manufacturing processes nor visual inspection capability can deliver a zero-defect quality level, interpretation of what constitutes “essentially free” has been the subject of inconsistent interpretation by both manufacturers and regulators. As a result, the U.S. Pharmacopeia adopted a new chapter under “General Chapters: Physical Tests and Determinations” <790> Visible Particulates in Injectables (14), the purpose of which is to provide a workable definition of essentially free and offer an archetypical approach for both an appropriate inspection environment and suitable visual inspection methods.

Acknowledging that visible particle detection is probabilistic, Chapter <790> also provides a method to establish that a batch conforms to the pharmacopeia requirements of “essentially free” if—subsequent to batch release after 100% inspection and an appropriate statistical sampling plan inspection—a unit containing a visible particle has been identified: If it becomes necessary to evaluate product that has been shipped to customers (e.g., because of a complaint or regulatory concern), sample and inspect 20 units. If no particles are observed in the sample, the batch is considered essentially free of visible particulates. If available, additional units may be inspected to gain further information on the risk of particulates in the batch.

Characterizing the Clinical Risk

By convention, most quality attribute assessments assign the criticality of the defect or variability of the attribute to one of three categories: minor, major, or critical (18). Once assigned, analytic techniques and resolution capabilities, control strategies, 100% inspection, and statistical sampling plan inspection paradigms can be established to support the risk classification for that attribute. In the case of macroscopic particulates in injectables occurring in units produced in a qualified and cGMP-compliant manufacturing process, visible particles should generally be characterized as a “major” defect. This classification is consistent with USP <790> and provides a risk-level commensurate with contemporary manufacturing and inspection capabilities. Establishing control strategies with macroscopic particles routinely classified as “critical” defects (19) simply does not conform to the overabundance of clinical experience and in the overwhelmingly majority of instances overstates the potential patient risk. Consequently, routine classification of visible particles as “critical defects” would result in excessive batch rejection of drug product and would be devoid of any meaningful impact to the product's benefit–risk profile. By the same token, establishing the defect as “minor” is inconsistent with contemporary cGMP-manufacturing capabilities and introduces a potential patient risk incremental to current clinical experience. However, there may be instances where stability-indicating intrinsic or inherent proteinaceous particles in biopharmaceuticals are measured, characterized, and present in clinical studies in humans—providing the context for a corresponding adjustment in the severity rating. Aldrich, Cherris, and Shabushnig (20) have recently published an excellent book on the technical aspects of visual inspection and particulate control that serves as an essential resource for parenteral manufacturing in outlining the important and fundamental components that should be evaluated. As noted above, the knowledge concerning the safety of injectables is predicated upon the consistency of the cGMP manufacturing and inspection capabilities of the manufacturing line.

Assessing the Clinical Risk

The above-compendia chapters highlight the requirements in conducting particle determinations. Importantly, these criteria are not a litmus test for product quality (i.e., a means to test quality into the product batch) but are one element of the confirmatory data set—including the checks and balances of an adequately defined control strategy such that the batch meets pharmacopeial provisions defining a cGMP-compliant process. The compendial acceptance criteria should serve as baselines for product quality and not absolute product-specific batch release specifications. Importantly, USP <790> includes this cogent stipulation: Used along with 100% inspection during the manufacturing process, this procedure is sufficient to demonstrate that the batch is essentially free of visible particulates. A complete program for the control and monitoring of particulate matter remains an essential prerequisite. (Emphasis added).

Yet, while implicit in context, the pharmacopeia general chapters cannot answer the one question that should serve as a fundamental consideration in establishing acceptance limits for particulate matter in parenteral products: What is the potential risk that macroscopic particulate matter might pose to the patient receiving the drug product? In conducting a clinical risk assessment, one should remain cognizant of the fact that the clinical assessment is simply one component feeding into the control strategy and defect criticality assessment. Considering the particles' attributes and patients' characteristics in the three broader categories of minor, major, and critical will allow focus on the more essential distinction—identifying instances where the clinical classification should move from major to critical. While it is certainly possible to utilize an expanded severity scale to characterize potential patient risk from zero to catastrophic, attempts to nuance that evaluation through the use of expanded gradations will only serve to obfuscate the assessment process.

In that regard, in order to adequately assess the patient safety-risk particles may introduce, generally, three factors should be considered:

• the source of the particle;

• particle-specific attributes; and

• characteristics of the intended patient population.

Sources of Particles

Particles in injectable preparations originate from one of three sources: extrinsic to the manufacturing process; intrinsic to the process or primary packaging; or, inherent to the drug product formulation. The definitions below are abstracted from USP <1790> (21), the informational chapter supporting USP <790>. Although the definitions have not been adopted as a whole by other major pharmacopeias, they do serve to provide relevant context for risk assessments. Extrinsic particles are exogenous: foreign to the manufacturing process. This includes materials such as hair, non-process related fibers, insect parts, and similar inorganic and organic matter. Units containing this debris should be rejected during the manufacturing-inspection process and if later identified, assessed for a potential increased risk of microbiological contamination, the introduction of organic or inorganic leachates or interaction with the drug substance.Intrinsic particles are those from ‘within’ the process and generally originate from processing equipment or primary packaging materials. These particles are typically sterile or sanitized, so the primary risk is related to the morphologic (form and structure) aspects of the particle and potential interaction with the drug product. Distinct subsets of intrinsic particles are “stability-related” or “stability-indicating” particulates that result from container–closure interaction, arising from changes to the drug formulation (i.e., insoluble degradation products), or are an effect of temperature, agitation, or sheer stress on the product formulation.Inherent particles are those particles effectively “intended” to be in the products, such as suspensions and emulsions or those designed as particle assemblies such as agglomerates and aggregates. In biopharmaceuticals, protein particles are considered inherent when their presence may be measured, characterized, and determined to be part of the clinical profile. (Emphasis added)

It is important that protein particles are appropriately identified as either inherent particles or stability-indicating intrinsic particles, as the control strategy to manage the formation of these particles and potential clinical implications may be different.

Particle-Related Attributes

Both particle-specific attributes and patient-specific factors may affect the ultimate impact of patient exposure to/from particulate matter in parenteral drug products. Generally, the potential effect of four particle-related sequelae should be considered:

• the impact on sterility (as a source of bioburden or a nidus for infection);

• the particles' composition and potential likelihood for organic or inorganic leachates;

• the possibility that the particle might promote intravascular occlusion (e.g., frankly obstructing the lumen of a vessel or as a locus to propagate thrombosis) or serve as a soft tissue or organ irritant; and

• the potential to express features that might induce or enhance an untoward immune response.

In most instances, the greatest risk to patients receiving injectable pharmaceutical preparations and subject to manufacturing control processes relates to a lack of product sterility. This is the reason that a robust inspection program is not simply directed to the presence of particles, but encompasses a broad array of inspection steps, microbiologic and analytical testing, and statistical sampling plans to cull out units that might pose a risk of bioburden or other related harms. While the presence of a single particle is unlikely to provide a sufficient substrate to support microbiologic proliferation, multiple units identified with extrinsic particles should raise an alarm as the source is likely from a non-sterile environment. If microbial contamination of the drug product occurs from the presence of particulate matter, then proliferative growth might proceed under favorable conditions that must include the following: the presence of adequate moisture; the presence of nutrients in the drug product; and, the absence of inhibitory substances (22). Thus, the first question considered when presented with a foreign particle should be: What is the potential nexus to product sterility?

The next question necessitates an understanding of the particle's composition as well as the constituents of the formulated drug product. Does the particulate contain organic or inorganic components that might leach from the particle and either pose a risk of organotoxicity or negatively interact with the drug product? While the actual risk of harm is likely to be very low due to the mass of material being quite small, there might be circumstances where this issue arises as a reasonable concern for inherent material particles. For example, the vial stopper has historically been identified as the largest source of particulate material in injectable products (23). The presence of a small, stopper fragment might not be of any practical concern for a routine single injection. However, for patients receiving multiple injections over an extended period of time, such as patients receiving therapy for hormone deficiencies or autoimmune conditions, accumulation of small fragments in the subcutaneous compartment over months or years might impart a theoretical risk, and therefore, a special consideration to the possible leachates from closure materials used with this product should be an important part of the development and validation processes (24). The medical literature assessing patients' immune response to surgical implants (vascular grafts or orthopedic prostheses, for example) provides a useful resource for understanding the potential focal or systemic impact of some product-contact materials (25).

Another point to consider is the potential for the particle to act as a vaso-occlusive agent by way of directly obstructing a vessel, serve as a locus to propagate thrombosis or if injected into the soft tissue, stimulate a focal immune response, or act as a nidus around which microbes might propagate and create an abscess.

When intrinsic particles are introduced into patients' subcutaneous tissue or muscle, there is generally little concern for infection by these materials, as they are typically sterile or sanitized. It is possible, however, that they might stimulate a local immunologic reaction consisting of macrophages and foreign body giant cells at the interface of the particulate material (e.g., a granuloma), but it is likely that such a reaction would remain asymptomatic. More commonly, complications from subcutaneous and intramuscular medications arise from the irritating properties of the drug product and are often drug-specific but generally have a minimal impact (26).

Intravenous (IV) infusion of particles may induce phlebitis due to particles causing direct traumatic irritation to the vein or chemical damage from undissolved particles (27). When introduced intravenously, a particle that exceeds the internal diameter of a pulmonary capillary (7–12 μm) would lodge in that capillary or a more proximal pulmonary arteriole. While clinical outcomes are generally benign, such matter introduced into the systemic circulation might stimulate a local immunologic reaction (e.g., a granuloma) in the lung (28). These isolated, small parenchymal lesions generally remain asymptomatic. In addition, as demonstrated in animal studies, the massive infusion of particulates has been accompanied by histologic evidence of injury to pulmonary capillary endothelial cells (29), microscopic thrombi in the pulmonary capillaries (30), pulmonary microscopic granulomata (31), and hepatic inflammatory effects (32). A comprehensive review of the clinical implications of particulate infusions is beyond the intended scope of this paper and most of the studies above infused a massive number of subvisible particles. Suffice it to say the intravascular infusion of particulate matter has the potential for pathophysiologic sequelae (33)¹ likely dependent on the size and quantity of the particles and should be included as an element in the clinical risk assessment (3).

A corollary to this question is, “Does the material composition of the particle (e.g., glass, stainless steel, or rubber etc.) raise the prospect for a greater or lesser traumatic tissue or vascular insult?” In all instances of injections, the size of the particle is a restricting attribute. The particle must be small enough (or non-rigid) to pass through the bore of the needle (in other words, the particle must be smaller than the internal diameter of the needle in at least two dimensions). Additionally, detection increases with size such that larger particles are more likely to be identified and the product either rejected during manufacture or found during dose preparation and discarded. Thus, the particles that are ultimately injected will generally be at or below the limit of detection of the qualified manufacturing environment, with the overwhelming majority of those particles remaining being subvisible. There is no compelling clinical evidence to support the supposition that particles composed of different materials pose fundamentally different traumatic risks.

For example, glass particles are often cited as presenting a greater risk to patient safety if injected (19). However, this opinion appears to emanate from the residuum of one's personal experience with glass and its propensity to harm when shattered, such as stepping on a broken bottle or falling through a window. The important notable exception to this discussion are glass lamellae, presumptively the result of the vessel contents interacting with the glass surface. Glass particles resulting from delamination point to a serious primary container-related deficiency, their presence categorically unacceptable for the administration and storage of parenteral therapies.

The more common finding, glass particles originating from the primary container or breakage on the filling line and found in parenteral product containers, are without exception either sanitized or sterilized. Any residual risk from these particles likely derives from its morphologic features. At the sub-200 μm range, the structure and configuration of glass, metal, or rubber appear surprisingly similar (34). In any case, pulmonary capillaries comprise single endothelial cells, and pulmonary arterial and systemic intravenous blood is contained in a low-pressure, low-flow environment. Therefore, the composition of the particle and its propensity to induce intravascular trauma is likely not a clinically relevant or differentiating factor.

The final particle-related consideration is the potential for particulate matter or aggregates of particulate matter to express features that might induce or augment an untoward immune response. Several reviews provide examples of aggregated proteins inducing immune responses, with clinical sequelae ranging from no apparent clinical impact to hypersensitivity reactions and the formation of cross-reactive and neutralizing antibodies (35). Additionally, regulatory guidance documents are available that provide for a risk-based approach to evaluating and mitigating immune responses associated with therapeutic protein products that affect their safety and efficacy (36, 37). Consideration of the potential immunogenic impact of protein-based aggregates is a critical component of the clinician's assessment of the drug product. It is essential, therefore, that inherent protein particles and stability-related intrinsic particles in biopharmaceuticals are measured, characterized, and present in correlative, investigational, new-drug application studies in humans. In conjunction with analytical data sets and in vitro studies, product-class knowledge, animal studies (although providing limited relevance to human immune responses), and appropriate literature reviews, the totality of the data can provide valuable insight into the potential patient impact. “In such studies, it is possible to learn the impact of multiple product attributes, including aggregates, and subvisible particles (SVPs), particularly if they are assessed in each lot of product administered to a defined group of patients and are correlated to an immunogenicity profile in those patients” (38)

Patient-Related Factors

Likewise, patient characteristics play an equally important role in the ultimate effect of parenteral exposure to particulate matter. Although the ensuing list is not meant to be exhaustive, the following patient-related factors should generally be considered:

• age;

• comorbid conditions;

• immunocompetency and/or alterations in immune status;

• presence of ischemic tissue (e.g., vascular occlusion, crush injury, or burns); and

• the route, volume, frequency, and duration of administration.

These patient-related aspects should be addressed in the context of the following: the label-indicated use; the known genotypic or phenotypic characteristics of the intended patient population; and the reasonably foreseeable off-label product use² for both indication and patient population (e.g., pediatrics or immunocompromised), among others. Because these criteria will be product- and patient-specific, each marketed product should have an individualized evaluation.

The three factors of age, comorbidities, and immune status, are essential to the discussion as age at the extreme ends (neonates and advanced elderly), especially as these individuals may exhibit comorbid conditions or a weakened immune response. In some situations, those patients may lack the residual functional capacity to tolerate pathophysiologic insults that otherwise healthy individuals might experience without untoward sequelae. For example, in other patients, an immune system in a heightened state due to an autoimmune disorder, might exhibit an exaggerated response, inducing a hypersensitivity or inflammatory reaction. Significant pulmonary injury and death have been reported from the presence of non-visible precipitates containing calcium, phosphorus, and organic material in total parenteral nutrition (39) as well as the co-administration of ceftriaxone and calcium-containing intravenous solutions in neonates (40). Therefore, understanding the relevant characteristics of the intended patients is a critical component of risk assessment.

The presence of ischemic tissues probably poses an independent risk and should be considered if the product—such as parenteral antibiotics—is intended for use in individuals with crush injuries, burns, significant peripheral vascular disease, or other conditions involving perfusion anomalies. Lehr et al. (41) systemically injected three different 1 g preparations of an antibiotic containing known particulate amounts into hamsters and measured capillary perfusion. A loss of capillary perfusion due to particle or microsphere injection was observed and found to be dependent on the extent of ischemia-/reperfusion-induced muscle injury, with more capillaries lost in the more severely compromised muscle areas. The authors surmised that particle contaminants may not pose a major threat to tissue that is intact, but it may severely compromise tissue perfusion in patients with prior microvascular compromise.

Finally, the route, volume, frequency, and duration of administration will affect the type and potential severity of any untoward responses. For example, daily high-volume continuous infusions have been estimated to contain more than a million particles that are >2 μm in size (2). The route and frequency of administration may also impact the risk of sensitization. Intradermal, subcutaneous, and inhalational routes of administration are generally linked with increased immunogenicity, compared to intramuscular and intravenous (IV) routes, with an IV route considered the least likely to elicit an immune response, while intermittent dosing may be more immunogenic than continuous infusions (36).

As previously noted, the above discussion is not intended as an exhaustive review; rather, it is meant to illustrate the basic elements that should be incorporated into the risk assessment—the source, particle factors, and patient characteristics. A number of excellent resources are available articulating the essential considerations for each of these components (2, 3).

Conducting the Assessment

The clinical risk assessment should be a signed cGMP-compliant report. The evaluation should include the relevant product, patient, and particle factors considered in the assessment; the rationale supporting the criticality determination; and, the risk-classification conclusion, if relevant. These report elements are necessary to enable auditors, investigators, and others to understand the factors considered supporting the original assessment.

Critical quality attributes, including particles, should be assessed at appropriate intervals tied to the continuous quality improvement plan. Assessments should be reevaluated at critical development milestones, during the periodic review of quality and manufacturing protocols where the assessment was used as a component of the control strategy (e.g., line clearance for broken vials) and line extensions providing for new treatment indications.

As an understanding of process and analytic variability is better established and patient data from clinical trials (CT) can be assessed, these data should be incorporated into the evaluation. Ideally, risk reviews should be conducted, at a minimum, on the following occasions.

First and foremost is an a priori risk assessment conducted during early development activities and updated though pivotal clinical studies when product attribute acceptance criteria and defect classifications are established. ICH Q8 (Pharmaceutical Development) (4), Q9 (Quality Risk Management) (5), and Q10 (Pharmaceutical Quality Systems) (6) provide a structured way to define product-critical quality attributes (CQAs), the design space and ultimately the manufacturing process and control strategies necessary to produce reliably consistent pharmaceutical products. An early formative assessment of the particulate burden–risk to patient safety can be utilized to better inform the manufacturing control strategy employed for that product, support the safety assessment surrounding acceptable common-cause variability within the design space, and the subsequent adoption of particulate matter acceptance criteria and defect-level classification. That, in turn, can help guide the utilization of informative orthogonal analytics (42, 43) and post-marketing monitoring activities to be incorporated into the risk management plan. This assessment provides the initial opportunity to consider the potential impact of intrinsic, stability-indicating particles, particles inherent to the formulation as well as subvisible particles. Conducting assessments pre-commercialization provides the better opportunity to identify if batch-related differences in the type or number of particles (or other product attributes) translate into a quantifiable difference in safety, efficacy, or immune-related outcomes. Clinical trial protocols and associated safety monitoring plans may be designed or adapted to identify product attribute-related and clinically relevant drug-event combinations.

In addition, this early assessment is critical to establishing the appropriate defect-level classification, as setting a defect classification too high unnecessarily consumes resources—both human and fiscal—that might be better utilized on higher priority activities. By the same token, underestimating the potential clinical implications of particulate matter for the product can unnecessarily impact the therapeutic benefit–risk calculus. Therefore, reassessing the available clinical data relative to batch-specific CT material product attributes at various milestone reviews (for example, when moving from Phase 2 to pivotal CT drug substance manufacturing) can be used to identify potential clinical indicators for purposeful observation and incorporated into the safety monitoring program.

Clinical input should also be used as an essential component of the control strategy and inspection plan for “difficult-to-inspect” products. These products, by the nature of their formulation (e.g., viscous, lyophilized, and suspensions) or their container (e.g., bag, amber glass etc.), present unique challenges for visual inspection. (18) Therefore, in order to support the control strategy that would be implemented, the clinical assessment is instrumental in the criticality evaluation since the upstream control stratagem will serve as a decisive factor determining the particulate burden in the final product.

A second opportunity to incorporate clinical input is during the creation of the defect library utilized for visual inspection. Confirmation that a rejected unit shares morphologic characteristics with the library sample may speak to the fact that subsequent findings of this type of material were not unexpected. However, the potential patient risk related to the presence of the particle is not answered by the mere fact that it has been previously catalogued. A clinical review of the defect library also provides a chance to engage clinicians in the manufacturing and inspection qualification processes and gain an understanding of line-related manufacturing and inspection capabilities.

A critical caveat: A prospective opinion should not provide for the acceptability of extrinsic particles. Extrinsic particles are foreign to the manufacturing process, are exogenous in origin, and include hair, non-process-related fibers, insect parts, and similar inorganic and organic materials. They should be rejected during the manufacturing process and if later identified, assessed for a potential increased risk of microbiologic contamination, leachates, or interaction with the drug substance.

A third, pre-commercial activity involves the assessment of intrinsic stability-related particles, aggregates, or agglomerates. These may form in a biotherapeutic as a function of time, temperature, or interaction with the primary packaging or other intrinsic materials. In a similar fashion, an understanding of particles that might be generated “in-use” during administration via a prefilled syringe (44), auto-injector, nebulizer, etc. could be beneficial in assessing any impact to patient risks. Providing characterization data in addition to the size and number of particles is essential to developing an informed clinical assessment. For these reasons, one could provide a strong argument that assessments for complex, biotherapeutic products should be conducted at more frequent intervals than small-volume chemical entities.

Two additional circumstances span the product life cycle: first, an important example involves the “one-off” batch-related safety question presented when particles are identified via an external complaint, regulatory inquiry or during the inspection of retention or stability samples. A clinical assessment is critical, as it involves product in distribution and available for patient use. It is therefore imperative that the clinician have full visibility to the investigation and suspected root cause to assess the implications to the batch benefit–risk profile.

Second, an assessment should be conducted when relevant process changes are adopted with the potential to impact the control strategy for particulates or aggregates, and when environmental stress or stability data identify a change in the subvisible population or the formation of stability-indicating particles. This assessment provides a critical opportunity to review the product's pharmacovigilance Risk Management Plan (RMP) and determine if the RMP provides for adequate pharmacovigilance measures or if it should be updated due to the proposed change. Consideration can be given to adopting a surveillance program within the routine pharmacovigilance plan as one method to proactively monitor post-change events.

Differentiating Clinical Risk Assessment and Product Risk Acceptance for Visible Particles

The clinical risk assessment should be viewed as only one component of the evaluation conducted to establish defect-criticality criteria. Other elements of that assessment would generally include an appraisal of the potential regulatory implications, cGMP and internal quality standards and expectations, and the risk to the product or company “brand.” This is where differentiating “risk assessment” from “risk acceptance” for visible particles is especially important.

Too often, a presumption is offered—unstated or otherwise unsupported—suggesting that an enhanced visual inspection method or a more refined inspection or statistical sampling plan calculus should be utilized because “it improves patient safety.” (45) Although the detection or rejection rates might well be enhanced, does the resulting delta translate into improved patient outcomes? Probably not. Rare macroscopic particles in injectables pose a remote risk of harm to patients.

Intuitively, offering these proposals “for the patient” elicits an affirmative and visceral effect. However, in the case of pharmaceutical science, intuitive theories can be misleading (10). Intuitive thinking contends that visible particles in injectables are inherently unsafe. And on that basis, extraordinary resources are expended on controlling visible particles, rejecting defective units and recalling entire batches of drug product for isolated findings. It is the presumption that the visible particles are unsafe—based upon one's intuition—that drives the subsequent actions. At the same time, batches are routinely released that carry a subvisible (or invisible in the case of difficult-to-inspect products) particle burden that might ultimately provoke consequential pathophysiologic effects. With quality-compliant manufacturing capabilities, the patient safety issues inherent in parenteral products generally arise from what we don't see, or, that the observed attributes are not identified as clinically important—both factors imposing constraints on any attempts to mitigate patient-related risk. The appropriate question is not whether the presence of macroscopic particles is unsafe, but rather, whether their presence sufficiently perturbs the benefit–risk profile of the drug product that manufacturing and inspection activities should be adjusted.

As a result, it is important that one should not overstate the risk to the patient in order to rationalize elevating the defect classification. From the patient risk perspective, the presence of rare, sterile, extraneous, macroscopic particles produced in a cGMP-compliant environment, for example, in products intended for subcutaneous injection, would not be viewed clinically as a critical defect. Because the presence of visible particles is probabilistic, attributing the basis of the criticality level to patient safety (i.e., elevating patient risk to serve as a proxy for one of the other elements discussed) will lead to a less informed approach in assessing product complaints or regulatory inquiries. However, within the manufacturer's quality culture or understanding of customer expectations or requirements, the final defect classification might be elevated to satisfy those provisos. It is important, however, that the rationale for the classification should be clearly articulated and documented.

When conducting a clinical risk assessment for particulate matter, it is essential to recognize that each product is different due to a number of factors, such as different formulation, container-closure, delivery system, route of administration, intended patient population, therapeutic indication, and duration of therapy. Assessing the clinical considerations should be conducted first. And if the sum of this assessment provides that the patient risk falls to the higher end of the risk-spectrum (i.e., neonatal population, administered via intra-umbilical artery, or ischemic tissue present, etc.), then one should resource the control strategy appropriately, classify visible particles as “critical,” and adopt an adjusted inspection and sampling plan. In these rare instances, the manufacturer should also evaluate the contemporaneous impact of subvisible particles and identify compatible in-line filter(s) for bedside application to mitigate the risk of particle administration.³

In conclusion, understanding the product quality attributes that hold the potential to impart a greater impact to the product's benefit–risk profile will provide an opportunity to mitigate product-related risks through a targeted risk and science-based control strategy and adaptive process capability or to inform via labeling and other risk-minimization activities. Differentiating risk assessment and risk acceptance is key to building a rational, quality system. The essential point is that the ultimate attribute-risk classification should be predicated on the appropriate evaluation category (safety, regulatory, and customer requirements) so informed decisions can be made in resourcing the relevant control strategy.

Subvisible Particles

To avoid introducing unnecessary confusion concerning the visible particle assessment process—a separate commentary would serve as a better vehicle to offer a proposed approach to evaluating the clinical risks related to subvisible particles, and so a comprehensive review of subvisible particles falls beyond the intended scope of this paper. The primary objective of this commentary is to outline one paradigm that might be utilized to assess the potential patient-risk implications from visible or macroscopic particles in injectables. Nevertheless, as noted at points throughout this discussion, it is likely in most instances that exposure to subvisible particles poses a greater risk than macroscopic particles, as subvisible particles will be injected or infused with every administration (3) and therefore a comprehensive clinical assessment should cover both visible and subvisible particles. Subvisible particles are evaluated using the methods and criteria found in USP <787> (16), <788> (9), and <789> (17). USP <788> provides subvisible particle acceptance limits that are generally utilized to establish parenteral drug product specifications.

Although the general questions one might posit in considering subvisible particle risks include some similarities to those considered in assessing visible particles, the risk-related outcomes are different enough that a separate analytic model should be utilized. As parenteral therapy has evolved to include an expanse of complex protein molecules, relying on the subvisible particle acceptance limits provided by the harmonized pharmacopeia should be further augmented by several factors such as a greater understanding of the product attributes, interactions with the primary container materials, impact of environmental stresses, and the intended patient population. Additionally, all current subvisible assessment methods necessitate destructive sample testing compared to non-destructive inspection methods for visible particles.

Conflict of Interest Declaration

The author is a shareholder of Eli Lilly and Company and declares that he has no other competing interests.