Most particulate matter recalls are not surprises. They are the visible end of a chain that started months earlier, usually at the glass vial itself, and the quality system either had the controls in place to catch it or it didn't. That is the lesson worth sitting with before we get to any specific case: a sterile injectable that ships with a piece of glass floating in it did not fail at the moment someone spotted the flake. It failed earlier, at container qualification, at inspection validation, or at the release testing bench, and the visible particulate was just the last domino.
On July 24, 2026, American Regent, Inc. announced a voluntary nationwide recall of a single lot of Papaverine Hydrochloride Injection, USP, 60 mg/2 mL, after particulate matter identified as glass was found in the product. I am not going to spend this article relitigating that specific event, and I'd encourage you to read the FDA recall notice directly if you want the particulars. What I want to walk through is the system of controls, both regulatory and operational, that exists specifically to prevent glass particulates from reaching a patient in the first place. Because this is not a rare or exotic failure mode. Glass particulate contamination in parenterals is one of the most recurring and most preventable categories of sterile injectable recalls in FDA's history, and the standards for catching it have existed, in detailed form, for over a decade.
Where Glass Particulates Actually Come From
There are really only three pathways a glass particulate takes into a finished vial, and each one has a different point of intervention.
The first is delamination, sometimes called lamellae formation. Certain drug formulations, particularly those with high pH, citrate or phosphate buffers, or extended dwell time in contact with the glass, can chemically attack the inner surface of a borosilicate vial. Thin flakes of glass, called lamellae, detach from the wall over the shelf life of the product. This is insidious because the vial can pass 100 percent visual inspection at time of fill and still shed particulates six, twelve, or eighteen months later while sitting in a warehouse or a pharmacy refrigerator. FDA became acutely aware of this in 2010 and 2011, when a cluster of injectable recalls, more than a dozen products across several manufacturers in a two-year span, were traced back to glass lamellae rather than any failure during manufacturing itself. FDA responded with an industry advisory on the issue, and USP subsequently developed General Chapter <1660>, "Evaluation of the Inner Surface Durability of Glass Containers," specifically to give manufacturers a validated method for screening vial-formulation combinations before they ever reach commercial production.
The second pathway is mechanical: cracked, chipped, or scored vials introduced during filling, capping, or transport, where fragments shed at the point of stress rather than through a slow chemical process. This is more classically a manufacturing-line problem, tied to filling equipment settings, vial handling, and incoming component inspection.
The third is contamination from an external source, glass debris introduced during compounding or from a broken container elsewhere on the line, which is really an environmental and procedural control failure rather than a container-chemistry problem.
Each of these has a distinct root cause, and a distinct set of preventive controls, which is exactly why lumping every particulate deviation into a single generic CAPA is how root cause investigations go sideways.
| Root Cause | Typical Detection Point | Preventive Control |
|---|---|---|
| Glass delamination (chemical attack on vial wall) | Late in shelf life, often post-distribution | USP <1660> container-formulation compatibility screening; accelerated and real-time stability with particulate testing |
| Mechanical cracking or scoring | At fill or during line inspection | Incoming component inspection (AQL sampling per ANSI/ASQ Z1.4); vial handling equipment qualification |
| External glass contamination | Line clearance or in-process checks | Environmental monitoring; line clearance procedures; broken-glass response SOPs |
| Inadequate inspection sensitivity | Should be caught at release, often isn't | Probability-of-detection studies per USP <1790>; semi-automated or automated inspection qualification |
The Regulation Nobody Reads Until It's Too Late
Here is the part most quality teams underestimate: particulate matter limits for injectables are not FDA guidance in the soft sense. They are compendial standards with the force of law. Under section 501(b) of the Food, Drug, and Cosmetic Act, a drug that is labeled as meeting USP standards is deemed adulterated if it fails to conform to those standards, regardless of whether FDA ever pulls a single vial for testing. That is the legal hook connecting USP General Chapters <788> and <790> to enforcement, and it is why a particulate excursion is never just a quality event. It is a statutory adulteration question the moment the product carries a USP designation.
21 CFR 211.167(a) requires that for each batch of drug product purporting to be sterile and/or pyrogen-free, there be appropriate laboratory testing to determine conformance to such requirements, with written test procedures that are followed and documented. FDA has long interpreted this to include particulate matter testing against the applicable USP chapters as part of a complete release testing package for parenterals. Layer on 21 CFR 211.94, which requires that drug product containers and closures not be reactive, additive, or absorptive so as to alter the safety, identity, strength, quality, or purity of the drug, and you have the specific regulatory basis a delaminating glass vial violates the moment it starts shedding material into solution. And 21 CFR 211.165(a) requires that each batch be tested and meet its final specifications, including any particulate matter specification, before release for distribution.
USP <790>, "Visible Particulates in Injections," became official on August 1, 2015, and it sets a straightforward but demanding bar: injections must be essentially free from visible particulates when examined under suitable conditions of visibility, and every unit produced is subject to inspection for this attribute, not just a sample. That last point trips people up. Non-visible-particulate attributes can be released on an AQL sampling plan under ANSI/ASQ Z1.4. Visible particulates, under <790>, cannot. The chapter effectively requires 100 percent inspection, whether by trained human inspectors or a validated automated system.
USP <788>, "Particulate Matter in Injections," governs the subvisible range, particles too small to see by eye but large enough to matter clinically, using light obscuration or microscopic particle count methods. For a small-volume injection like a 60 mg/2 mL vial, the limits under Method 1 are no more than 6,000 particles per container equal to or greater than 10 micrometers, and no more than 600 particles per container equal to or greater than 25 micrometers. USP <1790>, "Visual Inspection of Injections," is the companion informational chapter, official since December 1, 2017, and it is the one most quality units skip reading even though it contains the practical expectation FDA investigators actually look for during inspections: a documented probability-of-detection study proving your inspection method can reliably find defects of the size and type your product is prone to, not just a procedure that says "inspect visually."
| Chapter | Scope | Method | Key Requirement |
|---|---|---|---|
| USP <790> | Visible particulates (down to ~50-100 micrometers, human-eye detectable) | 100% visual inspection, manual or automated | Essentially free from visible particulates; every unit inspected |
| USP <788> | Subvisible particulates | Light obscuration (Method 1) or microscopic particle count (Method 2) | ≤6,000 particles/container ≥10 µm; ≤600 particles/container ≥25 µm (small-volume injections) |
| USP <1790> | Inspection method design and qualification | Probability-of-detection studies, defect libraries, inspector qualification | Documented evidence the inspection method actually detects the defects it's meant to catch |
| USP <1660> | Glass container inner-surface durability | Accelerated extraction and delamination-risk screening | Formulation-container compatibility confirmed before commercial use |
Published studies of manual visual inspection performance have consistently found that individual human inspectors detect small or low-contrast particulates well below the 90 percent probability FDA expects a validated method to demonstrate, which is precisely why <1790> pushes manufacturers toward documented detection-probability studies instead of an unvalidated assumption that a trained eye is enough.
Building an Inspection Program That Actually Catches This
If I'm auditing a sterile injectables line, the particulate matter section of my checklist has four questions, in this order.
First, was the glass qualified for this specific formulation before it went into commercial use? A vial type that performs fine with a neutral saline formulation can delaminate with a high-pH or citrate-buffered one. USP <1660> testing should exist for the specific vial-formulation-closure combination in production, not a generic assertion that "this is USP Type I glass."
Second, does the inspection method have a documented probability-of-detection study behind it, covering the specific defect types and sizes relevant to this product, including the smallest lamellae fragments the formulation has historically produced? A procedure that says "100% visual inspection performed" without a POD study attached is a paper compliance answer, not a real one.
Third, is the visible particulate check actually applied to 100 percent of units, with a defect library that inspectors or automated systems are qualified against, and documented reject-rate trending over time? A creeping reject rate on a specific defect category is often the earliest signal of a developing container or process issue, well before it becomes a customer complaint.
Fourth, does stability testing include particulate matter as a tested attribute at meaningful time points across shelf life, not just at release? Delamination is a time-dependent failure. A batch that passes <790> and <788> at time zero can still fail at month twelve if nobody is watching for it on stability.
A quality unit that can answer yes to all four of these has done the actual prevention work. A quality unit that can only point to a release-testing SOP has done the minimum paperwork and is one delamination-prone formulation change away from a recall.
What the Investigation Should Look Like When Particulates Turn Up Anyway
Even a well-controlled program will occasionally find a particulate, and how the investigation is conducted matters almost as much as the prevention controls upstream. 21 CFR 211.192 requires a thorough investigation of any unexplained discrepancy or failure of a batch to meet specifications, extending to other batches of the same product and other products that may have been associated with the specific failure. For a glass particulate, that means the investigation cannot stop at "one vial had a flake in it." It has to trace back to the specific glass lot, the specific fill-line time window, whether other lots filled with the same glass lot are affected, and whether the formulation itself has a delamination history that stability data would reveal.
This is also where CAPA effectiveness checks earn their keep. A corrective action that amounts to "retrained inspectors" without addressing whether the inspection method has adequate detection probability for the defect found, or without a formal <1660> risk assessment on the glass-formulation pairing, is not going to hold up under a subsequent FDA inspection, and it is not going to prevent a repeat event either.
A Short Compliance Timeline Worth Keeping on the Wall
USP <790> has been official and enforceable since August 1, 2015. USP <1790> has been the informational companion since December 1, 2017. USP <1660> for glass delamination risk assessment has been available since 2013, following directly from FDA's 2011 industry advisory on lamellae formation. None of this is new guidance. If your injectable product line has been in commercial production through any of these dates without a documented <1660> assessment or a POD-backed inspection qualification, that is a gap worth closing before an inspector or, worse, a patient finds it first.
Firms building or refreshing a GMP quality system for parenteral manufacturing are usually better served having an outside set of eyes stress-test the inspection and release program against these specific chapters rather than assuming general cGMP training has covered it. I work with manufacturers on exactly this kind of GMP gap assessment before it becomes a recall conversation, and separately on mock recall readiness exercises that test whether the traceability and investigation muscle described above actually works under time pressure, not just on paper.
FAQ
Does USP <790> require 100 percent inspection of every single vial, or can I sample? Yes, 100 percent inspection is required for the visible-particulate attribute specifically. Other quality attributes may be released on an AQL sampling plan under ANSI/ASQ Z1.4, but USP <790> does not permit sampling for visible particulates. Every unit has to be examined, whether by qualified human inspectors or a validated automated inspection system.
What's the actual difference between USP <788> and USP <790>? USP <790> covers particulates large enough to see with the human eye under proper lighting and background conditions. USP <788> covers subvisible particulate matter, measured by light obscuration or microscopic particle count, with numeric limits per container based on particle size and container volume. A product has to pass both; they are not interchangeable or substitutable tests.
Is glass delamination a manufacturing defect or a formulation problem? It's fundamentally a container-formulation compatibility issue, not a manufacturing line error. Certain formulations, especially high-pH or buffered ones, chemically attack the inner glass surface over time. That's why USP <1660> testing needs to happen during development or any formulation change, not just as a manufacturing quality check.
How do I know if my visual inspection method is actually good enough? You need a documented probability-of-detection study, per USP <1790>, that demonstrates your specific inspection method reliably finds the defect types and sizes your product is prone to. A written SOP describing the inspection process is not the same thing as evidence the method works.
What triggers a nationwide recall versus a market withdrawal for a particulate finding? That determination depends on FDA's risk classification of the specific defect and product, which considers factors like the route of administration, the potential for patient harm from the specific particulate type and size, and how widely the affected lot was distributed. Any confirmed particulate finding in a distributed lot should trigger a 21 CFR 211.192 investigation regardless of how it's ultimately classified.
Last updated: 2026-08-02
Jared Clark
Principal Consultant, Certify Consulting
Jared Clark is the founder of Certify Consulting, helping organizations achieve and maintain compliance with international standards and regulatory requirements.