A clear vial can look perfect.
No particles are visible. The seal is intact. The label is clean. A sterility result may even show no viable microorganisms under the test conditions.
And the product can still have an endotoxin problem.
That distinction moved from laboratory theory to current news this week. On September 13, 2026, the U.S. Food and Drug Administration posted a nationwide recall from Centric Compounding involving six lots of three injectable products: Glutathione 200 mg/mL, Myer’s Cocktail, and Tri-Immune Boost. According to the recall notice, the products had been compounded with a glutathione active pharmaceutical ingredient, or API, that contained elevated endotoxin levels.1
The company reported nine adverse-event reports involving fever, chills, chest pain, nausea or vomiting, headache, and/or malaise. FDA’s notice says injectable products with elevated endotoxin may cause significant reactions, potentially including fever, low blood pressure, inflammatory reactions, anaphylactic shock, and death.1
This is an important safety story. It is also a remarkably clear quality lesson.
The common issue did not begin with three unrelated finished-product labels. It began upstream, with an ingredient shared across those products. That makes the recall a useful case study in how pharmaceutical quality actually works: from supplier selection and raw-material verification through process control, finished-product testing, and lot traceability.
The central idea is simple:
A finished vial inherits the history of every material used to make it.
What FDA Recalled
The recall involved 30 mL multidose vials distributed nationwide to patients through home delivery and to healthcare prescribers. FDA listed the following affected lots.1
| Product | Lot number | Beyond-use date |
|---|---|---|
| Glutathione 200 mg/mL | 062326:B | August 7, 2026 |
| Glutathione 200 mg/mL | 070826:A | August 22, 2026 |
| Glutathione 200 mg/mL | 072726:D | September 25, 2026 |
| Myer’s Cocktail | 070226:D | August 16, 2026 |
| Myer’s Cocktail | 072126:A | September 4, 2026 |
| Tri-Immune Boost | 071626:B | August 30, 2026 |
The company announcement identifies glutathione API containing elevated endotoxin as the reason for the recall. It does not publish a completed root-cause investigation, so it would be inappropriate to speculate about exactly where or how the endotoxin entered the material.
What the public record does establish is enough to understand the quality pattern. One ingredient was used in several finished products. When that ingredient presented an endotoxin concern, the response had to follow it downstream into multiple lots.
Imagine a bakery using one shipment of flour in bread, muffins, and cake. If the flour is later found to be unsuitable, the problem is not limited to one recipe. Investigators need to know which bags of flour were used, on which days, in which products, and where those products went.
Pharmaceutical traceability serves the same purpose, with much higher stakes.
What Is Bacterial Endotoxin?
FDA describes bacterial endotoxins as lipopolysaccharides associated with the outer membrane of gram-negative bacteria.2 The scientific name matters, but the everyday idea is easier:
Endotoxin is a specific piece of bacterial material that can remain after the bacterial cell is no longer alive.
That is why the words sterile and endotoxin-free are not interchangeable.
Sterility testing asks whether viable microorganisms are present under defined test conditions. Bacterial-endotoxin testing asks whether specific bacterial remnants are present above an acceptable level.
A useful analogy is a campfire. Sterility testing asks whether flames are still burning. Endotoxin testing asks whether irritating ash remains after the flames are gone. Putting out the fire does not make the ash disappear.
FDA’s technical guide explains this directly. It notes that sterilizing or microbiological filters do not necessarily remove endotoxin and that reducing a high microorganism count does not produce a matching reduction in endotoxin.2
That means both of these results can be true at the same time:
| Laboratory question | Possible result |
|---|---|
| Are viable microorganisms detected under the sterility test conditions? | No |
| Is bacterial endotoxin present above the product’s allowable limit? | Yes |
There is no contradiction. The tests are looking for different things.
This is the first major lesson from the recall: one reassuring result cannot answer a question it was not designed to answer.
Why Endotoxin Matters More for an Injectable Product
The route of exposure changes the quality question.
An injectable product bypasses many of the body’s usual external barriers. That is why bacterial-endotoxin limits are tied to factors such as the route of administration and the maximum dose rather than one universal number for every product.2
FDA’s current guidance points manufacturers to USP Chapter <85>, Bacterial Endotoxins Test, and related standards. It describes established gel-clot, photometric, and kinetic approaches and recommends that appropriate components and finished products be tested for pyrogens and endotoxins.3
The public does not need to memorize those method names. What matters is the principle: an endotoxin result is only meaningful when the method is suitable for that product, the sample is representative, the dilution is appropriate, and the acceptance limit fits the product’s intended route and dose.
A number without that context is not a complete quality conclusion.
The Problem Can Begin Upstream
The phrase active pharmaceutical ingredient sounds as if it refers only to what the drug is supposed to do. In quality systems, the API is also a physical material with a history.
It has an original manufacturer. It may pass through a repackager. It carries a lot number, specifications, shipping records, storage history, test methods, and a stated intended use. It may later be divided, combined, or used across several finished formulations.
FDA’s supplier-quality page tells compounders to know their bulk-drug and excipient suppliers, understand what testing those suppliers perform, confirm that materials are suitable for use in drugs, use valid certificates of analysis, and maintain supply-chain transparency.4
FDA’s older technical guide identifies several possible endotoxin sources in parenteral manufacturing, including water, raw materials, chemicals, equipment, and packaging components. It specifically notes that when a raw material was the source of a pyrogen problem in certain injectable-product investigations, the active drug substance was a recurring concern.2
The practical message is not that every API is dangerous. It is that finished-product quality cannot be separated from incoming-material quality.
If a shared ingredient has a problem, that problem can follow the ingredient into every finished lot that used it. Conversely, accurate lot genealogy allows a firm to define the affected scope quickly instead of recalling blindly.
A traceable system should be able to answer five questions without guesswork:
| Question | Why it matters |
|---|---|
| Who originally manufactured the material? | Identifies the true source, not only the last seller or repackager. |
| Which incoming lot was received? | Connects documents and testing to the physical material. |
| Which finished batches used it? | Defines the downstream impact. |
| What testing was performed, by whom, and with which method? | Shows what the evidence actually supports. |
| Where were the finished lots distributed? | Enables a targeted, timely response. |
This is how a laboratory result becomes part of a functioning quality system rather than an isolated number.
Why a Supplier CoA Is Not the Finish Line
A certificate of analysis is important. It may report identity, assay results, specifications, lot information, and other attributes. It can help connect a shipment to a documented set of tests.
But a CoA is still a document. Its value depends on the integrity of the system behind it.
A strong receiving organization asks whether the supplier is qualified, whether the named manufacturer is transparent, whether the ingredient is intended for pharmaceutical use, whether the specifications fit an injectable application, whether the methods are suitable, and whether the material received matches the documented lot.
A 2019 FDA glutathione investigation demonstrates why those questions matter. FDA reported adverse events in patients who received compounded injectable products made with glutathione powder that had been labeled as a dietary supplement and was not intended for use in sterile drugs. FDA laboratory testing found excessive endotoxin in sampled powder, with some results as high as five times the limit calculated for the administered dose.5
FDA’s conclusion was not that testing is unnecessary. It was that testing alone should not be relied upon to create quality after the fact. Ingredients used in sterile injectable drugs should be manufactured under conditions and specifications appropriate for that use.5
That distinction is essential.
A CoA is one link in the evidence chain. Quality is the whole chain.
What a Defensible Endotoxin Result Requires
It is tempting to imagine endotoxin testing as a strip that turns red or green. Real testing requires more judgment.
A sample can interfere with the assay. Its acidity, alkalinity, color, turbidity, viscosity, or chemical components may inhibit or exaggerate the reaction. FDA therefore emphasizes method suitability and recovery studies to show that the assay can detect endotoxin in the presence of the product.2
Dilution can sometimes reduce interference, but there is a limit. FDA warns that excessive dilution can push a harmful endotoxin concentration below the method’s detectable range, making a negative result meaningless.2
Sampling also matters. The March 2026 FDA guidance says firms should consider possible contamination in raw materials, in-process materials, and finished products. Sampling plans should account for process consistency, hold times, endotoxin-removal steps, and release specifications.3
Even pooling samples requires care. Combining aliquots from several units can be efficient, but a high-endotoxin unit may be diluted by lower-endotoxin units. FDA therefore discusses adjusted dilution limits and generally suggests pooling no more than three small-volume parenteral units in one composite sample.3
| Quality question | What can go wrong if ignored? |
|---|---|
| Is the method suitable for this product? | The product matrix may hide or amplify the assay response. |
| Is the dilution within the validated range? | Too much dilution can turn a meaningful signal into an apparent negative. |
| Is the sample representative? | A result from the wrong location or stage may miss the actual risk. |
| Does pooling preserve sensitivity? | A failing unit can be averaged into a passing composite. |
| Is the acceptance limit justified? | A generic cutoff may not fit the product’s route or maximum dose. |
This is why competent testing is not merely pressing “run.” It is designing a method and sampling plan that make the result interpretable.
The Layered Control System
The most useful response to a recall is not a promise that one final test will catch everything. It is a system designed to prevent, detect, and trace problems at several points.
Supplier qualification establishes who made and handled the ingredient, whether the source is appropriate, and whether its quality history supports continued use.
Incoming-material controls confirm that the physical material received matches its identity and risk-based specifications. The test plan should fit the material and its intended use.
Process controls monitor places where endotoxin can enter or accumulate, including water, equipment, components, hold times, and bioburden before sterilization.
Method suitability demonstrates that the product matrix does not prevent the assay from recovering endotoxin accurately.
Finished-product release testing evaluates the final product against a justified endotoxin limit with a scientifically sound sampling plan.
Lot traceability connects every result backward to suppliers and incoming materials and forward to finished batches and distribution records.
Each layer has limits. Together, they reduce the chance that one blind spot becomes the entire decision.
What This Recall Does and Does Not Prove
A current recall can generate broad conclusions that outrun the evidence. The disciplined approach is narrower.
The recall does not establish that every glutathione product is contaminated. It does not establish that every compounded injectable has the same problem. It does not make a CoA worthless. It does not mean sterility testing failed at its assigned job.
It does show why an injectable-product quality system cannot stop at appearance, one test, one certificate, or the finished-vial label.
The recorded issue connected one upstream API to three product types and six lots. FDA’s broader guidance explains why supplier knowledge, incoming-component controls, method suitability, finished-product testing, and traceability belong in the same system.1 4
For clinics, med spas, pharmacies, developers, and informed consumers, the most productive questions are therefore not “Does this look clean?” or even “Does it have a CoA?” Better questions include:
Was the material intended for this use? Was the supplier qualified? Was the received lot verified? Were sterility and endotoxin addressed separately? Was the method shown to work in this product? Can the result be traced to the exact finished lot?
Those questions move the conversation from reassurance to evidence.
The Bottom Line
The recall started before the final vial.
A shared ingredient with elevated endotoxin moved into several compounded injectable products, and lot traceability allowed the affected products to be identified. The event demonstrates why raw-material quality is finished-product quality in waiting.
It also corrects a common misconception. A sterile result and a bacterial-endotoxin result are not competing versions of the same test. They answer different questions. A product can contain no detected viable microorganisms under sterility-test conditions and still contain bacterial remnants above an allowable endotoxin limit.
The solution is not one perfect test. It is an evidence chain:
Know the supplier. Verify the material. Control the process. Prove the method works. Test the finished product. Trace every lot.
At Vanguard Laboratory, the goal is not to turn complex quality questions into vague reassurance. It is to produce defensible analytical evidence about the material and product actually in front of us.
References
- U.S. Food and Drug Administration. “Centric Compounding Issues Nationwide Recall of Glutathione, Myer’s Cocktail, and Tri-Immune Boost Due to Elevated Endotoxin Levels.” Published September 13, 2026.
- U.S. Food and Drug Administration. “Bacterial Endotoxins/Pyrogens.” Inspection Technical Guide No. 40.
- U.S. Food and Drug Administration. Pyrogen and Endotoxins Testing: Questions and Answers, Edition 2. March 2026.
- U.S. Food and Drug Administration. “FDA to Compounders: Know Your Bulks and Excipients Suppliers.” February 4, 2026.
- U.S. Food and Drug Administration. “FDA Highlights Concerns with Using Dietary Ingredient Glutathione to Compound Sterile Injectables.” June 7, 2019.
- United States Pharmacopeia. “USP <85> Bacterial Endotoxins.”