An expiry date looks precise.
Month.
Year.
Done.
But the date is not the control.
It is the claim.
The control is everything underneath it: formulation knowledge, test methods, stability, packaging, warehousing, temperature data, excursion response, investigations, and the quality decision.
If those pieces do not work together, the date on the carton can look more certain than the evidence behind it.
WHAT FDA SAID
On September 22, 2026, FDA posted a warning letter to kdc/one Chatsworth, Inc. The letter was dated September 8 and followed a March-to-April 2026 inspection of the company's California drug manufacturing facility.
FDA said the company manufactures over-the-counter drug products.
FDA said the firm had not established scientifically sound finished-product specifications and did not include testing for a relevant contaminant as a release or stability requirement. The letter describes known degradation and reaction pathways through which the contaminant could form and increase over the product's shelf life.
FDA also said the firm did not adequately evaluate the U.S. impact of a contamination event that led a customer to recall products in Canada, even though products distributed in the United States and Canada used the same formulation, packaging, and labeling.
The letter describes inadequately investigated viscosity OOS results from stability testing.
It also says finished drug products were stored above labeled storage conditions and above the firm's internal temperature-monitoring limits. FDA said the response did not include an adequate retrospective assessment of the effect on distributed product, including the potential formation of impurities and degradants during heat exposure.
Those are the source facts.
Here is the operator lesson:
A shelf-life claim has to survive the warehouse.
STABILITY IS NOT A LABORATORY ISLAND
Stability teams can run a technically sound program and still be disconnected from the conditions the product experiences after release.
That is the trap.
The chamber is mapped.
The timepoints are scheduled.
The methods are approved.
The reports are filed.
Meanwhile, the warehouse has excursions.
Distribution has delays.
Packaging changes.
An OOS result gets explained as a documentation problem.
Each system may look manageable by itself.
The product experiences all of them together.
That is why stability cannot be treated as a laboratory calendar.
It is a lifecycle control.
THE WAREHOUSE IS PART OF THE STABILITY PROGRAM
Finished goods do not stop aging when they leave the manufacturing area.
They sit.
They move.
They wait for shipment.
They experience weekends, equipment failures, hot weather, open dock doors, staging delays, and the occasional sensor nobody noticed until Monday.
The warehouse is not just a logistics space.
It is part of the product's real stability environment.
That means storage controls need more than a temperature range in an SOP.
They need:
• Qualified storage areas
• Monitoring that represents actual product locations
• Alert and action limits with clear ownership
• Escalation that works after hours
• Defined product-impact assessment requirements
• Traceability from the excursion to every potentially affected lot
• Quality authority over disposition
• Evidence that the stability program can support the decision
A sensor alarm is not the end of the process.
It is the start of a quality question.
AN EXCURSION REVIEW HAS TO REACH THE PRODUCT
It is easy to close a temperature excursion at the facility level.
The HVAC was corrected.
The sensor returned to range.
The alarm is cleared.
Maintenance closes the ticket.
But the product question is still open.
Which lots were present?
Where were they located?
How long were they exposed?
What temperature did they experience?
What degradation pathways matter for those formulations?
Does the available stability evidence cover the excursion?
Could the packaging or container-closure system change the risk?
Are retain samples available and meaningful?
Does testing have enough sensitivity to answer the question?
If the excursion record never reaches those questions, the building may be back in control while the product assessment remains incomplete.
That is not closure.
That is a maintenance repair wearing a quality badge.
KNOWN CHEMISTRY HAS TO BECOME ROUTINE CONTROL
FDA's letter describes known pathways through which a contaminant could form and increase over shelf life.
That matters because known risk has to move into routine control.
Once the organization understands that a formulation, ingredient combination, temperature condition, or time period can create a quality risk, the information cannot stay inside a technical memo.
It has to change something.
The specification.
The method.
The stability protocol.
The warehouse limit.
The supplier requirement.
The packaging decision.
The investigation trigger.
The market-action assessment.
Knowledge that does not change control is just an interesting document.
Pharma has plenty of those already.
OOS RESULTS ARE NOT ISOLATED LAB EVENTS
The warning letter also describes viscosity failures during stability testing that were not adequately investigated.
An OOS result on stability is not only a laboratory issue.
It may be a method issue.
It may be a specification issue.
It may be a manufacturing issue.
It may be a packaging issue.
It may be a storage issue.
It may be the product telling you that the shelf-life claim is too optimistic.
The investigation has to follow the evidence across functions.
If the laboratory cannot establish a conclusive assignable cause, the review has to move outward into manufacturing history, raw-material variability, process capability, packaging, complaints, storage data, and other affected batches.
The goal is not to rescue the result.
The goal is to understand the product.
A FOREIGN RECALL IS STILL YOUR SIGNAL
FDA's letter says the firm knew a customer had recalled products in Canada because of potential contamination but did not adequately determine whether the recalled batches also entered U.S. distribution or evaluate the potential risk to U.S. consumers.
Quality signals do not respect market boundaries.
If the same formulation, process, packaging, or batch can reach more than one market, a signal in one region is not someone else's paperwork.
It is your signal.
The first questions should be simple:
• What is shared?
• What lots are affected?
• Where did they go?
• What evidence do we have?
• What action protects patients while we learn more?
The market map should already exist.
The escalation path should already exist.
The quality unit should not have to build either one during the event.
THE EXPIRY DATE HAS TO BE DEFENSIBLE
A shelf-life claim should connect the product's science to its real operating environment.
Leadership should be able to trace the claim through stability-indicating methods, relevant specifications, representative batches, marketed packaging, storage and distribution conditions, excursion history, OOS and complaint investigations, product-quality review, and quality-unit decisions.
If the evidence breaks at the warehouse door, the claim is not complete.
If an excursion cannot be connected to affected lots, the claim is not complete.
If known degradation risk never reached the testing program, the claim is not complete.
If a stability OOS is explained without understanding the product, the claim is not complete.
The date on the carton is the final line.
The real work is everything that makes that line true.
Clarity starts when stability, warehousing, supply chain, laboratory, and quality are looking at the same product story.
PRIMARY SOURCE
FDA Warning Letter to kdc/one Chatsworth, Inc., MARCS-CMS 733205, September 8, 2026: