Label expiry is not a stability study

A date printed on a label is the last day a manufacturer supports a specification under the storage they stated. It is not a stability study you performed on your shelf, and it is not permission to keep going. This selection guide helps you decide when that date is the wrong thing to negotiate, and, when a negotiation is even eligible, which evidence would be relevant. It does not authorise use past the date. That decision stays with the laboratory that owns the result.

Water grades illustrate the point sharply, because a certificate about ions is not a certificate about everything else. The split is explained in laboratory water types and where they fail. A buffer's pH is only evidence about pH when the meter is fit, which is preparing a buffer and checking pH. Replacement reagents are specified from the reagents and chemicals catalogue. Vessels are in beakers and flasks. A repurchase, when you choose one, goes on a quotation request with the grade and the storage you can actually provide.

What the printed date is

The manufacturer studied the product under named conditions: a temperature band, a closed container, sometimes protection from light. The date is the end of the period for which they will stand behind the specification. It is support, not a cliff in the laws of chemistry that every bottle obeys on the same morning. Two consequences follow. Before the date, storage that violated the label means you are already outside the claim. After the date, being inside the claim is something you can no longer ask the manufacturer to say. Your own data, if you collect any, are your data. They do not edit the printed date.

A laboratory stability study is a designed set of tests over time, with the properties and the acceptance windows chosen in advance. Looking at a bottle, finding it colourless, and writing "still fine" is not that study. This page will not teach you to imitate one. It will tell you which questions are eligible for a simple control and which questions should be replaced rather than tested.

When extension is the wrong question

Sterility is the first. You cannot see it, and pH, conductivity, and a clear meniscus do not measure it. An expiry on a sterile bottle is partly a statement about a closed, qualified container. Once that container is open, the statement is already about the past. Extending it because the liquid is still colourless imports a claim you have no method for. If the procedure allows you to filter-sterilise or autoclave a particular formula, that is a new process with its own records, not an extension of the old date.

Nuclease control is the second. Water, tubes, and tips sold as nuclease-controlled were tested, or produced under controls, at release. A PCR that happens to amplify is not a test for RNase or DNase. Inhibitors can hide a nuclease, and a hardy amplicon can survive a low one. If the method required the claim, open material that still carries it. Do not annex a passing research PCR as a quality certificate.

A calibration standard is the third. pH buffers and conductivity standards are references. NIST's calibration guidance exists because the standard is the thing you trust in order to trust the meter. An expired standard can look perfect and still sit off the value printed on it. Every buffer you set from that meter would then share the offset. Replace the standard. The slope you get with a fresh set is the control that matters, not a rescue of the old bottles.

Some chemical degradations belong in the same "do not extend" group when you have no assay for the hazard: a peroxide-forming solvent, a bicarbonate that has lost carbon dioxide, a light-sensitive cofactor the method says to protect. If you cannot test the property that actually fails, the date is not a debate. It is a stop.

When a control can be relevant evidence

A simple salt, or a buffer you prepared and labelled yourself, can be remeasured for the properties the use actually needs. The evidence is a control that still passes a window you wrote before you looked. For a working buffer, that might be appearance, pH at the stated temperature on a calibrated meter, and, if ionic strength was the specification, conductivity against a historical band for that recipe. For an enzyme, it is the activity control the method already trusts, run with a retained substrate, and accepted or rejected against the window the method already uses.

Write the window first. A pH of "about 7.4" decided after you see 7.6 is not a window. A defined tolerance, the one the assay was validated with, is a window. Record the lot or the batch, the date of the check, the temperature, and the use you are willing to put the material to. A pass supports a documented laboratory decision for that use. It does not restore a sterile claim, a nuclease claim, or a calibration role the bottle never had.

A fail ends the discussion. Do not average a failing control with a memory of last month. Do not apply a pH pass to a different property. The bottle that still reads the right pH is not thereby free of microbes, and the enzyme that still turns a substrate is not thereby the calibration standard for a pipette.

Select the question in order

Start with storage. If the label said 4 degrees and the bottle spent a hot month on the bench, you are outside the claim even if the printed date is next year. The selection then is whether the property you need can be retested, not whether the ink is still in date.

If storage was respected and the date has passed, ask whether the critical property is sterility, nuclease control, or calibration. If it is any of those, stop. Replace the material. If the critical property is pH, conductivity, appearance, or a functional control you already run, write the window and run that control once. A second property you did not test stays unclaimed.

If several properties matter, the strictest one dominates. A buffer used as a sterile calibrant for a pH meter would be a terrible object to extend, because both sterility and calibration are on the stop list. A non-sterile phosphate buffer used as a wash, with pH as the only requirement, is the case where a meter check can be relevant. The use decides. The nickname on the bottle does not.

A worked selection

A bottle of pH 7 calibration buffer expired last week. The liquid is clear. The selection is immediate: it is a calibration standard, so extension is the wrong question. Order a replacement and do not adjust today's Tris against the old bottle. The cost of a wrong pH is every assay that believes it.

A carboy of buffer you made, never claimed sterile, used to dilute a colorimetric reagent, expired according to your own four-week laboratory label. Storage was the refrigerator the recipe named. The properties that matter are pH at the labelled temperature and the absence of haze. You write the window you already use for that assay, measure, and either accept a documented further short use or discard it. You do not, in the same sentence, call it nuclease-free because the assay is not a nucleic-acid assay. You also do not stretch the decision to a colleague's culture medium.

A tube of water sold for RNA work is past the date and half empty. The selection is the nuclease stop. A preview PCR on a hardy transcript does not move it off that stop.

Heat, humidity, and a date that was never true

A hot store or a humid shelf can put a reagent outside its storage statement months before the printed date. Sodium hydroxide pellets liquefy. Anhydrous salts become hydrates. A cold-chain enzyme that thawed in a power cut has a history the date does not describe. Record the excursion. The selection then follows the same tree: if you can test the property with a real control, the laboratory may decide; if the property is sterility, nuclease control, or calibration, replace the material. Do not calculate a "remaining life" from the printed date minus the hot days. That arithmetic is fiction.

Safety and scope

Expired chemicals are not milder chemicals. An old bottle can be more hazardous, as with peroxides, or simply unknown. Use the safety data sheet and the laboratory waste rules. This page is not an instruction to administer anything, not a clinical shelf-life, and not a biosafety approval. A research reagent's date and a medicine's date are different regulatory objects. Do not carry a laboratory habit across that line.

What to put in the enquiry

Ask for the specification the date supports, the storage condition that keeps the date meaningful, and the certificate that states both. Ask which tests are release tests you could, in principle, repeat, and which are claims you should treat as closed when the seal opens. For water, name ions, organic carbon, endotoxin, or nuclease control as separate lines so a later argument does not collapse them into one expiry. When a bottle arrives, file the date and the storage requirement in the same note as the lot. The day you need the selection tree, those two lines are the first thing you read.