24.9 — How Expiry Dates Are Actually Calculated

A US Department of Defense study tested 122 drug products held in stockpiles past their expiry dates.

Around 88 percent were still within specification, and the average extension beyond the printed date was more than five years. Some products remained acceptable well over a decade past it.

Which raises an obvious question: where does the printed date come from, if not from when the product stops working?

The answer is that an expiry date is not a prediction of failure. It is a guarantee the manufacturer is prepared to make, based on how long they chose to run the test.

And knowing the difference between those two things is genuinely useful, because it tells you which dates to obey absolutely and which are conservative.

What a shelf life actually is

The period during which the manufacturer guarantees the product meets its specification when stored as directed.

Three things in that sentence do the work:

Meets its specification — not "is safe" and not "still works a bit". For a drug, that usually means at least 90 to 95 percent of the stated amount of active ingredient remains, with degradation products below defined limits.

When stored as directed — the date assumes the storage conditions on the label were maintained. A drug kept in a hot bathroom or a car glovebox has not had those conditions (Chapter 22.4).

Guarantees — it is a commitment with legal and regulatory weight, which is why companies are conservative.

How a drug shelf life is determined

Through stability testing, under internationally agreed rules — the ICH guidelines, which are followed by regulators in most major markets.

Real-time testing

Batches are stored under defined conditions and tested at intervals.

Long-term: 25 degrees and 60 percent relative humidity, for temperate climates.

Intermediate: 30 degrees and 65 percent humidity.

Accelerated: 40 degrees and 75 percent humidity.

And climate zones matter — a product sold in a hot humid country is tested at 30 degrees and 75 percent humidity as the long-term condition, which is why the same drug can carry different shelf lives in different markets.

Testing points are typically at 0, 3, 6, 9, 12, 18, 24 and 36 months for long-term studies.

What is measured at each point: the amount of active ingredient remaining; degradation products; dissolution — how fast the tablet releases the drug; appearance, colour and odour; moisture content; pH for solutions; microbial contamination; and container integrity.

A product passes if every one of those stays within specification.

Accelerated testing and the chemistry behind it

Real-time testing for a five-year shelf life takes five years, which is a problem when you want to launch a product.

So accelerated testing raises the temperature to speed up degradation, and the results are extrapolated.

And the extrapolation rests on the Arrhenius equation, which describes how reaction rate depends on temperature:

k = A\,e^{-E_a/RT}

Read aloud: the rate constant k equals a constant A multiplied by e raised to the power of minus the activation energy E_a divided by the gas constant R times the absolute temperature T.

What each symbol means: k is how fast the degradation reaction proceeds. A is the pre-exponential factor, essentially how often the molecules collide in a way that could react. E_a is the activation energy — the energy barrier a reaction must climb. R is the universal gas constant, 8.314 joules per mole per kelvin. T is the absolute temperature in kelvin.

The practical consequence, and the reason the whole method works: because temperature appears inside an exponential, small increases in temperature produce large increases in rate.

A widely used rule of thumb is that reaction rate roughly doubles for every 10 degree rise, which for typical activation energies is approximately right.

Which means storage at 40 degrees for 6 months produces roughly the degradation of around 2 years at 25 degrees.

Six months of accelerated data therefore supports a provisional shelf life of around 24 months, which is confirmed or corrected as the real-time data accumulates.

And the limits of the method matter. Arrhenius extrapolation assumes the same reaction mechanism operates at both temperatures. If heating triggers a different degradation pathway — melting, a phase change, a different reaction becoming dominant — the extrapolation breaks. Which is why accelerated data alone is never sufficient and real-time data always follows.

Biological products — vaccines, antibodies, insulin — are far less amenable to this, because proteins denature rather than degrading by simple chemistry, and their stability is dominated by temperature-dependent unfolding. They rely much more heavily on real-time testing and cold storage.

What makes drugs degrade

Hydrolysis — water splitting the molecule. The commonest pathway, and why moisture control and desiccant sachets exist. Aspirin hydrolyses to salicylic acid and acetic acid, which is why old aspirin smells of vinegar.

Oxidation — reaction with oxygen. Countered with antioxidants, nitrogen-filled headspace, and sealed blisters.

Photolysis — light-driven breakdown. Countered with amber glass and opaque packaging.

Thermal degradation.

Physical changes — tablets hardening so they no longer dissolve properly, creams separating, suspensions caking.

And the container is part of the product. Stability data is generated in the final packaging, which is why the shelf life applies to a sealed blister and not to a tablet decanted into a pill organiser.

After opening

Once opened, the tested conditions no longer apply.

Which is why separate in-use dates exist:

Eye drops — typically 28 days after opening, because of contamination risk rather than chemical degradation.

Insulin in use — around 28 days at room temperature (Chapter 22.4).

Reconstituted antibiotic suspensions — commonly 7 to 14 days, refrigerated.

Creams and ointments in jars — shorter than in tubes, because fingers introduce bacteria.

And the small open-jar symbol on cosmetics and some medicines gives the in-use period in months.

The exceptions worth knowing

Some drugs genuinely must not be used past their date:

Nitroglycerin tablets — lose potency quickly once the bottle is opened, and the one time you need them is an emergency (Chapter 22.14). This is why the spray is preferred.

Insulin and biologics — lose activity, and a diabetic relying on degraded insulin is in real danger.

Vaccines — reduced potency means reduced protection, with no way to tell.

Liquid antibiotics — both potency loss and microbial growth.

Adrenaline auto-injectors — potency falls, and this matters absolutely (Chapter 23.5). Discoloured or cloudy solution means replace it now.

Anything that has changed colour, smell or texture, or where a tablet crumbles or a solution has particles in it.

And the historical example that shaped the field: degraded tetracycline was linked to a form of kidney damage in the 1960s, which is one of the few documented cases of an expired drug being actively toxic rather than merely weaker. Modern formulations have addressed this, and it remains the standard reason not to treat "expired means weaker" as a universal rule.

Everything else, in a solid oral dosage form, stored properly, is usually about potency rather than safety — and using it is a decision to accept an unquantified reduction in dose, which is fine for a headache and not for anything that matters.

Food dates

A completely different system, based on microbiology rather than chemistry.

Use by — a safety date, set by testing how long it takes for pathogens to reach dangerous levels under specified storage. On high-risk foods: meat, fish, dairy, ready meals.

Do not eat past it. Listeria, salmonella and other pathogens produce no change in smell, taste or appearance (Chapter 24.3).

Best before — a quality date, set by sensory testing. When does it stop tasting right, look right, or have the right texture?

Safe afterwards. Use your senses.

How they are determined:

Challenge testing — deliberately inoculating the food with a known pathogen and measuring growth over time under realistic storage conditions.

Predictive microbiology — mathematical models of bacterial growth based on water activity, pH, salt, temperature and preservatives.

Sensory panels — trained tasters assessing samples at intervals.

Accelerated testing at elevated temperature, with the same caveats as for drugs.

And a safety margin is applied on top — typically shortening the date by a third or so, because manufacturers must account for imperfect home storage and for the fridge that runs warm.

Which is a large part of why so much edible food is discarded. Food waste from date confusion is substantial, and several countries have removed best-before dates from fresh produce entirely for that reason.

What determines how long a food lasts:

Water activity — the availability of water for microbial growth, not the total water content. Honey is 17 percent water and essentially never spoils, because sugar binds the water so tightly that nothing can use it. The same principle preserves jam, dried fruit and salted fish.

pH — most pathogens cannot grow below pH 4.6, which is the basis of pickling and why acidic canned foods are safe with less processing.

Salt and sugar, which lower water activity.

Preservatives — nitrites in cured meat, sorbates, benzoates, sulphites.

Packaging — vacuum, modified atmosphere, oxygen scavengers.

Processing — pasteurisation, sterilisation, canning, irradiation.

And temperature, which dominates everything. Refrigeration does not stop bacterial growth; it slows it. A fridge at 5 degrees rather than 8 makes a substantial difference to how long food lasts.

Foods that keep almost indefinitely: honey, salt, sugar, dried rice, vinegar, and hard liquor. All of them are hostile to microbial life for one of the reasons above.

Cosmetics and everything else

Sunscreen — typically 3 years, and the active ingredients genuinely degrade. An old bottle does not deliver its stated SPF, which matters because it is being relied on (Chapter 22.14).

Mascara — 3 months after opening, because of eye infection risk.

Batteries — self-discharge, and the date indicates when they will still hold most of their stated capacity.

Fire extinguishers, smoke alarms and carbon monoxide alarms — all have finite lives, and all are things people assume work forever. Smoke alarms are typically replaced every 10 years; carbon monoxide alarms every 5 to 7; and both should be tested regularly.

Car seats for children — expire, because plastic degrades with heat cycling and UV exposure.

Climbing rope, harnesses and helmets — retired on age as well as on use.

The practical rules

Obey use-by dates on food absolutely.

Treat best-before dates as advice and use your senses.

Obey expiry dates on the medicines that matter — adrenaline injectors, glyceryl trinitrate, insulin, vaccines, eye drops and anything liquid.

Store things as the label says, because the date assumes it. The bathroom cabinet is the worst place in the house for medicines (Chapter 22.4).

Check the emergency items twice a year. Adrenaline injector, aspirin, inhaler, smoke alarm, carbon monoxide alarm, fire extinguisher. Pick two dates you will remember — the clock changes work well.

And return unused medicines to a pharmacy rather than binning or flushing them.

And that is Volume V

It started with a single cell, and how the first ones assembled themselves out of chemistry (Chapter 1.9).

It went through the code that runs them, the four billion years of editing that shaped it, and the body that all of it built — every organ, its shape, its blood supply, its chemistry, and what happens when it fails.

Then what goes wrong, from the commonest to the rarest, and what medicine can actually do about each.

Then the drugs, what they do at the molecule and how to take them safely.

Then the twelve chapters that matter most: what to do in the minutes when someone's life is in your hands.

And it ends here, on food, movement, sleep and the ordinary decisions that decide how the decades go.

The point of all of it was never to make you a doctor.

It was so that you understand the thing you live inside — well enough to look after it, well enough to know when something is wrong, well enough to ask the right questions, and well enough to help somebody else on the worst day of their life.

You know. You survive. You help others survive.

That was the objective, and that is the volume.

What the next Volume covers

Volume VI turns from the body to the mind — psychology, how thinking actually works, how it goes wrong, and how to tell whether the thoughts in your head are your own.