Supplement expiration dates: what FDA requires, how shelf life is set and what a date changes after launch
FDA's manufacturing rule for supplements doesn't require a date on the label, but FDA says any date you print should be backed by data. Here is FDA's exact wording, the choice between expiration, best-by and no date, how stability studies set shelf life, the label amounts a date has to survive, and what a date changes for production runs, picking and subscriptions.
On this page
- What FDA says about supplement expiration dates
- Expiration date, best-by date or no date: which to print
- How shelf life is set: real-time and accelerated stability data
- Do supplements expire? What the label has to deliver until the date
- Storage conditions decide whether the date holds
- What a date changes after launch: run sizes, FEFO and subscriptions
- Before you print a date: a checklist
- Frequently asked questions
A supplement expiration date isn't required by FDA's manufacturing rule for supplements, 21 CFR part 111. Asked whether the rule requires an expiration date, a shelf date or a "best if used by" date, FDA's compliance guide answers "No." But if you print one, FDA says you should have data to support it, and the label amounts still have to hold when the product is tested.
What FDA says about supplement expiration dates
FDA answers the question in its Small Entity Compliance Guide for the dietary supplement CGMP rule, dated December 2010. Two of its questions settle most arguments about dates:
- Does the rule require you to establish an expiration date, a "shelf date" or a "best if used by" date? FDA's answer is "No."
- Should you have data to support any such date you put on a label? FDA's answer is "Yes." The rule doesn't require a date, but you "should have data to support any such date that you place on a product label," citing the rule's 2007 preamble (72 FR 34752 at 34855).
The verb is "should", the language of guidance, not the "must" of a regulation, so it would be wrong to claim FDA requires stability studies. It's still the standard to plan around: a date on a bottle is a statement about your product, and FDA expects data behind it.
Both questions also name all three kinds of date, so calling yours "best by" instead of "expiration" doesn't lower the bar. The rule's 2007 preamble groups them the same way where it explains how long records are kept: shelf-life dating "such as expiration dating, shelf life dating, or 'best if used by' dating".
A date also sets two clocks. Reserve samples of each packaged lot, and the records part 111 requires, are kept for 1 year past the shelf-life date if you use shelf-life dating, or for 2 years from distribution of the last batch if you don't (21 CFR 111.83, 111.605). A lot dated March 2029 keeps its samples and records until March 2030.
All of this is the US federal rule as of October 2026, and it isn't legal advice. Retailers, marketplaces and other countries may set date requirements of their own, which this post doesn't cover, so check each channel you sell through.
Expiration date, best-by date or no date: which to print
Here is what each choice asks of you, and what it tells the buyer:
Expiration date, best-by date or no date
What each choice means under the federal CGMP rule
Data expectedExpiration date
- FDA: you should have data to support it
- Reserve samples and records: kept 1 year past the date
- To a buyer: a hard stop
Fits when your stability data is solid and you want no ambiguity.
Data expectedBest-by date
- Named alongside expiration dates in FDA's guide: the same data expected
- Same clock: samples and records kept 1 year past the date
- To a buyer: a quality window
Softer wording, same evidence.
AllowedNo date
- Not required by the federal CGMP rule
- Reserve samples and records: kept 2 years from distribution of the last batch
- Label amounts must still meet FDA's minimums when tested
- A retailer, marketplace or other country may want one
Nothing to support, and nothing for the buyer to go on.
The choice comes down to the data you have. If your manufacturer holds stability results for this formula in this package, print the date they support, in whichever wording you prefer. If it doesn't, launching without a date is allowed, and you can add one once the results exist. The option that doesn't work is a date picked because 24 months sounds normal. Whatever you print, write the month in letters, as in MAR 2029, so no buyer has to guess whether 03/04 means March or April.
How shelf life is set: real-time and accelerated stability data
A shelf life comes from a stability study. Samples from production lots, in the container you sell, are stored under set conditions and tested at intervals for the amounts on the label. On real-time data, the date is the last test point at which everything still passes.
There are two kinds of study, and the clearest definitions come from drug regulation: ICH Q1A(R2), a stability guideline FDA published in 2003 for new drugs. Its long-term testing runs under the recommended storage condition for the shelf life proposed for the label, which in its general case means 25°C and 60% relative humidity (or 30°C and 65%). Its accelerated testing uses exaggerated storage conditions, 40°C and 75% humidity, to speed up chemical degradation and physical change.
Long-term (real-time) results measure the date; accelerated results predict it sooner. The guideline itself warns that accelerated results "are not always predictive of physical changes", and it was written for drugs, not supplements. So ask your manufacturer which conditions and test points its study used, and whether your date rests on real-time results or on an accelerated prediction. If it's a prediction, keep the real-time study running and be ready to shorten the date if the results fall short. Our post on choosing a supplement manufacturer puts these questions alongside the rest of the audit.
A study also shows whether your ingredients age at the same pace. Here is an illustrative capsule with two added ingredients, tested every six months against their label amounts:
Two ingredients, one capsule, one date
Illustrative stability results: amount found, as a share of the label amount
- Ingredient A
- Ingredient B
- Label amount
Months after manufacture
Show the numbers
| Months after manufacture | Ingredient A | Ingredient B | Label amount |
|---|---|---|---|
| 0 | 104% | 118% | 100% |
| 6 | 103.5% | 113% | 100% |
| 12 | 103% | 108% | 100% |
| 18 | 102.5% | 104% | 100% |
| 24 | 102% | 101% | 100% |
| 30 | 101.5% | 97% | 100% |
| 36 | 101% | 92% | 100% |
Ingredient A would carry a 36-month date on its own. Ingredient B passes at 24 months (101% of label) and fails at 30 (97%). Every ingredient has to meet its label amount, so the product gets a 24-month date: the weakest ingredient sets it, not the average.
The data only covers what was tested: that formula, in that package, from that manufacturer. Switch to a cheaper bottle, a different capsule shell, a new desiccant or another manufacturer, and what you sell is no longer what was tested. Before the old date goes on the new version, ask whether the old data still represents it, or start a study on the new one.
Do supplements expire? What the label has to deliver until the date
Whether a supplement "expires" is really a question about its label: how long the amounts printed on it stay true. FDA judges those amounts against fixed minimums (21 CFR 101.36(f)(1), which points to 101.9(g)):
- Added ingredients, the vitamins and extracts put into the formula, must be present at 100% or more of the declared amount, except for deviation attributable to the analytical method.
- Naturally occurring ingredients must be present at 80% or more. FDA's own example in its labeling guide: vitamin C added to a product falls under the 100% rule, while vitamin C that comes naturally from added rose hips must reach 80%.
- Ingredients without a Daily Value follow the same split.
- The sample is a composite of 12 packages or 10% of the packages in the lot, whichever is smaller, picked at random.
The rule also allows "reasonable excesses over labeled amounts" within good manufacturing practice, without naming a percentage. That's the room for an overage: making the product above the label so a fragile ingredient still meets it at the last test point. Ingredient B in the chart is declared at 50 mg and made at 59 mg, 118% of label. At month 24 the test finds 50.5 mg (101%); at month 30, 48.5 mg (97%), a fail for an added ingredient. A 30-month date would need a bigger overage, a more stable form or a better package. Ask your manufacturer to explain the overage it picks, because the rule allows only a reasonable excess.
The rule doesn't say when the minimum applies. Neither 101.36(f) nor 101.9(g) mentions shelf life or when the sample is taken. FDA describes them as its "requirements for the minimum quantity of dietary ingredients throughout the shelf life" of a supplement in a footnote to a 2018 draft guidance on labeling supplements that contain live microbials. A footnote in a draft isn't the regulation, but it is FDA's own reading, and the safe one to plan around: the panel should still be true when the date arrives.
The certificate of analysis at release tells you where a lot starts; the stability data tells you where it ends. The panel itself, line by line, is in our post on supplement label requirements.
A date passes through four stages before a bottle ships, and each one can break it:
From stability study to warehouse shelf
The four stages behind a supplement expiration date, and what breaks each one
1Run the stability study
What happens: Samples from production lots, in the container you sell, are stored at the label's conditions (and, for a faster read, at accelerated ones) and tested at set intervals for every amount on the label.
Why it matters: This is the data FDA says you should have before you print any date.
What goes wrong:
- A study on a different formula or package than the one you sell
- An accelerated prediction with no real-time study running behind it
- Only the headline ingredient tested
Read more: How to Choose a Supplement Manufacturer: cGMP Checks
2Choose the date
What happens: Take the last test point at which every ingredient still meets its minimum: 100% of label for added ingredients, 80% for naturally occurring ones.
Why it matters: In a 2018 draft guidance, FDA describes those minimums as applying throughout the shelf life, so the date marks the last point at which the data supports the panel.
What goes wrong:
- Dating by the average instead of the weakest ingredient
- Rounding up to a familiar 24 or 36 months the data doesn't reach
3Print the date and the storage statement
What happens: The date goes on the label in the wording you chose, with storage instructions that match the study's conditions.
Why it matters: The date also starts the retention clock: reserve samples and records are kept for 1 year past it.
What goes wrong:
- A storage statement that promises more than the study tested
- A date format buyers can misread
4Hold and ship by date
What happens: Keep stock under appropriate conditions of temperature, humidity and light, pick first-expired, first-out, and set the least shelf life you will ship.
Why it matters: A date is only supported for the conditions the study tested, and part 111 requires holding conditions that protect identity, purity, strength and composition.
What goes wrong:
- Summer heat in trucks and unventilated storage
- Short-dated bottles sent to subscribers
Read more: Reorder Point: Formula, Calculator and Safety Stock
Storage conditions decide whether the date holds
Two of part 111's rules for holding and distributing bear on the date: hold supplements under "appropriate conditions" of temperature, humidity and light, so their identity, purity, strength and composition aren't affected, and distribute them under conditions that protect against contamination and deterioration (21 CFR part 111, subpart M). The rule sets no numbers. Your label's storage statement and your stability study supply them.
The same statement decides how you keep reserve samples, the bottles held back from each lot "for use in appropriate investigations". They're kept in the same container-closure system you sell in, under the storage conditions on your label (or ordinary storage conditions if it states none), in at least twice the quantity needed for every test. If a question about a lot comes up later, those are the bottles kept to investigate it.
So three things have to agree: the storage statement on the bottle, the conditions in the study and the places the bottles actually sit. A date backed by a study at 25°C is backed for 25°C. If your stock spends August in a hot trailer, ask whether your manufacturer has data on that kind of excursion: Q1A(R2) names accelerated data as one way to evaluate short-term excursions outside the label conditions, "such as might occur during shipping". What a warehouse has to do for supplements is its own subject, covered in our post on supplement fulfillment.
What a date changes after launch: run sizes, FEFO and subscriptions
Once bottles carry a date, every week in the warehouse costs shelf life, and the size of a production run decides how much is left on the last bottle when it ships. Say your capsules carry a 24-month date, a run takes two months from manufacture to first sale (testing, release, freight), and you sell 2,000 bottles a month:
Shelf life left on the last bottle of a run
Illustrative: 24-month date, 2 months from manufacture to first sale, 2,000 bottles sold a month
The last bottle of the 40,000-bottle run ships with two months left. Put it in a subscriber's three-bottle, 90-day pack and the third bottle reaches its date around the day it's opened. Three habits prevent that:
- Pick first-expired, first-out (FEFO). The lot that dates out soonest ships first, whatever order the lots arrived in. Ask your 3PL whether it records lot and date for every location.
- Set a floor. Decide the least shelf life you'll ship, for example the supply in the box plus three months, and keep stock below it out of multi-bottle and subscription orders. For a 90-day pack that floor is six months, which in the example holds back the last 8,000 bottles of the 40,000 run.
- Size runs to the date, not only to the price. A lower unit price on a bigger run can be eaten by bottles that age out. The reorder point decides when the next run starts and how big it should be.
If you sell through retailers or marketplaces, ask each one whether it sets a minimum remaining shelf life for stock it receives, and plan runs around the strictest.
Before you print a date: a checklist
Go through it with your manufacturer before the label goes to print:
Before a date goes on the label
0 of 10 done
If the answers are thin, launch without a date and add one when the data is in. A bottle without a date is allowed; a date you can't defend is a claim you can't support.
Frequently asked questions
Does the FDA require stability testing for supplements?
Not as a requirement. FDA's compliance guide to the dietary supplement manufacturing rule says the rule doesn't require an expiration date, and that if you put one on a label you should have data to support it. In practice that data comes from a stability study of the product in its final package, so a brand that prints a date should be able to show the study behind it.
Is a best-by date on a supplement different from an expiration date?
To a buyer, yes: best by reads as a quality window, expiration as a hard stop. To FDA, no. Its compliance guide names expiration, shelf and best if used by dates together, says the manufacturing rule requires none of them, and says you should have data to support whichever you print. The retention clock is the same too: one year past the date.
Does a supplement without an expiration date last forever?
No. A missing date only means the brand chose not to print one, which the federal manufacturing rule allows. It says nothing about how long the contents keep their label amounts. Those amounts still have to meet FDA's minimums when the product is tested, and its records and reserve samples are kept for two years after the last batch is distributed.
Can a brand extend the date on stock it already has?
Only with data that covers the longer period. FDA's guidance says you should have data to support any date on a label, and a 30-month date is supported by stability results showing every ingredient still meets its label minimum at 30 months, for the same formula in the same package. Without that, sell the stock within its current date.
How long are records kept for a supplement lot with a date?
Under 21 CFR 111.605, the records part 111 requires are kept for one year past the shelf-life date when you use shelf-life dating, or two years after the last batch they cover is distributed when you do not. Reserve samples follow the same clock under 111.83. A lot dated March 2029 keeps both until March 2030.
What should I ask my manufacturer about an expiration date?
Ask what data supports the date it proposes: which formula and package were tested, at what storage conditions, at which time points, and whether the results are real-time or accelerated. Ask for every label ingredient at the last time point, not only the headline one, and whether the real-time study is still running. Then check the storage statement on the label against the study.
Sources
- Small Entity Compliance Guide: Current Good Manufacturing Practice in Manufacturing, Packaging, Labeling, or Holding Operations for Dietary Supplements — FDA
- Current Good Manufacturing Practice in Manufacturing, Packaging, Labeling, or Holding Operations for Dietary Supplements, final rule (72 FR 34752, 25 June 2007) — Federal Register
- 21 CFR 111.83: Requirements for reserve samples — eCFR
- 21 CFR 111.605: Requirements for the records you make and keep — eCFR
- 21 CFR part 111, subpart M: Holding and Distributing — eCFR
- 21 CFR 101.36: Nutrition labeling of dietary supplements — eCFR
- Dietary Supplement Labeling Guide: Chapter IV. Nutrition Labeling — FDA
- Policy Regarding Quantitative Labeling of Dietary Supplements Containing Live Microbials, draft guidance (September 2018) — FDA
- Q1A(R2) Stability Testing of New Drug Substances and Products (November 2003) — FDA