21 CFR 111.75: Finished Product Testing & Release Rules

Most dietary supplement firms don't get cited for skipping testing entirely. They get cited for testing the wrong thing, relying on a supplier's paperwork when the regulation doesn't allow it, or never writing down the specification they were supposedly testing against in the first place. 21 CFR 111.75 is the clause that governs whether a finished batch of dietary supplements is fit to release, and it's one of the most frequently misread sections in Part 111.

I've walked enough firms through FDA Form 483 responses to know where the confusion sits. It's not the concept of testing. It's the specific, narrow rule about identity testing that cannot be delegated to a supplier's Certificate of Analysis, and the broader rule about what "appropriate" testing actually means when your specification is vague. This guide breaks 111.75 down clause by clause, with the practical release-testing workflow underneath it.

What 21 CFR 111.75 Actually Requires

21 CFR 111.75 sits inside Subpart E of the dietary supplement CGMP regulation (21 CFR Part 111), the subpart covering production and process controls. Within Subpart E, 111.75's job is narrow: it tells you what you must do to determine whether the specifications you established under 21 CFR 111.70 are actually met, for in-process material and for the finished batch.

That distinction matters. 111.70 is where you write the specification. 111.75 is where you prove you hit it. A firm can have a beautifully documented specification for potency and still fail an inspection if nothing in its records shows the finished batch was actually tested against that number.

Under 21 CFR 111.75(c)(1), a manufacturer must confirm the identity of the dietary supplement and determine the purity, strength, and composition of, and limits on contaminants in, the finished batch. Identity confirmation is the part firms most often get wrong, because the regulation treats it differently than the other three attributes.

Identity Testing: The One Test You Cannot Outsource to a COA

Here is the fact that catches the most firms off guard: 21 CFR 111.75(c)(1) requires at least one appropriate test or examination to verify the identity of the finished dietary supplement, and this cannot be satisfied by relying on a supplier's Certificate of Analysis alone. You have to run the test, or have a qualified lab run it for you, on the finished batch itself.

Contract manufacturers and brand owners both stumble here, usually for the same reason: they've already tested the incoming raw materials for identity, and they assume that coverage carries through to the finished product. It doesn't. A raw material identity test tells you what went into the mixing tank. It says nothing about whether the correct material ended up in the correct bottle at the correct concentration after blending, encapsulation, and packaging. The finished batch is a different thing, legally and analytically, and it needs its own identity confirmation.

This is also the single most common 483 observation I see tied to this section: a firm testing identity on components but never on the finished, packaged product.

Purity, Strength, and Composition: Where You Have Options

The other three attributes named in 111.75(c)(1), purity, strength, and composition, plus limits on contaminants, work differently than identity. Here, the regulation allows you to determine compliance through your own testing or examination, or by relying on a Certificate of Analysis from the party that manufactured or tested the finished batch, provided you've qualified that party's COA as reliable. Qualification isn't a formality. It means periodically confirming the accuracy of the supplier's reported results against your own testing, and documenting that confirmation.

That "provided" clause is where firms build a real gap without noticing it. Accepting a COA at face value, filing it, and moving on is not the same thing as qualifying the supplier. If you can't produce records showing when and how you last verified that the supplier's COA data holds up, an investigator has grounds to say you never established the reliability the regulation requires.

In-House Testing vs. Certificate of Analysis Reliance

The table looks permissive for everything except identity. In practice, the qualification burden for COA reliance is heavy enough that a lot of smaller manufacturers find it's cheaper and more defensible to just run the tests themselves. Others contract a lab instead, one that will stand behind the data with method validation records you can inspect.

The Specification Comes First: 21 CFR 111.70

None of 111.75's testing obligations mean anything without a specification to test against, and that specification is built under 21 CFR 111.70. Under 111.70(e), you must establish specifications for the identity, purity, strength, and composition of the finished batch. Under 111.70(f), you must separately establish specifications for the types of contamination that may adulterate, or lead to the adulteration of, the finished batch.

Notice that contamination limits are a distinct specification from purity. I've seen firms conflate the two, writing one purity spec and assuming it implicitly covers heavy metals, microbial limits, and pesticide residue. It doesn't, and an investigator reviewing your master manufacturing record will look for a documented contaminant specification that stands on its own, tied to an actual test method and acceptance criterion.

In-Process Testing: The Step Before Finished Batch Release

21 CFR 111.75(a) covers a narrower, earlier obligation: determining whether the in-process specifications you established under 111.70(b) are met. This is the testing that happens mid-process, not at release. Weight checks during encapsulation, blend uniformity testing before compression, moisture checks before drying is complete. These in-process checks exist so a batch doesn't travel all the way to finished-product testing before you discover the blend was off.

A production and process control system that only tests at the finished batch stage is more expensive to run and slower to catch problems than one that checks in-process material along the way.

Reserve Samples: Your Evidence If Something Goes Wrong

21 CFR 111.83 requires you to collect and hold reserve samples of each lot of packaged and labeled dietary supplements you distribute, in a quantity sufficient to perform all required tests or examinations under 111.75 at least twice. If a complaint comes in six months after distribution, or FDA requests a sample during an inspection, the reserve sample is what lets you re-test rather than explain that nothing is left to check. Retention runs for one year past the shelf-life date, or two years past the date of distribution if no shelf-life date is assigned.

When Specifications Aren't Met

Testing under 111.75 only has teeth if you act on the result. When a specification isn't met, the batch cannot simply be shrugged through to distribution. Under the broader corrective-action requirements of Part 111, a firm has to investigate the failure, determine whether other batches are implicated, and document a disposition decision, whether that's rejection, reprocessing under a documented and justified procedure, or, in rare and well-supported cases, release with an explained deviation. An investigator reviewing a batch record wants to see that a failing result triggered an actual investigation, not a retest until a passing number showed up. Retesting into compliance without an investigation is one of the fastest ways to turn a testing question into a data-integrity question.

Common 483 Observations Tied to 111.75

Across warning letters and inspection findings, the pattern repeats:

• No identity test performed on the finished, packaged product, only on incoming components.
• Certificate of Analysis accepted from a supplier with no documented qualification of that COA's reliability.
• A written specification for purity or strength with no corresponding test method referenced or validated.
• Contaminant limits not established as their own specification, separate from purity.
• Reserve samples not retained in sufficient quantity to allow a second full round of release testing.
• Out-of-specification results resolved by retesting without a documented investigation into the root cause.

Any one of these, standing alone, is a fixable gap. Several of them together tend to indicate a quality system that was assembled to look compliant on paper rather than built to actually function during production.

Release Testing Checklist

• Written specification exists for identity, purity, strength, composition, and contaminant limits, each tied to 111.70(e) or 111.70(f).
• At least one identity test or examination is performed directly on the finished batch, not inferred from component testing.
• Any Certificate of Analysis relied on for purity, strength, or composition is backed by a documented supplier qualification record.
• In-process specifications under 111.70(b) are tested during production, not only at the end.
• Test methods are validated or verified for their intended use before they're used for release decisions.
• Reserve samples are retained in a quantity sufficient to repeat every required finished-batch test at least once.
• Out-of-specification results trigger a documented investigation before any disposition decision is made.
• Batch records tie test results, specifications, and the final release decision together in one traceable chain.

Building a Release Testing Program That Holds Up

A workable finished-product release program under 111.75 needs four things working together: a written specification for every attribute FDA expects (identity, purity, strength, composition, and contaminant limits), a validated or verified test method for each one, a documented chain of custody from in-process check to finished batch release, and a reserve sample program sized to survive a re-test request.

Where a firm relies on outside testing labs or contract manufacturers, the qualification of that supplier's data is not a one-time signature on an intake form. It's a periodic re-confirmation, and the record of that re-confirmation is what an investigator will ask to see. If you're structuring or auditing supplier relationships that feed into your release decisions, our guide on GMP supplier and vendor qualification procedures walks through what that qualification record needs to contain.

FAQ

Can I release a dietary supplement batch based only on a supplier's Certificate of Analysis?

For purity, strength, and composition, yes, if you've qualified that supplier's COA as reliable through documented periodic confirmation. For identity, no. 21 CFR 111.75(c)(1) requires you to perform at least one appropriate test or examination on the finished batch yourself.

What's the difference between 111.70 and 111.75?

111.70 is where you establish the specification itself, the numeric or qualitative criteria a batch has to meet. 111.75 is where you determine, through testing, whether the finished batch actually meets that specification. One writes the target; the other checks whether you hit it.

Do I need a separate specification for contaminants, or does a purity spec cover it?

They're separate. 21 CFR 111.70(f) requires its own specification for the types of contamination that may adulterate the finished batch, distinct from the identity, purity, strength, and composition specification under 111.70(e).

How long do I need to keep reserve samples?

Under 21 CFR 111.83, reserve samples must be retained for one year past the shelf-life date, or two years past the date of distribution if no shelf-life date applies, and in a quantity sufficient to run every required release test at least twice.

What happens if a finished batch fails its release testing?

The batch cannot simply be retested until it passes. A documented investigation into the failure is required, along with a disposition decision, whether that's rejection, a justified reprocessing procedure, or a fully explained and supported release. Skipping the investigation and retesting into compliance is a common source of data-integrity findings.

If you're preparing for an inspection and want a second set of eyes on your finished-product testing program before FDA sees it, our dietary supplement GMP page outlines how we work with manufacturers on exactly this kind of gap.

Last updated: 2026-08-24