Out-of-Specification Results in Supplement Testing: What FDA's cGMP Investigation Process Actually Requires

The call comes in on a Tuesday. Your potency result is at 74% of label claim — your specification requires a 90% minimum. Your first instinct is to ask the lab to retest. That instinct, while understandable, is also one of the most common ways supplement brands inadvertently create a regulatory problem.

Under 21 CFR Part 111 — the FDA’s current Good Manufacturing Practice regulation for dietary supplements — an out-of-specification (OOS) result isn’t a simple pass/fail event. It’s the beginning of a structured, documented investigation process. How that process is conducted determines everything from product disposition to your standing in an FDA inspection.

Most brands don’t know the details of that process until they’re already in the middle of one.

What Actually Defines an “Out-of-Specification” Result

Before getting into the investigation itself, it’s worth being precise about what OOS means — because not every surprising number qualifies.

An OOS result is any analytical result that falls outside the established specifications for a finished product, in-process material, or raw material. Those specifications might come from a USP monograph acceptance criterion, an FDA regulatory limit, or your own internal quality standards documented in a product specification sheet. If the result falls outside the range written in your spec, it’s OOS.

This is distinct from an atypical or anomalous result — one that’s statistically unusual but still within spec. It’s also different from an out-of-trend (OOT) result, which falls within spec but deviates from expected patterns over time. OOS is the most consequential category because it triggers mandatory investigation under §111.68 and §111.70 of 21 CFR Part 111.

Under §111.70(e), you must establish a written specification for each dietary supplement you manufacture. Under §111.68, your laboratory controls must include a system for reviewing OOS results. These aren’t suggestions — they’re the regulatory minimum.

Phase I: The Laboratory Investigation

FDA’s guidance on OOS investigations — originally published in June 2006 and broadly applicable to dietary supplement cGMP — describes a two-phase investigation framework. Phase I is entirely the lab’s responsibility, and it must happen before any retesting is authorized.

The goal of Phase I is to determine whether a specific, assignable laboratory error caused the failing result. Analysts and supervisors review:

• The calculation. Was there an arithmetic or unit conversion mistake in the reported value?
• The instrument. Were calibration standards within acceptance criteria? Does the instrument log show any aberration during the run?
• Sample preparation. Was the correct extraction solvent used? Was sample weight recorded accurately? Was there a dilution error?
• Reagents and standards. Were reference standards within their expiry date? Was the mobile phase prepared to spec?
• The chromatogram. Does peak integration look correct? Was there co-elution or baseline drift that would distort the result?

Here’s the critical rule: if a laboratory error is found and documented with objective evidence, the original result can be invalidated and a single confirmatory retest performed. If no assignable cause is found, the original result stands and Phase I ends — the investigation moves to Phase II.

What labs cannot do — and what constitutes a cGMP violation — is run additional tests first and then work backward to construct a justification. The investigation must drive the retest decision, not the other way around. This distinction matters enormously during an FDA inspection. Phase I is typically expected to conclude within 3 to 5 business days of the OOS result being identified.

Phase II: The Full Investigation

If Phase I finds no assignable laboratory cause, the investigation expands well beyond the bench. Phase II typically involves several parallel lines of inquiry.

Expanded retesting of the original sample. Using statistical approaches referenced in USP <1151>, labs commonly test 6 to 7 additional aliquots from the same retained sample. Every result generated — including additional failing values — becomes part of the investigation record. There is no mechanism under cGMP to exclude inconvenient data points from the package. That’s a rule some brands discover the hard way.

Batch record review. Were there any documented deviations during manufacturing? Did any in-process test result sit close to the specification limit? Were there equipment cleaning or calibration anomalies on the manufacturing date?

Raw material traceability. Are the COAs for every ingredient in this batch verified and on file? Was there a supplier change or a lot substitution that might have introduced variability into the finished product?

Resampling consideration. In some cases, a different sample from the same batch is tested to determine whether the OOS result reflects a localized distribution issue versus a batch-wide problem. Resampling has strict requirements — it must be scientifically justified, not simply used as a mechanism to generate a passing result.

At the conclusion of Phase II, a written investigation report is required. That report must document every test performed, every result obtained — passing and failing — the conclusions reached, and the disposition decision for the batch, whether that’s release, rejection, reprocessing, or destruction. The phrase “investigation complete, batch released” without underlying documentation to support it is an observation waiting to happen.

What Brands Should Be Asking Their Contract Lab

Most supplement brands outsource their cGMP testing to contract laboratories, which means the OOS investigation process is largely happening outside their direct view. That’s precisely why the time to understand your lab’s OOS procedure is before you receive a failing result — not after.

There are four questions worth asking any contract lab you work with.

Do you have a written OOS SOP? Under ISO 17025:2017 — the international accreditation standard governing how testing laboratories operate — labs are required to maintain documented procedures for managing nonconforming work and OOS results. If a lab can’t produce its OOS SOP on request, that’s a meaningful flag about their quality system overall.

What’s your timeline for Phase I completion? A Phase I investigation that drags on for two or three weeks without resolution is a sign it isn’t being driven by objective evidence. Clear, committed timelines are reasonable to expect.

Will I receive the full raw data package? Your investigation report should include original chromatograms, instrument logs, analyst worksheets, and all retest results — not just a reissued COA with a passing number attached. If you’re only receiving summary documents, ask specifically why.

How do you handle client retesting requests? This is the most revealing question. A quality lab will explain that retesting is governed by the investigation outcome — not client preference. If the response suggests retesting is a routine accommodation that can be requested on demand, that’s worth taking seriously.

At Qalitex, our OOS procedure is grounded in both ISO 17025:2017 requirements and the cGMP investigation framework described under 21 CFR Part 111. Every investigation generates a full traceability package — original data included — and retest authorization requires documented Phase I conclusions. That’s the baseline your testing partner should be able to demonstrate.

The Regulatory Stakes of Getting This Wrong

FDA Warning Letters related to dietary supplement cGMP frequently cite inadequate OOS investigation procedures. Common observations include: retesting without documented laboratory error as the trigger; failure to include all retest data in the final investigation report; disposing of batches without completing Phase II; and missing or incomplete investigation reports altogether. These findings appear under 21 CFR Part 111 violations and, in more serious cases, accompany injunction proceedings.

The consequences extend beyond FDA. A supplier audit, a retail partner quality review, or an Amazon marketplace compliance inquiry may ask specifically whether your testing program includes OOS procedures and how they’re executed. “We retest until it passes” is not a defensible answer in any of those contexts.

A properly conducted OOS investigation — even one that ends in a batch rejection — is actually evidence of a functioning quality system. FDA knows failures happen. What they’re evaluating during an inspection is whether your procedures, your documentation, and your decision-making hold up when a result doesn’t go your way.

One Practical Step to Take Before Your Next Test Result

Ask your contract lab for their OOS investigation SOP today. Read it specifically for three things: the criteria that distinguish Phase I from Phase II, the retesting authorization requirements, and the documentation deliverables at the end of an investigation. If the SOP is vague on any of those three points, ask how they fill the gaps.

A failing test result is a data point. What your laboratory does with that data point — and how thoroughly it’s documented — is where your actual cGMP compliance gets demonstrated.

---

Written by Nour Abochama, Vice President of Operations, Qalitex Laboratories. Learn more about our team

Talk to our team about your testing needs. Contact us

Related from our network

• FDA Inspection Readiness and cGMP Compliance Consulting — Aurora TIC helps supplement and device firms prepare for FDA audits, build CAPA systems, and meet 21 CFR documentation requirements.
• Raw Material COA Verification and Supplier Qualification Testing — Ayah Labs provides independent raw material testing and supplier qualification support for ingredient sourcing programs.