OOS vs OOT in Pharma: Difference with Examples

In short: An out-of-specification (OOS) result is a test result outside the limit that applies to the material or product, so it always needs a formal investigation. An out-of-trend (OOT) result may still be inside the limit but does not fit the pattern of earlier data, so it needs a proportionate look before it turns into an OOS. This guide is for QC analysts and freshers: definitions, a comparison table, three worked examples, how OOT is detected, and common mistakes.

The difference in one minute

Every QC analyst meets both words in the first month, and they are often mixed up. The simplest way to remember the difference is the reference point.

• OOS is compared with a fixed number: the specification or acceptance criterion. Fail the number, and you have an OOS.
• OOT is compared with a pattern: what earlier results for the same product, test and conditions would lead you to expect. Break the pattern, and you may have an OOT, even if every result is still inside the specification.

An OOS is a failure today. An OOT is an early warning that a failure may be coming, or that something in the method, the process or the storage has changed.

What is an OOS result?

FDA’s guidance Investigating Out-of-Specification (OOS) Test Results for Pharmaceutical Production (Revision 1, May 2022) uses “OOS” for any test result that falls outside the specifications or acceptance criteria set in a drug application, a drug master file, an official compendium, or by the manufacturer. The guidance covers chemistry-based laboratory testing of active ingredients, excipients, in-process materials and finished products, including testing against stability specifications. It is non-binding guidance, but inspectors use it as the benchmark.

The key rule is that an OOS result must not be put down to “analytical error” until an investigation has established a documented laboratory root cause. The full procedure is in our OOS investigation complete guide, and a ready-to-adapt format is in the OOS investigation form template.

What is an OOT result?

There is no single, worldwide regulatory definition of OOT. Regulators expect you to notice unexpected trends, but each company writes its own rule for what counts as “out of trend”. In practice an OOT result is a result that is still within specification but does not follow the expected pattern, for example when compared with earlier time points of the same stability batch, or with other batches of the same product.

Two regulatory anchors are worth knowing:

• EU GMP Part I, Chapter 6 (Quality Control), in operation since 1 October 2014: §6.7 lists, among the documents the QC department should have, a procedure for the investigation of out-of-specification and out-of-trend results. §6.9 says that some kinds of data, such as test results, yields and environmental controls, should be recorded in a manner that permits trend evaluation, and that out-of-trend or out-of-specification data should be addressed and investigated. §6.35, in the on-going stability section, says that out-of-specification results or significant atypical trends should be investigated.
• MHRA, in its GOV.UK page Out-of-specification investigations and the attached presentation Out of Specification & Out of Trend Investigations (October 2017), treats OOS, OOT and atypical results together. It describes an OOT result generally as a stability result that does not follow the expected trend. See our post on OOT results and MHRA expectations.

FDA’s OOS guidance is written for OOS results and does not set a numeric OOT rule, but a footnote in its introduction notes that much of the guidance may be useful when examining out-of-trend results.

Atypical (aberrant) results

An atypical or aberrant result is an unexpected result that may be inside the specification, for example a chromatogram with an unexpected peak, or a single replicate far from the others. It is not automatically OOT, because a trend needs several data points. But an atypical result is a reason to look closely at the run, and it is often how a problem is first seen.

OOS vs OOT: comparison table

Three worked examples

The numbers below are made up for teaching. They are illustrative and are not taken from any real product or regulatory limit.

Example 1: an assay OOS

A tablet has an assay specification of 98.0 to 102.0 percent of label claim. An analyst reports 96.8 percent for a release batch. This is outside the limit, so it is an OOS. The analyst stops, informs the supervisor, and does not discard the solutions or start retesting on a hunch. Phase I begins: check the calculation, the standard and sample weights, dilutions, the system suitability, the instrument and the chromatography data. If a documented laboratory cause is proved, the result may be invalidated and the test repeated as the SOP allows. If not, the 96.8 percent stands as a valid result and a full-scale Phase II investigation begins, including production and sampling.

Example 2: a stability impurity that is OOT but not OOS

A degradation impurity has a limit of NMT 0.50 percent. The batch shows 0.08, 0.10 and 0.12 percent at 0, 3 and 6 months, then 0.31 percent at 9 months. Every value is inside the limit, so there is no OOS. But the jump from 0.12 to 0.31 does not match the earlier slope of this batch, and the other batches of the product stayed close to 0.15 percent at 9 months. This is an OOT candidate. The team checks whether the sample and chamber records are in order, whether the integration and the method behaved normally, and whether the result is repeatable under the SOP. If the rise is real, QA decides whether to increase testing, assess shelf-life impact and open a CAPA, well before the 0.50 percent limit is reached.

Example 3: a dissolution trend

A product has a dissolution requirement of not less than 80 percent (Q) in 30 minutes. Average results across stability time points are 96, 94, 91, 86 and 83 percent at 0, 3, 6, 9 and 12 months. All are above 80 percent. Earlier batches stayed within about 3 percentage points of their starting value over the same period. The slow, steady fall is an adverse trend. It is not an OOS today, but a regression line would reach the limit before the expiry period ends. This is exactly the situation EU GMP §6.35 describes as a significant atypical trend that should be investigated. Possible causes include moisture uptake, a change in packaging or a change in the excipient grade.

How OOT is detected

Because no regulator prescribes one OOT formula, you choose a method that suits your data, justify it and write it in an SOP. Common approaches are generic:

• Control charts for release or in-process data: plot results in time order against a centre line and control limits, and flag points beyond the limits or unusual runs. Shewhart-type rules such as one point beyond three standard deviations, or several consecutive points on one side of the centre line, are typical examples.
• Regression for stability data: fit a line to the results of a batch over time and flag a new point that falls outside the prediction interval of earlier points. ICH Q1E, Evaluation of Stability Data, describes regression as a way of evaluating quantitative stability attributes.
• By time point: compare each new result with the results of earlier batches at the same time point.

Example in-house criterion: flag a stability result as potential OOT if it lies outside the 95 percent prediction interval of the regression line built from the batch’s earlier time points, or if the change between two consecutive time points is more than three times the typical change seen in earlier batches. These numbers are only an illustration of how a company might write its rule; the actual criteria must be justified for your product and method.

Records and data that support a batch should be kept at least one year after the expiry date of the respective product batch (EU GMP Part I Chapter 4 §4.11; 21 CFR 211.180(a)), or longer where local law or the marketing authorisation requires.

Investigation flow

For an OOS result

1. Stop and report to the supervisor. Preserve all solutions, preparations and raw data.
2. Phase I, laboratory investigation: the analyst and supervisor check for an assignable laboratory cause against a checklist, before any decision about the result.
3. If a documented laboratory cause is proved, invalidate the result as the SOP allows and repeat the test.
4. If not, the result is valid. Move to Phase II, full-scale investigation, led by the quality unit, covering production, sampling, raw materials and any extra laboratory work defined in advance.
5. Make the batch decision, extend the assessment to other batches and products, and plan CAPA. See the CAPA guide.

MHRA describes the laboratory stage as Phase Ia and Ib, with Phase II and III afterwards. The details are in the OOS complete guide.

For an OOT result

1. Confirm the data: transcription, calculation, correct time point, correct condition and correct batch.
2. Check the method and system: system suitability, instrument, standards and analyst, using records already available.
3. Compare with the trend method defined in your SOP. Decide whether the result is a real change or normal variation.
4. If it is a real trend, QA widens the evaluation to process, materials, packaging and storage, assesses the impact on shelf life and on batches already on the market, and decides on extra testing, increased monitoring or CAPA. EU GMP §6.35 expects confirmed OOS results or significant negative trends affecting marketed batches to be reported to the competent authorities.
5. Record the conclusion and the reasoning, even when the answer is “normal variation”.

For stability work and the study design behind trending, see the stability study guide (ICH Q1A).

Common mistakes

• Testing into compliance: running more tests until a passing number appears. The number of retests must be fixed in an SOP before testing, not adjusted by the result.
• Averaging a failing result with passing retests. FDA’s guidance warns that averaging can hide variability; the individual results should be evaluated.
• Blaming “analyst error” with no evidence. A laboratory cause must be demonstrated and documented.
• Ignoring OOT because “it passed”. A drifting trend is the cheapest warning you will ever get.

OOS vs OOT interview questions

1. What is the main difference between OOS and OOT? OOS is outside the specification limit; OOT is within the limit but does not fit the expected pattern of earlier data.

2. Is every OOT an OOS? No. An OOT may never become an OOS. But a trend left unexplained can lead to one.

3. Which regulation requires OOT investigation? EU GMP Chapter 6 asks for a procedure for OOS and OOT investigation (§6.7) and for investigation of out-of-trend data (§6.9). There is no single defined OOT limit; the rule is set in your SOP.

4. Can you retest an OOS sample? Only as the approved SOP allows, with the number of retests fixed in advance, and never to replace a valid failing result with a passing one.

5. How would you detect OOT in stability data? Compare each new time point with a regression line or prediction interval from earlier data or earlier batches, using the method defined in the SOP.

Frequently asked questions

Does OOT mean the batch has failed?

No. An OOT result inside the specification does not by itself fail a batch. It triggers an evaluation, and the conclusion may be normal variation or a real trend that needs action.

Do we need Phase I and Phase II for OOT?

The OOS phase terms belong to OOS. For OOT, MHRA’s material indicates a proportionate investigation along similar lines; the depth depends on risk. Define the steps in your SOP.

Is OOT only for stability data?

No. Stability is the most common use, but release results, in-process data, yields and environmental data can also be trended, as EU GMP §6.9 suggests.

References

1. US FDA, Guidance for Industry: Investigating Out-of-Specification (OOS) Test Results for Pharmaceutical Production, Revision 1, May 2022 (CDER). fda.gov/media/158416/download.
2. European Commission, EudraLex Volume 4, EU GMP Part I Chapter 6: Quality Control (in operation from 1 October 2014), §§6.7, 6.9, 6.32, 6.35. health.ec.europa.eu.
3. European Commission, EudraLex Volume 4, EU GMP Part I Chapter 4: Documentation, §4.11. health.ec.europa.eu.
4. MHRA, GOV.UK: Out-of-specification investigations, first published 28 August 2013, updated 26 February 2018; attachment “Out of Specification & Out of Trend Investigations”, October 2017. gov.uk.
5. ICH, Q1E: Evaluation of Stability Data, Step 4, 6 February 2003. database.ich.org.
6. US FDA, 21 CFR 211.180(a), General requirements for records. ecfr.gov.