Microbiological OOS results: investigation, evidence and CAPA

A microbiological OOS cannot be resolved by selecting the most favourable result. An investigation must explain what happened, how reliable the measurement is and which activities or products may be affected. Biological variability belongs in the assessment; it is not a ready-made cause.

1. Classify the event before interpreting it

An out-of-specification (OOS) result fails an applicable specification. An out-of-trend (OOT) result departs from the expected pattern and may still meet specification. An invalid result needs evidence justifying invalidity under the method and procedure; it does not mean an unwelcome result.

An environmental monitoring excursion belongs to its monitoring programme. Annex 1, 2022, §9.13 distinguishes responses to alert levels and action limits. Record test, matrix, current criterion, units, context and initial classification. For a positive sterility test, also use the dedicated investigation guide.

2. Preserve evidence and activity status

Notify responsible functions and apply established measures for potentially affected samples, activities and batches. Retain original data, calculations, available images, contemporaneous records, metadata and relevant change history. Protect useful samples and materials within stability, safety and approved-procedure constraints; unsuitable storage can change what you intend to examine.

Build a timeline covering collection, receipt, preparation, incubation, reading, communication and decisions. Separate observed facts from later reconstruction. Record who performed or checked each step and what can no longer be verified.

3. Examine the laboratory through hypotheses

Review sample representativeness and transport, matrix, method version, suitability, media, controls, equipment, incubation and reading. WHO good practices for pharmaceutical microbiology laboratories, 2011 provide a specific reference for these elements. Finding a deviation does not itself demonstrate that it caused the result.

The public summary of USP <1227>, public 2019 version connects microbial recovery with suitability in the presence of product. The full current chapter criteria were not accessed. Check the authorised method and applicable compendial text rather than inventing neutralisation criteria.

4. Interpret counts and isolate information

A count represents colony-forming units recovered under the method’s conditions, not every microorganism present. Uneven distribution, aggregates and the tested aliquot affect interpretation. Do not discard a high value simply by calling it “microbiological variability”.

Identification can guide an investigation, but finding the same species at two locations does not alone prove a common source. Assess discriminatory power, identification quality, timing and process connections. For limitations and records to retain, see microbial identification systems.

5. Extend the scope beyond one test

EU GMP Chapter 6, operative since 2014 requires OOS and OOT data to be addressed and investigated. Connect findings to relevant production, raw materials, shared equipment, other samples and batches, environmental trends and outsourced activities. Do not wait for every laboratory check to close if evidence already calls for process action.

Simulated case. A potentially interfering matrix yields an OOS count. Process contamination, contamination during testing, heterogeneity and counting error are plausible. Interference that reduces recovery does not automatically explain a high count. The team compares suitability, original data, controls and batch history; it does not conclude “analyst error” by elimination.

6. Give further testing a defined purpose

Before repeating a test, define the hypothesis, sample, method, justified number of tests, criteria and interpretation of every outcome. Retesting an available sample and resampling a batch answer different questions; a new sample does not necessarily recreate the original condition.

The FDA OOS guidance, May 2022 covers chemistry-based testing of CDER-regulated drugs; it is not presented here as a universal microbiological protocol. Avoid testing until a desired result appears and indiscriminate averaging of OOS and passing results. A favourable repeat neither erases the original nor automatically permits release.

7. Connect CAPA to effectiveness checks

EU GMP Chapter 1, 2013, §1.4(xiv) requires cause analysis and assessment of action effectiveness; attributing a problem to human error needs justification without overlooking system failures.

CAPA structure: evidence → demonstrated cause or declared residual hypothesis → immediate correction → action addressing the cause → owner and deadline → indicator, period and effectiveness criterion → review. Repeating a training course alone does not prove resolution.

• Are classification, impact and initial measures documented?
• Are original data, alternatives and contradictory findings included?
• Does each additional test have an approved purpose?
• Are conclusions, residual uncertainty and batch decisions justified?
• Do CAPAs have effectiveness criteria and defined ownership?