Out-of-Specification (OOS) Investigations in Peptide Testing

Educational information for a laboratory audience. Not medical advice, not a recommendation for human use. Peak Labs products are for laboratory research use only.

A certificate of analysis reports a number: a purity percentage, an identity confirmation, a moisture value. What it rarely shows is what happened when a number first came back wrong. Every analytical laboratory occasionally produces a result that falls outside its own acceptance criteria. What separates a rigorous testing operation from a careless one is not the absence of these events. It is how the laboratory investigates them before deciding what to report.

What Counts as an Out-of-Specification Result

An out-of-specification, or OOS, result is any test outcome that falls outside the acceptance criteria set for that assay, whether the criterion comes from a compendial method, an internal specification, or a validated in-house procedure. The term is distinct from an "out-of-trend" result, which stays within specification but deviates from a batch's historical pattern, and from an anomalous or atypical result observed during a test that has not yet produced a final value. In peptide testing, an OOS might appear as an HPLC purity figure below the stated threshold, a mass spectrometry result that does not match the expected molecular weight, or a moisture value outside the range expected for a lyophilised product.

Why the Investigation Matters More Than the Number

A single out-of-range number can have several origins: a genuine problem with the material, an error in sample preparation, a transient instrument fault, or an analyst mistake. Reporting the number without investigating its source treats all four causes as equivalent, which they are not. A structured investigation is what allows a laboratory to distinguish a true batch failure from a laboratory-generated artifact, and to do so with a documented, defensible rationale rather than a guess. This is the same logic that underlies reading a certificate of analysis critically: the number on the page is only as trustworthy as the process that produced it.

Phase 1: Laboratory Investigation

The first phase asks a narrow question: is there an identifiable, documented laboratory error that explains the result? This involves reviewing the analyst's technique, checking instrument performance logs and calibration status, confirming reagent and mobile phase preparation, re-examining the raw data (chromatograms, spectra, integration parameters) for anything inconsistent with correct method execution, and verifying that the sample itself was handled and prepared correctly. If a clear, documented assignable cause is found at this stage, for example a diluent error or a documented instrument malfunction with a maintenance record to support it, the result can be invalidated on that specific basis and retesting proceeds under a written protocol.

Phase 2: Full-Scale Investigation

If the laboratory-level review does not identify an assignable cause, the investigation broadens. This second phase looks beyond the immediate analysis to the manufacturing or synthesis history of the material itself: process records, raw material lots, in-process controls, and any deviations recorded earlier in production. Additional testing may be performed, including retesting of the original sample and, where justified, testing of a new sample. Retesting is not a mechanism for testing until a passing result appears; a sound investigation defines in advance how many retests are justified and how results will be interpreted, consistent with the statistical basis of the method.

Common Root Causes Distinguished During Investigation

• Sample preparation error: incorrect dilution, weighing error, or degradation of the prepared sample before analysis.
• Instrument or method variability: a system suitability failure, column degradation, or detector drift that affected the result independent of the sample itself.
• Genuine material nonconformance: the batch itself does not meet specification, which is a valid and reportable outcome, not a failure to be hidden.
• Documentation or transcription error: a correct result recorded or transferred incorrectly, distinct from an actual analytical problem.

Distinguishing between these requires objective evidence, not inference. A laboratory that cannot point to a specific, documented cause has no basis for invalidating a result, and the original value stands.

What Happens After the Investigation Concludes

A completed investigation ends in one of three outcomes: the original result is invalidated for a documented reason and replaced by a valid retest, the original result is confirmed and the batch is dispositioned accordingly (which may mean rejection, rework where scientifically justified, or downgrading), or the investigation is inconclusive, in which case regulatory and quality guidance generally treats the original result as valid. None of these outcomes involves quietly substituting a more favorable number. The investigation record, not the final figure alone, is what demonstrates that a laboratory's quality system functioned as intended.

Documentation a Research Buyer Can Reasonably Expect

A buyer evaluating a supplier's quality practices is not usually entitled to see every internal investigation report, but a supplier with a genuine quality system should be able to describe, in general terms, how it handles atypical results: whether it follows a written OOS or deviation procedure, whether investigations are documented and reviewed by someone independent of the original testing, and whether a batch with an unresolved or confirmed nonconformance is released for sale. The absence of any such procedure, or a supplier unable to describe one, is itself informative and pairs naturally with the kind of certificate scrutiny covered in frequently asked questions about testing and documentation.

How This Differs from Simply Rejecting or Hiding a Result

An investigation process is not a mechanism for making inconvenient data disappear. Two practices distinguish a legitimate OOS investigation from data manipulation: the original result is never deleted or overwritten, only invalidated with a documented, reviewable rationale, and the number of retests permitted is defined by a written procedure before the investigation begins, not chosen after seeing how a retest turns out. Buyers who understand this distinction are better equipped to interpret the small variances that can appear in reported values, a topic covered in more depth alongside the concept of measurement uncertainty in COA reporting, and to recognize the warning signs described in a related discussion of certificate of analysis red flags.

Sources and further reading

• FDA guidance document database, including "Investigating Out-of-Specification (OOS) Test Results for Pharmaceutical Production"
• ISO/IEC 17025:2017, General requirements for the competence of testing and calibration laboratories
• ICH Quality Guidelines, including Q10 Pharmaceutical Quality System

Research use only. Peak Labs products are supplied strictly for in-vitro laboratory research. They are not medicines or supplements, are not for human or veterinary use, and are not intended to diagnose, treat, cure, or prevent any condition.