Overview. Cannabis laboratory investigations require discipline because the product is heterogeneous and the analytical methods may be matrix-sensitive. An unexpected result may arise from the batch, the sample, the method, the instrument or the analyst. The investigation should identify the most scientifically credible explanation without testing into compliance.
Immediate assessment
The first review should verify calculations, standards, reagents, system suitability, equipment status, sequence integrity, sample preparation and analyst execution. Obvious assignable causes should be supported by evidence.
OOS versus OOT
An out-of-specification result exceeds an approved criterion. An out-of-trend result may remain within specification but differ materially from expected history. Both may indicate loss of control.
Cannabis-specific laboratory causes
Poor homogenisation, incomplete extraction, low water-activity sample handling, volatile terpene loss, matrix suppression, cannabinoid co-elution and unstable standards are common analytical risks.
Sampling and batch heterogeneity
Where the sample may not represent the batch, the investigation should assess the approved sampling plan and collection records. Resampling should address a documented hypothesis rather than serve as a second chance.
Hypothesis testing and retesting
Additional testing should test a specific explanation. Retesting plans should be predefined, scientifically justified and approved. Averaging original and repeat results to obtain a passing mean is generally inappropriate.
Manufacturing investigation
If no clear laboratory cause is demonstrated, the investigation should expand to cultivation, drying, mixing, extraction, filling, packaging, storage and supplier data.
Batch disposition and CAPA
The quality unit should consider all valid results and the strength of the root-cause evidence. CAPA should address the underlying failure, such as sampling design, training, method robustness or process control.
Practical reference table
| Unexpected result | Possible cause | Evidence to review |
|---|---|---|
| Low cannabinoid assay in flower | Non-representative or poorly homogenised sample | Sampling record and particle-size study |
| High CBN | Oxidation or ageing | Stability, storage and oxygen exposure |
| High microbial count | True contamination or handling error | Controls, media, sampling and facility history |
| Residual solvent failure | Process removal issue or vial leak | Manufacturing trend and headspace controls |
| Pesticide detection | Cultivation input or carryover | Supplier, field and blank data |
Control and decision path
Frequently asked questions
Can a passing retest invalidate an original OOS?
No. The original result remains valid unless a scientifically demonstrated assignable cause invalidates it.
What is an OOT result?
A result that differs unexpectedly from historical or process trends while possibly remaining within specification.
When is resampling acceptable?
When the original sampling process is shown or credibly suspected to be unrepresentative and a justified plan is approved.
Should all repeat injections be reported?
All relevant data and attempts should remain traceable and reviewable.
Who decides batch disposition?
The independent quality authority defined by the pharmaceutical quality system and applicable law.
Primary references and guidance
- FDA Guidance for Industry: Investigating OOS Test Results
- EU GMP Part I, Chapter 6
- MHRA OOS and Data Integrity expectations
- PIC/S PE 009
- ICH Q9(R1)
- ICH Q10
- EU GMP Annex 11
- European Pharmacopoeia general methods
Confirm the current effective revision and national applicability before operational or regulatory use.