Overview. Sampling is one of the most underestimated sources of uncertainty in cannabis quality control. A laboratory can produce a precise result from a sample that does not represent the batch. Botanical heterogeneity, localised contamination and variable moisture make sampling design critical.
What representative sampling means
A representative sample reflects the attributes and variability of the batch. The strategy should account for batch size, container count, production flow, segregation risk and known heterogeneity.
Flower and botanical heterogeneity
Cannabinoid concentration, moisture and microorganisms may vary between plants, branches, container locations and parts of a bulk lot. Stratified sampling across containers or positions is generally more defensible than a single grab sample.
Bulk extracts and oils
Liquids may appear homogeneous but can stratify, precipitate or retain concentration gradients. Mixing status, temperature, sampling location and hold time should be defined before sample collection.
Finished dosage forms
Capsules, bottles and sprays require sampling across the packaging run to detect start-up, middle and end-of-run variation. Statistical plans may be needed for content uniformity and packaging defects.
Microbiological and contaminant sampling
Localised contamination creates particular risk. Composite sampling can improve representativeness but may dilute a localised high result. The intended decision should determine whether individual or composite samples are appropriate.
Sample handling and chain of custody
Containers, transport, storage, labelling and hold time must preserve the sample. Light, heat, oxygen and moisture can change cannabis results between collection and analysis.
Sampling deviations
If a sampling error is suspected, the investigation should assess whether the original result remains valid. Resampling should be scientifically justified and not used simply to obtain a passing result.
Practical reference table
| Material | Main heterogeneity risk | Recommended approach |
|---|---|---|
| Bulk flower | Potency, moisture and microbes vary spatially | Stratified increments across containers |
| Milled flower | Segregation by particle size | Controlled mixing and multiple increments |
| Oil bulk | Concentration gradient or precipitation | Sample after validated mixing |
| Capsules | Fill drift across run | Beginning, middle and end sampling |
| Packaged flower | Container-to-container variation | Statistical container selection |
| Microbiology sample | Localised contamination | Risk-based individual/composite design |
Control and decision path
Frequently asked questions
Why is a grab sample often unsuitable for cannabis flower?
It may not capture spatial variation in potency, moisture or microbiology.
Should microbiological samples be composited?
Sometimes, but the risk of diluting localised contamination must be considered.
When should an oil bulk be sampled?
After validated mixing and under defined temperature and hold-time conditions.
Can resampling invalidate an OOS result?
No. Resampling requires scientific justification and does not automatically replace the original valid result.
Who should perform sampling?
Trained and authorised personnel following an approved plan and procedure.
Primary references and guidance
- EU GMP Part I, Chapter 6
- WHO Guidelines on Sampling of Pharmaceutical Products
- ISO 17025
- European Pharmacopoeia sampling principles
- WHO GACP
- ICH Q9(R1)
- Applicable national cannabis sampling rules
Confirm the current effective revision and national applicability before operational or regulatory use.