Cannabis products have been plagued by unreliable content labels and inaccuracies in the quantification of cannabinoids. Excluding fraudulent cases of intentional data manipulation, what can explain result-discrepancies between well-meaning and/or accredited laboratories? Why is ISO 17025 not enough? What further standards should the industry seek to apply?
ISO is an abbreviation of the International Standards Organisation. They are an independent body made up of an extensive network of individuals who are experts in different areas. An ISO standard is essentially an internationally recognised way of doing something. It means that everyone follows the same set of guidelines no matter where they are based, resulting in a safer, more consistent end result. This benefits both the organisation and the customer or end user. For companies, they are secure in the knowledge that this standard is followed and recognised worldwide. For customers, they know they are getting a product or service which is safe, good quality and trustworthy.
The ISO 17025 accreditation is the most important standard for testing and/or calibration laboratories. It confirms that a particular laboratory is able to produce precise and accurate test results and calibration data by evaluating, tracking and scoring key metrics and categories, including:
The tools by which it achieves this are documentation, processes and specifications combined into a Quality Management System (QMS). The QMS is what guides the analysts’ every action when performing a measurement and is designed to cover and record every stage of the process in a traceable manner.
Cannabis flos is a chemically complex matrix, with hundreds of volatile and semi-volatile compounds. Storing & handling conditions can have a significant impact on its chemical profile. When measuring cannabinoids, it is customary to provide the final result in weight-to-weight ratio (w/w %). This means that the total measured cannabinoids in mg will be given as a ratio to total sample weight used. It is therefore crucial to standardise the characterisation of the initial sample weight quantification: what is being weighted and at which stage of sample prep.? What ratio of flos to total biomass? What moisture content? This is a multifaceted question involving a number of concepts and processes, including sampling methodologies, sample preparation and uncertainty calculations. Last but not least, quantification of total cannabinoid content following cold-methodologies that do not decarboxylate the acids depend on a simplistic equation meant to calculate total cannabinoids by accounting for the loss carboxyl group. However, as was pointed out in detail by Dr. Roggen, this method fails to account for the loss of mass in the sample itself and therefore ultimately makes use of the wrong ratio to present the final result.
As any analytical laboratory scientist will testify, most time in the lab is spent in sample preparation activities. ‘Sample preparation’ refers to the actions taken by the analyst to prepare the sample for instrument injection. In the case of cannabis flos & cannabinoids, sample preparation will involve numerous steps, which can include: weighing, grinding, extraction, centrifugation / stirring, filtration, concentration and/or dilution and vial transfer. Aligning laboratories in their uncertainty assessments along the whole sample preparation and testing lifecycle is going to play a crucial part in achieving interlaboratory results consistency.
ISO 17025 is by design addressing internal consistency over objective assessment. This is not a criticism but an important feature to recognise when trying to diagnose and troubleshoot the discrepancies in cannabinoid laboratory results. After all, if one is performing a miscalculation, consistently, the relative error is not visible within the dataset. An external reference point is needed. Given the lack of interlaboratory testing programs as well as availability of cannabis reference materials, the ISO’s internal consistency bias is emphasized. The cumulative potential of the uncertainty accrued at every stage of the analysis, as well as definitional hazards is very significant and will account for most of the non-fraudulent laboratory results discrepancies.