(You can have cannabis science content such as this delivered directly to your inbox. Simply sign up here for our MJBiz Science newsletter.)
The National Institute of Standards and Technology (NIST) recently published the first results from last year’s Cannabis Quality Assurance Program (CannaQAP).
It’s an overdue program that was initiated because of the need for reliable and proficient laboratory testing.
One objective of the program was to study laboratory performance to improve test-measurement comparability across laboratories and to “provide feedback about performance that can assist participants in improving laboratory operations.”
The July 2021 report represents the first of at least three well-planned exercises that call for laboratory participation across the industry (for example, independent labs, regulatory bodies and forensic labs).
For this exercise, NIST prepared a series of three well-characterized and statistically significant “homogeneous” hemp oils that contained up to 17 cannabinoids.
116 labs registered for participation, and up to 86% submitted data that were analyzed for the report.
Labs were requested to store hemp oil samples at -4 degrees Celsius, use a sample size of 0.5 grams and employ in-house methods for sample preparation and testing.
Laboratory test results would be compared to the NIST “target” means and ranges and to the general consensus.
Of the 17 cannabinoids that were analyzed, the most interesting results were from delta-9 THC and CBD:
- For delta-9 THC in Hemp Oil No. 1, results from 65 (82%) labs “completely overlap(s) the target range.”
- In Hemp Oil No. 2, results from 72 (78%) labs overlapped.
- In Hemp Oil No. 3, results from 11 (58%) labs overlapped approximately 85% of the target range.
Laboratory results were also compared to NIST “target” ranges for total CBD (i.e., CBD and CBDA).
For Hemp Oil No. 1, 48% of labs reported mass fractions (%) of CBD outside the NIST range, compared with 11% outside the NIST range for Hemp Oil No. 2.
Finally, 100% were outside the NIST range for Hemp Oil No. 3.
‘Wide range of variability’
According to the NIST report, between-laboratory variability for CBD was higher in Oil No. 3 than in Oils Nos. 1 and 2 and also notes that the variability “may be an artifact of significantly fewer laboratories reporting results.”
The CBD results showed “a fairly wide range of variability,” researcher Walter Brent Wilson told Marijuana Moment.
He said the variances likely result from differences in calibration. Different testing methods or storage conditions could also explain dissimilar results, he said.
“Everybody wants a true value for something, but the truth is that the true value is unknown,” Wilson told the online news oulet.
“Somebody tells you the concentration is 0.3% – that’s the estimated value. They don’t know what the true value is.”
Although this column is reporting on only one granule of the plethora of information that can (and should) be garnered from the NIST report, additional explanations for the behavior observed with CBD include laboratories’ tendencies to create and design calibration curves three or four magnitudes to capture all unknowns.
Laboratories must remember that Ordinary Least Squares Regression is only a statistical model that is based on assumptions. If any of those assumptions are violated, the calibration is questionable, regardless of the coefficient of determination.
Additionally, limited data is available around variability when hemp oil samples are diluted during sample preparation rather than analyzing the extractant.
And minimal variability of THC is expected in samples that were derived from hemp where the concentration is <0.3%.
The < 0.3% factor
This also is a good opportunity to consider the ramifications of laboratories that draw conclusions on THC concentration in marijuana samples where the only available reference materials are < 0.3%.
The NIST report is a great start to the federal understanding of challenges facing the cannabis testing laboratories, including – but certainly not limited to – a shortage of standard test methodologies, appropriate calibration (including calibration curves) programs and reference materials.
CannaQAP Exercise No. 2 was launched earlier this year focusing on cannabinoids, moisture and toxic elements in hemp and marijuana plant samples. The exercise closed last month, and a final report is expected by year’s end.
Exercise No. 3 is in planning and will assuredly build on the lessons learned from Exercises Nos. 1 and 2.
Let’s hope these early exercises are laying the foundation for a long-term, NIST-driven proficiency program.
Susan Audino, Ph.D., is a chemistry consultant and instructor for the American Association for Laboratory Accreditation. She is based in Ohio.