Laboratory Analysis of THC Content in Industrial Hemp Seed
An MRAC supported project Manitoba Rural Adaptation Council Inc.
508-800 Portage Avenue Winnipeg, MB. R3G 0N4
Final Report Prepared By:
Hemp Oil Canada Inc.
PO Box 188, Ste. Agathe, MB. R0G 1Y0
Shaun Crew, President
Telephone: (204) 275-7616
Email: hempoilcan@escape.ca
March 10, 2000
Laboratory Analysis of THC Content in
Industrial Hemp Seed
An MRAC supported project
Project Purpose
The theory was that hemp seeds do not contain THC, but rather that the seed coats are
contaminated by the leaf matter surrounding the seed bud. However, this had never
been scientifically proven in a laboratory setting. Our project purpose was to determine if
hemp seeds do intrinsically contain THC by analyzing the whole seed, seed hulls, nut
membrane and hemp nut. This analysis would be conducted on the seed components
derived from actual hemp food processes (seed cleaning and dehulling) and by carefully
dissecting whole hemp seeds in a laboratory environment. Triplicate subsamples would
be analyzed in duplicate for the six hemp seed varieties being tested in this research
project. The results of this analysis will ultimately provide the answer as to whether THCFree status can be realistically be attained by commercial processes presently being
researched & developed by Hemp Oil Canada Inc.
Determining the level of THC, if any, in the various compositional parts of a hemp seed
would provide important information for,
(1) developing processes to eliminate and/or reduce the THC contaminating the seed,
(2) establish the realistic benchmark that would be attainable and
(3) assisting regulatory bodies in reviewing and establishing the acceptable levels of
THC in hemp food products.
The theoretic elimination of THC is thought to be the step necessary to elevate hemp
food products into the mainstream and gain the attention of the major food processing
companies. The degree to which hemp food markets develop will be directly related to
the hemp industry’s ability to eliminate or reduce the present trace levels of THC, below
that which could be detected in a drug screening test.
Determining the level of THC which can be ingested without failing a drug screening test
was not part of this project. However, it should be noted that several research projects
are underway in both Canada and the US to answer this question. The results of that
outside research & analysis will assist hemp food processors in determining the
maximum levels of THC that would be permissible in hemp food products, while at the
same time keeping these levels under those allowable by existing government
regulations.
Summary
The research work required to complete this project was provided by the applicant,
Hemp Oil Canada Inc. along with the project partner, Websar Laboratories Inc., of Ste.
Anne, Manitoba who provided the analytical services, sample preparation, seed
dissection and consultation during the project. Each company provided both direct and
“in kind” financial support along with the direct financial support provided by MRAC.
The research work conducted by the project participants determined that:
(1) THC is intrinsically found in all parts of the hemp seed, albeit at far lower levels in the
hemp seed nut than that found on the seed coats which receive the highest degree
of contamination.
(2) The levels of THC will vary between the different OECD approved industrial hemp
seed varieties.
(3) The detected levels of THC will vary between whole hemp seeds, hemp seed hulls,
hemp seed nuts, hemp fines and hemp screenings in each individual variety.
(4) The levels of THC in the various components of a hemp seed can be reduced to
below 1.0 mg/g (parts per million) when the whole seeds are properly prepared
(cleaned) for processing. This applied to both the manually and mechanically
separated fractions.
(5) In the majority of the samples analyzed, the THC levels were typically higher in the
mechanically cleaned & processed components as compared to the manually
dissected seed parts.
Attached to this report is the “final report” of the hemp seed analysis performed by
Websar Laboratories Inc., identifying the description of samples analyzed, sample
preparation, samples extraction, results & discussion and the conclusions drawn.
In conclusion, it would appear from the results that achieving a “THC Free” status is not
achievable in terms of a true zero. However, in terms of a relative zero, properly
prepared whole hemp seeds can be cleaned and mechanically processed in such a way
that the levels of THC are reduced to below 1.0 mg/g and in a number of cases to below
0.5 mg/g. Health Canada has regulated and limited the content of THC in hemp seed
derivatives to 10 mg/g (10 parts per million).
Attached is the analytical data compiled by Websar Laboratories in support of this
project.
Hemp Seed Dissection Project (with MRAC support to Hemp Oil Canada Inc.)
Introduction
There has been an ongoing controversy regarding the presence of D9-THC in hemp seed oil and the hemp seed nuts the oil is pressed from. Several studies have shown that consumption of commercially available hemp seed oil results in urine specimens testing positive for D9-THC in the USA (Callaway et al. 1997, Constantino et al. 1997, Struempler et al. 1997), Germany (Meier and Vonesch 1997) and Switzerland (Lehmann et al. 1997). Other studies have reported marijuana-positive urine test results following consumption of hemp seed derived food products such as snack bars (Alt and Reinhart 1998, Fortner et al. 1997) and hempen ale (Gibson et al. 1998). No information was given regarding the procedural details of seed cleaning prior to pressing to recover the commercial oils; however, the hempen ale was prepared from seed which had been washed twice before the brewing process.
The ongoing question concerns the origin of the D9-THC in the hemp seed nuts and hemp seed products. The presence in hemp seed nuts of D9-THC could severely limit the market potential for product derived from the processing of hemp seed nuts. On the other hand, if the nuts are essentially free of D9-THC, there is a significant market for oil pressed from those nuts, oil that can be consumed without the risk of the consumer testing positive for marijuana use.
Objectives of the Study
The objective of the study was to establish the intrinsic level of D9-THC in hemp seed nuts from six (five in the original proposal) varieties of industrial hemp. The intrinsic D9-THC content of the nuts determines the limit below which it is difficult to go for D9-THC in hemp seed oil. The D9-THC content in the hulls of the cleaned seeds indicates the residual D9-THC which oil production processing has to contend with.
Several tasks were undertaken to answer this question. The first task was to determine D9-THC levels in whole seed prior to cleaning to establish base-line levels of D9-THC for each variety. Analysis of mechanically dehulled hemp nut and hulls was also undertaken as was the analysis of manually dehulled hemp nut and the hulls so obtained.
Description of Samples
Seed samples of industrial hemp seed were submitted to Websar Laboratories, Inc. on October 13, 1999 representing the following six hemp varieties: Fedora 19, USO 14, Felina 34, Fin 314, Fasamo and Ferimon 12. Information on each variety is given in Table 1.
A second set of samples, this time of mechanically dehulled seed, was submitted on December 22, 1999:
· Fasamo (hulled hemp seed)
· Fin 314 (hulled hemp seed)
· Mixed French varieties (Ferimon, Fedora, & Felina) (hulled hemp seed)
These hulled samples had not yet been subjected to air washing to remove the hull pieces and the whole seeds from the mechanical dehulling. The hulled seed was manually separated from the mixture in each case to enable analysis of the nut and hulls. All separated material was stored in 15-mL amber bottles with Teflon lined screw-caps.
A third set of samples, this time of fractions from mechanical dehulling, was submitted on
January 18, 2000:
· Fin fines
· Fin screenings
· Fin nut
· Fasamo fines
· Fasamo screenings
· Fasamo nut
· Zolo fines
· Zolo screenings
· Zolo nut
The Zolo variety was not one of the original five varieties planned for the study.
Samples of Felina 34, Fedora 19, Ferimon 12, and USO 14 submitted on October 13, 1999 were given to Shaun Crew of Hemp Oil Canada Inc. to enable him to prepare mechanically dehulled nut and hull samples for analysis January 18, 2000. These samples were received for analysis by
Websar Laboratories Inc. on January 28, 2000.
Sample Preparation
Whole Seed Analysis
Three replicate samples of whole seed, weighing 10 grams, were taken from each bulk variety sample and separately ground in a coffee mill. Each ground sample was then transferred to an amber storage bottle (with a teflon-lined screw cap) and held at 4o C until analysis could be carried out. The coffee mill was cleaned between each sample.
Manual Dissection Fractions (nut and hulls)
Individual seeds were removed from the lot and viewed under a 3x magnifier. Only nuts that were undamaged were used. The seed was opened using a scalpel along the midpoint of the two halves and the seed was cut right through. The nut was removed using tweezers and placed in a screw capped amber bottle (see above). The hulls were placed in separate amber bottles. Enough seed was dissected to allow for triplicate sample lots. Duplicate samples from each lot were analyzed for each variety. All utensils used for manipulation of the seed, nut, and hulls were rinsed in toluene after each use on individual seeds. It was impracticable to remove the seed membrane from each seed; however, at the time of sampling for analysis, any membrane material that could be removed was manually removed before extraction.
The samples were stored at 4o C until analysis. Each sample was ground in a coffee mill prior to analysis. Mechanically Dehulled Seed fractions (nut, hulls, fines, and screenings)
These samples were prepared, extracted, and analyzed as whole seed samples.
Sample Extraction and Analysis A 200 mg subsample from each sample of the replicate set for each variety was weighed out for analysis in duplicate using the protocol for analysis of D9-THC in hemp oil and seed cake as developed by Websar Laboratories, Inc. Fresh solutions of 5% ethanolic potassium hydroxide were prepared for the saponification step. Final extract volume was 1mL. D9-THC-d3 was used as the internal standard. Analysis was carried out using a Varian Saturn 2000 GC/MS calibrated using against a 10 mg/g D9-THC calibration standard. Instrument performance must be within specification using control standards at both 4 mg/g and 30 mg/g before analysis can proceed. Injection volume was 1 mL. Duplicate injections were made for each analysis.
Results and Discussion
Whole Seed
Results of the analyses of whole seed used in the study are presented in Table 2.
Triplicate subsamples were analyzed in duplicate for each of the six hemp seed varieties
submitted for the study. Each of these extracts, in turn, was analyzed using duplicate injections, resulting in averaged instrument responses for six replicate samples per variety.
Some variation in D9-THC levels among replicates was noticed, especially in the case of Ferimon 12. This variation would be consistent with D9-THC contamination of the outer seed coating through dust where dust deposit would not be expected to be uniformly dispersed throughout the seed sample.
During the manual dissection, it was observed that there was a hard waxy coating on the hulls of the seeds which crazed when the dissection was being initiated. The waxy coating was removed with the hulls and was analyzed as part of the hulls fraction.
Manual Dissection Results of the analyses of manually dissected seeds are given in Table 3.
Triplicate subsamples were analyzed in duplicate for each of the six hemp seed varieties
submitted for the study. Each of these extracts, in turn, was analyzed using duplicate injections, resulting in averaged instrument responses for six replicate samples per variety.
Levels of D9-THC in the dissected out nut from the six varieties were in general low, i.e., well below 1 mg/g (ppm). The lowest level was in the Fedora 19 and the lowest variability was also seen here.
The relatively high level reported for Fin 314 nut must be viewed with the knowledge
that the variation is high. Hulls in general show higher levels of D9-THC than the nut, except for USO 14, where the levels are essentially equivalent.
Mechanically Dehulled Seed
Results for mechanically dehulled seed show (these results replace those for the combined French varieties in the Final Report) that
a) the mechanically dehulled nuts appear to contain somewhat more D9-THC than do the
manually dehulled nuts, except in the case of Fin 314. The differences are statistically
different, however, only in the case of Fedora and Fasamo.
b) the levels in the hulls are higher than those in the mechanically dehulled nuts in the case of Fedora, Felina 34, and Fin 314, and are essentially equivalent to those in the mechanically dehulled nuts for USO 14, Ferimon 12, and Fasamo.
c) the levels in the fines are higher than in the hulls for Felina 34, essentially the same for USO 14, Fin 314, and Ferimon 12, and lower for Fedora, and Fasamo.
d) the levels in the screenings are higher than in the fines for Fedora, Felina 34, Fin 314,
Fasamo, and Ferimon 12, and essentially the same for USO 14.
e) Hulls in all cases contain less D9-THC thane the screenings.
f) Fines in all cases (except Felina 34 where the fines are higher) contain essentially the same levels of D9-THC as the mechanically dehulled nuts.
The results are also shown in the six figures (one for each variety) at the end of the report.
The mechanically dehulled Zolo nut contained 0.125 ± 0.05 mg/g; the fines contained 0.25 ± 0.13 mg/g, and the screenings contained 0.27 ± 0.11 mg/g.
Conclusion
The results show that it will be feasible to prepare hemp seed products with levels of D9-THC below 1.0 mg/g, and for some varieties, well below 1.0 mg/g with the exception of Ferimon 12. The fines from the mechanical dehulling process also fit this category for Fedora, Fin 314, and Fasamo; the fines for USO 14, Felina 34, and Ferimon 12 do not. For each variety, the screenings also do not.
To see the tables click here and scroll to end of doc
