Journal of Chromatography B 1109 (2019) 76–83

Contents lists available at ScienceDirect

Journal of Chromatography B
journal homepage: www.elsevier.com/locate/jchromb

Development of a validated method for the qualitative and quantitative T

analysis of cannabinoids in plant biomass and medicinal cannabis resin
extracts obtained by super-critical fluid extraction
Aaron C. Elkins , Myrna A. Deseo, Simone Rochfort, Vilnis Ezernieks, German Spangenberg
⁎

Agriculture Victoria Research, AgriBio, Centre for AgriBioscience, Bundoora, VIC 3083, Australia

ARTICLE INFO ABSTRACT

Keywords: The social push for the therapeutic use of cannabis extracts has increased significantly over recent years.
Medicinal cannabis Cannabis is being used for treatment for conditions such as epilepsy, cancer and pain management. There are a
UHPLC range of medicinal cannabis products available, but the use of cannabis resin obtained by super critical fluid
Cannabinoids extraction, often diluted in oil, is becoming increasingly more prominent. Much of the research on cannabis has
Quantitation
focused on plant biomass or the final therapeutic product with a concerning lack of information on the inter-
Cannabis resin
Cannabidiol (CBD)
mediate resin. This study aims to bridge the gap between current methods of analysis for biomass and the final
therapeutic product by describing a fully developed and validated ultra-high-performance-liquid-chromato-
graphy method with diode array detection (UHPLC-DAD) for the qualification and quantification of the can-
nabinoids CBDA, CBD, CBN, THC, CBC and THCA, in medicinal cannabis biomass and resin obtained by super-
critical fluid extraction (SFE). The method was validated for specificity, linearity, limit of detection (LOD), limit
of quantitation (LOQ), precision, accuracy, robustness, spike recovery and stability in accordance with the
Validation of Analytical Procedures: Text and Methodology Q2 to meet the requirements of the International
Council for Harmonisation (ICH), Therapeutic Goods Authority (TGA) and the Food and Drug Administration
(FDA) test method validation regulations.

1. Introduction and antioxidant [4,8,9]. THC, responsible for the psychoactive prop-
erties [2,5,9,15], has been approved to control nausea and vomiting in
In recent years the social push for the therapeutic use of medicinal cancer treatment, appetite stimulation in AIDS patients [2,6,16,17] and
cannabis resin extracts has increased despite the limited number of in the treatment of glaucoma [18], migraine [19,20], anxiety and as an
clinical studies [1]. To date > 500 compounds, 120 cannabinoids, have analgesic [6,9].
been identified from cannabis plant material, Cannabis sativa L. (Can- In 2009 the United Nations Office on Drugs and Crime listed gas
nabaceae). Cannabinoids originate from the condensation of olivetolic chromatography mass spectrometry (GCMS) and GC coupled with a
acid, in the polyketide pathway; and geranyl pyrophosphate, from the flame ionisation detector (GC-FID) as recommended methods for the
methylerythritol pathway; to form cannabigerolic acid (CBGA). CBGA is analysis of cannabis and cannabis products [21]. The restriction with
then catalysed by three oxidative cyclases to form cannabidiolic acid using GC in the analysis is that derivatisation of the acidic cannabinoids
(CBDA), tetrahydrocannabinolic acid (THCA), and cannabichromic acid is required or they are likely to be decarboxylated in the GC injection
(CBCA) [2,3]. CBDA and THCA are decarboxylated to form cannabidiol port due to the temperatures required to volatilise the sample for
(CBD) and delta-9-tetrahydrocannibinol (THC), the two major canna- analysis [2,4,6]. As a result, the reportable levels of cannabinoid are
binoids of pharmacological interest [2,4–6] as the precursors, CBDA based solely on the neutral species and reported as a combined total
and THCA are not capable of passing the blood-brain barrier and value.
therefore considered pharmacologically inactive [7–9]. CBD has been Separation using ultra high performance liquid chromatography
reported to have the highest number of medicinal benefits having been (UHPLC) is preferred to GC as both the acid variant and neutral can-
applied in the treatment and management of epilepsy [6,10,11], as an nabinoid are able to be independently quantitated [2]. It is important to
antipsychotic [12,13], in anxiety management [14], and as an analgesic be able to measure the degree of decarboxylation if the final product is

⁎
Corresponding author at: Agriculture Victoria, AgriBio, Centre for AgriBioscience, 5 Ring Road, Bundoora, VIC 3083, Australia.
E-mail address: [email protected] (A.C. Elkins).

https://doi.org/10.1016/j.jchromb.2019.01.027
Received 7 December 2018; Received in revised form 29 January 2019; Accepted 29 January 2019
Available online 30 January 2019
1570-0232/ © 2019 Elsevier B.V. All rights reserved.
A.C. Elkins et al. Journal of Chromatography B 1109 (2019) 76–83

in oil form and will not be further heated by the consumer. There are cannabis inflorescences of a proprietary strain using a BioBotanical
several published HPLC methods for cannabinoid analysis, some of super-critical fluid liquid CO2 extractor (SFE) (Waters Corporation,
which require mobile phase at a specific pH [5,22], within 0.1 pH units Milford, MA). Resin samples were transferred to a pre-weighed vial and
[22], to obtain reproducible and reliable chromatographic separations. diluted 1:125 with methanol. Samples were sonicated for 5 min to so-
There are more robust methods that report that gradient profiles lubilise the resin. An aliquot was transferred for HPLC analysis.
[8,17,23–25] or simple isocratic mobile phase conditions, without pH
manipulation, can be employed to provide adequate separation 2.3. Method optimisation by LCMS
[9,26–28] while acknowledging the challenges of differentiating can-
nabinoids with similar chemical-physical properties. Coupling the Method development was performed on a Thermo Scientific
HPLC system with MS can improve the selectivity of the analysis due to (Waltham, MA) Q Exactive Plus Orbitrap mass spectrometer (MS)
the high sensitivity and resolving power of the MS. The enhanced coupled with Thermo Scientific Vanquish ultra-high-performance-li-
sensitivity of MS compared to DAD can also be advantageous for de- quid-chromatography (UHPLC) system equipped with degasser, binary
tection and quantitation of minor cannabinoids such as CBG and CBC. pump, temperature controlled autosampler and column compartment,
While the focus on utilising cannabis extract as a treatment option is and diode array detector (DAD). UHPLC-MS-DAD was used to ensure
increasing there is no published fully validated standard method for the separation of fourteen standards with no co-elution. The optimised
quantitation of cannabinoids in biomass or in cannabis resin extracts method focusing on the target compounds; CBDA, CBD, CBN, THC, CBC
intended for therapeutic use at the active pharmaceutical ingredient and THCA; was transferred to an Agilent Technologies (Santa Clara,
(API) stage. Research studies often focus on the plant material CA) 1290 UHPLC system for routine analysis with the method valida-
[5,6,18,22,24], synthetic cannabinoids [23,28], biological samples tion completed to meet the guidelines of the Therapeutic Goods
[17,27], cannabinoid isolates [26] or the final product in the presence Authority (TGA).
of oil for patient administration [8,9] without thought to the resin
produced during the extraction process. 2.4. Instrumentation and data analysis
The objective of this study was to develop and validate a UHPLC-
DAD quantitative method for the analysis of CBDA, CBD, CBN, THC, Analysis of both cannabis biomass and cannabis resin was per-
CBC and THCA in cannabis inflorescences harvested from five week old formed using an Agilent 1290 UHPLC system equipped with degasser,
plants; and medicinal cannabis resin obtained by super critical fluid binary pump, temperature controlled autosampler and column com-
extraction of the harvested inflorescences; in accordance with the partment, and UV DAD. Agilent OpenLab software (v2.1.0.433) was
Validation of Analytical Procedures: Text and Methodology Q2 to meet used for instrument control, data acquisition and data processing. All
the requirements of the International Council for Harmonisation (ICH) samples were analysed using a Phenomenex Luna Omega C18
and the Food and Drug Administration (FDA) test method validation 150 × 2.1 mm × 1.6 μm column with an injection volume of 3 μL. The
regulations [29] under Good Manufacturing Practice (GMP). column temperature was maintained at 30 °C. Chromatographic se-
paration was achieved using a multistep gradient consisting of solvent A
2. Materials and methods (water with 0.1% formic acid) and solvent B (acetonitrile with 0.1%
formic acid). The gradient program was conducted as follows: 0–2 min
2.1. Reagents and standards 40% B, 2–3 min 15–75% B, 3–10 min 75–90% B, 10–11 min 90–100%
B, 11–15 min 100% B, followed by equilibration to initial conditions for
All HPLC grade reagents, water with 0.1% formic acid (mobile 5 mins at a flow rate of 0.3 mL/min. The UV DAD acquisition wave-
phase A), acetonitrile with 0.1% formic acid (mobile phase B) and length was set to 280 nm providing significantly better linear range for
methanol were obtained from Fisher Scientific (Fair Lawn, NJ, USA). all standards when compared to the more commonly used 210 nm. UV
Primary standards for CBDA and THCA in acetonitrile, and CBD, CBN, spectra from 190 to 640 nm also collected for peak characterisation and
CBC, THC in methanol, at 1000 μg/mL, were commercially purchased purity assessment. Baseline separation for all cannabinoids was suc-
from Novachem Pty Ltd. (Heidelberg West, Australia) as distributor for cessfully achieved with described conditions for the standards, biomass
Cerilliant Corporation (Round Rock, Texas, USA). A mixed stock stan- and resin samples (Figs. 1, 2 and 3). Extracted spectrum λmax values of
dard at 125 μg/mL CBDA, CBN, CBC, THCA and 250 μg/mL CBD, THC the cannabinoid standard peaks were compared to the corresponding
in methanol was prepared with working standards at 0.5, 2.5, 5, 12.5, sample peak to assess purity.
25, 50, 100, and 125 μg/mL for CBDA, CBN, CBC and THCA; and 1, 5,
10, 25, 50, 100, 200 and 250 μg/mL for CBD and THC prepared as serial 2.5. Method validation
dilutions from the mixed stock. All standards were stored at −80 °C.
Method validation was conducted in accordance with the Validation
2.2. Sample preparation of Analytical Procedures: Text and Methodology Q2 to meet the re-
quirements of the International Council for Harmonisation (ICH) and
Dried and ground cannabis inflorescences were obtained from the the Food and Drug Administration (FDA) test method validation reg-
Victorian Government Medicinal Cannabis Cultivation Facility. ulations. The method was validated for specificity, linearity, accuracy
Inflorescences of two cultivars were used: one high CBD cultivar re- and precision, robustness, limit of detection (LOD), limit of quantitation
ferred to as Biomass 1 and one high THC cultivar referred to as Biomass (LOQ), and stability at acquisition wavelength 280 nm. All assessments
2. Biomass samples were cured at 120 °C for 2 h facilitating the con- were carried out using Agilent OpenLab validated software
version of CBDA to CBD and THCA to THC. Cured samples were ground (v2.1.0.433).
to a fine powder with liquid nitrogen using a SPEX SamplePrep 2010 To assess stability biomass and resin samples were exposed to UV
Geno/Grinder for 1 min at 1500 rpm. After grinding, 10 mg of each and visible light (photolysis) and thermally treated at 60 °C for 1 h.
sample was weighed into an Axygen 2.0 mL microcentrifuge tube on a Comparison was made to an untreated control sample to assess the
Sartorius BP210D analytical balance. Each sample was extracted with method specificity in the presence of products from the forced de-
1 mL of methanol, vortexed for 30 s, sonicated for 5 min and cen- gradation.
trifuged at 13,000 rpm for 5 min. The supernatant was transferred to a Extraction recovery was determined for cured biomass (10 mg) and
2 mL amber HPLC vial and diluted 1:3 for analysis. extracted resin (50 mg). Samples were spiked with a 1 mg/mL stock
Cannabis resin samples were provided by the Victorian Government solution to give final concentrations of 25 μg/mL (LS). 50 μg/mL (MS),
Medicinal Cannabis Cultivation Facility. Resin was extracted from and 100 μg/mL (HS) with comparison made to an un-spiked sample

77
A.C. Elkins et al. Journal of Chromatography B 1109 (2019) 76–83

Fig. 1. Chromatogram of 100 μg/mL standard used for system suitability and quantitation.

(NS). NS, LS, and MS samples were stored at −80 °C while HS samples 3. Results and discussion
were stored at ambient room temperature (20 °C). Samples (NS, LS, MS
and HS) and standards were analysed weekly across four weeks to de- System suitability was assessed to verify the resolution, column ef-
termine stability. ficiency and repeatability of the Agilent 1290 UHPLC, in accordance
with ICH guidelines. A summary of the system suitability parameters
and results are detailed in Table 1. The data was analysed against re-
2.6. Statistical analysis commendations detailed in the Centre for Drug Evaluation and Re-
search (CDER) Validation of Chromatographic Methods as a part of the
Samples exposed to forced degradation conditions (photolysis and United States Pharmacopeia [30] to comply with guidelines set by the
thermal) were compared to the control sample using a paired t-test with FDA and the ICH. All data was acquired at wavelength 280 nm. Ac-
p < 0.05 indicating statistical significance. ceptance criteria for retention time ± 0.2 min, resolution ≥ 2.0,

Fig. 2. Chromatogram of biomass 1 methanol extract.

78
A.C. Elkins et al. Journal of Chromatography B 1109 (2019) 76–83

Fig. 3. Chromatogram of super-critical fluid extracted cannabis resin.

tailing ≤ 2.0, k (retention factor) > 2, Plate number > 2000. System The biomass data suggests that the thermally treated samples show
suitability is clearly demonstrated in that the method exceeds these no sign of degradation whilst biomass samples exposed to UV and
acceptance criteria. The sub 2 μm column displays excellent retention visible light show a statistically significant (p = 0.04) decrease in CBD
time reproducibility (average variance is less that 0.02 of a minute) and concentration. Similar results were observed for THC although this was
plate number. The high plate number was expected due to the superior not statistically significant at p < 0.05. In resin, the levels of CBD and
separation power of modern UHPLC columns (with sub 2 μm particle THC are slightly reduced as in the biomass however this is not statis-
size) compared to older 5 μm HPLC columns. Column efficiency is in- tically significant (at p < 0.05). A peak purity test was performed
versely proportional to the particle size, as the particle size reduces the confirming each cannabinoid peak corresponds to one analyte. CBDA,
separation power increases. Another advantage of using the UHPLC CBD, THC, CBC and CBN are all deemed stable in the resin when ex-
columns are that for analysis times can be significantly reduced when posed to forced degradation conditions.
compared to HPLC columns [31].

3.2. Linearity, LOD and LOQ

3.1. Specificity
A series of known concentration chemical standards were prepared
The specificity and selectivity of the method was assessed by de- by serial dilution of the stock standard to assess linearity. All standards
termining if the cannabinoids of interest were impacted by the presence LOD and LOQ were determined at signal-to-noise ratios of 3:1, and 10:1
of known impurities, degradants, or the sample matrix. Table 2 de- respectively. Regression values, linear concentration range of the
scribes cannabinoid content in the samples which were exposed to working standards, LOD and LOQ are reported in Table 3. Previous
photolysis or thermally treated compared to a control. Known de- studies acquire data at wavelength 220 nm [5,25] or 235 nm [22] re-
gradation products include CBN from the oxidative aromatisation of sulting in lower LOD than reported but also has a decreased linear range
THC, cannabielsion (CBE) from pyrolysis of CBD, and cannabicyclol with an upper limit of 50 μg/mL [5,22]. Reported LOD and LOQ are
(CBL) formed by photolysis during storage [2,3]. CBN and CBC have subsequently lower at 220 nm compared to 280 nm however from the
been included in the analysis method as known impurities; CBE and cultivars analysed the quantitated values for CBDA, CBD, CBN, THC,
CBL have not been observed and are unlikely to be generated with CBC, and THCA are well within the range of this method. Analysis at
current methodology. 280 nm also decreases interference from impurities. Previously reported

Table 1
Summary of system suitability data across four weeks (n = 40).
Cannabinoid Retention Time (min) Resolution Tailing Factor k (retention factor) Plates

CBDA 6.89 ± 0.004 NA⁎ 1.0 ± 0.051 4.6 ± 0.007 110,787

CBD 7.38 ± 0.008 5.8 ± 0.031 1.1 ± 0.037 5.0 ± 006 117,855
CBN 8.69 ± 0.008 14.8 ± 0.065 1.0 ± 0.033 6.1 ± 0.007 143,938
THC 9.56 ± 0.008 9.2 ± 0.027 1.0 ± 0.041 6.8 ± 0.006 158,493
CBC 10.34 ± 0.016 8.1 ± 0.047 1.0 ± 0.037 7.4 ± 0.007 183,865
THCA 10.58 ± 0.017 2.4 ± 0.034 1.0 ± 0.032 7.6 ± 0.007 178,289

⁎
NA = Not applicable.

79
A.C. Elkins et al. Journal of Chromatography B 1109 (2019) 76–83

Table 2
Specificity - change in cannabinoid concentration after photolysis and thermal treatment (n = 4 samples per treatment).
Biomass 1 (μg/mL) Biomass 2 (μg/mL) Resin (μg/mL)

Control Photolysis Thermal Control Photolysis Thermal Control Photolysis Thermal

CBDA ND ND ND ND ND ND 2.34 2.30 2.21

CBD 57.87 52.62a 56.89 ND ND ND 72.12 69.24 67.24
CBN ND ND ND ND ND ND 0.40 0.39 0.38
THC 3.04 2.74 2.99 24.67 21.20 24.51 48.02 46.76 45.06
CBC 42.86 36.06 40.98 28.76 25.43 28.43 4.78 4.87 4.63
THCA ND ND ND ND ND ND ND ND ND

ND = not detected
a
CBD statistically significant degradation (p = 0.04) when exposed to forced degradation by photolysis.

Table 3
Linear concentration range, correlation coefficient, LOD and LOQ for cannabinoid standards tested.
Standard Concentration (μg/mL) Equation R2 LOD (μg/mL) LOQ (μg/mL)

CBD 1, 5, 10, 25, 50, 100, 200, 250 y = 1.5935x 0.99981 1.00 2.50
THC 1, 5, 10, 25, 50, 100, 200, 250 y = 1.8643x 0.99987 1.00 2.50
CBDA 0.5, 1, 5, 10, 25, 50, 100, 200, 250 y = 10.3529x 0.99986 0.15 0.50
CBN 0.5, 1, 5, 10, 25, 50, 100, 200, 250 y = 32.0981 0.99972 0.02 0.05
CBC 0.5,1, 5, 10, 25, 50, 100, 200, 250 y = 15.6780x 0.99962 0.10 0.33
THCA 0.5,1, 5, 10, 25, 50, 100, 200, 250 y = 14.4303x 0.99988 0.15 0.50

Table 4
Accuracy and precision of the method across 9 determinations at 3 different concentrations.
Compound CBD THC CBDA CBN CBC THCA

Mean conc. (μg/mL) 24.99 24.97 12.49 12.49 12.49 12.49

%RSD 1.05 1.04 1.04 1.01 1.05 1.03
% Conc. of standard 99.95% 99.87% 99.90% 99.91% 99.90% 99.91%
Mean conc. (μg/mL) 49.80 49.92 25.02 25.00 25.02 25.00
%RSD 0.72 0.64 0.55 0.65 0.59 0.63
% Conc. of standard 99.59% 99.85% 100.10% 99.90% 100.07% 99.97%
Mean conc. (μg/mL) 98.71 99.17 49.66 49.65 49.77 49.60
%RSD 0.49 0.66 0.56 0.73 0.82 0.41
% Conc. of standard 98.71% 99.17% 99.32% 99.30% 99.54% 99.21%

Table 5
Robustness, comparison of the method between analysts.
Compound CBD CBDA THC THCA

Analyst 1 Analyst 2 Analyst 1 Analyst 2 Analyst 1 Analyst 2 Analyst 1 Analyst 2

Mean (μg/mL) 100.00 100.00 100.00 100.00 100.00 99.99 100.00 99.99
%RSD 0.38 0.52 0.41 0.53 0.41 0.59 0.41 0.48

Comparison analyst 1 and analyst 2

Mean (μg/mL) 100.00 100.00 99.99 99.99
%RSD 0.44 0.46 0.49 0.43

HPLC-MS values for linear range between 0.5 and 100 ng/mL, LOD and 8–20% for MS analysis [9,17]. All of these methods fail the ac-
0.2 ng/mL and 0.5 ng/mL [9,17] well below the levels observed in ceptance criteria (≤2.0%RSD) specified by the ICH and FDA for the
biomass samples and extracted resin making UHPLC-DAD preferable to analysis of an active pharmaceutical ingredient (API).
HPLC-MS.
3.4. Robustness
3.3. Precision and accuracy
The robustness of the method was validated by two independent
The accuracy and precision of the method was assessed by analysing analysts who prepared independent system suitability standards at
three different standard concentrations (100 μg/mL, 50 μg/mL and 100 μg/mL CBD and THC and 50 μg/mL CBDA, CBN, CBC and THCA.
25 μg/mL CBD and THC; 50 μg/mL, 25 μg/mL, 12.5 μg/mL CBDA, CBN, The analyses were performed on two different UHPLC systems, on two
CBC and THCA) over three injections for a total of 9 determinations different columns (of the same type), on two different days. The second
(Table 4). All standards were ≤1% RSD with all determinations < 2% system used for robustness assessment was a Thermo Scientific
of the theoretical concentration. These results are well within the ac- (Waltham, MA, USA) Vanquish UHPLC system equipped with degasser,
ceptable limit of 20%. The precision of this method exceeds previously binary pump, temperature controlled autosampler and column com-
reported values range between 2 and 6% for DAD analysis [5,22,25] partment, and UV DAD detector. The method was determined to be

80
A.C. Elkins et al. Journal of Chromatography B 1109 (2019) 76–83

Table 6
Spike recoveries from two different biomass cultivars and super-critical fluid extracted cannabis resin.
Cannabinoid Sample conc. (μg/mL) Spiked amount (μg/mL) Total amount (μg/mL) % Recovery %RSD

Biomass 1
CBDA ND 25 26.03 104.1 4.1
50 49.57 99.1 0.9
100 97.41 97.4 2.6
CBD 57.87 25 81.97 96.4 1.1
50 107.37 99.0 0.5
100 153.65 95.8 2.7
THC 3.04 25 27.34 97.2 1.1
50 51.54 97.0 2.8
100 97.38 94.3 5.5
THCA ND 25 24.30 97.2 2.8
50 47.15 94.3 5.7
100 95.07 95.1 4.9

Biomass 2
CBDA ND 25 25.06 100.2 0.2
50 48.99 98.0 2.0
100 98.35 98.4 1.6
CBD ND 25 24.48 97.9 2.1
50 49.90 99.8 0.2
100 100.13 100.1 0.1
THC 24.67 25 48.60 95.7 2.1
50 71.79 94.3 3.8
100 116.70 92.0 6.4
THCA ND 25 24.27 97.1 2.9
50 47.51 95.0 5.0
100 95.26 95.3 4.7

Resin

Cannabinoid Sample conc. (μg/mL) Spiked amount (μg/mL) Total amount (μg/mL) % Recovery %RSD

CBDA 2.34 25 28.17 103.3 3.1

50 51.09 97.5 2.4
100 100.61 98.3 1.7
CBD 72.12 25 97.49 101.5 0.4
50 121.03 97.8 0.9
100 170.57 98.5 0.9
CBN 0.40 25 24.80 97.6 2.4
50 49.43 98.1 1.9
100 101.76 101.4 1.4
THC 48.02 25 73.39 101.5 0.5
50 98.16 100.3 0.2
100 149.08 101.1 0.7
CBC 4.78 25 29.23 97.8 1.8
50 54.32 99.1 0.8
100 107.53 102.8 2.6
THCA ND 25 25.17 100.7 0.7
50 48.59 97.2 2.8
100 98.01 98.0 2.0

resin was measured by comparing the theoretical and determined

Table 7
Stability of standards over a four-week assessment period (n = 30). concentrations. Samples were spiked with known concentration of
mixed standard and extracted with 1 mL methanol. Spike recovery for
Cannabinoid Mean conc. (μg/mL) Std. Dev. %RSD
the extraction process range from 92 to 104% for biomass samples and
CBDA 47.74 1.28 2.67 97–103% for resin (Table 6). Initial biomass extraction methods re-
CBD 96.87 2.38 2.45 quired 200 mg of plant material extracted with 90% methanol 10%
CBN 49.16 0.42 0.85 chloroform solvent that was agitated for 30 mins, dried and recon-
THC 95.78 2.81 2.93 stituted in 1:1 water/methanol solution for analysis providing re-
CBC 49.51 0.44 0.88
THCA 48.05 1.25 2.60
coveries of 90–110% [5]. The reconstitution solvent was then modified
to 65% methanol 35% water to increase chromatographic separation
and peak shape while having no effect on overall recovery [22]. Mudge
robust as a concentration difference of < 0.5% RSD between analysts et.al [25] were able to achieve recovery results from 90 to 99% using
was measured (Table 5). The robustness of the method between ana- 200 mg of biomass extracted with 25 mL of 80% methanol/water that is
lysts has not been previously described but the similarity of the inter- sonicated for 15 min with vortexing every 5 min removing chloroform
day results performed by different analysts further supports the validity from the process. Our method, 10 mg biomass extracted with 1 mL of
of this method. methanol, vortexed for 30 s, sonicated for 5 min and centrifuged at
13,000 rpm for 5 min, significantly increases sample throughput redu-
3.5. Recovery cing extraction time by 10 min with similar recovery values. There have
been no reported methods for the extraction and analysis of medicinal
The efficiency of the solvent extraction on biomass 1, biomass 2 and cannabis resin.

81
A.C. Elkins et al. Journal of Chromatography B 1109 (2019) 76–83

Table 8
Biomass and super-critical fluid extracted cannabis resin stability over a four-week assessment period (n = 20).
NS LS MS HS

Cannabinoid Mean conc. (μg/mL) %RSD Mean conc. (μg/mL) %RSD Mean conc. (μg/mL) %RSD Mean conc. (μg/mL) %RSD

Biomass 1
CBDA ND ND 25.33 1.64 48.28 1.54 90.38 5.04a
CBD 56.79 1.29 79.73 1.72 104.87 1.4 149.22 2.03
THC 2.99 1.26 27.07 1.57 50.45 1.29 93.43 2.49
THCA ND ND 23.54 1.87 45.71 1.76 83.19 9.18a

Biomass 2
CBDA ND ND 24.25 2.36 48.71 1.5 92.25 4.73a
CBD ND ND 23.21 3.28 49.17 1.56 96.82 2.63
THC 24.14 1.29 47.03 2.18 71.24 1.57 109.97 4.2
THCA ND ND 23.49 2.04 47.14 1.84 83.35 9.6a

Resin
CBDA 2.22 3.28 27.14 2.6 49.57 2.41 91.32 7.31a
CBD 69.47 3.15 94.35 2.89 118.09 2.93 166.25 2.51
CBN 0.39 1.72 24.58 0.92 49.4 1.06 101.42 0.66
THC 46.02 3.53 70.5 3.36 95.23 2.98 150.35 3.01
CBC 4.65 1.78 29.33 0.93 54.87 0.86 107.95 1.18
THCA ND ND 24.47 2.41 47.53 2.37 87.89 8.29a

a
HS samples stored at ambient temperature show increased rate of degradation compared to samples stored at −80 °C.

3.6. Stability with the ICH and FDA test method validation regulations for the ana-
lysis of an active pharmaceutical ingredient under GMP requirements.
Plant biomass, cannabis resin (NS, LS and MS) and standards were The use of DAD opposed to MS for the analysis of cannabis biomass and
stored at −80 °C with HS samples stored at ambient room temperature resin has several advantages in a quality control laboratory. The com-
(20 °C). These were analysed weekly and compared to a freshly pre- paratively low cost of the instrument and maintenance combined with
pared standard for a period of four weeks to assess stability. The results the superior reproducibility makes DAD the instrument of choice for
were extremely consistent indicating a range of 0.8–2.9%RSD for the routine analysis. The simple extraction process, minimal amount plant
calculated the standard; between 1.2 and 3.3% for the cured biomass; material and resin required makes this method ideal for sample analysis
and between 0.6 and 3.5% for the resin when stored at −80 °C. Storage in a GMP environment.
at ambient temperature significantly increases the rate of degradation
of CBDA (5.0% and 4.7% RSD) and THCA (9.2 and 9.6% RSD) in par- Conflict of interest
ticular (Tables 7 and 8).
Contradictory stability data has been reported for biomass samples Declarations of interest: none.
to date. Patel et.al [22] reports that extracts are stable for up to 48 h at
room temperature with a %RSD up to 7.1; while Citti et.al [9] reports Funding
that after three consecutive freeze-thaw cycles from −20 °C cannabi-
noids in oil and ethyl acetate are stable with RSD up to 8.2%. Both This research did not receive any specific grant from funding
authors claim that the acceptable limits are up to 15%RSD. Conversely, agencies in the public, commercial, or not-for-profit sectors.
Mudge et.al [25] suggest that samples should be considered unstable if
there is a > 5% loss from the initial analysis. They determined that Acknowledgements
there was significant change to the cannabinoid standard, particularly
CBDA and THCA, concentration when stored at 22 °C for 30 h whereas Authors would like to thank the staff at the Victorian Government
mixed standards stored at −20 °C were degraded after 48 h. Interest- Medicinal Cannabis Cultivation Facility for the supply of the dried
ingly, they also found that standards stored at 4 °C for 72 h were stable. ground plant biomass that has been used for validation of the test
Sample extracts in 80% methanol were not stable at 22 °C but stable for method.
48 h at 4 °C. There is no reported stability data on cannabis resin.
Based on our findings standards, methanol biomass extracts and References
cannabis resin is stable across the four week assessment period when
stored at −80 °C. [1] Medicinal Cannabis - Guidance Documents, Department of Health, Therapeutic
Goods Administration, 2018.
[2] S. Vladimir-Knežević, B. Blažeković, M. Bival Štefan, M. Kindl, Phytochemical
Aspects and Therapeutic Perspective of Cannabinoids in Cancer Treatment, (2017).
4. Conclusion [3] S. Farag, O. Kayser, The Cannabis Plant: Botanical Aspects, (2017).
[4] L. Allegra, H.Z. L., S.K. A., A review of methods for the chemical characterization of
cannabis natural products, J. Sep. Sci. 41 (2018) 398–415.
A fully validated UHPLC-DAD method for the analysis of cannabi-
[5] B. De Backer, B. Debrus, P. Lebrun, L. Theunis, N. Dubois, L. Decock, A. Verstraete,
noids in biomass samples and unadulterated cannabis resin obtained by P. Hubert, C. Charlier, Innovative development and validation of an HPLC/DAD
super-critical fluid extraction has been developed and rigorously tested method for the qualitative and quantitative determination of major cannabinoids in
for the quantitative determination of the cannabinoids CBDA, CBD, cannabis plant material, J. Chromatogr. B 877 (2009) 4115–4124.
[6] W. Mei, W. Yan-Hong, A. Bharathi, R.M. M., W.A. S., M. Zlatko, A. John, E.M. A.,
CBN, THC, CBC and THCA across a four-week period. The use of a K.I. A., Quantitative determination of cannabinoids in cannabis and cannabis pro-
multi-step gradient improves the resolution of all cannabinoid species ducts using ultra-high-performance supercritical fluid chromatography and diode
investigated, negating the need for mass spectrometry peak differ- array/mass spectrometric detection, J. Forensic Sci. 62 (2017) 602–611.
[7] S.H. Burstein, The cannabinoid acids: nonpsychoactive derivatives with therapeutic
entiation, particularly when comparing CBC and THCA elution. potential, Pharmacol. Ther. 82 (1999) 87–96.
Specificity, linearity, precision, accuracy, robustness, spike recoveries [8] C. Citti, M. Battisti Umberto, D. Braghiroli, G. Ciccarella, M. Schmid, A. Vandelli
and stability assessments have all been tested and proven compliant Maria, G. Cannazza, A metabolomic approach applied to a liquid chromatography

82
A.C. Elkins et al. Journal of Chromatography B 1109 (2019) 76–83

coupled to high-resolution tandem mass spectrometry method (HPLC-ESI-HRMS/ therapeutic implications in medicine and headache: what a long strange trip It's
MS): towards the comprehensive evaluation of the chemical composition of can- been …, J. Headache Pain 55 (2015) 885–916.
nabis medicinal extracts, Phytochem. Anal. 29 (2017) 144–155. [20] B.C. Lochte, A. Beletsky, N.K. Samuel, I. Grant, The use of Cannabis for headache
[9] C. Citti, G. Ciccarella, D. Braghiroli, C. Parenti, M.A. Vandelli, G. Cannazza, disorders, Cannabis Cannabinoid Res. 2 (2017) 61–71.
Medicinal cannabis: principal cannabinoids concentration and their stability eval- [21] United Nations Office of Drugs and Crime, Recommended Methods for the
uated by a high performance liquid chromatography coupled to diode array and Identification and Analysis of Cannabis and Cannabis Products, United Nations
quadrupole time of flight mass spectrometry method, J. Pharm. Biomed. Anal. 128 Publications, New York, 2009.
(2016) 201–209. [22] B. Patel, D. Wene, Z.T. Fan, Qualitative and quantitative measurement of canna-
[10] O. Devinsky, J.H. Cross, L. Laux, E. Marsh, I. Miller, R. Nabbout, I.E. Scheffer, binoids in cannabis using modified HPLC/DAD method, J. Pharm. Biomed. Anal.
E.A. Thiele, S. Wright, Trial of cannabidiol for drug-resistant seizures in the Dravet 146 (2017) 15–23.
syndrome, N. Engl. J. Med. 376 (2017) 2011–2020. [23] D. Borg, A. Tverdovsky, R. Stripp, A fast and comprehensive analysis of 32 synthetic
[11] B.E. Porter, C. Jacobson, Report of a parent survey of cannabidiol-enriched can- cannabinoids using agilent triple quadrupole LC–MS-MS, J. Anal. Toxicol. 41
nabis use in pediatric treatment-resistant epilepsy, Epilepsy Behav. 29 (2013) (2017) 6–16.
574–577. [24] W. Peschel, M. Politi, 1H NMR and HPLC/DAD for Cannabis sativa L. chemotype
[12] C. Rohleder, J.K. Müller, B. Lange, F.M. Leweke, Cannabidiol as a potential new distinction, extract profiling and specification, Talanta 140 (2015) 150–165.
type of an antipsychotic. A critical review of the evidence, Front. Pharmacol. 7 [25] E.M. Mudge, S.J. Murch, P.N. Brown, Leaner and greener analysis of cannabinoids,
(2016) 422. Anal. Bioanal. Chem. 409 (2017) 3153–3163.
[13] A.C. Campos, M.V. Fogaça, A.B. Sonego, F.S. Guimarães, Cannabidiol, neuropro- [26] C.E. Layton, A.J. Aubin, Method validation for assay determination of cannabidiol
tection and neuropsychiatric disorders, Pharmacol. Res. 112 (2016) 119–127. isolates, J. Liq. Chromatogr. Relat. Technol. 41 (2018) 114–121.
[14] M.V. Fogaça, A.C. Campos, L.D. Coelho, R.S. Duman, F.S. Guimarães, The anxiolytic [27] A. Molnar, S. Fu, J. Lewis, D.J. Allsop, J. Copeland, The detection of THC, CBD and
effects of cannabidiol in chronically stressed mice are mediated by the en- CBN in the oral fluid of Sativex(R) patients using two on-site screening tests and LC-
docannabinoid system: role of neurogenesis and dendritic remodeling, MS/MS, Forensic Sci. Int. 238 (2014) 113–119.
Neuropharmacology 135 (2018) 22–33. [28] M. Popławska, A. Błażewicz, K. Kamiński, E. Bednarek, Z. Fijałek, L. Kozerski,
[15] A.K. Hewavitharana, G. Golding, G. Tempany, G. King, N. Holling, Quantitative GC- Application of high-performance liquid chromatography with charged aerosol de-
MS analysis of Δ9-tetrahydrocannabinol in fiber hemp varieties, J. Anal. Toxicol. 29 tection (LC–CAD) for unified quantification of synthetic cannabinoids in herbal
(2005) 258–261. blends and comparison with quantitative NMR results, Forensic Toxicol. 36 (2018)
[16] M. Mücke, M. Weier, C. Carter, J. Copeland, L. Degenhardt, H. Cuhls, L. Radbruch, 122–140.
W. Häuser, R. Conrad, Systematic review and meta-analysis of cannabinoids in [29] ICH, Q2 (R1), Validation of analytical procedures: text and methodology, ICH
palliative medicine, J. Cachexia. Sarcopenia Muscle 9 (2018) 220–234. Harmonised Tripartite Guideline, International Conference on Harmonisation of
[17] R. Jamwal, A.R. Topletz, B. Ramratnam, F. Akhlaghi, Ultra-high performance liquid Technical Requirements for Registration of Pharmaceuticals for Human Use,
chromatography tandem mass-spectrometry for simple and simultaneous quantifi- Chicago, USA, 2005.
cation of cannabinoids, J. Chromatogr. B 1048 (2017) 10–18. [30] < 621 > Chromatography, United States Pharmacopeial Convention, Inc,
[18] M.A. ElSohly, D. Slade, Chemical constituents of marijuana: the complex mixture of Rockville, Md, 2016.
natural cannabinoids, Life Sci. 78 (2005) 539–548. [31] J.R. Mazzeo, U.D. Neue, M. Kele, R.S. Plumb, Advancing LC performance with
[19] P. Baron Eric, Comprehensive review of medicinal marijuana, cannabinoids, and smaller particles and higher pressure, Anal. Chem. 77 (2005) 460 (A-467 A).

83