Journal of Analytical Toxicology, 2016;40:350–359

doi: 10.1093/jat/bkw027
Article

Article

Simultaneous Analysis of Cannabinoid and

Synthetic Cannabinoids in Dietary Supplements
Using UPLC with UV and UPLC–MS-MS

Downloaded from https://academic.oup.com/jat/article/40/5/350/1750552 by guest on 01 June 2021

Seok Heo†, Geum Joo Yoo†, Ji Yeon Choi, Hyoung Joon Park, Jung-Ah Do,
Sooyeul Cho, Sun Young Baek, and Sung-Kwan Park*
Advanced Analysis Team, National Institute of Food and Drug Safety Evaluation, Ministry of Food and Drug Safety,
Osong Health Technology Administration Complex, 187 Osongsaengmyeong2-ro, Osong-eup, Heungdeok-gu,
Cheongju-si, Chungcheongbuk-do 28159, Republic of Korea
*Author to whom correspondence should be addressed. Email: skpark@korea.kr
†
These authors contributed equally to the study.

Abstract
The primary purpose of this study was to develop and validate a method based on UPLC with UV and
UPLC–MS-MS for the simultaneous analysis of different cannabinoids and synthetic cannabinoids in
food as well as in herbal and dietary supplements. The limits of detection and quantitation of the
method ranged from 0.1 to 0.3 and 0.3 to 0.9 μg/mL by UPLC with UV, respectively. The coefficient
of determination was >0.999; the intra- and interday precision of the method were 0.1–3.7 and
0.9–4.1%, respectively. The intra- and interday accuracy were 94.8–103.1 and 98.3–100.9%, respect-
ively. The mean recoveries of nine cannabinoids obtained from tablet samples ranged from 81.1 to
105.4%. The mean extraction recoveries of nine target cannabinoids obtained from various types of
samples (tablets, capsules, powders, liquids, cookies and candies) ranged from 82.26 to 112.40%.
The relative standard deviation (RSD) of the stability of the prepared sample solutions was
<1.80%. Identification and quantification of the nine cannabinoids were accomplished by ion spray
UPLC–MS-MS using multiple reaction monitoring. The UPLC–MS-MS method was validated for
linearity (R 2 > 0.99); the precision was 0.1–4.0% (intraday) and 0.1–2.8% (interday), and the accuracy
was 98.0–103.5% (intraday) and 97.1–103.2% (interday). The mean extraction recoveries of six types
of samples were 82.2–114.5% and the RSD of stability was <6.54%, complying with the established
international guidelines. The results indicated that the method can be used for rapid and accurate
screening of cannabinoids present in food.

Introduction including some narcotic drugs and cannabinoids (Korea Customs

In recent years, drug abuse has garnered significant international Service).
attention, as the abuse of substances categorized as narcotics and can- Synthetic cannabinoids are added to herbal mixtures to induce
nabinoids has become a global concern. The emergence of new narco- psychoactive effects (2). Synthetic cannabinoids are chemical com-
tics, herbal products and dietary supplements containing synthetic pounds that act on the cannabinoid receptors to produce psychoactive
cannabinoids is widespread in global black markets (1). For example, effects (3). The side effects of cannabinoids are well documented, and
the Korea Customs Service reported that since the Incheon Airport the abuse of cannabinoids and their synthetic analogs represent a
Customs began analyzing dietary supplements in 2012, numerous serious problem (4, 5). The addition causes of several adverse effects
products have been found to contain significant concentrations of of cannabis, as appeared and described in the literature, including
antihistamines, diuretics, weight loss substances and antidepressants, psychosis, intoxication, tachycardia, and changes in blood pressure

© The Author 2016. Published by Oxford University Press. All rights reserved. For Permissions, please email: journals.permissions@oup.com 350
Simultaneous Analysis of Cannabinoids by UPLC–UV and LC–MS-MS 351

(6–12) and atrial fibrillation (13, 14). Some reports declared a 1.7 µm) was utilized, and its temperature was maintained at 30°C
fatality directly related to the toxic effects by synthetic cannabinoid using a column oven. The flow rate was 0.25 mL/min and an injection
use (15–17). volume of 2 µL was utilized. The mobile phase was composed of
Recently, cannabinoids have been described as ‘a new trend’ (18) solvent A (0.1% formic acid in distilled water; D.W) and solvent B
and as ‘new narcotic drugs’ (19). In 2009, JWH-018 and its major (0.1% formic acid in acetonitrile). The gradient elution was as follows:
metabolites in urine were evaluated using liquid chromatography 0 min, 5% B; 1 min, 5% B; 5 min, 80% B; 9 min, 80% B; 9.1 min, 5%
(LC)–MS-MS. Sobolevsky et al. (20) also identified JWH-018 metabo- B and 12 min, 5% B. MS was conducted in electrospray ionization
lites in urine using gas chromatography and LC–MS after synthetic (ESI) mode. The desolvation gas flow and temperature of the
cannabinoids had been smoked. The determination of JWH-018 in positive-ion mode were 600 L/h (N2) and 400°C, respectively. The ca-
serum has also been described, and recently Teske et al. (21) described pillary voltage was 2.7 kV and the collision gas flow was off to achieve
the determination of several compounds in serum, including the optimal analytical conditions.
JWH-015, JWH-018, JWH-019, JWH-020, JWH-073, JWH-081,
JWH-200, JWH-250, and WIN-55212-2 (22). Other approaches Sample preparation
using time-of-flight mass spectrometry, which can be used to detect Forty-five samples were used in this study, including those from tablets
JWH210 and JWH122 in rat adipose tissue (23), have also been (7), capsules (10), powders (4), liquids (1), cookies (22) and candy (1);

Downloaded from https://academic.oup.com/jat/article/40/5/350/1750552 by guest on 01 June 2021

described. However, the aforementioned methods are limited to the all samples were collected from Korean markets. The entire content of
analysis of biological samples such as urine, blood and other bodily the package was crushed into a powder to homogenize. Approximate-
fluids. To date, the simultaneous analysis of such compounds in diet- ly 1 g of the homogenized samples was weighed in a 20-mL volumetric
ary supplements has not been reported. Since the dietary supplement flask. The samples were prepared in 100% methanol solution. The
market continues to grow owing to consumer concern over healthy obtained mixture and methanol were vortexed briefly and sonicated
products and food, it is important to check product quality and com- for 30 min, and then additional methanol was added to a 20-mL
position to evaluate potential health risks. To address this challenge, volumetric flask after cooling. The stock solution was filtered through
we developed and validated a rapid and reliable method for the a 0.22-µm PVDA filter (Millipore, Milford, USA) prior to UPLC
simultaneous detection of nine cannabinoids in food and various analysis.
types of dietary supplements by UPLC–UV and UPLC–MS-MS. The
validation process included the assessment of specificity, sensitivity, Method validation
accuracy, precision, linearity and recovery. The method developed for the determination of cannabinoids was
validated with regard to system suitability along with specificity, limit
Materials and methods of detection (LOD), limit of quantification (LOQ), linearity, accuracy
and precision, recovery and stability according to the requirements pub-
Chemicals and reagents
lished by the AOAC Official Methods of Analysis and International
AM2201, JWH018, JWH019, JWH073, JWH081, JWH122,
Conference on Harmonization of the Harmonized tripartite guideline,
JWH250, THC and XLR11 were purchased from USP (Rockville,
validation of analytical procedures: text and methodology (24).
MD, USA). The purity of the reference compounds was generally
Specific identification of each of the nine compounds was carried
≥98%. To prepare the stock solutions, the standards were dissolved
out by injecting a blank sample fortified with a standard solution of
in methanol at ∼1 mg/mL and stored in a refrigerator (4°C) until
each target compound. The linearity of the method was determined
use. HPLC grade methanol and acetonitrile were purchased from
at five concentrations by plotting the peak areas versus the corre-
Burdick and Jackson (Muskegon, MI, USA), and formic acid was
sponding concentrations of each analyte. The LOD and LOQ were
purchased from Sigma-Aldrich (St Louis, MO, USA). High-purity
determined by spiking the blank sample, and were defined as the low-
deionized water (18.1 MΩ) was obtained from a Milli-Q purification
est analyte concentration that gave signal-to-noise ratios of 3 : 1 and
system (Millipore, Bedford, MA, USA).
10 : 1, respectively. The recovery experiments were performed by add-
Instrumental conditions ing accurate amounts of mixed standards into a blank sample, which
was previously analyzed and found to be not contaminated at three
UPLC
different concentrations. The test was carried out three times, and
Method development and validation were performed on an Acquity
the average percentage recovery and relative standard deviation
UPLC™ system (Waters, Milford, CT, USA). The system consisted
(RSD) were determined for each compound. The intra- and interday
of a binary solvent manager, a sample manager and a photodiode
precision were assessed using the RSD; the accuracy was determined
array detector. The output signal was monitored and processed
as the recovery of the theoretical concentration of the target com-
using the Waters Empower 2 software. The column was a Waters
pounds. Sample stability was measured at several time points for
Acquity UPLC HSS C18 (2.1 mm × 150 mm, 1.8 µm), which was
24 h upon storing the solution at 4°C.
held at 30°C with a flow rate of 0.18 mL/min; the injection volume
was 10 µL. The UV detection was set at 210 nm. The mobile phase
consisted of 25 mM sodium phosphate and 0.01% sodium hexane Results and discussion
sulfonate in deionized water adjusted to pH 3 with phosphoric acid
Optimization of sample preparation
(solvent A) and acetonitrile (solvent B). The gradient elution was as
follows: 0 min, 60% B; 4 min, 80% B; 9 min, 100% B; 11 min, To determine the optimal sample preparation, all samples were pre-
100% B; 11.1 min, 60% B and 15 min, 60% B. pared in 100% methanol because the recovery efficiency following
methanol extraction was greater than that from ethanol. In addition,
UPLC–MS-MS the use of 100% methanol led to a better recovery than either 70 or
Chromatography was carried out using an Acquity UPLC™ system 50% methanol (v/v). The sonication time was varied from 10 to
equipped with Xevo TQ (Waters). For the LC–MS-MS analysis, a 60 min; sonication of most compounds for 30 min led to superior
Waters Acquity UPLC BEH C18 column (2.0 mm × 100 mm, recovery. Supplementary Data 1 summarizes the optimized sample
352 Heo et al.

preparation conditions, including extraction parameters, proportions UPLC–MS-MS

of solvent, sample weight and sonication time. Depending on the structural characteristics of each analyte, UPLC–
MS-MS using ESI was carried out in the positive-ion mode. We
Optimization of instrument conditions evaluated different mobile phases and buffers, including 0.1% formic
UPLC acid/0.1% formic acid in acetonitrile, 10 mM ammonium acetate/
The main objective of this study was to develop a UPLC–UV method acetonitrile and D.W/acetonitrile. We also evaluated several columns,
for the analysis of nine cannabinoids. We selected 210 nm as the including a Waters Acquity UPLC BEH C18 (2.1 mm × 150 mm,
detection wavelength for the analysis, during which we tried to obtain 1.7 µm), Waters Acquity UPLC BEH C8 (2.1 mm × 150 mm,
a good peak shape. Waters Acquity UPLC HSS C18 (2.1 mm × 150 mm, 1.7 µm) and Waters Acquity UPLC HSS T3 (2.1 mm × 100 mm,
1.8 µm), Waters Acquity UPLC BEH C8 (2.1 mm × 150 mm, 1.7 µm) 1.8 µm) to improve the peak shape and MS detection sensitivity
and Waters Acquity UPLC HSS T3 (2.1 mm × 100 mm, 1.8 µm) toward the target compounds. The best results were obtained upon
columns were tested owing to the good shape and symmetry of the ob- using 0.1% formic acid in D.W/0.1% formic acid in acetonitrile
tained peaks and resolution; the best column was the Waters Acquity with the Waters Acquity UPLC BEH C8 column in the positive-ion
UPLC HSS C18 (2.1 mm × 150 mm, 1.8 µm). The evaluated mobile mode. The mass spectrometer was operated in the multiple reaction
phases included the following: 0.05% phosphoric acid ( pH 2.3), monitoring (MRM) mode; the parameters are listed in Table I.

Downloaded from https://academic.oup.com/jat/article/40/5/350/1750552 by guest on 01 June 2021

0.5 mM sodium phosphate ( pH 2.1), 5 mM sodium phosphate ( pH
2.3), 20 mM potassium phosphate, 25 mM sodium phosphate and Validation
sodium hexane sulfonate ( pH 3.0). Notably, 0.01% sodium hexane Specificity
sulfonate led to the best peak resolution. The UPLC chromatogram The selectivity of detection was ensured by determining the retention
and UV spectra of nine cannabinoids are shown in Figures 1 and 2, time of each cannabinoid and by monitoring the UV spectra of the dif-
respectively. ferent components simultaneously. The results are shown in Figures 1

Figure 1. UPLC chromatogram of cannabinoids: (a) matrix blank and (b) matrix sample fortified with cannabinoids; 1. AM2201 (6.622); 2. JWH250 (7.292); 3. JWH073
(7.528); 4. XLR11 (7.647); 5. JWH018 (8.250); 6. JWH081 (8.493); 7. JWH122 (8.968); 8. JWH019 (9.112); 9. THC (9.870).
 Downloaded from https://academic.oup.com/jat/article/40/5/350/1750552 by guest on 01 June 2021
Simultaneous Analysis of Cannabinoids by UPLC–UV and LC–MS-MS 353
Figure 2. UV spectra of nine cannabinoids (AM 2201, XLR11, JWH 250, JWH 073, JWH 018, JWH 081, JWH 122, JWH019 and THC).
354 Heo et al.

Table I. Retention Time and MRM Conditions for Cannabinoids and 2. The obtained MRM transition parameters of UPLC–MS-MS for
the determination of the nine compounds are summarized in Figure 3.
Compound Ion Formula Precursor Product CE (eV) CV
mode ion ion (V)
Linearity
AM2201 + C24H22FNO 359.90 126.85 45 35 The linearity of the method was determined at five concentrations
154.90 25
fortified into the blank solid and liquid matrix. The linearity of the
231.98 25
relationship was evaluated for each matrix-matched calibration
JWH-018 + C24H23NO 341.90 126.87 40 40
154.90 25
compound in a given concentration range, including 2.5–50 µg/mL
JWH-019 + C25H25NO 355.90 126.88 40 25 and 0.01–10 µg/mL for UPLC–UV and UPLC–MS-MS, respectively.
154.90 25 The calibration curves were obtained using least square linear regres-
JWH-073 + C23H21NO 327.90 126.85 40 35 sion and the linearity was confirmed using R 2 values and quality
143.85 35 coefficients (25). Table II summarizes the concentration ranges of
154.85 25 nine components, as well as the calibration curves, R 2 values and
JWH-081 + C25H25NO2 371.94 184.90 25 35 quality coefficients obtained using UPLC–UV and UPLC–MS-MS.
214.00 25 The calibration curve for all components was linear within the chosen

Downloaded from https://academic.oup.com/jat/article/40/5/350/1750552 by guest on 01 June 2021

JWH-122 + C25H25NO 355.92 140.90 40 35
concentration range.
168.90 25
214.00 25
JWH-250 + C22H25NO2 335.90 90.80 20 30 Sensitivity (LOD and LOQ)
120.90 20 The LOD and LOQ were evaluated based on signal-to-noise ratios ob-
143.90 35 tained by the serial dilution of the cannabinoid reference solutions for-
THC + C21H30O2 315.94 192.92 20 30 tified in the blank solid and liquid samples. The LODs of the solid
259.92 20
samples ranged from 0.1 to 0.3 µg/mL and 0.00001 to 0.0003 µg/mL
XLR11 + C21H28FNO 329.93 96.90 25 40
for UPLC–UV and UPLC–MS-MS, respectively, whereas the LOQs
124.90 25
ranged from 0.3 to 0.9 µg/mL and 0.00009 to 0.0015 µg/mL for
143.85 35
231.98 25 UPLC–UV and UPLC–MS-MS, respectively. The LODs of the liquid
297.10 30 samples ranged from 0.1 to 0.4 µg/mL and 0.00001 to 0.0006 µg/mL
for UPLC–UV and UPLC–MS-MS, respectively, whereas the LOQs

Figure 3. LC–MS-MS chromatograms of cannabinoids; (a) AM 2201 (6.51 min), (b) XLR11 (7.10 min), (c) JWH 250 (6.86 min), (d) JWH 073 (7.01 min), (e) JWH 018
(7.56 min), (f ) JWH 081 (7.77 min), (g) JWH 122 (8.17 min), (h) JWH019 (8.32 min), (i) THC (9.00 min).
Simultaneous Analysis of Cannabinoids by UPLC–UV and LC–MS-MS 355

Downloaded from https://academic.oup.com/jat/article/40/5/350/1750552 by guest on 01 June 2021

Figure 3. Continued

Table II. Linearity of Five Concentrations, LOD, and LOQ for Cannabinoids by UPLC and UPLC–MS-MS

Compound Calibration curve Linear range R2 Solid (µg/mL) Liquid (µg/mL)

LOD LOQ LOD LOQ

UPLC–UV (μg/mL)
AM2201 y = 61552.27x + 861.69 2.5–30 0.9999 0.2 0.6 0.3 0.9
JWH018 y = 60640.86x + 6197.61 2.5–30 0.9999 0.2 0.9 0.2 0.6
JWH019 y = 60651.17x + 1913.48 2.5–30 0.9999 0.1 0.3 0.2 0.6
JWH073 y = 66577.82x + 2290.85 2.5–30 0.9999 0.3 0.9 0.3 0.9
JWH081 y = 53221.86x + 2093.03 2.5–30 0.9999 0.1 0.3 0.1 0.3
JWH122 y = 50382.10x + 2612.26 2.5–30 0.9999 0.2 0.6 0.4 1.2
JWH250 y = 32786.14x + 732.98 2.5–30 0.9999 0.3 0.9 0.3 0.9
THC y = 37512.71x + 2915.45 2.5–30 0.9999 0.2 0.6 0.4 1.2
XLR11 y = 28970.29x + 2118.52 2.5–30 0.9999 0.3 0.9 0.3 0.9
UPLC–MS-MS (µg/mL)
AM2201 y = 778.07x + 3370.35 0.01–1 0.998 0.0001 0.0003 0.0001 0.0003
JWH018 y = 3.80x − 26.43 0.1–5 0.998 0.0001 0.0003 0.0001 0.0003
JWH019 y = 0.84x + 4.85 0.25–10 0.998 0.0003 0.0009 0.0005 0.0015
JWH073 y = 6.19x − 12.76 0.1–5 0.999 0.0003 0.0009 0.0004 0.0012
JWH081 y = 10.60x + 335.39 0.05–5 0.999 0.00003 0.00009 0.00005 0.00015
JWH122 y = 7.57x + 300.43 0.25–10 0.999 0.00001 0.00009 0.00002 0.00006
JWH250 y = 857.46x + 11,635.30 0.01–1 0.999 0.0003 0.0009 0.0004 0.0012
THC y = 0.23x − 16.54 0.25–10 0.999 0.00015 0.0004 0.0006 0.0018
XLR11 y = 599.26x + 1336.11 0.01–1 0.998 0.0005 0.0015 0.0005 0.0015
356 Heo et al.

Table III. Intra- and Interday Validation of the Developed Method

Compound UPLC–UV UPLC–MS-MS

Concentration (µg/mL) Concentration (µg/mL)

5 10 50 0.05 0.1 0.5

AM2201
Intraday
Precision (RSD, %)a 0.3 0.7 0.7 1.2 1.3 0.2
Accuracy (%) 102.4 100.8 99.7 98.5 101.7 99.8
Interday
Precision (RSD, %) 3.4 3.5 1.1 1.2 1.3 0.2
Accuracy (%) 100.0 99.8 99.9 98.6 101.6 99.8
JWH018
Intraday
Precision (RSD, %) 0.4 0.3 0.7 1.2 1.3 0.2

Downloaded from https://academic.oup.com/jat/article/40/5/350/1750552 by guest on 01 June 2021

Accuracy (%) 102.6 100.5 99.7 97.8 102.5 99.7
Interday
Precision (RSD, %) 3.3 3.3 1.0 2.0 2.1 0.3
Accuracy (%) 100.3 99.7 100.1 97.1 103.2 99.6
JWH019
Intraday
Precision (RSD, %) 0.4 0.4 0.7 1.3 1.2 1.6
Accuracy (%) 102.8 100.6 99.7 98.5 101.5 98.8
Interday
Precision (RSD, %) 3.4 3.4 1.0 0.5 1.6 2.1
Accuracy (%) 100.4 99.9 100.1 98.8 100.4 97.9
JWH075
Intraday
Precision (RSD, %) 0.4 0.3 0.7 1.5 1.6 0.2
Accuracy (%) 103.1 100.8 99.6 98.0 102.3 99.8
Interday
Precision (RSD, %) 3.3 3.3 1.2 1.7 1.9 0.2
Accuracy (%) 100.8 99.8 99.9 98.5 101.7 99.8
JWH081
Intraday
Precision (RSD, %) 0.5 0.5 0.7 1.1 0.9 2.0
Accuracy (%) 101.4 99.9 99.9 99.8 99.9 97.2
Interday
Precision (RSD, %) 3.0 2.8 0.9 0.6 1.3 0.4
Accuracy (%) 99.3 99.3 100.2 98.5 101.2 98.2
JWH122
Intraday
Precision (RSD, %) 0.4 0.3 0.7 1.7 1.9 0.2
Accuracy (%) 103.0 100.7 99.6 98.1 102.1 99.8
Interday
Precision (RSD, %) 3.9 3.3 1.0 1.1 1.2 0.1
Accuracy (%) 100.3 99.9 100.0 98.6 101.6 99.8
JWH250
Intraday
Precision (RSD, %) 0.3 0.4 0.7 1.8 1.9 0.2
Accuracy (%) 102.4 100.0 99.9 96.9 103.5 99.6
Interday
Precision (RSD, %) 3.2 3.3 1.0 1.5 1.6 0.2
Accuracy (%) 100.1 99.2 100.2 97.6 102.7 99.7
THC
Intraday
Precision (RSD, %) 0.4 0.6 0.8 4.6 5.2 0.6
Accuracy (%) 103.3 100.6 99.6 100.8 99.1 100.1
Interday
Precision (RSD, %) 3.4 3.4 1.0 1.4 1.1 0.2
Accuracy (%) 100.9 99.9 100.0 100.4 101.2 100.1
XLR11
Intraday
Precision (RSD, %) 0.5 0.5 0.7 1.9 2.1 0.2
Continued
Simultaneous Analysis of Cannabinoids by UPLC–UV and LC–MS-MS 357

Table III. Continued

Compound UPLC–UV UPLC–MS-MS

Concentration (µg/mL) Concentration (µg/mL)

5 10 50 0.05 0.1 0.5

Accuracy (%) 101.5 98.8 100.2 98.6 101.6 99.8

Interday
Precision (RSD, %) 2.9 3.3 1.0 1.7 1.9 0.2
Accuracy (%) 99.4 99.7 100.5 98.7 101.5 99.8

a
RSD is defined as the standard deviation of a group of values divided by their mean.

Table IV. Recovery of Nine Cannabinoids in Dietary Supplement Samples (Tablet Type) by UPLC and LC–MS-MS

Downloaded from https://academic.oup.com/jat/article/40/5/350/1750552 by guest on 01 June 2021

Compound UPLC–UV UPLC–MS-MS
2.5 (μg/mL) 25 (μg/mL) 50 (μg/mL) 0.025 (μg/mL) 0.25 (μg/mL) 0.5 (μg/mL)

Recovery RSD Recovery RSD Recovery RSD Recovery RSD Recovery RSD Recovery RSD
(%) (%)a (%) (%) (%) (%) (%) (%) (%) (%) (%) (%)

AM2201 95.9 1.6 98.7 0.6 101.2 0.5 108.1 1.5 106.0 0.6 104.3 0.8
JWH018 95.0 0.7 88.1 0.5 99.5 0.8 99.4 3.5 99.0 3.7 100.4 0.6
JWH019 98.8 0.7 107.9 1.2 101.9 0.4 100.3 2.9 96.4 1.6 96.8 1.1
JWH073 94.6 0.8 91.8 0.5 94.7 0.6 97.7 4.6 101.8 1.3 98.7 1.2
JWH081 103.4 1.8 104.3 0.6 97.5 0.2 96.1 2.6 96.9 0.8 97.8 0.8
JWH122 104.3 1.9 105.4 1.6 97.8 0.6 93.4 2.8 104.6 4.2 105.4 2.1
JWH250 92.9 0.8 96.8 0.5 99.6 0.5 98.6 1.0 101.0 0.5 101.3 1.1
THC 101.0 3.9 95.7 2.6 103.0 0.6 100.3 4.5 98.0 0.4 96.8 2.0
XLR11 98.1 1.9 97.8 0.2 101.3 0.5 100.3 3.0 102.1 1.4 102.1 1.2

a
RSD is defined as the standard deviation of a group of values divided by their mean.

Table V. Stability of Cannabinoids Over 24 h, As Determined by UPLC–UV and UPLC–MS-MS

Compound UPLC–UV UPLC–MS-MS

a
Injection time (h)/compound concentration Mean ± RSD Injection time (h)/compound concentration Mean ± RSD
(%) (%)
0 2 4 6 12 24 0 2 4 6 12 24

AM2201 107.2 106.7 106.5 106.8 108.2 110.6 107.7 ± 1.5 86.8 87.6 85.5 93.0 98.3 81.9 90.2 ± 6.5
JWH018 104.0 103.8 103.5 104.0 105.4 107.7 104.7 ± 1.5 95.8 98.8 99.0 105.0 106.5 93.9 101.0 ± 5.0
JWH019 104.3 104.1 103.9 104.3 105.7 108.0 105.0 ± 1.5 94.0 99.3 97.5 107.1 106.8 94.9 101.0 ± 5.7
JWH073 105.1 104.4 103.4 104.7 106.2 107.6 105.2 ± 1.4 100.1 100.0 98.2 103.9 103.6 94.1 101.1 ± 3.6
JWH081 100.2 100.4 100.4 100.5 101.8 104.2 101.2 ± 1.5 97.4 94.4 96.2 105.6 109.5 93.3 100.6 ± 6.5
JWH122 104.2 102.5 102.9 102.3 105.2 106.0 103.8 ± 1.5 93.9 99.0 97.3 107.3 107.9 94.3 101.1 ± 6.2
JWH250 106.9 106.1 106.5 106.4 108.0 110.1 107.4 ± 1.4 101.6 101.3 97.2 101.7 103.4 96.7 101.0 ± 2.7
THC 104.1 103.8 104.9 104.3 106.8 108.0 105.3 ± 1.6 93.3 102.6 106.3 103.8 104.5 103.2 102.1 ± 4.5
XLR11 107.8 108.4 107.6 108.1 109.1 111.9 108.8 ± 1.5 91.7 104.5 102.6 101.1 102.7 97.0 100.5 ± 4.7

a
RSD is defined as the standard deviation of a group of values divided by their mean.

ranged from 0.3 to 1.2 µg/mL and 0.00006 to 0.0018 µg/mL for precision was determined by repeated analysis of the compounds on
UPLC–UV and UPLC–MS-MS, respectively (Table II). the same day (n = 3) for each concentration, whereas interday preci-
sion was determined by repeated analysis on 3 consecutive days. An
acceptable precision was obtained for all cannabinoids (Table III).
Accuracy and precision The mean RSD was 0.3–3.9% for UPLC–UV and 0.1–5.2% for
The accuracy of the method was evaluated by adding three different UPLC–MS-MS. The maximum RSD was obtained for THC at 5.2%
amounts (corresponding to 50, 100 and 150% of the test preparation con- with UPLC–MS-MS.
centrations) of cannabinoids for 3 days. For each amount, three solutions
were prepared and injected in duplicate, and the recovery percentage was
calculated. The mean accuracies of UPLC–UV and UPLC–MS-MS were Recovery
98.8–103.2 and 97.1–103.5%, respectively (Table III). The recoveries were determined using three known concentrations of
Precision is a measure of the relative error of the method and is each target compound by analyzing fortified blank samples and
expressed as the RSD of the repeatability and precision. Intraday calculating their concentrations using calibration curves, similar to
358 Heo et al.

Table VI. Extraction Recovery (mean ± RSD, %) of Nine Cannabinoids From Various Samples by UPLC–UV

Compound AM2201 JWH018 JWH019 JWH073 JWH081 JWH122 JWH250 THC XLR11

Tablet
Low 87.1 ± 1.4 83.3 ± 2.0 83.4 ± 0.7 88.9 ± 2.1 83.3 ± 1.7 83.2 ± 1.0 92.3 ± 0.7 89.4 ± 0.6 92.3 ± 1.6
Medium 87.2 ± 1.4 80.5 ± 0.5 80.5 ± 2.8 85.4 ± 2.3 85.6 ± 0.5 97.4 ± 2.2 109.8 ± 9.2 85.0 ± 0.4 109.3 ± 3.9
High 98.3 ± 0.9 82.4 ± 0.3 95.6 ± 0.6 102.1 ± 2.9 97.1 ± 0.3 90.6 ± 0.5 85.1 ± 2.0 92.7 ± 0.6 103.0 ± 1.6
Capsule
Low 86.2 ± 0.9 94.6 ± 2.7 96.7 ± 0.7 107.3 ± 0.9 99.7 ± 0.5 97.4 ± 0.4 99.1 ± 1.8 98.1 ± 0.5 94.6 ± 1.7
Medium 83.2 ± 1.5 85.9 ± 0.3 92.5 ± 4.5 96.7 ± 1.1 92.9 ± 4.2 101.5 ± 9.2 106.7 ± 3.3 92.0 ± 1.1 93.2 ± 1.9
High 96.3 ± 0.4 99.0 ± 8.4 81.0 ± 0.5 80.0 ± 0.2 92.2 ± 0.6 100.4 ± 5.6 95.4 ± 2.6 97.0 ± 0.6 87.1 ± 4.2
Powder
Low 81.9 ± 0.5 84.1 ± 1.9 84.0 ± 0.2 81.6 ± 0.7 82.7 ± 0.5 90.4 ± 0.2 80.9 ± 0.4 92.2 ± 0.1 106.2 ± 1.4
Medium 82.1 ± 0.1 87.4 ± 0.4 81.5 ± 1.4 84.3 ± 0.7 81.5 ± 1.4 84.0 ± 0.7 83.1 ± 0.2 85.4 ± 0.1 90.7 ± 5.4
High 96.1 ± 0.1 91.3 ± 0.1 96.1 ± 0.3 81.9 ± 0.6 90.4 ± 0.1 93.5 ± 0.0 86.5 ± 0.1 100.6 ± 0.1 97.1 ± 0.1
Liquid

Downloaded from https://academic.oup.com/jat/article/40/5/350/1750552 by guest on 01 June 2021

Low 96.0 ± 3.8 115.6 ± 4.5 89.3 ± 1.5 82.4 ± 4.3 101.8 ± 1.7 104.8 ± 1.1 100.2 ± 1.7 110.5 ± 0.4 101.5 ± 6.1
Medium 84.7 ± 0.1 84.4 ± 0.1 80.2 ± 0.4 105.9 ± 1.2 97.5 ± 5.0 92.9 ± 0.2 93.4 ± 1.0 97.4 ± 0.1 90.5 ± 0.7
High 98.2 ± 0.2 89.8 ± 3.8 92.8 ± 0.2 85.4 ± 1.1 107.4 ± 1.2 94.3 ± 0.0 88.1 ± 0.2 101.7 ± 0.1 86.1 ± 0.0
Cookie
Low 103.8 ± 5.1 82.6 ± 3.5 97.3 ± 3.4 89.3 ± 2.5 93.9 ± 4.2 96.6 ± 1.8 98.0 ± 2.8 118.0 ± 0.4 83.4 ± 0.8
Medium 95.2 ± 1.1 91.6 ± 1.8 104.3 ± 0.4 93.7 ± 1.5 108.7 ± 1.7 84.1 ± 0.9 95.7 ± 1.0 93.3 ± 0.5 86.7 ± 0.8
High 81.2 ± 0.3 89.6 ± 0.8 95.7 ± 3.7 90.6 ± 1.7 93.9 ± 1.9 91.7 ± 2.7 101.3 ± 9.2 86.4 ± 0.1 99.1 ± 5.6
Candy
Low 81.2 ± 0.6 102.7 ± 2.9 93.8 ± 2.6 95.7 ± 3.0 90.6 ± 3.7 100.5 ± 0.7 96.7 ± 1.2 108.1 ± 5.6 94.2 ± 1.6
Medium 85.1 ± 3.1 95.6 ± 1.0 108.1 ± 7.1 84.7 ± 0.5 104.2 ± 5.1 100.7 ± 0.2 111.9 ± 1.1 96.2 ± 0.2 91.8 ± 0.1
High 82.9 ± 0.7 88.5 ± 4.1 102.0 ± 1.8 87.3 ± 7.7 101.1 ± 2.1 100.1 ± 2.6 102.5 ± 4.6 85.3 ± 0.1 99.4 ± 2.3

Table VII. Extraction Recovery of Nine Cannabinoids From Various Samples by UPLC–MS-MS

Compound AM2201 JWH018 JWH019 JWH073 JWH081 JWH122 JWH250 THC XLR11

Tablet
Low 82.3 ± 3.6a 92.4 ± 4.4 93.1 ± 2.7 90.9 ± 1.8 102.6 ± 6.7 100.0 ± 6.5 90.6 ± 8.1 96.6 ± 4.6 92.5 ± 6.7
Medium 106.7 ± 1.4 97.8 ± 0.2 101.3 ± 1.6 102.0 ± 0.1 99.8 ± 2.7 101.5 ± 5.4 92.1 ± 2.3 99.4 ± 1.9 106.1 ± 1.6
High 91.4 ± 5.0 98.0 ± 0.8 97.4 ± 1.1 100.0 ± 1.9 96.6 ± 3.6 103.0 ± 2.4 97.8 ± 0.7 96.8 ± 1.9 95.5 ± 1.5
Capsule
Low 100.9 ± 1.5 104.2 ± 0.7 101.3 ± 2.9 99.3 ± 4.5 107.3 ± 3.3 112.4 ± 4.8 100.9 ± 5.0 93.7 ± 9.8 107.8 ± 4.5
Medium 95.8 ± 1.9 90.9 ± 4.8 93.6 ± 5.2 95.5 ± 2.5 102.5 ± 1.3 101.9 ± 1.8 93.0 ± 0.7 98.3 ± 2.4 92.2 ± 1.3
High 85.1 ± 2.2 96.6 ± 1.1 94.2 ± 0.8 93.8 ± 0.4 104.2 ± 1.0 105.5 ± 8.8 99.0 ± 1.9 96.6 ± 2.8 97.2 ± 0.5
Powder
Low 99.1 ± 71 99.0 ± 3.4 94.6 ± 0.5 102.1 ± 6.7 92.6 ± 7.3 106.2 ± 7.5 88.0 ± 3.4 88.5 ± 61 114.5 ± 5.4
Medium 94.1 ± 6.4 91.7 ± 7.6 95.9 ± 5.5 92.6 ± 4.5 96.8 ± 3.1 90.1 ± 4.0 91.2 ± 2.9 97.1 ± 0.8 94.5 ± 1.7
High 102.4 ± 1.3 99.1 ± 0.3 99.9 ± 1.6 99.1 ± 0.6 88.3 ± 5.1 98.6 ± 2.0 90.1 ± 1.4 97.3 ± 4.3 89.0 ± 0.9
Liquid
Low 94.3 ± 4.0 97.4 ± 1.0 95.0 ± 4.6 99.2 ± 1.3 84.2 ± 5.7 95.7 ± 8.0 99.0 ± 1.4 96.0 ± 6.1 100.4 ± 4.0
Medium 98.4 ± 1.5 98.1 ± 0.7 98.3 ± 0.3 99.0 ± 0.4 83.9 ± 7.4 100.1 ± 1.2 95.3 ± 2.8 99.3 ± 1.4 101.7 ± 1.9
High 91.6 ± 2.2 100.2 ± 0.1 86.9 ± 0.9 98.8 ± 0.4 91.3 ± 6.5 101.3 ± 3.4 101.9 ± 0.4 100.4 ± 0.9 98.8 ± 1.5
Cookie
Low 85.1 ± 6.9 83.1 ± 4.4 83.0 ± 1.0 83.6 ± 0.7 102.6 ± 3.6 104.8 ± 5.0 82.4 ± 3.8 101.2 ± 5.5 94.8 ± 2.1
Medium 85.1 ± 1.6 86.7 ± 0.6 91.1 ± 2.1 89.3 ± 3.4 101.2 ± 4.7 99.4 ± 4.0 84.7 ± 3.2 91.0 ± 2.2 93.3 ± 0.8
High 99.0 ± 0.9 94.6 ± 0.6 100.8 ± 1.2 95.7 ± 1.5 90.7 ± 2.2 105.9 ± 2.8 93.8 ± 2.0 99.2 ± 3.3 95.4 ± 4.4
Candy
Low 82.3 ± 6.0 85.8 ± 3.8 89.1 ± 1.9 85.3 ± 3.1 100.5 ± 7.1 106.4 ± 5.0 86.0 ± 8.8 80.5 ± 3.6 87.5 ± 7.7
Medium 87.1 ± 1.9 90.3 ± 0.4 90.9 ± 0.7 88.4 ± 0.4 96.2 ± 6.9 108.4 ± 2.2 93.4 ± 5.0 96.2 ± 1.9 87.1 ± 2.8
High 103.5 ± 3.1 101.3 ± 2.2 100.4 ± 1.5 98.5 ± 1.3 95.9 ± 2.2 104.7 ± 2.6 104.1 ± 4.3 97.6 ± 0.9 98.5 ± 2.5

a
Mean ± RSD (%), RSD is defined as standard deviation of a group of values divided by their mean.

the procedure that was used to obtain the accuracy profile. Table IV acceptable limits, indicating that the method was suitable for the ana-
summarizes the mean recoveries obtained for the tablet types of lysis of active substances in pharmaceutical preparations.
samples at each concentration. The mean recoveries ranged from
88.1 to 107.9% and from 93.4 to 108.1% in UPLC–UV and Stability
UPLC–MS-MS, respectively. The results are shown together with Table V shows the stability of the test preparation. The sample
RSD% for UPLC–UV and UPLC–MS-MS. All recoveries were within solution was stable for up to 24 h. The recoveries of the cannabinoids
Simultaneous Analysis of Cannabinoids by UPLC–UV and LC–MS-MS 359

ranged from 100.2 to 111.9% in UPLC–UV and from 81.9 to 107.9% 5. Ashton, A.C. (1999) Adverse effects of cannabis and cannabinoids. British
in UPLC–MS-MS. All RSD values were far below 6.5%, thus confirm- Journal of Anesthesia, 83, 637–649.
ing that the equipment worked well for the developed analytical 6. Grotenhermen, F. (2007) The toxicology of cannabis and cannabis prohib-
ition. Chemistry and Biodiversity, 4, 1744–1769.
method, and was highly reproducible.
7. Khiabani, H.Z., Mørland, J., Bramness, J.G. (2008) Frequency and irregu-
larity of heart rate in drivers suspected of driving under the influence of can-
Application with authentic samples nabis. European Journal of Internal Medicine, 19, 608–612.
Forty-five samples, including those from tablets (7), capsules (10), 8. Zimmermann, U.S., Winkelmann, P.R., Pilhatsch, M., Nees, J.A.,
powders (4), liquids (1), cookies (22) and candies (1), were analyzed Spanagel, R., Schulz, K. (2009) Withdrawal phenomena and dependence
with the validated UPLC–UV and UPLC–MS-MS method. To validate syndrome after the consumption of ‘spice gold’. Deutsches Arzteblatt
the method, extraction recovery was also investigated by analyzing 45 International, 106, 464–467.
commercial products that were fortified with the standard analytes. 9. Every-Palmer, S. (2010) Warning: legal synthetic cannabinoid- receptor
agonists such as JWH-018 may precipitate psychosis in vulnerable indivi-
Each sample with three different concentrations of analytes was ana-
duals. Addiction, 105, 1859–1860.
lyzed. The lowest concentration tested was near the LOQ, the medium
10. Muller, H., Sperling, W., Kohrmann, M., Huttner, H.B., Kornhuber, J.,
concentration was ∼10-fold above the LOQ and the highest concen-
Maler, J.M. (2010) The synthetic cannabinoid spice as a trigger for an
tration was ∼20-fold above the LOQ. Increased chromatographic acute exacerbation of cannabis induced recurrent psychotic episodes.

Downloaded from https://academic.oup.com/jat/article/40/5/350/1750552 by guest on 01 June 2021

areas of the analytes in fortified dietary supplements were compared Schizophrenia Research, 118, 309–310.
with the areas of the pure standards. The mean recoveries of nine 11. Simmons, J., Cookman, L., Kang, C., Skinner, C. (2011) Three cases of
targeted chemicals in different formulations are illustrated in Table VI ‘spice’ exposure. Clinical Toxicology, 49, 431–433.
(UPLC–UV) and Table VII (UPLC–MS-MS). 12. Benford, D.M., Caplan, J.P. (2011) Psychiatric sequelae of spice, K2 and
synthetic cannabinoid receptor agonists. Psychosomatics, 52, 295.
13. Kosior, D.A., Filipiak, K.J., Stolarz, P., Opolski, G. (2001) Paroxysmal at-
Conclusion rial fibrillation following marijuana intoxication: a two-case report of pos-
sible association. International Journal of Cardiology, 78, 183–184.
A sensitive and reproducible method based on UPLC–UV and UPLC–
14. Korantzopoulos, P., Liu, T., Papaioannides, D., Li, G., Gou-de venos, J.A.
MS-MS was developed for the detection and quantification of cannabi-
(2008) Atrial fibrillation and marijuana smoking. International Journal of
noids in various samples (tablets, capsules, powders, liquids, cookies Clinical Practice, 62, 308–313.
and candies). The method was fully validated according to the inter- 15. Bachs, L., Mørland, H. (2001) Acute cardiovascular fatalities following
national guidelines and found to specific, precise, accurate, linear and cannabis use. Forensic Science International, 124, 200–203.
stable. The developed UPLC–UV and UPLC–MS-MS method permits 16. Stoppacher, R. (2012) Synthetic cannabinoid use and sudden death. AFP
the simultaneous analysis of a large number of cannabinoids in dietary Journal, 2, 202–206.
supplement matrices with minimal sample preparation. Therefore, the 17. Shanks, K.G., Winston, D., Heidingsfelder, J., Behonick, G. (2015) Case re-
validated method should be useful, accurate and reliable for the contin- ports of synthetic cannabinoid XLR-11 associated fatalities. Forensic
Science International, 252, e6–e9.
ued monitoring of cannabinoids. The present method allows for the
18. Vardakou, I., Pistos, C., Spiliopoulou, C.H. (2010) Spice drugs as a new
most comprehensive targeted screening of cannabinoids in dietary sup-
trend: mood of action, identification and legislation. Toxicology Letters,
plement and foods available currently, and is expected to be useful for
197, 157–162.
adulterant inspection and sample analysis. 19. Mustata, C., Torrens, M., Pardo, R., Perez, C., Farre, M. (2009)
Spice drugs: cannabinoids as new designer drugs. Addicciones, 21,
181–186.
Supplementary data
20. Sobolevsky, T., Prasolov, I., Rodchenkov, G. (2010) Detection of JWH-018
Supplementary data are available at Journal of Analytical Toxicology metabolites in smoking mixture post-administration urine. Forensic Science
online. International, 200, 141–147.
21. Teske, J., Weller, J.P., Feiguth, A., Rothamel, T., Schulz, Y., Troger, H.D.
(2010) Sensitive and rapid quantification of the cannabinoid receptor agon-
Funding ist naphthalene-1-yl-(1-pentylindol-3-yl) methanone (JWH-018) in human
serum by liquid chromatography-tandem mass spectrometry. Journal of
This research was supported by a grant [Number 12181MFDS705] Chromatography B, 878, 2659–2663.
from the Ministry of Food and Drug Safety in 2014. 22. Dresen, S., Kneisel, S., Weinmann, W., Zimmermann, R., Auwarter, V.
(2011) Development and validation of a liquid chromatography-tandem
mass spectrometry method for the quantitation of synthetic cannabinoids
References of the aminoalkylindole type and methanandamide in serum and its
1. Strano-Rossi, S., Anzillotti, L., Castrignano, E., Romolo, F.S., Chiarotti, M. application to forensic samples. Journal of Mass Spectrometry, 46,
(2012) Ultrahigh performance liquid chromatography-electrospray 163–171.
ionization-tandem mass spectrometry screening method for direct analysis 23. Schaefer, N., Peters, B., Bregel, D., Maurer, H.H., Schmidt, P.T., Ewald, A.
of designer drugs, ‘spice’ and stimulants in oral fluid. Journal of H. (2014) Can JWH-210 and JWH-122 be detected in adipose tissue four
Chromatography, 1258, 37–42. weeks after single oral drug administration to rats? Biomedical
2. Merola, G., Aturki, Z., D’Orazid, G., Gottaedo, R., Macchia, T., Chromatography, 28, 1043–1047.
Tagliaro, F. et al. (2012) Analysis of synthetic cannabinoids in herbal blends 24. International Conference on Harmonisation (ICH). (2005) Harmonized
by means of nano-liquid chromatography. Journal of Pharmaceutical and Tripartite Guideline, Validation of Analytical Procedures: Text and Meth-
Biomedical Analysis, 71, 45–53. odology Q2 (R1).
3. Lemos, N.P., Ingle, E.A. (2011) Cannabinoids in postmortem toxicology. 25. de Beer, J.O., de Beer, T.R., Goeyens, L. (2007) Assessment of quality per-
Journal of Analytical Toxicology, 35, 394–401. formance parameters for straight line calibration curves related to the
4. Hall, W., Solowij, N. (1998) Adverse effects of cannabis. The Lancet, 352, spread of the abscissa values around their mean. Analytica Chimica Acta,
1611–1616. 584, 57–65.