Journal Pre-proofs

Stability characteristics of Cannabidiol for the design of pharmacological, bi-

ochemical and pharmaceutical studies

A.I. Fraguas-Sánchez, A. Fernández-Carballido, C. Martin-Sabroso Sofware,

A.I. Torres-Suárez

PII: S1570-0232(20)30123-9
DOI: https://doi.org/10.1016/j.jchromb.2020.122188
Reference: CHROMB 122188

To appear in: Journal of Chromatography B

Received Date: 23 January 2020

Revised Date: 9 May 2020
Accepted Date: 20 May 2020

Please cite this article as: A.I. Fraguas-Sánchez, A. Fernández-Carballido, C. Martin-Sabroso Sofware, A.I.
Torres-Suárez, Stability characteristics of Cannabidiol for the design of pharmacological, biochemical and
pharmaceutical studies, Journal of Chromatography B (2020), doi: https://doi.org/10.1016/j.jchromb.
2020.122188

This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover
page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version
will undergo additional copyediting, typesetting and review before it is published in its final form, but we are
providing this version to give early visibility of the article. Please note that, during the production process, errors
may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

© 2020 Published by Elsevier B.V.

 Stability characteristics of Cannabidiol for the design of pharmacological,
biochemical and pharmaceutical studies

A.I. Fraguas-Sánchez1; A.Fernández-Carballido1, 2; C Martin-Sabroso1, A.I. Torres-

Suárez1, 2

1 Department of Pharmacy and Pharmaceutical Technology, School of Pharmacy,

Complutense University of Madrid, Pl Ramón y Cajal s/n., 28040 Madrid, Spain.

2 Institute of Industrial Pharmacy, School of Pharmacy, Complutense University of

Madrid, Pl Ramón y Cajal s/n. Universidad Complutense de Madrid, 28040 Madrid,
Spain.

* Corresponding author:
Ana Isabel Torres Suárez
e-mail: galaaaa@ucm.es
Tel: +34 913941735

1
 ABSTRACT

Cannabidiol (CBD) is one of the most promising cannabinoids in therapeutics.

Nevertheless, the reported stability testing has been carried out with plant extracts and
not with CBD as a drug substance. The aim of this work was to evaluate the stability of
CBD in solution. A High-Performance Liquid Chromatography (HPLC) analytical
method, with CBD in ethanol, was previously validated for these stability studies. The
resulting method was linear and proportional in a range of concentrations from 1-150 µg
CBD/mL, as well as precise. It was also considered suitable to quantify CBD in aqueous
medium as reported in accuracy studies. The stability of CBD was influenced by multiple
factors. Temperature was one of the most critical parameters, with an activation energy
of 92.19KJ/mol. At room temperature, CBD was highly unstable (t95=117.13 days).
However, at 5ºC it was stable for at least 12 months. CBD was also sensitive to oxidation,
with a short t95 of 1.77 days in oxidizing environments, as well as to light. The photolytic
reaction seems to be oxidative. The solvent influences CBD stability, and the latter is
more stable in ethanol than in aqueous medium. In fact, in simulated physiological
conditions (pH 7.4 and 37ºC) 10% of CBD was degraded within 24 hours. These studies
indicate that CBD is highly unstable, and this should be taken into account in the
development of in vitro and in vivo studies of CBD activity and in the pharmaceutical
development of dosage forms.

Keywords: Cannabidiol, Degradation, HPLC, Shelf-life, Stability, Oxidation.

2
 1. Introduction

Stability studies provide information about the environmental conditions that

could trigger drug degradation and the mechanisms responsible for these processes. These
studies are conducted in accordance with the shelf life of a formulation in the best storage
conditions. However, stability studies are also essential during the pharmaceutical
development of a new drug, to guarantee its proper handling and formulation, and to avoid
mistakes in the evaluation, both in vitro and in vivo, of the activity or bioavailability due
to dosing problems or drug instability in physiological conditions [1].

In the past decades, cannabinoids, the active compounds of the “Cannabis plant”,
have attracted a great deal of interest as potential new therapeutic agents for the treatment
of different pathological conditions, including neurodegenerative disorders (e.g. multiple
sclerosis and epilepsy), pain, nausea, vomiting, anxiety or cancer [2-6]. Currently, some
formulations (tablets, capsules, oromucosal spray and creams) are available for the
management of several conditions including loss of appetite, nausea and vomiting related
to chemotherapy, convulsions, and pain [7]. Among all cannabinoids, cannabidiol (CBD)
is one of the most promising candidates.

Regarding cannabinoid stability, these compounds have generally shown high

instability. Most of the studies have been focused on Δ9-tetrahydrocannabinol (Δ9-THC),
the main cannabis constituent and the primary psychoactive compound of cannabis plant
preparations and have been performed with cannabis extracts obtained from cannabis
flowers and not with isolated cannabinoids. Δ9-THC has been shown to be thermolabile
[8], sensitive to oxidation [9], unstable in acid solutions [10] and photolabile [11]. In
respect of CBD, it has been reported that, while methanolic extracts stored at 5ºC showed
a high stability [12], aqueous preparations were highly unstable after 7 days of storage
even at 5ºC, with a CBD loss of around 60%. Oil preparations stored in the same
conditions were also unstable, although a lower CBD degradation was detected [13].
Nevertheless, a more recent study demonstrated that cannabis oil preparations showed a
high thermal stability with a slight (although significant) loss of CBD content of around
12% after one year of storage in darkness at 4ºC [14].

3
 Therefore, the aim of this work was to evaluate the stability of CBD in solution.
The effect of temperature, oxygen, radiations and physiological conditions on CBD
solutions were evaluated. Previously, a HPLC analytical method was validated for these
stability studies.

2. Materials and methods

2.1. Reagents and chemicals

Cannabidiol (CBD) and Δ9-THC were purchased from THC-Pharma (Frankfort,
Germany). The analytical reference standard solution of CBD in methanol (1mg/mL) was
supplied by Sigma-Aldrich (Supelco, Missouri, USA). Methanol and acetonitrile HPLC
grade were obtained from Fisher Scientific (Madrid, Spain); hydrochloric acid, sodium
hydroxide, ethanol absolute and Tween®-80 were provided by Panreac (Barcelona Spain).
Hydrogen peroxide (H2O2) was supplied by Acros Organics (Fisher Scientific, Spain).
Laboratory generated demineralized Milli-Q® water (Millipore, Spain) was used.

In order to avoid cannabidiol glass and plastic binding, all material was pre-treated
with Sigmacote® (Sigma, Missouri, USA).

2.2. Instruments and equipment

An Agilent 1200 series Liquid Chromatography system (Agilent technologies,

California, USA) equipped with a quaternary pump (G1311 Agilent 1200 series), a 1200
series Autosampler (G1329, Agilent 1200 series), a vacuum degasser, a thermostatted
Column Compartment (G1316 Agilent 1200 series) and a Diode Array Detector (G1315
Agilent 1200 series) was used. The data were processed using the Agilent Chemstation
software (Agilent technologies, California, USA).

Temperature and oxygen-dependent stability studies were performed in

thermostatic stability chambers (Memmert UN30, Germany) and photo-stability studies
in a photo-stability chamber (CCI: Xenoterm 1500-RF, CCI, Barcelona, Spain) equipped
with a D65/ID65 xenon lamp. Stability studies in simulated physiological conditions were
conducted using a thermostatic shaking bath (Nuve ST30, Herderson Biomedical,
London, UK).

4
2.3. Analytical method

The chromatographic conditions used were based on the analytical method

described by Munjal et. al [8]. The analytical column was a Mediterranea®‐Sea C18 (150
x 4.6mm i.d., 5μm) (Teknokroma, Barcelona, Spain). The mobile phase consisted of a
mixture of methanol, acetonitrile and water (52:30:18 v/v) adjusted to a pH of 4.5 with
acetic acid, and filtered through a Millipore® membrane filter (47mm membrane, 0.45μm
pore size, nylon) and degassed in an ultrasonic bath before use. An isocratic analysis at a
flow rate of 1.8mL/min was performed at room temperature. The injection volume was
20μL and the detection wavelength was set at 228nm. The quantification of CBD was
performed by measuring the peak area count compared to the analytical reference
standard solution of CBD in methanol at a concentration of 100μg/mL.

CBD stock solutions in ethanol (1mg/mL) were prepared and stored at -20ºC.
Validation characteristics were determined according to the ICH Q2 (R1) guidelines [15].

Linearity was evaluated using three freshly prepared calibration curves (1-
150μg/mL) prepared from three CBD stock solutions (1mg/mL).

The limit of detection (LOD) and limit of quantification (LOQ) were determined
using the following equations [16]:

LOD = 3.3SD/b and LOQ = 10 SD/b

where SD and b are, the standard deviation of the response (γ-intercept standard deviation
was used) and the slope of a calibration curve performed with 5 concentrations close to
LOQ (1-10μg/mL) respectively.

The precision of the method was evaluated as intraday variability (repeatability)

and interday variability (intermediate precision). Repeatability was determined by
measuring CBD concentrations of 5, 50 and 100μg/mL (standard solution), 3 replicates
of each on the same day. Interday precision was determined by measuring three replicates
of the CBD concentrations (5, 50 and 100μg/mL) on three consecutive days. The mean
values and variation coefficient (CV) were calculated in both cases.
Accuracy was determined from the analysis of three concentrations of CBD (25,
50 and 100μg/mL) in demineralized water with Tween 80 at 0.5% (v/v) (as a CBD
solubilizing agent), 3 replicates of each, prepared from three stock solutions (1mg/mL) in
the same solvent.

5
 Specificity was determined from the analysis of the chromatograms of the samples
of the different stability studies covering degradation products from light, heat and
oxidation. Purity of the peak corresponding to CBD was determined with a diode array.
Moreover, the ability of this analytical method to quantify CBD in the presence of Δ9-
THC was also evaluated. Briefly, the stock solutions of CBD and Δ9-THC in methanol
(1mg/mL) were prepared, mixed and diluted to a solution of CBD and Δ9-THC
100μg/mL.

System suitability was evaluated to confirm that the equipment was adequate for
the analysis to be performed. This test was carried out by replicate injections of the
standard solution (100μg/mL of CBD) and determining retention time, column efficiency
as a number of theoretical plates (N) and symmetry factor.

2.4. Temperature-dependent stability

Briefly, a standard solution of CBD in ethanol (100µg/mL) was prepared and stored
in glass ampoules at 5±3ºC and at 25, 40, 50, 60 and 70±2ºC. All samples were maintained
in darkness. At predetermined time intervals, ampoules of each condition (n=4) were
analyzed by HPLC. The degradation rate constant (K) was calculated for each storage
temperature and the relationship between K and the temperature (T) was established by
means of the Arrhenius equation:

ln K= ln A –Ea/RT

where Ea is the activation energy, A and R are the frequency factor and gas constant,
respectively, and T the absolute temperature (ºK).

The time interval required for the degradation of 5% of the drug in the solution (t95)
was extrapolated at 5 and 25ºC from the Arrhenius equation obtained with data at 25, 40,
50 and 60ºC. These values were compared with those calculated based on data from the
stability study at both temperatures.

2.5. Oxygen-dependent stability

Some authors have reported that by oxidation reactions, monomeric and dimeric
hydroxyquinones or cannabelsoic acid compounds can be formed from CBD [17, 18]. For
this reason, the stability of CBD against oxidation was also evaluated. Briefly, three CBD
solutions (100µg/mL) were prepared and the following was used as vehicle: i)
6
demineralized water with Tween 80 at 0.5% (v/v) (samples O±), ii) demineralized water
with Tween 80 at 0.5% (v/v) and H2O2 at 3% (v/v) (samples O+), and iii) demineralized
water with Tween 80 at 0.5% (v/v) and flushing with nitrogen to displace the dissolved
oxygen (samples O-).

CBD solutions were deposited into glass ampoules and placed in a stability
chamber at 25ºC±2ºC. At predetermined time intervals, four samples (n=4) of each
solution were analyzed by HPLC.

2.6. Photo-stability studies

Some authors have reported that cannabis oil preparations are unstable when
stored at light [19]. For this reason, photo-stability studies were performed according to
ICHQ1B guidelines. A standard solution of CBD in ethanol (100µg/mL) was prepared in
glass ampoules that were closed and placed in the photo-stability chamber for 72 hours
in order to obtain an overall illumination greater than 1.2 million of lux/h and an
integrated energy near ultraviolet (above 320nm) greater than 200Wh/m2 [20]. Ampoules
covered with aluminum foil and stored in the same chamber were used as controls. At
each time point (12, 24, 36, 48 and 72 hours), 4 ampoules were analyzed by HPLC. These
studies were undertaken at 5 and 25ºC.

Due to the influence of oxygen on CBD stability, nitrogen flushed solutions were
also analyzed in order to determine the participation of oxygen in light instability of CBD.
In photo-oxidative reactions, singlet oxygen can react with the unsaturated bonds present
in polynuclear aromatic rings of the drug structure, as is the case of CBD [21]. For this
study, ampoules were flushed with nitrogen before closing and stored in the photo-
stability chamber at 5ºC±3ºC. Ampoules stored at 25ºC were not analyzed due to the high
degradation of CBD detected in the thermo-stability study at 25ºC.

2.7. Stability in simulated physiological conditions

To simulate physiological conditions, a CBD solution (100µg/mL) was prepared
in a phosphate buffer solution (pH 7.4). Due the high lipophilicity of this compound (log
P= 6.33) Tween 80 at 0.5% (w/v) was added as a surfactant. This solution was degasified
in an ultrasonic bath for 40 minutes to eliminate the oxygen, placed in glass ampoules
and incubated at 37ºC±0.5ºC in a thermostatic sacking bath (agitation of 100rpm) for 10
days. At predetermined time points (1, 2, 3, 4, 5, 6, 7, 8, 9 and 10 days), samples were

7
extracted, filtered and analyzed by HPLC. Four samples (n=4) were analyzed at each time
point.

2.8. Data analysis

All the data are presented as the mean ± standard deviation. Stability profile fitting
was carried out based on the least squares method so that the slope of the curves directly
corresponds to the degradation rate constant. A statistical analysis was performed using
Statgraphics Centurion XVIII (Statgraphics Technologies, Virginia, USA). All graphs
were drawn using Origin 2017 software (Origin lab, Massachusetts, USA).

3. Results

3.1. Analytical method

As illustrated in Figure 1A, the chromatogram of the standard CBD solution

showed a large sharp peak separated from the solvent front with a retention time of 4.72
minutes.

The assay was linear from 1-150µg/mL with a determination coefficient of 0.999
(Table 1, Figure 1B). The F-lack of fit ANOVA (p > 0.05) showed that the two variables
(peak area and concentration) fitted a linear model. Moreover, the analytical method was
proportional, with a p value >0.05 in the t-test for the intercept. The LOD and LOQ were
estimated to be 0.2µg/mL and 0.7µg/mL, respectively.

Regarding precision, the CV values obtained for intraday and interday variability
were <2% (Table 2).

Table 3 shows the results obtained in the accuracy study carried out on samples
of CBD dissolved in water with 0.5% Tween 80, using the calibration curve in ethanol.
Recovery percentages were in the interval of 99.2-101.6%, with a mean recovery of 100.4
± 1.20%. Plotting of calculated concentration versus actual value was linear with a slope
of 0.997 ± 0.028 and an intercept of -0.003± 1.898 (Figure 2).

This analytical method was also suitable for quantifying CBD in the presence of
Δ9-THC, the main constituent of cannabis sativa, showing separate peaks that appeared
at 4.7 and 11.34 minutes, respectively.

8
 The system suitability parameters of the HPLC system are shown in Figure 1C.
The number of theoretical plates was around 16000, and the symmetry factor was 0.892,
in accordance with FDA limits (N > 2000 and T < 2) [22].

3.2. Temperature-dependent stability

Temperature is one of the critical factors that compromise cannabinoid stability.

In this work, the effect of temperature was evaluated by storing CBD solutions in ethanol
at different temperatures in the range of 5-70ºC. As shown in Figures 3A and 3B, the
higher the temperature the higher the degradation rate of CBD, except at 70ºC where no
differences were observed in comparison with 60ºC. The values of the remaining % of
CBD vs. time for each storage temperature were fitted to first-order kinetics, obtaining
the parameters shown in Table 4. A statistically significant correlation between both those
variables was detected for all temperatures except 5ºC. The t95 at 25ºC was 117±18 days.

Statistically significant differences were observed among degradation rate

constants (K) obtained at 25, 40, 50 and 60ºC. Therefore, all the K values obtained at
these temperatures were fitted to the Arrhenius equation obtaining the parameters shown
in Table 5. A high correlation between the lnK and each corresponding absolute
temperature (ºK) (r=-0.998) and a good fit to the proposed model (Flack of fit=1.34;

p=0.318) were observed. From the slope of the curve, a value of Ea=92.19KJ/mol was
calculated.

When the value of t95 at 25ºC was extrapolated from the parameters of the
Arrhenius equation, 119 days were obtained. Finally, the value of t95 at 5ºC was also
extrapolated from the Arrhenius equation, which was 4.8 years. Due to this high stability
at 5 ºC, CBD solutions should be stored in a refrigerator.

As shown in Figure 3C, numerous peaks corresponding to CBD degradation

products can be observed in the chromatograms of degraded samples as compared to
Figure 1A. As detailed in section 3. 6, the peaks that appeared around 2.8 and 5.6 minutes
are the most representative.

3.3. Oxygen-dependent stability

The results of the stability study of CBD against oxygen are shown in Figure 4A.
In all cases, CBD degradation was more intensive when H2O2 was incorporated in the
9
CBD solution (samples O+), especially in the first days of the assay (20% of drug
degradation was detected after 10 days of storage), and less intensive in samples flushed
with nitrogen (O-). These data were fitted to first order kinetics, obtaining the parameter
shown in Table 6. For the three CBD solutions, statistically significant correlation
between drug concentration and storage time was demonstrated. The values of the
degradation rate constant (K) allowed us to calculate the values of t95 included in the
Table 5. While t95 of the drug in an oxidizing environment was less than 2 days (t95=1.8
days), in an oxygen-free solution it was much higher (t 95=19.7 days). The t95 of
demineralized water with Tween at 0.5% (samples O+/-) showed an intermediate value
of 7.6 days.

As shown in the chromatogram in Figure 4C, the main degradation products of

CBD appeared around 2.9 and 5.8 minutes. In the N2-flushed solution (Figure 4B), these
degradation products were also detected, although their percentage in the sample was
significantly lower than in non-oxygen-free solutions.

3.4. Photo-stability studies

While CBD solutions stored in darkness remained stable during the whole
experiment, samples exposed to light showed CBD degradation. It was significantly more
intensive at 25ºC compared to 5ºC, with a loss of 25 and 18% of the drug, respectively
(Figure 5A). No correlation between CBD concentration and storage time higher than
25% was found in samples stored in darkness, while in samples irradiated, correlation
was higher than 87% ( Figure 5D).

. As shown in Figure 5B, no CBD degradation was found in nitrogen flushed light-
exposed samples within the 72 hours of the study and, as a consequence, no statistically
significant correlation was found between CBD concentration and exposure time.

3.5. Stability in physiological conditions

The CBD solution incubated in simulated physiological conditions (pH 7.4 and
37±0.5ºC) was highly unstable, and 10% of the drug was degraded within the first day of
the study. The degradation profile followed first-order kinetics (Figure 6A).

The main degradation products appeared around 2.9 and 5.8 minutes as illustrated
in (Figure 6B).

3.6 Degradation products

10
 As shown in Table 7, several CBD degradation products have been detected in
our study. The peak that appeared at 5.4-5.8 minutes constituted the main degradation
product of CBD. It appeared in all stability studies (temperature, light and oxygen) and
showed the highest ratio with the CBD peak area (5-35%). The peak detected at 2.6-2.8
minutes also appeared in all experimental conditions, being the second degradation
product, except in samples with H2O2. These two peaks, at 5.4-5.8 and 2.6-2.8 minutes,
could be related to thermal decomposition of CBD. In thermo-stability studies, their ratio
(with the area of CBD peak) significantly increased with the rise in temperature. The peak
detected at 4.1-4.2 minutes only appeared in thermo-stability samples. In samples with
H2O2, a peak at 3.2 minutes was also detected. It is exclusive for this experimental
condition and could consequently be related to the degradation of CBD in highly
oxidizing environments, representing the second major degradation product in this
experimental condition. This peak at 3.2 minutes could be related to the formation of the
monomeric and dimeric hydroxyquinones or cannabelsoic acid compounds described by
other authors.

4. Discussion

In this work, we validated an isocratic reverse phase HPLC method to evaluate

the stability of CBD in solution against different factors (temperature, oxygen and light).
This method provided an efficient separation of CBD from its main degradation products,
and THC. The retention time of CBD (4.72 minutes) allowed rapid analysis of a large
number of samples. The method was lineal and proportional in a range of 1-150µg
CBD/mL ethanol, so that a standard instead of a complete calibration curve was sufficient
to quantify the CBD content in problem samples. Interday and intraday variability were
<2% along the CBD concentration range of 5-100µg/mL, indicating adequate precision
of the analytical method for detecting a CBD degradation of up to 95% in the stability
studies. Three determinations are enough to research a result with a precision of ±2%.
The method was suitable for quantifying CBD in aqueous solution with 0.5% of Tween
80, showing an accuracy according to USP specifications (a slope close to 1 and an
intercept statistically equal to 0 for estimated curve concentrations vs. actual values),
which discarded proportional and constant errors.

11
 The reported results showed that CBD stability is influenced by multiple factors,
such as temperature, light and oxygen. Temperature is one of the most important factors,
since CBD degradation was detected in the study carried out on this drug in ethanol
solution stored at 25, 40, 50, 60 and 70ºC. The higher the storage temperature, the higher
the degradation rate constant (K), except at 70ºC, probably due to a change in the
degradation pathway. For this reason, the K70 value was discarded in the application of
the Arrhenius equation to evaluate the effect of temperature on the degradation rate
constant. The Ea value obtained (92.19KJ/mol) is characteristic of the degradation of
drugs in solution (usually between 40 and 125KJ/mol) and reflects a high sensitivity of
CBD against temperature changes. Using the K25 value extrapolated from the Arrhenius
equation, a t95 value (time at 25ºC for the degradation of 5% of CBD) of 119 days was
calculated. This value was statistically equal to that interpolated from real time data from
the study at 25ºC (117±18 days), which validates the Arrhenius equation to estimate the
stability of the CBD in solution at different temperatures. The high instability of the
solutions of CBD at room temperature (25ºC) requires its storage in a refrigerator. Since
in the stability study at 5ºC no significant degradation was detected after 12 months of
storage (the correlation between the variables CBD concentration vs. storage time was
not statistically significant), the shelf life (t95) of CBD solutions stored in a refrigerator
was estimated from the Arrhenius equation, obtaining a value of 4.8 years, which allows
the storage of solutions of CBD in ethanol in these conditions.

Our study also showed that CBD is more unstable in aqueous solvents than in
ethanol, with a t95 value of 7 days at 25ºC when water with 0.5% Tween 80 was used as
a solvent. This instability was partly due to the presence of dissolved oxygen, more
reactive than atmospheric oxygen. Therefore, when the aqueous solution of CBD was
saturated with an inert gas such as nitrogen, which displaced the oxygen, the stability of
CBD increased, obtaining a t95 value of 16.7 days. However, it is still a value far removed
from that obtained when ethanol was used as a solvent. On the other hand, when H2O2
was incorporated into the aqueous solution of CBD creating a strongly oxidizing
environment, rapid degradation of CBD occurred (t95 of 1.8 days), confirming that it is
an easily oxidizable substance.

Cannabinoid molecules have also showed to be sensitive to light, and must be

stored in darkness. In our study, CBD has demonstrated to be sensitive to light when
incubated at both 5 and 25ºC in a photostability chamber according to ICH guidelines.

12
Due to the influence of oxygen on CBD stability, nitrogen flushed solutions were also
analyzed at 5ºC in order to determine its participation in the light instability of CBD. In
this work, when the air of the ampoules containing CBD solution was replaced by an inert
gas (nitrogen), the degradation of CBD exposed to light was avoided. This suggests that
the photolytic reaction of CBD is oxidative.

Finally, in simulated physiological conditions (pH 7.4 and 37ºC), CBD was highly
unstable, with 10% of the drug degraded in 24h. When comparing CBD aqueous solution
(O±) incubated at 25ºC, a higher degradation was found. While at 25ºC the estimated
half-life (50% of degradation) was 19 days, the half-life estimated from data obtained at
37ºC in PBS was 7 days. This difference could be attributed not only to the increase in
the degradation rate with the incubation temperature but also to the contact with the air
of the dissolution during the assay in physiological conditions.

5. Conclusions

The present study provides data on degradation kinetics of CBD in solution under
different environmental conditions. The solvent used has an important influence on the
stability of the CBD, so that ethanolic solutions are much more stable than aqueous ones.
In spite of this, it is necessary to keep the ethanol solutions of CBD in the refrigerator in
order to guarantee a shelf life of more than one year, since at room temperature the
solutions remain stable for less than 4 months. The Arrhenius equation (with Ea=
92.19KJ/mol and A=6.65E+12d-1) is a validated tool to estimate the stability of CBD
alcoholic solutions at different temperatures. CBD is an easily oxidizable substance, and
its aqueous solutions must be saturated with an inert gas like nitrogen in order to increase
their stability from 7 to 19 days at room temperature. Light catalyzes the oxidation of
CBD in solution, so all solutions must be protected against it. The use of antioxidants in
CBD formulations should be considered. Finally, the instability of CBD in solution under
the test conditions must be taken into account in the design of both in vitro studies
simulating physiological conditions and in vivo studies.

Acknowledgements

13
 This work was supported by the Complutense Research Fund (Ref. FEI16/83) and
by the Santander-UCM research group “Formulation and bioavailability of new drug
products” (Ref. UCM-910939). A.I. Fraguas-Sánchez have been granted with a research
fellowship (Ref: FPU 14/06441) from the Spanish Ministry of Education.

REFERENCES

[1] N. Jordan, J. Zakrajsek, S. Bohanec, R. Roskar, I. Grabnar, Applying the methodology of

Design of Experiments to stability studies: a Partial Least Squares approach for evaluation of
drug stability, Drug development and industrial pharmacy, 44 (2018) 778-786.
[2] S. Pisanti, M. Bifulco, Medical Cannabis: A plurimillennial history of an evergreen, Journal of
cellular physiology, 234 (2019) 8342-8351.
[3] M.A. ElSohly, M.M. Radwan, W. Gul, S. Chandra, A. Galal, Phytochemistry of Cannabis sativa
L, Progress in the chemistry of organic natural products, 103 (2017) 1-36.
[4] E. Papaseit, C. Perez-Mana, A.P. Perez-Acevedo, O. Hladun, M.C. Torres-Moreno, R. Muga,
M. Torrens, M. Farre, Cannabinoids: from pot to lab, International journal of medical sciences,
15 (2018) 1286-1295.
[5] K. Cohen, A. Weizman, A. Weinstein, Positive and Negative Effects of Cannabis and
Cannabinoids on Health, Clinical pharmacology and therapeutics, DOI 10.1002/cpt.1381(2019).
[6] N. Maurya, B.K. Velmurugan, Therapeutic applications of cannabinoids, Chemico-biological
interactions, 293 (2018) 77-88.
[7] A.I. Fraguas-Sanchez, A.I. Torres-Suarez, Medical Use of Cannabinoids, Drugs, 78 (2018)
1665-1703.
[8] M. Munjal, M.A. ElSohly, M.A. Repka, Chemical stabilization of a Delta(9)-
tetrahydrocannabinol prodrug in polymeric matrix systems produced by a hot-melt method:
Role of microenvironment pH, AAPS PharmSciTech, 7 (2006) E114-e125.
[9] D.J. van Drooge, W.L. Hinrichs, K.A. Wegman, M.R. Visser, A.C. Eissens, H.W. Frijlink, Solid
dispersions based on inulin for the stabilisation and formulation of delta 9-
tetrahydrocannabinol, European journal of pharmaceutical sciences : official journal of the
European Federation for Pharmaceutical Sciences, 21 (2004) 511-518.
[10] F. Grotenhermen, Pharmacokinetics and pharmacodynamics of cannabinoids, Clinical
pharmacokinetics, 42 (2003) 327-360.
[11] R. Mechoulam, Chemistry of Cannabis Handbook of Experimental Pharmacology 55 (1981)
119-134.
[12] K.B. Scheidweiler, M. Andersson, M.J. Swortwood, C. Sempio, M.A. Huestis, Long-term
stability of cannabinoids in oral fluid after controlled cannabis administration, Drug testing and
analysis, 9 (2017) 143-147.
[13] R. Pacifici, E. Marchei, F. Salvatore, L. Guandalini, F.P. Busardo, S. Pichini, Evaluation of
cannabinoids concentration and stability in standardized preparations of cannabis tea and
cannabis oil by ultra-high performance liquid chromatography tandem mass spectrometry,
Clinical chemistry and laboratory medicine, 55 (2017) 1555-1563.
[14] R. Pacifici, E. Marchei, F. Salvatore, L. Guandalini, F.P. Busardo, S. Pichini, Evaluation of
long-term stability of cannabinoids in standardized preparations of cannabis flowering tops
and cannabis oil by ultra-high-performance liquid chromatography tandem mass spectrometry,
Clinical chemistry and laboratory medicine, 56 (2018) 94-96.
[15] I. Q1A(R2), Stability testing of new drug substances and products. CPMP/ICH/2736/99,
2003.

14
[16] P.W.N. Joachim Ermer Method Validation in Pharmaceutical Analysis: A Guide to Best
Practice, 2nd Edition ed., Wiley2014.
[17] N.M. Kogan, R. Rabinowitz, P. Levi, D. Gibson, P. Sandor, M. Schlesinger, R. Mechoulam,
Synthesis and antitumor activity of quinonoid derivatives of cannabinoids, J Med Chem, 47
(2004) 3800-3806.
[18] R. Mechoulam, L. Hanus, Cannabidiol: an overview of some chemical and pharmacological
aspects. Part I: chemical aspects, Chem Phys Lipids, 121 (2002) 35-43.
[19] I. Trofin, G. Dabija, D.-I. Vaireanu, F. Laurentiu, Long - term Storage and Cannabis Oil
Stability, Revista da Chimie, 53 (2012) 294.
[20] ICHQ1B, Stability Testing : Photostability Testing of New Drug Substances and Products,
1996.
[21] D. Raghuvanshi, G. Nkepang, A. Hussain, H. Yari, V. Awasthi, Stability study on an anti-
cancer drug 4-(3,5-bis(2-chlorobenzylidene)-4-oxo-piperidine-1-yl)-4-oxo-2-butenoic acid
(CLEFMA) using a stability-indicating HPLC method, Journal of pharmaceutical analysis, 7
(2017) 1-9.
[22] FDA, FDA. Reviewer Guidance: Validation of Chromatographic Methods, DOI (1994).

15
Table 1: Linearity statistical parameters. * p value<0.05

Parameter Result
Linearity range (µg/mL) 1-150
Correlation coefficient (r) 0.999
Determination coefficient (r2) 0.998
Slope "b" (mUA·mL·µg-1) 25.209
SD slope (mUA·mL·µg-1) 0.193
Intercept (mUA) -17.986
SD intercept (mUA) 13.167
" F " ANOVA test for regression (p) 17015.510 (0.0000)*
" F " ANOVA test for lineal 1.519 (0.2385)
model(p)
“t” Student's test for proportionality -1.366 (0.1863)
(p)

Table 2: Precision data for HPLC method validation.

CBD concentration Intra-day variability Inter-day variability

(µg/ml) (CV, %) (CV, %)

5 0.8 1.3
50 1.7 1.8
100 1.3 1.7

16
 Table 3: Statistical parameters of accuracy studies. * p value<0.05

PARAMETER RESULT
Mean recovery (%) 100.4
CV (%) 1.2
Correlation coefficient (r) 0.993
Determination coefficient (r2) 98.694
Slope "b" ± SD 0.997± 0.028
Intercept ± SD -0.003± 1.898
" F " ANOVA test for regression (p) 1208.82 (0.000)*
" F " ANOVA test for lineal model (p) 0.16 (0.609)
“t” Student's test for proportionality (p) -0,002 (0.998)

Table 4: Degradation rate constants of CBD at different temperatures after fitting the data
to first order kinetic. *p >0.05

Temperature (ºC)
T (ºC) 5 25 40 50 60 70
Correlation 0.3588 -0.807 -0.945 -0.951 -0.977 -0.989
coefficient (r)
Determination 12.88 65.12 89.35 90.39 95.49 97.71
coefficient (r2)(%)
F ANOVA regression 2.07* 31.74 76.62 123.17 165.57 382.96
Standard error --- 0.0359 0.1014 0.0500 0.117 0.079
K (day-1) --- 4.38E-4 2.682E-3 6.722E-3 0.023 0.022
Standard deviation --- 7.77E-5 4.14E-4 8.95E-4 2.093E-3 1.410 E-3
of degradation rate
constant (SDK)
t95 (days) 117.±18

17
Table 5: Parameters of Arrhenius equation.

Correlation coefficient (r) -0.998 K25extrap. (d-1) 4.316E-4

Determination 99.66 t95 (25ºC) (days) 118.85
coefficient (%)
Standard error (s) 0.118941 K5extrap. (d-1) 2.954E-05
Slope (b ) (1/ºK) -11107.4 t95 (5ºC) (days) 1736.03
Standard deviation of the 455.311
slope(SDb)(1/ºK)
Activation energy (Ea) 92.191
(KJ/mol)
Intercept (a) 29.525
Standard deviation of the 1.441
intercept (SDa)
Frequency factor (A) (d-1) 6.646E+12
F ANOVA regression (p) 595.13 (0.002)
F ANOVA lack of fit(p) 1.34 (0.318)

18
Table 6: Parameters obtained from the oxygen dependent study fitting to a fist order
kinetic.

O+ O+/- O-
Correlation coefficient -0.873 -0.816 -0.885
(r)
Determination 76.22 66.56 78.32
coefficient (r2) (%)
F ANOVA regression (p) 60.90 (0.00) 29.86 (0.0001) 72.26 (0.00)
F ANOVA lak of fit (p) 0.49 (0.8423) 3.50 (0.050) 1.90 (0.1505)
Standard error (s) 0.266 0.088 0.023
Degradation rate 0.029 0.007 0.003
constant (K) (days-1)
Standard deviation of 0.004 0.001 0.0003
degradation constant
(SDk )(days-1)
t95 (days) 1.8 7.6 19.7

Table 7: Ratio (%) [area of degradation product]/[area of CBD peak], calculated when a
CBD degradation around 30% was detected.

Area Ratio of CBD degradation products:

retention time (min)
Experimental 2.6-2.8 3.2 4.1-4.2 5.4-5.8
condition
40ºC 8.21±0.52 1.85±0.92 11.54±2.53
50ºC 10.6±1.10 4.59±0.12 15.94±7.04
60ºC 21.1±1.41 14.23±1.04 34.55±4.14
Oxygen (O+) 1.89±0.12 9.41±1.94 15.41±1.48
Photo-stability (25ºC) 2.72±0.81 4.95±0.41
PBS 5.13±0.92 6.50±1.31

19
FIGURES

Figure 1: Chromatogram of a CBD solution of 100 µg/ml (A), calibration curve of CBD
(1-150 µg/mL) (B) and system suitability parameters (C).

Figure 2: Graphical representation of actual drug concentration of CBD samples

dissolved in water with Tween 80 (0.5%) versus the values obtained from calibration
curve of linearity study.

20
Figure 3: Stability studies of CBD in solution stored at 5, 25 and 40ªC (A) and 50, 70,
and 60ºC (B). Representative chromatogram of a CBD solution of 100µg/ml after exposed
to 40ºC for 12 months (C).

21
Figure 4: Oxygen dependent stability studies of CBD in solution (A). Representative
chromatograms of CBD in solution (100µg/ml) from oxygen degradation studies (2
month samples) (B-D).

Figure 5: Photo-stability studies of CBD in solution: non-nitrogen flushed samples (A)

and nitrogen flushed samples (B). Representative chromatogram of CBD in solution

22
exposed to light at 25ºC for 72 hours (C). Statistical analysis of data (D). ** with the best
fit (Y^2/√X), *** with the best fit (1/Y-√X).

Figure 6: Stability of CBD in simulated physiological conditions and first order

degradation kinetic fitting (A). Representative chromatogram of CBD in phosphate buffer
solution (pH 7.4) incubated at 37ºC for 10 days (B).

23