Downloaded from https://journals.lww.com/pain by BhDMf5ePHKav1zEoum1tQfN4a+kJLhEZgbsIHo4XMi0hCywCX1AWnYQp/IlQrHD3ctdU4DsFBzRPCzfSl6Tv6/K99MsjyAR9/Bqk+3SqukI= on 06/24/2019
Downloadedfromhttps://journals.lww.com/pain by BhDMf5ePHKav1zEoum1tQfN4a+kJLhEZgbsIHo4XMi0hCywCX1AWnYQp/IlQrHD3ctdU4DsFBzRPCzfSl6Tv6/K99MsjyAR9/Bqk+3SqukI= on 06/24/2019
An experimental randomized study on the analgesic
effects of pharmaceutical-grade cannabis in
chronic pain patients with fibromyalgia
Tine van de Donk
, Marieke Niesters
, Mikael A. Kowal
, Erik Olofsen
, Albert Dahan
*, Monique van Velzen
In this experimental randomized placebo-controlled 4-way crossover trial, we explored the analgesic effects of inhaled
pharmaceutical-grade cannabis in 20 chronic pain patients with fibromyalgia. We tested 4 different cannabis varieties with exact
knowledge on their Δ
-tetrahydrocannabinol (THC) and cannabidiol (CBD) content: Bedrocan (22.4-mg THC, ,1-mg CBD;
Bedrocan International BV, Veendam, the Netherlands), Bediol (13.4-mg THC, 17.8-mg CBD; Bedrocan International BV,
Veendam, the Netherlands), Bedrolite (18.4-mg CBD, ,1-mg THC; Bedrocan International BV, Veendam, the Netherlands), and
a placebo variety without any THC or CBD. After a single vapor inhalation, THC and CBD plasma concentrations, pressure and
electrical pain thresholds, spontaneous pain scores, and drug high were measured for 3 hours. None of the treatments had an effect
greater than placebo on spontaneous or electrical pain responses, although more subjects receiving Bediol displayed a 30%
decrease in pain scores compared to placebo (90% vs 55% of patients, P50.01), with spontaneous pain scores correlating with the
magnitude of drug high (r520.5, P,0.001). Cannabis varieties containing THC caused a significant increase in pressure pain
threshold relative to placebo (P,0.01). Cannabidiol inhalation increased THC plasma concentrations but diminished THC-induced
analgesic effects, indicative of synergistic pharmacokinetic but antagonistic pharmacodynamic interactions of THC and CBD. This
experimental trial shows the complex behavior of inhaled cannabinoids in chronic pain patients with just small analgesic responses
after a single inhalation. Further studies are needed to determine long-term treatment effects on spontaneous pain scores,
THC–CBD interactions, and the role of psychotropic symptoms on pain relief.
Keywords: Cannabis, Chronic pain, Fibromyalgia, Pharmacokinetics, THC, CBD, Placebo cannabis, Pain models
In the current opioid epidemic, there is the need for pharmaceutical
alternatives to opioid treatment in patients with chronic pain. An
alternative may be found in the chemicals of the cannabis plant
(Cannabis sativa L.), which contains over 500 chemical compo-
nents, with more than 100 of them being cannabinoids.
Cannabinoids, or more specifically phytocannabinoids, are the
main active chemical components of the cannabis plant. They
exhibit most of their pharmacological effects via cannabinoid type 1
) G-protein-coupled receptors.CB
are located mainly in the central nervous system, whereas CB
receptors are mostly found on immune cells.
form part of the endocannabinoid system, a modulatory biological
system that influences the activity of different neurotransmitters with
their own ligands, the endocannabinoids, such as anandamide and
As for cannabis, its major cannabinoid is
-tetrahydrocannabinol (THC), a partial CB
-receptor agonist, that
produces a variety of effects including altered cognition and motor
function, analgesia, and psychotropic effects (eg, drug high).
Another key component of cannabis is cannabidiol (CBD) that,
while nonintoxicating, does affect mood and cognition.
It is a CB
receptor antagonist and additionally has agonist activity at the 5HT-
receptor and stimulates the vanilloid receptor type 1 with similar
efficacy as capsaicin.
In this experimental trial, we explored the effect of
pharmaceutical-grade cannabis in patients with chronic pain caused
bythefibromyalgia(FM)syndrome.Fibromyalgia is characterized by
chronic widespread pain, often accompanied by secondary
symptoms including sleep disturbance, tiredness, and cognitive
symptoms such as memory deficits.
This condition predominantly
affects women, with a worldwide prevalence of 2% to 8% and
conventional pharmacologic treatment is considered only mildly
We explored the analgesic effects of inhaled pharmaceutical-
grade cannabis using the cannabis plant with all its natural
components. We tested 4 different varieties with exact knowledge
on their THC and CBD content. The varieties used were Bedrocan
with a high THC/low CBD content, Bedrolite with a high CBD/low
Sponsorships or competing interests that may be relevant to content are disclosed
at the end of this article.
Department of Anesthesiology, Leiden University Medical Center, Leiden, the
Bedrocan International BV, Veendam, the Netherlands
*Corresponding author. Address: Anesthesia and Pain Research Unit, Department
of Anesthesiology, Leiden University Medical Center, H5-022, 2300 RC Leiden, the
Netherlands. Tel: 131 71 526 9111. E-mail address: firstname.lastname@example.org (A. Dahan).
Supplemental digital content is available for this article. Direct URL citations appear
in the printed text and are provided in the HTML and PDF versions of this article on
the journal’s Web site (www.painjournalonline.com).
PAIN 160 (2019) 860–869
Copyright ©2018 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf
of the International Association for the Study of Pain. This is an open-access article
distributed under the terms of the Creative Commons Attribution-Non Commercial-
No Derivatives License 4.0 (CCBY-NC-ND), where it is permissib le to download and
share the work provided it is properly cited. The work cannot be changed in any way
or used commercially without permission from the journal.
860 T. van de Donk et al.·160 (2019) 860–869 PAIN
THC content, Bediol with a combined high THC/high CBD content,
and a placebo variety without any THC or CBD content. This
approach enabled exploration of cannabis effects on pain relief
relative to placebo cannabis that was similar in smell, appearance,
and handling compared with the other varieties. We assessed relief
of experimental pressure pain, electrical pain, and spontaneous
pain (primary endpoints), as well as the subjective and psychotropic
effects. We hypothesized that compared with placebo treatment, all
THC-containing treatments would cause greater analgesic
responses for both spontaneous pain and evoked pain models.
2.1. Ethics and trial registration
This single-center, double-blind, placebo-controlled, 4-way
crossover study, with acronym Spirocan, was performed at the
Anesthesia and Pain Research Unit of the Department of
Anesthesiology at LUMC. The protocol was approved by the
local institutional review board and the Central Committee on
Research Involving Human Subjects in The Hague. The study
was registered at trialregister.nl under identifier NTR6091 and in
the European Union Drug Regulating Authorities Clinical Trials
(EUDRACT) database under identifier 2015‐003811‐39. Before
enrollment, all patients gave written informed consent.
2.2. Patients: inclusion and exclusion criteria
Female patients diagnosed with FM were approached to
participate in the study through announcements in local
newspapers and the web site of the association of patients
with FM. When patients indicated interest in the study and were
diagnosed with FM by a rheumatologist, they were queried for
inclusion and exclusion criteria. Inclusion criteria were: a pain
score $5 for most of the day (on a verbal pain scale from 0 5no
pain to 10 5most pain imaginable) and positive diagnostic
criteria of the 2010 American College of Rheumatology.
criteria include a widespread pain index (WPI) $7 (on a scale
from 0 to 19) and a symptom severity (SyS) score $5 (on a scale
from 0 to 12) or a WPI of 3 to 6 and a SyS score $9. The WPI
defines the number of body areas in which a patient experi-
enced pain during the past week; the SyS score indicates the
level of other main symptoms of FM such as fatigue, non-
refreshing sleep, and cognitive symptoms. The presence of
autonomic complaints such as diarrhea or obstipation, dizzi-
ness, dry mouth/eyes, etc. was not a reason for exclusion, as we
consider these symptoms consistent with the FM syndrome.
Exclusion criteria included age ,18 years, any medical,
neurological, or psychiatric illness, use of strong opioids or
other painkillers except paracetamol and/or ibuprofen, benzo-
diazepine use, any known allergies to study medication, illicit
drug or alcohol use, recent use of cannabis, pregnancy, breast
feeding, and the presence of pain syndromes other than FM. On
the day of screening and on the morning of each of the 4 study
days, the urine of the patient was tested for illicit drug use using
a dipstick (Alere Toxicology Plc, Oxfordshire, United Kingdom;
the stick tests for cocaine, amphetamine, cannabinoids,
phencyclidine, methadone, benzodiazepines, tricyclic antide-
pressants, and barbiturates). In case of a positive test, the
subject was excluded from the study. Subjects were instructed
not to eat for at least 6 hours and drink for at least 2 hours before
the study visit. Any foods or beverages containing caffeine such
as coffee, tea, or chocolate were not allowed for 24 hours before
the study visit.
2.3. Study design: drugs, inhalation, and blood sampling
Patients visited the research unit on 5 occasions. On their first
visit, the patients were screened (medical history, physical
examination, and urinalysis) and familiarized with the experimen-
tal setup (they were, for example, trained in the inhalation
process). On each of their next visits, the patients received 1 of 4
possible cannabis treatments (in random order) with at least 2
weeks between visits.
The active cannabis substances were composed of the dried,
milled, and homogenized flowers of the plant Cannabis sativa L.,
which were cultivated under standardized conditions in line with
the requirements of good manufacturing practices (GMP). We
used 4 distinct pharmaceutical-grade cannabis varieties, all
obtained from Bedrocan International BV (Veendam, the Nether-
lands) and all prepared by Proxy Laboratories BV (Leiden, the
Netherlands) under GMP conditions:
(1) Bedrocan: The Bedrocan cannabis variety contains 22% THC
(220 mg per gram) and less than 1% CBD. It was developed in the
Netherlands out of a requirement by the Dutch Health Ministry to
have a “high THC” variety available to patients. We used 100 mg
that contained 22.4-mg THC and less than 1-mg CBD.
(2) Bediol: The Bediol cannabis variety is characterized by the
combination of 6.3% THC (63 mg per gram) and 8% CBD (80
mg per gram). We used 200 mg that contained 13.4-mg THC
and 17.8-mg CBD.
(3) Bedrolite: This variety is composed of 9% CBD (90 mg per
gram) and less than 1% THC. We used 200 mg that contained
18.4-mg CBD and less than 1-mg THC.
(4) Placebo: The placebo was derived from the Bedrocan
cannabis variety after selective removal of the cannabinoids
by solvent extraction by Proxy Laboratories BV under GMP
conditions. After removal of the cannabinoids, the specific
terpene profile (responsible for smell and taste) was restored in
a subsequent manufacturing step. Consequently, the placebo
had a moisture content and terpenoid profile matching the
active drug (Bedrocan).
Study medication was analyzed for cannabinoid content,
terpene profile, and water content by an independent quality
control laboratory. In addition, tests were performed to ensure
that unwanted elements were absent such as adulterants,
microbes, heavy metals, and pesticides. The pharmacy and
ethics committee reviewed and approved the products’ quality
certificates before dispensing the cannabis to the research team.
During the study, all varieties were refrigerated at 2 to 8 ˚C in triple-
layer laminated foil pouches.
Patients were dosed with cannabis vapor. All cannabinoids
are mostly present in the plant in their acid form. Application of
heat is needed for decarboxylation of the cannabinoid acids
into their active forms (eg, THC acid into THC).
All 4 cannabis
varieties were vaporized using the Volcano Medic vaporizer
(Storz & Bickel GmbH & Co, Tuttlingen, Germany)—a safe and
reliable method of intrapulmonary administration of cannabi-
The Volcano heated the homogenized plant material
to 210 ˚C to allow for conversion of the THC acid and CBD acid
into THC and CBD vapor for inhalation. The vapor was collected
in an 8-L plastic balloon that, after inflation, was detached from
the vaporizer and subsequently equipped with a mouthpiece
for inhalation. For the purpose of blinding, the balloon was
covered with an opaque plastic bag so that no variation in
density of the vapor was visible between visits. The evaporation
process was performed by a member of the research team not
involved in the study proceedings. Before and after each
evaporation, the device was cleaned with alcohol. The
April 2019·Volume 160 ·Number 4 www.painjournalonline.com 861
complete content of the balloon was inhaled through the mouth
within 3 to 7 minutes, and each breath was held for 5 seconds
after each inhalation.
On each occasion, an arterial line was placed in the left or right
radial artery for blood sampling. Five milliliter of blood was
obtained at t50 (control sample, before inhalation), 5, 10, 20, 30,
40, 50, 60, 90, 120, and 180 minutes after the start of inhalation.
Blood was collected in EDTA tubes (covered with an aluminum
foil), centrifuged at 2000gat 4˚C; separated plasma was stored at
280˚C until analysis. The samples were analyzed by Analytical
Biochemical Laboratory BV, Assen, the Netherlands. All handlingof
the samples was done in a darkened room to prevent the cannabis
molecules from disintegrating. Determinationof the CBD, THC, and
its active metabolite 11‐hydroxy‐THC (11-OH-THC) plasma con-
centrations was performed using liquid chromatography with
tandem mass spectrometer detection (LC‐MS/MS). In Supple-
mental Materials 1 to 3, the analysis specifications including
chromatograms of the 3 cannabis varieties are given for 2 (low and
high) concentrations (available at http://links.lww.com/PAIN/A705).
2.4. Study design: pain tests, questionnaires, and safety
All subjects rated their FM pain on an 11-point visual analogue
scale (from 0 5no pain to 10 5most severe pain imaginable) at
baseline (before cannabis inhalation) and at 1, 2, and 3 hours after
Two experimental pain tests were performed:
(1) Pressure pain test
: A pressure algometer (FDN 100; Wagner
Instruments Inc, Greenwich, CT) was used to deliver pressure
pain on a skin area of 1 cm
between the thumb and index
finger; the affected area overlays the adductor pollicis muscle.
The algometer has a force capacity (6accuracy) of 100 62N
(10 60.2 kgf) and graduation of 1 N (100 gf), respectively. A
gradually increasing pressure was manually applied, and the
subjects were asked to indicate when the procedure became
painful (pressure pain threshold). All measurements were
obtained in triplicate at t50 (baseline), 12, 22, 32, 42, 62, 92,
122, 152, and 182 minutes after the start of inhalation. The 3
measurements were averaged for further analysis.
(2) Electrical pain test
: Electrical pain was induced using a locally
designed computer interfaced electrical currents stimulator
(CICS, Leiden University Medical Center, Leiden, the Nether-
lands). The stimulator was connected to 2 electrodes (surface
area 0.8 cm
) placed on the tibial surface of the right leg,
approximately 10 cm above th e medial malleolus. The stimulator
produced a stimulus train (stimulus duration 0.2 ms at 10 Hz)
that increased from 0 mA at 0.5 mA/second (cutoff 128 mA).
The subjects were instructed to press a control button when
pain was first felt (pain threshold) and when the pain became
unbearable (pain tolerance; this ended the stimulus train).
Measurements were obtained at t50, 10, 20, 30, 40, 60, 90,
120, 150, and 180 minutes after the start of cannabis inhalation.
Two questionnaires were taken to assess the effect of drug
treatment on mental and psychoactive cannabis effects:
(1) Bowdle questionnaire
: This questionnaire evaluates 3
psychedelic effects (drug high, alterations in internal perception,
and alterations in external perception) from 13 questions scored
on a 100-mm visual analogue scale (from 0, no effect, to 100,
maximum effect). Internal perception reflects inner feelings that
do not correspond with the reality and is derived from questions
regarding the hearing of unrealistic voices or sounds and having
unrealistic thoughts and paranoid or anxious feelings. The
external perception indicates a misperception of an external
stimulus or change in the awareness of the subject’s
surroundings and is derived from questions regarding the
perceptual change of body parts, the change of surroundings,
the altered passing of time, the difficulty of controlling thoughts,
and the change in color and sound intensity.
(2) Bond and Lader questionnaire
: The Bond and Lader
scales are calculated from sixteen 100-mm visual analogue
scales. The endpoints are set at antonymous word pairs such
as “alert–drowsy,” “well coordinated–clumsy,” “mentally
slow–quick witted,” and “incompetent–proficient.” The study
participant’s task is to make a mark on each scale at the point
that best describes how they currently feel considering that the
2 anchors reflect the greatest extent they experience each
state. Responses from these 16 scales are then scored to yield
3 main factors of alertness (alert, strong, clear‐headed,
coordinated, energetic, quick‐witted, attentive, proficient,
and interested), contentment (contented, happy, amicable,
gregarious, and tranquil), and calmness (calm and relaxed). A
high score indicates impairment.
The subjects were queried before drug inhalation and at 30-
min intervals after the start of inhalation. Adverse events and
serious adverse events were collected in the case record form. In
case of a serious adverse event, the event was treated, and no
further measurements were obtained. In case of an adverse event
(eg, nausea, vomiting, headache, and dizziness), no further action
was taken apart from supportive care.
2.5. Randomization, allocation, and blinding
Randomization was performed by the pharmacy using a computer-
generated randomization list. A distinct randomization sequence
was created for each subject; randomization sequence was
controlled withjust 2 subjects with an identical treatment sequence.
On the day before the experiment, the subject was allocated to
treatment by the pharmacy after receiving a fax message from the
investigators with the participant’s identifier code and study visit
number. Treatment was prepared on the day of the study and
collected by a technician from the pharmacy in a closed opaque
canister labeled with the patient’s identifier code and study visit
number; the contents of the canister were emptied in the vaporizer.
The study team was next presented with the filled opaque balloon
just before the actual cannabis inhalation. The investigators (and
patients) remained blinded until data analysis was complete (June
2018). The study was independently monitored ensuring that all
good clinical practice requirements were met.
2.6. Statistical analysis: sample size and assessment of
Considering the data from Wallace et al.,
we calculated the
need for 20 subjects to allow for a significant separation between
treatments with a power .0.9 and alpha 50.05. In case of
dropout after one visit, the data were discarded, and a new
subject was recruited. Before the data analyses, all variables were
screened for missing data, homoscedasticity, distribution abnor-
malities, and outliers. For both primary and secondary endpoints,
the effect of active treatment (Bedrocan, Bedrolite, or Bediol) on
the change in effect was compared between treatments using
a mixed model. Treatment was set as a fixed effect, a random
effect for the subject was added to account for repeated
measurements over time, and treatment order was added as
a covariate. For spontaneous pain, the responder rate was
determined for each treatment and compared with placebo
responder rates using a x
test. A responder was defined as
having a reduction in spontaneous pain score of at least 30% or
50% at one or more measurements. In addition, the change in
862 T. van de Donk et al.·160 (2019) 860–869 PAIN
spontaneous pain score relative to baseline was related to the
drug high score by Spearman’s r. The number of adverse events
between the 3 active treatments and placebo was analyzed using
test. SPSS (IBM Corp Released 2017; IBM SPSS Statistics
for Windows, Version 25.0, Armonk, NY: IBM Corp) was used for
all analyses with Pvalues ,0.05 considered significant. All data
are reported as mean 6SD, unless otherwise stated.
Twenty-five patients were recruited for participation. Five patients
ended their participation after their first study visit for unknown
reasons (n 51), side effects such as dizziness and nausea (n 53),
and fear of needles (n 51) (Fig. 1). All were replaced by another
patient according to the protocol. The 20 patients who completed
the trial were on average 39 613 years with an average weight of 82
620 kg and height of 169 67 cm (body mass index 29 67kg/m
At screening, patients reported an average verbal pain score of 7.20
61.24 units and were all diagnosed with FM with a WPI of 13.9 6
2.6, SyS of 9.2 61.3, and 14.9 62.9 of positive tender points.
Cannabis inhalation was achieved in (minutes:seconds) 5:03 6
2:54 (21 611 inhalations; Bedrocan), 6:57 64:05 (23 611
inhalations; Bediol), 5:30 62:37 (22 610 inhalations; Bedrolite),
and 2:48 61:40 (14 66 inhalations; Placebo). The complete
content of the balloon was inhaled by all subjects. All 3 active
treatments, but not placebo, were associated with several
adverse effects (Table 1), with frequent effects related to the
inhalation of cannabis (coughing during inhalation in 65%-70%,
sore throat and bad taste during inhalation in 25%-35% of
participants). Most adverse effects unrelated to the inhalation
process were drug high in 40% to 80%, dizziness in 15% to 20%,
and nausea in 5% to 30% of participants. Two patients reported
feelings of drug high after placebo treatment. There were no
differences in frequency of adverse effects between active
treatments (P.0.05). No serious adverse events occurred.
After inhalation of all 3 active treatments, THC, its metabolite
11-OH-THC, and CBD, were detectable with the following CMAX
and TMAX values (Fig. 2). Bedrocan: THC 82 620 ng/mL at t55
minutes, 11-OH-THC 5 63 ng/mL at 10 minutes, and CBD 0.2 6
0.3 ng/mL at 5 minutes; Bediol: THC 76 635 ng/mL at t55
minutes, 11-OH-THC 5 63 ng/mL at 10 minutes, and CBD 80 6
029 ng/mL at 5 minutes; and Bedrolite: THC 13 65 ng/mL at t5
5 minutes, 11-OH-THC 0.9 60.5 ng/mL at 10 minutes, and CBD
155 657 ng/mL at 5 minutes. No cannabinoids were detectable
after placebo inhalations.
None of the treatments had an effect greater than placebo on
spontaneous pain scores or electrical pain responses (Fig. 3 and
Table 2). By contrast, both Bedrocan and Bediol caused
a significant increase in tolerance to the pressure applied to the
skin over the adductor pollicis muscle for the duration of the
study. The largest effect was observed for the cannabis variety
that contained high doses of both THC and CBD (Bediol) with an
increase in tolerated pressure of 9 to 11 kgf from t520 to 90
minutes (P,0.001 vs placebo; t50 minutes is the start of
cannabis inhalation). Over this same time range, Bedrocan
increased the tolerated pressure by 7 to 9 kgf (P50.006 vs
placebo). With respect to spontaneous pain scores and tolerance
to pressure pain, Bediol had significantly greater effects than
Bedrolite (P50.04 for both endpoints, Table 2 and Fig. 3).
After placebo treatment, 11 and 6 patients had 30% and 50%
reduction in pain scores on at least one measurement period,
respectively. Comparing these responder rates to active treat-
ment, significantly more patients responded to Bediol with
a decrease in spontaneous pain by 30% (n 518, P50.01;
Fig. 4) but not with a decrease by 50% (n 59, P50.052). At both
responder rates, all other treatments had response profiles not
different from placebo (Fig. 4). Spontaneous pain scores were
strongly correlated with the magnitude of drug high for Bedrocan
(r520.5, P,0.001) and for Bediol (r520.5, P,0.001).
Psychoactive effects of treatment, as measured by the Bowdle
questionnaire, are given in Table 3. Bedrocan and Bediol caused
moderate drug high responses, on average just below 50% of the
maximum possible response (Fig. 3B), but significantly greater
than placebo (P,0.001). Bedrolite had less intense drug high
responses compared with either Bedrocan (P50.003) or Bediol
(P,0.001). Small effects were seen for changes in internal
perception (Bediol vs placebo, max. mean difference with
placebo 7 mm, P50.009, Table 3) and external perception
(Bedrocan and Bediol vs placebo, max. mean difference with
placebo 17 mm, P,0.001), indicative of limited psychosis-like
effects after Bedrocan and Bediol treatment. Bedrolite caused
smaller changes in internal perception than Bediol (P50.04) and
smaller changes in external perception than both Bediol (P5
0.004) and Bedrocan (P50.01). The responses to the Bond and
Lader questionnaire indicate mild deterioration in mood observed
during Bediol treatment (max. mean difference with placebo
Figure 1. Consort flow diagram. FM, fibromyalgia.
Incidence of adverse events.
Bedrocan Bediol Bedrolite Placebo
Drug high, n (%) 16 (80) 16 (80) 8 (40) 2 (10)
Coughing, n (%) 14 (70) 14 (70) 13 (65) 0 (0)
Sore throat, n (%) 2 (10) 7 (35) 1 (5) 0 (0)
Bad taste, n (%) 5 (25) 6 (30) 5 (25) 0 (0)
Dyspnoea, n (%) 0 (0) 1 (5) 0 (0) 0 (0)
Dizzy, n (%) 3 (15) 4 (20) 2 (10) 0 (0)
Headache, n (%) 1 (5) 2 (10) 3 (15) 1 (5)
Nausea, n (%) 3 (15) 6 (30) 1 (5) 0 (0)
Vomiting, n (%) 0 (0) 0 (0) 1 (5) 0 (0)
Sleepy, n (%) 1 (5) 0 (0) 1 (5) 0 (0)
n, number of patients affected.
April 2019·Volume 160 ·Number 4 www.painjournalonline.com 863
11 mm, P50.02, Table 3) and mild deterioration in alertness
during Bedrocan (max. mean difference with placebo 21 mm, P5
0.02). Some small differences in mood and alertness were
observed among the 3 active treatments (Table 3).
To assess whether blinding of active vs placebo treatment was
successful, we calculated Bang’s blinding index (Bang’s BI),
which translates correct vs random guessing into a single
number. Bang’s BI ranges between 21 and 1 with 0 a perfect
blinding and values .0.5 or ,20.5 indicative of failure of blinding
above random guessing in the majority of subjects. Bang’s BI
values were between 0.3 and 0.4 just after inhalation for the 3
active treatments (40% of patients correctly guessed that they
received active treatment, whereas 50% of patients were unable
to determine what treatment they received). At the end of the
experiment, more subjects correctly guessed that they received
active treatment after Bedrocan (Bang’s BI 0.85) or Bediol
(Bang’s BI 0.90) inhalation. After placebo treatment, Bang’s BI
was 20.05 just after inhalation and 0.45 at the end of the study.
Assessment of a possible order effect on the measured pain-
related endpoints did not show a significant effect (P.0.05),
indicating that starting with placebo or with active treatment had
no significant effect on outcome.
In Figure 5A, the plasma THC concentration vs Dpressure pain
for the 3 active cannabis varieties is plotted showing loops with
counterclockwise direction. Using a nonparametric collapsing
approach, we closed the loops to give the relationship between
the estimated THC effect-site (or steady-state) concentration and
Dpressure pain (Fig. 5B)
(Using ke0obj, written and kindly
provided by Dr. S.L. Shafer [Stanford University, Palo Alto, CA]).
The effect of Bedrocan (blue dots) is derived from just THC
(reference drug). The effect of Bediol (red dots) is lower than
expected from its steady-state THC concentration range,
indicative of an antagonist effect of CBD (when combined with
THC) on the pressure pain response. By contrast, when CBD is
administered without relevant THC content (Bedrolite, green
dots), a small THC-independent analgesic effect is apparent.
The main findings of this experimental study in chronic pain
patients with FM are that:
(1) none of the treatments had an effect greater than placebo on
spontaneous pain scores; (2) compared to placebo responder
rates, significantly more patients responded to Bediol (contain-
ing high doses of THC and CBD) with a decrease in
spontaneous pain by 30%; the 2 other active treatments had
response profiles not different from placebo; (3) the reduction in
spontaneous pain scores was correlated with the magnitude of
drug high; (4) pressure pain threshold increased significantly in
patients treated with Bedrocan and Bediol, 2 cannabis varieties
with a high THC content; (5) Bedrolite, a cannabis variety with
a high CBD content was devoid of analgesic activity in any of
the spontaneous or evoked pain models; and (6) CBD
increased plasma concentrations of THC but had an antago-
nistic effect on analgesia when combined with THC.
Major strengths of our study are the measurement of plasma
concentrations of the inhaled cannabinoids enabling the corre-
lation of plasma concentration rather than dose to effect, the use
of a placebo cannabis variety exempt from THC and CBD but with
the original terpene profile of the Bedrocan variety, and the testing
of well-defined cannabis varieties in a group of patients with
a well-defined chronic pain condition. Limitations of the study are
the short treatment period and lack of validation of the
experimental measures in FM.
Figure 2. Plasma concentrations of Δ
-tetrahydrocannabinol (THC), its metabolite 11-hydroxy-THC (11-OH-THC), and cannabidiol (CBD) after inhalation of 3
cannabis varieties, Bedrocan (A), Bediol (B), and Bedrolite (C). Data are mean 695% confidence interval.
Figure 3. Effect of cannabis varieties Bedrocan, Bediol, Bedrolite, and placebo cannabis on spontaneous pain scores (A), pressure pain threshold (B), and drug
high (C). Data are mean 6SEM and are relative to baseline. NRS, numerical rating score; VAS, visual analogue scale.
864 T. van de Donk et al.·160 (2019) 860–869 PAIN
Effect of treatment on experimental pain and spontaneous pain responses.
Treatment Measurement time since the start of cannabis inhalation (minutes)
10 20 30 40 60 90 120 150 180 Pvs placebo
Pressure pain (Dkgf) Bedrocan 4.28 68.6 7.18 68.7 9.64 69.7 8.83 69.9 9.46 611.8 6.81 69.9 5.88 610.4 2.24 69.0 3.83 68.3 0.006
Bediol 7.72 67.1 10.67 612.0 11.23 613.2 10.13 611.3 9.83 612.7 9.14 69.9 5.88 69.2 6.24 68.6 4.24 610.6 <0.001
Bedrolite* 2.69 65.6 5.27 67.4 4.64 67.5 5.42 68.5 4.11 69.5 2.49 68.1 1.99 68.6 2.01 69.5 3.24 69.7 0.095
Placebo 1.68 63.5 2.04 64.6 2.56 66.5 1.49 67.3 0.36 66.7 20.87 66.8 21.12 67.6 21.96 68.8 22.67 67.6
Electrical pain threshold (DmA) Bedrocan 20.15 62.6 0.55 62.7 0.39 62.5 1.3 62.7 0.9 62.9 0.38 62.9 20.1 62.8 20.43 63.5 20.35 63.6 0.383
Bediol 0.18 61.8 20.31 64.8 0.32 63.0 1.02 63.4 1.09 62.6 1.94 64.5 0.52 62.7 1.19 63.5 0.82 62.5 0.825
Bedrolite 0.92 62.7 0.99 62.8 1.42 63.3 1.32 63.1 0.69 62.6 0.47 63.31 0.47 63.0 0.27 64.2 0.72 63.5 0.954
Placebo 20.2 62.2 0.58 61.8 1.1 61.8 1.63 62.8 1.13 62.9 1.1 62.6 1.15 62.5 0.8 62.4 0.98 63.3
Electrical pain tolerance (DmA) Bedrocan 20.95 63.8 0.58 63.3 0.33 63.3 0.83 63.7 20.1 63.3 20.4 63.6 20.77 64.2 21.17 64.7 20.92 65.0 0.809
Bediol 20.71 (3.2) 20.44 (3.0) 20.6 (4.2) 20.48 (3.5) 20.68 (3.3) 21.13 (3.4) 22.18 (4.1) 21.65 (4.0) 22.13 (4.3) 0.581
Bedrolite 20.41 (3.2) 20.68 (2.5) 0.14 (3.1) 20.18 (4.3) 20.48 (3.7) 20.43 (3.2) 21.11 (4.0) 20.68 (4.2) 20.88 (4.2) 0.900
Placebo 20.37 (2.3) 21.07 (2.5) 20.2 (2.6) 20.05 (2.4) 0.03 (3.5) 20.9 (2.43) 20.72 (2.1) 21.47 (1.9) 21.12 (2.6)
Spontaneous pain score (Dcm) Bedrocan 21.56 61.6 21.81 62.1 21.50 61.6 0.678
Bediol 22.06 61.6 22.00 61.9 21.59 61.4 0.293
Bedrolite* 21.29 61.4 21.12 61.2 21.53 61.6 0.351
Placebo 21.35 61.1 21.41 61.3 21.53 61.6
Data are relative to baseline values.
Values are mean 6SD.
Bold values represent significant
50.04 vs Bediol.
April 2019·Volume 160 ·Number 4 www.painjournalonline.com 865
Over the past years, cannabis has become increasingly
popular for medical use. Currently, an increasing number of
countries legalized or are planning to legalize cannabis for
medicinal purposes. For instance, in the Netherlands, stan-
dardized cannabis has been available in pharmacies on
prescription since 2003. However, cannabinoids typically have
modest effects with small effect sizes and numbers needed to
In addition, the effect of cannabinoids in relieving
chronic pain seems to diminish over time.
Still, many patients
report using cannabis for the treatment of chronic pain with
We performed a small experimental study to
explore the acute analgesic effects on experimental measures of
3 cannabis varieties that ranged in THC and CBD content after
a single inhalation.
Our experimental study was not designed to provide direct
evidence for the clinical use of cannabis in FM but may be used
to design future clinical trials. In addition, our approach allows
to link the observed effect with THC and CBD plasma
concentrations and to detect possible pharmacokinetic and/
or pharmacodynamic interactions. Here, we discuss the
performance and outcome of the study with focus on the use
of placebo cannabis, pharmacokinetics, potential analgesic
efficacy of THC and CBD, and adverse effects.
4.1. Placebo cannabis
We used a placebo cannabis variety (ie, a cannabis plant devoid
of THC or CBD but with the full terpene profile) as a comparator
to ensure blinding of treatment. Cannabis placebo varieties
without cannabinoids have been used before.
plant material had a similar smell and appearance as the other
cannabis varieties. The importance of successful blinding in
clinical trials on cannabis analgesia has recently been
Although our approach theoretically allows for
blinding of treatment during inhalation, we cannot exclude that
lack of psychoactive symptoms from placebo inhalation during
the course of the study had some influence on the outcome in
some of the pain models. This is especially relevant given the
study crossover design. Indeed, at the end of the study, 13/20
(65%) patients guessed correctly that they had received
Effect of treatment on subjective feelings derived from the Bowdle questionnaire and Bond and Lader questionnaire.
Treatment Measurement time since the start of cannabis inhalation (minutes)
0 30 60 90 120 150 180 Pvs placebo
Bowdle drug high (mm) Bedrocan 0 43 637 34 630 27 627 19 626 10 615 4 66<0.001
Bediol 0 45 637 45 633 40 634 24 627 14 621 5 69<0.001
Bedrolite* 0.1 60.2 13 621 9 616 6 611 4 68264164 0.240
Placebo 0.2 60.5 0 613680610612061060
Bowdle internal perception (mm) Bedrocan 1 63467368266164162060 0.272
Bediol 0 616697619 7 617 7 617 3 681640.009
Bedrolite† 0 61061061061060061061 0.903
Placebo 0 61060060060060060061
Bowdle external perception (mm) Bedrocan 1 6618624 15 619 10 617 7 614 4 619 2 640.001
Bediol 1 6318618 16 618 13 620 9 617 6 613 2 64<0.001
Bedrolite* 1 62262162262162161061 0.765
Placebo 1 64162061061061060060
Bond and Lader calmness (mm) Bedrocan 18 619 9 613 8 612 9 614 7 698612 12 615 0.289
Bediol 24 621 13 623 14 625 14 626 14 622 16 621 13 619 0.562
Bedrolite 18 619 7 696698611 8 610 8 611 11 614 0.226
Placebo 18 622 12 617 14 619 13 619 14 621 12 619 14 629
Bond and Lader mood (mm) Bedrocan 8 612 12 613 13 6912691269968767 0.875
Bediol‡ 12 613 20 622 18 622 18 624 17 622 15 621 12 616 0.020
Bedrolite 11 612 8 67968967868868868 0.996
Placebo 11 611 9 6910698
Bond and Lader alertness (mm) Bedrocan 14 617 45 627 47 622 45 621 40 624 30 621 23 620 0.026
Bediol 19 618 46 626 47 623 43 623 40 623 34 626 25 622 0.010
Bedrolite* 16 617 25 618 27 616 25 618 25 619 23 619 19 617 0.399
Placebo 16 617 16 614 16 614 17 614 19 618 17 618 15 616
Values are mean 6SD.
Bold values represent significant
,0.01 vs Bedrocan and Bediol.
,0.05 vs Bediol.
,0.05 vs Bedrocan and Bedrolite.
Figure 4. Cannabis responder rates: (A) Percentage responders with
a decrease of at least 30% in spontaneous pain scores on at least one
measurement. (B) Percentage responders with a decrease of at least 50% in
spontaneous pain scores on at least one measurement.
866 T. van de Donk et al.·160 (2019) 860–869 PAIN
placebo treatment. On the other hand, the terpenes present in
the placebo plant may have exerted some effects. Terpenes are
assumed to interact with cannabinoids (entourage effect),
improving their pharmacodynamic effects (eg, by increasing
pulmonary uptake and change binding of cannabinoids to their
receptors), but also have effects of their own, including anti-
inflammatory, antidepressant, and analgesic effects.
then suggests that the placebo cannabis variety used in our
study is best considered an active placebo. Hence, the
observation of an appreciable placebo effect in the relief of
spontaneous pain is not surprising.
The pharmacokinetic analysis showed that peak THC con-
centrations were similar after Bedrocan and Bediol inhalation,
whereas the peak THC concentration after Bedrolite inhalation
was about one-sixth of that of the other 2 varieties (Fig. 2).
These are important observations and indicate that magnitude
of THC plasma concentrations was partly dependent on the
presence of CBD in the inhalant. In Bedrocan, 24-mg inhaled
THC (and ,1-mg CBD) produced a mean THC peak plasma
concentration of 82 ng/mL. In the other 2 cannabis varieties
with CBD contents of about 18 mg, THC plasma concen-
trations were at least 50% higher than expected from the
Bedrocan pharmacokinetic data. The THC–CBD pharmaco-
kinetic interaction may be explained by (1) a possible CBD-
induced increase in pulmonary THC uptake, for example, due
to an increase in pulmonary blood flow. We are unaware of any
data that support this mechanism; (2) CBD-induced inhibition
of THC metabolism. Although CBD potently inhibits THC
metabolism in the rat,
our data do not support any inhibition
of THC conversion to 11-OH-THC (Fig. 2); and (3) cyclizing of
CBD into THC. Because both compounds are chemically
related, CBD can convert into THC; this has been observed
after subcutaneous administration of CBD in the rat.
further improve our understanding of the pharmacokinetic
behavior of THC under different CBD conditions, we plan
a compartmental pharmacokinetic analysis of our data.
4.3. Outcome of the acute experimental pain tests
Two cannabis varieties, Bedrocan and Bediol, were analgesic in
the pressure pain model but had no effect in the electrical pain
model or on relief of spontaneous pain. The pressure pain test
seems especially suited for exploring treatment effects in FM pain,
as it elicits mechanical muscle stimulation through Ad- and C-
fiber activation and better reflects the symptoms of patients with
FM than electrical pain, which produces direct sensory nerve
We previously used electrical noxious stimulation
as a model of acute pain and showed high sensitivity of opioids in
alleviating transcutaneous electrical pain.
The current data
suggest that cannabis may have limited use in acute pain
Interestingly, when CBD and THC were combined (in Bediol),
CBD had antagonistic pharmacodynamic effects (Fig. 5B),
possibly because of an antagonistic or negative modulatory
action at the CB
Despite this pharmacodynamic
antagonism, the analgesic responses exceeded those of
Bedrocan, possibly because of the CBD-induced increase in
THC concentrations. The opposed direction of the pharmaco-
kinetic and pharmacodynamic CBD–THC interactions is an
indication of the complex pharmacological behavior of canna-
binoids in humans. When CBD is given without relevant THC
content (ie, Bedrolite, containing predominantly CBD), just
small analgesic effects not different from placebo became
apparent. This is somewhat surprising, as it is our experience
and that of others that patients with chronic pain report
beneficial effects from CBD treatment.
Possibly, such effects
are related to improvement of insomnia, anxiety, cognition,
and/or mood. In addition, it may well be that a single CBD
administration may be insufficienttoelicitanalgesicresponses,
or that the dose was too low.
4.4. Side effects
Some side effects of active treatment were observed. One-third
of patients reported sore throat and bad taste, whereas two-
thirds coughed during the 5- to 7-minute inhalation of the active
treatments. In the course of the study, one-third of patients
experienced nausea without vomiting. All symptoms were rated
Figure 5. (A) Plasma THC concentration (C
) vs the change in pressure pain threshold after treatment with Bedrocan (blue dots), Bediol (red dots), and Bedrolite
(green dots). The arrows indicate the direction of effect, starting at the large yellow circle. (B) Estimated steady-state or effect-site (C
) concentration vs the change
in pressure pain threshold for the 3 active cannabis varieties. THC, tetrahydrocannabinol.
April 2019·Volume 160 ·Number 4 www.painjournalonline.com 867
as mild. An important observation was that most patients disliked
the feeling of drug high after inhalation, although the intensity was
rated as moderate (Fig. 3C). Because this is a general
observation in chronic pain patients treated with psychedelic
medication, we recently studied the ability to temper the feeling of
drug high induced by racemic ketamine. We observed that drug
high intensity was reduced by 30% during administration of the
nitric oxide donor sodium nitroprusside.
Because cannabis and
ketamine produce their psychotropic effects through separate
pathways (N-methyl-D-aspartate receptor antagonism vs CB
receptor agonism), further studies are needed to discover viable
options to reduce THC-related drug high without reducing
analgesia. Still, it may be that this may have a negative effect on
analgesic efficacy because we observed that relief of spontane-
ous pain was correlated with drug high scores. This suggests that
some level of intoxication is required for an analgesic cannabis
effect, or that the lack of complete blinding due to the occurrence
of psychotropic side effects (or symptoms during inhalation)
influenced pain scoring to some extent.
In conclusion, in this experimental and highly controlled
study, we explored the pharmacokinetics and pharmacody-
namics of 3 active cannabis varieties in chronic pain patients
with FM. The most important observation is that when
simultaneously inhaled, THC and CBD interact in complex
fashions with synergistic pharmacokinetic but antagonistic
pharmacodynamic interactions. The analgesic efficacy of
active treatment was limited to varieties that contained THC
and was observed exclusively in the evoked pressure pain
model. None of the active treatments were effective in reducing
spontaneous pain scores more than placebo. Further studies
are needed to assess efficacy and safety (including addictive
behavior) in clinical trials with prolonged treatment periods and
explore the role of psychotropic effects in the development of
Conflict of interest statement
This investigator-initiated trial was performed in collaboration with
Bedrocan International BV (Veendam, the Netherlands). Bed-
rocan International BV was responsible for the production and
delivery of the cannabis products and the Volcano device for
cannabis inhalation. M.A. Kowal is an employee of Bedrocan
International BV, the Netherlands. He commented on the
protocol and final version of the paper. The other authors have
no conflict of interest to declare.
Appendix A. Supplemental digital content
Supplemental digital content associated with this article can be
found online at http://links.lww.com/PAIN/A705.
Received 15 August 2018
Received in revised form 29 November 2018
Accepted 6 December 2018
Available online 20 December 2018
 Bang H, Ni L, Davis C. Assessment of blinding in clinical trials. Control Clin
 Bisogno T, Hanu ˇs L, De Petrocellis L, Tchilibon S, Ponde D, Brandi I,
Moriello AS, Davis JB, Mechoulam R, Di Marzo V. Molecular targets for
cannabidiol and its synthetic analogues: effect on vanilloid VR1 receptors
and on the cellular uptake and enzymatic hydrolysis of anandamide. Br J
 Bond A, Lader M. The use of visual analogue scales in rating of subjective
feelings. Br J Med Psychol 1974;47:211–8.
 Bowdle TA, Radant AD, Cowley DS, Kharasch ED, Strassman RJ, Roy-
Byrne PP. Psychedelic effects of ketamine in healthy volunteers.
 Calandre EP,Rico-Villadermoros F,Slim M. An updateon pharmacotherapy
for the treatment of fibromyalgia. Expert Opin Pharmacother 2015;16:
 Campos AC, Fogac¸ a MV, Scarante FF, Joca SRL, Sales AJ, Gomes FP,
Sonego AB, Rodrigues NS, Galve-Roperh I, Guimara
˜es FS. Plastic and
neuroprotective mechanisms involved in the therapeutic effects of
cannabidiol in psychiatric disorders. Front Pharmacol 2017;8:269.
 Campos A, Guimara
˜es F. Involvement of 5HT1A receptors in the anxiolytic-
like effects of cannabidiol injected into the dorsolateral periaqueductal gray
of rats. Psychopharmacology (Berl) 2008;199:223–30.
 Clauw DJ. Fibromyalgia: a clinical review. JAMA 2014;311:1547–55.
 ElSohly M, Gul W. Constituents of Cannabis sativa. In: Pertwee R, editor.
Handbook of Cannabis. Oxford: Oxford University Press, 2014.
 Fiz J, Dur ´an M, Capell `a D, Carbonell J, Farr ´e M. Cannabis use in patients
with fibromyalgia: effect on symptoms relief and health-related quality of
life. PLoS One 2011;6:e18440.
 Gomes F, Resstel L, Guimara
˜es F. The anxiolytic-like effects of
cannabidiol injected into the bed nucleus of the stria terminalis are
mediated by 5-HT1A receptors. Psychopharmacology (Berl) 2011;213:
˜es AG, Serafini MR, Quintans-j ´unior LJ. Terpenes and derivatives
as a new perspective for pain treatment: a patent review. Expert Opin
Ther Pat 2014;24:243–65.
 Hazekamp A, Ruhaak R, Zuurman L, van Gerven J, Verpoorte R.
Evaluation of a vaporizing device (Volcano) for the pulmonary
administration of tetrahydrocannabinol. J Pharm Sci 2006;95:
 Hlozek T, Uttl L, Kadarabek L, Balikova M, Lhotkova E, Horsley RR,
Novakova P, Sichova K, Stefkova K, Tyls F, Kuchar M, Palenicek T.
Pharmacokinetic and behavioral profile of THC, CBD, and THC 1CBD
combination after pulmonary, oral, and subcutaneous administration in
rats and conformation of conversion in vivo of CBD to THC. Eur
 Jonkman K, van der Schrier R, van Velzen M, Aarts L, Olofsen E, Sarton E,
Niesters M, Dahan A. Differential role of nitric oxide in the psychedelic
symptoms induced by racemic ketamine and esketamine in human
volunteers. Br J Anaesth 2018;120:1009–18.
 Kowal M, Hazekamp A, Colzato L, van Steenbergen H, Hommel B.
Modulation of cognitive and emotional processing by cannabidiol: the role
of the anterior cingulate cortex. Front Hum Neurosc 2013;7:147.
 Krasselt M, Baerwald C. Fibromyalgie-Syndrom: aktuelle
empfehlungen zu diagnsostik und Therapie. Dtsch Med Wochenschr
 Martini CH, Proto P, Olofsen E, van Velzen M, Aarts L, Dahan A, Niesters
M. A randomized controlled trial and novel mathematical analysis of the
analgesic effect of oxycodone versus paracetamol orodispersible tablets.
Eur J Pain 2015;19:295–304.
 Olesen AE, Andresen T, Staahl C, Drewes AD. Human experimental pain
models for assessing the therapeutic efficacy of analgesic drugs.
Pharmacol Rev 2012;64:772–9.
 Olofsen E, Romberg R, Bijl H, Mooren R, Engbers F, Kest B, Dahan A.
Alfentanil and placebo analgesia: absence of sex differences.
 Pertwee R, Cascio M. Known pharmacological actions of delta-9-
tetrahydrocannabinol and of four other chemical constituents of cannabis
that activate cannabinoid receptors. In: Pertwee R, editor. Handbook of
Cannabis. Oxford: Oxford University Press, 2014.
 Ruso EB. Taming THC. Potential cannabis synergy and
phytocannabinoid-terpenoid entourage effects. Br J Phramacol 2011;
 Sexton M, Cuttler C, Finnell J, Mischley L. A cross-sectional survey of
medical cannabis users: patterns of use and perceived efficacy.
Cannabis Cannabinoid Res 2016;1:131–8.
 Stockings E, Campbell G, Hall WD, Nielsen S, Zagic D, Rahman R,
Murnion B, Farrell M, Weier M, Degenhardt L. Cannabis and
cannabinoids for the treatment of chronic non-cancer pain conditions:
a systematic review and meta-analysis of controlled and observational
studies. PAIN 2018;159:1932–44.
 Wal lace MS, Marcotte TD , Umlauf A, Gouaux B, Atki nson JH. Efficacy of
inhaled cannabis on painful diabetic neuropathy. J Pain 2015;16:
868 T. van de Donk et al.·160 (2019) 860–869 PAIN
 Wilsey B, Deutsch R, Marcotte TD. Maintenance of blinding in clinical
trials and the implications for studying analgesia using cannabinoids.
Cannabis Cannabinoid Res 2016;1:139–48.
 Wilsey B, Marcotte TD, Deutsch R, Zhao H, Prasad H, Phan A. An
exploratory human laboratory experiment evaluating vaporized cannabis
is in the treatment of neuropathic pain from spinal cord injury and disease.
J Pain 2016;17:982–1000.
 Wolfe F, Cla uwD J, Fitzcharles MA, Goldenberg GL, Katx RS, Mease P, Russell
AS, Russell IJ, Winfield JB, Yunus MB. The American College of Rheumatology
preliminary diagnostic criteria for fibromyalgia and measurement of symptom
severity. Arthritis Care Res (Hoboken) 2010;62:600–10.
 Zanelati T, Biojone C, Moreira F, Guimara
˜es F, Joca S. Antidepressant-
like effects of cannabidiol in mice: possible involvement of 5-HT1A
receptors. Br J Pharmacol 2010;159:122–8.
 Zuurman L, Roy C, Schoemaker R, Hazekamp A, den Hartigh J, Bender
J, Verpoorte R, Pinquier JL, Cohen AF, van Gerven JM. Effect of
intrapulmonary tetrahydrocannabinol administration in humans.
J Psychopharmacol 2008;22:707–16.
April 2019·Volume 160 ·Number 4 www.painjournalonline.com 869