Comment
Jonathan Young, *Matthew J Kempton, Philip McGuire
Department of Neuroimaging, (JY, MJK), and Department of
Psychosis Studies (MJK, PM), Institute of Psychiatry, Psychology,
and Neuroscience, Kings College London, London SE5 8AF, UK;
National Institute for Health Research Mental Health Biomedical
Research Centre at South London and Maudsley NHS Foundation
Trust, London, UK (PM)
matthew.kempton@kcl.ac.uk
We declare no competing interests.
1 Koutsouleris N, Kahn RS, Chekroud AM, et al. Multisite prediction of
4-week and 52-week treatment outcomes in patients with rst episode
psychosis: a machine learning approach. Lancet Psychiatry 2016; published
online Aug 25. http://dx.doi.org/10.1016/S2215-0366(16)30171-7.
2 Fleischhacker WW, Keet IP, Kahn RS, Committee ES. The European First
Episode Schizophrenia Trial (EUFEST): rationale and design of the trial.
Schizophr Res 2005; 78: 14756.
3 Koutsouleris N, Meisenzahl EM, Davatzikos C, et al. Use of neuroanatomical
pattern classication to identify subjects in at-risk mental states of psychosis
and predict disease transition. Arch Gen Psychiatry 2009; 66: 70012.
Cannabis and psychosis: understanding the smoke signals
There is no cure for schizophrenia. Accordingly,
clinicians must do their best at managing the various
symptoms that emerge during the course of the
disease. Therefore, identication of putative risk factors
that trigger relapse and contribute to poor prognosis
is crucial. Because cannabis might contribute to the
development of schizophrenia,1 the continued eect of
the drug on the course of the disease is of great clinical
interest and a focus on how to improve outcomes in
this disorder is essential.
In their observational study, Tabea Schoeler and
colleagues2 investigated the eects of dierent
patterns of cannabis (eg, use, frequency, and
potency) on risk of relapse in patients with rst
episode psychosis in South London, UK. Findings
showed that cannabis eects on outcome varied
depending on the particular cannabis-using prole of
the individual. More specically, cannabis users who
had continuous, high frequency and high potency
exposure had the worst outcome, in terms of risk
of relapse (OR 328, 95% CI 122918), number of
relapses (IRR 177, 95% CI 096325), time to relapse
(b 022, 95% CI 040 to 004), and more intense
psychiatric care after the onset of psychosis compared
with former users (OR 316, 95% CI 126809).
Further, high frequency, high potency users relapsed
more quickly than did patients who consumed high
potency cannabis at a lower frequency, hash-like
www.thelancet.com/psychiatry Vol 3 October 2016
4 Begg C, Cho M, Eastwood S, et al. Improving the quality of reporting of
randomized controlled trials. The CONSORT statement. JAMA 1996;
276: 63739.
5 Munafo MR, Kempton MJ. Has analytical exibility increased in imaging
studies of bipolar disorder and major depression? Psychol Med 2015;
45: 44951.
6 Asendorpf JB, Conner M, De Fruyt F, et al. Recommendations for increasing
replicability in psychology. Eur J Personality 2013; 27: 10819.
7 Kriegeskorte N, Simmons WK, Bellgowan PS, Baker CI. Circular analysis in
systems neuroscience: the dangers of double dipping. Nat Neurosci 2009;
12: 53540.
8 Kohavi R, Sommereld D. Feature subset selection using the wrapper
method: overtting and dynamic search space topology. https://www.aaai.
org/Papers/KDD/1995/KDD95-049.pdf (accessed Aug17, 2016).
9 Chekroud AM, Zotti RJ, Shehzad Z, et al. Cross-trial prediction of treatment
outcome in depression: a machine learning approach. Lancet Psychiatry
2016; 3: 24350.
10 McGuire P, Dazzan, P. The role of neuroimaging in predicting outcomes in
psychosis. World Psychiatry (In press).
cannabis users, and patients who were never (regular)
cannabis users.
This study oers several strengths over other
studies that have assessed the association between
cannabis and outcome in patients with psychotic
disorders. First, this study was prospective, and
Science Photo Library
assessed cannabis use trajectories during the rst
2 years after the onset of psychosis. A major limitation
of previous cross-sectional studies is their inability to
dierentiate between cause and eect. This caveat Published Online
August 22, 2016
has led to the self-medication hypothesis,3 an over- http://dx.doi.org/10.1016/
popularised explanation for addictive behaviours S2215-0366(16)30230-9
that posits that cannabis is used by patients to help See Articles page 947
mitigate the negative eects of the disorder as well as
side eects associated with antipsychotic treatment.
Inconsistent with this idea, ndings from Schoeler and
colleagues showed that cannabis might contribute
to symptomatology rather than alleviating it. This
nding underscores the importance of employing the
appropriate temporal and longitudinal paradigms
when assessing the eects of cannabis on outcome in
patients across the schizophrenia spectrum.
Cannabis is a dicult substance to quantify.
Accordingly, the researchers aimed to assess cannabis
exposure not only as a measure of continuity and
frequency, but also with respect to potency. Cannabis
is a very complex plant that can be bred to yield
hundreds of strains, each with a unique combination
909
 Comment
of cannabinoids. However, in most studies cannabis
is conceptualised as a uniform drug. Cannabis varies
substantially in the level of its two major cannabinoids:
delta-9-tetrahydrocannabinol (THC) and cannabidiol
(CBD). Levels of CBD can range from almost zero to
up to 40%.4 Notably, these two cannabinoids have
opposing properties and thus can produce divergent
eects. For example, THC is often associated with
deleterious eects such as impaired cognition, anxiety,
and psychotic-like symptoms, whereas CBD has been
associated with antipsychotic and anxiolytic-like
eects, and might even oer neuroprotection.57 In
view of these dierences, the researchers aimed to
decipher the type of cannabis their study sample
consumed by asking participants to describe their
preferred type of cannabis: skunk-like (high THC) or
hash-like (high CBD).
The investigators should be commended for
confronting this issue; however, the manner in which
it was addressed raises concern. Can patients with
rst episode psychosis accurately identify the type
of cannabis they are using? Objective quantitative
analysis might more reliably capture cannabis
potency versus qualitative methods. Additionally,
the assessment of potency as a binary variable might
be too simplistic for such a complex plant in view of
the fact that eects of cannabis depend not only on
THC potency, but also on the ratio of THC to CBD.8
The importance of studying the type of cannabis of
users has recently been highlighted in the scientic
literature.9 As cannabis research progresses, hopefully
investigators will begin to regard cannabis as a
conglomerate of cannabinoids rather than one
homogeneous substance.
Findings presented by Schoeler and colleagues2
support the idea that clinicians should actively
intervene and help patients to quit cannabis and
remain abstinent. Nevertheless, the investigators
suggest a possible harm reduction model in which
interventions could focus on persuading cannabis
users to reduce use or shift to less potent forms of
cannabis, especially patients with psychosis who are
otherwise unable to stop using cannabis. However, the
proposal of such a framework might be problematic.
In the same respect that not all cannabis is the same,
not all users are the same. That is, skunk users might
dier from hash users. Additionally, if high potency
910
users switch to low potency cannabis, conceivably,
they might use more cannabis in an attempt to chase
the familiar high. Better characterisation of these
dierent users is warranted to develop appropriate
and tailored interventions.
The investigators results highlight the association
between cannabis and subsequent relapse in patients
with rst episode psychosis, especially in those who are
frequently exposed to high potency preparations. As a
result of genetic modication and advanced methods
of cultivation, high THC cannabis is becoming
the norm, therefore addressing the increased risks
implicated with these strains of cannabis should be of
high priority.
*Rachel A Rabin, Tony P George
The Institute of Medical Sciences, University of Toronto, Toronto,ON,
Canada (RAR); Schizophrenia Division, Centre for Addiction and
Mental Health (CAMH), Toronto, Ontario Canada (RAR, TPG);
Division of Brain and Therapeutics, Department of Psychiatry,
University of Toronto, Toronto, Ontario Canada (TPG); Icahn School
of Medicine at Mount Sinai, New York, NY 10070, USA (RAR)
rachel.rabin@camh.ca
In the past 12 months, TPG has received industry sponsored grants from Pzer,
and served as a member of a Data Monitoring Committee for Novartis. RAR
declares no competing interests. RAR and TPG both contributed to the writing
of the Comment.
Copyright The Author(s). Published by Elsevier Ltd. This is an Open Access
article under the CC BY NC-ND license.
1 Moore TH, Zammit S, Lingford-Hughes A, et al. Cannabis use and risk of
psychotic or aective mental health outcomes: a systematic review. Lancet
2007; 370: 31928.
2 Schoeler T, Petros N, Di Forti M, et al. Eects of continuation, frequency, and
type of cannabis use on relapse in the rst 2 years after onset of psychosis:
an observational study. Lancet Psych 2016; published online Aug 22.
http://dx.doi.org/10.1016/S2215-0366(16)30230-9.
3 Khantzian EJ. The self-medication hypothesis of addictive disorders:
focus on heroin and cocaine dependence. Am J Psychiatry 1985; 142: 125964.
4 Hardwick S, King LA. Home Oce cannabis potency study 2008.
http://www.dldocs.stir.ac.uk/documents/potency.pdf (accessed
Aug 18, 2016).
5 Morrison PD, Zois V, McKeown DA, et al. The acute eects of synthetic
intravenous Delta9-tetrahydrocannabinol on psychosis, mood and
cognitive functioning. Psychol Med 2009; 39: 160716.
6 Zuardi AW, Crippa JA, Hallak JE, Moreira FA, Guimaraes FS.
Cannabidiol, a Cannabis sativa constituent, as an antipsychotic drug.
Braz J Med Biol Res 2006; 39: 42129.
7 McPartland JM, Duncan M, Di Marzo V, Pertwee RG. Are cannabidiol
and Delta(9)-tetrahydrocannabivarin negative modulators of the
endocannabinoid system? A systematic review. Br J Pharmacol 2015;
172: 73753.
8 Zuardi AW, Hallak JE, Crippa JA. Interaction between cannabidiol (CBD) and
(9)-tetrahydrocannabinol (THC): inuence of administration interval and
dose ratio between the cannabinoids. Psychopharmacology 2012;
219: 24749.
9 Hagerty SL, Williams SL, Mittal VA, Hutchison KE. The cannabis conundrum:
Thinking outside the THC box. J Clin Pharmacol 2015; 55: 83941.
www.thelancet.com/psychiatry Vol 3 October 2016