Association between formal thought disorder and

cannabis use: a systematic review and meta-analysis

Mathilde Argote, Guillaume Sescousse, Jérôme Brunelin, Eric Fakra, Mikail
Nourredine, Benjamin Rolland

To cite this version:

Mathilde Argote, Guillaume Sescousse, Jérôme Brunelin, Eric Fakra, Mikail Nourredine, et al.. As-
sociation between formal thought disorder and cannabis use: a systematic review and meta-analysis.
npj Schizophrenia, 2022, 8 (1), pp.78. �10.1038/s41537-022-00286-0�. �hal-03959183�

HAL Id: hal-03959183

https://hal.science/hal-03959183v1
Submitted on 27 Jan 2023

HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est

archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents
entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non,
lished or not. The documents may come from émanant des établissements d’enseignement et de
teaching and research institutions in France or recherche français ou étrangers, des laboratoires
abroad, or from public or private research centers. publics ou privés.

Distributed under a Creative Commons Attribution 4.0 International License

 REVIEW ARTICLE OPEN

Association between formal thought disorder and cannabis

use: a systematic review and meta-analysis
Mathilde Argote 1,2,3 ✉, Guillaume Sescousse1,2,3, Jérôme Brunelin 1,2,3
, Eric Fakra1,4, Mikail Nourredine2,5,6 and
Benjamin Rolland1,2,3,7

Formal thought disorder (FTD) is a multidimensional syndrome mainly occurring along the psychosis continuum. Cannabis use is
known to increase symptoms of psychosis, particularly positive symptoms. However, the impact of cannabis use on FTD in
individuals presenting symptoms along the psychosis continuum remains unclear. To address this knowledge gap, we conducted a
meta-analysis examining the association between cannabis use and FTD in those individuals. We hypothesized that cannabis would
worsen FTD. We conducted a systematic search of the PubMed, ScienceDirect, PsycINFO, Web of Science, Embase and Google
Scholar databases up to July 2022. The results were collated through a random-effects model using the statistical software R.
Reference lists of included studies were searched for additional relevant publications. Nineteen studies were included, totalling
1840 cannabis users and 3351 non-cannabis users. The severity of FTD was found to be higher in cannabis users (SMD = 0.21, 95%
CI [0.12–0.29], p = 0.00009). Subgroup analyses revealed that FTD severity was increased among cannabis users, regardless of the
disorder severity: healthy individuals (SMD = 0.19, 95%CI [0.05–0.33], p = 0.02); patients with first-episode psychosis (SMD = 0.21,
95%CI [0.01–0.41], p = 0.04); patients with schizophrenia (SMD = 0.25, 95%CI [0.11–0.38], p = 0.005). Between-group differences
1234567890():,;

were not significant. In line with its already known effect on positive symptoms in psychosis, cannabis use appears to be associated
with increased FTD severity all along the psychosis continuum. Future research should consider potential confounding variables
such as other substance use disorders and explore how FTD dimensions are impacted by cannabis use.
Schizophrenia (2022)8:78 ; https://doi.org/10.1038/s41537-022-00286-0

INTRODUCTION such as hallucinations and delusions15, whereas negative

Formal thought disorder (FTD) refers to the disruption of the symptoms such as avolition and anhedonia seem less impacted
process of thoughts and language production, clinically resulting by cannabis use16. FTD has a special place in the semiology of
in an alteration of effective communication1,2 and is regarded as a schizophrenia spectrum disorders, as it has been suggested to
multidimensional syndrome3. FTD prevalence and severity are constitute a core (psychopathological) dimension of the
proportionally higher in patients with psychotic disorders than in disorder17. FTD dimensional structure is currently debated in
healthy populations2. 6% of people in the general population and the literature, leading to a high heterogeneity of assessment
up to 80% of patients with psychotic disorders exhibit more or less scales and conceptual definitions of this syndrome. At least two
pronounced symptoms of FTD, suggesting that FTD is best dimensions emerge as a consensus: negative FTD, referring to a
conceptualized as a continuous dimension2. FTD is particularly deficit in speech and thought production; and positive FTD,
frequent in schizophrenia, as it is found in 27–80% of patients1, referring to an increased amount of produced speech. Positive
but it may also occur in other types of disorders such as FTD is consistently associated with the disorganization dimen-
depression and mania1,4. FTD encompasses particularly disabling sion in several models of schizophrenia symptoms1,7,18. Previous
symptoms that increase social isolation5, reduce quality of life6
research in individuals with schizotypal traits reported that the
and are associated with poorer clinical outcomes7. This disorder is
severity of disorganization was higher in cannabis users than in
predictive of conversion to psychosis in high-risk populations8,
and is the strongest predictor of conversion from first-episode non-users19, suggesting that cannabis worsens positive FTD in a
psychosis to schizophrenia9. non-clinical sample.
Cannabis use is one of the most prominent risk factors The current review and meta-analysis firstly aimed to
associated with the occurrence and severity of many psychotic determine whether FTD and cannabis use are associated at
disorders, including schizophrenia. In patients with schizophre- different stages of the psychosis continuum, i.e. no or attenuated
nia, lifetime cannabis use is estimated at 42.110 and 26.2% of symptoms (stages 0, 1a, 1b), first-episode psychosis (stage 2), and
patients with schizophrenia suffer from cannabis use disorder11. recurrence and treatment resistance of the disorder (stages 3,
Cannabis use doubles the risk of developing a psychotic 4)20. Subsequently, it aimed to investigate whether cannabis use
disorder in vulnerable individuals12, and triggers the onset of would have a distinct impact on FTD depending on the stage of
first-episode psychosis13 and psychotic disorders two to three the psychotic disorder. We expected to find a positive association
years earlier than in non-users14. Among patients with schizo- between cannabis use and FTD severity, all along the psychosis
phrenia, cannabis users present more severe positive symptoms, continuum.

1
PSYR2, CNRL, INSERM U1028, CNRS UMR5292, UCBL1 Bron, France. 2Université Claude Bernard Lyon 1, Lyon, France. 3Centre Hospitalier Le Vinatier, Bron, France. 4Pôle
Universitaire de Psychiatrie, CHU Saint-Etienne, Saint-Etienne, France. 5Service de biostatistique, Hospices Civils de Lyon, Lyon, France. 6Service hospitalo-universitaire de
pharmacotoxicologie, Hospices Civils de Lyon, Lyon, France. 7Service Universitaire d’Addictologie de Lyon (SUAL), HCL, CH Le Vinatier, Lyon, France.
✉email: [email protected]

Published in partnership with the Schizophrenia International Research Society

 M. Argote et al.
2
Records identified from
database searching

Identification
PubMed (n = 777) Records identified via
ScienceDirect (n = 653) the references of already
Web of Science (n = 21) selected articles (n = 4)
PsycINFO (n = 26)
Embase (n = 145)
Google Scholar (n = 200)

Duplicates removed (n =
Screening
Records screened
209)
(n = 1822)
Records excluded on title
and abstract
(n = 1416)

Records excluded:
FTD not measured (n = 99) Records assessed for
Association between FTD eligibility
and cannabis not reported (n = 201)
(n = 43)
Eligibility

Absence of FTD scores for

non-users group (n = 11)
Cannabis mixed with other
substances (n = 20)
Full text not in English (n =
5)
Observational data without
quantitative measure (n = 2)
1234567890():,;

Unable to retrieve data from Eligible studies after the

corresponding author systematic review
Included

(n = 2) (n = 21)

Studies included in
the meta-analysis
(n = 19)

Fig. 1 Prisma flow diagram describing the selection process of included publications.

METHODS Eligibility criteria

The reporting of this meta-analysis was guided by the standards of To be included, each study had to meet the following criteria: (1)
the updated Preferred Reporting Items for Systematic Reviews and report FTD scores of two distinct groups: one of cannabis users
Meta-Analyses guidelines (PRISMA)21. No pre-registration was and a control one of non-users; (2) include individuals presenting
performed for this meta-analysis. symptoms pertaining to the psychosis continuum; (3) use
validated tools for assessing FTD; and (4) be written in English.
Search strategy Studies were excluded if: (1) FTD was not specifically distinguished
from other psychotic symptoms, (2) cannabis use was not
PubMed, ScienceDirect, Web of Science, PsycINFO, Embase and
assessed, or not distinguished from the use of other substances,
Google Scholar were used as sources to retrieve publications. The
(3) the link between FTD and cannabis use was not reported in the
search syntax was as follows: [“thought disorder” OR “thought
analyses.
disturbance” OR “cognitive disorganization” OR “formal thought
disorder” OR “disorganized speech” AND “cannabis” OR “mar-
ijuana” OR “THC” OR “pot” OR “hashish” OR “bhang” OR “ganja”] Data extraction
(see supplementary materials “SearchSyntax” for details). A filter Data used for statistical analyses were extracted from the Results
was applied on ScienceDirect to only retrieve “research articles”. In section or the supplementary material part of selected studies. To
addition, only the first 200 relevant sources were retrieved from allow for a comparison between groups, data had to be extracted
Google Scholar, as recommended22. Each database supports into two groups: one group of participants using cannabis and
spelling variations, such as the terms “disorganization” and one of participants not using cannabis. In addition, both extracted
“disorganized” were automatically searched with British and groups of participants had to present symptoms pertaining to the
American spelling. Each source was last searched in July 2022. same stage of the psychosis continuum (i.e. patients with
Reference lists of the selected publications were reviewed to schizophrenia and not using cannabis compared to patients with
identify additional relevant studies. schizophrenia using cannabis). In cases where study designs
Duplicates were removed by manual screening. After duplicate included participants pertaining to multiple groups in terms of
removal, each record was first screened on title and abstract by their drug use (i.e. polysubstance users, cannabis mainly users,
two independent researchers (MA and GS). In case of disagree- alcohol mainly users, non-users), only the relevant groups were
ment, the inclusion of the study was decided by a third researcher extracted (i.e. cannabis mainly users and non-users). Similarly, in
(BR). A second selection round was based on full text reading and the event that multiple groups of participants were formed in
followed the same procedure. The flow-chart detailing the terms of the disorder severity (i.e. first-episode psychosis
selection process is provided in Fig. 1. compared to healthy controls), only data of the relevant group

Schizophrenia (2022) 78 Published in partnership with the Schizophrenia International Research Society
 M. Argote et al.
3
was extracted (i.e. first-episode psychosis using and not using moderate, and high heterogeneity, respectively35. Tau² was
cannabis). estimated using the Restricted Maximum Likelihood. In addition,
In cases where insufficient data was present for the calculation Hartung-Knapp adjustment36 was applied to the 95% confidence
of our primary effect size, the principal investigators were interval of the pooled effect size to reduce the chances of false
contacted via email13,23–25. Two of the contacted authors provided positives. The prediction interval was calculated to estimate the
the requested data13,23. reliability of the results in the context of future studies. The search
for outliers was performed to address between-study hetero-
Quality assessment geneity, using the “find.outliers” function included in the {dmetar}
The risk of bias was assessed by two independent researchers (MA package. A sensitivity analysis was conducted to exclude cases
and MR) using the Quality Assessment Tool for Observational that would influence too greatly the results, due to a large sample
Cohort and Cross-sectional studies26, a recommended tool for size. This method allows to visualize if the calculated SMD was
analytical cross-sectional studies27. This process ensures that the distorted by one large study influencing the true effect size. This
quality of included studies is good enough to provide reliable supplementary analysis ensures that the synthetized results were
results. Based on a series of questions, the goal is to identify not distorted and thus robust. In cases where the influence
potential flaws in the publication that may impact the outcome analysis detected studies affecting too greatly the results, the
measure we want to study. Quality of studies can be rated as statistical leave-one-out method was applied to confirm those
“poor”, “fair”, or “good”. Question 9 “Were the exposure measures cases. This method estimates the overall effect size omitting one
(independent variables) clearly defined, valid, reliable, and study at a time. In addition, we performed post hoc subgroup
implemented consistently across all study participants?” and analyses to investigate the effect of cannabis use on the different
question 11 “Were the outcome measures (dependent variables) populations included across publications (i.e. no or attenuated
clearly defined, valid, reliable, and implemented consistently symptoms, first-episode psychosis, schizophrenia). Publication bias
across all study participants?” were judged crucial as the focus for was visually inspected with a funnel plot and quantified with an
the current meta-analysis is the relationship between FTD Egger’s test measuring the asymmetry of the funnel plot.
(outcome) and cannabis use (exposure). Question 14 “Were key
potential confounding variables measured and adjusted statisti-
cally for their impact on the relationship between exposure(s) and RESULTS
outcome(s)?” was considered important as the presence of The flow diagram of the study selection process is provided in
confounding variables can distort the relationship between FTD Fig. 1. Searches yielded 1822 results, which were then refined by
and cannabis use. In cases where studies collected a “no” to the screening of titles and abstracts. After the screening process,
questions 9 and 11, the quality was not rated as “good”. 201 publications were assessed for eligibility based on a full-text
reading, leading to the selection of 21 relevant studies. Details
Meta-analysis methods regarding the reasons for study exclusion are provided in Fig. 1.
Analyses were conducted using the statistical software R version We finally included 19 publications in the meta-analysis because
4.1.1 (R Core Team 2021)28. Mean scores, standard deviations and data extraction could not be performed for two studies24,25.
binary outcomes (FTD present or absent) of each study were Across the 19 studies, a total of 5191 participants were included:
extracted for cannabis users and for non-users; they were 1840 in the cannabis-users group and 3351 in the non-
subsequently converted into between-groups Standardized Mean users group.
Differences (SMD) using the {esc} package, so that the results
obtained from different scales could be pooled together. In cases Effect measures
where studies reported the result of a two-sample t-test29, the In the various studies, FTD was referred to as “cognitive
t-statistic was converted into a SMD. Hedge’s g correction was disorganization”30,37–39, “disorganization“23,40–45, “thought distur-
applied to all SMDs to correct for small-sample size. In cases where bance”29, “FTD”46,47 or “positive FTD”13,48–51. This is linked to the
studies categorized groups into types of cannabis used (i.e. skunk, scale used (see supplementary material “Scales” for details). Each
herbal, mixed)30, or level of use (i.e. use, harmful use, Substance of these terms refers to a similar concept of FTD, encompassing
Use Disorder)13, we regrouped these cannabis users into a single speech abnormalities, poor attention, conceptual disorganization,
group for each study by calculating their mean scores and
and difficulties in thought process. For clarity, the umbrella term
standard deviations. SMDs were calculated comparing the
“FTD” will be used to refer to all these terms throughout this meta-
statistics of the cannabis-using groups with those of the non-
analysis. Eight studies23,30,37–39,42,44,45 assessed FTD globally using
using groups. In cases where SMD values are positive, cannabis-
using groups present higher FTD mean scores than non-using two questionnaires, the Oxford-Liverpool Inventory of Feelings
groups. When SMD values are negative, cannabis-using groups and Experiences52 and the Schizotypal Personality Question-
present lower FTD mean scores than non-using groups. The naire53, as well as three scales: the Positive and Negative
overall calculated effect size was then transformed back into Syndrome Scale54, the Psychotomimetic States Inventory52, and
natural units of two widely used scales (Positive and Negative the Structured Interview for Prodromal Symptoms45. Nine
Symptoms Scale and Scale for the Assessment of Positive studies13,29,40,41,46,48–51 evaluated subcomponents of FTD using
Symptoms) to facilitate interpretation of the results. This the Scale for the Assessment of Positive Symptoms55, the Brief
transformation is based on a recognized statistical method31–33; Psychiatric Rating Scale56, the Young Mania Rating Scale57, the
it was obtained by multiplying the calculated SMD by the pooled Operational Criteria for Psychotic Illness58 and the Symptom Onset
standard deviation of selected scales31,34. The interpretation in Schizophrenia59. The main subcomponent assessed via these
values for effect size (SMD) according to Cohen’s classification scales was positive FTD, characterized by disorganized speech.
are: 0.1–0.3 (small effect), 0.3–0.5 (moderate effect) and superior to One study reported results for both negative and positive FTD48.
0.5 (large effect). Since the other studies focused on either the global concept of
The meta-analysis was performed using the {meta} package. FTD13,29,30,37,47, which includes positive FTD, or only positive
Group combination, plot generation and outlier identification FTD23,46,50,51, this dimension was the only one considered in order
were performed using the {dmetar} package. Between-study to maintain homogeneity between compared outcomes. Some
heterogeneity was computed via a random effect model with I², studies also reported results as a proportion of participants in
where I² = 25%, I² = 50%, and I² = 75% are considered as low, whom FTD was either present or absent40,47,48,51.

Published in partnership with the Schizophrenia International Research Society Schizophrenia (2022) 78
 M. Argote et al.
4
Table 1. Studies included in review and meta-analysis.

Study reference - Design Extracted populationa Symptom (assessment mode) Key findings
country

Basu 199947 - India Retrospective 24 CIP Patients FTD (from case reports) Percentage of FTD per group:
20 acute schizophrenic episode CIP group: 15%
without history of use Acute schizophrenic episode
group: 60%
Boydell 200748 - UK Retrospective 182 SCZ + CU Positive FTD (OPCRIT checklist from Percentage of pFTD
552 SCZ nCU case reports) per group:
CU group: 32%
nCU group: 25%
Caspari 199929 - Longitudinal 39 SCZ + CUD Thought disturbance (BPRS) Thought disturbance:
Germany 39 SCZ nCU t-test between groups:
t = 2.25
Cohen 201123 - USA Cross-sectional 20 Schizotypy + CU Disorganization (SPQ-BR) Disorganization mean (sd)
74 Schizotypy nCU scores:
CU group: 31.31 (3.36)
nCU group: 30.58 (4.33)
Dubertret 200651 Cross-sectional 38 SCZ + CUD Positive FTD (SAPS) Percentage of pFTD
- France 121 SCZ nCU per group:
CUD group: 84%
nCU group: 85%
Gonzales-Blanco40 Retrospective 144 FEP + CU Disorganization (SOS) Percentage of
- Spain 70 FEP nCU Disorganization per group:
CU group: 63.9%
nCU group: 58%
Herzig 201537 - Cross-sectional 11 FEP + CU Cognitive Disorganization (PANSS) CogDis mean (sd) scores:
Switzerland 18 FEP nCU CU group: 6.10 (3.36)
nCU group: 4.45 (1.86)
Ho 201141 - USA Cross-sectional 52 SCZ + CUD Disorganization (SAPS subscore) Disorganization mean (sd)
183 SCZ nCU scores:
CU: 5.1 (3.0)
nCU: 4.8 (3.0)
Koen 200949 - Retrospective 245 SCZ + CU/CUD Positive FTD (SAPS) pFTD mean (sd) scores:
South Africa 302 SCZ nCU CU: 1.46 (1.58)
nCU: 1.02 (1.38)
Korver 201045 - The Cross-sectional 34 CU Disorganization (SIPS) Disorganization mean (sd)
Netherlands 29 nCU scores:
CU: 4.59 (2.3)
nCU: 5.17 (3.21)
Mackie 202130 - UK Cross-sectional 143 CU Cognitive disorganization (O-LIFE) CogDis mean (sd) scores:
323 nCU nCU:5.26 (2.7)
CU: 5.35 (3.03)
Mason 200839 – UK Repeated measures 140 CU Cognitive Disorganization (PSI) CogDis mean (sd) scores:
144 nCU CU: 3.7 (3.4)
nCU: 3.5 (4.0)
Nunn 200138 – UK Cross-sectional 49 CU Cognitive disorganization (O-LIFE) CogDis mean (sd) scores:
49 no drinking/drug-using CU: 13.20 (2.33)
no drink/drug: 12.76 (1.26)
O’Tuathaigh 202042 - Cross-sectional 181 CU Disorganization (SPQ) Disorganization mean (sd)
Ireland 563 nCU scores:
CU: 6.67 (4.03)
nCU: 5.29 (4.11)
Peralta 199250 - Spain Cross-sectional 23 SCZ + CUD Positive FTD (SAPS) pFTD mean (sd) scores:
72 SCZ nCU CUD: 2.9 (1.8)
nCU: 2.2 (1.7)
Pope 202146 - USA Cross-sectional 155 FEP + CU Positive FTD (SAPS) pFTD mean (sd) scores:
76 FEP nCU nCU: 6.4 (6.2)
CU: 6.9 (6.3)
Schiffman 200543 - Cross-sectional 43 CU Disorganization (SPQ-B) Disorganization mean (sd)
USA 146 nCU scores:
CU: 2.40 (1.78)
nCU: 1.71 (1.71)
Soler 201844 - Spain Cross-sectional 110 CU Disorganization (SPQ) Disorganization mean (sd)
275 nCU scores:
CU: 1.4 (1.31)
nCU: 1.06 (1.20)

Schizophrenia (2022) 78 Published in partnership with the Schizophrenia International Research Society
 M. Argote et al.
5
Table 1 continued
Study reference - Design Extracted populationa Symptom (assessment mode) Key findings
country

Stone 201413 - UK Longitudinal 207 FEP + CU Thought Disorder (YMRS) TD mean (sd) scores:
295 FEP nCU CU: 0.77 (1.04)
nCU: 0.52 (0.9)
Abbreviations: CIP Cannabis Induced Psychosis, CUD Cannabis Use Disorder, CU Cannabis Users, nCU non-Cannabis Users, AL Alcohol-using, SCZ Schizophrenia,
FEP First Episode Psychosis, SUD Substance Use Disorder; FTD Formal Thought Disorder, pFTD Positive Formal Thought Disorder, nFTD Negative Formal Thought
Disorder, TD Thought Disorder, sd Standard Deviation.
a
Refers to the population of interest extracted from original publications.

Study characteristics positive values (PI [0.10; 0.33]). A second sensitivity analysis was
The eligible studies included different types of population: (1) performed, excluding two studies considered as having a large
individuals with no history of diagnosed mental ill- influence on the results by the statistical leave-one-out
ness23,30,38,39,42–45; (2) individuals with a history of first-episode method30,42 (see supplementary material “InfluenceAnalysis”).
psychosis13,37,40,46; (3) individuals with schizophrenia29,41,48–51, The removal of these two studies did not significantly impact
according to validated diagnosis tools; and (4) individuals with the overall effect size, which means that their influence did not
cannabis-induced psychosis47, as determined by an expert. For the distort the calculated effect size in the first analysis (SMD = 0.22,
purpose of this meta-analysis, it was decided to form four distinct p < 0.0001, 95% CI [0.15; 0.29], I² < 1%, p = 0.59). Table 2 provides a
groups of individuals, according to the staging model of psychotic detailed explanation of the statistical results of the two distinct
disorders20: (1) individuals without symptoms, or attenuated sensitivity analyses performed.
syndrome (stages 0, 1a, 1b); (2) individuals with first-episode When the SMD was back transformed using the pooled standard
psychosis (stage 2); (3) individuals with schizophrenia (stages 3–4); deviation of the PANSS, the mean difference in Cognitive
and (4) individuals with cannabis-induced psychosis (Table 1). Disorganization score between cannabis-using and non-using groups
equals to 0.57 points; Cognitive Disorganization scores ranged from 3
to 21 points. Applied to the SAPS, the mean difference in positive
Risk of bias FTD score between cannabis-using and non-using groups equals to
The overall risk of bias was considered as low: the quality of four 0.93 points; positive FTD ranged from 0 to 45 points.
studies was rated as “good”29,37,44,51, that of fourteen studies as
“fair”13,23,30,38,40–43,45,46,48–50,52 and that of one as “poor”47. Poor
Subgroup analyses
quality was firstly determined by the imprecise criteria for
pertaining to the cannabis using group, i.e. “regular cannabis Subgroup analyses were performed to examine the effect of
use for at least 1 month prior to onset of psychosis”, and by the cannabis use on the groups formed in line with the staging model
imprecise measure of FTD in participants, i.e. “no psychopathology of psychotic disorders previously mentioned: (1) individuals with
scale could be used”47. Most studies were rated as “fair” quality as no or attenuated symptoms, (2) patients with first-episode
either question 9 or question 11 of the Quality Assessment Tool psychosis, and (3) patients diagnosed with schizophrenia. The
for Observational Cohort and Cross-sectional studies rated as “yes” cannabis-induced psychosis group was not represented as the
for these publications. Good quality was decided when the only study focusing on this population was the outlier and thus
questions 9, 11, and 14 rated as “yes”, as well as the other excluded. Individuals with no or attenuated symptoms were
questions of the tool. The detailed ratings can be found in the extracted from the general population and are thus considered a
supplementary material. prodromal population. First-episode psychosis patients were
described in the publications as having experienced psychotic
symptoms accompanied by a decrease in functioning over the
Cannabis Use and FTD month before inclusion37, being diagnosed with a primary non-
The overall effect size of the 19 pooled studies indicated that FTD affective psychotic disorder46, having a history of psychotic
severity was significantly increased in cannabis-users groups, symptoms that lasted for more than 7 days13, experiencing the
compared to non-using groups (SMD = 0.21, p = 0.00009, 95% CI first episode of schizophrenia, excluding brief psychotic episodes
[0.12; 0.29]). Between-group heterogeneity was moderate due to intoxication48, or experiencing a first occurrence of positive
(I² = 34%, p = 0.08). The estimated range of the prediction interval and negative, cognitive and affective symptoms40. Individuals
indicates that the effect size may lie between 0.07 and 0.34 in the using cannabis in the first group (no or attenuated symptoms)
context of future studies on the same issue. Original results are demonstrated significantly higher FTD scores compared to non-
presented in Fig. 2. users (SMD = 0.19, 95% CI [0.05;0.33], p = 0.02). Cannabis users in
the second group (first-episode psychosis) also had a higher FTD
Sensitivity analyses severity than non-users, the 2 groups showed a comparable effect
The study rated as being of “poor” quality was reported as an size (SMD = 0.21, 95% CI [0.01;0.41], p = 0.04). Although tests
outlier47. While its inclusion in the original analysis had a revealed that the differences in effect size were not significant
negligible impact on the overall results, it greatly affected between groups (p = 0.75), cannabis users with a diagnosis of
heterogeneity (Table 2. “Main” analysis for statistical details). schizophrenia displayed higher FTD severity compared with non-
Considering its outlier status and its large impact on hetero- users, with the highest effect size (SMD = 0.25, 95% CI [0.11;0.38],
geneity, it was excluded from the other analyses performed. Thus, p = 0.005).
the overall effect size of the 18 remaining studies was kept
throughout the current meta-analysis (SMD = 0.22, p = 0.000006, Publication bias
95% CI [0.14; 0.53]), with a low between-study heterogeneity Studies reporting low effect sizes may remain unpublished
(I² = 4.8%, p = 0.4). Prediction interval based on the 18 remaining because of non-significant results. Only studies reporting sig-
studies confirmed that the effect size may lie within a range of nificant results tend to be published, creating a publication bias.

Published in partnership with the Schizophrenia International Research Society Schizophrenia (2022) 78
 M. Argote et al.
6
Standardised Mean
Study Difference SMD 95%−CI

Population = No or attenuated symptoms

Korver et al. −0.21 [−0.70; 0.29]
Mackie et al. 0.05 [−0.15; 0.25]
Mason et al. 0.05 [−0.18; 0.29]
Cohen et al. 0.17 [−0.32; 0.67]
Nunn et al. 0.23 [−0.16; 0.63]
Soler et al. 0.28 [ 0.05; 0.50]
O'Tuathaigh et al. 0.34 [ 0.17; 0.51]
Schiffman et al. 0.40 [ 0.06; 0.74]
Random effects model 0.19 [ 0.05; 0.33]
Test for effect in subgroup: t 7 = 3.19 (p = 0.02)

Population = First−episode psychosis

Pope et al. 0.08 [−0.20; 0.35]
Gonzales−Blanco et al. 0.16 [−0.17; 0.48]
Stone et al. 0.26 [ 0.08; 0.44]
Herzig et al. 0.64 [−0.13; 1.40]
Random effects model 0.21 [ 0.01; 0.41]
Test for effect in subgroup: t 3 = 3.42 (p = 0.04)

Population = Schizophrenia
Dubertret et al. −0.04 [−0.55; 0.47]
Ho et al. 0.10 [−0.21; 0.41]
Boydell et al. 0.20 [−0.01; 0.41]
Koen et al. 0.30 [ 0.13; 0.47]
Peralta et al. 0.40 [−0.07; 0.88]
Caspari et al. 0.50 [ 0.05; 0.96]
Random effects model 0.25 [ 0.11; 0.38]
Test for effect in subgroup: t 5 = 4.76 (p < 0.01)

Population = Cannabis−induced psychosis

Basu et al. −1.16 [−2.04; −0.27]

Random effects model 0.21 [ 0.12; 0.29]

Prediction interval [ 0.07; 0.34]

Heterogeneity: I 2 = 34%, τ2 = 0.0027, p = 0.08 −2 −1 0 1 2

Test for subgroup differences: χ23 = 9.85, df = 3 (p = 0.02)

Fig. 2 Forest plot displaying results of original analysis combining individual studies and subgroup analyses, standardized mean difference
(SMD), its confidence interval (CI), and overall prediction interval (PI), plus the weight of each study.

test did not reveal significant presence of asymmetry, indicating

Table 2. Results from the two sensitivity analyses conducted: first
no evidence for a publication bias.
removing the outlier; then the high-influence studies from the original
analysis.

Analysis SMD 95%CI p 95%PI I² I² 95%CI

DISCUSSION
This meta-analysis aimed to review the relationship between
Originala 0.21 [0.12;0.29] 0.00009 [0.10;0.33] 34% [0.0%;61.9%] cannabis use and FTD along a continuum ranging from individuals
Main - 0.22 [0.15;0.29] <0.0001 [0.10;0.33] 5% [0.0%;52.4%] with no history of characterized mental disorders to patients
Outlier meeting the criteria for schizophrenia, what had never been
removedb performed to our knowledge. The main results of the global
Infl. 0.22 [0.15;0.29] <0.0001 [0.14;0.30] <1% [0.0%;52.3%] analysis showed that individuals using cannabis presented a
studies significantly higher severity of FTD than non-users. Although the
removedc effect was significant, the size of the effect (SMD = 0.20) should be
a
All studies included (k = 19). bOutlier excluded47(k = 18). cThree studies considered as small60. Subgroup analyses were then performed to
excluded30,42,47 (k = 16). Abbreviations: k, number of studies, CI Confidence investigate whether the relationship between cannabis use and
Interval; PI Predication Interval, Infl. studies Influence studies. FTD varied between individuals of the included populations: no or
attenuated symptoms, first-episode psychosis, and schizophrenia.
Cannabis users from all three subgroups exhibited a similar FTD
This bias can be visually and statistically assessed via the severity, suggesting a stable association independent of the
inspection of a funnel plot and the statistical Egger’s test. The underlying stage of the psychotic disorder. Thus, the association
visual inspection of the funnel plot indicates a relatively low between FTD and cannabis use should not be seen as an
asymmetry. However, one study appears far away from the ideal intermediate vulnerability marker for schizophrenia. However,
delineated area (see Fig. 3). It is the same study that was found to since both cannabis use and FTD severity were associated with
be an outlier and excluded from the analysis47. Egger’s test was increased risk for psychosis, we cannot rule out a role of the
performed on the nineteen studies to statistically assess the relationship between FTD and cannabis use in the psychotic
publication bias (supplementary material “EggerTest”). The Egger’s transition. Indeed, triggering factors of psychotic symptoms are

Schizophrenia (2022) 78 Published in partnership with the Schizophrenia International Research Society
 M. Argote et al.
7

0.0
O'Tuathaigh
Koenetetet
Stone al.
al.
al.

0.1
Mackie et al.
Boydell et al.
Soler et al.
Mason et al.
Pope et al.
Ho et al.
Gonzales−Blanco et al.
Schiffman et al.
Standard Error

0.2

Nunn et al.

Caspari et al.
Peralta et al.
Cohen
Korver et al. et al.
Dubertret et al.
0.3

Herzig et al.
0.4

Basu et al.

−2 −1 0 1 2

Standardised Mean Difference

Fig. 3 Funnel plot investigating publication bias.

complex and combine both biological features such as genetic Disorder (SUD) cannot be ruled out. SUDs involving cocaine,
traits, and environmental factors such as socio-economic or alcohol, stimulants, hallucinogens, or tobacco are associated with
migration status61. The results support the hypothesis that FTD severity13,66,67. Another identified limitation is that, even if the
cannabis use could be seen as one of the key factors explaining search syntax used in the current meta-analysis intended to
the severity of FTD. include all publications on FTD and its subcomponents, FTD is
The results are in line with previous findings where disorganiza- actually a heterogeneous construct which can be expressed using
tion symptoms were exacerbated by cannabis use in individuals many terms, and whose exact conceptual structure is still being
with schizotypal traits19. In addition, some studies described an explored using factor analyses of clinical scales7. Furthermore,
aggravation of disorganization when individuals at a prodromal there is a paucity of publications exploring the association
stage of psychosis present an acute cannabis intoxication62,63. between cannabis use and specific FTD subcomponents. Hence,
Findings from a ten-year longitudinal experiment, which investi- FTD subcomponents other than the positive dimension could not
gated the impact of cannabis use on the course of illness in be adequately represented in the current study. Previous works
schizophrenia, reported that compared to non-users, cannabis found a significant positive association between cannabis use and
users presented higher scores of disorganized symptoms at the 6- negative FTD48. This finding should be replicated, as it could
month, 4-year and 10-year follow-ups64, implying an aggravation broaden knowledge on this issue. In addition, subcomponents of
of FTD by cannabis use, in line with the findings presented in the FTD need to be more systematically explored in future research
current review. In the same study, however, a decrease in via the use of scales specifically designed to measure FTD severity
disorganized symptoms for some individuals was associated with such as the Thought And Language Disorder (TALD), the Thought
a higher likelihood of cannabis use64. This could mean that some Language and Communication Disorders (TLC), the Thought and
individuals use cannabis to alleviate their disorganized symp- Language Index (TLI), and the Thought Disorder Index (TDI)1.
toms64, demonstrating the need to distinguish between subjective Finally, a last limitation of the current meta-analysis is to not have
and objective FTD3,65. been preregistered in a public registry.
This meta-analysis has some strengths. Publication bias, as well
as the risk of bias in each individual study, and small-sample study
biases, were taken into account in our meta-analysis. Between- CONCLUSION
study heterogeneity is considered as very low, due to the use of a Previous studies found that cannabis use was associated with
random model and the outlier identification. All the relevant more severe positive symptoms of psychosis, within an epide-
methods that avoid bias were employed to provide the most miological continuum ranging from the general population to
reliable results. However, there are several factors that could not schizophrenia spectrum disorders. Moreover, accumulating pieces
be controlled for as this meta-analysis used aggregated data. of evidence suggest a causal role of cannabis in triggering and
Because the included studies were epidemiological investigations, aggravating positive symptoms, particularly in schizophrenia
it is impossible to ascertain a causal effect of cannabis use on the spectrum disorders. Since FTD has been suggested to constitute
severity of FTD; intermediary factors such as sociodemographic a core symptom of schizophrenia, it was theoretically expected
features, the use of cannabis for self-medication, the effect of that cannabis use would be associated with exacerbated FTD. The
antipsychotic treatment, or the presence of Substance Use results presented in this meta-analysis tend to support this

Published in partnership with the Schizophrenia International Research Society Schizophrenia (2022) 78
 M. Argote et al.
8
association, and the effect of cannabis use on FTD seems 18. Minor, K. S., Marggraf, M. P., Davis, B. J., Mehdiyoun, N. F. & Breier, A. Affective
consistent along the psychosis continuum, regardless of the systems induce formal thought disorder in early-stage psychosis. J. Abnormal
severity of the psychotic disorder. Intermediary factors need to be Psychol. 125, 537–542 (2016).
controlled for in future research on this issue. These findings 19. Szoke, A. et al. Association between cannabis use and schizotypal dimensions-a
meta-analysis of cross-sectional studies. Psychiatry Res. 219, 58–66 (2014).
encourage future research to assess FTD using specific scales and
20. McGorry, P. D., Hartmann, J. A., Spooner, R. & Nelson, B. Beyond the “at risk
support the clinical relevance of assessing FTD in early stages of mental state” concept: transitioning to transdiagnostic psychiatry. World Psy-
the psychosis continuum, specifically among individuals using chiatry 17, 133–142 (2018).
cannabis. 21. Page, M. J. et al. The PRISMA 2020 statement: An updated guideline for reporting
systematic reviews. BMJ. 372, n71 (2021).
22. Bramer, W. M., Rethlefsen, M. L., Kleijnen, J. & Franco, O. H. Optimal database
DATA AVAILABILITY combinations for literature searches in systematic reviews: A prospective
The data supporting the findings of this study are available within the paper and its exploratory study. Syst. Rev. 6, 245 (2017).
supplementary information files. 23. Cohen, A. S., Buckner, J. D., Najolia, G. M. & Stewart, D. W. Cannabis and
psychometrically-defined schizotypy: Use, problems and treatment considera-
tions. J. Psychiatric Res. 45, 548–554 (2011).
CODE AVAILABILITY 24. Kiang, M. et al. Association of abnormal semantic processing with delusion-like
ideation in frequent cannabis users: An electrophysiological study. Psycho-
The code written for the analyses will be available upon request.
pharmacology 225, 95–104 (2013).
25. Shakoor, S. et al. Psychotic experiences are linked to cannabis use in adolescents
Received: 5 May 2022; Accepted: 10 September 2022; in the community because of common underlying environmental risk factors.
Psychiatry Res. 227, 144–151 (2015).
26. National Institution of Health: U.S Department of Health and Human Services.
Quality assessment tool for observational cohort and cross-sectional studies.
Available online: https://www.nhlbi.nih.gov/health-topics/study-quality-assessm
ent-tools (2015).
REFERENCES 27. Ma, L. L. et al. Methodological quality (risk of bias) assessment tools for primary
1. Kircher, T., Bröhl, H., Meier, F. & Engelen, J. Formal thought disorders: From and secondary medical studies: What are they and which is better? Military Med.
phenomenology to neurobiology. Lancet Psychiatry 5, 515–526 (2018). Res. 7, 7 (2020).
2. Roche, E., Creed, L., MacMahon, D., Brennan, D. & Clarke, M. The epidemiology 28. R Core Team. R: A Language and Environment for Statistical Computing. Vienna,
and associated phenomenology of formal thought disorder: A systematic review: Austria: R Foundation for Statistical Computing. https://www.R-project.org/ (2021).
Fig. 1. SCHBUL 41, 951–962 (2015). 29. Caspari, D. Cannabis and schizophrenia: Results of a follow-up study. European
3. Mutlu, E. et al. The cognitive aspect of formal thought disorder and its rela- Archives Psychiatry Clin. Neurosci. 249, 45–49 (1999).
tionship with global social functioning and the quality of life in schizophrenia. 30. Mackie, C. J. et al. A latent class analysis of cannabis use products in a general
Soc. Psychiatry Psychiatr. Epidemiol. 56, 1399–1410 (2021). population sample of adolescents and their association with paranoia, halluci-
4. Cavelti, M., Kircher, T., Nagels, A., Strik, W. & Homan, P. Is formal thought disorder nations, cognitive disorganisation and grandiosity. Addictive Behav. 117, 106837
in schizophrenia related to structural and functional aberrations in the language (2021).
network? A systematic review of neuroimaging findings. Schizophr. Res. 199, 2–16 31. Schünemann, H. J. et al. Cochrane Handbook for Systematic Reviews of Interven-
(2018). tions Version 6.2 (eds. Higgins, J., Thomas, J. & Chandler, J. et al.) (Wiley, 2021).
5. de Sousa, P., Spray, A., Sellwood, W. & Bentall, R. P. ‘No man is an island’. Testing 32. Devji, T. et al. Presentation approaches for enhancing interpretability of patient-
the specific role of social isolation in formal thought disorder. Psychiatry Res. 230, reported outcomes (PROs) in meta-analysis: A protocol for a systematic survey of
304–313 (2015). Cochrane reviews. BMJ Open 7, e017138 (2017).
6. Tan, E. J. & Rossell, S. L. Building a neurocognitive profile of thought disorder in 33. Murad, M. H., Wang, Z., Chu, H. & Lin, L. When continuous outcomes are mea-
schizophrenia using a standardized test battery. Schizophr. Res. 152, 242–245 sured using different scales: Guide for meta-analysis and interpretation. BMJ 364,
(2014). k4817 (2019).
7. Ucok, A., Karakaş, B. & Şahin, O. Ş. Formal thought disorder in patients with first- 34. Thorlund, K., Walter, S. D., Johnston, B. C., Furukawa, T. A. & Guyatt, G. H. Pooling
episode schizophrenia: Results of a one-year follow-up study. Psychiatry Res. 301, health-related quality of life outcomes in meta-analysis-a tutorial and review of
113972 (2021). methods for enhancing interpretability: Enhancing interpretability in continuous
8. Demjaha, A. et al. Combining dimensional and categorical representation of meta-analysis. Res. Syn. Meth. 2, 188–203 (2011).
psychosis: The way forward for DSM-V and ICD-11? Psychol. Med. 39, 1943–1955 35. Higgins, J. P. T. & Thompson, S. G. Quantifying heterogeneity in a meta-analysis.
(2009). Stat. Med. 21, 1539–1558 (2002).
9. Ayer, A. et al. Formal thought disorder in first-episode psychosis. Comprehensive 36. Jackson, D., Law, M., Rücker, G. & Schwarzer, G. The Hartung–Knapp modification
Psychiatry 70, 209–215 (2016). for random‐effects meta‐analysis: A useful refinement but are there any residual
10. Green, B., Young, R. & Kavanagh, D. Cannabis use and misuse prevalence among concerns. Stat. Med. 36, 3923–3934 (2017).
people with psychosis. Br. J. Psychiatry 187, 306–313 (2005). 37. Herzig, D. A. et al. Hemispheric language asymmetry in first episode psychosis
11. Hunt, G. E., Large, M. M., Cleary, M., Lai, H. M. X. & Saunders, J. B. Prevalence of and schizotypy: The role of cannabis consumption and cognitive disorganization.
comorbid substance use in schizophrenia spectrum disorders in community and Schizophr. Bull. 41, S455–S464 (2015).
clinical settings, 1990–2017: Systematic review and meta-analysis. Drug Alcohol 38. Nunn, J. A., Rizza, F. & Peters, E. R. The incidence of schizotypy among cannabis
Dependence 191, 234–258 (2018). and alcohol users. J. Nervous Mental Dis. 189, 741–748 (2001).
12. Ortiz-Medina, M. B. et al. Cannabis consumption and psychosis or schizophrenia 39. Mason, O. J., Morgan, C. J. M., Stefanovic, A. & Curran, H. V. The Psychotomimetic
development. Int. J. Soc. Psychiatry 64, 690–704 (2018). States Inventory (PSI): Measuring psychotic-type experiences from ketamine and
13. Stone, J. M. et al. Cannabis use and first-episode psychosis: Relationship with cannabis. Schizophr. Res. 103, 138–142 (2008).
manic and psychotic symptoms, and with age at presentation. Psychol. Med. 44, 40. González-Blanco, L. et al. Impact of previous tobacco use with or without can-
499–506 (2014). nabis on first psychotic experiences in patients with first-episode psychosis.
14. Hasan, A. et al. Cannabis use and psychosis: A review of reviews. Eur. Arch. Schizophr. Res. 236, 19–28 (2021).
Psychiatry Clin. Neurosci. 270, 403–412 (2020). 41. Ho, B. C., Wassink, T. H., Ziebell, S. & Andreasen, N. C. Cannabinoid receptor 1
15. Schoeler, T. et al. Continued versus discontinued cannabis use in patients with gene polymorphisms and marijuana misuse interactions on white matter and
psychosis: A systematic review and meta-analysis. Lancet Psychiatry 3, 215–225 cognitive deficits in schizophrenia. Schizophr. Res. 128, 66–75 (2011).
(2016). 42. O’Tuathaigh, C. M. P. et al. Does cannabis use predict psychometric schizotypy via
16. Sabe, M., Zhao, N. & Kaiser, S. Cannabis, nicotine and the negative symptoms of aberrant salience? Schizophr. Res. 220, 194–200 (2020).
schizophrenia: Systematic review and meta-analysis of observational studies. 43. Schiffman, J., Nakamura, B., Earleywine, M. & LaBrie, J. Symptoms of schizotypy
Neurosci. Biobehav. Rev. 116, 415–425 (2020). precede cannabis use. Psychiatry Res. 134, 37–42 (2005).
17. Bleuler, M. [Origin and significance of Eugen Bleuler’s work: “Dementia praecox or 44. Soler, J. et al. The interaction between the ZNF804A gene and cannabis use on
the schizophrenia group”. A contribution to the history of Manfred Bleuler’s the risk of psychosis in a non-clinical sample. Prog. Neuropsychopharmacol. Biol.
Zuricher psychiatry]. Schweiz Rundsch Med. Prax 77, 1322–1326 (1988). Psychiatry 89, 174–180 (2019).

Schizophrenia (2022) 78 Published in partnership with the Schizophrenia International Research Society
 M. Argote et al.
9
45. Korver, N. et al. Symptomatology and neuropsychological functioning in cannabis 65. Sumner, P. J., Rossell, S. L. & Tan, E. J. On the assessment of subjective formal
using subjects at ultra-high risk for developing psychosis and healthy controls. thought disorder. Schizophr. Res. 228, 58–59 (2021).
Aust. N. Z. J. Psychiatry 44, 230–236 (2010). 66. Mauri, M. et al. Substance abuse in first-episode schizophrenic patients: A ret-
46. Pope, L. G., Manseau, M. W., Kelley, M. E. & Compton, M. T. Symptomatology and rospective study. Clin. Pract. Epidemiol. Ment. Health 2, 4 (2006).
neurocognition among first-episode psychosis patients with and without can- 67. Nagels, A. et al. Distinct neuropsychological correlates in positive and negative
nabis use in the three months prior to first hospitalization. Schizophr. Res. 228, formal thought disorder syndromes: The thought and language disorder scale in
83–88 (2021). endogenous psychoses. Neuropsychobiology 73, 139–147 (2016).
47. Basu, D., Malhotra, A., Bhagat, A. & Varma, V. K. Cannabis psychosis and acute
schizophrenia—A case-control study from India. Eur. Addict. Res. 5, 71–73 (1999).
48. Boydell, J. et al. A comparison of symptoms and family history in schizophrenia DATA AVAILABILITY
with and without prior cannabis use: Implications for the concept of cannabis The data supporting the findings of this study are available within the paper and its
psychosis. Schizophr. Res. 93, 203–210 (2007). supplementary information files.
49. Koen, L., Jonathan, R. & Niehaus, D. J. Cannabis use and abuse correlates in a
homogenous South African schizophrenia population. S. Afr. J. Psych. 15, 5 (2009).
50. Peralta, V. & Cuesta, M. J. Influence of cannabis abuse on schizophrenic psy-
COMPETING INTERESTS
chopathology. Acta Psychiatr. Scand. 85, 127–130 (1992).
51. Dubertret, C., Bidard, I., Adès, J. & Gorwood, P. Lifetime positive symptoms in The authors declare no competing interests.
patients with schizophrenia and cannabis abuse are partially explained by co-
morbid addiction. Schizophr. Res. 86, 284–290 (2006).
52. Mason, O., Claridge, G. & Jackson, M. New scales for the assessment of schizotypy. ADDITIONAL INFORMATION
Personality Individual Differences 18, 7–13 (1995). Supplementary information The online version contains supplementary material
53. Raine, A. & Benishay, D. The SPQ-B: A brief screening instrument for schizotypal available at https://doi.org/10.1038/s41537-022-00286-0.
personality disorder. J. Personality Dis. 9, 346–355 (1995).
54. Kay, S. R., Fiszbein, A. & Opler, L. A. The positive and negative syndrome scale Correspondence and requests for materials should be addressed to Mathilde Argote.
(PANSS) for schizophrenia. Schizophr. Bull. 13, 261–276 (1987).
55. Andreasen, N. C. Scale for the assessment of positive symptoms. https://doi.org/ Reprints and permission information is available at http://www.nature.com/
10.1037/t48377-000 (2018). reprints
56. Mueser, K. T., Curran, P. J. & McHugo, G. J. Factor structure of the Brief Psychiatric
Rating Scale in schizophrenia. Psychol. Assess. 9, 196–204 (1997). Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims
57. Young, R. C., Biggs, J. T., Ziegler, V. E. & Meyer, D. A. A rating scale for mania: in published maps and institutional affiliations.
Reliability, validity, and sensitivity. Br. J. Psychiatry 133, 429–435 (1978).
58. Bergman, H. et al. Operational Criteria Checklist for Psychotic Illness and Affective
Illness (OPCRIT+) for diagnosing schizophrenia in people with psychotic symp-
toms. Cochrane Schizophrenia Group, ed. Cochrane Database of Systematic
Reviews. https://doi.org/10.1002/14651858.CD011104 (2014). Open Access This article is licensed under a Creative Commons
59. Perkins, D. O. et al. Characterizing and dating the onset of symptoms in psychotic Attribution 4.0 International License, which permits use, sharing,
illness: The Symptom Onset in Schizophrenia (SOS) inventory. Schizophr. Res. 44, adaptation, distribution and reproduction in any medium or format, as long as you give
1–10 (2000). appropriate credit to the original author(s) and the source, provide a link to the Creative
60. Cohen J. Statistical Power Analysis for the Behavioral Sciences. Rev. ed. New York: Commons license, and indicate if changes were made. The images or other third party
Academic Press; 1977. material in this article are included in the article’s Creative Commons license, unless
61. Owen, M. J., Sawa, A. & Mortensen, P. B. Schizophrenia. Lancet 388, 86–97 (2016). indicated otherwise in a credit line to the material. If material is not included in the
62. Mason, O. et al. Acute cannabis use causes increased psychotomimetic experi- article’s Creative Commons license and your intended use is not permitted by statutory
ences in individuals prone to psychosis. Psychol. Med. 39, 951–956 (2009). regulation or exceeds the permitted use, you will need to obtain permission directly
63. Mokrysz, C., Freeman, T. P., Korkki, S., Griffiths, K. & Curran, H. V. Are adolescents from the copyright holder. To view a copy of this license, visit http://
more vulnerable to the harmful effects of cannabis than adults? A placebo- creativecommons.org/licenses/by/4.0/.
controlled study in human males. Translational Psychiatry 6, e961–e961 (2016).
64. Foti, D. J., Kotov, R., Guey, L. T. & Bromet, E. J. Cannabis use and the course of
schizophrenia: 10-year follow-up after first hospitalization. AJP 167, 987–993 (2010). © The Author(s) 2022

Published in partnership with the Schizophrenia International Research Society Schizophrenia (2022) 78