Journal of Affective Disorders 95 (2006) 103 – 110

www.elsevier.com/locate/jad

Research report
Cannabis use and expression of mania in the general population
Cécile Henquet a , Lydia Krabbendam a , Ron de Graaf b ,
Margreet ten Have b , Jim van Os a,c,⁎
a
Department of Psychiatry and Neuropsychology, South Limburg Mental Health Research and Teaching Network,
EURON, Maastricht University, Maastricht, The Netherlands
b
Netherlands Institute of Mental Health and Addiction, Trimbos Institute, Utrecht, The Netherlands
c
Division of Psychological Medicine, Institute of Psychiatry, London, United Kingdom
Received 23 March 2006; received in revised form 3 May 2006; accepted 5 May 2006
Available online 21 June 2006

Abstract

Background: Cannabis use is common in patients with bipolar disorder, however little is known about cannabis as a risk factor for
mania. In order to investigate the association between exposure to cannabis and subsequent development of manic symptoms
whilst controlling for psychotic symptoms, a longitudinal population-based study was carried out.
Methods: 4815 individuals aged 18 to 64 years were interviewed using the Composite International Diagnostic Interview at
baseline, 1 year follow up and 3 year follow up, including assessment of substance use, manic symptoms and psychotic symptoms.
Results: Use of cannabis at baseline increased the risk for manic symptoms during follow-up (adjusted OR 2.70, 95% CI: 1.54, 4.75),
adjusted for age, sex, educational level, ethnicity, single marital status, neuroticism, use of other drugs, use of alcohol, depressive
symptoms and manic symptoms at baseline. The association between cannabis use and mania was independent of the prevalence and the
incidence of psychotic symptoms. There was no evidence for reverse causality, as manic symptoms at baseline did not predict the onset of
cannabis use during follow-up (OR = 0.35, 95% CI: 0.03, 3.49).
Limitations: As 3 years is a relative short period of follow-up, long-term effects of cannabis use on mania outcomes could not be detected.
Conclusion: The results suggest that cannabis use may affect population expression of manic symptoms (and subsequent risk to
develop bipolar disorder [Regeer, E.J., Krabbendam, L., R, DE Graaf, Ten Have, M., Nolen, W.A., Van Os, J., 2006. A prospective
study of the transition rates of subthreshold (hypo)mania and depression in the general population. Psychol Med, 1-9.]). These
findings may not be due to the emergence of psychotic symptoms or the effects of self-medication.
© 2006 Elsevier B.V. All rights reserved.

Keywords: Tetrahydrocannabinol; Bipolar disorder; Psychosis; General population; Cohort study

1. Introduction

Manic symptoms are common in patients diagnosed

with schizophrenia and, conversely, psychotic symptoms
often occur in those with bipolar disorder. The two co-
⁎ Corresponding author. Department of Psychiatry and Neuropsy-
morbid but separable symptom dimensions of mania and
chology, Maastricht University, PO Box 616 (DRT 10), 6200 MD
Maastricht, The Netherlands. Tel.: +31 43 3875443; fax: +31 43 psychosis (McGorry et al., 1998; Peralta and Cuesta, 1999)
3875444. also display a degree of overlap in genetic and non-genetic
E-mail address: [email protected] (J. van Os). aetiological influences (Murray et al., 2004; Walker et al.,
0165-0327/$ - see front matter © 2006 Elsevier B.V. All rights reserved.
doi:10.1016/j.jad.2006.05.002
104 C. Henquet et al. / Journal of Affective Disorders 95 (2006) 103–110

2002). Increased risk for schizophrenia has been reported studies have actually investigated the temporal sequence
in relatives of patients with bipolar illness characterised by of substance use and bipolar disorder (Escamilla et al.,
a high familial loading (Valles et al., 2000) and twin studies 2002). To our knowledge, no prospective study to date has
suggest overlap in the genes contributing to schizophrenia, investigated the hypothesis of cannabis as a risk factor for
schizo-affective and mania syndromes (Cardno et al., mania outcomes, disentangling co-morbidity with psy-
2002). High rates of mental illness among minority groups chotic symptoms, potential confounding variables (such
are not specific to schizophrenia and have been described as use of other drugs), and reverse causality (i.e. the self-
in mania as well (van Os et al., 1996). Furthermore, medication hypothesis).
neuroticism has been associated with the development of The aims of the current study, therefore, were to
both schizophrenia (van Os and Jones, 2001)/psychotic investigate prospectively (i) if baseline cannabis use
symptoms (Krabbendam et al., 2002) and bipolar disorder increases the risk for development of manic symptoms,
(Angst et al., 2003a,b). Other risk factors however, such as (ii) if the association between cannabis and mania is
obstetric complications (Browne et al., 2000) and independent of the emergence of psychotic symptoms,
urbanicity (Mortensen et al., 2003) have been associated and (iii) if baseline mania predicts cannabis use at follow-
with schizophrenia/psychotic symptoms, but not with up (the self-medication hypothesis).
bipolar disorder.
Although evidence is accumulating that cannabis is a 2. Methods
risk factor for schizophrenia/psychotic symptoms
(Andreasson et al., 1987; van Os et al., 2002), little is 2.1. Sample
known about cannabis as a shared risk factor for both
mania and psychosis. In studies of psychotic outcomes, The Netherlands Mental Health Survey and Incidence
there is evidence that exposure to cannabis plays a role not Study (NEMESIS), is a prospective study with three
only in the expression of psychotic disorder, but also in measurement points over a period of 3 years (Bijl et al.,
the emergence of psychotic experiences at lower levels of 1998a,b). A multistage, stratified, random sampling pro-
severity in non-clinical samples (Henquet et al., 2005; van cedure was used to first select 90 municipalities, then a
Os et al., 2002; Verdoux et al., 2003). Results from sample of private households, and finally a Dutch-
population-based studies furthermore suggest that canna- speaking individual aged 18–64 years within each house-
bis use interacts synergistically with pre-existing liability hold. Selected households were sent an introductory letter
to psychosis, indicating that the risk-enhancing effect of by the Minister of Health, inviting them to participate. A
cannabis is much stronger in individuals with prior total of 7076 individuals provided informed consent and
evidence of psychosis diathesis (Caspi et al., 2005; was interviewed at baseline, representing a response rate
Henquet et al., 2005; van Os et al., 2002). Patients with of 69.7%. At T1, 5618 subjects participated at the first
bipolar disorder have elevated levels of substance use follow-up and at T2, 4848 subjects participated at the
(Regier et al., 1990; Strakowski and DelBello, 2000), second follow-up; 4815 individuals had completed the
including cannabis (Sherwood Brown et al., 2001). There mania section of the CIDI at both follow ups. The sample
is also evidence that substance use in these patients is was found to be representative of the Dutch population in
associated with poor treatment response and poorer clini- terms of gender, marital status and level of urbanisation
cal outcome (Sonne et al., 1994; Tohen et al., 1990). There (Bijl et al., 1998b), with the exception of a slight under-
are no data shedding light on whether associations bet- representation of individuals in the age group 18–24 years.
ween cannabis use and mania may be causal (Strakowski
and DelBello, 2000). Clinical data, however, suggest that 2.2. Measures
in many patients the use of substances precedes the onset
of bipolar disorder (Strakowski et al., 1998). Several case Subjects were interviewed at home using the Com-
studies, on the other hand, report that patients may start posite International Diagnostic Interview (CIDI) version
using cannabis to moderate their manic symptoms 1.1 (Smeets and Dingemans, 1993). The CIDI was
(Grinspoon and Bakalar, 1998; Khantzian, 1997; Stra- designed for trained interviewers who are not clinicians
kowski and DelBello, 2000). Prospective studies, how- and has been found to have high inter-rater reliability
ever, have not provided evidence to support the self- (Cottler et al., 1991; Wittchen et al., 1991) and high test–
medication hypothesis, as patients with prior histories of retest reliability (Wittchen, 1994). Ninety interviewers
substance use often did not resume their substance use experienced in systematic data collection collected the
after onset of the disease (Strakowski and DelBello, 2000; data, having received a 3-day training course in recruiting
Strakowski et al., 1998). Only few population-based and interviewing, followed by a 4-day course at the
 C. Henquet et al. / Journal of Affective Disorders 95 (2006) 103–110 105

WHO-CIDI training centre in Amsterdam. Extensive 2.5. Psychotic symptom scores

monitoring and quality checks took place throughout the
entire data collection period (Bijl et al., 1998b). CIDI Baseline lifetime ratings from the 17 CIDI core
assessment at baseline yielded lifetime ratings, assess- psychosis sections on delusions (13 items) and halluci-
ment during follow up referred to the period between nations (four items) were used (items G1–G13, G15,
baseline and T1 and between T1 and T2. G16, G20, G21). These concern classic psychotic
symptoms involving, for example, persecution, thought
2.3. Drug use assessment interference, auditory hallucinations and passivity
phenomena. All these items can be rated in six ways:
Cannabis use was assessed using the CIDI-L section, “1”—No symptom, “2”—Symptom present but not
which includes a variety of substances. Two types of clinically relevant (not bothered by it and not seeking
cannabis use at baseline were constructed: (1) lifetime any help for it), “3”—Symptom result of ingestion of drugs,
use (hereafter: baseline cannabis use and (2) lifetime “4”—Symptom result of somatic disease, “5”—Symp-
frequency of use (hereafter: baseline frequency of tom present, bothered by it/seeking help for it, “6”—
cannabis use). Baseline frequency of cannabis use Symptom may not really be a symptom because there
referred to the period of heaviest use on a 1 to 5 scale appears to be some plausible explanation for it.
(less than once a month; 1–3 days/month; 1–2 days/week; Individuals with at least one positive rating on any of
3–4 days/week; nearly every day). Similarly, cannabis use the CIDI psychosis items at baseline, irrespective of the
at either T1 or at T2 was combined into a single variable qualitative symptom score of to “2” to “6”, were con-
‘cannabis use during follow up’ (i.e. use of cannabis from sidered as having psychotic symptoms at baseline
baseline to T1 and/or use of cannabis from T1 to T2). (hereafter: baseline psychosis). The justification for this
Lifetime use of psychostimulants, cocaine, phencyclidine broad rating was derived from a previous study, where it
(PCP) and psychedelics at baseline, was combined into was shown that the five qualitative ratings of “2” to “6” on
one variable ‘use of other drugs’. Alcohol use at baseline the CIDI psychosis items were in fact strongly associated
was analysed as frequency of use during the past with each other (van Os et al., 2000). In addition, the five
twelve months (no use; less than once a month; 1–3 qualitative ratings independently showed a similar pattern
days/month; 1–2 days/week; 3–4 days/week; nearly of associations with known risk factors for psychosis (van
every day). Os et al., 2000, 2001).
At T1 and T2, the same 17 CIDI psychosis items were
2.4. Manic symptom scores used in the interviews (covering the time intervals
between interviews), yielding T1 and T2 measures of
The CIDI mania section (section F) consists of two psychosis defined similarly as the baseline measure
probe items asking about the presence of experiences described above. A single follow-up psychosis measure
during a period of at least 2 days of (1) feeling so happy or was constructed similar to the single follow-up mania
excited that it caused the individual to get into trouble or measure, by combining the T1 and T2 psychosis measure
family or friends to be worried about it, or a doctor saying into one item (hereafter: follow-up psychosis).
the subject was manic or (2) being unusually irritable so
that the subject complained, started arguments or shouted 2.6. Statistical analyses
at or hit people. A positive rating on at least one of these
probe items, subsequently directed to further mania items 2.6.1. Associations between cannabis and mania
(9 items), to explore the specific nature of the manic outcome
symptoms. All of these 9 items can be rated either “yes” Associations between baseline cannabis use and
(1) or “no” (2). At baseline, the mania outcome was follow-up mania were expressed as odds ratios using
defined as having at least one positive rating on any of the logistic regression analyses in the STATA statistical
11 items of the CIDI core mania section (F) (hereafter: software program (Stata Corporation, 2003). All analyses
baseline mania). Similar mania scores were constructed were a priori adjusted for age (five groups), sex (Escamilla
during follow-up, and a single follow-up mania outcome et al., 2002), educational level (four levels), ethnicity (van
was constructed consisting of a positive rating on at least Os et al., 1996) (0, subject and both parents born in the
one of the mania items at T1 or at T2 (hereafter: follow-up Netherlands; and 1, other), single marital status, neurot-
mania). These measures of isolated manic symptoms were icism (Angst et al., 2003b), lifetime use of other drugs,
validated, in terms of risk of transition to bipolar disorder, alcohol use during the past twelve months, baseline
in a previous study (Regeer et al., 2006). depression (total score on the CIDI-section E)(Stefanis et
106 C. Henquet et al. / Journal of Affective Disorders 95 (2006) 103–110

Table 1
Patterns of baseline cannabis use and follow-up manic symptoms
Cannabis Any manic symptom at follow-up in whole Any manic symptom at follow-up in risk set Any manic symptom at follow-up in risk
exposure risk set with no manic symptoms at baseline set with no manic or psychotic symptoms at
at baseline
baseline
N outcome− N outcome+ N outcome− N outcome+ N outcome− N outcome+
(n = 4679) (n = 118) (n = 4560) (n = 63) (n = 3924) (n = 34)
No use 4284 (98.2%) 78 (1.8%) 4174 (99.1%) 40 (0.9%) 3641 (99.4%) 23 (0.6%)
Any use 413 (91.2%) 40 (8.8%) 386 (94.4%) 23 (5.6%) 283 (96.3%)) 11 (3.7%)

al., 2002) and baseline mania. To examine whether the cannabis preceded the incidence of manic symptoms. In
effect of baseline cannabis use on follow-up mania was addition, analyses were restricted to individuals who at
independent of psychotic symptoms, associations be- baseline had a score of lifetime absence on all the
tween baseline cannabis use and follow-up mania were individual items on both the CIDI mania section and the
additionally adjusted for both baseline and follow-up CIDI psychosis section (n = 3958), in order to investigate
psychosis. In order to examine whether any association whether the use of cannabis was specifically associated
with follow-up mania was due to recent intoxication with the mania outcome and independent of baseline
rather than the long-term effects of cannabis exposure, prevalence of psychotic symptoms. The analysis to
baseline cannabis use and cannabis use during follow-up investigate reverse causality was repeated for the
were entered jointly in the adjusted model. individuals who were lifetime cannabis-naive at baseline.
Reverse causality (i.e. the self-mediation hypothesis) Sensitivity analyses were conducted to examine
was investigated by assessing the association between whether differential attrition in the sample as a whole
baseline mania and cannabis use during follow-up. Odds (7076 at baseline, 4815 at follow-up) could have biased
ratios were adjusted for the a priori selected confoun- the findings. This was done by multiple imputation of
ders, but not for follow-up mania, nor for baseline and missing values of CIDI follow-up mania using the
follow-up psychosis. HOTDECK command in STATA. The HOTDECK
procedure is used several times within a multiple
2.6.2. Risk sets and sensitivity analyses imputation sequence since missing data are imputed
All analyses were conducted in the group of individuals stochastically rather than deterministically. One thousand
who successfully completed the mania items of the CIDI imputation sequences were run, yielding 1000 data sets in
interview at T2 (n = 4815). Analyses were repeated for a which the average odds ratio of the cannabis mania
risk set of individuals who at baseline had a score of association was estimated within the HOTDECK proce-
lifetime absence on all the individual items of the CIDI dure. Imputation of missing values was stratified by
mania section (n = 4623), in order to ensure that the use of known correlates of mania, namely age, sex, educational

Table 2
Associations between baseline cannabis use and follow-up manic symptoms
Cannabis exposure Any manic symptom at follow-up in whole risk set Any manic symptom at follow-up in whole risk rest,
OR a (95% Confidence Interval) adjusted b for a priori covariates OR a (95% Confidence Interval)
Baseline any use 5.32 (3.59–7.89) 2.70 (1.54–4.75)
Cumulative frequency
No use c 1 1
bonce a month 1.09 (0.27–4.51) 0.90 (0.20–4.11)
1–3 days per month 4.48 (2.01–10.00) 2.23 (0.82–6.07)
1–2 days per week 5.23 (2.54–10.73) 3.78 (1.59–8.97)
3–4 days per week 13.52 (5.39–33.90) 6.94 (2.00–24.06)
Nearly every day 9.01 (5.07–16.01) 3.43 (1.42–8.26)
Linear trend d 1.62 (1.47–1.79) 1.37 (1.17–1.59)
a
OR, odds ratio.
b
Adjusted for age, sex, educational level, ethnicity, single marital status, neuroticism, lifetime use of other drugs, alcohol use, baseline depression
and baseline mania.
c
Reference category, those subjects who did not use cannabis at baseline.
d
The increase in risk with one unit change in cannabis frequency.
 C. Henquet et al. / Journal of Affective Disorders 95 (2006) 103–110 107

level, ethnicity, single marital status, neuroticism, lifetime the whole sample and OR = 0.35, 95% CI: 0.03, 3.49 for
use of other drugs, alcohol, baseline depression and the risk set with no cannabis use at baseline, adjusted for
baseline mania. The HOTDECK procedure replaces the a priori selected variables).
missing values in the relevant variables by values ran- Based on 1000 imputation sequences in which missing
domly sampled from complete lines in the same stratum. values of CIDI follow-up mania in the whole sample were
imputed stochastically, the estimated average adjusted
3. Results association between baseline cannabis exposure and
follow-up manic symptoms remained large and statisti-
The sample consisted of 4815 subjects. The mean age cally significant (OR = 6.07, 95% CI: 3.81, 9.67).
of the sample at baseline was 41.2 years (S.D. = 11.9)
and 2573 (53.4%) were women. The rate of baseline 4. Discussion
(lifetime) mania was 192 (4.0%) and 118 (2.5%) over
the follow-up period (36 months). The rate of baseline The results suggest that use of cannabis increases the
psychosis was 784 (16.3%) and of follow-up psychosis risk for subsequent manic symptoms. This association
325 (6.8%). Use of cannabis at baseline (lifetime) was remained significant after adjustment for possible con-
admitted to by 453 subjects (9.4 %) of the risk set and by founders, such as use of other drugs and pre-existing
187 individuals (3.9%) during the follow-up. Subjects symptoms of depression and mania, both by adjustment in
with a mania outcome during follow-up had higher rates the regression equation and by sample restriction. The data
of cannabis use at baseline (Table 1). were furthermore suggestive of a dose–response relation-
Cannabis use at baseline was associated with follow- ship between frequency of exposure and mania outcome.
up mania (OR = 5.32, 95% CI: 3.59, 7.89). The The relationship between cannabis and mania was
association was reduced substantially but remained mediated by neither the prevalence nor the incidence of
significant after correction for age, sex, educational psychotic symptoms. Cannabis use during the follow-up
level, ethnicity, single marital status, neuroticism, use of period was not significantly associated with manic
other drugs, alcohol use and manic symptoms at baseline, symptoms at follow-up, suggesting that the association
with the strongest confounder being depressive symptoms between cannabis use and the incidence of manic
at baseline (OR = 2.70, 95% CI: 1.54, 4.75). After symptoms is not explained by the acute effects of
adjustment for baseline and follow-up psychotic symp- cannabis, but is more likely the result of its longer-term
toms, the association between baseline cannabis use and exposure. There was no evidence to support the self-
follow-up mania was reduced slightly (OR = 2.51, 95% medication hypothesis, as manic symptoms at baseline did
CI: 1.38, 4.59). The risk for manic symptoms increased not, directionally or statistically, predict the onset of
with increased baseline frequency of cannabis use. The cannabis use during follow-up.
effect size of cannabis use was largest in individuals using
cannabis 3 to 4 days/week and smaller in those using 4.1. Methodological issues
cannabis less frequently, apart from daily use (Table 2).
When entering both distal and recent cannabis use We studied the broader mania phenotype (i.e. sub-
together in the same model, baseline cannabis use threshold manic symptoms in the general population)
remained significantly associated with the mania outcome rather than bipolar disorder, for which the current analyses
(OR = 2.11, 95% CI: 1.06, 4.20), whereas cannabis use would be low in statistical power. Similar to subthreshold
during follow-up did not display a significant association psychotic symptoms (Johns and van Os, 2001; van Os
(OR = 1.64, 95% CI: 0.69, 3.88). et al., 2001), research suggests that expressions of mania
After exclusion of all individuals with baseline mania, outside the realm of clinical disorder have a distribution in
baseline cannabis use still predicted incident follow-up the general population (Akiskal, 2003; Angst et al., 2003a;
mania (OR = 2.86, 95% CI: 1.34, 6.09, adjusted for the a Krabbendam et al., 2004). Symptoms are more prevalent
priori selected variables and baseline and follow-up than their corresponding DSM-IV disorder-counterparts
psychosis). Similarly, in the risk set of individuals without and therefore may have greater sensitivity to detect subtle
both baseline mania and baseline psychosis, the associ- changes in mania induced by common risk factors.
ation between baseline cannabis use and follow-up mania Studying the general population and not clinical samples
remained strong (OR = 3.14, 95% CI: 1.10, 8.98, adjusted has furthermore the advantage of avoiding Berkson's bias,
for the a priori selected variables). which refers to the phenomenon that patients with
Manic symptoms at baseline did not predict cannabis comorbid bipolar and substance use disorders are more
use during follow-up (OR = 0.56, 95% CI: 0.28, 1.15 for likely to seek psychiatric treatment than patients with
108 C. Henquet et al. / Journal of Affective Disorders 95 (2006) 103–110

either condition (Strakowski and DelBello, 2000). Recent antagonism, is effective in both psychotic and manic
findings suggest that subthreshold expressions of mania disorder. There is some evidence that the catechol-O-
show continuity with clinical bipolar disorder (Kwapil et methyltransferase gene, which regulates dopamine break-
al., 2000; Lewinsohn et al., 2003; Thomas, 2004) down, is weakly associated with schizophrenia (Kunugi et
(Akiskal, 2003; Regeer et al., 2006). The association al., 1997) and also affects the rate of bipolar disorder
between cannabis and manic symptoms as described in the (Kirov et al., 1998). The association between cannabis and
current study may thus not only apply to the lower ends of both bipolar and psychotic disorder may be linked to a
the continuum, but may play a role in the expression of process commonly referred to as “sensitisation”. Sensiti-
clinical bipolar disorder as well. sation, in this case dopamine sensitisation, refers to the
process whereby repeated, intermittent stimulant exposure
4.2. Strengths and weaknesses produces a permanent change in dopaminergic responses
(Robinson and Becker, 1986; Wolf et al., 1993). A
Our measure for subthreshold mania required the dysregulated, hyperdopaminergic state may consequently
presence of manic symptoms during at least 2 days. It has lead to stimulus-independent release of dopamine which
been argued that hypomania according to DSM-IV is very may take over the normal process of contextually driven
difficult to assess in the general population (Akiskal, salience attribution. This mechanism has been suggested
2002), suggesting that the criterion of subthreshold mania in relation to the development of psychosis (Kapur, 2003;
may have yielded a significant number of false negatives Tsapakis et al., 2003), but may apply to mania as well.
as well. If this were the case however, than this would Once established, it represents a permanent change in the
result in an underestimation rather than an overestimation central nervous system, so that cannabis may be necessary
of the risk associated with use of cannabis. The issue of to initiate initial manic or psychotic vulnerability, but once
misclassification may also apply to our assessments of sensitised, the individual will display continuing symp-
drug use, as these were based on self-report and were not toms without additional substance use (Strakowski and
confirmed by urine screens. However, personal use of DelBello, 2000). Interactions between genetic vulnerabil-
cannabis is legal in the Netherlands, which makes ity and other environmental factors may further determine
underreporting unlikely, and any false negatives would if a person becomes psychotic or manic at one point in his
also have resulted in an underestimation of the actual life (Murray et al., 2004). Research focusing on
association between cannabis and mania. individuals with pre-existing vulnerability to dysregula-
One could argue that 3 years is a relative short period tion of the dopaminergic system in relation to cannabis
of follow-up. In the current analyses though, we found exposure is needed to further investigate these interactions
that the association between cannabis use and manic in the aetiology and co-morbidity of mania and psychosis.
symptoms was much stronger for the long-term effects
of cannabis exposure than for its proxy effects. The fact Acknowledgements
that such a strong association was observed after the
relatively short period of 3 years suggests that the This work was supported by a grant from the Dutch
follow-up period was adequate. Ministry of Health.

4.3. Co-morbidity between manic and psychotic

References
symptoms
Akiskal, H.S., 2002. Classification, diagnosis and boundaries of
The results indicate that cannabis is a shared, but bipolar disorder. In: Maj, M., Akiskal, H.S., Lopez-Ibor, J.J.,
independent risk factor for psychosis and mania, suggest- Sartorius, N. (Eds.), Bipolar Disorder. Wiley, London, pp. 1–52.
ing that cannabis may play a role in the co-morbidity of Akiskal, H.S., 2003. Validating ‘hard’ and ‘soft’ phenotypes within the
both symptom dimensions and clinical disorders. It has bipolar spectrum: continuity or discontinuity? J. Affect. Disord. 73,
1–5.
been suggested that both disorders may share a genetic Andreasson, S., Allebeck, P., Engstrom, A., Rydberg, U., 1987.
vulnerability to dysregulation of the dopaminergic system Cannabis and schizophrenia. A longitudinal study of Swedish
due to social or pharmacological stress, for example conscripts. Lancet 2, 1483–1486.
induced by cannabis (Murray et al., 2004). It is generally Angst, J., Gamma, A., Benazzi, F., Ajdacic, V., Eich, D., Rossler, W.,
known that acute symptoms associated with both 2003a. Toward a re-definition of subthreshold bipolarity: epide-
miology and proposed criteria for bipolar-II, minor bipolar
psychosis and mania can be induced by dopamine-relea- disorders and hypomania. J. Affect. Disord. 73, 133–146.
sing drugs such as cannabis, amphetamine and cocaine. Angst, J., Gamma, A., Endrass, J., 2003b. Risk factors for the bipolar
Antipsychotic medication, through dopamine receptor and depression spectra. Acta Psychiatr. Scand., Suppl. 15–19.
 C. Henquet et al. / Journal of Affective Disorders 95 (2006) 103–110 109

Bijl, R.V., Ravelli, A., van Zessen, G., 1998a. Prevalence of Kwapil, T.R., Miller, M.B., Zinser, M.C., Chapman, L.J., Chapman, J.,
psychiatric disorder in the general population: results of The Eckblad, M., 2000. A longitudinal study of high scorers on the
Netherlands Mental Health Survey and Incidence Study (NEME- hypomanic personality scale. J. Abnorm. Psychology 109, 222–226.
SIS). Soc. Psychiatry Psychiatr. Epidemiol. 33, 587–595. Lewinsohn, P.M., Seeley, J.R., Klein, D.N., 2003. Bipolar disorders
Bijl, R.V., van Zessen, G., Ravelli, A., de Rijk, C., Langendoen, Y., during adolescence. Acta Psychiatr. Scand., Suppl. 47–50.
1998b. The Netherlands Mental Health Survey and Incidence McGorry, P.D., Bell, R.C., Dudgeon, P.L., Jackson, H.J., 1998. The
Study (NEMESIS): objectives and design. Soc. Psychiatry dimensional structure of first episode psychosis: an exploratory
Psychiatr. Epidemiol. 33, 581–586. factor analysis. Psychol. Med. 28, 935–947.
Browne, R., Byrne, M., Mulryan, N., Scully, A., Morris, M., Kinsella, Mortensen, P.B., Pedersen, C.B., Melbye, M., Mors, O., Ewald, H.,
A., McNeil, T.F., Walsh, D., O'Callaghan, E., 2000. Labour and 2003. Individual and familial risk factors for bipolar affective
delivery complications at birth and later mania. An Irish case disorders in Denmark. Arch. Gen. Psychiatry 60, 1209–1215.
register study. Br. J. Psychiatry 176, 369–372. Murray, R.M., Sham, P., Van Os, J., Zanelli, J., Cannon, M., McDonald,
Cardno, A.G., Rijsdijk, F.V., Sham, P.C., Murray, R.M., McGuffin, P., C., 2004. A developmental model for similarities and dissimilarities
2002. A twin study of genetic relationships between psychotic between schizophrenia and bipolar disorder. Schizophr. Res. 71,
symptoms. Am. J. Psychiatry 159, 539–545. 405–416.
Caspi, A., Moffitt, T.E., Cannon, M., McClay, J., Murray, R., Harrington, Peralta, V., Cuesta, M.J., 1999. Dimensional structure of psychotic
H., Taylor, A., Arseneault, L., Williams, B., Braithwaite, A., Poulton, symptoms: an item-level analysis of SAPS and SANS symptoms in
R., Craig, I.W., 2005. Moderation of the effect of adolescent-onset psychotic disorders. Schizophr. Res. 38, 13–26.
cannabis use on adult psychosis by a functional polymorphism in the Regeer, E.J., Krabbendam, L., DE Graaf, R., Ten Have, M., Nolen, W.A.,
catechol-O-methyltransferase gene: longitudinal evidence of a gene Van Os, J., 2006. A prospective study of the transition rates of sub-
X environment interaction. Biol. Psychiatry 57, 1117–1127. threshold (hypo)mania and depression in the general population.
Cottler, L.B., Helzer, J.E., Mager, D., Spitznagel, E.L., Compton, W. Psychol. Med. 1–9.
M., 1991. Agreement between DSM-III and III-R substance use Regier, D.A., Farmer, M.E., Rae, D.S., Locke, B.Z., Keith, S.J., Judd,
disorders. Drug Alcohol Depend. 29, 17–25. L.L., Goodwin, F.K., 1990. Comorbidity of mental disorders with
Escamilla, M.A., Batki, S., Reus, V.I., Spesny, M., Molina, J., Service, alcohol and other drug abuse. Results from the Epidemiologic
S., Vinogradov, S., Neylan, T., Mathews, C., Meza, L., Gallegos, Catchment Area (ECA) Study. Jama 264, 2511–2518.
A., Montero, A.P., Cruz, M.L., Neuhaus, J., Roche, E., Smith, L., Robinson, T.E., Becker, J.B., 1986. Enduring changes in brain and
Leon, P., Freimer, N.B., 2002. Comorbidity of bipolar disorder and behavior produced by chronic amphetamine administration: a
substance abuse in Costa Rica: pedigree- and population-based review and evaluation of animal models of amphetamine
studies. J. Affect. Disord. 71, 71–83. psychosis. Brain Res. 396, 157–198.
Grinspoon, L., Bakalar, J.B., 1998. The use of cannabis as a mood Sherwood Brown, E., Suppes, T., Adinoff, B., Rajan Thomas, N.,
stabilizer in bipolar disorder: anecdotal evidence and the need for 2001. Drug abuse and bipolar disorder: comorbidity or misdiag-
clinical research. J. Psychoact. Drugs 30, 171–177. nosis? J. Affect. Disord. 65, 105–115.
Henquet, C., Krabbendam, L., Spauwen, J., Kaplan, C., Lieb, R., Smeets, R.M.W., Dingemans, P.M.A.J., 1993. Composite International
Wittchen, H.U., van Os, J., 2005. Prospective cohort study of Diagnostic Interview (CIDI) version 1.1. World Health Organiza-
cannabis use, predisposition for psychosis, and psychotic symp- tion, Amsterdam/Geneva.
toms in young people. BMJ 330, 11–15. Sonne, S.C., Brady, K.T., Morton, W.A., 1994. Substance abuse and
Johns, L.C., van Os, J., 2001. The continuity of psychotic experiences bipolar affective disorder. J. Nerv. Ment. Dis. 182, 349–352.
in the general population. Clin. Psychol. Rev. 21, 1125–1141. Stata Corporation, 2003. STATA Statistical Software, Release 8.0.
Kapur, S., 2003. Psychosis as a state of aberrant salience: a framework College Station, Texas.
linking biology, phenomenology, and pharmacology in schizo- Stefanis, N.C., Hanssen, M., Smirnis, N.K., Avramopoulos, D.A.,
phrenia. Am. J. Psychiatry 160, 13–23. Evdokimidis, I.K., Stefanis, C.N., Verdoux, H., Van, O.J., 2002.
Khantzian, E.J., 1997. The self-medication hypothesis of substance Evidence that three dimensions of psychosis have a distribution in
use disorders: a reconsideration and recent applications. Harv. Rev. the general population. Psychol. Med. 32, 347–358.
Psychiatry 4, 231–244. Strakowski, S.M., DelBello, M.P., 2000. The co-occurrence of bipolar
Kirov, G., Murphy, K.C., Arranz, M.J., Jones, I., McCandles, F., Kunugi, and substance use disorders. Clin. Psychol. Rev. 20, 191–206.
H., Murray, R.M., McGuffin, P., Collier, D.A., Owen, M.J., Strakowski, S.M., Sax, K.W., McElroy, S.L., Keck Jr., P.E., Hawkins,
Craddock, N., 1998. Low activity allele of catechol-O-methyltrans- J.M., West, S.A., 1998. Course of psychiatric and substance abuse
ferase gene associated with rapid cycling bipolar disorder. Mol. syndromes co-occurring with bipolar disorder after a first
Psychiatry 3, 342–345. psychiatric hospitalization. J. Clin. Psychiatry 59, 465–471.
Krabbendam, L., Janssen, I., Bak, M., Bijl, R.V., de Graaf, R., van Os, Thomas, P., 2004. The many forms of bipolar disorder: a modern look
J., 2002. Neuroticism and low self-esteem as risk factors for at an old illness. J. Affect. Disord. 79 (Suppl 1), S3–S8.
psychosis. Soc. Psychiatry Psychiatr. Epidemiol. 37, 1–6. Tohen, M., Waternaux, C.M., Tsuang, M.T., 1990. Outcome in Mania.
Krabbendam, L., Myin-Germeys, I., De Graaf, R., Vollebergh, W., A 4-year prospective follow-up of 75 patients utilizing survival
Nolen, W.A., Iedema, J., Van Os, J., 2004. Dimensions of depression, analysis. Arch. Gen. Psychiatry 47, 1106–1111.
mania and psychosis in the general population. Psychol. Med. 34, Tsapakis, E.M., Guillin, O., Murray, R., 2003. Does dopamine
1177–1186. sensitization underlie the association between schizophrenia and
Kunugi, H., Vallada, H.P., Sham, P.C., Hoda, F., Arranz, M.J., Li, T., drug abuse? Curr. Opin. Psychiatry 16 (Suppl 2), S45–S52.
Nanko, S., Murray, R.M., McGuffin, P., Owen, M., Gill, M., Valles, V., Van Os, J., Guillamat, R., Gutierrez, B., Campillo, M.,
Collier, D.A., 1997. Catechol-O-methyltransferase polymorphisms Gento, P., Fananas, L., 2000. Increased morbid risk for schizo-
and schizophrenia: a transmission disequilibrium study in multiply phrenia in families of in-patients with bipolar illness. Schizophr.
affected families. Psychiatr. Genet. 7, 97–101. Res. 42, 83–90.
110 C. Henquet et al. / Journal of Affective Disorders 95 (2006) 103–110

van Os, J., Jones, P.B., 2001. Neuroticism as a risk factor for Walker, J., Curtis, V., Murray, R.M., 2002. Schizophrenia and bipolar
schizophrenia. Psychol. Med. 31, 1129–1134. disorder: similarities in pathogenic mechanisms but differences in
van Os, J., Takei, N., Castle, D.J., Wessely, S., Der, G., MacDonald, A.M., neurodevelopment. Int. Clin. Psychopharmacol. 17 (Suppl 3),
Murray, R.M., 1996. The incidence of mania: time trends in relation to S11–S19.
gender and ethnicity. Soc. Psychiatry Psychiatr. Epidemiol. 31, Wittchen, H.U., 1994. Reliability and validity studies of the WHO-
129–136. Composite International Diagnostic Interview (CIDI): a critical
van Os, J., Hanssen, M., Bijl, R.V., Ravelli, A., 2000. Strauss (1969) review. J. Psychiatr. Res. 28, 57–84.
revisited: a psychosis continuum in the general population? Wittchen, H.U., Robins, L.N., Cottler, L.B., Sartorius, N., Burke, J.D.,
Schizophr. Res. 45, 11–20. Regier, D., 1991. Cross-cultural feasibility, reliability and sources
van Os, J., Hanssen, M., Bijl, R.V., Vollebergh, W., 2001. Prevalence of variance of the Composite International Diagnostic Interview
of psychotic disorder and community level of psychotic symptoms: (CIDI). The Multicentre WHO/ADAMHA Field Trials. Br. J.
an urban–rural comparison. Arch. Gen. Psychiatry 58, 663–668. Psychiatry 159, 645–653.
van Os, J., Bak, M., Hanssen, M., Bijl, R.V., de Graaf, R., Verdoux, H., Wolf, M.E., White, F.J., Nassar, R., Brooderson, R.J., Khansa, M.R.,
2002. Cannabis use and psychosis: a longitudinal population-based 1993. Differential development of autoreceptor subsensitivity and
study. Am. J. Epidemiol. 156, 319–327. enhanced dopamine release during amphetamine sensitization.
Verdoux, H., Sorbara, F., Gindre, C., Swendsen, J.D., van Os, J., 2003. J. Pharmacol. Exp. Ther. 264, 249–255.
Cannabis use and dimensions of psychosis in a nonclinical
population of female subjects. Schizophr. Res. 59, 77–84.