Secondary Schizophrenia (68 page)

through the dopaminergic system. More

These possibilities have not always been distin-biological research on these mechanisms is
guished, and, until recently, our ability to decide
necessary.

between them has been hampered by a lack of sophistication in research design
[3, 7, 8, 9, 10].
Because
there is currently compelling evidence that cannabis

use and psychosis are associated, the debate now con-There are good reasons for suspecting that regu-cerns whether cannabis use is or is not a contribu-lar cannabis use may be a contributory cause of
tory cause of psychosis and indeed, what we mean by
schizophrenia. First, tetrahydrocannabinol (THC), the
“cause”
[10].

169

Organic Syndromes of Schizophrenia – Section 3

Cannabis use as a precipitant

5.5) for those who had used cannabis between 1 and
of psychosis

50 times, and 6.0 (4.0, 8.9) for those who had used
cannabis more than 50 times. The size of these risks
There is consistent evidence that cannabis use occurs
was reduced by adjustment for confounding variables
at a much higher rate among patients with schizophre-that were related to the risk of developing schizophrenia than among their age peers in the community
nia (namely, having a psychiatric diagnosis at con-

[11].
There is also good epidemiological evidence for
scription and having parents who had divorced). Nev-an association between schizophrenia and drug abuse
ertheless, the relationship between cannabis use and
and dependence in the community. The Epidemio-schizophrenia remained statistically significant and
logical Catchment Area (ECA) study from the USA
still showed a dose-response relationship.

[12]
found that the lifetime prevalence of drug abuse
Andreasson and colleagues
[15]
and Allebeck
[16]

and dependence was 6.2% (7.7% among men and
argued from the earlier study that cannabis use precip-4.8% among women), with abuse and dependence
itated schizophrenia in vulnerable individuals, reject-on cannabis the most common form of drug abuse
ing as implausible the hypothesis that cannabis use was
and dependence, with a lifetime prevalence of 4.4%.

a consequence of emerging schizophrenia. First, the
The National Survey of Mental Health and Wellbeing
cannabis users who developed schizophrenia had bet-

(NSMHWB) conducted in Australia in 1997 included a
ter premorbid personalities, a more abrupt onset, and
screening questionnaire for psychotic symptoms
[13].

more positive symptoms than the nonusers who devel-A diagnosis of cannabis dependence increased the
oped schizophrenia
[16].
Second, although more than
chances of reporting psychotic symptoms 1.7 times,
half of the heavy cannabis users (58%) had a psychi-after adjusting for age, affective and anxiety disorders,
atric diagnosis at the time of conscription, there was
smoking status, and alcohol dependence
[14].
In the
still a dose-response relationship between cannabis use
NSMHWB, 11.5% of those who reported that they had
and risk of schizophrenia among conscripts who had
been diagnosed with schizophrenia met ICD-10 crite-psychiatric symptoms at baseline.

ria for a cannabis use disorder in the past 12 months.

Critics suggested a number of alternative explana-After adjusting for confounding variables, those who
tions for the findings by Andreasson and colleagues
met criteria for cannabis dependence were 2.9 times
[15]
. First, there was a long delay between self-reported
more likely to report that they had been diagnosed
cannabis use – at age 18–20 – and the diagnosis of
with schizophrenia than those who did not.

schizophrenia – over the next 15 years
[17, 18],
and
it was not clear that these individuals continued to
use cannabis up until the time that schizophrenia
Longitudinal studies

was diagnosed. Andreasson and colleagues showed
The first convincing evidence that cannabis use may
that self-reported cannabis use at age 18 was strongly
precipitate schizophrenia was provided by a study
related to the risk of attracting a diagnosis of drug
of Swedish conscripts
[15].
These investigators con-abuse, suggesting that cannabis use at age 18 was preducted a 15-year prospective study of 50,465 Swedish
dictive of continued drug use.

conscripts to investigate the relationship between self-A second explanation was that the excessive rate
reported cannabis use at age 18 and the risk of
of “schizophrenia” among the heavy cannabis users
receiving a psychiatric diagnosis of schizophrenia in
was due to cannabis-induced schizophreniform psy-the subsequent period, as indicated by inclusion in
choses
[17, 18].
Andreasson and colleagues addressed
the Swedish psychiatric case register. Their results
this criticism by a small study of the validity of the
showed that the relative risk of receiving a diagno-schizophrenia diagnoses, which suggested that 80%

sis of schizophrenia was 2.4 (95% confidence inter-of those so diagnosed met DSM-III criteria, among
val [1.8, 3.3]) for those who had ever tried cannabis
which is a requirement of at least 6 months’ duration
compared with those who had not. There was also a
that would exclude transient drug-induced psychoses.

dose-response relationship between the risk of a diag-Third, the relationship between cannabis use
nosis of schizophrenia and the number of times that
and schizophrenia may have been confounded
the conscript had tried cannabis by age 18. The relative
by the use of other psychoactive drugs. Frequent
risk of developing schizophrenia was 1.3 (0.8, 2.3) for
cannabis use in late adolescence predicts the later
170

those who had used cannabis 1 to 10 times, 3.0 (1.6,
use of amphetamine and cocaine
[19, 20]
which can
Chapter 11 – Schizophrenia secondary to cannabis use

produce paranoid psychoses
[21, 22, 23, 24, 25].

wards. The relationship was a little stronger in cases
The major illicit drug of abuse in Sweden during
observed in the first 5 years, probably reflecting the
the study period
[26, 27, 28],
was amphetamine,
decline in cannabis use with age.

making it difficult to exclude the hypothesis that
Zammit’s findings have been supported by van Os
amphetamine rather than cannabis explained the
and colleagues
[30]
in a 3-year longitudinal study of
association between cannabis use and schizophrenia.

the relationship between self-reported cannabis use
A fourth criticism was that Andreasson and col-and psychosis in a community sample of 4,848 peo-leagues had not ruled out the possibility that cannabis
ple in the Netherlands. Subjects were assessed at base-use at age 18 was a symptom of emerging schizophre-line on cannabis and other drug use and psychotic
nia. Statistical adjustment for a psychiatric diagno-symptoms were assessed using a computerized diag-sis at conscription did not eliminate the relationship
nostic interview. A diagnosis of psychosis was vali-between cannabis use and schizophrenia, but it sub-dated in positive cases by a diagnostic telephone inter-stantially reduced its size, because more than half of
view with a psychiatrist or psychologist. A consensus
the heavy cannabis users had received a psychiatric
clinical judgement was made on the basis of the inter-diagnosis by age 18.

view material as to whether individuals had a psychotic
A number of longitudinal studies have since been
disorder for which they were in need of psychiatric
reported that have supported the findings of the
care.

Andreassen and colleagues study and addressed many
Van Os and colleagues substantially replicated the
of the criticisms made of their study. Zammit and
findings of the Swedish cohort and extended them in a
colleagues
[29]
reported a 27-year follow-up of the
number of important ways. First, cannabis use at base-Swedish cohort that covered most of the risk period
line predicted an increased risk of psychotic symptoms
for the onset of psychotic disorders in the cohort. This
during the follow-up period in individuals who had
study improved on the Andreassen and colleagues
[15]

not reported psychiatric symptoms at baseline. Sec-study in a number of ways. The psychiatric register
ond, there was a dose-response relationship between
provided a more complete coverage of all cases diag-frequency of cannabis use at baseline and risk of psy-nosed with schizophrenia; there was better statisti-chotic symptoms during follow up. Third, the rela-cal control of more potential confounding variables,
tionship between cannabis use and psychotic symp-including other drug use, IQ, and known risk factors
toms persisted when they statistically controlled for
for schizophrenia and social integration; the study dis-the effects of other drug use. Fourth, the relationship
tinguished between cases in the first 5 years of the
between cannabis use and psychotic symptoms was
study and those that occurred more than 5 years after-stronger for persons with more severe psychotic symp-wards in order to look at the possible role of a pro-toms who were judged to be in need of psychiatric
drome; and the study undertook separate analyses in
care. Van Os and colleagues estimated the attributable
those who only reported using cannabis at the initial
risk of cannabis to psychosis was 13% for psychotic
assessment.

symptoms and 50% for cases with psychotic disor-Zammit and colleagues found that cannabis use
ders in need of treatment. Fifth, those who reported
at age 18 predicted an increased risk of schizophre-any psychotic symptoms at baseline were more likely
nia during the follow-up period, with a dose-response
to develop schizophrenia if they used cannabis than
relationship between frequency of cannabis use at age
were individuals who were not so vulnerable. They
18 and risk of being diagnosed with schizophrenia dur-estimated that cannabis use accounted for 80% of the
ing the follow-up. The relationship between cannabis
increased risk of developing a psychotic disorder that
use and schizophrenia persisted after statistically con-warranted treatment among vulnerable individuals.

trolling for the effects of other drug use and other
A 1-year follow-up of a cohort of 2,437 adoles-potential confounding factors, including psychiatric
cents and young adults between 1995 and 1999 in
symptoms at age 18. They estimated that 13% of cases
Munich
[31]
has substantially replicated the Swedish
of schizophrenia could be averted if all cannabis use
and Dutch studies. Subjects were assessed at base-were prevented. The same relationships were observed
line on cannabis use and psychotic symptoms using
in the subset of the sample who only reported cannabis
a questionnaire; psychotic symptoms were assessed
use at baseline and among cases diagnosed in the first
in early adulthood using the Computerized Interna-

171

5 years after assessment and for the 22 years after-tional Diagnostic Interview. Henquet and colleagues
Organic Syndromes of Schizophrenia – Section 3

[31]
found a dose-response relationship between self-tional polymorphism of the COMT gene that codes
reported cannabis use at baseline and the likelihood
for dopamine. They found that the 25% of the cohort
of reporting psychotic symptoms. As in the Dutch
who were homozygous for a polymorphism and used
cohort, young people who reported psychotic symp-cannabis were 10.9 times more likely to have devel-toms at baseline were much more likely to experi-oped a schizophreniform disorder than peers with the
ence psychotic symptoms at follow-up if they used
same polymorphism who did not use cannabis. In the
cannabis than were peers who did not have such a
absence of this polymorphism, young adults who used
history.

cannabis did not seem to be at any increased risk of a
Arseneault and colleagues
[32]
reported a prospec-psychosis.

tive study of the relationship between adolescent
Fergusson, Horwood, and Swain-Campbell
[34]

cannabis use and psychosis in young adults in a New
have reported a longitudinal study of the relation-Zealand birth cohort (N
=
759) whose members had
ship between cannabis dependence at age 18 and the
been assessed intensively on risk factors for psychotic
number of psychotic symptoms reported at age 21 in
symptoms and disorders since birth. Psychotic disor-the Christchurch birth cohort in New Zealand. They
ders were assessed according to DSM-IV diagnostic
assessed cannabis dependence using DSM-IV criteria
criteria, with corroborative reports from family mem-and psychotic symptoms were assessed by 10 items
bers or friends on social adjustment. They assessed
from the SCL-90. Because this was a birth cohort that
psychotic symptoms at age 11 before the onset of
had been assessed throughout childhood and adoles-cannabis use and distinguished between early and late
cence, Fergusson and colleagues were able to adjust
onset cannabis use. They also examined the speci-for a large number of potential confounding variables,
ficity of the association between cannabis use and psy-including self-reported psychotic symptoms at the pre-chosis by conducting analyses of the effects of: i) other
vious assessment, other drug use, and other psychiatric
drug use on psychotic symptoms and disorders; and
disorders. They found that cannabis dependence at age
ii) cannabis use on depressive disorders.