Transient cannabis-induced psychotic symptoms
It is possible that cannabis use might temporarily trigger some symptoms of psychosis among some users. Such symptoms are clinically (and significantly) distinct from a psychotic disorder such as schizophrenia.
Other drugs such as amphetamine have also been shown to have the potential to trigger psychotic symptoms among some users 
. Double-blind provocation studies using intravenous THC and related compounds in healthy controls are providing insights into the neurobiological correlates of cannabis-related transient psychotic symptoms and neuro-cognitive impairments 
Several cross-sectional studies have examined the relationship between cannabis use and self-reported psychotic experiences or psychotic symptoms in the general population. All have found that cannabis use (or cannabis use disorders) were more common among people reporting such experiences; and these associations persisted after controlling for other variables 
. Although these findings provide important clues to the mechanisms of action linking cannabis use and persistent psychotic symptoms and/or clinical diagnoses, these outcomes are less of a concern for the research community.
It is not always clear whether the psychotic symptoms endorsed in studies assessing the relationship between cannabis use and “psychosis” occurred only in the context of cannabis intoxication, or whether the symptoms were a more distal outcome of previous cannabis use. For example, the Fergusson et al. study 
assessed the relationship between psychotic symptoms in the past month with cannabis use in the past year. It remains possible that the psychotic symptoms endorsed may have been experienced only while intoxicated. The instruments used to measure psychotic outcomes in the Arsenault et al. 
, Henquet et al. 
, and van Os et al. 
studies contain instructions not to include psychotic symptoms that only occur in the context of intoxication. Further, the authors of the Swedish conscript study 
maintain that it is unlikely that substance-induced intoxication would have been misdiagnosed as schizophrenia. We turn now to the evidence relating to more persistent symptoms or disorders.
Schizophrenia and other psychotic disorders
In case-control studies 
, patients with schizophrenia are more likely to use cannabis than other psychiatric patients or normal controls 
. The prevalence of use in patients with schizophrenia has varied between studies but it is generally higher than rates in the general population 
Cross-sectional community surveys of psychiatric disorders have also documented higher rates of substance use disorders among persons with schizophrenia 
. Nearly half of the patients identified with schizophrenia in the US ECA study had a diagnosis of substance abuse or dependence (28% for an illicit drug disorder) 
. In an Australian population-based survey, 11.5% of those who reported that they had been diagnosed with schizophrenia met ICD-10 criteria for a cannabis use disorder in the past 12 mo, and 21.2% met criteria for an alcohol use disorder. After adjusting for confounding variables, those who met criteria for cannabis dependence were 2.9 times more likely to report that they had been diagnosed with schizophrenia than those who did not 
The first evidence that cannabis use may precipitate schizophrenia came from a 15-y prospective study of cannabis use and schizophrenia in 50,465 Swedish conscripts 
. This study investigated the relationship between self-reported cannabis use at age 18 y and the risk of being diagnosed with schizophrenia in the Swedish psychiatric case register during the next 15 y. Those who had tried cannabis by age 18 y were 2.4 times more likely to receive a diagnosis of schizophrenia than those who had not. The risk of a diagnosis of schizophrenia was related to cannabis use in a dose-response way to the number of times cannabis had been used by age 18. Compared to those who had not used cannabis, the risk of developing schizophrenia was 1.3 times higher for those who had used cannabis one to ten times, three times higher for those who had used cannabis between one and 50 times, and six times higher for those who had used cannabis more than 50 times. These results remained after statistical adjustment for two variables that were related to the risk of developing schizophrenia (personal history of psychiatric disorder and parental divorce).
A number of longitudinal studies have since been reported that have all supported the findings of the Andreassen et al. study. Zammit et al. reported a follow up of the Swedish cohort study, reporting on risk over a 27-y follow up that covers most of the risk period for the onset of psychotic disorders in a cohort that was first studied when 18–20 y old 
. This study improved on the earlier study in a number of ways. The psychiatric register provided more complete coverage of all cases diagnosed with schizophrenia; there was better statistical control of a larger number of potential confounding variables, including other drug use, IQ, known risk factors for schizophrenia, and social integration; the study distinguished between cases that occurred in the first 5 y of the study period and those that occurred more than 5 y afterwards in order to look at the possible role of a syndrome; and the study undertook separate analyses in those who only reported using cannabis at the initial assessment.
Zammit et al. 
also found cannabis use at baseline predicted an increased risk of schizophrenia during the follow-up period. There was a dose-response relationship with frequency of use, which persisted after statistical control for confounders, including a history of psychiatric symptoms at baseline. The same relationships were observed in the subset of the sample who only reported cannabis use at baseline and among cases diagnosed in the first 5 y after assessment and for the subsequent 22 y.
Zammit et al.'s findings were consistent with those of a study conducted by Van Os and colleagues 
. This was a 3-y longitudinal study of the relationship between self-reported cannabis use and psychosis in a community sample of 4,848 people in the Netherlands. Participants were assessed at baseline on cannabis and other drug use. Psychotic symptoms were assessed using a computerised diagnostic interview. A diagnosis of psychosis was validated in positive cases by a diagnostic telephone interview with a psychiatrist or psychologist. A consensus clinical judgement was made on the basis of the interview material as to whether individuals had a psychotic disorder for which they were in need of psychiatric care.
Van Os et al. replicated and extended the findings of the Swedish cohort in a number of important ways. First, cannabis use at baseline predicted an increased risk of psychotic symptoms during the follow-up period in individuals who had not reported psychiatric symptoms at baseline. Second, there was a dose-response relationship between frequency of cannabis use at baseline and risk of psychotic symptoms during the follow-up period. Third, the relationship between cannabis use and psychotic symptoms persisted when they statistically controlled for the effects of other drug use. Fourth, the relationship between cannabis use and psychotic symptoms was stronger for cases with more severe psychotic symptoms that were adjudged to need psychiatric care. Fifth, those who reported any psychotic symptoms at baseline were more likely to develop schizophrenia if they used cannabis than were individuals who were not so vulnerable.
A study by Henquet et al. 
replicated the Swedish and Dutch studies in a 4-y follow up of a cohort of 2,437 adolescents and young adults between 1995 and 1999 in Munich. Their participants were assessed at baseline on cannabis use and psychotic symptoms using a questionnaire. Psychotic symptoms were assessed in early adulthood using the Composite International Diagnostic Interview. They found a dose-response relationship between self-reported cannabis use at baseline and the likelihood of reporting psychotic symptoms. As in the Dutch cohort, young people who reported psychotic symptoms at baseline were much more likely to experience psychotic symptoms at follow up if they used cannabis than were peers who did not have such a history.
Arseneault et al. reported a prospective study of the relationship between adolescent cannabis use and psychosis in young adults in a New Zealand birth cohort (n
759). Participants were assessed intensively on risk factors for psychotic symptoms and disorders since birth 
, and psychotic disorders were conservatively assessed according to DSM-IV diagnostic criteria, with corroborative reports from family members or friends on social adjustment. They assessed psychotic symptoms at age 11 y before onset of cannabis use and distinguished between early and late onset of cannabis use. They also examined the specificity of the association between cannabis use and psychosis by conducting analyses of the effects of: (1) other drug use on psychotic symptoms and disorders; and (2) cannabis use on depressive disorders.
Arseneault et al. found a relationship between cannabis use by age 15 y and an increased risk of schizophreniform disorder by age 26 y. Controlling for other drug use did not affect the relationship. The relationship was no longer statistically significant after adjustment for reporting psychotic symptoms at age 11 y, which probably reflected the small number of psychotic disorders observed in the sample. The small number of cases also limited the ability of the study to examine predictors of psychotic disorders at age 26 y. The measurement of cannabis and other drug use was crude (viz, none, 1–2 times, and 3 or more times), although this was more likely to work against finding relationships.
There was also specificity in the effects of cannabis on schizophreniform disorder: there was no relationship between other drug use and psychotic disorders, and no relationship between cannabis use and depression. There was also an interaction between psychosis risk and age of onset of cannabis use, with earlier onset being more strongly related to psychosis. There was also the suggestion of an interaction between cannabis use and vulnerability, with a higher risk of psychosis among cannabis users who reported psychotic symptoms at age 11 y.
Caspi and colleagues subsequently used the cohort to examine an interaction between cannabis use and a functional polymorphism of the COMT
gene that codes for dopamine in their effects on the risk of psychosis 
. They found that the 25% of the cohort who were homozygous for the polymorphism and used cannabis were 10.9 times more likely to have developed a schizophreniform disorder than peers with the same polymorphism who did not use cannabis. In the absence of this polymorphism, young adults who used cannabis were not at any increased risk of psychosis.
Apart from clinical diagnoses, several longitudinal studies have also examined the relationship between cannabis use and subclinical (or isolated) psychotic symptoms. Fergusson, Horwood, and Swain-Campbell have reported a longitudinal study of the relationship between cannabis dependence at age 18 y and the number of psychotic symptoms reported at age 21 y in the Christchurch birth cohort in New Zealand 
. They assessed cannabis dependence using DSM-IV criteria and psychotic symptoms were assessed by ten items from the SCL-90. Because this was a birth cohort that had been assessed throughout childhood and adolescence Fergusson et al. were able to adjust for a large number of potential confounding variables, including self-reported psychotic symptoms at the previous assessment, other drug use, and other psychiatric disorders. They found that cannabis dependence at age 18 y predicted an increased risk of psychotic symptoms at age 21 y (relative risk [RR] of 2.3). This association was smaller but still significant after adjustment for potential confounders (RR of 1.8). More recently, Fergusson and colleagues examined the association between cannabis and psychotic symptoms until age 25 y with the same cohort of young adults, using a more sophisticated structural equations modelling design that accounted for both observed and nonobserved confounding factors 
. As with their earlier study, they concluded that the association between cannabis and psychosis did not appear to be explained by confounding factors, and that the direction of the association appeared to be from cannabis use to symptoms of psychosis rather than vice versa.
One study of high risk young people has failed to report an association between cannabis use and psychosis risk. This study identified 100 young people at “ultra high” risk for psychosis 
because of family history or prodromal symptoms of psychosis (on the basis of one or more of the following: schizophrenia in a first degree relative; the presence of attenuated psychotic symptoms; or a brief limited psychosis) in whom 18% reported symptoms of cannabis dependence in the past year. They assessed whether cannabis users were more likely to develop psychosis in the following year, but did not find any association, regardless of the frequency of cannabis use.
Increasingly, researchers in the field of psychosis are examining the concept of psychotic spectrum features as risk factors for psychosis 
. Recent work has found that these symptoms are common in the general population distribution and can persist over a 20-y period. Two major trajectories have been identified: persistent “schizophrenic nuclear symptoms” (which resemble psychosis) and persistent “schizotypal symptoms” (more closely resembling schizotypal personality disorder) 
. Cannabis use during adolescence has been found to be associated with “high load” schizophrenia nuclear symptoms during adulthood—but not so for the schizotypal symptom cluster. More frequent cannabis use was more strongly associated with persistent high load symptoms for the entire follow-up period. These findings suggest that there may be different aetiological dimensions for these two symptom dimensions, with an interaction between biological vulnerability and unique psychosocial risk factors for each symptom cluster; limitations of the study included the small number of cases, the use of open ended interviews, and the use of multiple analyses.