Cannabis is a drug used widely for “recreational” purposes, with over 75% of New Zealanders have reported using cannabis at least once by age 25, and with 12.5% showing use consistent with the Diagnostic and Statistical Manual of Mental Disorders (DSM)[[1]] criteria for dependence.[[2]] The active components of cannabis are cannabinoids, with two cannabinoids having the highest concentration: delta-9-tetrahydrocannabinol (THC) and cannabidiol (CBD).[[3]] THC is responsible for most of the desirable and less desirable effects of cannabis on anxiety, euphoria, perception and memory.[[4]] The average THC potency in cannabis has increased over the last two decades from 4% THC in 1994 to 12% THC in 2014, whereas the potency of CBD has decreased from 0.5% in 2004 to <0.2% in 2014.[[5]]

Cannabis intoxication can cause a drug-induced psychosis-like state which is short lived and resolves after a period of abstinence.[[6]] However, there is evidence that regular cannabis use may be associated with an increased risk of psychotic symptoms when the individual is no longer intoxicated.[[7]] Psychotic symptomatology is a defining characteristic of schizophrenia spectrum disorders and can result in significant periods of disability.[[8]] The defining positive symptoms of psychosis are hallucinations; perception-like experiences that occur without external stimuli; and delusions, fixed beliefs that are not amendable despite conflicting evidence.[[9]] In a sample of patients treated for cannabis-induced psychosis, Arendt and colleagues[[10]] reported that 45% of patients were later diagnosed with schizophrenia spectrum disorder, and that this occurred at an earlier average age than patients who did not use cannabis. The landmark study by Andreasson and colleagues,[[11]] who examined over 45,000 male participants in Sweden, found that even after adjustment there was dose-response relationship between cannabis exposure and the development of schizophrenia spectrum disorder.

Further evidence suggests that adolescence appears to be a particularly vulnerable period for the relationship between cannabis use and increased risk of developing psychosis.[[12]] Evidence suggests that regular cannabis use in adolescence may disrupt the normal development of the pre-frontal cortex and dopaminergic system to increase the risk of developing schizophrenia or psychosis.[[12]] THC exposure in susceptible users may lead to permanent alterations in neurotransmitter functions, leading to psychotic illness in the long-term.[[13]]

While the associations between cannabis and psychosis are relatively well-established, the question of a causal conclusion is widely debated.[[14]] There are several alternative models, implicating social, demographic and genetic factors. One alternative model is that those with a genetic, or other predisposition to psychosis, may either self-medicate their prodromal symptoms with cannabis or may hasten the onset of a psychotic illness to which they are predisposed. Caspi and colleagues[[15]] found that the association between early cannabis use and psychosis was limited to those with a particular genotype (the Val/Val variant of the COMT gene, involved in dopamine regulation). Furthermore, Henquet et al[[16]] found that baseline cannabis use predicted an increase in the risk for later psychotic symptoms, and that this association was stronger for those predisposed to psychosis. Finally, a recent study by Wainberg and colleagues,[[17]] using a polygenic risk score (PRS) for schizophrenia, found a much stronger link between the use of cannabis and later psychotic symptomatology in those with high schizophrenia PRS than in those with low schizophrenia PRS (67% with psychosis vs 7%).

The present study used data from the Christchurch Health and Development Study (CHDS). The CHDS has studied a cohort of over 1,000 participants since their birth in 1977 to age 40. The cohort were questioned about their cannabis use and experiences of psychotic symptoms from age 16 onwards. The data from this cohort has been used to produce extensive research on cannabis use and its potential harms.[[2,7,18]]

Fergusson et al.[[7]] investigated the causal link between cannabis use and psychotic symptoms in the CHDS. After control for confounding, daily cannabis use resulted in a significant increase in rates of psychotic symptoms (1.6–1.8 times higher) compared to non-cannabis users. Furthermore, the analysis showed evidence for the causal role of cannabis in psychosis through the specificity of association, robustness to control for confounding and dose-response.[[19]] However, no research has examined whether there is a qualitative difference in the kinds of symptoms reported by regular cannabis users and non-users. It could be argued that if psychosis is caused by cannabis use (reflected in a higher rate of symptoms), then we might also expect to observe more serious symptoms of psychosis (such as thought insertion, delusions and hallucinations) among regular cannabis users.

The present research aimed to reanalyse the data reported by Fergusson et al.[[7]] to examine the patterns of psychotic symptoms reported from the CHDS cohort at ages 18, 21 and 25, in order to answer the following question: was there a significant qualitative difference between the patterns of psychotic symptoms reported by regular cannabis users, as compared to those who did not use cannabis?

Methods

Participants

The data were gathered from the CHDS, a longitudinal study of a cohort of participants who were born in Christchurch, New Zealand in mid 1977. The original cohort had a total of 1,265 participants (635 males, 630 females) which represented 97% of births in Christchurch during the period of recruitment. The cohort was composed of approximately 87% European and 13% Māori/Pasifika ethnicity at birth. The cohort has been studied at birth, age four months, one year and then at annual intervals until age 16, and again at ages 18, 21, and at five yearly intervals from age 25 to 40. The present study used data obtained from the age 18, 21, and 25 assessments. Sample sizes ranged between 1,025 (age 18); 1,011 (age 21); and 1,003 (age 25) representing between 79–81% of the original birth cohort.

Measures

Psychotic symptomatology

At each assessment at age 18, 21 and 25 years the participants took part in a comprehensive mental health interview which was designed to assess multiple aspects of the individual’s psychosocial adjustment and mental health, including any current (within the past month) psychotic symptoms.[[7]] The assessment tool was the symptom checklist 90 (SCL-90),[[20]] from which ten items were used to represent symptoms of psychosis.[[21]] Participants were asked to respond “yes”, “no”, or “maybe” to whether they had experienced each symptom in the month prior to the assessment. The answer options “yes” and “maybe” were counted as a positive response, and these were summed for the 10 items at each assessment to generate a total symptoms of psychosis score for ages 18, 21 and 25 years. These items are shown in Table 1, along with the percentage of respondents endorsing each item.

Frequency of cannabis use

At each assessment at age 18, 21 and 25 years the participants were questioned on their cannabis use. This information classified participants on a five-point scale based on their average frequency of cannabis use over the 12-month period. The scale consisted of: Group 1=non-cannabis user; Group 2=used cannabis a few times; Group 3=used cannabis on less than monthly basis; Group 4=used cannabis on at least a monthly basis; and Group 5=used cannabis on at least weekly (or more often) basis. To assess the accuracy of the participants self-reporting, a nominated informant also reported the individual’s cannabis use at ages 18, 21 and 25 years. There was appropriate agreement between the participant and the informant (r=0.68; p<0.001). For the purposes of the present analysis, those who indicated using cannabis “at least weekly” or more often during an assessment period were classified as “regular cannabis users” during that assessment period.

Data analysis

In the first step of the analysis, in order to show differences in reporting across assessment ages, the percentage of positive responses for each psychotic symptom was calculated at age 18, 21 and 25. Chi-squared tests were performed to investigate differences between the frequencies of psychotic symptoms reported at the three assessment periods. A Šidák correction for multiple significance testing using correlated data was employed, setting the p-value for statistical significance at p=0.01.

In the second step of the analysis, the psychotic symptoms at each assessment point were ranked in order from most frequently reported to least frequently reported, within each level of cannabis use (classified as noted above, ranging from “no cannabis use” to “regular (at least weekly) cannabis use”). These were then aggregated over the period 18–25 years by calculating the average percentage of positive responses across the three assessments, for each level of cannabis use. In order to compare the “no cannabis use” with the “regular cannabis use” group across the three assessments of psychotic symptoms, the aggregated ranks were tested for rank-order differences using a non-parametric Mann–Whitney U test, at p<0.05 significance level (two-tailed).

Results

Frequency of psychotic symptoms

The incidence of each psychotic symptom in the CHDS cohort, represented by the percentage of positive respondents from each age group, is shown in Table 1.

The three most frequently reported symptoms were the same across all three assessment points. These were: “other people being aware of your private thoughts”; “having ideas or beliefs that others do not share”; and “never feeling close to another person”. Similarly, the three least frequently reported symptoms were also the same across all three assessment points. These were: “the idea that someone else can control your thoughts”; “hearing voices that other people do not hear”; and “the idea that something serious is wrong with your body”.

Some symptoms were significantly reported more or less frequently at different time points. Some symptoms decreased in frequency as age increased: “the idea that someone else can control your thoughts”; and “never feeling close to another person”. Finally, reports of the symptom “having ideas or beliefs that others do not share” were significantly more frequently reported at age 21, than at ages 18 or 25.

Rank order of psychotic symptoms relative to cannabis use at each time point

The rank orders of psychotic symptoms reported from non-cannabis users and regular (at least weekly) cannabis users at each time point is shown in Table 2. At all ages, the same three symptoms were reported most frequently for both groups. These were: “having ideas or beliefs that others do not share”; “feeling that you are being watched or talked about by others”; and “feeling that other people cannot be trusted”.

The least frequently reported symptoms were also similar between groups. At 18 years, the symptom “having thought that are not your own”, was one of the three least frequently reported for both groups. At both 21 years and 25 years, both groups shared the same three least frequently reported symptoms, these were: “the idea that someone else can control your thoughts;” “hearing voices that other people do not hear”; and “having thoughts that are not your own”.

Aggregated rank order of psychotic symptoms across ages 18, 21 and 25

The rank order of psychotic symptoms reported from non-cannabis users and regular cannabis users across ages 18, 21 and 25 is shown in Table 3. The rank order is formed from aggregated percentages of the positive responses across the three ages (see Methods).

For both non-cannabis users and regular cannabis users the same three symptoms were most commonly reported. These were: “having ideas or beliefs that others do not share”; “never feeling close to another person”; and “other people being aware of your private thoughts”. Furthermore, non-cannabis users and regular cannabis users reported overlapping symptoms that were the least commonly reported: “hearing voices other people do not hear”; “the idea that someone else can control your thoughts”; and “the idea that something is seriously wrong with your body”.

As noted in the Methods section, the symptom rankings for the non-user and regular user groups were compared using a non-parametric Mann–Whitney U test. The results of this analysis indicated that the distribution of symptom rankings for regular cannabis users did not differ significantly from that of the non-cannabis users (p=0.85). These results suggest that participants who used cannabis regularly at some point during the period of 18 to 25 years reported a pattern of psychotic symptomology that did not differ from non-cannabis users.

View Tables 1–3.

Discussion

This project utilised data from a longitudinal study of a birth cohort (the Christchurch Health and Development Study), gathered from ages 18, 21 and 25 to examine whether there was a significant difference between the patterns of psychotic symptoms reported from regular cannabis users compared to non-cannabis users. Previous research using this cohort has shown that those who use cannabis regularly (“at least weekly”) reported a significantly greater number of symptoms of psychosis during the period 18-25 years than those who did not use cannabis.[[7,21]] The present analyses sought to determine whether there was a qualitative difference in the kind of symptoms being reported by the two groups.

The primary findings of this analysis were that, while Fergusson et al.[[7]] had previously informed that regular cannabis users reported symptoms of psychosis at a rate that was 1.6 to 1.8 times higher than non-users over the period 18 to 25 years, there was no discernible difference in the pattern of symptom reporting over the same observation period. Both groups tended to report common, low-level symptoms (such as “having ideas or beliefs that others do not share”), and neither group was likely to report what would be considered as more severe[[22]] positive symptoms of psychosis, such as “hearing voices that other people do not hear”. It is clear from the pattern of results that regular cannabis users were not more likely than non-users to report symptoms that would be considered severe, irrespective of the fact that they reported a greater number of symptoms overall.

One of the main features of the literature on the potential causal role of cannabis use in the development of psychosis has been the use of a variety of methods for measuring psychotic illness/symptomatology.[[13]] For example, the Dunedin Multidisciplinary Health and Development Study has used schizophreniform disorder as an outcome measure,[[15]] while other studies have employed psychotic symptomatology as measured by the CIDI.[[16]] The present study used the psychoticism subscale of the SCL-90,[[20]] which differs markedly from the CIDI measure of psychotic symptoms, in terms of the nature of the items, and the lack of overlap between SCL-90 Psychoticism items and DSM symptom criteria for disorders with psychotic features. Indeed, several studies have shown that the SCL-90 is not particularly reliable in terms of distinguishing those with differing levels of symptoms (as measured by clinicians, or via alternative measures).[[23]] Perhaps more importantly, however, the SCL-90 has been shown to have poor positive predictive value for diagnosing psychotic illness.[[24]] Collectively, the results suggest that while those who were regular cannabis users reported a significantly greater number of symptoms than non-users, the symptom profile between the two groups did not differ, showing that there was no evidence of greater “severity” among regular cannabis users. Furthermore, it could also be argued that the measure was not, in fact, measuring psychotic symptomatology (as described in both DSM and ICD systems) in any reliable manner.

The present study also found that the incidence of symptom reporting indicated that there was a difference in the psychotic symptoms reported at different ages. While the incidences for six of the psychotic symptoms were stable over age, four symptoms had significant age differences. Incidence of reporting the symptom “having thoughts that are not your own” increased with age, whereas incidences of the symptoms “the idea that someone else can control your thoughts” and “never feeling close to another person” decreased with age. Finally, reporting of the symptom “having beliefs that others do not share” was highest at age 21.

This trend may be explained by the brain maturing over adolescence and early adulthood, resulting in a reduced frequency of particular symptoms. Adolescence is a period marked by significant brain development, with the prefrontal cortex (PFC) being one of the last brain areas to mature. The PFC is responsible for higher level cognitive functioning such as behaviours of goal-orientated planning and decision-making.[[25]] This is of significance because the age of onset for schizophrenia and other related disorders is most commonly in early adulthood, with peak age of onset being 20–29 years.[[26]] An understanding of the incidence of psychotic symptoms at different ages is helpful for health professionals to improve their care as well as enhance their vigilance of psychotic symptoms in young people. Of course, as noted above, some measure of caution should be used in the interpretation of SCL-90 symptoms as representing psychotic illness.

Patients who use cannabis have been shown to have increased likelihood of psychiatric hospital admission, compulsory treatment, and increased duration of admission.[[27]] The present study suggests that while cannabis users may have an increased incidence of psychosis, there is little reason to expect that the extent of cannabis use is related to the severity of psychotic disorder. These findings could aid health practitioners to further understand the nuanced nature of the relationship between cannabis use and psychotic symptoms, providing an opportunity for more tailored or personalized care for particular patients.[[28]].

Though there is considerable evidence for a dose-dependent relationship of cannabis on psychotic symptoms,[[13]] far less research has been done on the patterns of psychotic symptoms associated with cannabis exposure. The present study is the first (to our knowledge) to examine whether there was a qualitative difference between regular cannabis users and non-users in terms of their experiences of symptoms of psychosis. It would be particularly useful for other prospective studies that hold data on cannabis and psychotic symptomatology to undertake similar analyses, as it could be argued that the results of these investigations may differ according to the measure of psychotic symptomatology used.

A key strength of the present study was that it analysed data collected using a prospective, longitudinal design, allowing for an investigation of the impact of cannabis exposure on psychotic symptomology over multiple time points.[[29]] Additionally, the CHDS cohort has been followed since birth, reducing potential selection bias and recall bias.

The present study also has a number of limitations. The first, as mentioned previously, is that the use of the SCL-90 psychoticism subscale is problematic in that the scale cannot be reliably used to identify individuals with a psychotic illness. Also, the present study only held data on cannabis use and psychotic symptoms to age 25, so longer-term associations were not able to be investigated. However, given that both cannabis use and psychotic symptomology are most prevalent in young adulthood, using a younger age group is appropriate.[[2,26]] In addition, it should be noted that the “regular user” group was relatively small (n=64), which may have affected the reliability of observations within this group, as compared with the non-user group. Finally, the present study used data collected in 1995, 1998 and 2003. ElSohly et al.[[5]] found that the potency of cannabis as represented by the THC content of the plant increased significantly between 1994 and 2014. As THC is implicated in the link between cannabis and psychosis, the increased potency has potential implications on the relationships investigated in the present study,[[30]] with the possibility that the consumption of cannabis containing higher average THC levels may be more strongly associated with symptom reporting. Therefore, more recent research is warranted to investigate the effects of higher potency exposure to cannabis on patterns of psychotic symptomatology.