Cannabis, cannabinoids and psychosis: a balanced view.

The laws legalizing recreational cannabis use, the increasing strength of cannabis and cannabis derivatives, and the growing availability and commercialization of cannabis call attention to the possible implications for mental health, and specifically for the incidence of psychosis. Several lines of evidence suggest that exposure to cannabis and synthetic cannabinoids may contribute to the risk for psychosis1. The spectrum of psychosis outcomes linked to cannabis and cannabinoids range from short-lived psychotic states to chronic psychotic disorders. In addition to observational data, experimental laboratory studies provide compelling evidence that cannabis, its principal psychoactive constituent delta-9-tetrahydrocannabinol, and synthetic cannabinoids induce acute brief psychotic states characterized by positive, negative and cognitive symptoms resembling the symptoms of schizophrenia2. Cannabis may induce a psychotic disorder (cannabis-induced psychotic disorder, CIPD) lasting days to weeks, that often requires clinical intervention, resolves with the termination of use, and recurs with re-exposure. In Denmark, the increasing potency of cannabis has been associated with an increased incidence of CIPD3. Interestingly, up to 50% of patients diagnosed with CIPD are re-diagnosed years later with schizophrenia or bipolar disorder, suggesting that CIPD may be a harbinger of a chronic psychotic disorder4. Furthermore, the rate of “conversion” to schizophrenia seems greatest for CIPD relative to other substance-induced psychoses. However, whether CIPD evolves into schizophrenia, or whether CIPD and schizophrenia are related yet distinct, remains unclear. Beyond the above brief syndromes, epidemiological studies link cannabis exposure to a higher risk (2- to 4-fold) for schizophrenia. The dose-response relationship is linear, such that more frequent and heavier use, and use of higher potency cannabis, carries a greater risk. Other moderating factors include an earlier age of exposure, childhood trauma, and exposure to other drugs. Whether the relationship between cannabis and psychosis is causal should be considered separately for the psychosis outcomes in question. The tight temporal relationship between exposure to cannabinoids and the emergence of psychotic states observed in experimental studies provides strong evidence to support a causal relationship. Likewise, with CIPD, the emergence of psychosis with exposure to cannabis, its resolution with abstinence, and its recurrence with resumption, also make a compelling case for causality. In contrast, the evidence linking cannabis and schizophrenia is mostly from epidemiological studies, which are not without limitations. While these studies attempt to adjust for confounders, including other drug use, latent psychosis, pre-existing cannabis exposure, and other psychiatric disorders, any accurate estimate of causality is limited, and dependent on capturing and measuring all relevant known and unknown confounders. The relationship between cannabis and schizophrenia fulfills, to varying degrees, a number of the classic Hill criteria of causality, including the strength, consistency, specificity, temporal characteristics, direction, biological gradient, coherence and plausibility of the association and supporting experimental evidence. Regarding the strength of the evidence, there is a linear dose-response relationship (2- to 9-fold) driven by the frequency, amount of use and potency of cannabis5. For perspective, cigarette smokers are 15-30 times more likely to get lung cancer or die from lung cancer than non-smokers, and ~85% of cases of lung cancer are linked to smoking. In terms of specificity, while cannabis use increases the risk of depression, the evidence is strongest for psychosis. Furthermore, while other drugs (e.g., amphetamines) are associated with psychosis, the risk of psychosis seems greatest with cannabis. The fact that the endocannabinoid system is involved in neurodevelopmental processes supports the biological plausibility for cannabis exposure during adolescence to disrupt neurodevelopmental processes and, in doing so, increase the risk for schizophrenia. However, regarding temporality, the reverse causation hypothesis has been proposed, according to which the risk for schizophrenia confers risk for cannabis use rather than the risk for cannabis conferring risk for schizophrenia. Genome-wide association studies (GWAS) provide evidence for a bidirectional causal relationship between cannabis use and schizophrenia, but suggest a larger contribution of reverse-causal mechanisms and common genetic risk for both schizophrenia and cannabis use (genetic confounding)6. The classic criteria for causality have limitations, especially when applied to multifactorial disorders. Schizophrenia, or the “group of schizophrenias” as termed by Bleuler, is likely heterogenous in etiopathogenesis. Perhaps cannabis exposure might be linked to a specific psychosis subtype that is buried within the group of schizophrenias. Furthermore, in line with a multifactorial etiopathogenesis, cannabis is neither necessary nor sufficient to cause schizophrenia. More likely cannabis is partly causal, interacting with other factors such as genetic liability to confer greater risk for schizophrenia. If cannabis confers a higher risk for developing psychosis, then changes in the cannabis landscape should be accompanied by increased rates of psychosis. Indeed, some studies suggest increasing rates of psychosis linked to cannabis7. However, it may be too early for the public health impact of the changes to be fully realized. In this regard, the history of cigarettes and lung cancer may be instructive. While suspected, it took half a century to recognize that cigarettes cause lung cancer. Despite compelling epidemiological data, evidence from animal studies, cellular pathology and chemical analyses was necessary to establish that cigarettes cause lung cancer. In contrast, schizophrenia has no signature pathology and is likely heterogenous, and the risk for psychosis with cannabis is lower than that of lung cancer with smoking. Therefore, at the present time it may be unrealistic to expect the same level of evidence and/or certainty that cannabis causes schizophrenia. More likely, evidence may accumulate linking cannabis to a subtype of schizophrenia. While the spotlight has been on whether cannabis causes new cases of psychosis, we risk overlooking the impact of cannabis on those with established psychotic disorders. The high rates of cannabis use by individuals with schizophrenia have been attributed to “self-medication”, but there is little evidence to support that hypothesis8. In contrast, there is clear evidence of cannabis having a negative impact on the course of schizophrenia, with greater positive symptoms, relapse rates, emergency department visits, hospitalizations, homelessness and legal problems. Schizophrenia ranks amongst the top fifteen leading causes of disability and is financially burdensome. Therefore, additional costs from the consequences of comorbid cannabis use may be substantial. To conclude, it is tempting to speculate what impact reducing or eliminating cannabis exposure in adolescents might have on the rates of psychosis. For example, the EU-GEI study found that, if high-potency cannabis was no longer available, depending on the region, a substantial proportion (12-50%) of first-episode psychosis cases could be averted5. Reducing the rates of psychosis by just 10% is well worthwhile. Towards that end, we need to identify factors that place individuals at greatest risk for developing psychosis in the context of exposure to cannabis. Furthermore, the public needs to be educated about the contribution of cannabis use to the risk of psychosis. Likewise, greater efforts are necessary to educate patients with psychosis about the negative impact of cannabis on illness course, to discourage their use of cannabis, and to develop effective treatments. In what may be an ominous development, as cigarette sales decline worldwide, the tobacco industry, with its vast experience in mass-production, advertising, marketing, lobbying and legal defense, is investing in the cannabis industry! With the cannabis landscape continuing to evolve, we must remain concerned about the risk of psychosis outcomes related to cannabis.