Journal reference: Large M, Sharma, S, Compton MT, et al. Cannabis use and earlier onset of psychosis. Arch Gen Psychiatry 2011: published online 7 February 2011 ^[1]

Link: http://dx.doi.org/10.1001/archgenpsychiatry.2011.5

Published: 7 February 2011

Evidence cookie says...

Cannabis may be causally related to psychotic illness but the evidence remains unclear. This meta-analysis demonstrates an association between cannabis use and earlier onset of psychosis. There are limitations in the reliability of the results; the association should not be considered definitive and causal inferences should made with caution. See below for more details.

Article details:

Study design:

systematic review, meta-analysis of observational studies

Study aim:

to examine the association between the use of cannabis, alcohol, and other psychoactive substances and the age of onset of psychosis

Study conclusion:

The results of the meta-analysis provide evidence for a relationship between cannabis use and earlier onset of psychotic illness, and they support the hypothesis that cannabis use plays a causal role in the development of psychosis in some patients. The results suggest the need for renewed warnings about the potentially harmful effects of cannabis ^[1].

Critical appraisal:

Are the trial results valid?

Internal validity: Wikipedia

Is it unlikely that important, relevant studies were missed?

Yes.

• systematic search of 5 electronic databases (CINAHL, EMBASE, MEDLINE, PsycINFO, and ISI Web of Science)
• two authors independently selected studies according to inclusion criteria
• any paper considered likely to contain demographic data about cohorts of substance-using and non-substance using participants were examined in full-text
• references of included studies and those of previous reviews of associations between cannabis and psychosis were hand searched for further studies
• 83 studies included in the meta-analysis → included 8 studies that met inclusion criteria after further data was provided by the primary researchers
  • note: two authors of similar studies could not provide further data and their studies were not included
• Limitations:
  • English language only

Were the criteria used to select articles for inclusion appropriate?

Yes.

• age of onset of psychotic disorder reported
• compared cohorts of patients who reported the use of a psychoactive substance (other than tobacco), compared to a control group who did not use these substances
• unpublished studies and studies from non-peer-reviewed sources not included

Were the included studies sufficiently valid for the type of question asked?

Unclear.

• the majority of the studies were in patients with schizophrenia or first onset psychosis, mostly inpatients and some outpatients
• no reporting of the quality of the included studies, or whether quality had been assessed
• the included studies are probably valid for substance use in general
• however, users of individual substances (alcohol and cannabis) may have been taking other substances as well.
  • this places substantial limitations on the causal inferences that can be made regarding individual substances

Were the results similar from study to study?

Unclear.

• study heterogeneity for the age at onset of psychosis grouped by substance remained substantial: I² = 86.9
  • interpretation: 86.9% of the variation seen was due to differences between the studies rather than due to chance

What were the results?

Outcomes:

Effect on the age of onset of psychosis of the substance as compared to group not on the substance:

• cannabis:
  • -2.70 years (i.e., first onset of psychosis was on average 2.7 years earlier in the cannabis groups)
  • effect size (standardised mean difference):
    • -0.414, (95% CI -0.526 to -0.301)
    • interpretation: a small to medium effect size

• alcohol:
  • -0.28 years
  • effect size (standardised mean difference):
    • -0.038, (95% CI -0.196 to -0.120)
    • interpretation: no statistically significant effect detected

• substance use:
  • -2.00 years (i.e., first onset of psychosis was on average 2 years earlier in the substance use groups)
  • effect size (standardised mean difference):
    • -0.315, (95% CI -0.405 to -0.225)
    • interpretation: a small effect size

Will the results help me care for my patient?

External validity: Wikipedia

Are the participants different to my patient?

• the majority of the study participants were in retrospective observational studies of schizophrenia or first onset psychosis
• there is otherwise very limited demographic data available to make a judgement

Were all the clinically important outcomes considered?

• the age of onset of use of substances was not included in the analysis (not possible)
• in its absence, there is still considerable ambiguity regarding the presence or direction of causality between cannabis and psychosis

Study weaknesses (summary)

• weaknesses primarily in the limitations of the included studies
• little description of the demographics of the participants in the included studies limits the applicability of the findings

Study strengths (summary)

• meta-analytic approach
• appears to have been an extensive/exhaustive review of the available data

Biases and conflicts of interests

• none declared and none seem obvious

Clinical relevance to primary health care

Cannabis is a widely used in the Australian community. It is the third most commonly used substance behind alcohol and tobacco. There is a known association between substance use and psychosis, and more specifically, between cannabis use and psychosis, but substantial uncertainly remains on whether this association is causal. That is, we do not know the reason behind the observed phenomenon that people who end up with psychosis used more cannabis. It might be that cannabis can cause psychotic illness (the causal hypothesis), but it might also be that cannabis use is more common in people with other risk factors for psychosis.

This meta-analysis demonstrates an association between cannabis use and earlier onset of psychosis, and supports the hypothesis that there is a causative link.

Caution should be taken in not over-interpreting the results. The evidence from this study does not by itself provide any convincing evidence of causation. The limitations of making causative inferences from the source observational studies remain and it is likely, if not probable, that some of the effect size seen is due to confounding factors. For instance, the cannabis group included the participants of a study who used cocaine but whom also happened to all use cannabis.

It is reasonable for Australian general practitioners to advise their patients in taking a precautionary approach. Cannabis use may be causally associated with psychosis. It is prudent to avoid its use, particularly for those at risk of psychotic illness (e.g., family history).

References

1. Large M, Sharma, S, Compton MT, et al. Cannabis use and earlier onset of psychosis. Arch Gen Psychiatry 2011: published online 7 February 2011