Causal association between cannabis and psychosis: examination of the evidence
Arseneault L, Cannon M, Witton J and Murray RM;
Commented by , 25 Feb 2004
Aim of the study
Although it is well established that excessive cannabis use can lead to transient psychotic episodes and to exacerbation of pre-existing psychotic disorders, there is still a debate as to whether cannabis use is an individual risk factor for the development of schizophrenia. The authors therefore reviewed and meta-analytically summarised four prospective epidemiological studies that examined the relationship between cannabis use and subsequent development of psychosis.
Methods
The authors firstly reviewed the available evidence from a number of well designed cross-sectional studies that examined the association between cannabis use and psychosis. In a second step they reviewed and meta-analytically combined the results of four prospective studies that looked at a causal relationship between cannabis use and later development of psychosis.
Results
Five cross-sectional national surveys (from the USA, Australia and The Netherlands) and five British local surveys provided evidence that rates of cannabis use are higher among people with schizophrenia than among the general population. Quantitatively the use of cannabis was about twice as high in people with schizophrenia.
However, these associations between schizophrenia and cannabis use can only be considered to be hypothesis generating concerning the question as to whether cannabis is a cause for the development of schizophrenia. It may as well be that cannabis use is just a consequence of schizophrenia, e.g. as an attempt of affected people to cope with their symptoms. As the authors explain, a prerequisite of a causal relationship is temporality, i.e. cannabis use must precede the later development of psychosis. Such an effect can only be examined by prospective studies.
Four such prospective studies from the Netherlands, Sweden, and two from New Zealand were identified. All four prospective studies showed that cannabis use in adolescence was associated with later development of schizophrenia, even when a number of potentially confounding factors were controlled for.
Discussion
In their careful examination of the literature, the authors do conclude that cannabis use is a causal risk factor for the later development of schizophrenia even after controlling for a number of confounding factors such as low IQ, nicotine use, poor social integration, sex, age, ethnic group, educational status, unemployment, marital status and previous psychotic symptoms.
The authors also list a number of methodological short-comings of the available studies. E.g. cannabis use was assessed by self-reports in the studies, objective measures such as urine tests or hair analysis were not taken. However, this probably let to an underreporting of cannabis use so that the strength of the association might have been underestimated rather than overestimated.
There was also only limited information on the use of other illegal drugs, e.g. stimulants, which are known to be even more psychotogenic than cannabis, although this factor was controlled for when information was available. Even three of the four prospective studies could not rule out with absolute certainty that prodromal signs of schizophrenia preceded cannabis abuse so that in a way cannabis use may still be a consequence of schizophrenia rather than a cause.
This possibility was considered in one of the four prospective studies, however. The authors therefore concluded that although cannabis use alone is neither a sufficient nor a necessary cause of schizophrenia, the prevention of its use could reduce the incidence of schizophrenia by eight percent.