In an Australian multicenter cohort, 1237 participants with an established ICD-10 diagnosis of psychotic disorder were categorised according to history of lifetime cannabis use (non-users, n = 354; cannabis users, n = 221; cannabis dependency, n = 662). Groups were analyzed according to available indices of cognitive ability: the National Adult Reading Test — Revised (NART-R) for ability prior to illness onset; and the Digit Symbol Coding Test (DSCT) for current ability. Two-way analysis of variance was conducted without any covariate, followed by a two-way analysis of covariance (using age, age at onset of psychiatric illness, premorbid IQ and the Socio-Economic Index for Areas (SEIFA) rankings).
Whilst there appeared to be a significant association between cannabis use and mean DSCT (higher DSCT scores in cannabis using groups) F(2,1080) = 9.478, p < 0.001, 畏2 = 0.017), once covariates were used in the analysis there were no significant differences between groups in mean DSCT scores (F(2,1011) = 0.929, p = 0.395, 畏2 = 0.002). Similarly there were no differences between groups in mean NART scores once, age, age at illness onset and SEIFA rankings were used as covariates (F(2,1032) = 1.617, p = 0.199, 畏2 = 0.003).
Confounding variables underpin the association between cannabis use and cognitive function in psychotic disorders. Taken together, it would appear that cannabis use or dependence has no additive effect on cognitive dysfunction in these disorders.