L. John Horwood, David M. Fergusson, Carolyn Coffey, George C. Patton, Robert Tait, Diana Smart, Primrose Letcher, Edmund Silins, Delyse M. Hutchinson

This study presents an integrative data analysis of the association between frequency of cannabis use and severity of depressive symptoms using data from four Australasian cohort studies. The integrated data comprised observations on over 6900 individuals studied on up to seven occasions between adolescence and mature adulthood. Repeated measures data on frequency of cannabis use (not used/<monthly/≥monthly/≥weekly) and concurrently assessed depression scores were pooled over the four cohorts. Regression models were fitted to estimate the strength of association between cannabis use and depression. Fixed effects regression methods were used to control for confounding by non-observed fixed factors. Increasing frequency of cannabis use was associated with increasing depressive symptoms (p<0.001). In the pooled data weekly users of cannabis had depression scores that were 0.32 (95%CI 0.27-0.37) SD higher than non-users. The association was reduced but remained significant (p<0.001) upon adjustment for confounding. After adjustment depression scores for weekly users were 0.24 (95%CI 0.18-0.30) SD higher than non-users. The adjusted associations were similar across cohorts. There was a weak age×cannabis use interaction (p<0.05) suggesting that the association was strongest in adolescence. Attempts to further test the direction of causality using SEM methods proved equivocal. More frequent cannabis use was associated with modest increases in rates of depressive symptoms. This association was stronger in adolescence and declined thereafter. However, it was not possible from the available data to draw a definitive conclusion as to the likely direction of causality between cannabis use and depression.