The neuroanatomy of cannabis use: does gender matter? Findings from the ENIGMA addiction working group

Background: Cannabis is the most widely used illicit drug globally. Structural neuroimaging (sMRI) studies of cannabis users versus controls show neuroanatomical alterations in areas dense in cannabinoid receptors, but the location and direction of the alterations are inconsistent. These discrepancies may be driven by gender differences that have not been fully explored given the small-sized and mostly-male samples, as well as by heterogeneous cannabis use levels and confounders that have been inconsistently accounted for (e.g. IQ, nicotine and alcohol use). Here, we aimed to examine the neuroanatomical correlates of cannabis use, and the role of gender and severity of use in the largest cohort of 270 participants examined to date, from the ENIGMA Addiction Working Group. Methods: 145 cannabis users (48 females) and 125 controls (36 females) from seven distinct research sites underwent sMRI. We selected eight regions of interest (ROIs) - amygdala, hippocampus and cerebellum (gray/white matter) volumes, and orbitofrontal (medial/lateral), frontal pole and insular cortical thickness extracted locally, in each site, using standard pipelines from Freesurfer v5.3. Mixed-effect models were run to test the impact of the following variables on the ROIs (i) group, gender, and group-by-gender interaction; (ii) cannabis use level (age of onset and monthly dosage) and (iii) cannabis dependence in a subsample (n = 228), comparing 70 dependent users (25 females), 51 non-dependent users (20 females), and 107 controls (34 females). All models were adjusted for assessment site, age, IQ, intracranial volume and alcohol and tobacco monthly use. Results: Cannabis users showed trend-level thinner medial orbitofrontal cortex (p<0.08) than controls. Dependent users had thinner medial (p<0.01) and lateral (p<0.05) orbitofrontal cortex, and smaller cerebellar white matter volume (p<0.05) compared to both non-dependent users and controls. No differences were found between non-dependent users and controls. Female cannabis users had trend-level smaller cerebellar white matter volume (p<0.06) relative to female controls, an effect not found in males. Consistently, female dependent users had smaller cerebellar white matter (p<0.01) and thinner orbitofrontal cortex (p<0.05) compared to female non-dependent users and controls, while no effect was found among males. Higher monthly cannabis dosage and earlier cannabis use onset predicted smaller volumes of cerebellar grey (p<0.05) and white (p<0.05) matter respectively, in male users only. By contrast, in female users, trend-level associations were found between earlier onset of cannabis use and thinner OFC (p<0.08) and frontal pole (p<0.06). Conclusions: We showed that overall, cannabis users had marginal neuroanatomical alterations compared to controls limited to the orbitofrontal cortex. Accounting for gender and dependence revealed more marked reductions in the same region (i.e., orbitofrontal cortex) and extended also to other regions (i.e., cerebellar white matter) which appeared driven by women cannabis users and dependent users. Our results warrant a stratified approach to the study of cannabis use neurobiology and point to a neurobiological vulnerability in women cannabis users and dependent cannabis users and suggest that different neurobiological targets for treatment may be required in these groups.