Characterization of Childhood Trauma, Hippocampal Mediation and Cannabis Use in a Large Dataset of Psychosis and Non-Psychosis Individuals

Abstract Background Cannabis use (CA) and childhood trauma (CT) independently increase the risk of earlier psychosis onset; but their interaction in relation to psychosis risk and association with endocannabinoid-receptor rich brain regions, i.e. the hippocampus (HP), remains unclear. The objective was to determine whether lower age of psychosis onset (AgePsyOnset) is associated with CA and CT through mediation by the HP, and genetic risk, as measured by schizophrenia polygene scores (SZ-PGRS).

Methods: Cross- sectional, case-control, multicenter sample from 5 metropolitan US regions.

Participants: (n=1185) included 397 controls not affected by psychosis (HC); 209 participants with bipolar disorder type-1; 279 with schizoaffective disorder; and 300 with schizophrenia (DSM IV-TR). CT was assessed using the Childhood Trauma Questionnaire (CTQ); CA was assessed by self-reports and trained clinical interviewers. Assessment included neuroimaging, symptomatology, cognition and calculation of the SZ polygenic risk score (SZ-PGRS).

Outcomes: In survival analysis, low CT and CA are associated with lower AgePsyOnset. At high CT or CA, CT or CA are individually sufficient to affect AgePsyOnset. CT relation with AgePsyOnset is mediated in part by the HP in CA users before AgePsyOnset. CA before AgePsyOnset is associated with higher SZ-PGRS and correlated with younger age at CA usage.

Interpretation: CA and CT interact to increase risk when moderate; while severe CT and/or CA abuse/dependence are each sufficient to affect AgePsyOnset, indicating a ceiling effect. Probands with/out CA before AgePsyOnset differ on biological variables, suggesting divergent pathways to psychosis.

Funding: MH077945; MH096942; MH096913; MH077862; MH103368; MH096900; MH122759. Research in context Evidence before this study Cannabis use (CA) and childhood trauma (CT) independently increase the risk of earlier development of psychosis. Scarce evidence exists on the interaction between CA and CT and the neurobiological substrate of their interaction. Added value of this study Analysis of a large transdiagnostic sample of psychosis probands and controls (N=1288) indicates synergy of CT and CA and small but significant contribution of the posterior hippocampus. Data further indicate existence of two populations of probands with psychosis, those with and those without CA after CT before psychosis onset. CT and CA before psychosis onset interact according to a stepwise increase up to reaching a ceiling effect. Implications of all the available evidence Clinically, youth with low, medium CT need to be targeted for intervention before CA onset.