BackgroundGenetic vulnerability to psychosis is polygenic, involving multiple genes with small individual effects (Psychiatric Genomics Consortium (PGC), 2014). The risk of psychosis is also related to environmental factors, such as childhood trauma (Lardinois et al, 2011). Although the onset of psychosis is thought to result from the interaction of genetic and environmental risk factors (Walker & Diforio, 1997), the extent to which the influence of childhood trauma depends on genetic susceptibility remains unclear. We sought to address this issue in a large prospective study of people at clinical high risk (CHR) for psychosis. These individuals present with psychotic and affective symptoms, and are at increased risk of developing both schizophreniform and affective psychoses.MethodsWe studied subjects of European ancestry, drawn from EU-GEI, a large multi-centre prospective study of people at CHR for psychosis. At baseline, DNA was obtained from subjects who met the CAARMS criteria for the CHR state (n=266) and healthy controls (HC; n=42). Childhood trauma was assessed using the childhood trauma questionnaire (CTQ), which comprises 5 subdomains: emotional abuse, physical abuse, sexual abuse, physical neglect, and emotional neglect. Polygenic risk scores (PRSs) for schizophrenia (SCZ), bipolar disorder (BD) and major depressive disorder (MDD) were constructed separately, using results from meta-analyses by the corresponding Disorder Working Groups of the PGC. The CHR subjects were clinically monitored for up to 5 years and clinical outcomes were assessed in terms of transition to psychosis (as defined by the CAARMS), remission from the CHR state (subject no longer meets CAARMS inclusion criteria) and level of functioning (GAF Disability Scale). Logistic regression models were used to investigate the association between each PRSs and childhood trauma as predictors of transition and remission, adjusted by population stratification using the first 10 principal components, age, sex and site. All findings are reported at p<0.017, Bonferroni-corrected for the 3 PRSs.ResultsWithin the CHR sample, the onset of psychosis during follow up was related to interactions between the BD PRS and the total childhood trauma score (OR=0.959, 95% CI 0.930–0.988, p=0.006), and between the BD PRS and physical abuse (OR=0.787, 95% CI 0.689–0.900, p<0.001). Remission from the CHR state was related to an interaction between the SCZ PRS and childhood sexual abuse (OR: 1.110, 95% CI 1.004–1.226, p=0.041).DiscussionThese data indicate that clinical outcomes in CHR subjects are related to interactions between the polygenic risk for psychotic disorders and childhood adversity. The measurement of interactions between genomic and environmental risk factors may help to predict individual outcomes in people at high risk in a clinical setting.