LAUSR

Abstract Background The importance of inherited factors in the development of affective psychosis, which includes bipolar disorder and major depression disorder, is widely accepted, but the fact that monozygotic twin concordance is substantially less than 100% suggests that environmental factors (ERF) are likely to play an important role as well. While the link between a variety of ERF and schizophrenia-spectrum disorder is well established, less is known about how these ERF impact in affective psychosis. In the current study we aim to analyse the role of environmental risk factors in the expression of affective disorder compared with schizophrenia-spectrum disorder, and its interaction with the genetic risk by using polygenic risk score (PRS). Methods DNA was obtained from most participants (73.6% of 1130 cases and 78.5% of 1499 controls) among 16 European cities as part of the EUGEI case-control study. PRS for SZ, BD and MDD were built using the latest available data from the Psychiatric Genomic Consortium (PGC). Multinomial logistic regression models were used to test whether the association of genetic load (by PRSs) with different diagnostic categories based on DSMIV from OPCRIT items were greater if there was also evidence of ERF (urbanicity, migration, cannabis use and childhood trauma) through the inclusion of interaction terms between the different PRSs and the ERF. Analyses were conducted for each environmental factor separately and for a combined poly-environmental risk score based on Maudsley Environmental Score (MPES) will be calculated. Results Being 1st generation migrant was not associated with any of the diagnostic categories, nor independently nor in interaction with PRSs. Living in urban environment increases the risk of SSD (RRR=1.68, 95% CI 1.06 – 2.67), but without interacting with any genetic measure. Regarding cannabis use, having ever used cannabis is independently significantly associated with SSD (RRR=2.26, 95% CI 1.69 – 3.02) and BD (RRR=5.3, 95% CI 2.69 – 10.46), showing as well in the latter group an interaction with PRS MDD (RRR=2.3, 95% CI 1.18 – 4.49). Although daily use of cannabis strongly predicted risk of SSD and BD, having use more than once a week only increased risk for SSD. Neither having used cannabis more than once a week or daily interacted with any of the PRSs. Having been exposed to any childhood trauma was independently significantly associated with all three diagnostic groups, but did not show any significant interaction with PRSs. Lastly, despite MPES increased risk for SSD and BD, it didn’t interact significantly with any PRSs. Discussion These results suggest that despite evidence of both PRSs on one hand and urbanicity, cannabis and childhood trauma overall increase risk of belonging to any psychotic diagnostic category separately, we only found some suggestion of potential interaction between genetic vulnerability to MDD and cannabis use associated with BD. Nonetheless, due to most of the interactions showing the expected trend, analyses examining interactions between PRSs and ERF with the different diagnostic groups with bigger samples are warranted.