The role of loneliness and social isolation in mediating the relationship between childhood maltreatment and schizophrenia: A genetically informed approach.

Abstract

Observational studies have found loneliness and social isolation to mediate the relationship between childhood maltreatment and schizophrenia. Limitations with observational studies (e.g., confounding and reverse causation), however, have meant the robustness of these relationships has thus far not been explored. To address this gap, the current study utilized genomic structural equation modeling (genomic SEM) and Mendelian randomization (MR) to perform a genetic mediation analysis between childhood maltreatment, loneliness/isolation, and schizophrenia, using summary statistics from three genome-wide association studies (sample sizes 105,318–487,647). While we observed a putative effect of both childhood maltreatment (inverse variance weighted OR = 3.44 per standard deviation increase, 95% confidence interval [CI] [1.66–7.13], p < .001) and loneliness/isolation (OR = 2.98, 95% CI [1.37–6.46], p = .006) on schizophrenia, our hypothesis that loneliness/isolation would mediate the relationship between childhood maltreatment and schizophrenia was not supported (genomic SEM indirect effect = −0.05, SE = 0.05, p = .255; MR indirect effect = 0.10, SE = 0.11, p = .369). Furthermore, reverse mediation analysis indicated that the effect may be in the opposite direction (genomic SEM indirect effect = 0.11, SE = 0.02, p < .001; MR indirect effect = 0.01, SE = 0.00, p < .001), accounting for 20.3%–28.9% of the total effect. The current results suggest that intervening in loneliness/isolation in individuals with a history of childhood maltreatment is unlikely to reduce schizophrenia risk. On the contrary, targeting loneliness/isolation in individuals with a genetic predisposition toward schizophrenia may diminish childhood maltreatment risk.