Abstract
Suicide, suicide attempts, and suicide ideation account for a significant proportion of deaths, disability, and suffering worldwide. There is currently little knowledge about what factors causally influence risk for suicide. Schizotypy—a set of cognitive-perceptual, interpersonal, and disorganised indicators that reflect liability for schizophrenia—predicts concurrent and future suicidal ideation and attempts. Some suggest the schizotypy-suicidality relationship reflects the influence of shared risk factors, or that both are non-specific indicators of the severity of psychopathology. However, schizotypy is significantly related to suicidality after adjustment for confounding, suggesting an independent role. The overarching aim of the present thesis was to test whether schizotypy is causally related to suicidal ideation and attempts.
In Studies 1 and 2, the causal relationship of schizotypy with suicidality was tested using Mendelian randomisation in two independent cohorts, the Philadelphia Neurodevelopmental Cohort (n = 4767) and the FinnTwin12 cohort (n = 1213). No associations were found between a schizophrenia polygenic risk score (PRS) and schizotypy, or between the schizophrenia PRS and suicidal ideation in either study. The schizophrenia PRS was associated with suicide attempts in the FinnTwin12 cohort. As the schizophrenia PRS was not validated as a proxy measure of schizotypy, causal inferences could not be clearly made. In Study 3, the causal relationship of schizotypy and suicidal ideation was examined with a co-twin control analysis in the FinnTwin12 cohort. Results were consistent with a causal relationship between interpersonal features of schizotypy and suicidal ideation. In contrast, there was no evidence cognitive-perceptual or disorganised features were causally related to suicidal ideation. Finally, in Study 4, to determine whether schizotypy was specifically related to certain subpopulations of suicide ideators (represented by distinct growth trajectories), growth mixture modelling was used to model suicidal ideation trajectories over a 20-year period in the Dunedin study cohort (n = 1037). Schizotypy predicted a persistent passive ideation trajectory more so than a transient ideation trajectory.
Together, these results provide some evidence consistent with the view that schizotypy causally influences suicidality. However, some sources of confounding (e.g., by non-shared environmental variables) could not be controlled. Results were consistent with conceptualisations of schizotypy and suicidality as heterogenous constructs and provided preliminary evidence that the schizotypy-suicidality relationship may be stronger in certain subpopulations of suicidal ideators.