What Causes the Onset of Psychosis in Individuals at Clinical High Risk? A Meta-analysis of Risk and Protective Factors

• Published in: Schizophrenia bulletin Vol. 46; no. 1; pp. 110 - 120
• Main Authors: Oliver, Dominic, Reilly, Thomas J, Baccaredda Boy, Ottone, Petros, Natalia, Davies, Cathy, Borgwardt, Stefan, McGuire, Philip, Fusar-Poli, Paolo
• Format: Journal Article
• Language: English
• Published: United States Oxford University Press 04.01.2020
• Subjects: Humans; Protective Factors; Psychotic Disorders - epidemiology; Regular; Risk Factors; Schizophrenia - epidemiology; symptoms; functioning; risk; schizophrenia; clinical high risk; psychosis
• ISSN: 0586-7614; 1745-1701
• DOI: 10.1093/schbul/sbz039

Twenty percent of individuals at clinical high risk for psychosis (CHR-P) develop the disorder within 2 years. Extensive research has explored the factors that differentiate those who develop psychosis and those who do not, but the results are conflicting. The current systematic review and meta-analysis comprehensively addresses the consistency and magnitude of evidence for non-purely genetic risk and protective factors associated with the risk of developing psychosis in CHR-P individuals. Random effects meta-analyses, standardized mean difference (SMD) and odds ratio (OR) were used, in combination with an established stratification of evidence that assesses the association of each factor and the onset of psychotic disorders (from class I, convincing evidence to class IV weak evidence), while controlling for several types of biases. A total of 128 original controlled studies relating to 26 factors were retrieved. No factors showed class I-convincing evidence. Two further factors were associated with class II-highly suggestive evidence: attenuated positive psychotic symptoms (SMD = 0.348, 95% CI: 0.280, 0.415) and global functioning (SMD = -0.291, 95% CI: -0.370, -0.211). There was class III-suggestive evidence for negative psychotic symptoms (SMD = 0.393, 95% CI: 0.317, 0.469). There was either class IV-weak or no evidence for all other factors. Our findings suggest that despite the large number of putative risk factors investigated in the literature, only attenuated positive psychotic symptoms, global functioning, and negative psychotic symptoms show suggestive evidence or greater for association with transition to psychosis. The current findings may inform the refinement of clinical prediction models and precision medicine in this field.