The objective of this systematic review was to investigate which specific and especially neurocognitive factors related to schizophrenia, were associated with homicide risk.

A systematic English-French Medline literature search of cohort studies, case-control studies and transversal studies published from January 1999 to December 2009 was performed combining the MeSH terms “schizophrenia”, “homicide”, “violence”, “mental process”, “cognition”, “risk”, “risk factors”,. Abstract selection was based on the STROBE checklist for observational studies and on the consort statement for clinical trials.

Of the 366 selected studies, 65 observational or prospective studies, 10 systematic reviews and meta-analysis and 2 interventional studies met the selection criteria and were included in the final analysis. Firstly, we highlighted that historical (past violence, juvenile detention, physical abuse, parental arrest record), dispositional (male gender, young age, low socioeconomic status) and contextual (recent divorce, unemployment, victimisation) factors could be considered as general homicide-related factors. Clinical factors (clinical paranoid, delusions of persecution or thought insertion, substance abuse, disorganized thinking, long duration of untreated psychosis, stopped monitoring or treatment) were more schizophrenia-specific factors for homicide. Most of the excess risk appears to be mediated by substance abuse. Secondly, our results suggested that schizophrenics with a history of aggressive behaviour compared to those without such history, had better performances on global neuropsychological tests exploring executive functions but performed more poorly as considering orbitofrontal functions.

We suggest that every comprehensive psychiatric assessment should explore the risk of homicide, including historical, dispositional, contextual, clinical and neurocognitive (low insight capacity, impaired frontal functions) factors of violence.