Introduction: There is increasing evidence of excess mortality in schizophrenia but less information on other non-affective psychoses. We therefore generated standardised mortality ratios (SMRs) for community-dwelling people with schizophrenia and other non-affective psychoses, relative to the general population, and examined changes to the SMR over time.

Methods: We conducted a systematic review in which Pubmed, CINAHL, EMBASE, Google Scholar and PsycINFO were searched for publications that reported SMRs for all-cause mortality among community-dwelling people with schizophrenia and psychotic disorders. Meta-analyses of SMRs were conducted, pooled across genders and then separately by gender. Sub-group analyses were conducted for diagnostic group, global region, decade and risk of study bias.

Results: We were able to include 34 studies covering 1,724,906 participants. The gender pooled SMR for schizophrenia and psychotic disorders was 3.08 (95%CI 2.88–3.31). Schizophrenia and broader psychotic disorders had similar SMRs. Stratification by decade of observation suggests that the difference in SMR is not declining and may possibly be widening. Analyses showed high levels of heterogeneity.

Conclusions: The appearance of a static or widening mortality gap over time between people with schizophrenia and psychotic disorders and the general population is of concern. However, whether it is an increase over time is unclear, as there are insufficient studies to confirm this.