Predictors of Mortality Following a Schizophrenia Spectrum Diagnosis: Evidence From the 20-Year Follow-up of the OPUS Randomized Controlled Trial

Research output: Contribution to journalJournal articleResearchpeer-review

BACKGROUND AND HYPOTHESIS: The life expectancy of patients diagnosed with schizophrenia is 10-12 years lower than in the general population and the mortality gap seems to be worsening. Many of these deaths might be avoidable. We aimed to determine mortality rates and causes of death after a first-episode psychosis, and to examine if clinical characteristics at baseline or during illness could predict mortality.

STUDY DESIGN: The OPUS study was a randomized controlled trial of 578 patients first diagnosed with schizophrenia spectrum disorders. Patients were clinically assessed after 2, 5, 10, and 20 years. Information about time and cause of death was obtained from the Danish Cause of Death Register. Hazard ratios were used to assess predictors of death.

STUDY RESULTS: In total, 82 (14.4%) participants died during 20 years of follow-up. The most common cause of death was suicide (27%). At baseline employment (HR 0.47 P = .049), psychotic disorder other than schizophrenia (HR 0.36, P = .017), and longer duration of untreated psychosis (HR 0.57 P = .042) predicted lower mortality while substance use predicted higher mortality (HR 2.56, P < .001). During follow-up, symptom remission without antipsychotic medication and recovery predicted lower mortality (HR 0.08 P = .013 and HR 0.21, P = .028) while substance use (HR 3.64 P < .001), and all chronic illnesses predicted increased risk.

CONCLUSIONS: There is an increased risk of early mortality in schizophrenia compared to the background population, and there is an urgent need for new efforts to improve the disparities in health that lead to this increased mortality.

Original languageEnglish
JournalSchizophrenia Bulletin
Issue number5
Pages (from-to)1256–1268
Number of pages13
Publication statusPublished - 2023

Bibliographical note

© The Author(s) 2023. Published by Oxford University Press on behalf of the Maryland Psychiatric Research Center. All rights reserved. For permissions, please email:

ID: 362543569