Background and aim: A British study in a large database population reported that people with schizophrenia and diabetes, compared to those with diabetes alone, had a 52% increased risk of death after adjusting for a limited range of confounding variables.1 We therefore investigated whether having schizophrenia independently increased the risk of all-cause mortality in individuals with type 2 diabetes (T2DM) in a community-based cohort with excellent phenotypic characterisation.
Patients and methods: We studied 1,296 residents with T2DM (mean±SD age 64.0±11.3 years, 48.6% male, median [inter-quartile range] diabetes duration 4.0 [1.0-9.0] years) from the longitudinal observational Fremantle Diabetes Study (FDS). All deaths, hospitalisations and mental health outpatient clinic attendances in the state of Western Australia (WA) are recorded in the WA Data Linkage System. A diagnosis of schizophrenia was identified from ICD coding in the hospital and mental health outpatient clinic databases. The main outcome measure was death from any cause after FDS study entry (1993-6) until end-December 2012. Cox proportional hazards modelling, with age as the timeline, was used to identify conventional risk factors for all-cause mortality. After adjusting for the most parsimonious model, schizophrenia was entered.
Results: At FDS study entry, 5 participants had schizophrenia. During 12.9±6.1 years’ follow up, 738 (56.9%) participants died. Male sex, Aboriginality, exercise, smoking, BMI, SBP, DBP, urinary ACR, eGFR, retinopathy, ischaemic heart disease, and peripheral arterial disease independently predicted all-cause mortality. After adjusting for these variables, schizophrenia independently increased mortality (hazard ratio (95% CI): 3.65 (1.33-9.99), P=0.012). All 5 with schizophrenia died (3 from cardiac causes, 2 from infections) at a mean age of 71.9±9.1 years, compared with 78.7±9.0 years in the remaining 733 who died (P=0.09).
Conclusion: Schizophrenia independently increased the risk of death more than threefold in people with type 2 diabetes resulting in premature mortality. Causes of death were diabetes-attributable.