Effect of age, family history of diabetes, and antipsychotic drug treatment on risk of diabetes in people with psychosis: a population-based cross-sectional study

Abstract

Background: Psychosis is associated with an increased risk of diabetes mellitus. A positive synergy between antipsychotic drug effects and a pre-existing liability to diabetes mellitus might explain the especially high relative risk of diabetes mellitus in young adults with psychosis. We aimed to assess the individual and joint effect of age, family history of diabetes mellitus, and currently prescribed antipsychotic drug treatment on risk for diabetes mellitus. Methods: In this study, we used data from the 2010 Australian National Survey of Psychosis—an observational study done at seven sites in five Australian states. We included data from 1155 people with psychosis aged 18–64 years who were in contact with psychiatric services and who gave a fasting blood sample to test for current diabetes mellitus. Using logistic regression, we modelled the association of diabetes mellitus with age, family history of diabetes mellitus, and current antipsychotic drug treatment. We compared model fit with and without two-way and three-way interaction terms and used likelihood ratio tests to establish which terms to include in the final model. Findings: After adjustment for older age, which was an independent risk factor, compared with not taking antipsychotic drugs, antipsychotic drug treatment was associated with diabetes mellitus only in those without a family history of diabetes mellitus (clozapine adjusted odds ratio [OR] 7·22, 95% CI 1·62–32·20, p=0·01; quetiapine 5·91, 1·33–26·30, p=0·02; aripiprazole 5·06, 0·86–29·64, p=0·07; risperidone 4·17, 0·90–19·24, p=0·07; and olanzapine 2·23, 0·45–11·06, p=0·32). Antipsychotic drug treatment was not associated with additional risk of diabetes mellitus in those with a family history (clozapine adjusted OR 1·51, 95% CI 0·64–3·54, p=0·34; quetiapine 1·09, 0·49–2·43, p=0·82; aripiprazole 0·43, 0·12–1·49, p=0·18; risperidone 1·12, 0·48–2·63, p=0·79; and olanzapine 0·67, 0·26–1·71, p=0·39). Interpretation: People with psychosis are at increased risk of diabetes mellitus if they have a family history of diabetes mellitus or if they have no family history of diabetes mellitus but are taking antipsychotic drugs. Increasing age increases risk but independently of family history or antipsychotic drug treatment. Clinicians should not think the absence of a family history of diabetes mellitus protects their patients from the diabetic side-effects of antipsychotics. Funding: Australian Federal Government and Orygen.