BARCELONA — It remains unclear whether people with psychiatric problems taking atypical antipsychotic medicines develop diabetes more frequently than do other people, according to the results of a systematic evidence review presented at an international congress on prediabetes and the metabolic syndrome.
Several previously published case reports and cross-sectional studies have suggested that there is an association between atypical antipsychotic agents and type 2 diabetes, said Jeffrey Johnson, Ph.D., of the Institute of Health Economics at the University of Alberta, Edmonton. So Dr. Johnson and a colleague, Lauren Brown, set out to do a systematic review of available evidence on this issue to examine whether the risk of type 2 diabetes in people with psychotic illnesses who are taking atypical antipsychotic agents is actually raised compared with individuals not taking these medications.
Dr. Johnson and Ms. Brown searched several electronic literature databases, including the Cochrane Library and Medline, to collect evidence for their study. The researchers identified 228 studies relating to this issue, of which 22 matched the inclusion criteria. The studies included those looking at populations of individuals diagnosed with schizophrenia or schizoaffective disorder who had been treated with clozapine, olanzapine, quetiapine, or risperidone, with type 2 diabetes diagnosis as an outcome. Selected studies were restricted to case-control trials or those with randomized controlled or cohort design.
Of the 22 studies selected for evaluation in the systematic review, 17 were retrospective cohort design and 5 were case-control studies. According to the researchers, the resulting group of investigations included a heterogeneous mix of psychotic illnesses, drugs, dosages, and comparison treatments, making the analysis of underlying effects difficult to ascertain. Heterogeneity between studies was related to study design, length of study, the stage at which the illness was diagnosed, medication dose, age of study population, and study quality.
By pooling the results of the studies with all treatments, Dr. Johnson and Ms. Brown calculated that the odds ratio for a diagnosis of diabetes in people taking any atypical antipsychotic agent compared with controls was 1.16 (95% confidence interval 1.01–1.33) in favor of the treatment, suggesting that the treatment slightly increased the risk of type 2 diabetes. But when Dr. Johnson and Ms. Brown did a separate analysis looking at diagnoses of type 2 diabetes in people taking just two drugs—risperidone or olanzapine—the odds ratio was 1.10 (95% CI 0.96–1.27), with the effect slightly bigger with olanzapine.
“Based on inconsistent results of the available observational studies, it is unclear whether atypical antipsychotic agents increase risk of diabetes compared with controls,” Dr. Johnson said. He also cautioned that the overall effect sizes must be interpreted with care “due to a significant heterogeneity between studies, and because the overall effect sizes were calculated based on unadjusted odds ratios.”
Dr. Johnson added that “Until more information regarding the relationship between atypical antipsychotic agents and diabetes is available, individuals taking atypical antipsychotics should have baseline and follow-up metabolic evaluations.”