Depression is associated with increased insulin resistance: a meta-analysis

Background Depression has been linked to an increased incidence of clinical comorbidity, such as diabetes mellitus (DM) [1]. Moreover, a bidirectional relationship between depression and DM has been described, with depressive symptoms being predictive of poor glycemic control in type 2 DM and DM increasing the incidence of depression [2]. To date, increased levels of insulin, indicating metabolic alterations, have been described in schizophrenia and bipolar disorder [3,4]. However, the presence of increased levels of insulin in depression remains unclear. The aim of this study is to verify if peripheral insulin levels are altered in acute or remitted depression compared to healthy controls. We also seek to determine whether there are changes on insulin levels after clinical resolution of an index depressive episode following antidepressant treatment. Methods A systematic search for all eligible peer-reviewed articles was conducted using Medline, the Cochrane Library and Scopus according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses statement (PRISMA Statement). No language, year or country restrictions were used. The last search was performed in August of 2016. We included studies assessing serum or plasma fasting insulin in adults with depression alongside a control group of healthy volunteers. Two reviewers performed the study selection and data extraction. We identified 8,769 studies, of which thirty-eight fulfilled inclusion criteria. We carried out three between-group meta-analyses examining differences in serum or plasma insulin levels between healthy controls and: a) adult subjects with acute depression with psychiatric medication, b) adult subjects with acute depression without psychiatric medication and c) adult subjects with remitted depression with psychiatric medications. We also conducted a within-group meta-analysis comparing changes on insulin levels before and after treatment of an index depressive episode with antidepressants. Standardized mean difference estimates in insulin levels were calculated for all meta-analyses. Sensitivity analyses and meta-regressions according to mood state and psychiatric medication status were also performed. As the analyses showed that the studies were heterogeneous, effect size results from different studies were pooled according to the inverse variance method accounting for random effects. Results Thirty-eight studies were included, providing data on 31,763 participants. Of all the participants, 7,142 were subjects with depression and 24,623 were healthy controls. Insulin levels were moderately increased in subjects with acute depression compared with healthy controls, both in patients treated with psychiatric medication (g=0.38, 95% CI 0.17 to 0.59, p Conclusions Our findings show that insulin levels are increased in acute depression and normal in remission. These results highlight a potential use of insulin as a biomarker of disease activity in depression. Further studies are warranted.