Psychiatric disorders as risk factors for type 2 diabetes: an umbrella review of systematic reviews with and without meta-analyses

Nanna Lindekilde, Femke Rutters, Jan Erik Henriksen, Mathias Lasgaard, Miranda T Schram, Katrine Hass Rubin, Mika Kivimäki, Giesje Nefs, Frans Pouwer

Research output: Contribution to journalJournal articleResearchpeer-review

75 Downloads (Pure)


Having a psychiatric disorder may increase the risk of developing type 2 diabetes[T2D] and this umbrella review aims to determine whether people with a psychiatric disorder have an increased risk of developing T2D and to investigate potential underlying mechanisms. A literature search was performed to identify systematic reviews of longitudinal studies investigating different psychiatric disorders as risk factors for incident T2D in humans (≥18 years). A total of 8,612 abstracts were identified, 180 full-text articles were read, and 25 systematic reviews were included. Six categories of psychiatric disorders were identified. Except for eating disorders, all psychiatric disorders were associated with increased risk of incident T2D ranging from RR = 1.18 [95% CI 1.12-1.24] to RR = 1.60 [95% CI 1.37-1.88] for depression; from RR = 1.27 [95% CI 1.19-1.35] to OR = 1.50 [95% CI 1.08-2.10] for use of antidepressant medication; from OR = 1.93 [1.37-2.73] to OR = 1.94 [1.34-2.80] for use of antipsychotic medication; from RR = 1.55 [95% CI 1.21-1.99] to RR = 1.74 [95% CI 1.30-2.34] for insomnia, and finally showed OR = 1.47 [95% CI 1.23-1.75] for anxiety disorders. Plausible underlying mechanisms were discussed, but in most reviews corrections for mechanisms did not explain the association. Notable, only 16% of the systematic reviews had a high methodological quality.

Original languageEnglish
Article number108855
JournalDiabetes Research and Clinical Practice
Number of pages14
Publication statusPublished - Jun 2021


Dive into the research topics of 'Psychiatric disorders as risk factors for type 2 diabetes: an umbrella review of systematic reviews with and without meta-analyses'. Together they form a unique fingerprint.

Cite this