Impaired glucose tolerance is observed in depressed patients, and patients suffering from depression have an increased risk to develop diabetes mellitus. In depressed and diabetic patients, studies have shown both a beneficial effect of antidepressants on glucose homeostasis and the opposite. This review aims to structure the conflicting data and focuses on the question, which effect specific antidepressants have on glucose homeostasis. We therefore performed a systematic review of all available studies referenced in Medline from 1960 to 2011. We included antidepressant agents indexed in the Anatomical Therapeutic Chemical (ATC) classification system of the WHO in 2011 and searched for studies investigating their effects on glucose metabolism in clinical samples as well as in healthy subjects. Of 876 studies screened we included 66. Most studies had small sample sizes and lacked a placebo group limiting conclusions about antidepressant effects on glucose tolerance. However, some evidence points to beneficial effects on glucose homeostasis of hydrazine-type monoamine oxidase inhibitors (MAOIs) and selective serotonin reuptake inhibitors (SSRIs). In case of SSRIs, the effect is more pronounced in diabetic patients or patients with comorbid depression and diabetes mellitus. Noradrenegic substances (and possibly also dualacting antidepressants), in contrast, may deteriorate glucose tolerance. They can be used in depressed patients when favorable effects on mood outweigh adverse metabolic effects, but in depressed diabetics this can be at the expense of worsening of glycemic control. The effects of other antidepressants, like bupropione, mirtazapine or newer agents, require further investigation before reliable conclusions can be made. The synthesis of the findings is discussed in light of the specific pharmacodynamic properties of the antidepressants as well as the pathophysiological changes in depression and impaired glucose homeostasis, including animal studies.