Background: The association between depression and metabolic syndrome is becoming more obvious. Waist circumference (WC) might be the most important metabolic syndrome (MetS) feature in relation to late-life depression, with a possible mediating role for adiponectin. Methods: Cross-sectional population based survey of 1277 participants (50-70 years). We measured all components of MetS, plasma adiponectin levels and depressive symptoms using Beck Depression Inventory (BDI). Principal components analysis on the BDI items revealed two factors, representing a cognitive-affective and a somatic-affective symptom-cluster. Multiple linear regression models with the BDI sum score and both depression symptom-clusters as dependent variables, respectively, were used to examine the association with each component of metabolic syndrome adjusted for confounders. We explored sex-differences as well as a hypothesised mediating effect of adiponectin. Results: The presence of MetS as well as number of metabolic risk factors were significantly associated with BDI sum score. In men WC, triglycerides and HDL cholesterol explained variance in depressive symptoms, whereas in women this effect was confined to WC. Moreover, irrespective of sex, all associations were primarily driven by the somatic-affective symptom cluster, Adiponectin neither mediated nor moderated any of the associations found. Limitations: Cross-sectional design limits causal interpretation. Being a population based survey, some selection bias might have occurred toward healthier part of population. Conclusions: Although pathophysiological mechanisms underlying the association between metabolic disturbances and depression remains to be elucidated, our study points to sex differences as well as a specific phenotype of depression that is associated with metabolic disturbances. (C) 2013 Elsevier B.V. All rights reserved