In two European countries with a different prevalence of depression, namely Denmark (high) and.Spain (low), we assessed whether the mediation effect of emotional eating between depression and Body Mass Index (BMI) as found in earlier studies can be replicated and whether this mediation effect is contingent on 1) change in appetite and 2) gender. Mediation and moderated mediation was assessed with Hayes' PROCESS macro in SPSS. Emotional eating (DEBQ: Dutch Eating Behavior Questionnaire), depressive symptoms (CES-D: Center for Epidemiologic Studies Depression Scale), change in appetite, weight and height were self-reported. In both countries, emotional eating acted as a mediator between depression and BMI (Denmark: B = 0.03 (SE = 0.01), 95% CI, [0.03, 0.05]; Spain: B = 0.03 (SE = 0.01), 95% CI, [0.02, 0.04]). In Denmark this mediation effect was stronger for participants with increased appetite and for females than for participants with decreases/no change in appetite and for males (more appetite: B = 0.08, (SE = 0.03), [0.03, 0.15]; decreased appetite/no change in appetite: B = 0.03 (SE = 0.01), [0.02, 0.04]); females: B = 0.05 (SE = 0.01), [0.03, 0.07]; males: B = 0.01 (SE = 0.01), [0.004, 0.04]. This supports depression with atypical features as an underlying mechanism in the mediation effect of emotional eating. In Spain there was no support for depression with atypical features as underlying mechanism because the mediation effect was neither moderated by change in appetite nor by gender. Instead, post hoc analyses suggested 'stress of unemployment' as possible explanatory factor of the mediation effect, with stronger mediation effects for unemployed than for employed people (unemployed: B = 0.05 (SE = 0.01), [0.03, 0.07]; employed B = 0.02 (SE = 0.01), [0.01, 0.04]). The mediating effect of emotional eating between depressive symptoms and body mass index in both countries suggests that obesity interventions should take emotional eating into account. (C) 2016 Elsevier Ltd. All rights reserved.