ObjectiveTo examine whether symptoms of depression or anxiety predict glycaemia and incident diabetes complications four years later, and whether diabetes self-care behaviours mediate these associations, in adults with type 1 diabetes (T1DM). MethodsData of 205 adults with T1DM from the 2011 and 2015 Diabetes MILES–Australia surveys were analysed. Variables of interest were: baseline depression and anxiety (PHQ-8 and GAD-7, respectively) symptoms; HbA1c and incident complications at four-year follow-up; and self-care behaviours at both time points. Longitudinal associations were analysed using structural equation modelling. ResultsForty-two participants (20.6%) reported incident complications. Baseline depressive symptoms predicted higher HbA1c at follow-up indirectly via less optimal self-care at follow-up (β = 0.19, P = 0.011). Baseline anxiety was not independently associated with HbA1c or self-care at follow-up (P ≥ 0.64). Neither depressive nor anxiety symptoms predicted incident complications, although depressive symptoms were associated with less optimal self-care at baseline (β = -0.67, P < 0.001), and this predicted microvascular complications (β = -0.38, P = 0.044); however, the indirect association via self-care was not significant (β = 0.25, P = 0.067). LimitationsParticipants were self-selected; all study variables were assessed using self-report measures; and adjusting for baseline HbA1c was not possible. ConclusionsDepressive symptoms predicted suboptimal self-care behaviour and glycaemic outcome four years later, while anxiety symptoms did not. The findings suggest that tailored diabetes care should take the potential impact of comorbid depression into consideration to help people improve their diabetes self-care and achieve best possible health outcomes.