RESUMEN
BACKGROUND: Diabetes and depression are reciprocally linked, but few studies modeled their interplay considering the influence of affective temperaments (AT) and demographic factors. METHODS: Participants with type 1 and type 2 diabetes (T1DM and T2DM, n=279) recruited from Diabetes Units were assessed with the Beck Depression Inventory (BDI), Temperament Evaluation of Memphis, Pisa, Paris and San Diego-autoquestionnaire version (TEMPS-A), Morisky Medication Adherence Scale (MMAS), Diabetes Distress Scale (DDS) and Cumulative Illness Rating Scales (CIRS). Glycosylated hemoglobin levels (HBA1C) was used as index of glycemic control. The bi-directional association between glycemic control, depression and candidate mediators was examined with Structural Equation Modeling, testing the impact of moderator variables (AT, diabetes type, age and gender) with multigroup comparison. RESULTS: The association between HBA1C and depressive symptoms was mediated by diabetes-related distress,, while there was no definite evidence of depression influencing HBA1C through changes of adherence, tiredness, appetite, alcohol intake or smoking. Among individuals with AT, distress was unrelated to HBA1C and had a higher impact on depression; adherence was inversely association with HBA1C. Moreover, physical comorbidities impacted on depression. While diabetes type had a moderation role, age and gender did not affect the model. LIMITATIONS: Cross sectional design, lack of objective measures of diet and physical activity. CONCLUSIONS: Glycemic control seem to influence the severity of depressive symptoms, but the reciprocal association seems non-significant. AT and diabetes type may shape this relationship influencing distress and adherence to medications. Findings may aid interventions aimed at improving patients' care and quality of life.