Depression-related hyperglycemia in type 1 diabetes: a mediational approach.

OBJECTIVE Depression is linked with hyperglycemia and with an increased risk for diabetes complications, but the mechanisms underlying these relationships have not been established. In this study, we applied mediational analysis methods to determine whether the hyperglycemic effect of depression could be mediated by poor diabetes self-care. METHODS Depression symptoms and diabetes self-care activity were assessed in a primary care sample of 188 patients with type 1 diabetes by using the Hopkins Symptom Checklist-90 (SCL-90) and the Summary of Diabetes Self-Care Activities (SDSCA). A composite score of self-care activity was formed from SDSCA ratings for diet amount, exercise, and glucose testing. Degree of hyperglycemia (level of glycosylated hemoglobin [HbA1c]), weight, insulin dose, and other clinical characteristics were obtained from electronic medical records. Ordinary least-squares regression was used to determine the effect of depression on HbA1c level controlling for weight and insulin dose. The SDSCA score was then added to the regression model to determine whether it attenuated the effect of depression symptoms on HbA1c level, thus providing suggestive evidence of mediation from these cross-sectional data. RESULTS Depression symptoms, poor diabetes self-care, and hyperglycemia were correlated with one another in univariate analyses (p <.05). Depression symptoms were associated with higher HbA1c after controlling for weight and insulin dose (parameter estimate for depression 0.53, t = 3.6, p <.001). Inclusion of SDSCA in the model minimally attenuated the effect of depression symptoms (adjusted parameter estimate for depression 0.50, t = 3.3, p = .001). CONCLUSIONS These findings do not support mediation of the depression-hyperglycemia relationship by diabetes self-care behavior. Other pathways, including psychophysiological mechanisms, should be investigated.