Abstract
CONTEXT Neural substrates that may be responsible for the high prevalence of depression in type 1 diabetes mellitus (T1DM) have not yet been elucidated. OBJECTIVE To investigate neuroanatomic correlates of depression in T1DM. DESIGN Case-control study using high-resolution brain magnetic resonance images. SETTINGS Joslin Diabetes Center and McLean Hospital, Massachusetts, and Seoul National University Hospital, South Korea. PARTICIPANTS A total of 125 patients with T1DM (44 subjects with ≥1 previous depressive episodes [T1DM-depression group] and 81 subjects who had never experienced depressive episodes [T1DM-only group]), 23 subjects without T1DM but with 1 or more previous depressive episodes (depression group), and 38 healthy subjects (control group). MAIN OUTCOME MEASURES Spatial distributions of cortical thickness for each diagnostic group were compared with the control group using a surface-based approach. Among patients with T1DM, associations between metabolic control measures and cortical thickness deficits were examined. RESULTS Thickness reduction in the bilateral superior prefrontal cortical regions was observed in the T1DM-depression, T1DM-only, and depression groups relative to the control group at corrected P < .01. Conjunction analyses demonstrated that thickness reductions related to the influence of T1DM and those related to past depressive episode influence were observed primarily in the superior prefrontal cortical region. Long-term glycemic control levels were associated with superior prefrontal cortical deficits in patients with T1DM (β = -0.19, P = .02). CONCLUSIONS This study provides evidence that thickness reduction of prefrontal cortical regions in patients with T1DM, as modified by long-term glycemic control, could contribute to the increased risk for comorbid depression.