Differential association of insulin resistance with cognitive and somatic symptoms of depression

To examine the associations of depressive symptoms with insulin resistance, evaluating somatic and cognitive depressive symptoms separately.

A total of 328 individuals (mean age 60 years) referred for exercise stress testing, taking part in the Mechanisms and Outcomes of Silent Myocardial Ischemia study, completed the Beck Depression Inventory II. A fasting venous blood sample was collected for assessments of insulin and glucose level; the HOMA-IR (homeostatic model assessment of insulin resistance) was calculated. In principal component analysis, Beck Depression Inventory II items were forced to load onto two components (somatic and cognitive depressive symptoms). Adjusting for age, sex, BMI, medication use, smoking, physical activity, diabetes and cardiovascular disease, general linear model analyses were conducted to examine the associations between the components and log HOMA-IR .

Principal component analysis showed that nine items loaded onto a cognitive depressive symptoms component and 10 items loaded onto a somatic depressive symptoms component. When examined separately, both components were significantly associated with log HOMA-IR however, when including both components simultaneously in the model, only somatic depressive symptoms remained significantly associated with log HOMA-IR. Back-transformed, a one-unit change in somatic depressive symptoms was associated with a 1.07 (95% CI 1.002, 1.14) change in HOMA-IR and a one-unit change in cognitive depressive symptoms was associated with a 1.03 (95% CI 0.97, 1.14) change in HOMA-IR.

Somatic depressive symptoms seem to be more strongly associated with insulin resistance than do cognitive depressive symptoms. Monitoring somatic depressive symptoms may be more appropriate than monitoring cognitive depressive symptoms among depressed individuals with high insulin resistance.