Longitudinal effects of depression on glycemic control in veterans with Type 2 diabetes.

OBJECTIVES

To examine the longitudinal effects of depression on glycemic control in veterans with Type 2 diabetes.

METHODS

Data on 11,525 veterans with Type 2 diabetes were analyzed. A person-period dataset for each subject to cover 3-month intervals (36 time intervals) from April 1997 to March 2006 was created. Subjects were classified as depressed based on ICD-9 codes for depression. General linear mixed model regression was used to examine changes over time in HbA(1c) levels and whether the changes from baseline were different in depressed and nondepressed diabetic veterans, sequentially adjusting for baseline age, demographic variables and comorbidities (coronary heart disease, stroke and hypertension). Pooled t-tests were used to compare unadjusted mean HbA(1c) at each time point across the depressed and nondepressed groups. SAS was used for statistical analysis.

RESULTS

Ninety-seven percent were men, 48% were white, 27% were blacks and 25% were other. Mean age was 66 years and mean follow-up period was 4.1 years. Six percent (696/11,525) of the sample had diagnosed depression. Unadjusted mean HbA(1c) values were significantly higher in depressed vs. nondepressed subjects at all time points. The adjusted mean HbA(1c) values over time in the final mixed model were significantly higher in depressed vs. nondepressed subjects (mean difference of 0.13; 95% CI [0.03; 0.22]; P=.008). In all adjusted models, differences in mean HbA(1c) values were significantly higher in depressed vs. nondepressed subjects with Type 2 diabetes.

CONCLUSION

This study of veterans with Type 2 diabetes demonstrates that there is a significant longitudinal relationship between depression and glycemic control as measured by HbA(1c) and that depression is associated with persistently higher HbA(1c) levels over time.