The effect of long-term past glycemic control on executive function among patients with type 2 diabetes mellitus.

OBJECTIVE: Patients with type 2 diabetes mellitus (T2DM) show more executive dysfunction than nondiabetics. However, how long poor glycemic control affects executive function remains unclear. Thus, we aimed to investigate the relationships in a cross-sectional study. METHODS: We studied 118 T2DM outpatients (age, ≥ 60 years; excluding history of stroke, dementia and severe hypoglycemia). HbA1c values were recorded every ≤ 12 weeks for ≥ 5 years. All patients underwent verbal-fluency tests (reflecting executive function) and Mini-Mental State Examination (MMSE). The correlation between past glycemic control values and both cognitive tests scores was investigated. As markers of past glycemic control, we used average hemoglobin A1c (HbA1c) values and glycemic control variability [coefficient of variation (CV) of HbA1c values (HbA1c-CV)]. RESULTS: Verbal-fluency tests scores correlated with HbA1c-CV, but not with average HbA1c values, after adjusting for age, years of education and sex. Verbal-fluency tests scores correlated with HbA1c-CV for the past 5 years, best compared with HbA1c-CV for past < 5 years. MMSE scores were also related to only HbA1c-CV for the past 3 years in an adjustment model. CONCLUSIONS: Five-year HbA1c variability affected executive function in T2DM patients, but not average HbA1c values. Long-term longitudinal studies may be required.