Several studies have found significant cognitive impairment in patients with diabetes compared with nondiabetic patients. Only a few studies have evaluated the change in cognition over time, but they had inadequate controls for depression and education level. In 1999, a comprehensive review of the published literature suggested a correlation between type 2 diabetes and cognitive impairment affecting memory and executive function. In an earlier review, the most commonly affected test of cognitive ability was verbal memory. Because of these conflicting results, Kanaya and associates assessed the cognitive performance in older adults according to glucose-tolerance status.

The study included men and women 42 to 89 years of age who were followed over four years. Cognitive function was assessed on at least three different occasions using the Mini-Mental State Examination (MMSE), a verbal fluency test, and the Trail-Making Test B. Participants were divided into normal, impaired glucose tolerance, and diabetic groups based on World Health Organization criteria. Other data collected included demographic information, health-related behaviors, clinical history, presence of coronary artery disease, and medication use. Height, weight, calculated body mass index, and blood pressures were recorded. In addition, patients were assessed for depression using 18 of 21 items from the Beck Depression Inventory. Laboratory data included fasting glucose levels and glucose levels two hours after a 75-g oral glucose load, lipid panels, and other markers. Participants also were assessed for nephropathy and retinopathy.

A total of 999 white men and women participated in the study. No differences in cognitive function scores were noted among the three glucose-tolerance groups at baseline. After four years, women in the diabetes group had a fourfold increase in the risk of cognitive decline on the verbal fluency test compared with nondiabetic women. The mean verbal test scores at the end of the study were 15.2 for women with diabetes, 16.7 for women with impaired glucose tolerance, and 17.2 for women with normal glucose tolerance (see accompanying figure). There was no difference in cognitive decline in men when comparing the three groups. Lipid levels, blood pressure, and the existence of microvascular or macrovascular disease had no impact on the decline in verbal fluency in women during the study. There were no significant differences in MMSE or Trail-Making Test B scores among women in the three glucose-tolerance groups.

The authors conclude that older women with diabetes have a more rapid decline in their performance on the verbal fluency test compared with women who have impaired or normal glucose tolerance. They add that better glucose control may reduce this decline in verbal function in women with diabetes.