Cognitive impairment and dependence of patients with diabetes older than 65 years old in an urban area (DERIVA study)

Emiliano Rodríguez-Sánchez, Sara Mora-Simón, María C. Patino-Alonso, Diana Pérez-Arechaederra, José I. Recio-Rodríguez, Manuel A. Gómez-Marcos, Luis F. Valero-Juan, Luis García-Ortiz, for the DERIVA Group

We analyzed the associations between diabetes and cognitive impairment (CI) and dependence in a population of patients 65 years or older. Cross-sectional study. We randomly selected 311 participants over the age of 65 living in an urban area of Spain. The mean age of the cohort was 75.89 ± 7.12 years, and 69 of the individuals (22.2 %) had diabetes. Two questionnaires were used to assess cognitive performance (MMSE and Seven Minute Screen Test), and two assessments were used to evaluate patient dependence (Barthel Index and Lawton-Brody Index). Clinical information and sociodemographic data were also gathered. Nearly one quarter of patients with diabetes (21.7 %) lived alone. Diabetic patients were more sedentary (p = .033) than non-diabetic patients. Roughly one sixth (15.3 %) of the diabetics and 10.1 % of the non-diabetics were depressed (p = .332). CI was present in 26.1 % of the diabetics and 14.5 % of non-diabetics (p = .029). Diabetic patients had a MMSE score that was significantly worse than non-diabetics (24.88 ± 4.74 vs 26.05 ± 4.03; p <.05), but no differences were found in the Seven Minute Screen Test. Logistic regressions revealed that the presence of diabetes was independently associated with CI (adjusted for age, gender, years of education, sedentary lifestyle, body mass index, diastolic blood pressure, cholesterol, and depression (OR = 2.940, p = .013). Patients with diabetes showed greater dependence, as measured by the Barthel Index (p = .03) and Lawton-Brody Index (p <.01). Nevertheless, when dependence (dependence or not dependence for each questionnaire) used as a dependent variable in the logistic regression analyses, no significant association with diabetes was found, after adjusting for confounding variables. Diabetic patients over the age of 65 are more likely to present CI but not dependence. These findings support the need to include both a functional and cognitive assessment as necessary components in a standard evaluation in both clinical guides and randomized trials of therapeutic interventions in patients with diabetes.