The GERODIAB follow-up survey is the first French multi-centre, prospective, observational study designed to analyze the influence of glycaemic control on morbidity-mortality over 5years in type 2 diabetic patients aged 70years and over. This report analyses the factors associated with cognitive decline at inclusion.

A total of 987 patients with well-maintained autonomy were consecutively included in 56 French diabetic centres. Cognition was systematically explored using the Mini-Mental State Examination (MMSE). Cognitive disorders were defined by previously known dementia and/or MMSE scores≤24/30. Results were analyzed using the Kruskal–Wallis or the chi² tests. Multivariate analysis was a stepwise logistic regression.

Two hundred and eighty-four patients (28.8%) had a cognitive disorder; they were older (78.4±5 vs. 76.5±4.8year; P<.001) and were more frequently women (60.2 vs. 48.8%; P<.001). They had a similar duration of diabetes (17.6±11.1 vs. 17.9±10.7years; P=.651), but a higher HbA_1c level (7.8±1.5 vs. 7.5±1.2%; P=.002). They were treated with insulin more often (65.1% vs. 54.3%; P=.002) and less often with metformin (40.1% vs. 52.3%; P<.001). In multivariate analysis, cognitive disorders were associated with age, gender and HbA_1c, successively (concordance 64.1%, P<.001). Of the complications of diabetes, cognitive impairment was associated with peripheral neuropathy (37.3 vs. 24.5%; P<.001), cerebrovascular involvement (20.8 vs. 13.8%; P=.007) and heart failure (14.1 vs. 8.5%; P=.009).

Cognitive disorders were associated with HbA_1c level and some complications of diabetes, suggesting similar pathological mechanisms. Systematic screening for cognitive impairment could be beneficial in elderly type 2 diabetic patients.