Background: Patients with type 2 diabetes (DM2) are a greater risk of cognitive dysfunction, ranging from cognitive decrements to dementia. Several mechanisms have been implicated, including oxidative stress, insulin-resistance, hypoglycemia, and vascular/neuronal injury. However, pathogenesis is poorly understood and there is lack of effective treatment, though GLP1-RA appears promising in preclinical studies. Aim: To identify factors associated with mild cognitive decrements in DM2 patients. Materials and methods: The study was performed on DM2 patients consecutively enrolled in the TESEO study. We excluded patients with history of TIA, stroke, carotid revascularization, >70% internal carotid stenosis, cognitive dysfunction affecting daily activities, liver cirrhosis, and hypoglycemic episodes. We collected clinical data, screened for complications (eGFR, ACR, fundus oculi, ECG, echocardiography, Doppler-TSA) and administered the Repeatable Battery for the Assessment of Neuropsychological Status (R-BANS) test to assess cognitive performance. Logistic regression analysis was used to identify variables independently associated with a low R-BANS (L-RBANS<80). Results: Patients (n=329, age 61.5±7.9 years, 56.5% male) had a short DM2 duration (3.7±5.0 years) and good metabolic control (HbA1c 6.56%±1.02). Mean R-BANS score was 95.4±16.9 and L-RBANS was found in 24% of the subjects. L-RBANS was less frequent in men (p=0.012), in people with higher education (p<0.001) and in those treated with GLP1-RA (p=0.029). People on GLP1-RA (n=99, mono/combination therapy) had significantly higher R-BANS score (99.4±16.7 vs 93.7±16.7, p=0.005) and performed significantly better in Language, Immediate Memory, and Delayed Memory, while no association was found with other antihyperglycemic drugs. In logistic regression analysis, higher education and GLP1-RA use were associated with 85% and 64% lower OR of L-RBANS, respectively, independent of age, sex, BMI, diabetes duration, HbA1c, ACR, CVD and valvular disease, atrial fibrillation, and treatment with hypoglycemic drugs. Conclusions: The present study showed that education and use of GLP1-RA are independently and significantly associated with mild cognitive decrements in DM2 patients.