Interventions for Cognitive Impairment in Type 2 Diabetes Mellitus: An Umbrella Review of Pharmacological and Lifestyle Strategies, Mechanisms, and Clinical Implications

Background: Type 2 diabetes mellitus (T2DM) is one of the fastest-growing epidemics associated with cognitive impairment and dementia. This reduces quality of life among patients with diabetes. Several trials on the management of cognitive impairment have been conducted, although their effectiveness and mechanisms are still unclear. Aim: To synthesize evidence from systematic reviews and meta-analyses of RCTs examining interventions for cognitive impairment in T2DM, stratified by intervention category, comparator, patient characteristics, mechanisms, and clinical outcomes. Methods: We identified 11 systematic reviews (n = 213,604 participants) reporting the effects of cognitive impairment interventions through a comprehensive literature search. Using random-effects models, a meta-meta-analysis was conducted and heterogeneity assessed using I² statistics. AMSTAR and GRADE were used to assess the quality of reviews. Subgroup analyses were conducted for intervention category, cognitive severity, age, diabetes duration, sex, and follow-up duration. Results: Sodium glucose co-transporter inhibitors showed the strongest cognitive improvement effect (pooled OR = 0.44, 95% CI: 0.35–0.55; SUCRA = 94.5%), followed by GLP-1 receptor agonists (RR/OR = 0.60, 95% CI: 0.50–0.72; SUCRA = 92.1%). Among lifestyle interventions, multi-component exercise was the most effective (SMD = 0.69–1.04). The lower baseline cognition (MMSE < 23) benefited more from GLP-1 receptor agonists (Cohen's d = -3.00, p < 0.001). Advanced age (≥60 years) and prolonged intervention time (≥24 weeks for medications; ≥6 months for exercise) were associated with greater effects. Female gender showed greater benefit from GLP-1 receptor agonists (HR = 0.63 vs 0.81 for male gender). Conclusions: SGLT2 inhibitors and GLP-1 receptor agonists have the greatest effect on cognition in T2DM, and multi-component exercise is the most effective lifestyle component. The review therefore recommends individualized selection of interventions based on patient characteristics and baseline cognitive status. This calls for high-quality trials with prolonged follow-up to strengthen the evidence base.