Abstract
BACKGROUND: Type 2 diabetes (T2D), affecting approximately 12% of the global population and over 30% of older adults, is among the most prevalent and fast-growing risk factors for neurodegenerative disorders. Evidence is lacking on whether specific glucose-lowering agents may reduce the risk of neurodegeneration onset in people living with T2D. METHODS: In this retrospective cohort study, we utilized the TriNetX platform, which contains electronic health records of over 170 million people worldwide. We propensity-score matched (1:1) people with T2D lacking evidence of neurodegeneration who initiated glucagon-like peptide-1 receptor agonists (GLP-1 RAs) or dipeptidyl peptidase-4 inhibitors (DPP4i) (2010-2021). In a separate analytical cohort, we compared individuals initiating GLP-1 RA with those initiating basal insulin. Follow-up continued for ≤ 5 years. We used Cox proportional-hazard regression models to assess the risk of the composite outcome of developing new neurodegenerative conditions, including Alzheimer's disease, Parkinson's disease, dementia subtypes, and other synucleinopathies. We also assessed each component individually. Analyses were repeated among subgroups defined by sex, age, and the specific GLP-1 RA initiated. RESULTS: Overall, 214,442 matched individuals initiated GLP-1 RAs or DPP4i (109,731 women, mean age 58.6 years [SD 12], and mean HbA1c 7.7% [1.4]). During a 4.0-year mean follow-up, neurodegenerative disorder onset occurred in 2,393 (2.2%) and 3,062 (2.9%) people initiating GLP-1 RAs and DPP4i, respectively (hazard ratio of 0.81 [95% CI 0.77 to 0.86]; absolute risk difference - 0.6% [- 0.8 to - 0.5]). The associations were separately observed among women (0.78 [0.72 to 0.84]) and men (0.90 [0.83 to 0.98]), individuals aged ≥ 65 years old (0.82 [0.78 to 0.87]) or < 65 years old (0.84 [0.70 to 1.00]), and in those initiating semaglutide (0.75 [0.67 to 0.84]), liraglutide (0.77 [0.70 to 0.84]), or dulaglutide (0.82 [0.77 to 0.88]). The hazard ratios for dementia, Alzheimer's disease, vascular dementia, and Parkinson's disease onset were 0.76 [0.72 to 0.81], 0.77 [0.68 to 0.87], 0.75 [0.67 to 0.85], and 1.04 [0.93 to 1.17] with GLP-1 RAs versus DPP4i, respectively. The results were in the same direction when comparing individuals initiating GLP-1 RAs with those initiating basal insulin. CONCLUSIONS: In a real-world cohort of people living with T2D with a multinational representation, the initiation of GLP-1 RAs, compared to DPP4i or basal insulin, was associated with a lower risk of new-onset neurodegeneration. These data support the rationale for dedicated clinical trials to assess the potential neuroprotective properties of GLP-1 RAs in this population.