BACKGROUND: Parkinson's disease (PD) is an increasingly prevalent neurodegenerative condition, particularly among individuals with type 2 diabetes mellitus (T2DM). While prior studies have suggested that GLP-1 receptor agonists (GLP-1RAs) and metformin may confer neuroprotective effects, most were limited by small sample sizes, short follow-up, uncontrolled designs, or lacked direct comparisons between therapies-making their findings inconclusive for clinical decision-making.
METHODS: Using the TriNetX Global Collaborative Network, we identified 92 485 patients with T2DM initiating GLP-1RA therapy and matched them 1:1 to new metformin users using propensity score matching. Patients with prior antidiabetic therapy, PD, or dementia were excluded. Incident PD was the primary outcome; all-cause mortality served as a secondary endpoint. Adjusted hazard ratios (aHRs) were estimated using Cox models. Competing risk and time-stratified (≤5, 5-10, >10 years) analyses were conducted. Positive and negative outcome/exposure controls were employed to validate internal consistency.
RESULTS: The overall PD risk was comparable between GLP-1RA and metformin groups (aHR, 0.91; 95% CI, 0.79-1.05). However, between years 5 and 10, GLP-1RA use was associated with a significantly lower PD risk (aHR, 0.56; 95% CI, 0.34-0.93). Mortality risk did not differ significantly. Validation analyses confirmed the specificity of the findings.
CONCLUSIONS: By addressing key limitations of earlier studies through a large-scale, active-comparator, new-user design, this study provides novel evidence of a delayed neuroprotective effect of GLP-1RAs. These findings support incorporating neurologic outcomes into long-term diabetes management and may inform therapy selection in high-risk populations.
