Genetic Insights Into Protective Effects of GLP-1 Receptor Agonists on Stroke and Alzheimer Disease Across Ancestries

Abstract

Background and Objectives Stroke and dementia are leading causes of death and disability worldwide, requiring enhanced preventive strategies. Glucagon-like peptide-1 receptor agonists (GLP-1RA), widely used for type 2 diabetes (T2D) and obesity, have demonstrated pleiotropic effects, including cardiovascular benefits, with mounting evidence also suggesting potential beneficial effects on brain health. The objective of this study was to investigate the genetically proxied effect of GLP-1RA on stroke, Alzheimer disease (AD), and related endophenotypes, in a cross-ancestry setting. Methods We conducted a 2-sample Mendelian randomization (MR) study using both a locus-based and a protein quantitative trait loci (pQTL)–based approach. The study leveraged summary statistics from large-scale genome-wide association studies (GWAS). These GWAS comprised participants of European, East-Asian, and African ancestry with and without T2D, stroke, and AD, and with measurements of fasting glucose, glycated hemoglobin, stroke, and AD endophenotypes. Sample size ranged from 3,026 to 1,812,017 individuals. Genetically proxied GLP-1RA exposure was instrumented using independent T2D-associated variants (p < 5 × 10−8) within GLP1R (locus-based MR) and pQTLs for circulating GLP-1R protein levels (pQTL-based MR). Instruments were validated for effects on glycemic traits and GLP1R gene expression. Primary outcomes were stroke (any, ischemic, cardioembolic, small vessel, large artery, and intracerebral hemorrhage) and AD. Secondary analyses included etiologic factors and endophenotypes of stroke and AD: heart failure, atrial fibrillation, MRI markers (white matter hyperintensity volume, perivascular space burden, and hippocampal volume), and CSF biomarkers (beta-amyloid, tau). Outcomes were stratified by ancestry and meta-analyzed. Results In Europeans, locus-based MR indicated significant protective associations of GLP-1RA with any stroke and ischemic stroke (pFDR < 0.05) and nominally (p < 0.05) with AD. Cross-ancestry meta-analyses strengthened protective effects of GLP-1RA on any stroke (OR, 0.75; 95% CI 0.63–0.89), ischemic stroke (0.72; 0.56–0.91), cardioembolic stroke (0.55; 0.35–0.86), and AD (0.51; 0.31–0.83), all with pFDR < 0.05. pQTL-based MR analyses suggested protective effects of genetically proxied GLP-1RA on atrial fibrillation (pFDR < 0.05) and, nominally (p < 0.05), on hippocampal perivascular space burden and CSF β-amyloid levels. Discussion GLP-1RAs may confer protective effects against stroke (especially ischemic and cardioembolic) and AD risk, across ancestries. These findings warrant further investigation in experimental models and clinical trials to confirm benefits on brain health and understand underlying mechanisms.