Mendelian randomization analysis with the GEO database: Exploring the molecular mechanism underlying insulin therapy for perioperative neurocognitive disorders

Central insulin resistance is a significant factor in perioperative neurocognitive disorders (PND), yet the therapeutic effects and underlying molecular mechanisms of insulin remain unclear. We conducted a bidirectional Mendelian randomization (MR) study to investigate the causal relationships between insulin use and PND. The forward MR analysis revealed that insulin use considerably reduced the risk of delirium (odds ratio [OR] = 0.01, 95 % confidence interval [CI]: 0.0007–0.22, P = 0.003) and enhanced cognitive performance (OR = 8.03, 95 % CI: 2.83–22.78, P = 0.001). Importantly, the reverse MR analysis indicated that cognitive impairment or delirium did not causally affect insulin use (OR ≈ 1.0, P > 0.5). Utilizing GEO datasets, we identified 63 differentially expressed genes (DEGs) common to both postoperative delirium and insulin treatment. Subsequent analyses, including the construction of a protein-protein interaction network and Gene Ontology (GO) and KEGG pathway analyses, identified EPN2, DNAJC6, ARFGAP1, and HIP1R as key hub genes that could serve as potential therapeutic targets for PND. Further research showed that these hub genes are part of a multi-pathway interactive network, which may significantly contribute to the onset and progression of PND, with insulin also affecting their regulation. Our results establish a unidirectional genetic causality for the therapeutic effects of insulin in PND and reveal new molecular targets, thereby improving our understanding of the mechanisms behind insulin treatment for PND.