Transcutaneous auricular vagus nerve stimulation in perioperative neurocognitive disorders: clinical evidence and mechanistic perspectives

Background Perioperative neurocognitive disorders (PND) are common neurological complications in older surgical patients and are associated with impaired recovery, loss of independence, and worse quality of life. Transcutaneous auricular vagus nerve stimulation (taVNS) is a non-invasive neuromodulatory intervention with plausible effects on inflammatory, autonomic, sleep, pain, and cognitive-network pathways. Methods We conducted a PROSPERO-registered systematic review of direct perioperative randomized clinical evidence and, when outcomes were sufficiently comparable, an exploratory meta-analysis. PubMed/MEDLINE, Embase, Web of Science Core Collection, and CENTRAL were searched from inception through 5 August 2026. Risk of bias was assessed with RoB 2 and certainty with GRADE. Mechanistic and translational evidence was synthesized separately using a structured narrative approach. Results Four randomized trials enrolled or randomized 480 participants, of whom 469 contributed to reported analyses. Two trials involving 220 participants assessed related but non-identical early postoperative cognitive-decline constructs: one used dNCR and the other investigator-defined traditional POCD. Both studies individually favored taVNS, and the individual study estimates were treated as the primary evidence. A post hoc exploratory cross-construct synthesis yielded an RR of 0.34 (95% CI 0.18–0.64). Because the contributing trials used non-equivalent cognitive definitions, stimulation regimens, and comparator conditions, this secondary pooled estimate should not be interpreted as representing either a single standardized cognitive outcome or the effect of a single standardized taVNS intervention. A separate multicenter trial reported postoperative delirium in 12.0% versus 25.3% of participants (risk ratio 0.47, 95% confidence interval 0.23–0.98), and another trial showed more favorable postoperative Montreal Cognitive Assessment trajectories. Beyond clinical outcomes, the broader mechanistic literature supports a multilevel neurophysiological rationale for taVNS, consistent with possible modulation of auricular vagal afferent–nucleus tractus solitarius signaling, locus coeruleus and autonomic regulation, inflammatory reflex pathways, splenic sympathetic–cholinergic immune modulation, blood–brain barrier/neurovascular unit protection, glial responses, sleep–arousal regulation, and cognitive-network plasticity. However, most mechanistic evidence remains derived from preclinical or non-perioperative studies, and no causal pathway linking target engagement to cognitive improvement has been established. Conclusion Current evidence suggests a promising but preliminary clinical signal for taVNS in perioperative neurocognitive disorders. However, certainty remains low to very low, and routine clinical implementation is not yet supported. Available adverse-event data are insufficient to establish perioperative safety or to exclude uncommon cardiovascular or autonomic adverse events. Standardized PND outcomes, credible sham conditions, optimized stimulation dose and timing, systematic adverse-event ascertainment, multimodal target-engagement biomarkers, and risk-enriched confirmatory trials are priorities for future development. Systematic review registration https://www.crd.york.ac.uk/PROSPERO/view/CRD420261468039 , Identifier: CRD420261468039.