Vagus nerve stimulation neuromodulation for refractory epilepsy: a systematic review and network meta-analysis of randomized controlled trials
Abstract
Background: Refractory epilepsy is a condition that requires multimodal and specialized clinical management. In this context, neuromodulation therapies have emerged in recent years, notably two different approaches involving the same nerve structure: Transcutaneous Auricular Vagus Nerve Stimulation (taVNS) and Invasive Vagus Nerve Stimulation (iVNS). Objective: To review the effectiveness of the two primary forms of Vagus Nerve stimulation in the treatment of refractory epilepsy. Methods: A network meta-analysis conducted according to PRISMA 2020 guidelines. The PICO strategy guided the search in PubMed/MEDLINE, Embase, and Cochrane Library in February 2026, without language or date restrictions. Randomized Controlled Trials (RCTs) involving patients diagnosed with drug-resistant epilepsy treated with iVNS or taVNS versus sham control were included. Statistical analysis was performed using R software (5.4.2). The dichotomous outcome evaluated was the responder rate (reduction in at least 50% of seizures), using Risk Ratio (RR) as the effect measure, analyzed via the Mantel-Haenszel (M-H) method in a random-effects model (DerSimonian-Laird). Heterogeneity was measured by tau and I² (Cochran's Q). P-Scores were calculated to rank therapies in the network model. Risk of bias was assessed using the RoB 2.0 tool. Results: Six RCTs were included (3 comparing taVNS vs. Sham; 3 comparing iVNS vs. Sham), totaling 622 patients. A significant difference was found favoring taVNS over Sham (RR=1.58; 95% CI [1.04; 2.40]; p=0.031) regarding responder rates, which was not observed for iVNS vs. Sham (RR=1.51; 95% CI [0.94; 2.42]; p=0.089). Overall heterogeneity was low (I² = 10.2%; tau = 0.13). No significant difference was found in the network comparison between iVNS and taVNS (RR=0.96; 95% CI [0.51; 1.80]). P-Scores ranked taVNS as the superior treatment (0.7704), followed by iVNS (0.6993) and Sham (0.0303). Regarding the overall risk of bias: 1 study was "low risk," 3 had "some concerns," and 2 were "high risk”. Conclusion: Both taVNS and iVNS show promising results in managing refractory epilepsy. However, the limited number of RCTs and the lack of direct head-to-head comparisons are significant limitations. Larger, methodologically robust RCTs are necessary for more definitive conclusions.