Left vagus nerve magnetic stimulation facilitates nasogastric tube removal in post-stroke patients with dysphagia: a prospective observational cohort study

Background Post-stroke pharyngeal dysphagia (PSD) frequently leads to aspiration, aspiration pneumonia, and malnutrition, significantly compromising patient outcomes. Although standard swallowing rehabilitation training is widely applied, many patients show slow recovery with prolonged nasogastric tube placement. Recently, neuromodulation techniques such as vagus nerve magnetic stimulation (VNMS) have shown promise; however, observational evidence supporting its real-world clinical efficacy in post-stroke dysphagia patients remains limited. Methods This single-blind prospective observational cohort study was conducted from June 2024 to June 2025 at Guangdong Sanjiu Brain Hospital, China. A total of 78 stroke patients with dysphagia were enrolled (33 in intervention group, 45 in control group). The intervention group received left-sided VNMS combined with standard swallowing rehabilitation training, while the control group received standard training alone. Multiple confounder adjustment and causal inference approaches were applied to ensure the robustness of the results. Results Among 78 patients, after confounder adjustment, the intervention group demonstrated a significantly higher nasogastric tube removal rate, with an adjusted hazard ratio of 1.850 (95% CI: 1.103–3.101, p = 0.021). The intervention group removed nasogastric tubes on average 3.75 days earlier than the control group (95% CI: −5.01 to −2.50, p = 0.000), and the direction of effect was consistent across all analyses. Although swallowing safety assessment showed initial improvement, effects on overall swallowing function were inconsistent across analyses. Conclusion In this prospective observational cohort, left-sided VNMS significantly accelerated nasogastric tube removal in post-stroke dysphagia patients, with an average advancement of 3.75 days. However, the improvement in swallowing physiological parameters was inconsistent, suggesting that VNMS’s effects may be primarily concentrated on functional outcomes. Long-term efficacy and optimal candidate populations require further investigation. Randomized controlled trials are recommended to verify causal relationships and confirm clinical utility.