A network meta-analysis on the optimal treatment of malignant gastric outlet obstruction.

Background: Gastric outlet obstruction (GOO) caused by malignancy significantly impairs patient quality of life. Surgical gastrojejunostomy (SGJ), endoscopic stenting (ES), and endoscopic ultrasound-guided gastroenterostomy (EUS-GE) are the three main palliative treatments. However, the optimal approach remains unclear because of variations in study results and limited studies comparing treatment types. Objectives: To evaluate and compare the effectiveness and safety of EUS-GE, ES, and SGJ in the management of GOO. Design: Network meta-analysis. Data sources and methods: The PubMed, Cochrane Library, and Web of Science databases were systematically searched for full-length articles of randomized controlled trials and cohort studies in English comparing SGJ, ES, and EUS-GE in adult patients with malignant GOO. Studies with unbalanced baseline characteristics or technical variations within a single modality were excluded. The primary outcome was the reintervention rate for recurrent GOO. Secondary outcomes included clinical success and adverse event rates. Results: Fifty-four studies involving 6110 patients were analyzed (SGJ, n = 1974, ES, n = 3226, EUS-GE, n = 910). Compared with ES, both SGJ (odds ratio (OR): 0.32, 95% confidence interval (CI): 0.22-0.46) and EUS-GE (OR: 0.29, 95% CI: 0.17-0.51) significantly reduced the risk of reintervention. EUS-GE achieved a higher clinical success rate than ES (OR: 2.46, 95% CI: 1.50-4.04) and also lower adverse event rate than both SGJ (OR: 0.33, 95% CI: 0.19-0.56) and ES (OR: 0.61, 95% CI: 0.37-0.99). Heterogeneity was moderate across outcomes, with no major inconsistency or publication biases detected. Conclusion: Of the three treatments for GOO, EUS-GE demonstrated the most favorable profile in terms of efficacy, safety, and durability. SGJ remains a viable alternative, particularly in centers lacking expertise in advanced endoscopy. These findings may inform future clinical guidelines and support the broader adoption of EUS-GE in appropriate settings.