Efficacy and Safety of Endoscopic Gastroplasty versus Lifestyle Modification for Obesity: A Meta-Analysis of Randomized Controlled Trials with Technique-Specific Subgroup Analysis

Background Endoscopic gastroplasty (EG) offers a minimally invasive alternative to bariatric surgery for obesity management. While multiple randomized controlled trials (RCTs) have compared EG with lifestyle modification (LM), the extent of benefit and influence of procedural technique remain uncertain. Materials and Methods We conducted a meta-analysis restricted to RCTs comparing EG with LM in adults with obesity. The primary outcome was total body weight loss (TBWL) at 12 months. Secondary outcomes included TBWL and excess weight loss (EWL) at 6 months, and EWL at 12 months. Pooled mean differences (MDs) were estimated using a random-effects model. Subgroup analyses were performed based on techniques, such as OverStitch ESG, Endomina, and Primary Obesity Surgery Endoluminal (POSE). Trial sequential analysis (TSA) and assessment of adverse events and publication bias were also conducted. Results Five RCTs involving 696 participants were included. EG resulted in significantly greater TBWL at 12 months compared to LM (MD: 7.67%; 95% confidence interval [CI]: 4.38-10.96; I 2 = 92.5%). TSA confirmed the conclusiveness of this finding. EG also led to superior TBWL at 6 months (MD: 6.98%; 95% CI: 3.76-10.20) and greater EWL at 6 months (MD: 24.85%; 95% CI: 13.44-36.27) and 12 months (MD: 26.24%; 95% CI: 11.19-41.29). Subgroup analysis showed highest efficacy with OverStitch ESG (MD: 10.82%), followed by Endomina (6.50%) and POSE (5.61%). The pooled incidence of any adverse event was 10.8%, while serious adverse events were rare (2.9%). Conclusion EG is more effective than LM for weight loss at 6 and 12 months, with an acceptable safety profile. Among techniques, OverStitch ESG demonstrated the highest efficacy. These findings support the incorporation of EG into the obesity treatment algorithm, especially for patients unsuitable for surgery.