Prevention strategies of esophageal stenosis after endoscopic resection for superficial esophageal cancer: a Bayesian network meta-analysis.

[INTRODUCTION] What interventions effectively prevent postoperative stenosis following endoscopic resection (ER) of superficial esophageal cancer? This study aimed to identify effective interventions or combinations through a systematic review and network meta-analysis.

[METHODS] Six databases were systematically searched for eligible studies up to 30 April 2023, on interventions to prevent esophageal stenosis post-ER. Odds ratios (ORs) evaluated stenosis rate (primary outcome) and complications (secondary outcome), while mean differences (MD) evaluated endoscopic balloon dilatation (EBD) sessions post-stenosis.

[RESULTS] Twenty-three studies involving 1271 patients and 11 different interventions were included. Eight interventions were effective in preventing post-ER stenosis: oral hydrocortisone sodium succinate and aluminum phosphate gel (OHA) (OR: 0.02, 95% credible interval [CrI]: 0.00-0.11), polyglycolic acid (PGA) + ST (OR: 0.02, 95% CrI: 0.00-0.23), oral tranilast (OT) + preemptive endoscopic balloon dilatation (PEBD) (OR: 0.08, 95% CrI: 0.01-0.77), botulinum toxin (BT) (OR: 0.10, 95% CrI: 0.03-0.32), ST (OR: 0.08, 95% CrI: 0.01-0.67), oral steroid (OS) (OR: 0.11, 95% CrI: 0.05-0.28), endoscopic triamcinolone injection (ETI) + OS (OR: 0.17, 95% CrI: 0.07-0.42), and ETI (OR: 0.18, 95% CrI: 0.11-0.30). Five interventions significantly reduced EBD sessions: PGA + ST (MD: -5.78, 95% CrI: -11.04 to -1.21), ETI + OS (MD: -3.27, 95% CrI: -5.37 to -0.72), OS (MD: -6.18, 95% CrI: -9.43 to -3.38), ETI (MD: -3.81, 95% CrI: -5.74 to -1.99), and BT (MD: -2.16, 95% CrI: -4.12 to -0.40). None of the interventions significantly increased complications.

[CONCLUSIONS] This study confirmed the efficacy of OS, ETI, and ETI + OS and verified five other interventions (OHA, PGA + ST, OT + PEBD, BT, and ST) in preventing stenosis. Notably, PGA + ST and BT also reduced the number of EBD sessions.