Capnoperitoneum may occur during gastric endoscopic submucosal dissection (ESD) due to type IV deep mural injury when resecting mucosal lesions, or during exposing endoscopic full-thickness resection (e-EFTR) for subepithelial lesions (SELs). In some cases, it may progress to tension capnoperitoneum, resulting in gastric lumen collapse, respiratory compromise, or haemodynamic instability. Prompt recognition and effective decompression are essential for safe endoscopic practice. We report outcomes of a standardised protocol for the management of significant capnoperitoneum.

We performed a retrospective analysis of a prospectively maintained database of consecutive gastric endoscopic resections undertaken at a tertiary referral centre between November 2016 and November 2025. All procedures were performed under general anaesthesia with endotracheal intubation, and with continuous monitoring of end-tidal CO₂, ventilation pressures, and abdominal distension. Significant capnoperitoneum was managed using a standardised protocol involving left upper quadrant percutaneous needle decompression, endoscopic clip closure of the muscle defect, and reinflation of the stomach to perform a leak test. Bubbling through an attached water-filled syringe catheter indicated residual leakage, and prompted careful re-assessment to identify and close full thickness muscular injury. After a negative leak test, a 14-French nasogastric tube was placed on free drainage.

Among 354 gastric endoscopic resections, 14 patients (4.0%) required needle decompression for significant capnoperitoneum (9/252 mucosal lesions, 5/102 SELs). Median age was 72 years (IQR 63–77), 57% were female, and median lesion size was 32.5 mm (IQR 20–50). R0 resection was achieved in 12/14 (86%). Median full-thickness defect size was 10 mm (IQR 8–12.5) with e-EFTR and 3 mm (IQR 2–7) with ESD. Complete endoscopic closure was achieved in all cases using a median of 5 clips (IQR: 4-6) in the ESD group and 10 clips (IQR: 8-11.5) in the e-EFTR group, with negative leak testing confirmed in every case. Post-procedure CT imaging was performed in 3/14 patients (21.4%) for abdominal pain, demonstrating expected capnoperitoneum without contrast leak. One patient (7%) developed transient fevers managed with antibiotics alone, and one (7%) required patient-controlled analgesia.

Clear fluids and solid diet were resumed at a median of 1 days (IQR 1–2) and 3 days (IQR 2.5–3.5), respectively. The NG tube was removed after a median of 2 days (IQR: 1-3). Median length of stay was 3 days (IQR 2.5–3.5). No surgical intervention was required, and there were no long-term sequelae or mortality. Readmission rate was zero. 

Full-thickness defects with tension capnoperitoneum during gastric ESD or e-EFTR can be safely managed using a standardised protocol incorporating needle decompression, endoscopic closure and confirmatory leak testing. This approach achieved 100% technical success with no surgical rescue, rapid diet advancement, and short inpatient stay. Our findings support routine integration of this protocol into advanced gastric endoscopic resection practice.