Leaks and endoscopic assessment of break of integrity after NOTES gastrotomy: the LEAKING study, a prospective, randomized, controlled trial

Abstract

Gastric leak testing after natural orifice transluminal endoscopic surgery (NOTES) gastrotomy closure may help reduce the risk of leaks after transgastric procedures. To develop a novel endoscopy-based system to determine the presence of a leak after NOTES gastrotomy and to compare this system prospectively with radiographic leak testing. Prospective, randomized, controlled trial. Academic Medical Center laboratory. Fifty swine. During the pretrial phase, an endoscopic system for the measurement of intragastric pressure was developed. In the trial phase, swine with a NOTES gastrotomy were randomized to endoscopic versus radiographic leak testing. If a leak was demonstrated, the gastrotomy was reclosed by using a second-generation prototype T-anchor system. The primary outcome was leak detection after gastrotomy closure. The secondary outcome variables included necropsy findings, peritoneal fluid analysis, histologic examination, and clinical outcome. Fourteen swine were included in the pretrial phase and 36 in the randomized trial. Swine were survived for a mean of 9 days postoperatively. Endoscopic pressure monitoring demonstrated a reproducible change in intragastric pressure with insufflation; r = 0.735, P = .001 and r = 0.769, P < or = .000 for the total and maximum pressures, respectively. Post-peritoneoscopy, there was a detectable and significant decrease in the mean total and mean maximum pressures versus baseline (P = .006 and P = .009). There was no significant difference between the radiologic and endoscopic arms in leak detection rate (4/18 vs 3/18, respectively, P = .500). Clinical outcomes and mean weight gain were equivalent. There was 1 operative abdominal wall injury and no deaths. Animal study. Endoscopic pressure monitoring was reproducible, demonstrated the presence of gastric leak, and was as reliable as contrast-based radiographic leak testing.