The Leaks & Endoscopic Assessment of Break of Integrity After a NOTES® Gastrotomy (LEAKING) Study, a Prospective Randomized Controlled Trial: Initial Results

Background: Intragastric pressure monitoring may provide a sensitive, real time, endoscopic test for the exclusion of a leak following a NOTES gastrotomy. A highly reliable endoscopic leak test will be important for clinical NOTES procedures. Design: A prospective randomized controlled trial of contrast-based radiologic leak testing versus pressure-monitoring-based endoscopic leak testing. Methods: During the pre-trial phase, a system for the measurement of intragastric pressure was developed using a clinical hemodynamic blood pressure monitoring device and pressure transducer. In the trial phase, 40kg swine were block randomized to radiologic or endoscopic arms. All gastrotomies were created using a needle knife and an 18mm dilating balloon. In the radiologic arm, leak testing was assessed by pooling 120cc of dilute gastrograffin over the gastrotomy and obtaining abdominal radiographs in AP, lateral, and oblique views. Radiographic images were interpreted by a GI radiologist. After delineation in the pre-trial phase, no-leak in the endoscopic group was defined as a pressure ≥ 15 mmHg during 120 sec of insufflation. If a leak was demonstrated, the gastrotomy was reclosed using a second generation prototype T-anchor system. At necropsy, peritoneal fluid analysis, histologic examination, and a water-fill test were also performed. Results: There were 59 total swine, 23 in the pre-trial phase and 36 in the randomized trial. Swine were survived for a mean of 9 days post-operatively. By total and max gastric pressure over 120 sec, endoscopic pressure monitoring demonstrated a reproducible change with insufflation (r=0.735, p<0.001 & r=0.769, p<0.000). In the post-peritoneoscopy state, there was a detectable and significant decrease in the mean total and mean max pressure versus baseline (p=0.006 & p=0.009). After endoscopic closure, there was no significant difference in mean total and mean max gastric pressures compared to baseline (p=0.480 & p=0.380). There was no significant difference between the radiologic and endoscopic arms in the detection of leaks (4/18 vs 3/18 p=0.500). Mean weight gain was equivalent (p=0.998). There was one operative abdominal wall injury and no deaths. There were no leaks by the water fill test at necropsy. Conclusion: In the largest randomized trial of natural orifice surgery to date, endoscopic gastric pressure monitoring demonstrated the presence of a gastric leak following NOTES procedures and was as reliable as contrast based radiology. This is the first randomized trial demonstrating the utility of an endoscopic pressure-based test for the detection of a gastric leak during natural orifice surgery.