Weight regain following roux-en-y gastric bypass affects up to 30% of individuals. Regain is often due to dilation of the gastrojejunostomy over time, leading to decreased restriction and postprandial satiety. Endoscopic gastrojejunostomy revision (EGJR) is an incisionless procedure that narrows the anastomosis and restores restriction. Several suture patterns have been described for EGJR, but their efficacy has not been compared experimentally. This study utilized an ex vivo porcine model to compare failure pressure and procedure time for different EGJR suture patterns.


Gastrojejunostomies were created between an ex vivo porcine stomach and a small intestine with a 33mm circular stapler. The ostomy was revised with one of five suture patterns: simple-interrupted, vest-over-pants, figure-of-eight, purse-string, or ‘hairpin.’ The ‘hairpin’ pattern is a two-layer simple-running stitch that was designed to keep tension away from the anchors that hold a completed suture pattern in place. After revision, the stomachs were filled with water under continuous manometric monitoring. Pressure was steadily increased until failure, which was defined as a disruption of the suture, a severed connection between the suture and anchor, or a tear in the tissue due to tension from the revision.


Procedure time, failure cause and failure pressure were recorded for eight trials per pattern. Average failure pressure from lowest to highest was (in mmHg): simple interrupted (34.9), vest-over-pants (46.8), figure-of-eight (48.5), purse-string (53), and hairpin (62.8). By ANOVA, the pattern series were statistically different (p<0.01), but by Tukey’s-HSD, only hairpin failure pressure was higher than the simple-interrupted (p<0.01). There were no failures of the suture itself, the suture-anchor connection failed 16 times, and tissue tore 24 times. By t-test, failure pressure of the suture-anchor connection was lower than that of tissue failure (p<0.01). Time to execute the different patterns averaged about 16 minutes, except for the purse-string, which averaged 8 minutes and was significantly faster than the others (p<0.05).


This study successfully utilized an ex vivo porcine model to compare performance of suture patterns used for EGJR. The results seem to indicate that the more durable patterns are those that spread tension along the length of the suture. The four, short sutures used for an interrupted revision are inferior to one-suture patterns such as purse-string or hairpin. Technically purse-string was simpler and twice as fast to perform than hairpin, and therefore, may be a superior technique for EGJR.