Feasibility and preliminary results of modeling a clinically meaningful anastomotic leak in pigs

Animal models are essential for investigating the pathobiology of anastomotic leakage (AL). To be clinically meaningful, AL must occur despite a technically adequate anastomosis and reflect a phenotype relevant to surgical practice. This feasibility pilot study aimed to develop a clinically relevant porcine AL model by combining a Western diet (WD), segmental ischemia, perioperative antibiotic prophylaxis (PAP), and luminal exposure to pathogenic bacteria under optimal surgical conditions.

Twenty-one female domestic pigs were randomized into four groups. Group 1 received standard chow, PAP, and underwent a stapled colorectal anastomosis. Group 2 received WD, PAP, and perioperative luminal exposure to viable human and murine pathogens via colonoscopy. Group 3 received the same treatment with added ischemia-reperfusion (I/R) injury. Group 4 served as a negative control, receiving all exposures with autoclaved pathogens. Animals underwent a technically optimal stapled colorectal anastomosis. Healing was assessed clinically, endoscopically, and macroscopically on postoperative days 3 and 7.

Gross healing (P = 0.0027) and adhesion scores (P = 0.0067) differed significantly between groups, with the highest scores in pigs exposed to WD, I/R, and viable pathogens. These changes did not reach the threshold of clinically overt anastomotic failure (Clavien-Dindo grade III), and endoscopic scores showed no significant intergroup differences. A strong association was observed between gross healing and adhesion scores (r = 0.909).

While the model did not progress to clinically overt AL (Clavien-Dindo grade III), it reproducibly induced subclinical impairment of anastomotic healing under compounded perioperative stress. These findings support its feasibility and biological relevance for studying early determinants of anastomotic integrity prior to clinical failure.