Clip closure versus endoscopic suturing versus thoracoscopic repair of an iatrogenic esophageal perforation: a randomized, comparative, long-term survival study in a porcine model (with videos).

BACKGROUND

Esophageal full-thickness wall repair is an important but unsolved issue in endoscopy. It is unknown how well endoscopic clip closure (ECC) and endoscopic closure with suturing (ECS) perform compared with the criterion standard of thoracoscopic closure (TC).

OBJECTIVE

Comparison of technical success, feasibility, long-term patency, complications, and histological quality of the different closure techniques (ECC, ECS, TC) for esophageal perforations.

DESIGN

Comparative animal study.

SETTING

Approved animal facility.

SUBJECTS: Eighteen pigs.

INTERVENTIONS: Eighteen pigs were randomized, 6 each into 3 groups (ECC, ECS, TC). After endoscopic wall incision and mediastinoscopy, closure was performed by using 1 of the 3 techniques. After 8 to 12 weeks, pre-euthanasia endoscopic, necropsy, histological, and morphometric analyses were performed.

MAIN OUTCOME MEASUREMENT: Long-term survival and histological quality of the repair.

RESULTS

The closure of the esophageal incisions was successful in all pigs. On days 2 and 6, 1 animal died of mediastinitis, 1 in the ECS group because of reflux of gastric contents into the mediastinum before the repair and 1 in the TC group because of leakage of the sutured closure (P = 1.0). No strictures were seen on prenecropsy endoscopy. At necropsy, 1 mediastinal abscess was found in an ECS animal (P = 1.0). Minor complications included periesophageal adhesions and reactive lymph nodes in 3 of 6 (ECC group) and 5 of 6 (TC and ECS groups). Histology showed muscle layer defects up to 12 mm in width and 21 mm in length, with a trend toward smaller defect size of width and length in the ECS group of animals.

LIMITATIONS: Animal study of limited size.

CONCLUSIONS

Overall, ECS and ECC performed similarly to TC. ECS showed the smallest histological defects in the long-term repair.