Clinical features of postoperative anastomotic bleeding after gastrectomy and efficacy of endoscopic hemostasis

Postoperative anastomotic bleeding(PAB) is rare, but can be lethal. Between January 2004 and May 2013, 16,591 patients underwent gastrectomy for gastric cancer at Our Center. As a case group, 36 patients with PAB were enrolled. Each subject was matched by ratio of 1:5 with randomly selected individuals without bleeding. The clinical outcomes and risk factors of PAB and results of endoscopic hemostasis(EH) were analyzed. The incidence of PAB was 0.22%. Median bleeding time was 34.5 hours. EH was attempted to 25 patients, surgical approach to 6 patients, and conservative management to 5 patients. PAB-related death occurred in 3 patients. In multivariate analysis, the type of gastrectomy showed significant odds ratio for PAB (OR 3.803 (95% CI, 1.285-11.260), p=.016). Primary EH was achieved in 64%. Endoscopic tools used were hemoclips(52%), epinephrine injection(32%) and glue injection(24%). In the EH failure group, secondary EH and reoperation were performed in 2 and 4 patients, respectively. Regarding bleeding site at the anastomotic ring, the success rate of EH at anterior wall side(100%) was higher than that at posterior wall side(50%). Even though PAB is an infrequent and potentially life-threatening complication, endoscopic procedure can be considered as a useful method for both confirmation of bleeding and therapeutic intervention.