Work

Novel algorithm for the diagnosis and treatment of different types of bile leakage following liver surgery

Year: 2026

Type: article

Abstract: BACKGROUND: Posthepatectomy bile leakage (PHBL) after liver resection with bilioenteric anastomosis may originate from the anastomosis (anastomotic leak (AL)) or the transection surface (surface leak), including leakage from isolated bile ducts (IBDLs). Differentiating these sources remains challenging, and standardized management guidelines are lacking. METHODS: All patients undergoing liver resection with bilioenteric anastomosis from 2008-2023 were analysed. PHBL was classified by origin (AL, surface leak, IBDL) and timing (early versus late onset). Drain fluid bilirubin concentration, drain-to-serum bilirubin ratio, and daily bilirubin loss were evaluated by leak type. Outcomes of conservative, interventional, and surgical management were compared. RESULTS: Among 2936 liver resections, 229 involved bilioenteric anastomosis. PHBL occurred in 96 patients, including 44 ALs (46%), 47 surface leaks (49%), and 5 IBDLs (5%). AL showed higher peak drain bilirubin than surface leak/IBDL (38.3 versus 18.6 mg/dl; P < 0.001). Receiver operating characteristic analysis identified cut-offs for AL diagnosis: drain bilirubin > 20.85 mg/dl, drain-to-serum ratio > 10.57, and bilirubin loss > 160.93 mg/day. Management success varied by type and timing: AL and IBDL responded best to reoperation (success rates 78% and 100%), whereas surface leak was more effectively managed without surgery (65%). In AL specifically, reoperation was more successful in early-onset leaks (65% versus 38% in late-onset; P = 0.069), whereas interventional therapy was more successful in late-onset leaks (58% versus 15% in early-onset; P = 0.003). CONCLUSIONS: Distinct diagnostic thresholds help differentiate AL from other PHBL types. AL and IBDL benefit from reoperation, whereas surface leak is better managed without surgery. Timing influences therapeutic success, supporting a novel type- and timing-based management algorithm for PHBL.

Source: BJS Open

DOI: 10.1093/bjsopen/zrag110