Combined biliary obstruction and gastric outlet obstruction management: comparison of various strategies of biliary stenting and duodenal derivation in terms of reintervention risk: an expert single-center experience.

Abstract

Background: Double malignant biliary and duodenal obstruction is a frequent and debilitating complication of advanced biliopancreatic cancers. Endoscopic ultrasound (EUS)-guided biliary and enteric drainage techniques have emerged as alternatives to ERCP and duodenal stenting (DS), but optimal therapeutic sequencing remains unclear.

Patients and methods: We conducted a retrospective monocentric study including 34 patients treated for DMBO-DDO between February 2015 and December 2023. Inclusion required histologically confirmed tumors with both distal malignant biliary obstruction (DMBO) and high-grade gastric outlet obstruction (GOO). Treatment sequences were guided by the first symptom and validated by a multidisciplinary team. The primary endpoint was biliary reintervention rate according to different therapeutic combinations. Secondary endpoints included technical and clinical success, adverse events, and global reintervention rates.

Results: Mean patient age was 75 years, with pancreatic adenocarcinoma being the most common (58.8%). ERCP was the first procedure in 67.6% of cases. Technical success was 100% for all biliary interventions. The overall biliary reintervention rate was 52.9%, significantly higher in ERCP-based strategies (up to 77.8%) compared to EUS-guided approaches (18.2%-40%; p = 0.03). Lower initial bilirubin levels (<150 μmol/L) were also associated with increased reintervention rates (77.2% vs. 23.5%; p = 0.019). Complication rates were low (4.4% for grade ≥ III).

Conclusions: ERCP-first strategies for DMBO in the association of GOO showed significantly higher biliary reintervention rates compared to EUS-guided approaches. These findings suggest that EUS-based drainage should be favored in selected patients. Prospective multicenter studies are needed to refine treatment algorithms and confirm predictive factors for reintervention.