It is unclear which is the best approach for the drainage of malignant distal biliary obstruction (MDBO) after failed endoscopic retrograde cholangiopancreatography(ERCP). We compared endoscopic ultrasound (EUS)–guided gallbladder drainage (GBD) and EUS-guided choledocoduodenostomy (CDS) with lumen-apposing metal stents (LAMS) as rescue treatment in the case of ERCP failure.

This was an international multicenter retrospective observational study at 28 tertiary care centers through August 2024. Outcomes were compared using propensity score matching.  The primary outcome was the comparison of the clinical success rates between EUS-GBD vs. EUS-CDS. Clinical success was defined by a >15% decrease in total bilirubin levels within 24 h and a >50% decrease within 14 days after the procedure; secondary outcomes were technical success, defined as adequate LAMS placement with visualization of bile flow, AE and severe AE (SAE), defined as >grade II AE according to the AGREE classification, and overall survival.

Five hundred twenty-nine patients underwent EUS-guided drainage, of which 136 underwent EUS-GBD, and 393 underwent EUS-CDS. After 1-to-1 propensity score matching, 224 patients were selected (112 per group). The mean age was 76±4 years in both groups (p=0.9), and 49 (43.7%) and 48 (42.8%) male patients were treated with EUS-GBD and EUS-CDS, respectively (p=0.9). EUS-GBD and EUS-CDS had similar technical success (97.3% and 91%; p=0.08) and clinical success rates (83% and 85.7%; p=0.17). Mean bilirubin levels decreased from 17.5 mg/dL (SD6) at baseline to 8.1 mg/dL (SD 6.54) at 24 hours to 2.8 mg/dL (SD: 2.15) at 14 days in the EUS-GBD group and from 18.1 mg/dL (SD 8) at baseline to 7.9 mg/dL (SD 8.25) at 24 hours to 3.2 mg/dL (SD 1.35) at 14 days in the EUS-CDS group. AE rate was 19.6% in the EUS-GBD group and 12.5% in the EUS CDS group (p=0.20), of which 10 (8.9%) and 7 (6.2%) severe AEs respectively (p=0.61). Bleeding occurred in seven patients (6.1%) after EUS-GBD and 3 patients (2.5%) after EUS-CDS. Two cases of perforation (1.7%) were observed after EUS-CDS and both were successfully  managed endoscopically.  Stent occlusion occurred in 6 patients (5%, of which 4 cases within 48 hours) after EUS-GBD and  4 patients (3.5%, all of them beyond one week from the procedure) following EUS-CDS (p=0.74);  in all cases these events were treated with double pigtail placement into the LAMS.Stent migration was observed in 2 (1.7%) and 1 (0.8%) patients in the two groups, respectively (p=0.98). Two of these cases were intraprocedural and were treated with removal and replacement  of the LAMS, while one case occurred 90 hours after EUS-GBD and required surgery.  Two cases of mild acute pancreatitis were observed after EUS-GBD (1.7%), probably due to the previous ERCP attempt. Five infectious events were registered after EUS-GBD (4.4%), of which 1 was severe and 3 within 48 hours; on the other hand, 4 cases (3.5%), of which 2 were severe and  3 within 48 hours occurred after EUS-CDS (p=0.29). All these infectious events were treated  successfully with antibiotics. whereas 5 infectious events were registered after EUS-GBD (4.4%) and 4 cases (3.5%) after EUS-CDS (p=0.29). No treatment-related deaths were observed. Median overall survival was 5 months (4-7) in the EUS-GBD group and 5.5 months (5-7) in the EUS-CDS group (p=0.55).

Our study showed that in patients with MDBO after failed ERCP, EUS-GBD or EUSCDS were comparable with similar rates of efficacy and safety. EUS-GBD could  represent an easy and safe option in MDBO patients without previous cholecystectomy  and with a clear patency of the cystic duct.