Endoscopic ultrasound-guided choledochoduodenostomy versus hepaticogastrostomy combined with gastroenterostomy in malignant double obstruction (CABRIOLET_Pro): A prospective comparative study

Abstract

Objectives: Malignant double obstruction, defined as the simultaneous presence of biliary and gastric outlet obstruction, represents a challenging clinical scenario. Previous retrospective experiences have demonstrated shorter dysfunction-free survival (DyFS) of endoscopic ultrasound-guided choledochoduodenostomy (EUS-CDS) versus EUS-hepaticogastrostomy (EUS-HGS) in this setting, but no prospective evidence is available.

Methods: Twenty consecutive patients with malignant double obstruction, treated with EUS-gastroenterostomy (and EUS-guided biliary drainage, following a previously failed ERCP, were enrolled in a prospective observational study (ClinicalTrials.gov NCT04813055) comparing EUS-CDS versus EUS-HGS. Efficacy and safety were evaluated, with Biliary Dysfunctions as the primary outcome and DyFS using Kaplan-Meier estimates as a primary measure.

Results: Twenty patients (75% with pancreatic cancer, 50% with metastatic disease) with EUS-gastroenterostomy were included (seven EUS-CDS and 13 EUS-HGS). No significant difference was detected at baseline. Technical success was 100% in both groups. EUS-CDS compared to EUS-HGS showed similar clinical success (100% vs. 92.3%, p = 0.5), a higher rate of post-procedural adverse events (42.9% vs. 7.7%, p = 0.067, mostly related to severe/fatal cholangitis in the EUS-CDS group) and a higher rate of biliary dysfunctions during follow-up (71.4% vs. 16.7%, p = 0.002).DyFS was significantly shorter in the EUS-CDS group (39 [15-62] vs. 268 [192-344] days, p = 0.0023), with a 30-days DyFS probability of 57.1% vs. 100% (hazard ratio = 7.8 [1.4-44.2]).

Conclusions: In this prospective comparison of patients with malignant double obstruction undergoing EUS-gastroenterostomy, treating jaundice with EUS-CDS versus EUS-HGS resulted in a reduced probability of survival without biliary events and an increased risk of biliary dysfunctions (number needed to harm = 1.8), with detection of severe/fatal cholangitis.

Keywords: cholangiopancreatography; cholangitis; endoscopic retrograde; gastric outlet obstruction; jaundice; stent.

FIGURE 1

(a–c) Endoscopic ultrasound (EUS)‐guided gastroenterostomy (EUS‐GE). (a) Endosonographic appearance of the operative window for EUS‐GE before lumen‐apposing metal stents (LAMS) deployment (see the tip of the catheter in the left upper corner). The operative space includes an adequately distended jejunal loop, orientated in the direction of the operative channel of the scope, with the shortest distance possible between the gastric and jejunal lumen; (b) fluoroscopic image showing an oro‐jejunal tube looped in the first jejunal loop, through which the jejunum is distended (in this case with saline mixed with contrast) and a LAMS released between the stomach and the jejunal loop. (c) Visualization of the jejunal folds through the LAMS after stent release and dilation. (d–f) EUS‐guided choledochoduodenostomy (EUS‐CDS). (d) Endosonographic appearance of a LAMS placed between the choledochus and the duodenal bulb; (e) endoscopic image of a released proximal flange with bile flowing in the duodenum; (f) fluoroscopic image of a double obstruction managed by EUS‐GE + EUS‐CDS (with aerobilia). (g–i) EUS‐guided hepaticogastrostomy. (g) Endosonographic appearance of a 19G needle accessing a branch of a mildly dilated left biliary three through a cushion of hepatic parenchyma; (h) contrast injection confirming the puncture of a branch of the left biliary tree; (i) endoscopic appearance of a gastroenteric LAMS and of a hepatogastric SEMS.

FIGURE 2

(a–c) Endoscopic ultrasound‐choledochoduodenostomy (EUS‐CDS) dysfunction. (a) During the follow‐up of a combination of EUS‐gastroenterostomy (EUS‐GE) + EUS‐CDS, the patient experienced cholangitis, and a gastrointestinal contrast series showed reflux of contrast medium inside the biliary tree through the lumen‐apposing metal stents (LAMS). (b) At endoscopic reintervention, food impaction was visible inside the choledocho‐bulbar LAMS; (c) balloon swipes were performed to extract food debris and sludge from the biliary duct. (d–f) EUS‐hepaticogastrostomy (EUS‐HGS) dysfunction. (d) During follow‐up of a combination of EUS‐GE + EUS‐HGS, the patient experienced recurrent jaundice and cholangitis. At endoscopic revision, the hepatogastric SEMS was visibly obstructed and showed an outflow of pus. (e) The SEMS was cannulated with a Fogarty balloon over the wire and balloon swipes were performed to clear accumulated debris; (f) at subsequent cholangiography, the end of the hepatogastric SEMS was angulated with respect to the direction of the left hepatic duct, and therefore biliary drainage was facilitated by anterograde placement of a transpapillary SEMS.

FIGURE 3

Kaplan‐Meier analysis of overall (a) and dysfunction‐free survival (b). (a) No difference in overall survival was noticed between the EUS‐choledochoduodenostomy (EUS‐CDS) and the EUS‐hepaticogastrostomy (EUS‐HGS) groups. (b) In the setting of combined biliary and alimentary obstruction, the latter treated by EUS‐gastroenterostomy, the EUS‐CDS group showed a significantly shorter survival without biliary events compared to the EUS‐HGS group (log‐rank test p = 0.0023).