Prevention of Variceal Rebleeding by EUS-guided vs Conventional Endoscopic Therapy in Hepatocellular Carcinoma Patients (EUS-SPV)

Prevention of Variceal Rebleeding by EUS-guided vs Conventional Endoscopic Therapy in Hepatocellular Carcinoma Patients

Secondary Prevention of Variceal Rebleeding by Endoscopic Ultrasound-guided Therapy Versus Conventional Endoscopic Therapy in Hepatocellular Carcinoma Patients: a Randomized Controlled Trial

Rebleeding rate is high in hepatocellular carcinoma (HCC) patients with variceal bleeding despite conventional endoscopic therapies for esophageal and gastric varices (EV, GV). Secondary prevention of variceal rebleeding was reported to improve outcomes of HCC patients, but the optimal endoscopic approach is not well defined. In this difficult-to-manage population, variceal rebleeding rates remain substantial after conventional endoscopic therapies. n recent studies by others and our group on direct EUS-guided therapy for varices in cirrhotic patients, high technical success (90 - 100%), low post-treatment rebleeding rate (3 - 11%) and low adverse event rate (~3%) have been reported for GV treatment by cyanoacrylate glue injection, coiling or a combination of both, and for cyanoacrylate glue injection or coiling of EV refractory to variceal band ligation (VBL). This study aims to compare rebleeding rates after secondary prevention by EUS-guided therapy or conventional endoscopic therapy in HCC patients with recent variceal bleeding.

Chronic liver diseases and cirrhosis are common health problems in the Asia-Pacific region. In 2015, 54.3% of global deaths due to cirrhosis occurred in the Asia-Pacific region. Acute variceal bleeding is a life-threatening complication of cirrhosis that occurs at a rate of 10 - 15% per year, with a 6-week mortality rate up to 25%. In patients with successful acute bleeding control by endoscopic therapy and vasoactive agents, rebleeding is common in those without subsequent secondary prevention by non-selective beta blocker and/or endoscopic therapy (e.g. variceal band ligation (VBL) for esophageal varices (EV) and glue injection for gastric varices (GV)). Apart from variceal bleeding, hepatocellular carcinoma (HCC) (liver cancer) is another important complication of cirrhosis. HCC patients with prior variceal bleeding are at high risk of rebleeding due to significant portal hypertension and frequent presence of portal vein thrombosis (PVT). In this difficult-to-manage population, variceal rebleeding rates remain substantial after conventional endoscopic therapies. In a large multicenter study comparing clinical outcomes after EV bleeding in patients with or without HCC, lack of secondary prevention for rebleeding was found to be frequent in HCC patients and was associated with a higher rate of rebleeding and mortality. In HCC patients with EV bleeding, secondary prevention failure by conventional endoscopic therapy was significantly higher (50% vs 31%, P = 0.001) when compared with patients without HCC. However, the optimal endoscopic approach for secondary prevention in HCC patients has not been well defined. Whether commonly used endoscopic techniques such as VBL for EV and cyanoacrylate glue injection for GV being performed at intervals of 3 - 4 weeks for secondary prevention can achieve durable variceal control in HCC patients remains unclear. Direct endoscopic ultrasound (EUS)-guided variceal interventions by cyanoacrylate glue injection, coiling, or a combination of both using a therapeutic curvilinear echoendoscope is a novel endoscopic technique that have attracted clinical attention due to its high efficacy in variceal control. In recent studies by others and our group on direct EUS-guided therapy for varices in cirrhotic patients, high technical success (90 - 100%), low post-treatment rebleeding rate (3 - 11%) and low adverse event rate (~3%) have been reported for GV treatment by cyanoacrylate glue injection, coiling or a combination of both, and for cyanoacrylate glue injection or coiling of EV refractory to VBL. In a retrospective study published by our group in 2020, we compared outcomes in 27 HCC patients with variceal bleeding who underwent secondary prevention by EUS-guided glue injection every 12 weeks and 33 HCC patients without secondary prevention after control of acute variceal bleeding. The technical success of EUS-guided therapy was 100%. The overall procedure-related adverse event rate was low (3.7%) and no radiographic evidence of glue-lipiodol embolization was observed. The EUS-guided therapy group was found to have a significantly lower 90-day death-adjusted cumulative incidence of rebleeding and a significantly higher variceal bleeding-free survival at 3 and 6 months. As such, it would be clinically important to conduct a prospective randomized controlled study to confirm the benefits of EUS-guided therapy for secondary prevention in HCC patients.

EUS would be performed with a curvilinear echoendoscope based on protocol described in our prior study. Because EUS-guided glue injection can be applied to both EV and GV and is less technically demanding than coiling or the combination approach, it is chosen as the EUS guided intervention in our study. The target EV or GV size will be measured by the caliper function on the EUS machine. After confirmation of blood flow in the target varix by Doppler, EUS-guided glue injection would be performed for EV or GV ≥ 3mm in diameter using a standard 19G needle. Each injection will contain a 1.2ml mixture of 0.5ml glue (Histoacryl, n-butyl-2-cyanoacrylate, B. Braun Surgical, Germany) + 0.7ml lipiodol. Flow obliteration in the treated varix will be confirmed on EUS Doppler. If blood flow is still observed on Doppler after the first injection, additional injection of glue-lipiodol mixture would be repeated (up to 4 doses) until flow obliteration is achieved.

In patients with prior EV bleeding, EV with high-risk stigmata (regardless of size) or EV of medium or large size detected on study EGD will be treated with VBL using a multi-band ligator fitted on the gastroscope for secondary prevention. In patients with prior GV bleeding, if compressible GV suggestive of incomplete obliteration from prior glue treatment is noted on study EGD, cyanoacrylate glue injection using a 1.2ml mixture of 0.5ml glue (Histoacryl, n-butyl-2-cyanoacrylate, B. Braun Surgical, Germany) + 0.7ml lipiodol will be performed for secondary prevention.

conventional endoscopic therapy (VBL for EV or glue injection for GV) for secondary prevention of EV or GV rebleeding.

90-day death adjusted cumulative incidence of variceal rebleeding after secondary prevention by EUS-guided therapy or conventional endoscopic therapy

Cumulative incidence of variceal rebleeding at 90-day after study procedures will be analyzed after adjustment of incidence of death from HCC or cirrhosis

30-day death adjusted cumulative incidence of variceal rebleeding after secondary prevention by EUS-guided therapy or conventional endoscopic therapy

Cumulative incidence of variceal rebleeding at 30-day after study procedures will be analyzed after adjustment of incidence of death from HCC or cirrhosis

bleeding-free survival at 3 months after secondary prevention by EUS-guided therapy or conventional endoscopic therapy

To determine the proportion of patients who do not have rebleeding and are alive at 3 months after study procedures

procedure related adverse events are defined as perforation, glue-lipiodol embolization, post-banding ulcer bleeding, infection, death related to procedure

Inclusion Criteria: Consecutive HCC patients age 18 or older with recent (within 4 weeks of the episode) EV or GV bleeding successfully controlled by conventional endoscopic therapies (VBL for EV or glue injection for GV) Able to provide written informed consent to participate in the study and comply with the study procedures Exclusion Criteria: Unable to provide written informed consent Contraindications for endoscopy due to underlying comorbidities HCC patients with non-variceal source of gastrointestinal bleeding Refractory coagulopathy (INR>1.5) or refractory thrombocytopenia (platelets <50,000) despite blood product transfusion Moribund patients from terminal illness