Endoscopic ultrasound-guided vascular intervention for portal hypertension

Abstract

Introduction

Portal hypertension (pHTN) can result from cirrhosis or thrombosis of major venous system (e.g., portal vein thrombosis, splenic vein thrombosis, Budd-Chiari syndrome, etc.).1,2 Gastroesophageal variceal bleeding is a life-threatening complication in patients with pHTN and cirrhosis. While advances in the management of acute variceal bleeding (e.g., early use of vasoactive pharmacological therapy such as somatostatin analogue and terlipressin, antibiotic prophylaxis for spontaneous bacterial peritonitis, timely endoscopy, etc.) in the recent decades have led to improved clinical outcomes, the mortality rate remains as high as 20%.1,2 Conventional vascular diagnostic and therapeutic interventions in patients with pHTN are generally performed with standard endoscopy (e.g., endoscopic band ligation [EBL] for esophageal varices [EV], glue injection for gastric varices [GV]), interventional radiology (IR) based procedures (e.g., assessment of portal hemodynamics, creation of transjugular intrahepatic portosystemic shunt [TIPS], balloon-occluded retrograde transvenous obliteration [BRTO]), or surgery (e.g., surgical shunting).1,2 As endoscopic ultrasound (EUS) technology improves, the diagnostic and therapeutic applications of EUS in patients with pHTN have been increasingly explored. This article aims to review the current status of EUS-guided vascular interventions in pHTN.

Detection of Gastroesophageal Varices and Collateral Veins

Although direct endoluminal examination by esophagogastroduodenoscopy (EGD) is well established for detection of endoscopically visible varices, EUS can offer additional information by detection of “deep varices”, collateral veins/perforating feeding veins, and differentiating between thickened gastric folds and genuine GV.3 Gastroesophageal varices, collateral veins/perforating feeding veins, and paraesophageal/paragastric varices can be observed endosonographically with catheter based miniprobe, mechanical echoendoscope or electronic echoendoscope with Doppler.3–6Fig. 1 illustrate GV observed by radial echoendoscope and linear echoendoscope. Fig. 2 illustrate EV and perforating vein observed by linear echodendoscope. In a study including 52 cirrhotic patients and 166 dyspeptic patients, EUS was found to be more sensitive than EGD in diagnosing GV (16/52 vs 9/52), and could detect extraluminal venous abnormalities (e.g., paraesophageal/paragastric collateral veins) in 48/52 cirrhotic patients.6 In another study with 66 cirrhotic patients and 32 healthy control patients, paraesophageal varices detected by EUS were a more sensitive predictor of cirrhosis than EV detected by EGD (97% vs 74%, P < 0.0001).5

Figure 1

Gastric varices (arrows) detected by radial echoendoscope and linear echoendoscope. (A) B-mode imaging with radial echoendoscope. (B) B-mode imaging with linear echoendoscope. (C) Color Doppler imaging with linear echoendoscope.

Figure 2

(A) Esophageal varices detected by linear echoendoscope with Doppler. (B) Perforating vein (arrow) of esophageal varices detected by linear echoendoscope with Doppler.

Evaluation of Variceal Recurrence, Rebleeding Risk, and Efficacy of Pharmacotherapy

EUS can provide important information on the risk of variceal recurrence and rebleeding after endoscopic therapy and pharmacotherapy. In a study with 40 patients treated with EBL for EV bleeding, finding of large (≥ 5 mm) paraesophageal varices on EUS 4 weeks after EBL was shown to predict EV recurrence and rebleeding.7 Fourteen patients had large paraesophageal varices, while 26 patients had none/small paraesophageal varices on EUS 4 weeks after EBL.7 During the 12-month follow-up period, the risk of EV recurrence (93% vs 46%, P = 0.0019) and rebleeding (43% vs 12%, P = 0.044) were significantly higher in patients with large paraesophageal varices on EUS at 4 weeks.7 In another study evaluating the utility of EUS monitoring in cyanoacrylate (CYA) injection by EGD to obliterate GV, clinical outcomes were compared between 54 patients who underwent biweekly EUS monitoring and repeated CYA injection by EGD until complete GV obliteration and 47 patients who underwent “on-demand” CYA injection in response to new episode of GV bleeding.8 While there was no difference in the rate of early (< 48 hours) rebleeding, the rebleeding rate ≥ 48 hours was significantly reduced in the group with EUS monitoring and repeated CYA injection (44.7% vs 18.5%, P = 0.0053).8 Volumetric change in paraesophageal varices on EUS has also been correlated with response to non-selective beta-blocker and risk of EV recurrence. In a randomized study of 66 patients who achieved EV eradication by EBL, the relation between volumetric change of paraesophageal varices detected by EUS and EV recurrence in patients with or without propranolol as secondary prophylaxis was evaluated.9 During the 2-year follow-up period, the volumetric change of paraesophageal varices detected by EUS between the 2 groups was found to become significant at the third month.8 Of 33 patients in the propranolol group, 20 patients showed regression of paraesophageal varices on EUS, with no EV recurrence during the 2-year follow-up, indicating response to propranolol.8

EUS-guided Therapy of Gastroesophageal and Ectopic Varices

Conventional endoscopic therapies for patients with bleeding EV include EBL, or injection sclerotherapy with a sclerosing agent or tissue adhesive such as CYA.1,2 When compared to conventional endoscopic therapies for varices, EUS-guided therapy can offer the following advantages: (1) to ensure intra-variceal delivery of injection therapy, (2) to allow real-time monitoring of variceal obliteration, and (3) to provide injection therapy under pure EUS guidance when the target varix is endoscopically obscured due to torrential bleeding.10–16Fig. 3 illustrate GV obliteration by EUS-guided CYA injection with Doppler monitoring.

Figure 3

Endoscopic ultrasound (EUS)-guided cyanoacrylate injection of gastric varices. (A) B-mode imaging. (B) Color Doppler imaging. (C) Real-time EUS Doppler confirmation of flow obliteration in treated gastric varices.

EUS-guided injection therapy for gastroesophageal and ectopic varices using sclerosing agent, tissue adhesive (e.g., CYA), coil, or a combination of coil and CYA have been studied.10–16 In general, a 19 G or 22 G standard EUS needle can be used for EUS-guided intravascular puncture of the target varices.10–16 Tissue adhesives in the form of CYA (e.g., N-butyl-2-CYA, 2-Octyl-CYA) with or without lipiodol, or sclerosing agents (e.g., ethanolamine) have been used for EUS-guided injection therapy of gastroesophageal varices.10–16 While fluoroscopy is not required during injection therapy of varices with CYA under EUS guidance, fluoroscopy can be beneficial during EUS-guided coiling procedures with or without CYA injection. After injection therapy, EUS Doppler is used to provide real-time monitoring of variceal flow obliteration. Table 1 summarizes the major studies on EUS-guided injection therapies.

In an early study comparing conventional endoscopic sclerotherapy with ethanolamine or EUS-guided ethanolamine injection to esophageal collateral veins in 50 patients with EV, there was no difference in the number of sessions until EV eradication and EV recurrence during the follow-up period.11 Persistence of esophageal collateral vessels after sclerotherapy was found to be a risk factor for EV recurrence.11 Subsequent studies on EUS-guided injection therapies focus more on the treatment of GV bleeding. When compared to EV, GV bleeding tends to be more severe and carries a higher mortality.1,2 EUS-guided injection of CYA in perforating feeding veins of GV was reported in 2007 in 5 patients with GV bleeding.12 The mean number of sessions required for GV obliteration was 1.6 session, with no rebleeding during a mean follow-up period of 10 months.12 One of the feared serious complications of CYA injection is distant CYA embolization. Combining coil with CYA injection may theoretically reduce the risk CYA embolization and the amount of CYA needed during treatment. In a study of 30 patients with fundal GV (2 with active bleeding and 14 with stigmata of recent bleeding) published in 2011, EUS-guided transesophageal treatment of fundal GV with combined coil and CYA injection was evaluated.13 The mean volume of CYA injected was 1.4 mL per varix.13 During the mean follow-up of 193 days of 24 patients, 96% of the treated GV was obliterated after a single treatment session based on surveillance EUS with Doppler.13 Of the patients, 16.6% had rebleeding from a nonvariceal source while none had procedure-related complications or CYA embolization.13 In a recently published retrospective review of 152 patients with fundal GV treated with EUS-guided combined coil and CYA injection (5% with active bleeding, 69% with recent bleeding, 26% with treatment for primary prophylaxis) over 6 years, the combined coil and CYA injection technique was found to be effective in treatment of acute bleeding, secondary and primary prophylaxis.14 In 100 patients with follow-up EUS, 93% had complete fundal GV obliteration based on EUS Doppler and 3% had post-treatment bleeding from obliterated fundal GV during follow-up.14 Among 125 patients with clinical and/or endoscopy follow-up, 4 patients experienced post-procedure abdominal pain, and 1 had pulmonary embolism.14 When compared with CYA injection, EUS-guided coil for GV appears to carry a smaller risk for adverse event, mainly in the reduction of pulmonary embolism. In a retrospective study comparing the clinical outcomes of 11 patients treated with EUS-guided coil with those of 19 patients treated with EUS-guided CYA injection for GV, while the overall GV obliteration rate was similar between 2 groups, the overall adverse event rate (symptomatic and asymptomatic) is higher in the CYA group when compared to the coil group.15 Eleven patients in the CYA group had adverse events (9 with asymptomatic pulmonary CYA embolisms detected on computed tomography, 1 with chest pain and 1 with fever) after the EUS procedure, while 1 patient in the coil group had adverse event from EV bleeding after the EUS procedure.15 EUS-guided therapies for ectopic varices have also been described. In a case series of 14 patients, 3 patients with duodenal varices and 5 patients with choledochal varices who failed conventional therapies were treated with EUS-guided coil with or without CYA injection.16 While there was no rebleeding from the treated duodenal varices, 3 out of the 5 treated patients with choledochal varices developed rebleeding, requiring placement of biliary covered metal stents by ERCP +/− repeated EUS-guided coil for hemostasis.16

While EUS-guided injection therapies provide several advantages when compared to conventional injection therapies by EGD, complications such as pulmonary CYA embolism, post-CYA injection ulcer, development/enlargement of varices at another location (e.g., enlargement of EV after treatment of GV), rebleeding from the treated varices may still occur. In patients who undergo EUS-guided injection with CYA and lipiodol mixture, post-procedure chest X-ray and abdominal X-ray should be obtained to monitor for distant CYA embolism (e.g., pulmonary CYA embolism). Surveillance EGD +/− EUS for EV and GV should be arranged, following similar principles used in patients who undergo conventional endoscopic treatment for EV or GV. For example, in patients with both GV and EV with EUS-guided injection therapy for GV, follow-up EGD +/− EUS would be arranged in about 3 to 4 weeks to monitor for any exacerbation of EV after treatment of the GV. If follow-up EGD + EUS show durable obliteration of GV, without exacerbation of EV, then lengthening of the follow-up interval can be considered.

EUS-guided Assessment of Portal Hemodynamics and Creation of Intrahepatic Portosystemic Shunt

While conventional assessment of portal hemodynamics and creation of intrahepatic portosystemic shunts (IPS) are generally performed by IR techniques, EUS-based portal vein catheterization, portal pressure measurement, and creation of IPS have been described.17–22 In 2004, the feasibility of EUS-guided extrahepatic portal vein catheterization and portal pressure measurement using a 22 G needle was first demonstrated in a porcine model involving 7 normal pigs and 14 pigs with induced pHTN.17 The mean portal pressures measured by EUS-guided and transhepatic catheterization were found to be well correlated under all conditions (r = 0.91).17 In another study involving 5 pigs, EUS-guided catheterization of intrahepatic portal vein for angiography and pressure measurements using a 19 G needle was shown to be feasible without complication.18 Recently, the feasibility of rapid EUS-guided needle access and portal pressure measurement using a 22 G needle and a digital pressure wire was demonstrated in porcine model with good correlation with measurement obtained by conventional transjugular catheterization.19 In another recent feasibility study using a porcine model, a new EUS-guided portal pressure measurement system using a 25 G needle and a newly developed compact manometer was found to produce measurements with excellent correlation with standard transjugular approach in low, medium, and high pressure conditions.20

The feasibility of EUS-guided IPS as an alternative to TIPS creation by IR has been explored in animal model. In 2009, EUS-guided IPS creation was successful in a porcine model using a 19 G needle for access of hepatic vein/portal vein and a standard uncovered biliary self-expandable metal stent (SEMS).21 No bleeding or injury to intraperitoneal organs was found in necropsy after acute and survival experiment.21 In another recent study, the feasibility of simultaneous EUS-guided direct portal pressure measurement and creation of IPS using a lumen apposing metal stent (LAMS) with a mean procedure time of 43 minutes has been reported in a porcine model.22 No bleeding was observed in necropsy.22 Further studies would be needed to assess the long term clinical outcomes of these novel vascular intervention for pHTN.

Conclusion

Applications of EUS in pHTN have been increasingly studied over the past decade. EUS has been shown to be useful in variceal and perforating feeding veins identification, prediction of variceal recurrence/rebleeding, and assessment of response to pharmacological therapy for pHTN. EUS-guided therapy for varices and/or perforating feeding veins with coil, CYA injection, or a combination of both was demonstrated to be a valuable addition to conventional endoscopic therapy in recent studies. While the feasibility of EUS-guided assessment of portal hemodynamics and creation of IPS has been reported, further studies would be needed to assess the long term clinical outcomes of these novel EUS-guided vascular interventions for pHTN.

Article information

References