Endo-hepatology: Updates for the clinical hepatologist

Abstract

INTRODUCTION

Endo-hepatology is an emerging field in which diagnostic and therapeutic endoscopic ultrasound (EUS) techniques are used to assist in the diagnosis and management of acute or chronic liver diseases.^1–3 Endo-hepatology provides the opportunity to complete diagnostic and therapeutic procedures in one encounter for patients with liver disease.^2,4,5 As with any diagnostic and/or therapeutic intervention, the clinician should recognize the advantages and limitations of EUS-guided techniques and how they compare with, and potentially transform, the current standard of care. With increasing interest in endo-hepatology, hepatologists should be familiar with the many potential applications; 2 of which will be critically reviewed herein: EUS-portosystemic pressure gradient measurements and EUS-guided liver biopsies; and a brief mention of EUS-guided elastography. Other relevant EUS-guided procedures not covered in this review include coil and/or cyanoacrylate embolization of gastric varices.⁵

EUS-Portosystemic pressure gradient (PPG) measurement

EUS-PPG technical review

EUS-PPG was first described in porcine models demonstrating efficacy and safety along with high concordance with porcine HVPG measurements.^6,7 In 2014, the first case report of successful EUS-guided portosystemic pressure gradient (PPG) measurement was published for a patient with Noonan syndrome, who had multiple HVPG measurements that could not confirm portal hypertension.⁸ Direct pressure measurements of both portal and hepatic veins may provide more accurate PPG assessments, thereby improving diagnostic and prognostic capabilities across a broader range of chronic liver diseases.^1,4,6

Portal pressure gradient measurements using endoscopic ultrasound require a manometry apparatus that includes a 22 or 25 g fine needle aspiration (FNA) needle, noncompressible tubing, a digital manometer, and heparinized saline.¹ The manometry apparatus most used is the Cook EchoTip Insight PPG Measurement System, which includes a 25 g needle, flexible tubing and luer locks, and a digital self-calibrating transducer that displays in mmHg. Patients are positioned supine and administered either monitored anesthesia care or general anesthesia.¹ Under EUS guidance, the FNA needle is introduced in a transgastric approach through hepatic parenchyma to target the hepatic vein. Doppler flow is utilized to confirm the typical multiphasic waveform of hepatic venous flow (Fig. 1). If a hepatic vein cannot be accessed due to thrombosis or diminutive size, the intrahepatic portion of the inferior vena cava can typically be substituted.^1,6,7,9,10 Once punctured, the heparinized saline is flushed, and the manometer is zeroed to the patient’s midaxillary line. Once pressure readings stabilize, 3 measurements are taken, the mean of which is utilized as the hepatic vein pressure in mm Hg.¹ The needle is then slowly withdrawn from the vein into the parenchyma and back into the needle sheath. Using Doppler, the parenchyma is carefully assessed for bleeding. Traversing hepatic parenchyma prior to the puncture of the hepatic vein helps tamponade the site and reduces the risk for extrahepatic bleeding. An intrahepatic branch of the portal vein is then targeted, again in a transgastric approach, utilizing Doppler to confirm venous flow (Fig. 1).¹ Similar techniques including measurements in triplicate are recorded. PPG is then calculated by subtracting the hepatic vein pressure from the portal vein pressure. Prophylactic antibiotics can be given for 3–5 days but are not universally utilized.¹

FIGURE 1.

Endoscopic ultrasound images showing Doppler images of the middle hepatic vein (A) followed by puncture of the middle hepatic vein using a 25 g FNA needle (B); Doppler images of the left portal vein (C) followed by puncture of the left portal vein using a 25 g FNA needle (D).

EUS-PPG success and safety

Given the ability to directly visualize the target veins and avoid adjacent structures, the success rate of EUS-PPG is quite high in reported literature, ranging from 96% to 100% (Table 1). While the reported human experience remains limited, EUS-PPG has been found to be safe with few reported adverse events in several case series (ie, abdominal pain requiring analgesia) without reported severe adverse events.^9,11 In 2018, Huang et al published the first prospective study of 28 patients, 19 with “high clinical evidence” for cirrhosis, who underwent EUS-PPG using a 25 g needle with 100% technical success and no adverse events.¹¹ In another retrospective analysis of 83 patients, there was 100% technical success with no adverse events.⁹ In all reported series, patients with thrombocytopenia (platelet count < 50) and prolonged international normalized ratio (INR) (>1.5) were excluded. Some studies additionally administered periprocedural antibiotics, while others do not report antibiotic use (Table 1).^4,10,11

TABLE 1.

Overview of publications on endoscopic ultrasound-guided portal pressure gradient measurement (EUS-PPG).

Abbreviations: EUS, endoscopic ultrasoundt; FNA, fine needle aspiration; HV, hepatic vein; HVPG, hepatic vein pressure gradient; mHV, middle hepatic vein; NR, not reported; PV, portal vein; PPG, portal pressure gradient measurement; RHV, right hepatic vein.

Comparing EUS-PPG to HVPG

The current gold standard for measurement of portal hypertension requires transjugular i.v. access to measure the free and wedged hepatic vein pressures, with the difference being the HVPG. Portal hypertension is defined as an HVPG greater than 5 mm Hg and is considered clinically significant (CSPH) when greater than or equal to 10 mm Hg.^12,13 Despite their invasive nature, wedged hepatic vein pressures and HVPG are indirect measurements of portal venous pressure and, thus, may be invalid in cases of presinusoidal or prehepatic portal hypertension. Other disadvantages include exposure to radiation (fluoroscopy) and i.v. contrast; respiratory variability with sedation; and significant interoperator variability due to unsuccessful wedge pressure or unrecognized portocaval or other intrahepatic shunting.^1,4,12

HVPG was initially validated for HCV and alcohol-associated cirrhosis. It may not have the same precision in patients with other etiologies of chronic liver disease, particularly NASH, which is rapidly increasing in prevalence.⁷ Patients with NASH were found to have lower wedged hepatic vein pressures and HVPG compared to patients with HCV for each stage of fibrosis.^14,15 In addition, another study of patients with HVPG between 6 and 10 mm Hg, in which the majority had NASH cirrhosis, found that over half had decompensated liver disease at first presentation despite their subclinical HVPG measurements, raising suspicion that HVPG may be underestimating PPG in these patients.¹⁶

Due to the limitations or lack of availability of HVPG, clinicians may favor alternative diagnostic methods, such as EUS-PPG. One challenge against the use of EUS-PPG is that this procedure has not yet been validated against HVPG for most etiologies of liver disease. Initial reports on porcine models found a strong correlation between EUS-PPG and HVPG (r value 0.985–0.99).^6,7 In the only reported human study comparing EUS-PPG and HVPG, Zhang et al found a strong correlation (r value 0.923, P<0.001) between EUS-PPG with a 22 g needle (mean 18.07 +/- 4.3 mm Hg) compared to HVPG (mean 18.82 +/- 3.43 mm Hg) in a small cohort of 9 patients with sinusoidal obstruction syndrome. The procedural time was comparable (38 minutes for EUS-PPG versus 37 minutes for HVPG).¹⁰ However, further studies are needed to validate EUS-PPG compared to HVPG for other causes of portal hypertension. A note of caution, however, in the NASH population where the current gold standard of HVPG has been questioned, prospective outcome studies are needed.¹⁵ A large multicenter international clinical trial utilizing EUS-PPG has recently concluded and may further strengthen support for this technique (NCT04668664).

EUS-PPG clinical implications

As the EUS-PPG technique evolves, our understanding of its clinical correlation with complications of liver disease will be further defined. In Huang et al, EUS-PPG was significantly higher in patients with high pretest probability for CSPH (10.33 mm Hg versus 3.81 mm Hg, P=0.005), varices (14.37 mm Hg versus 4.26 mm Hg, P=0.0002), and portal hypertensive gastropathy (12.76 mm Hg versus 6.09 mm Hg, P=0.007).¹¹ Another retrospective analysis found that EUS-PPG was significantly higher in patients with known cirrhosis, esophagogastric varices, and/or thrombocytopenia.⁹ Furthermore, they found that EUS-PPG greater than or equal to 5 mm Hg was significantly associated with biopsy-proven fibrosis stage 3 or greater, with a likelihood ratio of 27 (P=0.004).^9,17 However, a prospective study of 24 patients undergoing EUS-PPG and EUS-liver biopsy demonstrated that, while higher EUS-PPG significantly correlated with transient elastography findings, FIB-4, and APRI score, there was no correlation between EUS-PPG and fibrosis stage.⁴

EUS-PPG disadvantages

As an emerging discipline, the disadvantages of EUS-PPG arise from our incomplete understanding of various procedural factors that may influence portal pressure measurements. As outlined above, there is a strong need to not only validate EUS-PPG compared to HVPG but also to understand its clinical implications and prognostic capability for a variety of liver diseases. Further investigations are needed to establish the feasibility and reproducibility of EUS-PPG for various underlying etiologies of liver disease or degrees of portal hypertension. Studies assessing acute and chronic effects of medical therapy for portal hypertension (eg, nonselective beta blockers) on EUS-PPG are also needed. One major limitation has been the exclusion of patients with the most advanced stages of portal hypertension, typically associated with platelet count <50, prolonged INR, and/or large ascites. Whether this technique can be safely applied to this population remains unknown. Furthermore, the availability of advanced endoscopists with expertise in EUS-PPG measurements is currently limited to select centers. Current literature has not defined the impact of the type of anesthesia on EUS-PPG, similar to HVPG. Therefore, our current knowledge has yet to identify the ideal patient population for which EUS-PPG would be a preferred procedure over the traditional transjugular HVPG measurement.

EUS-liver biopsy (EUS-LB)

Liver biopsies remain the gold standard for the diagnosis and staging of liver disease.^18,19 The adequacy of liver biopsy is dependent on three factors: total width and length of the specimen, intact specimen length (as opposed to fragments), and the number of complete portal tracts (CPTs). Benchmarks for successful biopsy, according to the American Association for the Study of Liver Diseases (AASLD), are those 20 mm in length with 11 or more CPT.¹⁸

EUS-LB technical review

Since the emergence of EUS-guided liver biopsy (EUS-LB) roughly 10 years ago, many techniques (dry suction, wet suction, slow pull, modified wet suction, and heparin) have been described (Table 2.) Techniques with more reliable yield involve one or more passes and actuations using wet suction with either saline or heparin and a newer generation 19 g core needle, preferably with a Franseen tip.^19–21 Some endosonographers prefer completing 2 passes with 3–5 actuations each to increase yield. Both passes are typically obtained from the left lobe; however, bilobar sampling can be considered if disease findings may be patchy. The left lobe is accessed through the transgastric approach and the right lobe through the transduodenal approach. The 19 g needle is prepared by removing the stylet and flushing it with heparin. A syringe is then primed with 20 mL of saline with a vacuum, termed “wet suction.” The use of wet suction generates greater pulling force and increases tissue yield while decreasing bleeding.²² Heparin is additionally used as it decreases clot formation and the formation of a “blood noodle.” After the needle and syringe are prepared, the endoscopic exam begins with a linear EUS scope with the 19 g needle in the endoscopic working chamber. Using the Doppler waveform, an area of hepatic parenchyma of at least 4 cm without intervening vasculature or biliary structures is targeted. The heparin syringe is then detached from the needle, and 1 pass with up to 5 actuations is completed in 1 quick motion, often with a fanning technique (Fig. 2). The activated wet suction syringe is then attached to the proximal port of the needle. Suction is created by rotating the stopcock 180°, allowing a heparin splash into the suction syringe, indicating that tissue is now in the needle. The suction syringe stopcock is then closed, allowing for the withdrawal of the needle back into its sheath. Doppler ultrasound is used to observe for bleeding or fistula tract creation after the needle is withdrawn. Once the tissue is obtained, the sample is placed directly into formalin from the needle to decrease tissue fragmentation and increase specimen quality. Patients are sedated with monitored anesthesia care or general anesthesia throughout the procedure. After the procedure, patients are observed for 1 hour. If patients have altered anatomy, such as a Roux-en Y, the left lobe may still be accessed through a proximal transgastric approach.²¹

TABLE 2.

Overview of publications on endoscopic ultrasound-guided liver biopsy (EUS-LB).

Abbreviations: FNA, fine needle aspiration; FNB, fine needle biopsy; NR, not reported.

FIGURE 2.

Endoscopic ultrasound image showing core biopsy of the left lobe of the liver using a 19 G FNB needle (arrow).

EUS-LB safety

EUS-LB has been verified as a safe technique. Most endoscopists will proceed with EUS-LB with platelets > 50,000, INR <1.5, and the absence of large ascites.^23,24 In 1 meta-analysis of 437 patients undergoing EUS-LB, bleeding occurred in 1.2%, with other adverse events, defined as any deviation from expected postprocedure course, in 2.3%.²⁵ Another meta-analysis found no differences in adverse event rates between EUS-LB, interventional radiology (IR) LB, and percutaneous (P) LB.²⁶ In one single-center comparison of 152 liver biopsies, IR-guided LB, which included transjugular and percutaneous, had more complications compared to EUS-LB (P=0.03).²⁷ One randomized prospective study, reporting interim results for 80 patients in abstract form, suggested that EUS-LB had significantly lower rates of delayed (24–48 h) postprocedure pain (P=0.001) and shorter postprocedural stay of 2.2 h (SD 1.6) compared to 4 h after P-LB (SD 1.6) (P=0.001). However, these investigators noted a decreased histologic yield for EUS-LB as described in the following section.²⁰

Comparing EUS-LB to traditional methods of liver biopsy—yield

Historically, concerns about adequate yield have dampened enthusiasm for EUS-LB as early experience noted suboptimal specimen lengths, heavily fragmented samples, and few CPTs.²⁸ While there has yet to be universal protocolization of EUS-LB, recent advances have increased diagnostic yield, including the utilization of a 19 g Franseen core needle, heparinization, wet suction, and the fanning pass and actuation technique. One recent randomized control trial (BLOCS trial), presented in abstract form, found that modified wet suction compared with dry slow pullback technique demonstrated increased total fragment length (46.5 mm versus 34.5 mm, P<0.0001), increased length of the longest fragment (14 mm versus 11 mm, p 0.0002), and increased number of portal tracts (16 versus 11.5, p = 0.0006).²¹

Several studies have found that the histologic yield of EUS-LB is now comparable to percutaneous and transjugular LB (Table 2). Most studies suggest the diagnostic yield of an EUS-LB ranges from 93% to 100% compared to percutaneous (P-LB) and transjugular LB (TJ-LB).^24,29 However, there are two studies demonstrating the histologic superiority of P-LB or TJ-LB over EUS-LB. One single-center randomized prospective study, reporting interim results in abstract form, found P-LB obtained more CPT compared to EUS-LB (18.9 versus 12.5, P=0.009); however, overall sample length was similar (19.4 mm versus 19.6 mm, P=0.93).²⁰ Another randomized prospective study found that P-LB may have greater diagnostic yield compared to EUS-LB although both seemed unacceptably low (58% versus 24%, P=0.028).³⁰ Critiques of this study cite no statistical difference in CPT obtained between the two methods. Furthermore, EUS-LB samples were collected without using techniques, such as suctioning and fanning, which are now considered standard of care.^31,32 A meta-analysis, which incorporated these two prospective studies along with retrospective studies, concluded that EUS-LB and P-LB are statistically similar in diagnostic adequacy and adverse events although P-LB may obtain more CPT; however, this advantage dissipates when analyzing only prospective randomized clinical trial data.³³ Results of ongoing studies comparing P-LB and TJ-LB to EUS-LB utilizing current standard of care techniques are eagerly awaited (NCT04751045, NCT05118308, and NCT05817994).

EUS-LB—cost

The comparative cost of EUS-LB has yet to be fully elucidated, but, as endo-hepatology continues to mature, many anticipate a “one stop shop” model for variceal screening, portal pressure gradient measurements, and liver biopsy. One of the randomized prospective studies noted above suggested P-LB is significantly less costly compared to EUS-LB (US$1824 versus US $3240, P<0.001).³⁰ In a simulated model of 50 patients with NAFLD, presented only in abstract form, the cost of EUS-LB with monitored anesthesia compared to P-LB with conscious sedation was $2,610 versus $2,660, respectively, after estimating costs of complications and specimen inadequacy.³⁴ Given the requirements for anesthesia care and endosonography for EUS-LB, it appears unlikely to be more cost-efficient compared to P-LB alone but may be more advantageous for patients with additional indications for endoscopic examination.

Additional considerations for EUS-LB compared to P-LB and TJ-LB

EUS-guided FNA can also target hepatic lesions that are difficult to otherwise access from conventional US or CT-guided percutaneous techniques. The diagnostic rate of EUS-FNA of hepatic lesions ranges from 82% to 100%.^35,36 Furthermore, Nguyen et al found that CT demonstrated only 21% of the liver lesions subsequently detected and biopsied via EUS (93% malignant).³⁷ By modifying the staging of various malignancies, EUS-guided FNA of hepatic lesions has been shown to change the management in 59%–86% of patients.^28,35

Limitations of EUS-LB

As an emerging discipline, the largest drawback for EUS-LB is that the techniques have not been universally protocolized and are not a formal part of all EUS training, leading to variable diagnostic yields based on the technical skill of each endoscopist. EUS-LB additionally requires sedation, typically with an anesthesiologist, which may not be preferable to some patients and providers. Just as with percutaneous or transjugular biopsies, the benefits of histologic evaluation of liver parenchyma must be weighed against the risk of bleeding in patients with significant coagulopathies.

Exploratory applications of EUS-elastography

One promising field of endo-hepatology not covered in this review is endoscopic ultrasound shear wave elastography (EUS-SWE). EUS-SWE allows noninvasive assessment of liver fibrosis in patients for whom transabdominal elastography is less accurate (ie, obesity). A randomized controlled trial comparing the diagnostic accuracy of EUS-SWE and transabdominal vibration–controlled transient elastography to liver histology found that EUS-SWE has similar AUROC to vibration–controlled transient elastography for patients with advanced fibrosis and cirrhosis without any statistical difference.³⁸ Furthermore, 8 of the 42 patients in the study had unreliable vibration–controlled transient elastography but consistent EUS-SWE.³⁸ Future prospective studies are needed to expand on this proof-of-concept study.

Summary and future directions

The field of endo-hepatology continues to grow with the potential to advance the diagnostic and therapeutic management of patients with acute and chronic liver diseases. EUS-PPG is an evolving technique that directly measures portosystemic pressure gradients, which may have specific advantages over current methods that only estimate these pressure gradients (Table 3). Future studies are needed to validate EUS-PPG with HVPG measurements and prospectively assess prognostic capabilities, especially for patients with NASH where the gold standard may be suboptimal. We look forward to the presentation of results from ongoing clinical trials in this area.

TABLE 3.

Potential advantages, disadvantages, and patient populations best served by endoscopic ultrasound-guided liver biopsy (EUS-LB) and portal pressure gradient (EUS-PPG).

Abbreviation: NRH, nodular regenerative hyperplasia.

On the other hand, there is a robust literature supporting the use of EUS-LB with second-generation core needles in clinical practice that demonstrates both adequacy in biopsy yield and safety, and tolerability for patients. Through close collaboration with locally advanced endoscopists and increasing use of EUS techniques as a “one stop shop,” both patients and clinicians may favor EUS-LB for histologic sampling in the future (Table 3). Future studies could investigate the economic impact of a single endoscopic procedure potentially addressing several clinical concerns (ie, screening/treating varices, measuring PPG, and sampling hepatic parenchyma) that endo-hepatology offers. We expect these future investigations will further clarify the optimal patient population that endo-hepatology serves, given the changing burden of chronic liver disease.