Improved clinical outcomes following embolization of extrahepatic portosystemic shunts in cirrhotic patients with recurrent hepatic encephalopathy

Abstract

BACKGROUND

Hepatic encephalopathy (HE) affects more than 30% of patients with cirrhosis. Extrahepatic portosystemic shunt (EHPSS) has been suggested to be a contributing factor to HE recurrence and mortality. Therefore, early detection and intervention in EHPSS may improve patient outcomes.

AIM

To evaluate the effects of shunt embolization on mortality and HE recurrence.

METHODS

In this retrospective case-control study, 16 cirrhotic patients with HE treated at a tertiary care center from January 2012 to August 2022 were included. Outcomes in eight patients who underwent embolization of EHPSS were compared with those in eight patients receiving standard care without embolization. Data on baseline characteristics, HE recurrence, and overall survival were collected and analyzed using Kaplan-Meier and log-rank tests.

RESULTS

Baseline characteristics were comparable between the groups. The 1-year overall survival rate was significantly higher in the treatment group (0.50) than in the control group (0.33). The HE recurrence-free rate was also higher in the treatment group (1.00) than in the control group (0.17). The median survival duration was longer in the treatment group {not reached [95% confidence interval (CI): 23.84 to not available (NA)]} than in the control group [15.02 months (95%CI: 9.86 to NA)] (P = 0.006). Similarly, the recurrence-free duration was longer in the treatment group [63.09 months (95%CI: 63.09 to NA)] than in the control group [9.21 months (95%CI: 4.47 to NA)] (P = 0.006). EHPSS embolization significantly reduced 1-year HE recurrence (hazard ratio = 0.09; 95%CI: 0.01-0.75; P = 0.026).

CONCLUSION

EHPSS embolization significantly improves 1-year survival and prevents recurrence of HE in cirrhotic patients. Routine computed tomography and early embolization are clinically beneficial.

Key Words: Liver cirrhosis; Hepatic encephalopathy; Extrahepatic portosystemic shunt; Embolization; Survival; Recurrence

Core Tip: Extrahepatic portosystemic shunt (EHPSS) is an underrecognized contributor to hepatic encephalopathy (HE) recurrence and mortality in patients with cirrhosis. In this study, EHPSS embolization significantly improved 1-year survival (50% vs 33%) and prevented HE recurrence (100% vs 17%) compared with no intervention. Embolization also extended both survival and recurrence-free duration. These findings highlight the importance of routine abdominal computed tomography for early EHPSS detection and timely embolization to improve outcomes in cirrhotic patients with recurrent HE.

INTRODUCTION

Hepatic encephalopathy (HE) is common in patients with chronic liver disease, with a prevalence of 20%-80% in patients with liver cirrhosis[1]. Some patients with HE experience chronic and recurrent disease, which impacts not only the quality of life of these patients but also places a burden on health economics due to high resource utilization[2]. Gut-derived neurotoxins are not removed by the liver owing to vascular shunting; decreased hepatic mass allows these toxins to reach the brain. Many of these patients have spontaneous portosystemic shunts that are often large enough to divert a major portion of the portal blood flow[2-4]. Previous reports have suggested that 40%-70% of patients with refractory HE demonstrate an extrahepatic portosystemic shunt (EHPSS) upon radiological screening[5]. These shunts were regarded as a compensatory safety valve in progressive portal hypertension, diverting blood flow and thereby partially decompressing the portal venous tributaries. However, studies including a multicenter analysis supported by the Baveno VI consensus group have revealed that large spontaneous portosystemic shunts are associated with increased risk of portal hypertension-related complications such as HE and death[6]. Recent studies have highlighted EHPSS embolization as a viable intervention for managing refractory HE. Laleman et al[2] demonstrated that embolization of a large EHPSS significantly alleviated HE symptoms in most patients, improving autonomy and reducing hospitalization rates. The study also identified the model for end-stage liver disease (MELD) score as a predictor of post-procedural outcomes, emphasizing the importance of patient selection.

An et al[7] further corroborated these findings, showing that EHPSS embolization not only reduced HE recurrence but also enhanced liver function and overall survival of patients with relatively preserved hepatic reserve (MELD score < 15). However, an improvement in overall survival was observed only in a limited subgroup within the overall study population. They emphasized the importance of timely intervention and restoration of hepatic portal perfusion to mitigate the effects of portosystemic shunting. Moreover, recent studies, including meta-analysis by Song et al[8], have confirmed the therapeutic benefit of EHPSS embolization in reducing HE recurrence, and supported its role as an effective treatment in selected cirrhotic patients[9,10]. However, the survival benefit of the procedure was not well demonstrated. In this study, we aimed to compare the therapeutic efficacy and safety between embolization and control groups and to retrospectively determine the therapeutic efficacy and safety of embolization.

MATERIALS AND METHODS

Study patients

Nine patients underwent portosystemic shunt embolization between January 2012 and August 2022. Patients in the treatment group experienced more than two episodes. Twenty patients in the control group were also screened during the same period. We retrospectively reviewed the patients’ medical records. This study was approved by the Institutional Review Board of Chungbuk National University Hospital, and the requirement for written informed consent was waived.

The inclusion criteria were: History of admission, age between 19 and 80 years, liver cirrhosis with not less than two episodes of HE, and early-stage and cured hepatoma. The exclusion criteria were: Age < 19 or > 80 years, Child-Turcotte-Pugh (CTP) score > 13, advanced-stage hepatoma (not less than stage II), large esophago-gastric varices (grade > 2), recipient with surgical portocaval shunt, transjugular intrahepatic portal systemic shunt, liver transplantation, and other cognitive disorders such as cerebrovascular accident, Parkinson’s disease, and dementia.

Each episode of HE was graded with the West Haven criteria[11]. Only episodes classified as grade 2 or higher were counted, and patients who had at least one such episode were eligible for inclusion. Because the study population consisted solely of individuals with overt HE, psychometric and neurophysiological assessments were not performed. Grade 2 HE was defined by the presence of one or more of the following: Disorientation in time, place, or person; bizarre behavior; obvious asterixis; or impaired serial-subtraction performance. All participants underwent portal-venous-phase multidetector computed tomography to identify and characterize portosystemic shunts. Extra-hepatic shunts considered relevant were splenorenal communications and recanalized para-umbilical veins. Shunt embolization was considered eligible in patients with an EHPSS diameter greater than 8 mm. Before embolization, every patient received optimized medical therapy - including correction of precipitating factors and administration of lactulose and/or rifaximin - according to standard practice.

EHPSS embolization was attempted angiographically in nine patients; the shunts were completely occluded in all patients (treatment group). One 31-year-old female patient was excluded from the treatment group due to other cognitive disorders and a CTP score of 13. Twenty patients who did not undergo EHPSS embolization were initially screened. Five of these patients were excluded because they had advanced hepatocellular carcinoma; the other seven were excluded because they had less than two episodes of HE. Finally, eight patients were included in the control group (Figure 1). The baseline was defined as the date of the first shunt in the treatment group and the date of the second hospitalization at our institution for the management of the control group. The patients were followed up for survival, HE recurrence, liver function parameters, and liver volume. The HE recurrence was defined as hospitalization or emergency room visits for management. The occurrence or worsening of ascites or esophagogastric varices was also assessed. All clinical and laboratory data necessary for the analysis were collected as much as possible, but the presence of ascites was evaluated based only on imaging studies.

Figure 1 Consort diagram of patient selection. EHPSS: Extrahepatic portosystemic shunt; CTP: Child-Turcotte-Pugh; HCC: Hepatocellular carcinoma; HE: Hepatic encephalopathy.

Shunt embolization procedures

In all cases, large (20-40 mm) shunts were suspected on computed tomography and confirmed via arteriography. Ultrasonography revealed splenorenal shunts. Angiographic access for embolization was performed via the femoral vein in seven splenorenal and one mesocaval shunts. All procedures were carried out under local anesthesia. Shunt occlusion was achieved with either an Amplatzer vascular plug (AVP) (Medical, Golden Valley, MN, United States) or a combination of detachable coils and gelatin-sponge (Gelfoam, Upjohn, Kalamazoo, MI, United States) with up to six steel coils of 5-10 mm in diameter. The coils were placed within the sinus lumen through a femoral catheter in all patients.

All patients underwent initial diagnostic angiography under local anesthesia and analgesics via the transfemoral approach in four splenorenal shunts and the transhepatic approach in four mesocaval shunts using a 5-F vascular sheath. The shunts were catheterized with a 5-F angiographic catheter (Cobra, Cook, Bloomington, IN, United States) and a 2.4-F microcatheter (Progreat, Terumo, Somerset, NJ, United States) in four patients, whereas a 9-F guiding sheath (Flexor Ansel, Cook Medical, Bloomington, IN, United States) was introduced to deploy the vascular plug in four patients. Embolization was performed using vascular plugs (AVP; Medical, Golden Valley, MN, United States) or coils (Interlock, Boston Scientific Co., Natick, MA, United States) combined with gelatin sponges (Gelfoam, Upjohn, Kalamazoo, MI, United States). The AVP II was utilized. Furthermore, correctly sized detachable coils (average 11.1 ± 5.3, range 5-22) were deployed.

Outcomes

The primary outcome was overall survival. The secondary outcome was the recurrence of overt HE for 3 months. HE of grade 2 or higher was considered a recurrence, regardless of the presence of an identifiable precipitant. Changes in liver function, plasma ammonia levels, and safety laboratory tests were evaluated before and after the procedure. MELD scores and CTP scores were assessed before and 1 year after the shunt embolization.

Statistical analysis

Since the data did not follow a normal distribution and sample sizes of both the treatment (n = 8) and control groups (n = 8) were small, a non-parametric statistical analysis was applied. Categorical variables were compared and analyzed using Fisher’s exact test, and continuous variables were compared using the Mann-Whitney U test. Kaplan-Meier analyses were performed for survival and cumulative HE recurrence rates, and differences between the two groups were compared and tested using the log-rank test. The Cox proportional hazards model was used to identify the risk factors associated with patient mortality and HE recurrence. The variables included in the Cox hazard model were group (treatment vs control), age, sex, previous HE episodes within 3 months, HE grades (II vs III/IV), albumin, total bilirubin, creatinine, PT-international normalized ratio, CTP score (B vs C), and MELD score. Statistical significance was defined as P < 0.05; all analyses were performed using R version 4.4.1 (R Foundation for Statistical Computing, Vienna, Austria).

RESULTS

Baseline characteristics of patients

The baseline characteristics of the treatment (n = 8) and control (n = 8) groups were comparable across all demographic and biochemical parameters (Table 1). The etiology of cirrhosis in the treatment group included four, one, and two cases of alcoholic cirrhosis, autoimmune hepatitis, and cryptogenic disease, respectively. The control group had one case of hepatitis B virus and seven cases of alcoholic cirrhosis, respectively. The median age was 63 years (range, 45-74) in the treatment group and 57 years (range, 42-71) in the control group (P = 0.27). The incidence of esophageal varices was 50% in the treatment group and 87.5% in the control group (P = 0.11). Albumin levels were similar between the groups, measuring 3.10 g/dL (range, 2.85-3.25) in the treatment group vs 3.15 g/dL (range, 2.75-3.55) in the control group (P = 0.64). The bilirubin levels were 2.51 mg/dL (range, 0.26-4.76) and 2.88 mg/dL (range, 0.85-4.91) in the treatment and control groups, respectively (P = 0.35). Creatinine levels were 0.93 mg/dL (range, 0.48-1.38) in the treatment group and 1.38 mg/dL (range, 0.17-2.59) in the control group (P = 0.67). The international normalized ratio was 1.42 (range, 1.12-1.72) in the treatment group and 1.41 (range, 1.22-1.60) in the control group (P = 0.875). The MELD scores were 11.08 (range, 5.32-16.84) vs 14.08 (range, 7.57-20.59) (P = 0.25). The CTP scores in the two groups were 8.88 (range, 7.33-10.43) vs 9.38 (range, 8.08-10.68) (P = 0.43).

Table 1 Baseline characteristics of the participants, n (%).

Group	Control group (n = 8)	Treatment group (n = 8)	P value
Age (years)	57 (42-71)	63 (45-74)	0.269
Male	6 (75.0)	3 (37.5)	0.143
Etiology
HBV	1 (12.5)	0	0.044
Alcohol	7 (87.5)	5 (62.5)
Others	0	3 (37.5)
Number of HE episodes during the previous 12 months
1-2	4 (50.0)	5 (62.5)	0.626
≥ 3	4 (50.0)	3 (37.5)
HE grades
II	4 (50.0)	5 (62.5)	0.626
III-IV	4 (50.0)	3 (37.5)
Esophageal varix (%)	87.5	50.0	0.106
Albumin (g/dL), median (IQR)	3.15 (2.75-3.55)	3.10 (2.85-3.25)	0.635
Bilirubin (mg/dL), median (IQR)	2.88 (0.85-4.91)	2.51 (0.26-4.76)	0.345
Creatinine (mg/dL), median (IQR)	1.38 (0.17-2.59)	0.93 (0.48-1.38)	0.674
INR, median (IQR)	1.41 (1.22-1.60)	1.42 (1.12-1.72)	0.875
MELD × 10, median (IQR)	14.08 (7.57-20.59)	11.08 (5.32-16.84)	0.248
CTP score, median (IQR)	9.38 (8.08-10.68)	8.88 (7.33-10.43)	0.431

HBV: Hepatitis B virus; HE: Hepatic encephalopathy; IQR: Interquartile range; INR: International normalized ratio; MELD: Model for end-stage liver disease; CTP: Child-Turcotte-Pugh.

Patient survival and recurrence of HE

The 1-year overall survival rate was significantly higher in the treatment group than in the control group (0.50 vs 0.33). The median survival duration was longer in the treatment group than in the control group {not reached (NR) [95% confidence interval (CI): 23.84 to not available (NA)] vs 15.02 months (95%CI: 9.86 to NA)} (P = 0.006) (Figure 2A). The recurrence-free rate was significantly higher in the treatment group than in the control group (1.00 vs 0.17). The recurrence-free duration was longer in the treatment group than in the control group [63.09 months (95%CI: 63.09 to NR) vs 9.21 months (95%CI: 4.47 to NR)] (P = 0.006) (Figure 2B).

Figure 2 Kaplan-Meier curves for overall survival and recurrence-free survival in patients with extrahepatic portosystemic shunt. A: Kaplan-Meier curves for overall survival in patients with extrahepatic portosystemic shunt; B: Kaplan-Meier curves for recurrence-free survival in patients with extrahepatic portosystemic shunt.

Predictors of mortality and HE recurrence in the overall study subjects

Using univariate analysis, EHPSS embolization treatment was a significant predictor of reduced HE recurrence in all patients [hazard ratio (HR) = 0.09; 95%CI: 0.01-0.75; P = 0.026]. The presence of severely decompensated cirrhosis - CTP C was a significant predictor of HE recurrence (HR = 5.61; 95%CI: 1.02-30.9; P = 0.048). No significant predictors of mortality were identified in univariate analysis (Table 2). Because only a limited number of HE-recurrence events occurred, multivariate analysis was not performed, so as to avoid overfitting.

Table 2 Predicting mortality and recurrence in patients with hepatic encephalopathy: A Cox proportional hazards model analysis.

Characteristic	Mortality			HE recurrence
	HR	95%CI	P value	HR	95%CI	P value
Group
Control group	-	-		-	-
Treatment group	0.30	0.06-1.49	0.139	0.09	0.01-0.75	0.026a
Age (years)	0.97	0.91-1.04	0.430	0.95	0.88-1.02	0.147
Sex
Female	-	-		-	-
Male	1.45	0.34-6.15	0.611	1.54	0.34-6.99	0.573
Number of HE episodes during the previous 3 months	0.00	0.00, infinity	0.999	0.98	0.34-2.89	0.978
HE grades
II	-	-		-	-
III/IV	0.77	0.18-3.24	0.721	2.86	0.55-15.0	0.213
Albumin	0.35	0.09-1.37	0.132	0.98	0.19-5.06	0.979
Total bilirubin	1.13	0.80-1.61	0.479	1.20	0.88-1.64	0.245
Creatinine	0.71	0.24-2.10	0.531	1.12	0.55-2.28	0.760
INR	3.07	0.15-61.8	0.464	0.92	0.03-25.3	0.961
MELD	0.99	0.32-3.08	0.991	1.19	0.39-3.67	0.758
CTP class
B	-	-		-	-
C	2.35	0.50-11.0	0.277	5.61	1.02-30.9	0.048a

^aP < 0.05.

HR: Hazard ratio; CI: Confidence interval; HE: Hepatic encephalopathy; INR: International normalized ratio; MELD: Model for end-stage liver disease; CTP: Child-Turcotte-Pugh.

Embolization outcome

Four patients with splenorenal shunts were approached via the transfemoral route; a coil was used for embolization in three of them, and a vascular plug was used in one patient. Four patients with mesocaval shunts were approached transhepatically; a coil was used for embolization in two patients, and a vascular plug was used in the remaining two patients. Embolization was successful in all patients during the first attempt. No procedure-related complications occurred. No encephalopathy or variceal bleeding was observed after embolization. The transfemoral approach was used in four splenorenal patients, and the transhepatic approach was used in four patients with mesocaval shunts (Figure 3). Plasma ammonia levels decreased substantially after embolization, with the median falling from 161.5 μmol/L to 86 μmol/L (Table 3).

Figure 3 Angiographic images of splenorenal shunt embolization and mesocaval shunt embolization. A: After a 9-F vascular sheath was placed in the splenorenal shunt, the shunt was accessed coaxially using a 5-F cobra catheter; B: The vascular plug was deployed at the narrowest region of the splenorenal shunt through the vascular sheath and completion venography showed complete occlusion of the splenorenal shunt; C: Transhepatic superior mesenteric venogram showed the large size of mesocaval shunt; D: Coil placement showed complete closure of mesocaval shunt.

Table 3 Plasma ammonia levels before and after shunt embolization of the treatment group.

Subject number	Plasma ammonia level before embolization		Plasma ammonia level after embolization
	Date of measurement	Lab result (μmol/L)1	Date of measurement	Lab result (μmol/L)1
1	August 10, 2023	100	August 17, 2023	92
2	May 10, 2022	130	February 17, 2022	90
3	May 22, 2021	218	June 16, 2021	60
4	December 26, 2020	213	December 29, 2020	95
5	January 17, 2020	143	March 12, 2020	118
6	September 8, 2017	180	October 11, 2017	62
7	April 4, 2014	237	April 18, 2014	45
8	June 27, 2023	83	July 27, 2023	82
		Median (161.5)		Median (86)
		mean ± SD (163.0 ± 57.5)		mean ± SD (80.5 ± 23.5)

¹Reference range: 0-65 μmol/L.

DISCUSSION

HE is a debilitating complication of cirrhosis. Alteration of mental status leads to the inability to make correct judgments and eventual job loss. Standard medical treatments cannot prevent recurrence. The mainstays of treatment include non-absorbable disaccharides such as lactulose and antibiotics such as rifaximin, which decrease ammonia production in the gut[12]. These treatments effectively reduce the incidence and severity of HE, improve patient outcomes, and reduce healthcare costs[13]. There remains an unmet need for a more concrete or definitive treatment method before liver transplantation[13]. A retrospective cohort study of 1729 patients showed that EHPSS were linked to higher rates of gastroesophageal varices, variceal bleeding, ascites, and hepatorenal syndrome. Moreover, transplant-free survival was poorer in the EHPSS group, with an adjusted HR for death or liver transplantation of 1.6 (95%CI: 1.33-1.93) compared with patients of similar MELD score who lacked an EHPSS[6].

Our study showed that the embolization procedure could reduce overall mortality and HE recurrence. The 1-year survival rates in the embolization and control groups were 75% and 38%, respectively. No safety concerns were observed in the embolization group. These findings align with those in previous reports. Ke et al[14] demonstrated improved HE-free survival following embolization in a multicenter cohort using an inverse probability of treatment weighting analysis, with safety confirmed in appropriately selected patients. Similarly, Rajesh et al[15] showed 1-year survival of 87% and a significant reduction in HE recurrence. However, concerns remain that occluding a large shunt may acutely raise portal pressure, thereby aggravating ascites or gastroesophageal varices and increasing the risk of variceal bleeding in a subset of patients, which may necessitate the addition of non-selective β-blockers or prophylactic endoscopic therapies. Accordingly, the worldwide uptake of EHPSS embolization as routine therapy is still limited, and further prospective, controlled studies are needed to define which patients derive net benefit[2].

Our findings are in agreement with those in earlier investigations, demonstrating that neurological symptoms and HE recurrence were reduced after treatment when shunt embolization was combined with standard therapy for recurrent HE accompanied by EHPSS[16]. An et al[7] reported that embolization prevented HE recurrence and improved survival in patients with preserved liver function, whereas no benefit was observed in those with advanced liver dysfunction. Seven patients showed poor outcomes after embolization, with four deaths due to sepsis or variceal bleeding, and only two patients survived, one of whom required liver transplantation[5]. A European multicenter study involving 37 patients found that embolization of EHPSS controlled HE in 60% of cases at three months; half of the cohort maintained an HE-free status two years after the procedure[2]. Two predictors may affect recurrence. Severely decompensated cirrhosis - CTP C and embolization were significant predictors of HE recurrence in both patient groups. Many studies have shown that hepatic reserve function is an important predictor of therapeutic outcomes[2,7,17].

In this study, embolization was performed by an interventionist at the authors’ institution. This procedure was successful on the first attempt, after which no early or late procedure-related complications were observed. There were two types of EHPSS: Four splenorenal and four mesocaval shunts. We suggest that the findings in our study were more consistent and robust than those reported by other institutions, because variations in interventional expertise may have biased treatment outcomes. The types of shunts were diverse in each study. The most common type was splenorenal, ranging between 21% and 100%[18-21].

In the embolization group, one patient was ineligible due to a high CTP score. This patient, who showed HE recurrence after embolization, already had six HE episodes in the year before embolization, a MELD score of 14.71, and a CTP score of 13. The effectiveness of shunt embolization may be reduced in patients with HE; this might have severely reduced baseline liver function. This study confirmed that the prognosis after embolization in patients with liver cirrhosis and HE symptoms could be predicted based on preprocedural hepatic reserve function. Univariate analysis showed that severely decompensated cirrhosis - CTP C and embolization were significant predictors of HE recurrence in all patients. Several previous studies have suggested that the effectiveness of embolization might decrease depending on the baseline cirrhosis status before shunt embolization. An et al[7] analyzed patients with MELD score < 15 and no hepatocellular carcinoma, finding a markedly superior survival rate in the embolization group (100% vs 60%) among those who did not undergo the procedure. Inoue et al[19] reported that serum albumin levels < 2.8 mg/dL were linked to decreased survival. Accordingly, shunt embolization is usually reserved for the small subset of HE patients who still have relatively preserved liver function.

This study has several limitations. First, the small sample size of eight patients in each group reflects the rarity of HE due to EHPSS and might limit the generalizability of our findings. Hence, this issue requires further study, especially in prospective controlled trials. A prospective study would require approximately 33 patients to detect a significant difference in 6-month recurrence-free survival rates of 45% in the control group and 22% in the embolization group. Second, as a single-center study conducted in a medium-sized tertiary hospital, external validity might be limited. Third, although all procedures were performed by a single interventionist, this ensured consistency in technique throughout the study. Given that the procedure does not require highly specialized equipment or exceptional expertise, it may be reproducible in similar clinical settings. Finally, the retrospective design introduces potential selection bias. Nonetheless, we applied strict and consistent inclusion criteria and matched controls by age, sex, and CTP class to improve comparability and reduce confounding.

CONCLUSION

Shunt embolization improves 1-year survival and HE recurrence-free rates in patients with cirrhosis. Embolization and baseline decompensated cirrhosis are associated with HE recurrence. Early detection and timely embolization of EHPSS might improve outcomes in selected patients.

ACKNOWLEDGEMENTS

We would like to thank Dr. Suk Ran Kim for her work in editing this manuscript.