Transjugular intrahepatic portosystemic shunt combined with viral suppression reverses decompensated cirrhosis: Two case reports

Abstract

BACKGROUND

Decompensated cirrhosis, particularly when complicated by portal hypertension, has traditionally been regarded as an irreversible condition with a progressive clinical course.

CASE SUMMARY

We present 2 male patients (aged 43 and 50) with chronic hepatitis B and hepatitis C, respectively. Both patients had confirmed decompensated cirrhosis and significant splenomegaly, and were admitted for acute variceal hemorrhage refractory to endoscopic variceal ligation. Each underwent successful transjugular intrahepatic portosystemic shunt (TIPS) creation using a covered stent, which led to an immediate and substantial reduction in the portosystemic pressure gradient. While the second patient required shunt revision at 6 months due to stent stenosis, both were followed for over a decade. During this period, neither experienced recurrence of variceal bleeding, ascites, or hepatic encephalopathy. Liver function improved from Child-Pugh class B to A, viral loads remained undetectable under antiviral therapy, and spleen size regressed progressively. Remarkably, even after eventual TIPS occlusion in one case, the portosystemic pressure gradient remained within the normal range, and angiography revealed no visible esophageal varices. Sequential liver biopsies demonstrated significant regression of hepatic fibrosis and the disappearance of cirrhotic architecture, indicating substantial fibrosis regression.

CONCLUSION

This report suggests that in carefully selected patients, the combination of TIPS and effective etiological control can facilitate the long-term reversal of portal hypertension and promote histological regression of cirrhosis itself.

Key Words: Decompensated cirrhosis; Portal hypertension; Transjugular intrahepatic portosystemic shunt; Antiviral therapy; Regression; Case report

Core Tip: This case series demonstrates that decompensated cirrhosis—long thought to irreversible—can be clinically and histologically reversed. The key is combining transjugular intrahepatic portosystemic shunt to alleviate portal hypertension with sustained antiviral therapy to control the underlying etiology.

INTRODUCTION

Decompensated cirrhosis represents the end stage of chronic liver disease and is defined by severe complications of portal hypertension and hepatic dysfunction, including ascites, gastroesophageal variceal bleeding, hepatic encephalopathy, and jaundice. The transition from compensated to decompensated cirrhosis is a pivotal prognostic inflection point: Median survival declines from more than 10 years to approximately 2 years after the first decompensating event[1,2]. Contemporary care largely emphasizes prevention and treatment of complications, whereas therapies that meaningfully reverse the underlying pathophysiology remain limited[3].

Portal hypertension is the core hemodynamic abnormality in decompensated cirrhosis. It drives variceal formation and rupture, promotes ascites, and contributes to a hyperdynamic circulatory state that can precipitate end-organ dysfunction[4]. Transjugular intrahepatic portosystemic shunt (TIPS) is an established therapy for refractory portal hypertensive complications, particularly recurrent variceal bleeding, and difficult-to-control ascites[5,6]. Although traditionally regarded as a purely mechanical decompressive intervention, accumulating evidence suggests that TIPS may also exert systemic effects such as reducing gut bacterial translocation, attenuating systemic inflammation, and potentially disrupting the reciprocal amplification between inflammation and portal hypertension in advanced liver disease[7].

Here, we describe 2 patients with virus-related decompensated cirrhosis treated with TIPS and sustained antiviral therapy who achieved unusually favorable long-term outcomes. Over more than a decade of follow-up, neither patient experienced recurrent decompensation. Remarkably, in 2025, both patients demonstrated normal portosystemic pressure gradients (PPG) despite radiologically confirmed TIPS occlusion, together with histological evidence of fibrosis regression and loss of cirrhotic architectural features. The Baveno VII consensus defines “recompensation” as the sustained absence of decompensating events for at least 1 year after etiologic control, accompanied by stable improvement in liver function[5]. Both patients not only met these criteria but also demonstrated normalization of portal pressure and marked improvement in liver architecture. Collectively, these observations challenge the traditional view of decompensated cirrhosis as uniformly irreversible and suggest that, in carefully selected patients, effective hemodynamic control combined with durable etiologic suppression may support not only clinical recompensation but also structural hepatic recovery.

CASE PRESENTATION

Chief complaints

Case 1: A 43-year-old man was admitted in April 2012 with recurrent hematemesis and melena for 2 months.

Case 2: A 50-year-old man was admitted in July 2015 with hematemesis and melena for 6 days.

History of present illness

Case 1: In February 2012 (2 months before admission), the patient experienced his first episode of hematemesis without an obvious trigger. He vomited approximately 500 mL of bright red blood with clots, accompanied by dark red bloody stools totaling about 300 mL. He also reported dizziness, fatigue, and palpitations. He was admitted to a local hospital, where gastroscopy showed severe esophageal varices and endoscopic variceal ligation was performed. He also received portal pressure-lowering therapy, blood transfusions, and supportive care. The bleeding ceased and he was discharged. One month later, he developed recurrent small-volume hematemesis and persistent melena. He was transferred to our hospital in April 2012 for further management. On admission, he reported mild fatigue and abdominal distension, without abdominal pain, fever, or other symptoms.

Case 2: Six days before admission, the patient developed hematemesis and tarry stools. He received conservative treatment at another hospital for presumed variceal bleeding, but hemorrhage persisted. He was transferred to our hospital for further evaluation and intervention on the day of admission.

History of past illness

Case 1: The patient had chronic hepatitis B for more than 10 years and had not received regular antiviral therapy. He denied hypertension, diabetes, coronary heart disease, and other chronic illnesses. There was no history of surgery, trauma, or blood transfusion.

Case 2: The patient had chronic liver disease. Serology indicated prior hepatitis B virus (HBV) exposure, and he had active hepatitis C virus (HCV) infection.

Personal and family history

Case 1: His mother is a HBV carrier; his father is healthy. There was no family history of hereditary disease.

Case 2: He denied a family history of liver disease.

Physical examination

Case 1: Temperature 36.8 °C; pulse 88 beats/minute; respiratory rate 18 breaths/minute; blood pressure 112/74 mmHg. The patient was alert and cooperative. Mild jaundice of the skin and sclera was present. Several spider angiomas were noted on the anterior chest, and palmar erythema was present. The abdomen was flat, without abdominal wall venous collaterals. It was soft, with no tenderness or rebound tenderness. The liver was not palpable below the costal margin. The spleen was enlarged, palpable approximately three fingerbreadths (about 5 cm) below the left costal margin; it was firm, smooth, and non-tender. Shifting dullness was equivocal. Bowel sounds were 4/minute. There was no lower-extremity edema.

Case 2: On admission, the patient was alert but appeared pale and fatigued consistent with acute blood loss. Abdominal examination revealed splenomegaly palpable below the left costal margin. No other notable findings were documented.

Laboratory examinations

Case 1: Complete blood count: White blood cells 3.2 × 10⁹/L; hemoglobin 78 g/L; platelets 65 × 10⁹/L. Stool testing: Tarry stool; fecal occult blood (+++). Liver biochemistry: Alanine aminotransferase 68 U/L; aspartate aminotransferase 75 U/L; total bilirubin 32 μmol/L; albumin 31 g/L. Virology: Hepatitis B surface antigen (+), hepatitis B e-antibody (-), HBV DNA 1.5 × 10⁵ IU/mL. Antibodies to HAV, HCV, and HEV were negative. Child-Pugh score: 7 (class B).

Case 2: Complete blood count: Hemoglobin 65 g/L; platelets 78 × 10⁹/L. Stool testing: Melena present; fecal occult blood strongly positive. Liver biochemistry: Albumin 28 g/L. Virology: Serology consistent with prior HBV exposure; HCV RNA 5.96 × 10⁴ IU/mL, confirming active HCV infection. Child-Pugh score: 8 (class B).

Imaging examinations

Case 1: Ultrasound: Abnormal liver morphology with irregular surface and coarse, heterogeneous echotexture; portal vein diameter about 1.5 cm; marked splenomegaly (length about 16.6 cm, thickness about 5.6 cm); small amount of free fluid. Contrast-enhanced computed tomography (CT): Cirrhosis and portal hypertension; marked splenomegaly; esophageal and gastric varices; small-volume ascites.

Case 2: Ultrasound: Reduced liver volume with irregular surface; disproportionate lobar volumes; coarse, heterogeneous parenchymal echotexture; portal vein diameter 1.4 cm; splenomegaly (length 14.4 cm, thickness 5.0 cm); small-volume ascites. Contrast-enhanced CT: Cirrhosis, portal hypertension, esophageal and gastric varices, splenomegaly, and small-volume ascites.

FINAL DIAGNOSIS

Case 1

Decompensated hepatitis B-related cirrhosis with esophagogastric variceal bleeding.

Case 2

Decompensated hepatitis C-related cirrhosis with esophagogastric variceal bleeding.

TREATMENT

Case 1

In April 2012, a TIPS was successfully created using an 8 mm × 80 mm Fluency^® covered stent. The portal pressure gradient (PPG) decreased from 30 mmHg pre-procedure to 9 mmHg post-procedure. Long-term nucleos(t)ide analogue therapy with entecavir was initiated on postoperative day 2; follow-up assessments indicated good adherence and sustained viral suppression. The postoperative course was uneventful, with no recurrent bleeding.

Case 2

On July 10, 2015, TIPS was successfully performed with placement of an 8 mm × 80 mm Fluency^® covered stent, reducing PPG from 26 mmHg to 12 mmHg, and the patient stabilized. In December 2015, he was readmitted for recurrent gastrointestinal bleeding. Angiography demonstrated occlusion of the TIPS stent. On December 15, 2015, TIPS revision was performed: A 10 mm × 100 mm Fluency^® stent was deployed within the occluded stent, followed by balloon dilation. Patency was restored and PPG decreased from 25 mmHg to 9 mmHg. The patient did not receive direct-acting antiviral therapy; instead, he achieved spontaneous viral clearance, with persistently undetectable HCV RNA on follow-up.

OUTCOME AND FOLLOW-UP

A baseline liver biopsy was not performed for either patient for two reasons: First, both patients presented with clinical emergencies; and second, according to the prevailing clinical guidelines at the time, the diagnoses of decompensated viral hepatitis-related cirrhosis were adequately supported by their clinical histories, imaging findings consistent with cirrhosis, and complications of portal hypertension, including esophagogastric variceal bleeding. Therefore, baseline histological assessment was not considered necessary in these clinical contexts. All biopsy specimens were independently re-reviewed by two senior hepatopathologists blinded to clinical information; discrepancies were resolved by consensus to minimize assessment bias.

Both patients were followed for more than 10 years. Neither developed recurrent variceal bleeding, clinically significant ascites, hepatic hydrothorax, or hepatic encephalopathy. Liver synthetic function remained stable and improved over time; both patients improved from Child-Pugh class B (scores 7 and 8) to class A (score 5). Viral replication became undetectable, and spleen size decreased substantially in both patients (Table 1).

Table 1 Comparison of study participants’ variables at baseline and final follow-up.

Variables	Patient 1		Patient 2
	Base line	Final follow-up	Base line	Final follow-up
Virus quantitative	HBV DNA 1.5 × 105 IU/mL	Negative	HCV RNA 5 × 104 IU/mL	Negative
Ascites	Small amount	None	Small amount	None
Hepatic hydrothorax	None	None	Small amount in both sides	None
Child-Pugh score	7	5	8	5
Size of the spleen	5.6 cm × 16.6 cm	3.3 cm × 12.7 cm	5.0 cm × 14.4 cm	4.0 cm × 10.7 cm
Esophageal and gastric varices	Severe	None	Severe	None
PPG	30 mmHg	5 mmHg	26 mmHg	6 mmHg

HBV DNA: Hepatitis B virus DNA; HCV RNA: Hepatitis C virus RNA; PPG: Portosystemic pressure gradient.

At the 10-year follow-up, surveillance imaging (ultrasound and CT) demonstrated TIPS occlusion in both patients. Despite this finding, direct catheter-based pressure measurements through the tract showed normal PPG values (5 mmHg in patient 1 and 6 mmHg in patient 2). Angiography revealed no gastroesophageal varices or portosystemic collateral vessels. Liver biopsies showed marked histologic regression, with fibrosis improving to Ishak stage 3 in both patients (Figure 1). Importantly, biopsies no longer displayed architectural hallmarks of cirrhosis, such as pseudolobule formation (Figure 2), consistent with substantial hepatic remodeling.

Figure 1 Patient follow-up and evaluation timeline. Patient 1: Underwent transjugular intrahepatic portosystemic shunt (TIPS) placement combined with antiviral therapy in April 2012, followed by regular follow-up and imaging surveillance. In June 2025, imaging revealed stent occlusion, prompting angiography with portal pressure gradient (PPG) measurement and liver biopsy. Patient 2: Underwent TIPS placement in July 2015 and required TIPS revision in December 2015 due to recurrent bleeding. The patient then continued regular follow-up and imaging surveillance. In July 2025, imaging demonstrated stent occlusion, leading to angiography with portal pressure gradient measurement and liver biopsy.

Figure 2 Imaging and histopathology illustrate portal pressure changes and hepatic fibrosis regression after transjugular intrahepatic portosystemic shunt. A: Baseline contrast-enhanced computed tomography showing cirrhosis with a small-volume ascites (yellow arrow) and esophagogastric varices (blue arrow); B: Final follow-up computed tomography demonstrating resolution of ascites and varices despite occlusion of the transjugular intrahepatic portosystemic shunt (TIPS) stent; C: Direct portography performed 6 months after TIPS for recurrent gastrointestinal bleeding showing stent occlusion with recurrent varices; portal pressure gradient was 25 mmHg, and TIPS revision was performed; D: Direct portography 10 years after TIPS showing stent occlusion without varices; portal pressure gradient was 6 mmHg, and no intervention was undertaken; E: Hematoxylin and eosin-stained liver tissue shows hepatocytes arranged in cords with preserved sinusoidal architecture. Interlobular septa are maintained with residual fibrous septa; there are no histologic features of cirrhosis or chronic inflammation; F: Masson’s trichrome staining reveals branched fibrous septa with expansion of portal fibrous tissue, without definite nodular remodeling or classic cirrhotic features. The fibrosis stage is Ishak 3.

DISCUSSION

These two patients with virus-related decompensated cirrhosis experienced unexpectedly favorable long-term outcomes after TIPS placement combined with sustained antiviral therapy. Both remained clinically stable for more than a decade without further decompensation. Reassessment in 2025 revealed normalized PPG despite TIPS occlusion, accompanied by histologic findings consistent with significant fibrosis regression and architectural improvement. While causality cannot be established from two cases, these observations support an emerging concept: In carefully selected patients, durable etiological control combined with effective portal decompression may enable sustained physiological improvement and structural liver remodeling.

Decompensated cirrhosis has long been considered an inexorably progressive condition. The first decompensating event signals a major prognostic shift, and clinical management typically focuses on prevention of complications and timely transplant evaluation[3]. More recently, the concept of “recompensation” has gained traction, particularly among patients who achieve etiological cure or long-term suppression. Our cases extend this paradigm by suggesting that recompensation may, in rare circumstances, coincide with hemodynamic normalization and histological regression even after decompensation. Future research should prioritize identifying and validating clinical and biological predictors of cirrhosis regression. Based on prior literature and the characteristics of our case series, it is plausible that a subset of patients with decompensated cirrhosis may be more likely to benefit when the underlying etiology is clearly defined and amenable to durable control (e.g., viral hepatitis), hepatic functional reserve is relatively preserved (Child-Pugh class A or B), and patients are younger (≤ 50 years).

In the current standard clinical pathway for virus-related decompensated cirrhosis, management primarily centers on: (1) Timely initiation of etiological therapy; (2) Prevention and treatment of portal hypertension-related complications; and (3) Appropriate initiation of liver transplant evaluation after the first decompensating event. Within this framework, TIPS is generally reserved for established indications (e.g., refractory variceal bleeding or refractory ascites) rather than being used with the intent of long-term disease modification. Our observations raise the hypothesis that, in carefully selected patients with durable etiologic control and sufficient hepatic reserve, earlier portal decompression with TIPS in conjunction with etiological therapy may facilitate sustained clinical stabilization and potentially delay - though not replace - transplantation. This combined approach should therefore be viewed as a potential adjunct to existing pathways, to be validated in larger prospective studies.

Portal hypertension arises from both structural and dynamic components. Structural contributors include fibrosis and vascular remodeling, whereas dynamic contributors include potentially reversible increases in intrahepatic vascular tone driven by endothelial dysfunction and disordered vasoregulation. TIPS is generally used as rescue therapy for refractory complications, and its benefit is commonly attributed to mechanical decompression. However, accumulating data suggests additional effects, including reduced bacterial translocation and attenuation of systemic inflammation[7]. The durable stability and normal PPG in our patients - despite shunt occlusion - raise the possibility that the long-term benefit extended beyond simple diversion of portal flow.

One plausible explanation is that an initial period of portal decompression interrupts a self-reinforcing cycle of sinusoidal hypertension, endothelial injury, and vascular remodeling. By lowering portal pressure, TIPS may reduce microvascular shear stress and inflammatory activation, facilitating recovery of endothelial function and improved vasoregulatory balance[8]. Over time, this could reduce intrahepatic vascular resistance. If the primary injurious stimulus is simultaneously removed, hemodynamic gains achieved during shunt patency might persist even after occlusion. This hypothesis is consistent with the recognized role of dynamic components in portal pressure regulation, particularly when inflammatory drivers are controlled[9].

Sustained antiviral therapy likely contributed substantially. Viral suppression reduces ongoing hepatocyte injury and necroinflammation, thereby limiting stellate cell activation and extracellular matrix deposition. Although fibrosis regression after etiological control is well documented in chronic viral hepatitis, marked regression following decompensation is reported less frequently. In these cases, prolonged etiological control and sustained clinical stability may have enabled continued matrix remodeling and partial restoration of hepatic architecture.

The gut-liver axis may also be relevant. Portal hypertension promotes intestinal congestion and increased permeability, facilitating bacterial translocation and systemic inflammation. A period of portal decompression may improve mucosal barrier function and reduce the translocation of pathogen-associated molecular patterns, thereby decreasing inflammatory signaling that can perpetuate fibrogenesis. In conjunction with durable etiological control, reduced inflammation could support microvascular and extracellular matrix remodeling, consistent with contemporary models that position inflammation as an active driver of cirrhosis progression[10].

From a clinical standpoint, these cases underscore that radiologic TIPS occlusion does not necessarily equate to recurrent clinically significant portal hypertension. Follow-up often prioritizes shunt patency assessment by Doppler ultrasound or CT, and occlusion commonly triggers consideration of revision. However, in asymptomatic patients with sustained stability and controlled etiology, management may be better guided by physiological assessment - such as repeat PPG measurement - integrated with symptoms, laboratory trends, and evidence of portal hypertensive complications. A hemodynamics-informed strategy could help avoid unnecessary procedures and their attendant risks. Nonetheless, these cases should not be interpreted as grounds to ignore an occluded shunt when patients exhibit recurrent bleeding, refractory ascites, or objective evidence of portal hypertension.

The final liver biopsy demonstrated Ishak stage 3 fibrosis, which is consistent with substantial regression of cirrhotic remodeling. However, Ishak stage 3 still denotes persistent fibrous septa and therefore does not indicate complete restoration of normal liver architecture. In addition, the spatial heterogeneity of hepatic fibrosis and the inherent limitations of biopsy sampling should be acknowledged, as a single-site specimen may not fully capture the liver’s overall fibrotic burden. Accordingly, we interpret this finding as an important and encouraging indicator rather than definitive evidence of complete architectural normalization. Greater weight is therefore placed on the concordance between histological improvement and the patient’s sustained clinical remission, normalization of liver biochemical parameters, and marked resolution of portal hypertension-related complications (e.g., no recurrent gastrointestinal bleeding). Taken together, this multimodal assessment provides more robust support for meaningful fibrosis regression and architectural improvement.

Several limitations warrant emphasis. This report includes only two patients and is vulnerable to selection bias, limiting generalizability. The lack of baseline liver histology prevents direct within-patient comparison of fibrosis stage over time. In addition, sampling variability and spatial heterogeneity of fibrosis constrain definitive conclusions regarding “reversal” of cirrhosis. Finally, the mechanistic explanations proposed here remain speculative and require systematic testing.

Future work should examine prospective cohorts receiving TIPS alongside durable etiological therapy, incorporating serial hemodynamic measurements, noninvasive fibrosis assessment (e.g., elastography), and paired histology where feasible. Translational studies focused on endothelial remodeling, stellate cell plasticity, inflammatory signaling, and gut-liver axis pathways may help define mechanisms of structural recovery and identify patients most likely to benefit from combined portal decompression and sustained etiological control.

CONCLUSION

This report indicates that, in appropriately selected patients, combining TIPS with effective etiologic management may enable sustained reversal of portal hypertension and support histologic regression of cirrhosis.