This multicentre, prospective cohort study aims to enrol 375 patients with cirrhosis who meet the following criteria:

1. effective management or elimination of the primary aetiological factor (i.e., sustained virological response in chronic hepatitis C, complete viral suppression in chronic hepatitis B, or long-term alcohol abstinence in alcohol-related liver disease),
2. liver stiffness measurements under 25 kPa, and
3. absence of varices as confirmed by endoscopy

Following the cessation of non-selective beta-blockers, patients will undergo a follow-up endoscopy at the one-year mark. The study's primary endpoint is recurrent varices in 1 year. Success will be defined as the upper bound of the 95% confidence interval for recurrent varices being <5%. Should this criterion be met, the study will extend to predefined analyses of variceal haemorrhage and hepatic decompensation at 3-and 10-year intervals (funded through local resources; the General Research Fund will cover patient recruitment costs for 1 year). Clinical assessments, laboratory tests, liver and spleen stiffness measurements will be performed at baseline, 3 months, 6 months and 12 months to identify potential predictors of variceal recurrence and assess the feasibility of early identification.

1.1 Significance of cirrhosis and portal hypertension

Liver cirrhosis is the common final pathway of various chronic liver diseases. Globally, mortality due to cirrhosis have risen from 899 000 in 1990 to 1.32 million deaths in 2017. Additionally, as of 2017, 112 million and 10.6 million people were living with compensated and decompensated cirrhosis, respectively. Portal hypertension stands out as the primary cause of cirrhotic complications, such as ascites and variceal haemorrhage. Accurate diagnosis of clinically significant portal hypertension is crucial because the administration of non-selective beta-blockers (NSBBs) can reduce the risk of variceal haemorrhage and hepatic decompensation.

Chronic liver disease is highly prevalent in Hong Kong. A community screening study in Hong Kong in 2015-2016 indicated that the prevalence of hepatitis B surface antigen was still 7.8%. Studies by our group have further highlighted the extent of the problem, revealing that 26% of the local population and a staggering 73% of patients with type 2 diabetes suffer from metabolic dysfunction-associated steatotic liver disease. These chronic liver diseases carry a significant risk of progressing to cirrhosis and hepatic decompensation.

1.2 Non-invasive assessment of portal hypertension

Historically, the gold standard for assessing portal hypertension requires the measurement of hepatic venous pressure gradient, an invasive procedure rarely performed outside research settings. However, accumulating data support the use of non-invasive tests to assess portal hypertension. Specifically, liver stiffness measurement (LSM) through vibration-controlled transient elastography reflects the degree of liver fibrosis and the severity of cirrhosis. Furthermore, the platelet count reflects whether a patient with cirrhosis has hypersplenism secondary to portal hypertension. Additionally, spleen stiffness measurement (SSM) may be an even more direct evaluation of portal hypertension and has been used to ascertain improvements in portal hypertension following the initiation of NSBB. Based on existing data, the latest Baveno VII consensus has endorsed the use of these non-invasive tests to rule out (if LSM <15 kPa and platelet count >150×10e9/L) and to confirm (if LSM ≥25 kPa) the presence of clinically significant portal hypertension.

In a randomised controlled trial involving 548 patients with radiological cirrhosis, our group demonstrated that LSM+SSM was non-inferior to routine endoscopy in detecting varices and preventing future variceal haemorrhage. Furthermore, we validated the use of LSM-based criteria for excluding the presence of varices requiring treatment across a range of chronic liver diseases, including patients with or without hepatocellular carcinoma. Additionally, we reported the natural history of patients falling within the grey zone according to the Baveno VII criteria. Importantly, we identified factors associated with hepatic decompensation in this specific patient population.

1.3 Emerging concept: hepatic recompensation and reversal of cirrhosis and portal hypertension

As a number of chronic liver diseases can now be well controlled or cured, it is increasingly apparent that cirrhosis can resolve over time when the primary aetiology is controlled. Reports have also emerged, suggesting an improvement in portal hypertension and even hepatic recompensation in such circumstances. This has led to the practical recommendation on discontinuing NSBB in patients with LSM <25 kPa after removal or suppression of the primary aetiological factor, in the absence of varices. Removal or suppression of the primary aetiological factor includes sustained virological response in chronic hepatitis C, complete hepatitis B virus DNA suppression in chronic hepatitis B patients, and long-term abstinence from alcohol in those with alcohol-related liver disease. Nonetheless, it is crucial to note that these recommendations are based on tenuous evidence (C.2 recommendation). While cessation of NSBB in patients no longer requiring it can reduce unnecessary expenses (prescription and monitoring) and side effects, it remains imperative to confirm the safety of this approach through well-designed, prospective studies.

Inclusion Criteria:

• Aged 18 years or above
• Evidence of cirrhosis, based on either radiological and/or clinical features
• History of varices, variceal haemorrhage or portal hypertension warranting NSBB
• Removal or suppression of the primary aetiological factor (i.e., sustained virological response in chronic hepatitis C, complete hepatitis B virus DNA suppression in chronic hepatitis B, and long-term alcohol abstinence in alcohol-related liver disease)
• LSM <25 kPa
• Provision of written informed consent

Exclusion Criteria:

• Active aetiological factors not addressed in the inclusion criteria (e.g., autoimmune or hereditary liver diseases). However, hepatic steatosis-commonly coexisting with other liver diseases-is not an exclusion criterion unless accompanied by high plasma alanine aminotransferase >40 U/L or confirmed steatohepatitis.
• Recent hepatic decompensation within the past year (total bilirubin >50 μmol/L, prothrombin time >1.3 times the upper normal limit [unless attributable to use of anti-coagulation], albumin <35 g/L, or presence of ascites, variceal haemorrhage or hepatic encephalopathy). Notably, the Baveno VI criteria used LSM <20-25 kPa and normal platelet count >150×109/L to rule out varices requiring treatment (19). However, in line with the Baveno VII consensus, which sets an LSM <25 kPa as the threshold for considering NSBB discontinuation, thrombocytopenia is not an exclusion criterion to allow the cohort to include a wider spectrum of patients.
• Current or history of hepatocellular carcinoma.
• Radiological evidence of portal vein thrombosis.
• History of other malignancies (unless in complete remission for >5 years).
• History of liver transplantation or liver resection.
• Contraindications to undergoing endoscopy.
• Other clinical indications for NSBB (e.g., cardiovascular disease, arterial hypertension)

Eligible patients will be identified from the general medical and hepatology clinics in Hong Kong and collaborating countries.

Data may be shared upon reasonable request.

6 months after the first publication until 15 years after the end of the study

By email communication