Portal hypertension (PHT) is defined by an elevated pressure gradient between the portal vein and the hepatic veins ≥ 5 mm Hg, and is the main vector of complications in cirrhosis.

When the hepatic venous pressure gradient (HVPG) is ≥ 10 mm Hg, it is considered as a " clinically significant PHT ": ascites and oesophageal varices (EV) may occur.

Above 12 mm Hg, there is a risk of variceal bleeding. Carvedilol, a non-selective beta-blocker (NSBB), is recommended in all the patients with cirrhosis and clinically significant PHT in order to prevent decompensation of cirrhosis.

Nevertheless, 40 % of patients are NSBB non-responders, i.e. they do not show a significant decrease in HVPG. In addition, NSBB responders treated for primary prophylaxis have an incidence of variceal bleeding of approximately 10% per year, with a six-week mortality of 20%. Therefore, there is an unmet need for PHT in patients with cirrhosis who do not respond to NSBB, and also for an increase in efficacy in responders. In a randomised pilot study, Rittig et al. observed a mean change in HVPG of -2,9 mm Hg in 16 patients with cirrhosis and HVPG ≥ 12 mm Hg, not treated with NSBB, 90 minutes after ingestion of 1000 mg metformin.

The study will be a prospective, national, multicentre, phase II, superiority comparative randomized (1:1) simple-blinded clinical trial with two parallel arms: metformin versus placebo.

The main objective is to evaluate the effect of metformin versus placebo during 28 days on HVPG, in patients with cirrhosis and a HVPG ≥ 12 mm Hg already treated with carvedilol.

Subjects randomized in the metformin group or placebo group will receive metformin ou placebo, one pill of 500 mg per os twice a day (one in the morning and one in the evening, during or at the end of the meal) for 28 days.

Portal hypertension (PHT) is defined by an elevated pressure gradient between the portal vein and the hepatic veins ≥ 5 mm Hg, and is the main vector of complications in cirrhosis.

When the hepatic venous pressure gradient (HVPG) is ≥ 10 mm Hg, it is considered as a " clinically significant PHT ": ascites and oesophageal varices (EV) may occur.

Above 12 mm Hg, there is a risk of variceal bleeding. Carvedilol, a non-selective beta-blocker (NSBB), is recommended in all the patients with cirrhosis and clinically significant PHT in order to prevent decompensation of cirrhosis.

Nevertheless, 40 % of patients are NSBB non-responders, i.e. they do not show a significant decrease in HVPG. In addition, NSBB responders treated for primary prophylaxis have an incidence of variceal bleeding of approximately 10% per year, with a six-week mortality of 20%. Therefore, there is an unmet need for PHT in patients with cirrhosis who do not respond to NSBB, and also for an increase in efficacy in responders. In a randomised pilot study, Rittig et al. observed a mean change in HVPG of -2,9 mm Hg in 16 patients with cirrhosis and HVPG ≥ 12 mm Hg, not treated with NSBB, 90 minutes after ingestion of 1000 mg metformin.

The study will be a prospective, national, multicentre, phase II, superiority comparative randomized (1:1) simple-blinded clinical trial with two parallel arms: metformin versus placebo.

The main objective is to evaluate the effect of metformin versus placebo during 28 days on HVPG, in patients with cirrhosis and a HVPG ≥ 12 mm Hg already treated with carvedilol.

There are several secondary objectives in this research listed below:

• evaluate the safety and tolerability of metformin in patients with cirrhosis and a HVPG ≥ 12 mm Hg
• evaluate the rate of change in HVPG after 28 days of treatment
• evaluate the rate of patients with a clinically significant improvement in HVPG
• assess the change in systemic haemodynamics after 28 days of treatment
• assess the change in liver steatosis after 28 days of treatment
• assess the performance of liver and spleen stiffness by vibration-controlled transient elastography (VCTE) using FibroScan® (Echosens, Paris, France) to estimate the haemodynamic response to the treatment
• assess the performance of liver and spleen stiffness by 2D-shear wave elastography using Aixplorer® (SuperSonic Imagine, Aix-en-Provence, France) to estimate the haemodynamic response to the treatment
• assess the effect of metformin on systemic inflammation, coagulation, hepatocyte stress, and endothelial function.

Subjects randomized in the metformin group or placebo group will receive metformin ou placebo, one pill of 500 mg per os twice a day (one in the morning and one in the evening, during or at the end of the meal) for 28 days.

Inclusion Criteria:

• - Age ≥ 18 years

• Written informed consent to participate in the study

• Medical insurance coverage

• For child-bearing aged women, contraception using oestroprogestative, progestative, intrauterine device, or mechanical contraception

• Diagnosis of cirrhosis based on a liver biopsy, or on clinical, biological, endoscopic, and radiological evidence

• Active cause of cirrhosis, or resolution (alcohol cessation, sustained virological response to direct-acting antiviral treatment for HCV, initiation of nucleoside/nucleotide analog treatment for HBV) for at least 6 months

• Child-Pugh A or B

• High likelihood of HVPG ≥ 12 mm Hg based on investigator's judgement, or on the following criteria:

  1. Investigator's judgement

  2. active cause of cirrhosis and:

     • History of clinical ascites
     • Or history of variceal bleeding
     • Or liver stiffness by VCTE ≥ 35 kPa on carvedilol in the last two years
     • or spleen stiffness by VCTE ≥ 55 kPa on carvedilol in the last two years
     • or liver surface nodularity ≥ 2,9 in the last two years
     • or HVPG > 16 mm Hg prior to starting NSBB
     • or Laennec 4c cirrhosis on histology

  3. or resolution of the cause of cirrhosis for at least 6 months and:

     • history of clinical ascites in the last 6 months
     • or history of variceal bleeding in the last 6 months
     • or liver stiffness by VCTE ≥ 35 kPa on carvedilol in the last 6 months
     • or spleen stiffness by VCTE ≥ 55 kPa on carvedilol in the last 6 months
     • or liver surface nodularity ≥ 2,9 in the last 12 months
     • or Laennec 4c cirrhosis on histology in the last 12 months

• Treatment with carvedilol (≥ 6,25 mg/day) at a stable dose for at least one month

• Absence of hepatocellular carcinoma outside at least one nodule > 3 cm in diameter, or more than 3 nodules, on ultrasound, CT-scan or MRI performed during the previous 6 months

Exclusion Criteria:

• Serum total bilirubin > 50 µmol/L
• Prothrombin ratio < 50 %
• Transaminases > 5 ULN
• Need for at least one paracentesis for ascites fluid evacuation in the last 6 months
• Expected follow-up < 3 months
• Known hypersensitivity to the active substance or any of the excipients
• History of lactic acidosis, diabetic acidocetosis, or diabetic precoma
• Ongoing condition that may lead to acute kidney injury or hypoxia: dehydration, severe infection, shock, cardiac decompensation, respiratory failure, or myocardial infarction within the past month
• Known hypersensitivity to all the iodin-containing contrast agents
• Known hypersensitivity to lidocaine for local anesthesia
• Known hypersensitivity to beta-lactam antibiotics if the patient has a history of valve replacement
• Alcohol consumption > 14 units/week for women or > 21 units/week for men, current or abstinent for less than 6 months
• Biliary cirrhosis
• Hepatocellular carcinoma with at least one nodule > 3 cm in diameter, or more than 3 nodules
• Cholangiocarcinoma
• Extra-hepatic cancer without remission
• Severe chronic kidney disease defined as estimated glomerular filtration rate < 30 mL/min/1,73m2 using the MDRD-6 formula
• Ongoing treatment with metformin, or discontinued for less than 3 months
• Treatment with statins started or discontinued for less than 3 months
• Treatment with nucleoside/nucleotide analogue for HBV, or direct-acting antiviral treatment for HCV, started for less than 6 months
• Complete portal vein thrombosis (main portal trunk, or right branch), or portal cavernoma
• History of TIPS (transjugular intrahepatic portosystemic shunt) / surgical portosystemic derivation / liver transplantation / major hepatectomy
• Ongoing participation in another interventional therapeutic trial
• Pregnant or breastfeeding women
• Patients unable to give consent (under guardianship or curatorship)
• Non-randomisation criteria: HVPG < 12 mm Hg at the catheterism performed during the first follow-up visit