Chronic liver diseases related to viral hepatitis, metabolic syndrome or excessive alcohol consumption can evolve towards cirrhosis. Cirrhosis is responsible for 170 000 deaths per year in Europe. Initially asymptomatic and called "compensated" it can become "decompensated" with the developement of acute complications such as infections, ascites or variceal bleeding. The transition from compensated to decompensated cirrhosis is associated with a reduction in survival from 95 to 55% at 1 year.

The only curative treatment for cirrhosis is liver transplantation (LT). Liver transplants are allocated according to the severity of the patients. Despite a modest prognostic value (area under the ROC curve = 0.7 to predict the risk of death), graft allocation is based on the MELD (Model for End-Stage Liver Disease) score including INR, bilirubin and serum creatinine. In 2014, 11.5% of registered patients died on the liver transplant waiting list, illustrating the need for biomarkers that predict death and improve MELD-based prediction.

Microvesicles are membrane vesicles released in extracellular space during cell activation or apoptosis. Our team showed that circulating levels of hepatocyte microvesicles increase with the severity of cirrhosis and predict survival at 6 months independently of MELD score in a cohort of 242 patients with cirrhosis.

Type 1 interferons (IFN-1) are mediators of inflammation, which is excessively activated in cirrhosis. Our team has shown that a gene signature (IFN score) measured in the immune cells of 101 patients with cirrhosis is able to predict 6 month-survival independently of the MELD score.

Thus, the investigators hypothesize that a composite score combining the level of circulating hepatocyte microvesicles, the IFN score and the MELD score could improve the prediction of survival in patients with severe cirrhosis.

The aim of this study is to compare the prognostic performance for the cumulative incidence of death at 6 months of a composite score including MELD, hepatocyte microvesicle level and IFN score with that of the MELD score alone, in patients with Child B or C cirrhosis, considering liver transplantation as a competitive risk.

To address this question, peripheral blood from 335 patients with Child B or C cirrhosis will be obtained and hepatocyte microvesicle levels and IFN score will be measured using ELISA/filtration and Real Time-quantitative PCR.

Natural history of cirrhosis

Cirrhosis is the end stage form of chronic liver disease. It is estimated that 200,000 to 500,000 people in France have cirrhosis. Cirrhosis is responsible for more than 170,000 deaths per year in Europe.

The main causes of cirrhosis include excessive alcohol consumption, the leading cause in Europe, and chronic viral hepatitis B and C, which are the leading causes in Asia and Africa. In addition, the current epidemic of obesity and type 2 diabetes worldwide has led to a sharp increase in the incidence of non-alcoholic steatohepatitis associated with metabolic syndrome, particularly in North America.

Cirrhosis is defined histologically as a diffuse alteration in the architecture of the liver by annular fibrosis associated with regeneration nodules. With the emergence of non-invasive tests allowing fibrosis measurement, cirrhosis is now considered highly probable when the liver elasticity measured by Fibroscan® is > 15 kPa.

Patients with cirrhosis may be perfectly asymptomatic and have normal liver function. In this case, the Child-Pugh score, which rates the severity of cirrhosis, is "A". However, the severity of cirrhosis may increase, resulting in the development of signs of liver failure, such as jaundice, decreased coagulation factors and hypoalbuminemia, and/or ascites or hepatic encephalopathy. Cirrhosis is then classified as Child-Pugh B or C.

The progression of cirrhosis involves various processes: on the one hand, a slow and continuous worsening associated with the persistence of hepatic aggression by one or more causal agents (alcohol and/or metabolic syndrome and/or virus) and on the other hand, rapid worsening occurring with acute complications of the disease such as variceal bleeding or bacterial infections. Indeed, cirrhosis is associated with major alterations in innate immunity that promote severe bacterial infections, which can lead to organ failure. Such a complication represents a turning point in the natural history of cirrhosis since the transition from compensated to decompensated cirrhosis is associated with a reduction in 1-year survival from 95 to 60%. When these complications are associated with organ failure, short term mortality is dramatically increased with one-third of patients dying within one month. The association of acute decompensation of cirrhosis with one or more organ failures has recently been defined as a new syndrome named Acute-on-Chronic Liver Failure (ACLF).

The only curative treatment for cirrhosis is liver transplantation (LT). However, each year, between 10 and 15% of patients on the French liver transplant waiting list cannot access transplant and die prematurely from an acute complication of cirrhosis. In addition, many patients die before they are even put on the transplant waiting list, as evidenced by the results of a recent French multicenter study showing that the 6-month mortality rate for patients with decompensated cirrhosis, Child C, remains equal to 25%[8].

Predicting of the outcome is therefore a major challenge in the management of cirrhotic patients.

Prediction of the outcome of patients with cirrhosis: clinical scores.

Currently, prediction of the outcome of cirrhotic patients is based on the MELD (Model for End-Stage Liver Disease) score. MELD score includes three objective laboratory criteria (INR, serum bilirubin and serum creatinine). It was established in 2000 in an American cohort to predict the risk of death at 3 months after TIPS placement (Transjugular Intrahepatic Portosystemic Shunt). Its use was very quickly extended to the prediction 3-month survival of cirrhotic patients on the transplant waiting list. Since 2007, in France, graft allocation has been based on the MELD score, with patients with the highest score being transplanted first.

Since that time, the predictive value of MELD has been studied in large cohorts, showing that MELD's performance in predicting death is modest, with areas under the ROC curves (AUROC) ranging from 0.64 to 0.72[3, 7, 19, 20]. In 2014, 7 years after the introduction of the MELD score, it was found that 11.5% of patients on the liver transplant waiting list still died before they could be transplanted. These data illustrate the need to improve our ability to predict death of patients with advanced cirrhosis.

In this perspective, some authors have tried to improve the MELD score by adding one or more variables, such as serum sodium (MELD-Na=MELD-Na-[0.025×MELD× (140 - Na)] + 140). The contribution of these modified MELD scores was not considered sufficient to promote their use in clinical practice, leaving the objective of improving MELD unmet.

Prediction of the outcome of patients with cirrhosis: novel biomarkers

Beside clinical scores, some biomarkers look promising. Indeed, as part of the close collaboration between the clinical departments of Beaujon Hospital (Hepatology and Anesthesia-Resuscitation) and the Inserm teams (Dr E Weiss, Dr R Moreau, Inserm U1149, Centre de recherche sur l'inflammation, Paris) and Pr PE Rautou, Dr C Boulanger, Inserm U970 ; Centre de recherche cardiovasculaire, Paris), two types of biomarkers have recently been identified that predict the survival of patients with cirrhosis.

Microvesicles are vesicles released into the extracellular medium during cell activation or apoptosis, two phenomena strongly involved in cirrhosis. The Inserm U970 team (Pr PE Rautou, Dr C Boulanger) showed in 2012 that the circulating concentrations of hepatocyte microvesicles increase with the severity of cirrhosis. In addition, microvesicles in patients with child B or C cirrhosis promote portal hypertension by inducing arterial hyporeactivity to vasoconstrictive agents. Based on these results, a prospective study including 242 patients with cirrhosis showed that circulating concentrations of hepatocyte microvesicles are able to predict the survival at 6 months of patients with cirrhosis independently of the MELD score.

The susceptibility of cirrhotic patients to severe infections is related to abnormalities in their innate immune response and in particular to the occurrence of an excessive inflammatory response, responsible for organ failure. Type 1 interferons (IFN 1) are inflammatory mediators whose impact on the prognosis of sepsis could be harmful. To determine whether type 1 IFN pathway activity was involved in the innate immune response deregulation and in the susceptibility to sepsis of patients with cirrhosis, the Inserm U1149 team measured, using Real-Time quantitative PCR, a gene signature (IFN score) reflecting the activation of this pathway in the immune cells of 99 patients with decompensated cirrhosis hospitalized at the liver department at Beaujon Hospital. The investigators have thus shown that the IFN score is able to predict the risk of death of cirrhotic patients independently of the MELD score with AUROC equal to 0.79 (0.7 for MELD).

It should be noted that these two biomarkers (hepatocyte microvesicles and IFN score) reflect different and complementary pathophysiological processes: hepatocyte microvesicles reflect persistent liver injury while IFN score reflects innate immune response. It should also be noted that these biomarkers, protected by two international patents (EP121528133.5 and EP2016/077129) can be measured on peripheral blood using simple and readily available techniques (ELISA/filtration, qPCR).

Hypotheses

In this project, the investigators hypothesize that prediction of death of patients with advanced cirrhosis (Child-Pugh B or C) by a composite score combining circulating concentrations of hepatocyte microvesicles, IFN score and MELD score will be better than that assessed using the MELD score alone.

This composite score will help physicians to judge the emergency of liver transplantation and thus to refine criteria for the allocation of liver transplants. This composite score may also identify a patient population at particular risk of death that should be subjected to increased surveillance while waiting for liver transplantation. Through these two approaches (better prioritization of patients for liver transplantation and increased monitoring of high-risk patients), this new composite score will reduce mortality in cirrhosis patients waiting for liver transplantation.

Finally, new therapeutic approaches to improve survival of patients with severe cirrhosis awaiting liver transplantation (e. g. by reducing the occurrence of complications) are regularly tested. The composite score the investigators propose could also be used as a "companion" to monitor the effectiveness of these new treatments.

Population

Patients with Child-Pugh B or C cirrhosis (proven by liver biopsy, by non invasive method of fibrosis quantification or by a combination of clinical, biological and radiological signs, see above) will be included during a two-year period.

Patients will be included during follow-up outpatient visits in hepatology or during scheduled hospitalization for example for digestive endoscopy to screen signs of portal hypertension or for ascites removal. They will therefore be in a stable condition.

Presence of acute events or events that may modify the risk of death or liver transplantation, and/or the inflammatory response or release of microvesicles will be among non-inclusion criteria (see above).

Biomarker measurements

After informed consent, fifteen to twenty milliliters of peripheral venous blood (3 EDTA tubes and 3 citrate tubes) will be collected from patients participating in the study

For hepatocyte microvesicles, citrated tubes will be centrifuged twice and then frozen at -80°C until analysis.

For the IFN score, blood contained in the 3 EDTA tubes will be used to isolate Peripheral Blood Mononuclear Cells (PBMCs). PBMCs will be lysed to obtain RNA that will be stored at -80°C until the IFN score is measured.

At the end of the 2-year inclusion period, measurements of hepatocyte microvesicles and IFN score will be performed for all patients during a 1-year period. Plasma concentrations of hepatocyte microvesicles will be determined on platelet-free plasma using ELISA/filtration. To measure IFN score, RNA from previously isolated PBMCs will first be reverse transcribed. Then, the expression of five mRNAs of interest will be measure to calculate the IFN score.

Primary endpoint

The primary endpoint is the area under the ROC curve of the composite score including MELD, hepatocyte microvesicle concentration and IFN score for predicting the cumulative incidence of death at 6 months (considering liver transplantation as a competitive risk) compared to the area under the ROC curve of the MELD score alone.

Measurement of the primary endpoint will be performed according to the recommendations specific for survival analysis in patients with cirrhosis, using a survival analysis that considers liver transplantation as a competitive risk. Data from patients who did not die or have not been transplanted will be censored at the end of follow-up or on the date of last news (if the patient is lost to follow-up) or at the date of termination of participation (in the event of termination of participation). Follow-up period will be one year after inclusion (maximum 13 months for patients who have been transplanted in the 12th month after inclusion). A follow-up visit 6 and 12 months after inclusion will be scheduled. Data on possible deaths or transplants will be obtained during this visit, using medical records, or following telephone calls to patients or their relatives. For subjects who are lost to follow-up, requests will be made to the CépiDc to obtain their vital status. Data concerning possible transplants will be obtained via a request for access to the "Cristal" database (https://www.sipg.sante.fr) developed by the Agence de la biomédecine.

Inclusion Criteria:

• Age between 18 and 90 years

• Child-Pugh B or Child-Pugh C cirrhosis diagnosed on the basis of one or several of the following elements :

  • Liver Biopsy
  • Liver stiffness>15kPa measured by Fibroscan
  • Combination of clinical, laboratory and imaging criteria characteristic of cirrhosis (association of signs of portal hypertension, liver failure and abnormal liver morphology in a patient with at least one cause of cirrhosis)

Exclusion Criteria:

• Acute renal failure (increase in serum creatinin level by more than 1.5 times the baseline value or by more than 26 μmol/l from the baseline value) within 15 days before inclusion
• Bacterial infection proven or highly suspected on the basis of clinical-laboratory features within 15 days of inclusion
• Digestive bleeding within 15 days before inclusion
• Alcoholic hepatitis in the previous month
• History of porto-systemic shunt or liver transplant
• Primary sclerosing cholangitis
• Primary biliary cirrhosis
• Budd-Chiari Syndrome
• Hepatocellular carcinoma outside the Milan criteria contraindicating transplantation
• Active extrahepatic neoplasia
• HIV or known immune deficiency or immunosuppressive treatment
• Pregnant or breastfeeding woman
• Protected populations: persons under guardianship or curatorship
• Patient not affiliated to social security
• Patient who did not signed consent
• Ongoing participation in an intervention research whose protocol could, according to the literature, modify the release of hepatocyte microvesicles or the IFN score

Patients aged over 18 and under 90 with stable Child B or C cirrhosis