bwge2017

Authors
C. MINSART (1), C. LIEFFERINCKX (1), S. RORIVE (2), A. LEMMERS (3), C. DRESSEN (4), E. QUERTINMONT (5), E. TRÉPO (6), C. MORENO (6), J. DEVIÈRE (7), S. GORIELY (8), I. LECLERCQ (9), R. MOREAU (10), T. GUSTOT (7) / [1] Université Libre de Bruxelles Faculté de Médecine, Brussels, Belgium, Laboratory of Experimental Gastroenterology , [2] Erasme Hospital, Anderlecht, Belgium, Department of Pathology , [3] Erasme Hospital, Anderlecht, Belgium, Department of Gastroenterology, HepatoPancreatology and Digestive Oncology, [4] Université Libre de Bruxelles Faculté de Médecine, Brussels, Belgium, Laboratory of Physiology and Pharmacology, [5] Université Libre de Bruxelles Faculté de Médecine, Brussels, Belgium, Laboratory of Experimental Gastroenterology, [6] Erasme Hospital, Brussels, Belgium, Department of Gastroenterology, HepatoPancreatology and Digestive Oncology, [7] Erasme Hospital, Brussels, Belgium, Department of Gastroenterology, HepatoPancreatology and Digestive Oncology , [8] Université Libre de Bruxelles Faculté de Médecine, Brussels, Belgium, Institute for Medical Immunology, [9] Université catholique de Louvain (UCL), Brussels, Belgium, Laboratory of Hepato-Gastroenterology, [10] Hôpital Beaujon, Service d'Hépatologie, INSERM U-481, Clichy, France, Centre de recherche sur l'Inflammation (CRI)

Introduction
Voluntary or accidental acetaminophen (N-acetyl-p-aminophenol, APAP) overdose can induce a hyperacute form of liver failure potentially responsible for multiple organ failures and death. The mechanisms of this hepatotoxicity are incompletely understood. In "normal" conditions, high-mobilty group box 1 (HMGB1) is a small nuclear protein who binds DNA and regulate many transcriptional events by modulating transcription factor-DNA interactions. In case of overdose, APAP induces hepatocytes necrosis and thus HMGB1 is released. Extracellular HMGB1 acts like damage-associated molecular pattern (DAMPs) and contributes to APAP-induced liver injury but the mechanisms associated with this activity are incompletely understood or controversial.

Aim

The aim of the present study was to investigate the early effects of HMGB1 in APAP-induced liver injury, its direct effect on hepatocytes and its role in the propagation of necrosis process.

Methods
APAP hepatotoxicity was induced in vivo by intraperitoneal injection in C57Bl/6 mice and in vitro on cultured HepaRG cells. HMGB1 was quantified by ELISA or immuno-staining. Cell death was determined by MTT, ALT, LDH and caspase-3 assays. Glycyrrhizin (GL) and ethyly pyruvate (EP) was used to inhibit HMGB1. Liposomal clodronate was administrated to mice to deplete Kupffer cells (KC). Expression of HMGB1 receptors was assessed by RT-PCR and flow cytometry. Expression of proteins who participate to necroptotic process was demonstrated by western blot. Dabrafenib and necrostatin-1was used to inhibit receptor-interacting protein (RIP)3 and RIP1 respectively.

Results
We confirmed that, in APAP-challenged mice, inhibition of HMGB1 by glycyrrhizin improved survival and reduced further HMGB1 release. Depletion of Kupffer cells by liposomal clodronate in mice exacerbated APAP-induced hepatocyte necrosis and HMGB1 release suggesting that HMGB1 did not act through Kupffer cells activation. Based on these results, we hypothesized that a feed-forward circuit between HMGB1 and hepatocytes exist. In vitro, addition of APAP on cultured HepaRG cells induced cell necrosis characterized by lactate dehydrogenase release without caspase-3 activation, and HMGB1 release. Inhibition of HMGB1 by glycyrrhizin or ethyl pyruvate reduced APAP-induced HepaRG cell necrosis and further HMGB1 release. Exposure of HepaRG cells to recombinant human HMGB1 (rhHMGB1) resulted in cell death, supporting the hypothesis that HMGB1 acts directly on hepatocytes. Inhibition of RIPK3 by dabrafenib prevented APAP- and rhHMGB1-induced HepaRG cell death but inhibition of RIPK1 by Necrostatin-1 did not, suggesting the contribution of necroptosis. Moreover, inhibition of TRIF by Pepinh- TRIF reduced rhHMGB1-induced HepaRG cell death and Trif mutant mice were partially protected against APAP-induced liver injury.

Conclusions
In conclusion, these data support the hypothesis that HMGB1 contributes to the amplification of APAP-induced liver injury through feed-forward circuit with hepatocytes. This pathway seems to be independent of resident macrophages and, at least partially, dependent of TRIF/RIPK3 necrosis pathway resulting in the propagation of the liver injury.

Authors
T. CLAES (1), A. GEERTS (1), H. VAN VLIERBERGHE (1), X. VERHELST (1) / [1] Ghent University Hospital, Gent, Belgium, Department of Gastroenterology and Hepatology
Introduction Primary biliary cholangitis (PBC), formerly known as primary biliary cirrhosis is a relatively rare autoimmune liver disease. PBC can result in cirrhosis with decompensation and HCC development. Currently, ursodeoxycholic acid (UDCA) is the only registered and accepted drug for PBC treatment. Responsive patients in which treatment is started in early stages (I and II) of disease show similar survival rates to age and sex matched groups of the general population. Unfortunately, up to 40% of patients respond suboptimally to treatment. Internationally validated scoring systems predict probability of survival and are able to identify non-responders. Non-responsive patients could profit from other treatment options.
Aim
In this work, first we tried to characterize PBC patients in a single center tertiary Hepatology referral clinic at the Ghent University Hospital. Second we validated two prognostic scores, the GLOBE-score and the Paris-II criteria, in this cohort.
Methods In this retrospective study 67 PBC patients in the Ghent University Hospital were included between 1985 and 2015. Baseline characteristics, biochemical parameters, and outcome data were collected from the Patient Medical Record. The GLOBE-score and the Paris-II criteria were calculated. The Kaplan-Meier method was used to compute observed and expected event-free survival, transplant-free survival and probability of hepatocellular carcinoma (HCC) occurrence. Log Rank test was performed to compare event-free survival, transplant-free survival and probability of HCC occurrence between responders and non-responders according to the GLOBE-score and Paris-II criteria, AMA status, variant/non-variant presentation.
Results There were 50 women and 17 men. Thirty patients (44.8%) were symptomatic at the time of diagnosis. Throughout follow-up 62 patients (92.5%) received UDCA treatment. The mean value of the GLOBE-score was -0.29 (SD=1.66). The GLOBE-score identified 26.1% of patients as non-responders. The Paris-II criteria identified 44.0% of patients as non-responders. A total of 30 patients (44.8%) either suffered from liver related death, underwent liver transplantation or had at least 1 occurrence of a clinical event during follow-up. Twenty-four patients (35.8%) developed biopsy-confirmed liver cirrhosis during follow-up. Liver decompensation occurred in 15 of cirrhotic patients (22.4%). A total of 7 patients (10.3%) developed HCC. During follow-up 9 patients (13.4%) died and 14 patients (20.6%) underwent liver transplantation. The GLOBE-score was able to predict event-free survival (P< 0,001) and HCC occurrence (P ≤ 0.01), but not transplant-free survival (P=0.066). The Paris-II criteria were able to predict both event-free survival (P<0.01) and transplant-free survival (P<0.05).
Conclusions We characterized a single center cohort of PBC in a tertiary Belgian

Authors
N. LANTHIER (1), Q. ETIENNE (1), V. LEBRUN (1), L. POEKES (1), Y. HORSMANS (1), I. LECLERCQ (1) / [1] Université catholique de Louvain, , Belgium, Laboratory of Gastroenterology and Hepatology, Institut de Recherche Expérimentale et Clinique
Introduction Non-alcoholic fatty liver disease (NAFLD) is characterized by steatosis (accumulation of triglycerides in the liver) and insulin resistance. A subgroup of patients can develop a more serious condition called non-alcoholic steatohepatitis (NASH) with increased risk of fibrosis development. Innate immunity, cell injury, lipid metabolism and severity of insulin resistance constitute potential mechanisms underlying disease progression. Fetuin-A , an emerging player in insulin resistance in type 2 diabetic patients, is described as a liver-derived protein increased in human NAFLD.
Aim
Here, we explore the effect of a high fat diet on the expression of fetuin-A and its relation with the development of steatosis, cell injury and liver macrophage (Kupffer cell) activation in a mouse model of obesity and NASH.
Methods Male foz/foz mice were fed a normal diet (ND) or a high fat diet (HFD) for 12 (long term HFD or LHFD) or 30 weeks (very long term HFD or VLHFD) (n=4/group) to induce early or definite fibrosing NASH, respectively. Liver tissue homogenates were prepared for Western blot protein studies and total RNA was extracted for gene expression analysis. Liver paraffin-embedded sections were used for hematoxylin and eosin staining, Sirius red staining and double immunofluorescence detection of F4/80 and fetuin-A.
Results Compared to foz/foz mice fed a ND, HFD-fed foz/foz mice developed obesity, insulin resistance and either steatosis (LHFD) or steatohepatitis with steatosis, hepatocyte ballooning, inflammation and fibrosis (VLHFD). In ND fed mice, fetuin-A staining was positive in the cytoplasm of zone 3 centrilobular hepatocytes while F4/80+ Kupffer cells were located in the sinusoids of the intermediate lobular zone 2. In LHFD fed mice, lipid deposition occurred in the hepatocytes of the zone 3 centrilobular areas. Fetuin-A protein was also located in the cytoplasm of these zone 3 centrilobular hepatocytes. F4/80+ macrophages distributed mainly in the sinusoids of the intermediate lobular zones 2, as seen in ND fed mice. However, liver m-RNA expression showed a 2-fold increased level of F4/80+ macrophage mRNA compared to ND (p<0.05), suggesting activation. In VLHD, we observed a loss of zonation of liver steatosis with the presence of fat loaded hepatocytes in all liver lobular zones. Fetuin-A was highest in periportal fat-ladden hepatocytes and next to inflammatory infiltrates. There was a 4-fold F4/80 mRNA increased level upon VLHFD compared to ND (p<0.05). Three types of F4/80+ cells were recognized on the morphology: elongated cells located in liver sinusoids compatible with liver resident Kupffer cells, cells forming lipogranuloma together with fat loaded hepatocytes and small inflammatory cells located in inflammatory foci compatible with recruited macrophages. Interestingly, F4/80+ cells from lipogranuloma were positive for fetuin-A protein staining. Liver fetuin-A mRNA levels remained unchanged either in LHFD or VLHFD compared to ND. Similarly, liver fetuin-A protein level was also stable under HFD.
Conclusions We demonstrate that VLHFD foz-foz mice develop NASH together with zonal changes of steatosis, liver macrophage activation and fetuin-A expression in fatty hepatocytes and macrophages. A shift of steatosis and fetuin-A from the centrilobular region in ND and LHFD to the periportal zone was observed in VLHFD, together with macrophage activation, recruitment and fetuin-A co-localization in macrophages forming the lipogranuloma. The stable liver fetuin-A protein level could be compatible with a redistribution of this protein and/or the profile of a secretory factor. Taken together, we could imagine that lipid deposition and macrophage infiltration may be important factors in the liver tissue remodeling observed during NASH development. Further work is planned to delineate whether fetuin-A presence in macrophages is linked with a production and/or a simple storage in those cells in this model as well as the role of this protein in NASH progression and insulin resistance pathogenesis.

Authors
S. LEFERE (1), F. VAN DE VELDE (2), L. DEVISSCHER (1), M. BEKAERT (2), S. RAEVENS (1), X. VERHELST (1), Y. VAN NIEUWENHOVE (3), M. PRAET (4), A. HOORENS (4), C. VAN STEENKISTE (5), H. VAN VLIERBERGHE (1), B. LAPAUW (2), A. GEERTS (1) / [1] Ghent University, Ghent, Belgium, Gastroenterology and Hepatology, [2] Ghent University, Ghent, Belgium, Endocrinology, [3] Ghent University, Ghent, Belgium, Gastrointestinal Surgery, [4] Ghent University, Ghent, Belgium, Pathology, [5] Maria Middelares Ziekenhuis, Gent, Belgium, Gastroenterology and Hepatology

Introduction
Non-alcoholic fatty liver disease (NAFLD) is the most common chronic liver disease worldwide and is strongly associated with obesity, dyslipidemia and insulin resistance. NAFLD often presents as simple steatosis (NAFL) but can progress to non-alcoholic steatohepatitis (NASH) and fibrosis. Current non-invasive biomarkers are not tailored to identify significant (≥F2) fibrosis, although recent guidelines recommend a stringent follow-up of this patient population. We and others have reported on the role of pathological angiogenesis in the pathogenesis of NAFLD, highlighting pro-angiogenic factors as potential diagnostic markers.

Aim

We aimed to investigate the applicability of angiogenic factors as non-invasive diagnostic tools for NASH-associated fibrosis in obese patients.

Methods
Serum protein levels and visceral adipose tissue gene expression of endothelial dysfunction and angiogenic markers were analyzed by multiplex bead-based assay and quantitative RT-PCR, respectively, in sixty-one morbidly obese male patients undergoing bariatric surgery and in thirty-five control patients.

Results
We identified serum vascular cell adhesion molecule-1 (VCAM-1) as an independent predictor for ≥F2 fibrosis (median 14.0 vs. 8.7 ng/ml in patients with and without significant fibrosis; P < 0.0001) with an area under the receiver operating characteristics curve (AUROC) of 0.80. The cut-off point of 13.2 ng/ml showed a sensitivity of 80% and specificity of 83%. The AUROC increased to 0.86 when VCAM-1 and the presence of type 2 diabetes were combined. These AUROCs were higher than those for the simple fibrosis risk scores FIB-4 and BAAT. In line with these results, VCAM-1 visceral adipose tissue gene expression was also elevated in patients with fibrosis (P = 0.030).

Conclusions
Serum VCAM-1 levels were able to accurately predict significant (≥F2) fibrosis in obese men with NAFLD.

Authors
A. VANDIERENDONCK (1), F. FORNARI (2), O. GOVAERE (3), D. LAUKENS (1), X. VERHELST (1), A. GEERTS (1), C. AMPE (4), M. VAN TROYS (4), T. ROSKAMS (5), V. HANS (1), L. DEVISSCHER (1), Y. VANDEWYNCKEL (1) / [1] Ghent University, Ghent, Belgium, Gastroenterology and hepatology, [2] Azienda Ospedaliero-Universitaria Policlinico S. Orsola-Malpighi e Università di Bologna, , Italy, Centro di Ricerca Biomedica Applicata (CRBA), [3] Newcastle University, Newcastle upon Tyne, United Kingdom (the), Fibrosis lab, Institute of Cellular Medicine, [4] Ghent University, Ghent, Belgium, Biochemistry, [5] KU Leuven, , Belgium, Translational Cell and Tissue Research, Department of Imaging and Pathology
Introduction The development of hepatocellular carcinoma (HCC) and surrounding micro-environment cause cellular stress. This compromises endoplasmic reticulum (ER)-dependent protein folding and results in ER stress and unfolded protein response (UPR) activation. The UPR aims to restore protein homeostasis or induces cell death via CHOP. Protein degradation is enhanced by UPR-induced proteasome stimulation, a process that is fine-tuned by deubiquitinases (DUBs). DUBs are critical in the regulation of proteins involved in cellular processes and are proposed as potential oncotargets. However, their exact role in HCC development and progression is currently unknown.
Aim
Here, we investigated the expression of Ubiquitin carboxy-terminal hydrolase L1 (UCHL1), which is known to be involved in proteasome-dependent pathways, in human and experimental HCC and the effect of UCHL1 inhibition by LDN57444 on liver tumour cell survival during ER stress in vitro.
Methods Micro-array data of human HCCs and corresponding non-neoplastic liver samples (GSE59259) were used for the expression analysis of the UCH-family of DUBs. Micro-array data and RNA samples of diethylnitrosamine (DEN)-induced HCC livers of rats were provided by dr. Fornari (University of Bologna, Italy). The effect of ER stress and UCHL1 inhibition by LDN57444 on the expression of UPR markers and UCHL1 and cellular viability was analysed in HepG2 cells by using RT-qPCR and MTT assays, respectively.
Results Micro-array analysis revealed that UCHL1 is the only DUB of the UCH-family to be significantly increased in human HCC (p<0.05). This upregulation was confirmed in a DEN-induced HCC rat model, both by micro-array analysis and RT-qPCR. In vitro, the ER stress inducer thapsigargin upregulated the expression of UCHL1 and reduced cell viability (p<0.05). Interestingly, loss of viability was even more pronounced in addition of LDN57444 (p<0.0001). The observed reduced viability might be UPR-mediated since combined treatment of thapsigargin and LDN57444 resulted in significantly enhanced mRNA upregulation of CHOP and its downstream effector GADD34 compared to each mono-treatment.
Conclusions UCHL1 is upregulated in both human and experimental HCC. UCHL1 is induced upon ER-stress in HCC cells and renders those cells more sensitive to ER stress-induced cell death following UCHL1 inhibition. Further in vivo studies will have to reveal if UCHL1 inhibition might be an attractive therapeutic strategy for HCCs characterized by ER stress.

Authors
R. BOTTERMAN (1), E. GLORIEUS (1), S. LEFERE (1), X. VERHELST (1), P. VAN DE VOORDE (2), S. HACHIMI IDRISSI (2), A. GEERTS (1), E. PADALKO (3), P. DE PAEPE (2), H. VAN VLIERBERGHE (1) / [1] Ghent University, Ghent, Belgium, Gastroenterology and Hepatology, [2] Ghent University, Ghent, Belgium, Emergency Department, [3] Ghent University, Ghent, Belgium, Clinical Biology

Introduction
The US Preventive Servies Task Force recommends one-time hepatitis C virus (HCV) screening of all baby boomers (born 1945-1965). Since about the half of the HCV patients in Belgium are not aware of their disease status, this study investigated if the baby boom cohort effect could be present in our country. It is investigated if age or other variables are predictive factors for HCV.

Aim

The aim is to investigate if HCV screening is opportune in the baby boom cohort in Belgium.

Methods
A cross-sectional study was performed from 05/09/2016 until 30/11/2016 at the emergency unit in the Ghent University Hospital. In 1106 patients admitted at the emergency department, after signing informant consent and in need for a blood sample, a HCV ELISA antibody test was performed. A questionnaire on general risk factors for HCV infection was presented to those patients able to fill in this questionnaire.

Results
Data of 1106 patients (606 men/500 women) were collected, 376 (34%) were born between 1945-1965, 494 (44.7%) after and 236 (21.3%) before. There was a HCV positive prevalence of 1.9% in the entire patient population. In the baby boom cohort, the prevalence of HCV was 1.6%. In the younger and older tested population respectively 2.02% and 2.12% were positive. A significant correlation was found between positive HCV results and a history of IV drug use. There was no significant correlation between a positive HCV test and age, sex, blood transfusion before 1990 or tattoos and piercings.

Conclusions
There is a prevalence of 1.9% of HCV positive patients in a Belgian University hospital. This is higher than has been published previously. 44.7% of the tested patients were part of the baby boom cohort. However, there was no significant correlation between HCV positive results and age or age groups. Only screening in the baby boom group seems not opportune and in contrast to what is seen in the US study, a substantial number of HCV positive patients could be missed by the application of this recommendation. Other approaches need to be studied (eg. screening of the entire population, screening of populations at risk such as previous IV drug use).

Authors
J. LAMBRECHT (1), P. POORTMANS (1), H. REYNAERT (2), I. MANNAERTS (1), L. VAN GRUNSVEN (1) / [1] Vrije Universiteit Brussel, Jette, Belgium, Basic Biomedical Sciences, Liver Cell Biology Lab, [2] UZ Brussel, Jette, Belgium, Department of Gastro-enterology and Hepatology

Introduction
Liver fibrosis is the pathological condition of the liver, resulting from sustained scar formation in response to chronic liver injury, such as chronic infection with hepatitis B (HBV) and hepatitis C (HCV) virus. The main effector cell in this pathology is the hepatic stellate cell (HSC) which will undergo a myofibroblastic transdifferentiation process towards an activated state, in which it will excessively produce and deposit extracellular matrix (ECM). Till date, the diagnosis of liver fibrosis occurs by several non-invasive techniques, which are however insensitive and unable to detect early disease stages, and liver biopsy, known to be invasive and associated with some minor and major complications. Novel non-invasive diagnostic scoring systems are being developed, but none of these have yet proven their sensitivity for the detection of early-stage liver fibrosis, nor their potential to discriminate between the various fibrotic stages.

Aim

Our aim was to analyze the potential of circulating miRNAs, with emphasis on extracellular vesicle (ECV)-associated miRNAs as novel biomarkers for early-stage liver fibrosis.

Methods
This study included patients with liver fibrosis by chronic HBV (n=19) or HCV (n=20) infection, which were identified as early-stage fibrotic (≤ F2) by transient elastography (Fibroscan). Relative expression levels of miRNAs in circulating ECVs and total plasma were analyzed by use of qRT-PCR. MiRNAs were selected based on their significant dysregulation during HSC activation (miRNA-192, -200b, -150), or their known presence in the circulation (miRNA-21, 92a, -122). To analyze the potential of these circulating ECVs to represent the presence or absence of activated HSCs in the liver, primary murine HSCs were cultured in vitro to induce spontaneous activation. ECVs were extracted from the culture medium of quiescent- (day 0 till day 2) and activated HSCs (day 8 till day 10).

Results
Analysis of miRNA levels in total plasma confirmed the up-regulation of miRNA-122 during early-stage fibrosis. With the exception of up-regulated levels of miRNA-192 in early-stage HBV-induced fibrosis, no other miRNAs showed significant changes in early-stage fibrosis by HBV or HCV infection. In contrast, miRNA-analysis of circulating ECVs identified significant changing levels of miRNA-150, -192, -200b and -92a during early-stage liver fibrosis by HBV and HCV infection. Especially the down-regulated levels of ECV-associated miRNA-192 (HBV AUC: 0.9802; HCV AUC: 0.9762) and miRNA-200b (HBV AUC: 0.9699; HCV AUC: 0.9841) seem to have an inherent diagnostic potential for early-stage fibrosis. Comparison of the miRNA levels from circulating ECVs with the ECVs extracted from in vitro activating HSCs showed a similar trend in the down-regulation of miRNA-192, suggesting that ECV-associated miRNA-192 levels might represent the activation status of HSCs in the liver.

Conclusions
Circulating ECV-associated miRNAs could be used as novel tools for the diagnosis of early-stage liver fibrosis, through their potential to identify the absence or presence of activated HSCs in the liver.

Authors
X. VERHELST (1), A. GEERTS (2), D. VANDERSCHAEGHE (3), X. ROGIERS (4), A. VANLANDER (4), F. BERREVOET (4), N. CALLEWAERT (3), R. TROISI (4), H. VAN VLIERBERGHE (2) / [1] Ghent University Hospital, Ghent, Belgium, Department of Gastroenteroly and Hepatology, [2] Ghent University Hospital, Ghent, Belgium, Gastroenterology and Hepatology, [3] VIB, Gent, Belgium, Medical Biotechnology Center, [4] Ghent University Hospital, Ghent, Belgium, Hepatobiliary and Transplant Surgery
Introduction Primary non function (PNF) is a rare but major complication after liver transplantation requiring urgent retransplantation. It is associated with the use of extended-criteria donors. The donor risk index is a clinical score that can guide the estimation of graft quality but lacks the power to predict PNF risk in individual patients. Perfusate analysis is an attractive tool for assessment of donor liver function before implantation. Glycomic assessment of serum has proven useful in the diagnosis of liver disease.
Aim
The aim of this study was to identify a specific glycomic signature in perfusate that is associated with PNF after liver transplantation.
Methods In this prospective monocentric study 66 consecutive liver transplantations between October 2011 and July2013 were included. Perfusate samples were collected after flushing of the hepatic veins before implantation of the liver graft. All donor grafts were transported using cold static storage. Based on an optimized DNA sequencer technology we performed glycomic analysis of these perfusate samples and searched for glycomic alterations in PNF patients.
Results One single glycan, an agalacto core-alpha-1,6-fucosylated biantennary glycan (NGA2F) was significantly increased in the perfusate of the 3 patients that developed PNF after liver transplantation. It could identify PNF patients with 100% accuracy. This glycomarker was the only predictor of PNF in a multivariate analysis including donor risk index and perfusate AST/ALT levels (p<0.0001).
Conclusions In this cohort, patients who developed PNF after liver transplantation showed a specific glycomic signature in perfusate (before liver transplantation) that could distinguish them from non-PNF patients with 100% accuracy. This approach could guide the removal of donor grafts at risk for PNF from the donor pool, especially when the use of highrisk organs is considered.

Authors
X. VERHELST (1), A. GEERTS (2), X. ROGIERS (3), A. VANLANDER (3), F. BERREVOET (3), N. CALLEWAERT (4), R. TROISI (3), H. VAN VLIERBERGHE (2) / [1] Ghent University Hospital, Ghent, Belgium, Department of Gastroenteroly and Hepatology, [2] Ghent University Hospital, Ghent, Belgium, Gastroenterology and Hepatology, [3] Ghent University Hospital, Ghent, Belgium, Hepatobiliary and Transplant Surgery, [4] VIB, Gent, Belgium, Medical Biotechnology Center
Introduction Graft failure after liver transplantation (LT) remains a challenge for transplant professionals and sometimes requires retransplantation. Pretransplant estimation of graft function using scores like donor risk index has limited use in individual patients. Diagnosis of early allograft dysfunction after liver transplantation by clinical criteria can predict graft survival. However, biomarkers that reliably identify patients at risk for graft failure after LT are lacking. Analysis of N-glycans in serum (glycomics) has shown to reflect the underlying liver function in liver disease but has never been assessed after liver transplantation.
Aim
The aim of this study was to assess the potential of serum glycomics as predictive markers for graft and patient survival after liver transplantation.
Methods Serum glycomic profiles were collected before and after liver transplantation in 127 consecutive liver transplant recipients between 1 December 2012 and 31 December 2014 and analysed using an optimized glycomic technology on a DNA sequencer. The major outcome parameters (graft and patient survival) were related to the observed glycomic alterations.
Results The assessment of 2 serum glycans NG1A2F (an agalacto, core-alpha-1,6-fucosylated biantennary glycan structure) and NA3 (a triantennary glycan), combined as log(NA3/NG1A2F) on day 7 after liver transplantation was strongly associated with graft loss (hazard ratio = 7.222; p<0.001; 95% CI 2.352-22.182) and patient death (hazard ratio = 3.885; p=0.30; 95% CI 1.127-13.276) during the first year after liver transplantation (cox regression analysis). In a multivariable cox regression model including early allograft dysfunction (according to Olthoff) and Donor Risk Index, this glycomic marker, called GlycoTransplantTest, was the only independent predictor of graft survival (p=0.003).
Conclusions Assessment of GlycoTransplantTest, a glycomic serum marker, on day 7 post liver transplantation is a strong and independent predictor of graft survival during the first year after liver transplantation.

Authors
S. VAN HEES (1), S. BOURGEOIS (2), H. VAN VLIERBERGHE (3), T. SERSTÉ (4), P. MICHIELSEN (1), H. REYNAERT (5), J. HENRION (6), S. NEGRIN-DASTIS (7), L. LASSER (8), F. JANSSENS (9), G. ROBAEYS (10), P. STÄRKEL (11), C. MORENO (12), F. NEVENS (13), T. VANWOLLEGHEM (1) / [1] Antwerp University Hospital, Edegem, Belgium, Department of Gastroenterology and Hepatology, [2] ZNA Antwerpen, , Belgium, Department of Gastroenterology and Hepatology, [3] Ghent University Hospital, Ghent, Belgium, Department of Gastroenterology and Hepatology, [4] CHU Saint-Pierre, Brussels, Belgium, Department of Gastroenterology and Hepatology, [5] University Hospital Brussels, Vrije Universiteit Brussel, , Belgium, Department of Gastroenterology and Hepatology, [6] Centre Hospitalier de Jolimont-Lobbes., La Louvière, Belgium, Department of Gastroenterology and Hepatology, [7] Grand Hopital de Charleroi, Charleroi, Belgium, Department of Gastroenterology and Hepatology, [8] CHU Brugmann Brussels, Brussels, Belgium, Department of Gastroenterology and Hepatology, [9] Jessa Hospital, Hasselt, Belgium, Department of Gastroenterology and Hepatology, [10] ZOL, Genk, Belgium, Department of Gastroenterology and Hepatology, [11] Cliniques universitaires St-Luc, Brussels, Belgium, Department of Gastroenterology and Hepatology, [12] CUB Hôpital Erasme, Bruxelles, Belgium, Department of Gastroenterology, Hepatopancreatology and Digestive Oncology, [13] University Hospital Leuven, KU Leuven, , Belgium, Department of Gastroenterology and Hepatology

Introduction
Cessation of Nucleo(s)tide analogues (NA) therapy after HBeAg seroconversion is associated with a high degree of relapse, but evidence in Caucasian patients is scarce.

Aim

We investigated relapse rates and clinical outcomes after NA stop in a Belgian cohort of HBeAg positive Chronic Hepatitis B (CHB) patients.

Methods
This is a pooled analysis of non-immune-suppressed HBeAg-positive, mono-infected CHB patients from 13 hospitals in Belgium, treated with different NA for ≥ 3 months. Data were collected between 1998 and 2016. HBeAg seroconversion was defined as the loss of HBeAg and the appearance of anti-HBeAg on two time points ≥1 month apart. Virological relapse was defined as HBV DNA>2000 IU/mL; biochemical relapse as ALT>2xULN (with ULN defined as 40 IU/mL). Clinical events were defined as the appearance of hepatic decompensation, HCC or liver-related death. Cox regression model was used to identify predictive factors for relapse. Follow-up time was calculated as time from HBeAg seroconversion until relapse or end of follow-up.

Results
A total of 326 patients (74.8% male; 63% Caucasian; 17% African) were included; 96 of whom showed HBeAg seroconversion. Treatment was stopped in 57/96 patients (of whom 8 were cirrhotic at baseline) after HBeAg seroconversion with a subsequent median consolidation therapy of 7.5 months. The median follow-up after treatment stop was 2.9 years during which 25 patients showed relapse (14 solely virological, 11 combined biochemical and virological), necessitating retreatment in 15 cases. HBeAg reversion was observed in 3/25 (12%) relapsed patients. Cox regression model showed that neither the presence of cirrhosis (HR 3.386; p=0.116) at start of treatment, nor Caucasian ethnicity (HR 0.509; p=0.133) were significantly associated with relapse after treatment stop. Relapse was accompanied by hepatic failure in two cases leading to liver-related death. Treatment was continued after HBeAg seroconversion in 26 patients (of whom 9 were cirrhotic at baseline) for a median of 4.1 years. Three patients (all cirrhotic) developed ascites in the latter group, but recovered thereafter. No patient died.
Conclusions: Treatment cessation after HBeAg seroconversion led to relapse in 44% of predominantly Caucasian patients within a median follow-up 1056 days. Two relapsed patients showed severe clinical events leading to liver-related death.

Conclusions
Treatment cessation after HBeAg seroconversion led to relapse in 44% of predominantly Caucasian patients within a median follow-up 1056 days. Two relapsed patients showed severe clinical events leading to liver-related death.

Authors
B. DELIRE (1), P. HENRIET (2), P. LEMOINE (2), I. LECLERCQ (1), P. STÄRKEL (3) / [1] Institut de Recherche Expérimentale et Clinique (IREC), Catholic University of Louvain (UCL), , Belgium, Laboratory of Hepato-Gastroenterology, [2] De Duve Institute, , Belgium, Cell Biology Unit, [3] Saint-Luc Academic Hospital and Institute of Clinical Research, Catholic University of Louvain, , Belgium, Department of Gastroenterology
Introduction Liver fibrosis is the main risk factor for hepatocarcinoma (HCC). Mechanisms linking fibrosis and hepatocarcinogenesis remain however poorly understood. In many malignant diseases, inflammatory cells that infiltrate the tumor are key players in cancer development.
Aim
Our aim was to study in a mouse orthotopic transplantation model the impact of fibrosis on HCC development and local tumor infiltration and explore potential roles of macrophages and neutrophils.
Methods The HCC cell line Hepa 1-6 is syngenic with the C57Bl/6 mouse strain. Hepa 1-6 cells were injected into non-fibrotic livers (normal liver group-NLG) and in severe fibrotic livers (severe fibrosis group-SFG) without immunosuppressive therapy. Severe fibrosis was induced by CCl4 for 7 weeks. Mice were sacrificed 2 weeks post HCC cell injection. The liver was sliced and examined for the presence of tumor (nodule ≥ 1mm). The tumor volume and the liver to body weight ratio (LW/BW) were used as parameters of tumor burden. A part of each tumor was used for histological analysis, proteins and RNA preparation.
Results A tumor nodule was observed in 60% of animals in the NLG but in 100% of them in SFG. The tumor volume and the LW/BW were significantly higher in the SFG (p<0.0001; p=0.005) compared to the NLG. Tumor macrophages infiltration was evaluated by F4/80 immunohistochemistry: while F4/80 positive cells were mainly located around the tumor in NLG livers, macrophages infiltrated deeper the HCC nodules in SFG livers. F4/80 mRNA expression (p<0.0003) as well as CD11b expression (p<0.0003), a marker of recruited macrophages, were higher in SFG than in NFG tumors. Similarly, we observed a higher NIMP-R14+ neutrophils infiltration in tumors that developed in SFG compared to those in NLG (p=0.0289). Many tumor-associated macrophages and neutrophils-derived molecules such as matrix metalloproteinase (MMP-2) and MMP-9 are involved in tumor progression and invasiveness. Compared to NLG, tumors in SFG livers expressed higher levels of Mmp2 (p=0.0019) and Mmp9 (p=0.0047). Mmp2 mRNA was significantly higher in the tumor compared to adjacent liver parenchyma in both groups (NLG: p=0.0006;SFG: p=0.0002) while high Mmp9 expression in tumor compared to adjacent parenchyma was only seen in SFG livers (p=0.0006) but not in NLG livers. Furthermore, tumor volume positively and significantly correlated with intra-tumor Mmp2 (rS=.571, p=0.026) as well as with intra-tumor Mmp -9 (rS=.741, p 0.002) mRNAs. Zymography evaluates pro- and active MMP-2/ -9: pro- and active MMP-2 and -9 were significantly higher in SFG tumors compared to NLG tumors (pro-MMP-2:p=0.0007, active MMP-2:p=0.008; pro-MMP-9:p=0.008, active MMP-9: p<0.05). Similarly to gene expression, MMP-2 and -9 enzymes were significantly more active in tumor than in adjacent parenchyma (MMP-2: p<0.05; MMP-9: p<0.05). MMP-2 and -9 are known activators of transforming growth factor β (TGFβ), an inflammatory cytokine that promotes tumor cells growth. TGFβ mRNA expression was higher in SFG than in NLG tumors (p=0.0012). Moreover, there were higher amounts of active (cleaved) TGFβ protein, measured by ELISA, in the SFG tumors compared to the NLG tumors.
Conclusions Liver fibrosis promotes HCC development in a mouse orthotopic transplantation model. Our results suggest that a fibrotic liver background favors a higher infiltration of tumor associated macrophages and neutrophils in the developing tumor. These secrete and activate molecules such as MMP-2, MMP-9 and TGFβ that promote tumor progression.

Authors
S. FRANCQUE (1), P. LEFEBVRE (2), F. LALLOYER (2), M. PAWLAK (2), E. BAUGÉ (2), C. GHEERAERT (2), H. DEHONDT (2), J. VANHOUTTE (2), N. HENNUYER (2), C. CLAIRE MAZUY (2), B. DERUDAS (2), A. DRIESSEN (3), G. HUBENS (4), L. VONGHIA (1), W. KWANTEN (1), T. VANWOLLEGHEM (1), P. MICHIELSEN (1), J. EECKHOUTE (2), A. VERRIJKEN (5), L. VAN GAAL (5), B. STAELS (2) / [1] ANTWERP UNIVERSITY HOSPITAL, Edegem, Belgium, Gastroenterology Hepatology, [2] Univ. Lille, CHU-Lille, Institut Pasteur de Lille, Lille, France, Inserm, [3] ANTWERP UNIVERSITY HOSPITAL, Edegem, Belgium, Pathology, [4] ANTWERP UNIVERSITY HOSPITAL, Edegem, Belgium, Abdominal Surgery, [5] ANTWERP UNIVERSITY HOSPITAL, Edegem, Belgium, Endocrinology, Diabetology and Metabolic Diseases

Introduction
Pathogenic mechanisms leading to progression from simple steatosis towards active non-alcoholic steatohepatitis (NASH) and fibrosis are poorly defined.

Aim

We investigated the liver transcriptome in a human cohort of histologically staged NASH patients both at baseline and follow-up to identify key components of progression of disease and hence potential targets for therapy.

Methods
Obese patients were prospectively screened for presence of NASH and if suspected, liver biopsy was proposed. Patients entered a weight management program, including bariatric surgery (BarSur) in some, and were re-evaluated after 1 year including biopsy. Liver biopsy was scored using the NASH CRN scoring system. Gene profiling (Affymetrix GeneChip arrays + functional annotation and enrichment) was performed. Paired analysis of the liver transcriptome before and 1 year after BarSur identified genes dysregulated in NASH and fibrosis and whose expression was normalized upon regression of lesions. A meta-analysis with publicly available datasets with comparable histology was carried out to even more stringently identify genes dysregulated in NASH and fibrosis. Data were further crossed with transcriptomic data from NASH and fibrosis mouse models.

Results
Analysis was performed in 87 patients with paired biopsies. Progressive baseline histological damage from steatosis to NASH to NASH+fibrosis were characterized by gene expression patterns successively reflecting altered functions in metabolism, inflammation and epithelial-mesenchymal transition. The molecular signature for active NASH+fibrosis contained 193 upregulated genes (immune responses and ECM homeostasis) and 58 downregulated (metabolic pathways). Of these, 103 and 36 were normalized after BarSur, leading to a 139-gene signature of NASH+fibrosis normalized upon resolution. Comparison with existing databases led to a 24 BarSur-sensitive human NASH+fibrosis signature strongly enriched with ECM matrix formation and inflammatory responses. Comparison with NASH and fibrosis gene signatures of MCD and CCl4 mouse models respectively resulted in a 16-gene set of NASH+fibrosis with normalisation upon regression. This analysis pointed towards dermatopontin (DPT) as an important player.

Conclusions
Liver damage during NASH progression is characterized by deregulated expression of a restricted set of inflammation- and ECM-related genes. Targeting DPT may be a valuable strategy to reverse the hepatic fibrotic process.

Authors
R. MANCO (1), L. CLERBAUX (1), I. LECLERCQ (1) / [1] Université Catholique de Louvain, Brussels, Belgium, Laboratory of Hepato-Gastroenterology
Introduction Self-renewal of mature hepatocytes supports homeostasis and regeneration of adult liver. Recent studies indicate that liver progenitor cells (LPC) are recruited upon injury as a facultative reservoir for generation of hepatocytes, although only a small number of mature hepatocytes were shown to derive from LPC in vivo. Models used for these studies do not recapitulate long lasting chronic hepatocellular damage and fibrosis seen in human chronic liver disease and cirrhosis.
Aim
Our aim is therefore to follow the dynamics of ductular reaction (DR) and the LPC’s fate during chronic liver injury in mice.
Methods We used tamoxifen-inducible Osteopontin-Cre (OPN-CreERT2) mice crossed with yellow fluorescent protein (YFP) reporter mice to follow the fate of LPC and biliary cells with an efficiency >85%. Long-term chronic injury was induced by repeated carbon tetrachloride (CCl4) injections 3x/week for 4, 6, 8, 16 and 24 weeks, resulting in chronic fibrosis and eventually cirrhosis. Livers from 8 and 16 weeks were also analysed after 4 weeks and 2 and 4 weeks of CCl4-free recovery period, respectively.
Results After 4 weeks CCl4, DR is minimal with few ck19+/YFP+ positive cells in periportal area and LPC-derived hepatocytes (traced as YFP+ hepatocytes) are inconspicuous. After 6 weeks, DR is similar in intensity but small foci of YFP+ hepatocytes adjacent to portal area are readily seen; these have a median size of 3010µm². As fibrotic disease increases in severity, the DR is negligible while patches of YFP+ hepatocytes become larger (median size of 3850µm² and 7040µm² at 8 and 16weeks, respectively) and extend to into the parenchyma. In the cirrhotic liver (24 weeks CCl4) some regenerative nodules are entirely composed of YFP+ hepatocytes. The number of YFP+ hepatocytes does no rise accordingly to the size of the patches as they represent 4.2 ± 2.4% of the lobule area in 6 weeks’ samples, increases up to 11.5 ± 3.8 % in 8 weeks’ samples and stabilizes around 5% thereafter, suggesting that not all YFP+ hepatocytes expand into growing patches. At 6 weeks, YFP+ hepatocytes are significantly smaller cells than YFP- native hepatocytes (750 vs 981 µm²) but in 16 weeks’ samples YFP+ and YFP- hepatocytes have the same size (996 and 1001 µm²). The dynamic of the YFP+ hepatocytes was also evaluated upon recovery: in the 4 weeks following 8 weeks of CCl4, the area occupied by YFP+ hepatocytes has a tendency to decrease from 11.5 ± 3.8% to 5.03 ± 3.8% (p=NS), while in the 2 and 4 weeks of recovery after 16 weeks of CCl4 the area significantly increases from 4.58 ± 1.7% to 7.7 ± 3.3% (p=NS), up to 13.8 ± 0.7 % (p<0.001), respectively. Whereas, upon recovery the size of the YFP+ hepatocytes, in all the different time points, is the same of the native hepatocytes.
Conclusions Our data demonstrate a significant contribution of LPC to the hepatocytes regeneration in a model of chronic liver injury leading to cirrhosis. The kinetic study supports that when DR is present, LPC differentiate into small hepatocytes, some of these subsequently increase in number, to form growing patches, and in size, becoming undistinguishable from the native hepatocytes. Upon recovery the growth of the patches of the LPC-derived hepatocytes depends on the severity of the underlying injury. Clonality studies are ongoing to test this hypothesis.

Authors
E. VERLY (1), X. VERHELST (1), H. VAN VLIERBERGHE (1), A. GEERTS (1) / [1] Ghent University Hospital, Ghent, Belgium, Department of hepatology & gastroenterology

Introduction
Variceal bleeding is a severe complication of cirrhosis. The treatment of variceal bleeding is based on proper supportive care, vasoactive medication and endoscopic therapy. Since 2010, early TIPS placement has shown improved survival in patients variceal bleeding with a high risk for rebleeding as defined by Garcia Pagan et al. (NEJM 2010).

Aim

The aim of this study was to retrospectively review the use of TIPS in variceal bleeding in our centre.

Methods
This retrospective monocentric study was performed in a tertiary referral centre for liver disease and liver transplantation (Ghent University Hospital). All patients admitted with variceal bleeding between January 2010 and December 2014 were included. Clinical data and results were retrieved from the medical files. Outcome was assessed at hospitalisation, 3 and 12 months after variceal bleeding.
Statistical analysis was performed using SPSS (version 23).

Results
In this cohort 56 patients were identified with variceal bleeding, 16 female and 40 male patients between the ages of 22 and 84. Forty-nine (87,5%) patients survived the hospitalisation, 48 (85,7%) were alive after 3 months and 1-year survival was 73,2% (41 patients). 17 patients had a CHILD-PUGH classification of A, 24 CHILD-PUGH B and 12 were CHILD-PUGH C. Of 3 patients, the CHILD-PUGH score could not be calculated due to missing variables. All patients received supportive care, vasoactive medication and endoscopy within 12 hours of admission.
In this cohort, 20 patients were treated with TIPS placement. 6 of these patients were classified as CHILD-PUGH A, 9 as CHILD-PUGH B and 5 as CHILD-PUGH C. Eleven patients (19.6%) received TIPS in the early-TIPS strategy after initial bleeding, 7 (12.5%) due to a rebleeding episode. In two patients (3.6%) TIPS placement was postponed after the 72 hours time window but was given as an early-TIPS placement, and not due to rebleeding. In the early-TIPS group, 3 month and one year survival was respectively 92,3% and 84,6%. Transient encephalopathy after TIPS placement was observed in 7 patients (35,0%). In the early TIPS group, 4 patients (30,8%) had transient encephalopathy.

Conclusions
The implementation of the early TIPS protocol for variceal bleeding is safe and shows excellent one-year survival rates in this high-risk population. Serious Adverse events were rare and manageable in the majority of patients.

Invited lecture by Thierry Gustot (Erasme, ULB)

Authors
A. DILI (1), V. LEBRUN (2), C. BERTRAND (3), I. LECLERCQ (2) / [1] CHU ULC Namur and Laboratory of Gastroenterology, Istitut de Recherche Expérimentale et Clinique, UCL, Brussels, Brussels, Belgium, Surgery and Laboratory of Hepato-Gastroenterology, Institut de Recherche Expérimentale et Clinique, Université catholique de Louvain, Brussels, Belgium, [2] Laboratory of Hepato-Gastroenterology, Institut de Recherche Expérimentale et Clinique, Université catholique de Louvain, Brussels, Belgium, Brussels, Belgium, Laboratory of Hepato-Gastroenterology, Institut de Recherche Expérimentale et Clinique, Université catholique de Louvain, Brussels, Belgium, [3] CHU UCL Namur, Yvoir, Belgium, Surgery
Introduction ALPPS is a surgical technic that combines portal vein ligation (PVL) and parenchymal transection followed by resection of the deportalized liver within 2 weeks. ALPPS achieves rapid hypertrophy of the future liver remnant (FLR) protecting patients from liver failure after extended otherwise non-viable hepatectomy (small for size syndrome-SFSS). SFSS is related to portal hyperperfusion of a very small hepatic parenchyma, with a compensatory constriction of the common hepatic artery (hepatic arterial buffer response-HABR) believed lied to desarterialisation of FLR and postoperative liver failure. In ALPPS, PVL and parenchymal transection redirect the whole portal flow through a small FLR. Despite a growing use of the ALPPS procedure in clinics, consequences on arterial flow and underlying mechanisms for accelerated regeneration and protection from SFSS are still unknown.
Aim
There are reports on animal models for ALPPS, but none accurately mimics the human procedure: rodent models either do not achieve liver resection leaving a small, insufficient for survival, FLR or propose hepatic resection during the first step of ALPPS. Differences in volume of FLR and in surgical events may introduce bias in our understanding of pathophysiological mechanisms. This study aims to develop a model mimicking ALPPS with minimal FLR and to analyze hepatic hemodynamics.
Methods In rodents, the left median lobe (LML), represents 10% of the liver volume. Px90 represents a total hepatectomy except LML, transection (T) a hepatotomy between the right and left segment of median lobe and PVL a ligation of all portal branches except those that perfuse LML. PVLT followed by Px9O is a strict copy of conventional human ALPPS. The first experiment (group A) studied the volume hypertrophy of LML after a unique procedure (T, PVL, PVLT and sham); rats were harvested at 6hours,1,2,3,7days. The second experiment (group B) analyzed mortality and volume hypertrophy after Px90 and two step procedures, PVL-Px90 and PVLT-Px90. Flow rate in portal trunc and common hepatic artery (HA) were measured by US-Doppler in Sham, PVL, PVLT and Px90.
Results In group A, hypertrophy of FLR was greater at day 2 and 3 after PVLT compared to PVL (p<0,05) but not at day 7, suggesting that PVLT accelerated initial hypertrophy. Hepatocyte proliferation, assessed by Ki67 and BrdU IHC, was significantly higher at day 2 and 3 in PVLT remnants (p<0,05). We observed no hypertrophy after T. In group B, ALPPS was associated with a low seven day mortality rate (29.41%) compared to Px90 (77.7%) or PVL-Px90 (38.46%) (p<0.05). Acceleration in regeneration was confirmed by a significantly higher kinetic growth ratio in 1st and 2nd stage ALPPS (PVLT, PVLT-Px90) compared to PVL and PVL-Px90 (p<0,005). Total portal vein flow was similarly reduced after PVL, PVLT and Px90 compared to sham (p<0,001). However, because 90% of the liver parenchyma was excluded from the portal circulation in PVL, PVLT and Px90, the portal flow in the FLR was increased by a factor 4 to 5 compared to flow reaching LML in sham animals (p<0.0001). A decrease in HA flow occurred after PVL and PVLT compared to sham (p<0.001) and was further lowered after Px90 (p<0.5 vs PVLT; p<0.01 vs PVL) suggesting a HABR concommitent to portal hyperperfusion in all 3 procedures. While arterial blood is distributed in the entire liver in PVL and PVLT, it only enters the 10% FLR in Px90, in consequence, effective arterial flow into FLR is increased after Px90, but is halved after PVLT (p<0.05) and decreased in a lesser extend after PVL (p=ns). Immunohistochemistry using pimonidazole (an ischemia marker) demonstrated a significantly higher ischemia at day 1 in PVLT compared to sham, PVL and Px90 (p<0,05).
Conclusions We describe the first animal model with minimal FLR, leading to high mortality due to SFSS unless ALPPS is applied. The degree of liver growth and kinetic growth ratio confirm that ALPPS boosts liver hypertrophy more than PVL. Hemodynamic study suggests that even if HABR exists in Px90, the SFSS consecutive to this kind of marginal hepatectomy is not related to parenchymal desarterialisation; on the contrary, reduction of arterial parenchymal perfusion as observed in PVLT (the first step in ALPPS procedure) may protect the FLR from hepatocellular failure and stimulate regeneration. This model reproduces the objectives intended in human conventional ALPPS and should be valuable for study of physiological mechanisms.

Authors
R. BIELEN (1), C. MORENO (2), H. VAN VLIERBERGHE (3), S. BOURGEOIS (4), J. MULKAY (5), S. FRANCQUE (6), W. VERLINDEN (6), C. BRIXKO (7), J. DECAESTECKER (8), C. DE GALOCSY (9), F. JANSSENS (10), M. COOL (11), L. VAN OVERBEKE (12), C. VAN STEENKISTE (13), F. D'HEYGERE (14), K. WUYCKENS (1), F. NEVENS (15), G. ROBAEYS (16) / [1] University Hasselt, Hasselt, Belgium, Faculty of medicine and life sciences, [2] Erasme Hospital, Brussels, Belgium, Gastro-Enterology and Hepatopancreatology, [3] UZ Gent, Gent, Belgium, Gastro-Enterology and Hepatology, [4] ZNA Stuivenberg Hospital, , Belgium, Gastro-Enterology and Hepatology, [5] Hôpital Saint-Pierre, , Belgium, Gastro-Enterology and Hepatology, [6] Antwerp University Hospital, Edegem, Belgium, Gastro-Enterology and Hepatology, [7] CHR La Citadelle, , Belgium, Gastro-Enterology and Digestive Oncology, [8] AZ Delta, Roeselare, Belgium, Gastro-Enterology and Hepatology, [9] Hôp. Iris Sud Bracops, Bruxelles, Belgium, Gastro-Enterology and Hepatology, [10] Jessa Hospital, Hasselt, Belgium, Gastro-Enterology and Hepatology, [11] AZ Damiaan, Oostende, Belgium, Gastro-Enterology and Hepatology, [12] AZ Sint-Maarten, Mechelen, Belgium, Gastro-Enterology and Hepatology, [13] AZ Maria Middelares, Ghent, Belgium, Gastro-Enterology and Hepatology, [14] AZ Groeninge, Kortrijk, Belgium, Gastro-Enterology and Hepatology, [15] University Hospitals Leuven, KU Leuven, Leuven, Belgium, Gastro-Enterology and Hepatology, [16] Ziekenhuis Oost-Limburg Genk, Genk, Belgium, Gastro-Enterology and Hepatology

Introduction
Direct antiviral agents (DAA) have made HCV treatment very effective and safe these last years. Recently, concerns were raised of high rates of HCC recurrence in patients treated with DAA.

Aim

We investigated the HCC occurrence and recurrence rates within six months after treatment with DAA with or without Pegylated Interferon (PEG-IFN).

Methods
This is a national, retrospective, multicenter cohort trial, executed in 15 hospitals distributed across Belgium. Data were available from two earlier trials investigating the outcome of treatment with DAA with and without PEG-IFN. A new data collection based on the patient files was executed by medical doctors. Populations were matched based on fibrosis score starting from F3. Patients with a Child-Pugh score ≥ B were excluded. In total, 472 patients were included in this trial, of whom 72 were treated with DAA with PEG-IFN from 2008 to 2013 and 400 with DAA without PEG-IFN from 2013 until November 2015. In this cohort also an analysis of the rates of follow up by radiographic analysis was performed.

Results
Patients treated with DAA with PEG-IFN (53y±8) were younger than patients treated with DAA without PEG-IFN (59y±12) (p=0.001). 48% (38/72) of patients treated with DAA with PEG-IFN were in the F4 stage versus nearly 65% (259/399) of patients treated with DAA without PEG-IFN (p=0.004). The rates of radiographic follow up were 77.8% (n=56/72) in patients treated with DAA with PEG-IFN, and 78.0% (n=312/400) in patients treated with DAA without PEG-IFN. The early occurrence rate of HCC in patients treated with DAA with PEG-IFN was 3.6 % (n=2/55) and 1.1% (n=3/277) in patients treated with DAA without PEG-IFN. The early recurrence rate was 0% (n=0/1) in patients treated with DAA with PEG-IFN, and 20.0% (n=7/35) in patients treated with DAA without PEG-IFN.

Conclusions
There is no difference in early occurrence of new HCC between patients treated with DAA with and without PEG-IFN. We did observe a high early recurrence rate of HCC in patients treated with DAA without PEG-IFN. However, we cannot state that this difference is significant to patients treated with DAA with PEG-IFN, especially since there were significant differences in patient characteristics such as age and fibrosis stage. In 20%, screening for HCC was inadequate. More efforts are necessary as we need to remain vigilant when treating high risk patients.

Authors
M. VAN DE GARDE (1), S. PAS (2), G. VAN OORD (3), L. GAMA (4), Y. CHOI (5), R. DE MAN (3), A. BOONSTRA (3), T. VANWOLLEGHEM (6) / [1] Erasmus Medical Center, Rotterdam, Netherlands (the), Gastroenterology and Hepatology, [2] Erasmus Medical Center, Rotterdam, Netherlands (the), Department of Viroscience, [3] Erasmus Medical Center, Rotterdam, Netherlands (the), Department of Gastroenterology and Hepatology, [4] Johns Hopkins University, Baltimore, United States (the), Department of Molecular and Comparative Pathobiology, [5] Center for Disease Control and Prevention, Atlanta, United States (the), Division of Viral Hepatitis , [6] Erasmus Medical Center, Rotterdam, Netherlands (the), Gastroenterology&Hepatology

Introduction
Safe and effective antiviral options are needed for ribavirin intolerant, immunocompromised patients with chronic Hepatitis E Virus (HEV) genotype (gt) 3 infections. Pegylated interferon (pegIFN) has been used extensively to treat chronic viral hepatitis infections and baseline intrahepatic IFN-stimulated gene (ISG) expression has been linked to treatment success.

Aim

We studied the antiviral potency of pegIFN against HEV gt3, HEV gt1 and HBV gtA infections in an immunocompromised small animal model and modelled intrahepatic ISG responses pre- and post-treatment.

Methods
65 uPA+/+Nod-SCID-IL2Rγ-/- mice were transplanted with one of three human hepatocyte donors. Human liver-chimeric mice were challenged with HEV gt3, HEV gt1 or HBV gtA. Infected mice received either a single or twice weekly injections with pegIFNa-2b for 2 or 4 weeks. Quantification of HEV RNA was performed in liver, bile and feces using RT-qPCR. Human gene expression of human-chimeric mouse livers was analyzed using RT-qPCR and the nanostring nCounter® human-immunology panel for respectively 10 and 578 genes. 5 Non-chimeric mice were used as controls. Human CXCL10 was measured in mouse serum.

Results
HEV gt3 infections were cleared from liver and feces after 8 and 4 pegIFN doses, but relapsed in 2/4 mice after a single pegIFN injection. PegIFN anti-HEV activity was confirmed in HEV gt1 infected mice with complete clearance from liver and feces after 4 injections. In contrast, HBV gtA infected mice showed a 6 log IU/gr liver) at the end of a 2 week pegIFN treatment course. Baseline pre-treatment ISG expression was evaluated in 20 HEV gt3 and 10 HEV gt1 infected chimeric-mouse livers and revealed no ISG induction compared to 8 control chimeric mice. An in-depth gene expression array on 14 HEV gt3 infected chimeric-mice confirmed the absence ISG induction, irrespective of time point after inoculation, hepatocyte donor or HEV strain. Post- pegIFN treatment a clear human specific ISG induction was observed in liver (>10-fold CXCL10 mRNA increase), which led to increased circulating human CXCL10 levels in mouse serum.

Conclusions
HEV gt1 and gt3 infections do not induce innate intrahepatic immune responses and are extremely sensitive to pegIFN in immunocompromised humanized mice. This might inform treatment strategies for ribavirin resistant HEV.

Authors
R. BIELEN (1), G. ROBAEYS (2), S. SCHELFHOUT (3), D. MONBALIU (4), S. VAN DER MERWE (3), J. PIRENNE (4), F. NEVENS (3) / [1] University Hasselt, Hasselt, Belgium, Faculty of medicine and life sciences, [2] Ziekenhuis Oost-Limburg Genk, Genk, Belgium, Gastro-Enterology and Hepatology, [3] University Hospitals Leuven, KU Leuven, Leuven, Belgium, Gastro-Enterology and Hepatology, [4] University Hospitals Leuven, KU Leuven, Leuven, Belgium, Abdominal Transplant Surgery

Introduction
Intravenous Hepatitis B Immunoglobulins (HBIG) in combination with nucleos(t)ide analogues (NAs) are the cornerstone of prophylaxis against Hepatitis B recurrence after liver transplantation (LT). Long-term use of IV HBIG has a high cost and the regular admission in the hospital is inconvenient. NAs alone does not always prevent HBsAg recurrence and can be nephrotoxic. SC HBIG can be self-administered. The optimal dose of SC HBIG without concomitant use of NAs has never been studied.

Aim

To study the optimal dose of SC HBIG without concomitant use of NAs.

Methods
This is an investigator driven, prospective trial. All patients receiving IV HBIG were switched to SC HBIG (Zutectra®) without NAs. The doses and interval of SC HBIG administration were aimed to keep HBsAg and HBV DNA undetectable. First dosage of Zutectra® was based on the guidelines of the manufacturer (< 75 kg: 500 IU/week; ≥ 75 kg: 1.000 IU/week). Thereafter, the titer of HBsAb was monitored regularly and if the titer was higher than the target levels at 2 successive occasions, a dose reduction was executed. In patients with low risk of recurrence (pts with undetectable HBV without antiviral therapy before LT, pts with acute liver failure and Delta hepatitis co-infected pts), the targeted titer was ≥ 100 IU/l and in the other patients ≥ 200 IU/l. The tolerance of the patients (IV or SC) was assessed by a specific questionnaire.

Results
44 patients were included in this trial. One patient preferred to switch again to IV HBIG, all the others (n=43) preferred SC HBIG, they did not report side effects and the compliance was 100%. The mean time after LT was 9 ± 6 years. Mean follow up time was 2 years ± 7 months. None of the patients had a relapse of HBsAg or HBV DNA. The mean HBsAb titer before the study was 332 ± 173 IU/l. The mean HBsAb titer at the end of the follow up period was 253 ± 121IU/l in the low risk group (n=14) and 281 ± 91IU/l in the high risk group (n=21). In 76% (n=33) doses reductions were possible. The total combined dose at the start was reduced from 118.000 IU /month to 68.135 IU/month. The median frequency of injections reduced from 1/w to 1/2w (range 2/w -1/3 w).

Conclusions
All except one patient preferred subcutaneous HBIG. SC HBIG without NAs had a 100% success rate in the long-term prevention of HBsAg and HBV DNA reappearance. Doses adaptation based on pre LT risk factors for HBV recurrence resulted in the vast majority of the pts in reduction of doses and/or prolongation of the interval and together with the self-administration and the no use of NAs induced a significant reduction of cost.

Authors
P. LEFEBVRE (1), S. FRANCQUE (2), F. LALLOYER (3), E. BAUGÉ (4), A. VERIJKEN (5), L. VAN GAAL (5), B. STAELS (3) / [1] INSERM, , France, Intitut Pasteur de Lille, [2] UZA, Edegem, Belgium, Department of Gastroenterology and Hepatology, University Hospital Antwerp, [3] Univ Lille 2, , France, Institut Pasteur de Lille, [4] Institut Pasteur, Lille, France, N/A, [5] UZA, Edegem, Belgium, Department of Endocrinology, Diabetology and Metabolism, University Hospital Antwerp,

Introduction
Non Alcoholic Fatty Liver Disease (NAFLD) is associated to obesity and predisposes to liver- and extrahepatic-related morbidities such as cirrhosis, hepatocarcinoma and cardiovascular diseases. A key step in NAFLD progression is fibrosis, whereby abnormal deposition of extra-cellular matrix (ECM) components occurs in the space of Disse.

Aim

Identifying molecular mechanisms leading to increased ECM deposition, and defining molecular pathways amenable to pharmacological manipulation would be decisive in fighting NAFLD progression.

Methods
A comparative analysis of liver transcriptome from NASH patients and murine models of nonalcoholic steatohepatitis (NASH) was carried out. Candidate genes whose expression was correlated to the severity of NAFLD (NASH CRN score) were selected. A gene whose expression increased in NASH/fibrosis and was normalized by the activation of hepatic peroxisome proliferator activated receptor alpha (PPARα) was identified. Its contribution to the fibrotic response was studied by gene deletion studies in mice.

Results
Comparative transcriptomic studies in NASH patients and murine models of NASH or fibrosis identified a response characteristic of hepatic stellate activation. A subset of genes was identified as a potential target of the TGFb, CTGF or the PPAR pathway and involved in ECM homeostasis using data mining strategies in ChIP-Seq databases and gene ontology term enrichment. Among them, dermatopontin (Dpt) was identified as a novel contributor to the fibrotic response. Gene deletion showed decreased ECM deposition in Dpt KO mice submitted to a pro-fibrotic insult (CCl4). In various models of rodent NASH, Dpt expression was lowered by PPARα activation. Furthermore, Dpt expression was normalized by bariatric surgery in human NASH patients.

Conclusions
Dpt is an important contributor to the fibrotic response and its expression is amenable to pharmacological control.

Invited lecture by Marc Hautekeete

Authors
L. DEVISSCHER (1), C. SCOTT (2), S. LEFERE (1), S. RAEVENS (1), E. BOGAERTS (1), A. PARIDAENS (1), X. VERHELST (1), A. GEERTS (1), M. GUILLIAMS (2), H. VAN VLIERBERGHE (1) / [1] Ghent University, Ghent, Belgium, Gastro-enterology and Hepatology, [2] Ghent University, Ghent, Belgium, Department of Biomedical Molecular Biology

Introduction
Kupffer cells (KCs) and liver infiltrating bone-marrow (BM) monocyte-derived macrophages (mo-Mf) have been denoted as key players in the pathogenesis of non-alcoholic steatohepatitis (NASH). Despite this, to date it has not been possible to accurately discriminate between these two populations due to the lack of specific markers. Additionally, KCs were believed to be derived solely from embryonic progenitors, which are maintained by self-renewal, however, it has recently been demonstrated that BM monocytes can differentiate into bona fide KCs (mo-KCs) when required. To date, it is also unclear if mo-KCs are present during NASH. Understanding which of the distinct macrophage populations are present and the roles they play in NASH is crucial to furthering our understanding of NASH pathogenesis and the development of novel therapies.

Aim

By using newly defined specific markers including Clec4F and Tim4 alongside BM chimeras, we aimed at accurately characterize the dynamic changes and origins of the distinct liver macrophage subsets in experimental-induced NASH and recovery.

Methods
Immunohistopathology and flow cytometry were used to determine the level of steatosis, steatohepatitis and recovery in methionine and choline deficient (MCD) diet fed mice. Flow cytometric analysis including the specific markers Clec4F and Tim4 and BM chimeras were applied to identify the distinct liver macrophage subsets and their origins.

Results
Mice fed the MCD diet for 8 weeks gradually developed severe steatohepatitis while replacement of MCD diet by normal chow resulted in full recovery after 4 weeks. Ly6CloClec4F-Tim4- infiltrated mo-Mf were observed from week 2 of MCD feeding, further increased during MCD feeding and returned to baseline during recovery. The absolute number of KCs, characterized as Ly6CloClec4F+Tim4+ cells, did not differ significantly between mice fed either MCD or the control diet (CD) over the duration of feeding or during recovery. However, an increased proportion of Ki-67+ proliferating KCs were observed in mice fed MCD diet compared with mice fed control diet. In line with this, we observed the development of a new population of Ly6CloClec4F+Tim4- KCs, only typically present in minor numbers in steady state, previously identified as mo-KCs. Mo-KCs developed from week 4 on the MCD diet and remained present during recovery. As lack of Tim4 expression is only a temporal marker of mo-KCs, with mo-KCs gradually gaining Tim4 expression after their differentiation, we utilised BM chimeras to both validate the presence of mo-KCs and determine their longevity. Interestingly, while these cells do develop from monocytes during MCD feeding and peak during initial recovery, they do not have the capacity to self-renew as their numbers are reduced by week 4 recovery.

Conclusions
Our findings demonstrate that during NASH pathogenesis and recovery the KC pool is maintained by proliferation and the differentiation of short-lived mo-KCs in the MCD diet model.

Authors
S. VAN HEES (1), S. BOURGEOIS (2), H. VAN VLIERBERGHE (3), T. SERSTÉ (4), P. MICHIELSEN (1), H. REYNAERT (5), J. HENRION (6), S. NEGRIN-DASTIS (7), L. LASSER (8), F. JANSSENS (9), G. ROBAEYS (10), P. STÄRKEL (11), C. MORENO (12), F. NEVENS (13), T. VANWOLLEGHEM (1) / [1] Antwerp University Hospital, Edegem, Belgium, Department of Gastroenterology and Hepatology, [2] ZNA Antwerpen, , Belgium, Department of Gastroenterology and Hepatology, [3] Ghent University Hospital, Ghent, Belgium, Department of Gastroenterology and Hepatology, [4] CHU Saint-Pierre, Brussels, Belgium, Department of Gastroenterology and Hepatology, [5] University Hospital Brussels, Vrije Universiteit Brussel, , Belgium, Department of Gastroenterology and Hepatology, [6] Centre Hospitalier de Jolimont-Lobbes., La Louvière, Belgium, Department of Gastroenterology and Hepatology, [7] Grand Hopital de Charleroi, Charleroi, Belgium, Department of Gastroenterology and Hepatology, [8] CHU Brugmann Brussels, Brussels, Belgium, Department of Gastroenterology and Hepatology, [9] Jessa Hospital, Hasselt, Belgium, Department of Gastroenterology and Hepatology, [10] ZOL, Genk, Belgium, Department of Gastroenterology and Hepatology, [11] Cliniques Universitaires St Luc, Brussels, Belgium, Department of Gastroenterology and Hepatology, [12] CUB Hôpital Erasme, Bruxelles, Belgium, Department of Gastroenterology, Hepatopancreatology and Digestive Oncology, [13] University Hospital Leuven, KU Leuven, , Belgium, Department of Gastroenterology and Hepatology

Introduction
High relapse rates are seen when Nucleos(t)ide Analogue (NA) treatment is discontinued after HBeAg seroconversion, but this might be accompanied by significant rates of HBsAg loss.

Aim

We studied whether NA stop after HBeAg seroconversion is associated with increased HBsAg loss rates in a Belgian, predominantly Caucasian cohort of Chronic Hepatitis B patients.

Methods
This is a pooled analysis including mono-infected, non-immune-suppressed patients treated with different NA for ≥3 months from 13 hospitals in Belgium. All patients were HBeAg positive at start of NA treatment. HBeAg seroconversion was defined as the loss of HBeAg and the appearance of anti-HBeAg on two time points ≥1 month apart. Follow-up time was calculated as time from baseline (date of HBeAg seroconversion) until HBsAg loss or end of Follow-up. A Cox regression model was used to determine predictive factors for HBsAg loss.

Results
A total of 326 NA treated patients (74.8% male; 63% Caucasian; 17% African) were included. Patients were treated for a median of 3.4 years using either lamivudine, adefovir, tenofovir, entecavir or telbivudine. Ninety six patients (median age at HBeAg seroconversion 38 years) showed HBeAg seroconversion after a median treatment duration of 15.5 months. NA were stopped in 57/96 patients after a median consolidation therapy of 7.5 months. HBsAg loss was observed in 10 patients on-treatment and 8 patients off-treatment. COX model revealed that stopping NA was significantly associated with a decreased chance of HBsAg loss (HR 0.263; p=0.006), whereas presence of cirrhosis at start-of-treatment (HR 0.478; p=0.147), age at HBeAg seroconversion (HR 0.795; p=0.380) and length of consolidation therapy (HR 1.608; p=0.409) were not. Results remained unchanged when adjusted for time to response. Stopped patients had a longer follow-up time after HBeAg seroconversion (median 4 years vs. 2 years; p=0.003) and had less cirrhosis (40.6% vs 12.3%; p=0.001) compared to continuously treated patients. There was no difference in age at time of HBeAg seroconversion (median 43 vs 35 years; p=0.567).

Conclusions
Cessation of NA treatment post-HBeAg seroconversion was associated with a decreased chance of HBsAg loss. In addition, longer consolidation therapy had no significant effect on the chance of HBsAg loss.

Authors
A. MAROT (1), J. VIEIRA BARBOSA (2), A. DENYS (3), P. DELTENRE (2) / [1] Centre Hospitalier Universitaire Vaudois, Lausanne, Switzerland, Gastro-enterology, [2] Centre Hospitalier Universitaire Vaudois, Lausanne, Switzerland, Division of Gastroenterology and Hepatology, [3] Centre Hospitalier Universitaire Vaudois, Lausanne, Switzerland, Division of Radiology

Introduction
Chronic portal vein occlusion (PVO) can be associated with gastrointestinal bleeding (GIB), portal biliopathy or intestinal ischemia. Portal vein recanalisation (PVR) is a technique able to treat or prevent complications related to portal hypertension (PH) by addressing PVO itself. However, failure of PVR and stent thrombosis are challenging.

Aim

Identify factors associated with PVR failure and evaluate short and long-term stent patency in non-cirrhotic patients with chronic PVO.

Methods
The charts of patients with chronic PVO in which placement of a stent has been attempted and using a trans-hepatic approach were reviewed. Extension of occlusion was assessed by portography before PVR.

Results
15 patients were included (12 men, median age 49 years [95% CI: 39-57]). Indications for PVR were GIB (n=5), portal biliopathy (n=2), the need for reducing PH before surgery (n=5) and other reasons (n=3). A procoagulate state was identified in 36% and a local prothrombotic factor in 47%. Occlusion involved the main portal vein, either without (n=8) or with (n=7) the mesenteric and/or the splenic veins. Regarding the intra-hepatic extension of PVO, patients were classified into 3 groups: “type 1” with occlusion limited to the main portal vein (n=6), “type 2” with involvement of portal bifurcation and extension to segmental branches (n=7), and “type 3” with extension to distal branches (n=2). PVR was successful in 13 cases (87%). Failure of PVR occurred in 2 patients: one with type 2 and one with type 3 PVO. The second patient with type 3 developed stent thrombosis 24 hours after PVR. Overall, failure of PVR or stent occlusion within the first 24 hours occurred in 100% in patients with type 3 PVO vs. 8% in those with types 1 or 2 (p=0.002). One patient suffered from liver capsule perforation. The median follow-up was 38 months (95% CI: 12-60). Anticoagulation was given to 10 patients after PVR (77%, median duration: 105 days [95% CI: 57-1000]). In per-protocol analysis performed in the 13 patients in which PVR was feasible, the actuarial probability of stent permeability was 82% at 2 years (95% CI: 59-100, 100% vs. 60% in patients who received and who did not receive anticoagulation, respectively, p=0.1). Ninety percents of the patients had resolution of manifestations related to PH.

Conclusions
PVR is feasible in most patients with PVO unless there is no extension to distal branches. Most patients in which PVR was successful have a permeable stent at 2 years. Anticoagulation seems to prevent secondary thrombosis. PVR has a place in the management of complications related to PVO.

Authors
N. OPDEWEEGH (1), W. BUDTS (2), J. VAN CLEEMPUT (2), T. ROSKAMS (3), J. PIRENNE (4), B. MEYNS (5), F. NEVENS (6) / [1] University Hospital Leuven, KU Leuven, , Belgium, Department of Gastroenterology & Hepatology, [2] University Hospitals Leuven, KU Leuven, Leuven, Belgium, Department of Cardiology, [3] University Hospitals Leuven, KU Leuven, Leuven, Belgium, Department of Pathology, [4] University Hospitals Leuven, KU Leuven, Leuven, Belgium, Department of Abdominal transplant surgery, [5] University Hospitals Leuven, KU Leuven, Leuven, Belgium, Department of Cardiac surgery, [6] University Hospitals Leuven, KU Leuven, Leuven, Belgium, Departement of Gastroenterology & Hepatology

Introduction
The survival of patients with chronic heart disease has significantly improved, especially in children. Chronic elevated right heart pressure might provoke cardiac induced liver disease and finally cardiac cirrhosis, which is overall a rare condition.

Aim

We studied the prevalence of cardiac induced liver disease in patients with longstanding elevated right heart pressure.

Methods
The study population consists of 120 patients. The suspicion of a cardiac induced liver disease was based on lab tests and abdominal ultrasound and the risk of liver related mortality was assessed by the VAST-score (0-3). The first study group were 98 young adult patients who underwent a Fontan procedure during childhood. Mean time post-Fontan was 17 ± 6 years. The second study group contained of 22 patients of middle age with end staged cardiac disease who were possible candidates for heart transplantation and with suspicion of associated liver disease based on lab tests. The presence of cardiac cirrhosis in this group was investigated with transjugular liver biopsy.

Results
In the Fontan patients 4/98 (4%) needed a heart transplantation; 9/98 (9%) had a VAST-score ≥ 2. In the majority, the Fontan intervention was performed > 16 year before. None of these patients needed a combined heart-liver transplantation. In the second group, 9/22 (36%) received a heart transplantation; 8/22 (36%) had a VAST-score ≥ 2 and 4/22 (18 %) patients had histological proven cardiac cirrhosis. A combined heart-liver transplantation was necessary in 5/9 (55%).

Conclusions
Cardiac induced liver disease is not uncommon in patients with chronic elevated right heart pressure. It occured in 9% of our adult Fontan patients after long-term follow-up. In patients with end stage cardiac failure and disturbed liver test, the incidence of cardiac cirrhosis was 18% and combined heart-liver transplantation should be considered in these patients.