Diagnosis, prognosis, and treatment of varices


Active bleeding at endoscopy is not a prognostic factor in Child-Pugh B patients with variceal bleeding

Dominique Thabut1, Marika Rudler1, Nina Dib2, Nicolas C ar bonell7, Philippe Mathurin5, Faouzi Saliba6, Alain Mallet1, Julien Mossord1, Brigitte Bernard-Chabert4, Frederic Oberti2, Christophe Bureau3

1Hepatology ICU, AP-HP, hôpital Pitié-Salpêtriére, Paris, France; 2CHU Angers, Angers, France; 3CHU Toulouse, Toulouse, France; 4CHU Reims, Reims, France; 5CHU Lille, Lille, France; 6CHU Paul Brousse, Villejuif, France; 7AP-HP hôpital St Antoine, Paris, France

Background and aims: Active bleeding at endoscopy has been reported as a prognostic factor in the setting of variceal bleeding (VB) in cirrhotic patients. Therefore, an early placement of transjugular intrahepatic portosystemic shunt (TIPS) has been proposed in Child-Pugh B patients with active bleeding as well as in Child-Pugh C patients with VB. However, there are no strong data confirming that active bleeding at endoscopy is predictive of bad survival in CPB patients. The aim of this studywas to determine, in a cohort of consecutive patients included in a prospective multicenter French trial (Baveno IV trial), if active bleeding at endoscopy influences the prognosis of ChildPugh B patients with acute VB. Methods: Ancillary study of Baveno IV study including 365 cirrhotic patients with upper gastrointestinal bleeding. Cirrhotic patients with VB were included. Results: 245 patients presented with acute VB (male gender: 74%; age: 56±12 yrs; aetiology of cirrhosis: alcohol=82%, viral = 10%, other= 8%; oesophageal varices: 85%/gastric varices: 15%; MELD=14±6). Forty-four (18%) were Child-Pugh A, 107 (44%) were Child-Pugh B and 94 (38%) Child-Pugh C. 166 patients (68%) presented with active bleeding at endoscopy, and among them, 75 were Child-Pugh Overall actuarial probability of survival at Day-42 was 86 ± 2%, and was significantly better in Child-Pugh A/B than in Child-Pugh C patients (Child-Pugh A=100%, Child-Pugh B= 93 ± 2%, Child-Pugh C=71 ± 4%, p<0.0001). In Child-Pugh B patients, survival was excellent and did not differ between patients with or without active bleeding at endoscopy (92 ± 3% vs 93 ± 4%, p=0.73). In multivariate analysis, factors associated with worse survival in Child-Pugh B patients were the presence of hepatocellular carcinoma (p=0.004), infection (p=0.03) and hepatic encephalopathy (p=0.01) at inclusion. Conclusion: Active bleeding at endoscopy is not associated with a worse survival in Child-Pugh B patients with variceal bleeding. It is necessary to refine which Child-Pugh B patients should benefit from early TIPS placement. Whether early placement of TIPS is justified in patients with active bleeding remains to be proven.


Marika Rudler - Speaking and Teaching: Gilead Sciences, BMS, Gore, Eumedica

Philippe Mathurin - Board Membership: Schering-Plough, Janssen-Cilag, BMS, Gilead; Consulting: Roche, Bayer, Boehringer

Faouzi Saliba - Advisory Committees or Review Panels: Novartis, Roche, Genzyme; Grant/Research Support: Astellas; Speaking and Teaching: Schering

Plough, Gambro, MSD

Frederic Oberti - Speaking and Teaching: LFB, gore Christophe Bureau - Speaking and Teaching: Gore

The following people have nothing to disclose: Dominique Thabut, Nina Dib, Nicolas Carbonell, Alain Mallet, Julien Massard, Brigitte Bernard-Chabert


Early TIPs in patients with acute variceal bleeding: an external validation

Marika Rudler1, Philippe Cluzel2, Nadjib Hammoudi3, Hedi Benosman1, Thierry Poynard1, Dominique Thabut1;

1Gastroterology and Hepatology, Pitie-Salpéfrére Hospital, Paris' France; 2Radiology, Pitié-Salpêtriére Hospital, Paris, France; 3Cardiology, PitiéSalpêtrière Hospital, Paris, France

Background and aims Recent data suggest that early TIPS placement improved rebleeding and reduced one-year mortality in high-risk cirrhotic patients with variceal bleeding (i. e. Child-Pugh C cirrhosis or with Child-Pugh B cirrhosis and active bleeding). We aimed to perform an external validation of this therapeutic attitude in real-life setting. Methods We prospectively included all consecutive patients admitted in our ICU of Hepatology between March 2011 and February 2013 for variceal bleeding, who fulfilled the following inclusion criteria: Child-Pugh C 10-13 cirrhosis or Child-Pugh B with active bleeding at initial endoscopy; no prior history of cardiac decompensation; no advanced HCC. All patients were treated with vasoactive drugs and antibiotics, and band ligation was performed within the 6 hours after admission. TIPS placement was performed within 72 hours (Viatorr TIPS endoprothesis, diameter 8 mm, Gore). Patients were matched according to sex, age, Child-Pugh score, MELD score and cardiovascular risk factors to patients of our historical cohort hospitalized for variceal bleeding and who were not performed TIPS. Results 31/128 cirrhotic patients (male gender 77.4%, mean age 53.2 ± 9.0 years, MELD score 20.9 ± 6.9, Child-Pugh C: 77.4%, alcohol etiology: 77.4%; shock 52%; active bleeding at endoscopy: 68%) admitted for acute variceal bleeding between March 2011 and February 2013 (TIPS+group) were matched to 31 historical patients (TIPS-group). Clinical and biological characteristics were not different between the 2 groups. Early TIPS placement was performed in a mean delay of 39.0±30.4hrs. The mean portal-pressure gradient dropped from to 18.5±4.9 to 7.0±3.7 mmHg. Median follow up was 8.7 months (1-24.6). Rebleeding happened in 0/31 (0%) patient in TIPS+group vs 14/31 (45.2%) in TIPS-group (p<10-3). Actuarial 1-month, 3month and 6-month survival were not different in TIPS+ or TIPSgroups (83.9±0.07 vs 90.1 ±0.5%, p=0.45, 76.4±0.07 vs ±0.08%, p=0.97 and 68.1±0.1 vs 72.2±0.08%, p=0.19, respectively). Neither development of hepatic encephalopathy (50 vs 45.2%, p=0.71) nor acute cardiac failure was different in TIPS+ or TIPS- group (22.6% vs 6.5%, p=0.07). Seven out of 31 patients (22.6%) were transplanted within one year after variceal bleeding in TIPS+ group, as compared to 4/31 (12.9%) patients in TIPS-group (p=0.19). Conclusion We confirm that early-TIPS placement is effective in preventing rebleeding in high-risk patients with variceal bleeding. Survival was not statistically different between the 2 groups. This might be related to the high proportion of patients with Child-Pugh C cirrhosis in our cohort.


Marika Rudler - Speaking and Teaching: Gilead Sciences, BMS, Gore, Eumedica

Thierry Poynard - Advisory Committees or Review Panels: Merck; Grant/Research Support: BMS, Gilead; Stock Shareholder: Biopredictive

The following people have nothing to disclose: Philippe Cluzel, Nadjib Hammoudi, Hedi Benosman, Dominique Thabut


Thrombelastography decreases hemoderivates requirement before invasive procedures in cirrhotic patients with coagulation disorders. A randomized controlled trial

Marcello Bianchini1,Lesley De Pietri2, Tommaso Di Maira1, Nicola De Maria1, Bruno Begliomini2, Giorgio Enrico Gerunda3, Erica Villa1

1Liver Unit - Department of Internal Medicine, AO-U Policlinico di Modena, Modena, Italy; 2Anetshesiology I - Department of Surgery, AOU Policlinico di Modena, Modena, Italy; 3Transplantation Unit- Department of Surgery, AO-U Policlinico di Modena, Modena, Italy

Background and aims. Although cirrhosis is clearly characterized by a thrombophylic state, leading i. e. to portal vein thrombosis, bleeding is greatly feared in end-stage liver diseases. Therefore, hemoderivates are commonly used before invasive procedures in cirrhotics and the risk of bleeding is still assessed by some conventional coagulation tests (platelets count and INR). Thromboelastography (TEG) has been shown to be effective in detecting signs of hypo-hypercoagulability. Aim of this study was to evaluate the efficacy of TEG before invasive procedure as a guide for hemoderivates transfusion in cirrhotics. Methods. Patients with cirrhosis and coagulation disorders (INR>1.8 and/or PLTs <50x10-3/mmc) undergoing invasive procedures were eligible for this controlled trial randomized by blocks of 15.Exclusion criteria were: ongoing bleeding, thrombotic events, anticoagulant or antiaggregant medications, sepsis and renal replacement therapy. Patients were randomly allocated either to TEG group (TEG-G), receiving fresh frozenplasma (FFP 10 ml/kg) in case of R>40 mm and/or platelets (PLTs 1 unit/10kg) for MA<30 mm before procedure, or to PerProtocol Group (PPG), receiving PLTs and/or FFP before procedure according to internal guidelines. Transfusion requirement, side effects, and related costs were recorded. Results. Forty patients were enrolled so far, 22 in the TEG-G and 18 in the PPG. No differences at baseline parameters were found to be different, in particular age (57.5±810.5 vs 61.2±12.3), INR values (2.07±0.67 vs 2.25±0.78) or platelets level (60.5±36.1 vs 66.1 ±45.8x103/mmc), neither in the type of procedures performed (p>0.05). No post-procedural bleeding or complications occurred in both groups. As expected, all subjects in the PPG received transfusions as compared to 5 in the TEG-G (100% vs 22.7%, p=0.000). In the PPG 13 patients (72.2%) required FFP, 3 (16.7%) PLTs, and 2 (5%) both PLTs and FFP. In the TEG-G 1 FFP transfusion has been performed (4.5%), 1 needed PLTs (4.5%), 3 both PLTs and FFP (13.6%). A transfusionrelated reaction to FFP was recorded in the PPG, none in the TEG. Mean transfusion cost for the TEG-G was 149.6 USD/patient (including TEG cost), 318.9 USD/patient for the PPG (p=0.000). Conclusion. TEG is safe and effective directing hemoderivates requirements before invasive procedures in cirrhotics with coagulation disorders. The use of TEG as a guide for transfusion before invasive procedure can reduce transfusion requirement, risk of transfusion-related side effects, and medical costs.


The following people have nothing to disclose: Marcello Bianchini, Lesley De Pietri, Tommaso Di Maira, Nicola De Maria, Bruno Begliomini, Giorgio Enrico Gerunda, Erica Villa


Spleen Stiffness Measurement for the Detection of Esophageal Varices: Systematic Review and Metaanalysis of Diagnostic Accuracy

Siddharth Singh1, John Eaton1, M. Hassan Murad2, Hironori Tanaka3, Hiroko lijima3, Jayant A. Talwalkar1

1Gastroenterology and Hepatology, Mayo Clinic, Rochester, MN; 2Knowledge and Evaluation Research Unit, Mayo Clinic, Rochester, MN; 3Division of Hepatobiliary and Pancreatic Diseases, Department of Internal Medicine, Hyogo College of Medicine, Hyogo, Japan

Background: Spleen stiffness measurement (SSM) is a promising non-invasive alternative to esophagogastroduodenoscopy (EGD) for the diagnosis of esophageal varices (EV) in cirrhosis. However, its overall diagnostic accuracy in various clinical settings is unknown. Aims: We conducted a systematic review and meta-analysis of diagnostic accuracy studies comparing SSM with EGD for the diagnosis of any and clinically significant EV in patients with chronic liver disease. Methods: Through a systematic search of bibliographic databases, conference proceedings and contact with authors, we identified 12 studies (1497 patients, 735 patients with EV) which reported the diagnostic accuracy of SSM using elastography imaging compared with EGD for presence of (i) any and/or (ii) clinically significant EV in adults with chronic liver disease. Using meta-analysis, we combined measures of test performance of individual studies. Results: The pooled estimates for diagnostic accuracy of SSM for presence of any EV were: sensitivity 78% (95% confidence intervals [CI], 75%-81%), specificity 76% (95% Cl, 72%-79%), positive likelihood ratio (LR) 3.4 (95% CI, 2.3-4.9), negative lR 0.2 (95% Cl, 0.1-0.4) and diagnostic odds ratio (〇R) 19.3 (95% Cl, 7.5-49.8). Studies conducted in Asian populations showed a significantly better diagnostic performance of SSM, as compared to the Western population; likewise, real-time tissue elastography appeared to be significantly superior to transient elastography and acoustic radiation pulse force imaging for SSM. On meta-analysis of 9 studies examining clinically significant EV, the pooled estimates for diagnosis were: sensitivity 81% (95% CI, 76%-86%), specificity 66% (95% CI, 61%69%), positive LR 2.5 (95% CI, 1.7-3.9), negative LR 0.2 (95% CI, 0.1-0.5) and diagnostic OR 12.6 (95% CI 5.5-28.7). The mean area under receiver operator curve for diagnosis of any and clinically significant EV were 0.86 (95% CI, 0.79-0.93) and 0.80 (95% CI, 0.75-0.85), respectively. There was significant heterogeneity observed, partly explained by elastography technique and location of study. The performance was similar on using the fitted bivariate regression model as well as on limiting analysis to patients with compensated cirrhosis. Conclusions: Based on the systematic review and meta-analysis, we observed that the current techniques of SSM remain sub-optimal for use in clinical practice to screen patients for EV.


Jayant A. Talwalkar - Consulting: Lumena; Grant/Research Support: Intercept, Salix, Gilead

The following people have nothing to disclose: Siddharth Singh, John Eaton, M. Hassan Murad, Hironori Tanaka, Hiroko lijima


Statins are Associated with a Decreased Risk of Variceal Bleeding in Compensated Cirrhosis

Eric S. Orman1, Christopher Martin1, Michael D. Kappelman2, Paul H. Hayashi1,Millie D. Long1

1Department of Medicine, Division of Gastroenterology & Hepatology, University of North C ar olina, Chapel Hill, NC; 2Department of Pediatrics, Division of Pediatric Gastroenterology, University of North Carolina, Chapel Hill, NC

BACKGROUND: Statins have been shown to reduce portal pressure in patients with cirrhosis and portal hypertension, but the effects of statins on clinical outcomes have not been described. We examined the relationship between statin use and variceal bleeding in patients with compensated cirrhosis. METHODS: We performed a nested case-control study using administrative data from the IMS LifeLink® Information AssetsHealth Plan Claims Database from 1997 to 2011.All patients aged 18-63 with at least 2 years of continuous health plan enrollment and 2 separate claims for cirrhosis were identified. We excluded patients with diagnoses, procedures, or medications indicative of hepatic decompensation (ascites, encephalopathy, variceal bleeding, or other related conditions) during the first year following the cirrhosis diagnosis. After this 1-year run-in period, we identified cases with an inpatient diagnosis of variceal bleeding; these were matched to 5 cirrhotic controls without variceal bleeding on age, gender, region, and calendar time. Exposure to any statin medication was ascertained in the 6-month period prior to the outcome date for cases (or index date for matched controls). Conditional logistic regression was used to examine the association of statin use with variceal bleeding, accounting for potential confounders including health care utilization. Exposure to hydrochlorothiazide (HCTZ) was also examined to control for any potential healthy user bias. RESULTS: 27, 146 patients with cirrhosis and adequate follow-up were identified, of which 16, 021 were excluded because of codes for decompensation. In the cohort of 11, 125 compensated patients, 214 cases with variceal bleeding were identified and matched to 1, 063 controls without variceal bleeding. Median age was 54 years (interquartile range 48-58), and 66% were male. Five cases (2.3%) were exposed to statins, compared to 111(10.4%) controls (odds ratio [〇R] 0.20; 95% confidence interval [Cl], 0.08-0.49). After adjusting for potential confounders, statin use remainedprotective against variceal bleeding (〇R 0.16; 95% CI 0.040.58). Sensitivity analysis using alternate definitions of variceal bleeding and statin exposure demonstrated similar effect estimates. In comparison, there was no significant association between variceal bleeding and HCTZ (〇R 0.85; 95% CI, 0.54- 35; adjusted OR 1.01; 95% Cl, 0.45-2.28). CONCLUSION: Statin use is associated with a decreased risk of variceal bleeding in patients with compensated cirrhosis. Randomized trials of statins in patients with chronic liver disease are needed to establish the efficacy of these medications for complications of cirrhosis and portal hypertension.


Michael D. Kappelman - Consulting: GSK, Cubist; Stock Shareholder: GSK, Johnson and Johnson, Merck

The following people have nothing to disclose: Eric S. Orman, Christopher Martin, Paul H. Hayashi, Millie D. Long


A prospective, double-blind, randomized placebo-controlled trial of carvedilol for early primary prophylaxis of esophageal varices in cirrhosis

Ankit Bhardwaj1, Chandan K. Kedarisetty2, Manoj Kumar2, Shiv K. Sarin2

1Research,Institute of Liver and Biliary Sciences, New Delhi, India; 2Hepatology, Institute of Liver and Biliary Sciences, New Delhi, India

Background & Aims: Gastro-esophageal variceal (GEV) hemorrhage is a major complication of portal hypertension resulting from cirrhosis of the liver. The risk of bleeding from small (≤5 mm) varices is 7% at 2 years. Median rate of progression to larger varices is 5-23% per year. Carvedilol, a non-cardio selective vasodilating beta-blocker, has been claimed to be as effective in preventing variceal bleeding as EVL. We compared the efficacy and safety of carvedilol against placebo in for prevention of growth of small to large varices. Methods: 175 consecutive patients with cirrhosis and portal hypertension with small esophageal varices (<5 mm in size and no red color signs) were prospectively enrolled from November 2010 to December 2012.88 patients received carvedilol (Group A) and 87 patients received matched placebo (Group B). Upper gastrointestinal endoscopy was done at baseline, 6 and 12 months. Fibroscan and HVPG were done at baseline and at 1 year of follow-up. The primary endpoint of the study was prevention of development of large (>5 mm) varices. The secondary endpoints were adverse events and mortality in both groups. Results: 63 patients in Group A and 69 patients in Group B completed 1 year follow up. The mean CTP and MELD in the carvedilol arm was 7.31 ± 2.1 and 13.65 ± 5.4 and in placebo arm, 7.48 ± 2.2 and 14.2 ± 5.28 respectively. The mean carvedilol dose was 11.923±2.05 mg/day and target heart rate achieved in Group A was 58±3 beats per minute. Baseline HVPG in Group A and B were 15.3 ± 3.9 and 16.06 ± 5.7 respectively. After 1 year of treatment, there was no statistical difference in reduction of HVPG in the two groups (p0.224). There was no correlation between change in HVPG and fibroscan in carvedilol and placebo group after the period of 1 year (p=0.415 versus p=0.799). Twelve (19%) patients in Group A and 14 (20.1%) in Group B developed large esophageal varices (p=ns). The actuarial probability of prevention of development of large esophageal varices at 30 month was also not different between the groups (p- 0.218). The kappa index for assessment of variceal size was <5%. There were no differences in the adverse events reported (7.9% in Group A versus 1.4% in Group B, p=0.109). No patient bled from varices in either group. Conclusion: This is the first RCT (NCT01196507) showing that while carvedilol was well tolerated, it was not effective in preventing growth of small to large esophageal varices. The drug was also not effective in achieving significant reduction in HVPG in this group of patients at 12 months.


The following people have nothing to disclose: Ankit Bhardwaj, Chandan K. Kedarisetty, Manoj Kumar, Shiv K. Sarin