Meeting Reports
Portal Hypertension and Variceal Bleeding: An AASLD Single Topic Symposium1

Norman D. Grace¹, Roberto J. Groszmann², Guadalupe Garcia-Tsao², Andrew K. Burroughs³, Luigi Pagliaro⁴, Robert W. Makuch⁵, Jaime Bosch⁶, Gregory V. Stiegmann⁷, J. Michael Henderson⁸, Roberto de Franchis⁹, Judith L. Wagner¹⁰, Harold O. Conn¹¹, and Juan Rodes⁶

From the ¹ Division of Gastroenterology, Faulkner Hospital and Tufts University School of Medicine, Boston, MA; ² Section of Digestive Diseases, West Haven, CT and Yale University School of Medicine, New Haven, CT; ³ Division of Liver Transplantation and Hepatobiliary Medicine, Royal Free Hospital, London, England; ⁴ Universita degli Studi di Palermo, Istituto di Medicina e Pneumologia, and Ospedale V. Cervello, Palermo, Italy; ⁵ Division of Biostatistics, Yale University School of Medicine, New Haven, CT; ⁶ Hepatic Hemodynamic Laboratory, Liver Unit, Department of Medicine, Hospital Clinic i Provincial, University of Barcelona, Barcelona, Spain; ⁷ Department of General Surgery, University of Colorado Health Sciences Center, Denver, CO; ⁸ Department of General Surgery, The Cleveland Clinic Foundation, Cleveland, OH; ⁹ Servizio di Gastroenterologie ed Endoscopia Digestiva, Universita Degli Studi di Milano, Milan, Italy; ¹⁰ Section of Health Services Evaluation, Mayo Clinic, Rochester, MN; and the ¹¹ Department of Medicine, Yale University School of Medicine, New Haven, CT.

Introduction

In June 1996, the American Association for the Study of Liver Diseases sponsored a single topic workshop combining a two-day symposium on liver microcirculation in health and diseases¹ followed by a two day consensus workshop on portal hypertension and variceal bleeding. The goal of the combined conference was to identify areas of critical importance in the understanding and treatment of portal hypertension and to foster future collaborative research projects.

The portal hypertension-variceal bleeding conference consisted of eight panel discussions, each highlighting a specific topic related to evaluation and treatment of portal hypertension. The chair of each panel was charged to summarize the current state of knowledge in the field and to suggest areas for future investigation. In addition, there were three invited lecturers on specific areas of interest. This report will summarize the conclusions of each of the panels. In comparing the summary statements from the differentpanels, there may be differences in emphasis, definitions of endpoints, or choices for therapy. These differences are a reflection of the state of the field where areas of disagreement exist.

ISSUES INVOLVING THE NATURAL HISTORY OF PORTAL HYPERTENSION

Garcia-Tsao G (Chair), Blei A, de Franchis R, Dudley F, Lebrec D, and Pagliaro L

Portal hypertension is usually initiated by conditions that increase resistance to portal blood flow, most frequently cirrhosis. Portal hypertension leads to the development of portosystemic collaterals of which the most clinically significant, because of their tendency for hemorrhage, are those that form gastroesophageal varices. Variceal hemorrhage has a high mortality and constitutes a life-threatening complication of portal hypertension.

Studies that have contributed most to our current knowledge of the natural history of portal hypertension are those leading to the establishment of a portal pressure threshold for the development of varices^2-4 as well as studies analyzing the course of patients after variceal hemorrhage.^5-7 More recently, prospective studies identifying risk factors for variceal hemorrhage⁸ and studies analyzing the characteristics and course of untreated cirrhoticpatients^9-11 have advanced our knowledge further and have led to the proposal of some practical recommendations.

Development/Growth of Varices

Any patient with cirrhosis or chronic liver disease is at risk of developing esophageal varices as long as a minimal portal pressure (portosystemic gradient) threshold of 10 to 12 mmHg isreached.^2-4

Patients with cirrhosis should be screened endoscopically for the presence of varices. The prevalence of varices in cirrhotics is proportional to the severity of liver disease.⁹ Therefore,those who are Child’s Class A should be screened if and when thereis clinical evidence of portal hypertension, e.g., a low plateletcount (<140,000), an enlarged portal vein diameter (>13 mm), or evidence of collateral circulation on ultrasound. Cirrhotic patients who are Child’s Class B or C at the time of diagnosis of cirrhosis should be screened for varices as soon as the diagnosis is made. The rate of growth of varices in cirrhotic patients is also proportional to the severity of liver disease.^9,10 Therefore, patients who have no varices on screening endoscopyshould be rescreened every 2 years if their liver function isstable or every year if there are signs of liver function deterioration. Because the development of large varices is greater in patients with small varices on initial endoscopy compared with patients with no varices,¹⁰ patients who have small varices on screening endoscopy should be rescreened every year.

In patients with chronic liver disease which has not yet progressed to cirrhosis, such as chronic viral hepatitis, primary biliary cirrhosis, and primary sclerosing cholangitis, neither the prevalence of varices nor the proportion of patients who develop a portal pressure gradient >12 mmHg has been established. Based on our current knowledge, screening for varices in this population is only indicated if there is clinical suspicion of cirrhosis. It is also unclear if any particular cause of cirrhosis or chronic liver disease is associated with a higher risk of development/growth of varices, and this remains to be established. Another issue that requires further study is the role of sequential portal pressure measurements and their timing in cirrhotic patients and in those with chronic liver disease.

Development of Variceal Hemorrhage

The development of variceal hemorrhage is variable and depends on the presence of risk factors. Three risk factors for first variceal hemorrhage have been established: size of varices, presence of red signs on varices, and the severity of liver disease.⁸ The rebleeding risk is very high in survivors of a first episode of variceal hemorrhage. It reaches a peak in the first week following the index bleed but remains high until the second or third month.^5,6 Only the severity of liver disease has been recognized as a risk factor for both early (within 6 weeks) and late rebleeding.^5,9

Given that only about a third of patients that present with variceal hemorrhage have risk factors predictive of hemorrhage,⁸ it is necessary to define new predictive factors that, used in combination with recognized factors, will lead to a more powerful prediction of variceal hemorrhage. It will also be important to determine if a patient who has not rebled for more than three months after an index hemorrhage is a baseline risk and should be managed as a patient who has never bled from varices.

PROBLEM AREAS

Burroughs AK (Chair), Boyer T, Grace ND, Groszmann RJ, Pagliaro L, and Stiegmann GV

This session was dedicated to discussing problem areas in portal hypertension as highlighted in previous consensus conferences of Baveno I and Baveno II.^12-14

Active Bleeding from Esophageal Varices at First Endoscopy

Oozing or spurting were felt to be indicative of active bleeding. However, there was no uniform opinion regarding the prognostic significance of active bleeding in relation to early rebleeding from varices within 48 hours or within 5 days, or death within6 weeks. Moreover, there was no consensus as to whether the prognostic significance might vary with the timing of endoscopy or the increasing use of vasoactive drugs before the diagnostic endoscopy. However, if a vasoactive drug was used before endoscopy, then seeing active bleeding was considered a failure of the trial drug.

The disparate views on this apparently fundamental concept of the prognosis of active bleeding was surprising, considering that in peptic ulcer bleeding, active bleeding at endoscopy has been a well known prognostic factor for early rebleeding and death. Thus, data need to be collected on the clinical relevance of active bleeding, particularly as its indication for the use of endoscopic therapy, or as a failure criterion for a vasoactive drug given before endoscopy.

Clinically Significant Bleeding

Based on a consensus at the Baveno II meeting,^12,13 clinically significant bleeding was defined as a transfusion requirementof 2 units of blood or more within 24 hours of time zero (admission to hospital; or if in hospital, the time of manifestation of bleeding), together with the following at time zero, systolic blood pressure <100 mm Hg, or a postural change >20 mm Hg, and/or pulse rate >100/min. Most investigators used this definition to determine a change in therapy but only a few felt it was useful as a prognostic indicator.

Paradoxically, when the conference participants were asked when a bleeding episode was counted as a real event, the evaluation of bleeding varied considerably. Clearly this disparity would make it very difficult to be sure that bleeding and rebleeding events are arithmetically summed in the same way in different centers and more importantly in randomized clinical studies.

Failure to Control Acute Bleeding

The Baveno II definition for failure to control acute bleeding within 6 hours requires any of the following:

1.	Four unit blood transfusion or more
2.	Inability to increase systolic blood pressure by 20 mm Hg from baseline or to 70 mm Hg or more
3.	Inability to reduce pulse rate <100/min or a reduction of 20/min from baseline pulse rate13

Most participants used this definition to determine a change in therapy but not as a prognostic indicator.

The failure to control acute bleeding after 6 hours requires any of the following:

1.	Hematemesis
2.	Reduction >20 mm Hg systolic blood pressure from 6 hour point
3.	Increase pulse rate >20 beats/min from 6 hour point on 2 consecutive readings an hour apart
4.	Transfusion of 2 units of blood or more to increase hematocrit >27% or hemoglobin >9 g/dL13

When asked about the efficacy of acute treatment, most participants replied that an effective treatment should result in less blood transfusion. The consensus was to recommend that blood transfusion differences should always be considered in the evaluation of treatment effects. To ensure the comparisons are more precise, the indications for transfusion should be stated.

There was agreement in the recording of the following clinical descriptors for patients admitted with portal hypertensive bleeding: bilirubin, prothrombin time, albumin, ascites, hepatic encephalopathy, creatinine, hemoglobin or hematocrit, platelet count, blood pressure, pulse rate, previous bleeding, time intervals, manifestation of bleeding to admission and admission to endoscopy, presentation with hematemesis, and alcoholic etiology.

Rebleeding. To define the time at which rebleeding occurs, it is important to define the duration of the bleeding episode. Clinically significant rebleeds should be evaluated as a separate end point, and all rebleeding regardless of severity should be counted in evaluating rebleeding.¹³

The evaluation of rebleeding has not been uniform, in that different sources and severity of rebleeding are counted as events. Definitions vary from defining rebleeding only if it was related to portal hypertension, to any hematemesis and/or melena.

Most panelists agreed with the Baveno II definition¹² of a rebleeding index: number of rebleeds +1, divided by the number of months of follow up, gives a measure of distribution in time and permits statistical analysis.¹³

Hemodynamic Indices and Monitoring

Recent studies have shown the value of target reduction in portal pressure in determining the therapeutic efficacy of drugtherapy.^15-17 The panel recommended that the hemodynamic measurement of choice is the HVPG. Once pharmacologic therapy is initiated, follow-up measurements should be made by 3 months with a uniform time interval within centers. Target reduction in HVPG should be to 12 mm Hg or a reduction of at least 20% from the baseline HVPG. Ideally, these measurements should be performed in all patients being treated for prevention of first variceal bleeding or rebleeding.

DESIGN OF CLINICAL TRIALS

Pagliaro L (Chair), Burroughs AK, Everhart JE, Grace ND, and Makuch R

The design of RCTs is governed by an evolving set of general guidelines. To draw suggestions for tailoring the application of these guidelines to RCTs of treatments for portal hypertension in cirrhosis, eight points of special interest were discussed. The following is a summary of the points discussed, of the background justifying their selection, and of the final statements reached in the general session.

Nonrandomized/Pilot Trials

There was agreement that nonrandomized/pilot studies are acceptable in the preliminary exploration of the efficacy of new treatments. In addition, some participants suggested that nonrandomized studies may be acceptable for certain specific situations such as the use of TIPS or other emergency procedures in uncontrolled variceal bleeding.

Nonrandomized trials to establish efficacy may be justified in special situations, if certain requirements are fulfilled: 1) No alternative treatments available to be used as a control; 2) Uniformly unfavorable outcome of the clinical condition (treated or untreated); 3) Low side effect rate and severity of the experimental treatment; 4) Expected therapeutic benefit large enough to make the results of the nonrandomized trial unambiguous; and 5) Scientific rationale sufficiently strong to support a positive result of the nonrandomized trial.¹⁸

Use of Placebo

A placebo is often used as comparator in RCTs of new treatments, even when standard active treatments of established efficacy are available to be used in controls.^19,20 The participants agreed that a placebo or nonactive controltreatment was not justified: 1) When effective active controltreatments are available; 2) In RCTs for control of active bleeding;3) For prevention of first bleeding in patients with medium-sizedor large varices; and 4) For prevention of rebleeding.

However, all responders favored a trial design comparing the association of 2 active treatments with the association of one of the two plus a placebo (e.g., sclerotherapy + octreotide vs.sclerotherapy + placebo).

The panel accepted beta-blockers as the standard effective treatment for prevention of first bleeding in patients with medium-sized or large varices and beta-blockers or endoscopic ligation of varicesfor prevention of rebleeding.

Equivalence Trials

RCTs can be aimed to assess the equivalence in efficacy between a standard active treatment used as a comparator, and a new treatment, usually less invasive or costly.^21,22 Such equivalence trials usually require a sample size larger than trials aimed to assess whether a treatment is more efficacious than another.²²

However, as shown by a recently published trial claiming equivalence between octreotide and injection sclerotherapy for control of bleeding,²³ equivalence trials are often undersized and lack the ability to rule out treatment differences of moderate size.^24,25 It has been suggested that even undersized trials can be informative in the era of meta-analysis, because their results can be pooled to assess the efficacy of a treatment.²⁶

But this rationale was not deemed acceptable for planning an equivalence trial. Indeed, the participants agreed that whenever possible, in equivalence trials the sample size should be large enough to achieve the required power. Most tended to accept preliminaryundersized studies, particularly in clinical conditions whereit is difficult to gather a large sample size (e.g., active varicealbleeding).

Competing Outcomes

Classical survival analysis estimates the proportion of the total original sample remaining free of the end point of interest (usually bleeding in RCTs for portal hypertension) at any follow-up time. If the original sample is being reduced by a competing event (death from causes different from bleeding), the estimated proportion of the sample free of bleeding may be falsely high because nonbleedingevents are not counted as failures.²⁷ Therefore, a harmful treatment increasing the rate of nonbleeding death may appear falsely effective in reducing bleeding risk.

A majority of participants felt that the problem of competing outcomes did not distort trials of treatments for portal hypertensionsufficiently to make their conclusions fundamentally invalid.Some participants suggested that this problem could have affectedsome long-term trials (e.g., for prevention of first bleedingor rebleeding). The preferred statistical procedures to face theproblem were to: 1) use as the primary end point the occurrenceas a first event of either nonfatal bleeding or death, and combinethe 2 categories or 2) perform the usual actuarial estimate, i.e., censoring deaths without bleeding at any follow-up time. Both these procedures might be used to test their concordance.

Interaction Between Baseline Risk and Treatment Effect

There is increasing awareness that within a trial population, though defined by relatively stringent inclusion and exclusion criteria, there may be groups of patients in whom the treatment has a beneficial, neutral, or even a harmful effect. An important factor of the treatment effect is the baseline risk^28,29: when it exceeds a threshold harm or risk of the treatment, the net therapeutic result is beneficial; at low baseline risk, the rate of adverse events resulting from the treatment can make the net treatment effect neutral or even harmful (i.e., net therapeuticbenefit = baseline risk × therapeutic risk reduction  harm³⁰). For example, prophylactic sclerotherapy might be beneficial in patients with a baseline bleeding risk higher than 35% to 40% per year and ineffective or harmful at lower bleeding.^31,32

It was agreed that baseline risk, expected therapeutic risk reduction, and harm should be taken into account in interpreting RCTs, particularly when there is substantial patient heterogeneity.

Reporting Trials

There is evidence that trial results reported as reduction in relative risk or odds ratio are interpreted by clinicians more optimistically than the same results reported as measures of absolute risk reduction or number needed to be treated to prevent an adverseevent.^33-36

The number of patients that needed to be treated³⁷ was recommended as the most informative and understandable measure by several responders, but absolute risk reduction should also be given.

Trial Reports and Meta-Analysis

Structured reports are more informative and improve the amount and quality of information.^38,39 Guidelines providing checklists of items to be included in papers reporting clinical trials were recently published by two cooperative groups.^40,41 Several medical journals now require that authors submittingarticles reporting clinical trials comply with these guidelines.Incorporating data relevant for portal hypertension (e.g., bleedingrisk, liver dysfunction) in the structured model for reportingtrials in the field could be of help.

Meta-analysis (MA) could explore the treatment effects not evaluable in undersized individual trials; or in subgroups of patients, evaluable in each individual trial. MA with use of individual data (MAID) could be quite informative, particularly in the presence of substantial between-trial patient heterogeneity.^42-44 However,individual data from clinical trials are difficult to obtain,and relying on subsets of trials from which individual data areavailable can introduce a retrieval bias.

The participants agreed that a data set relevant for portal hypertension should be defined by consensus and recommended for reporting trials. Trial results in major prognostic subgroups should be reported. Some participants suggested caution in interpreting subgroup analyses of trial results and recommended to accept differences across subgroup responses only if the definition of subgroups is biologically plausible, if the hypothesis of difference preceded rather than followed the trial results, and if the trials are large and the differences important and consistent across trials.^45,46

Individual data from trials to perform MAID could be made available by: 1) standardized data set and credits to collaborative centers; 2) a specific software universally available; 3) a central data repository; 4) friendship (personal contact); and 5) registries and collaborative groups.

Scientific Integrity and Sponsored Trials. An increasing number of trials of drugs or devices, most often multicentric, are industry-sponsored and financially supported.Concern has been raised about the risk that promotional interestby the industry can affect the scientific content of the trials^47-49 or prevent their publication if negative.⁵⁰

The participants felt that this concern is justified, and a code of prescriptions is imperative. The panel suggested that: 1) all terms of the contract between industry and investigators should be defined before the trial starts; 2) the final analysis should be performed by statisticians and clinicians independent from the industry; and 3) the results should be published even when the trial is prematurely ended or negative. In addition, some participants suggested that all financial incentives and connections between researchers and the sponsoring industry should be disclosed.⁵¹

DATA ANALYSIS, COST ANALYSIS

Makuch RV and Wagner JL

Random Versus Fixed Effects for Meta-Analysis. MA uses statistical summaries, from published or unpublished sources, to provide a numerical consensus regarding issues inclinical medicine and epidemiology. Data from several studiescan produce a stronger conclusion (narrower confidence intervalsfor effect sizes, smaller P values) than can data of a singlestudy, although the credibility of the MA results depends on thequality of the study even more than on the sample sizes.

Of concern in a MA is the nature of the inference: is the generalization to be about the currently available studies and patients or about a potentially unlimited population of future studies and future patients? In the former case, fixed effects analyses are used whereas the broader generalization calls for random effects models. Working with the same data, random effects models generally yield conclusions of lesser statistical significance with wider confidence intervals than do fixed effects models: broad generalizations are less certain than narrower ones. DerSimonian and Laird⁵² were among the first to apply random effects models to MA. A good discussion of their approach and other random effects models for MA appears in Draper et al.⁵³

As discussed in Fleiss,⁵⁴ DerSimonian and Laird’s work did not resolve the issue about what MA method to choose, and controversy continues. Bailey⁵⁵ and DeMets⁵⁶ point out that what one doesdepends on the question: “Will the treatment, on average, havean effect?” (a question about the future) or “Has the treatmentproduced an effect in the studies being analyzed?” Meier⁵⁷ supports the consistent use of random effects models while Peto⁵⁸ continues to support the use of fixed effects models. Fleiss⁵⁴ agrees with Meier and says “The random effects model anticipates better than the fixed effects model the possibility that some studies not included in the analysis are under way, are about to be published, or perhaps have even already been published in an obscure journal, and that the results in some of the non-included studies are different from the results in most of the meta-analyzed studies.” In some of his many published MAs of medical topics, the late Dr. Thomas Chalmers included both sorts of analysis, each displayed in the same sort of diagram, but with the wider confidence intervals of the random effects model clearly visible.

Outcomes Research and Cost-Effectiveness Analysis

Increasingly, medical practices must show cost-effectiveness for acceptance in the new marketplace for health care. What do the purchasers of health care mean when they request information on outcomes or cost-effectiveness, and how should serious clinical researchers respond to this demand for new kinds of information?

Although clinical trials offer a unique opportunity for the collection of comparative cost and quality-of-life information associated with interventions, cost-effectiveness analysis is in some ways unsuited to the rigid design structure of a trial. As a method developed to assist public decision makers in making public investments, cost-effectiveness analysis is both comprehensive in its capture of monetary and health effects and uncertain in its measurement and valuation of such effects.^59,60 An essential element of carcinoembryonic antigen is the explicitrecognition of uncertainty about both costs and effects; thisrecognition requires formal exploration of the sensitivity ofcost-effectiveness results to alterations in assumptions aboutuncertain elements of cost and effect. Clinical trials, on theother hand, are designed to minimize uncertainty about the extentof the effect, but in so doing, they are subject to time censoringand other restrictions that limit their generalizability outsidethe confines of the experimental setting. Incorporating cost andquality-of-life measurement into the design of clinical trialscan reduce the uncertainty about certain effects but it is unlikelyto eliminate uncertainty altogether. How one goes about structuringa cost-effectiveness analysis with these limitations in mind isa new challenge for clinical researchers.

INTERPRETATION OF DATA FROM MULTIPLE TRIALS: META-ANALYSIS OR COMMON SENSE

Conn HO

MA, the process of combining the results of multiple, similar investigations, was devised and named by G. V. Glass.⁶¹ There are two types of MA: the first is a narrative overview; the second is a formal MA. The first technique was used until formal MA wasintroduced. Overviews describe the results of the individual studiesand reach a narrative conclusion, e.g., “These investigationsall appear to show that branched-chain amino acid (BCAA) therapyis not consistently effective in chronic hepatic encephalopathy(HE).” Overviews tend to be subjective, qualitative, and lessrigorous than formal MAs that tend to be more rigorous, more objective, quantitative analyses, e.g., “BCAA therapy is beneficial in the treatment of HE and is almost free of adverse effects (P < .01and < .5, respectively).” MA based on the data from individual patients from different randomized, controlled trials (RCT) is thought to be more reliable than the summaries of data from RCTs.⁶²

A well done MA is a complex, labor-intensive process that requires a carefully designed search strategy, establishment of the high quality of the many components of the RCTs, and meticulous standardized analysis and recording of many clinical and methodological aspects of the disorder and patients under analysis.⁴⁴ A MEDLINE search is the usual starting point for almost all MAs. It can be supplemented by searches of pharmaceutical and appropriate societal registers of RCTs.⁶³ The RCTs to be included must becarefully screened. The inclusion of nonpeer reviewed reports,such as abstracts, theses, chapters, workshops, and correspondence,appear to diminish the validity of the analysis. RCTs to be includedshould be truly randomized, should be analyzed on an intent-to-treatbasis, and the recording of adverse events should be blindly performed. Only the first treatment period from crossover trials can be included, and no RCT selected for inclusion can retrospectively be excluded. All details of the disease and its treatment must be recorded along with those patients lost-to-follow-up, and the reasons for loss or exclusion must be available to the analysts for blind, independent analysis.

Retrospective examination indicates that the five most critical criteria for a sound MA are:

1.	The existence of a detailed protocol
2.	Contact by the analysts with the primary investigators to obtain additional information and to clarify ambiguities
3.	Analysis according to intent-to-treat principle
4.	Sensitivity analysis
5.	Publication of the results of the MA in a peer-reviewed journal

Avoidance of bias of all sorts is essential, as is the need to identify all RCTs on the subject. Heterogeneity of results should be investigated by sensitivity analyses. A recently published analysis of five large, published MAs on important investigative subjects, by an experienced, dedicated analyst showed that these MAs contained many serious, often inexplicable flaws and biases,⁶⁴ reached conclusions that may not have been justified, reported heterogeneous results, managed data inconsistently, and omitted needed subgroup analyses.⁶⁵

MA represents an important statistical advance in the combining of data from similar trials. Because it is so simple in concept and appears so easy to conduct, we must redouble our efforts to perform objective, well-conceived analyses to avoid killing the goose that laid the golden eggs.

MANAGEMENT OF THE BLEEDING PATIENT: SUPPORTIVE CARE

Rodes J

Emergency therapy in patients bleeding from esophageal varices is essential to avoid hypovolemia, shock, and death. Therefore, the early medical measures to ensure an appropriate general vital support are the most important factors that may favorably influence the prognosis of these patients.⁶⁶ It is considered that portal hypertension patients with gastrointestinal bleeding must be urgently taken to the closest medical center available to prevent and/or to correct the systemic hemodynamic alterations. Peripheral venous catheterization with a large cannula for rapid perfusion of blood, blood derivatives, and fluids, as well as monitorization of central venous pressure are mandatory. The extraction of venous blood samples to analyze ABO blood type and to determine the prothrombin ratio, platelet count, hematocrit, and hemoglobin levels must be made as soon as possible. The placement of a nasogastric tube to empty gastric content lowers the risk of possible tracheobronchial aspirations, prepares the ground for the performance of diagnosis and/or therapeutic upper gastrointestinal endoscopy, and can also be useful in the evaluation of hemorrhage size. On the observation of important hypovolemic signs (systolic blood pressure < 90 mm Hg, tachycardia > 120 b/min, and signs of peripheral hypoperfusion) oxygen should be administered (4 L/min), as should rapid infusion of plasma volume expanders and blood transfusion. In patients with severe hemodynamic alterations and encephalopathy, orotracheal intubation and mechanical ventilation could be required. In those without hypovolemic signs, only standard intravenous hydration should be undertaken. It is essential to avoid over transfusion because it increases portal pressure and can, therefore, negativelyinfluence hemorrhage control. Upper gastrointestinal endoscopy,which should be carried out when the patient is in a stabilizedhemodynamic condition, allows diagnosis and decision of specifictreatment in around 90% of cases. In addition, it is importantto prevent clinical complications which are frequently seen inportal hypertension patients with gastrointestinal bleeding. Thisincludes the administration of lactulose or lactitol for the treatment of hepatic encephalopathy and oral nonabsorbable antibiotics (norfloxacin or others) during 5 days to reduce the incidence of severe bacterial infections produced by microorganisms of enteric origin.

PHARMACOLOGICAL THERAPY

Bosch J (Chair), Dudley F, Groszmann RJ, Lebrec D, Pomier-Layrargues G, and Planas R

Brief state of the art lectures provided an excellent update of the basic concepts underlying drug therapy for portal hypertension as a background for achieving consensus on where we are and what needs to be done regarding the pharmacological treatment of portalhypertension in three different clinical scenarios: the preventionof the first bleeding, the treatment of the acute bleeding episode,and the prevention of recurrent variceal bleeding. The proposedconsensus statements were derived from the answers to a questionnaire and were further elaborated within the panel.

The proposed consensus statements were presented with supporting evidence and subject to general discussion with the panelists and the audience. Consensus was reached in most aspects, but obviously not in every subject. Consensus statements are listed below.

Prevention of First Bleeding. 1) Pharmacological therapy is the established therapy for the prevention of first bleeding^11,42; 2) nonselective beta adrenergic blockers are the recommendeddrugs^31,67; and 3) the following needs to be done: 1) studies on preprimaryprophylaxis (prevention of varices); 2) RCTs with new drugs ordrug combinations^68,69; and 3) definition of factors that determine an optimal therapeuticresponse.

Treatment of Acute Variceal Bleeding. 1) Drug therapy is an established therapy for acute bleeding; 2) several drugs including terlipressin, vasopressin plus nitroglycerin, and somatostatin have been shown to be effective. More studies are required with octreotide¹¹ 3) wider availability of thesedrugs is desirable; and 4) drug therapy may be used as initialtreatment before sclerotherapy or ligation.^70,71

Prevention of Rebleeding. 1) If pharmacological treatment is chosen, nonselective beta blockers are the drugs of choice^11,72 combination of beta blockers with isosorbine-5-nononitrate may be more effective and needs further testing in RCTs^16,73; 2) assessment of the HVPG response is advised during pharmacologicaltherapy for the prevention of rebleeding.^16,73 3) The following needs to be done: 1) RCT testing new drugs or combinations versus beta blockers and versus banding ligation; 2) RCT with use of the concept of targeted therapy (in terms of fall in HVPG); and 3) new methods for noninvasive assessment of the effects of therapy.

Most Important Advances in Past 5 Years in the Field of Pharmacological Therapy. 1) Better understanding of the basic mechanisms leading to portal hypertension and bleeding; 2) demonstration that drug therapy is effective treatment; and 3) concept of combination therapy.

ENDOSCOPIC THERAPY

Stiegmann GV (Chair), de Franchis R, Laine L, Miller L, Plannas R, and Pomier-Layrargues G

Endoscopic therapy for esophageal varices has evolved considerably over the past three decades. Sclerotherapy emerged as a viable treatment in the 1970s and was evaluated in multiple controlled trials in the 1980s and early 1990s. Endoscopic band ligation was introduced in the late 1980s and has achieved wide application and confirmation of efficacy, particularly in the elective treatment setting.

General agreement was reached on defining failure of endoscopic treatment by uncontrolled bleeding (4-6 units transfusion or more) from varices or treatment (sclerotherapy/banding) induced ulcers despite repeated endoscopic treatment, balloon tamponade,terlipressin, somatostatin, or octreotide infusion. Consensuscould not be reached regarding the number of recurrent but controlled bleeding episodes which should define failure, nor was there agreement on a time frame which would distinguish acute as opposed to long-term failure. Many participants felt that two or three clinically significant (requiring blood transfusion) variceal bleeding episodes or more, despite two or three endoscopic treatments, should define failure of endoscopic treatment at any time following the index bleed.

The Role of Endoscopic Prophylaxis Against a First Variceal Bleed. Endoscopic sclerotherapy as prophylaxis against a first variceal bleed has been extensively studied.⁷⁴ In contrast to some early trials which showed benefit, more recent and larger studies have shown either no value or a deleterious effect of sclerotherapy.⁷⁴ There was overwhelming consensus that endoscopic sclerotherapy is definitely not indicated for prevention of a first variceal hemorrhage in cirrhotics. The value of band ligation as prophylaxis against a first bleed is currently under investigation.⁷⁵ Participants could not agree that future trials of prophylactic banding should be confined to cohorts of patients with large varices and greater risk of initial hemorrhage, as opposed to all patients with esophagogastric varices, because ability to predict which patients will experience a first bleed is not precise.⁸ There was consensus that patientsin the control arm of future trials of endoscopic prophylaxismust be treated with effective pharmacological therapy. Improvementin our current capability of predicting a first variceal hemorrhage,i.e., by incorporating other parameters (ultrasound, noninvasivevariceal pressure measurements) into existing prognostic indices,should help identify a subset of patients most likely to benefitfrom endoscopic intervention.

The Role of Endoscopic Treatment for Acute Variceal Bleeding. The optimal timing of endoscopic therapy for treatment of an acute bleeding episode remains unsettled. Ideally, the endoscopist who makes the diagnosis of bleeding varices should initiate treatment but may not be trained in sclerotherapy or band ligation. Moreover, there is little data to support the view that immediate endoscopic treatment is superior to that performed within the ensuing day. There was consensus that endoscopic measures are the first choice of treatment for control of acute variceal bleeding and that success in definitive control of bleeding should be achieved in from 80% to 90% of cases.⁷⁶ Some participants felt that endoscopic sclerotherapy was easier to perform in the setting of active variceal bleeding whereas others believed that band ligation was both easier to perform and safer. Based on the available data for patients treated for active variceal bleeding, participants agreed that endoscopic ligation and sclerotherapy are equally effective for treatment of active bleeding.^76-82 Future research should be directed toward randomized trials comparing endoscopic treatments (sclerotherapy or ligation) to vasoactive drugs or endoscopic treatment combined with vasoactive drugs as compared with endoscopic treatment alone.

The Role of Endoscopic Treatment for Prevention of Recurrent Bleeding. Endoscopic sclerotherapy has been shown effective for prevention of recurrent variceal bleeding when compared with observation or pharmacological treatments. In trials which compared sclerotherapy with shunt surgery, there was less rebleeding in the surgical arm but no consistent differences in mortality. Eight recent trials have compared sclerotherapy with band ligation for prevention of recurrent hemorrhage.^76-83 Based on a recent meta-analysis,⁷⁶ participants concluded that band ligation is the preferred endoscopic treatment for prevention of rebleeding from esophageal varices. Future randomized trials should attempt to determine the optimal method for prevention of recurrent variceal bleeding: endoscopic ligation, pharmacological therapy, combinations of endoscopic and pharmacologic therapy TIPS or shunt surgery.

Cost Efficacy and Outcomes for Endoscopic Treatment. Randomized trials suggest that endoscopic ligation is superior to sclerotherapy in control of bleeding, mortality, local complications, and number of sessions required to obliterate varices.⁷⁶ A preliminary report of a cost comparison of band ligation with sclerotherapy in patients without active hemorrhage found that band ligation was less expensive.⁸⁴ These data should translate into decreased cost and improved outcome for patients treated with band ligation. Participants agreed that the cost benefits and effectiveness of band ligation as compared with sclerotherapy and of endoscopic therapy as compared with other treatment methods have not been firmly established. A formal cost analysis of endoscopic treatment compared with nonendoscopic treatment is needed.

The Role of Endoscopic Ultrasonography in Patients With Esophageal Varices. Esophagogastroduodenoscopy is the current clinical standard for assessing esophageal varices and predicting risk of hemorrhage. Endoluminal sonography may be an accurate method for determining presence of varices and measuring variceal size and wall thickness.^85-88 In addition, endosonography may be useful to determine the presence or absence and size of para-esophageal veins which may have predictive value in terms of effectiveness of endoscopic therapy.^89,90 Participants did not believe there was an adequate body ofinformation to conclude that endosonography had defined valueat this time but did agree that serial endosonography may allowdetection of changes in size and wall thickness of varices whichcould result in more accurate prediction of the risk of hemorrhage.There was enthusiastic support for continued assessment of thevalue of this method for determining effects of other treatments(e.g., endoscopic, pharmacologic) and definition of the accuracyof endoluminal sonography for prediction of a first variceal hemorrhage.

Future Research.

Future research will need to address many new questions. Is the efficacy of combined endoscopic ligation and sclerotherapy, administered either synchronously or metachronously, superior to band ligation alone? Will addition of pharmacological agents to band ligation therapy improve outcomes? Do the new multiple fire ligation devices provide significant advantage to justify their increased cost? What is the most effective method for identifying patients who are at high risk for recurrent bleeding? Is endoscopic therapy the most cost effective intervention for prevention of recurrent variceal bleeding and if not, what is? Does the higher rate of recurrence of varices associated with endoscopic ligation offset the greater efficiency of initial eradication of varices by using this method? These and other emerging issues will occupy future consensus agendas. At this time, it is appropriate to conclude that endoscopic therapy is well established and effective treatment for bleeding esophageal varices. Subsequent improvements in methods will probably not be dramatic and will require carefully performed and executed randomized trials.

VARICEAL BLEEDINGSURGICAL THERAPY AND TIPS

Henderson JM (Chair), Blei A, Bosch J, Orloff MJ, Rikkers L, Sanyal AJ, and Sarfeh IJ

Liver transplantation and TIPS have dramatically altered the surgical approach to management of variceal bleeding in the 1990s. However, each of these therapies are, at the present time, limited in their availability or applicability, but have helped refocus some attention on the role of surgical decompression as an excellent way to stop variceal rebleeding. Future research should involve clinical trials comparing TIPS with endoscopic therapy and TIPS with surgical shunts.

Liver Transplantation. Liver transplant has significantly improved the outcome of patients with end-stage liver disease.⁹¹ One year survival greater than 80%, and 5-year survival in the 65% to 75% range are to be expected in the late 1990s. Availability of donor organs is the major limiting factor in wider application, leading to the need for prudent patient selection and wise use of this limited resource.⁹² The indication for use of liver transplant for patients with portal hypertension and variceal bleeding is end stage liver disease. Not every patient with variceal bleeding has end stage disease, so it requires full evaluation and documentation of a patient’s disease and its progression over time to reach a decision to transplant such patients. Approximately 25% of patients receiving transplants have variceal bleeding as a component of their end stage disease.^93-98

The panel recommended liver transplantation for the patient with portal hypertension and variceal bleeding with end stage liver disease. Selection criteria may vary by the disease. Patient and graft survival are not adversely affected by a prior surgical shunt, but operative time and blood requirement may be increased.^99-108

Surgical Shunts. Surgical shunts have received renewed interest in the 1990s because of their effectiveness in preventing rebleeding.^{96,98,109-114} Bleeding is controlled in 90% to 95% of patients. Data on surgical shunts in the 1990s indicate better outcomes than earlier reported series in the literature primarily because of better patient selection. Most recent series of surgical shunts are in Childs Class A and B patients.

Surgical variceal decompression can be achieved by 1) total portal systemic shunts, which divert all portal flow from theliver,^109,110 2) partial portal systemic shunts, which reduce portal hypertensionto 12 mm Hg, and maintain some portal flow in 80% of patients,^111,112,115 or 3) selective variceal decompression (distal splenorenal shunt) which decompresses the spleen and gastroesophageal varices, but maintains portal hypertension.^{96,98,113,114,116,117} Data indicate 90% control of bleeding with all types of surgical shunt. Encephalopathy and liver failure rates are dependent on underlying liver disease and the loss or maintenance of portal perfusion.

Devascularization procedures are used in the 1990s as an alternative surgical method for variceal bleeding. In Japan, experience has been extensive, and this method is widely used with mortality from rebleeding being 6%.¹¹⁸ Experience in the rest of the world has been less favorable with higher rebleeding, and mortality rates.¹¹⁹ It is acknowledged that these procedures may offer the only method for some patients in whom all other therapies have failed.

The panel felt that surgical shunts were highly effective in control of variceal bleeding in patients with portal hypertension. Surgical decompression is the preferred method to prevent rebleeding in patients with good liver function when endoscopic and/or pharmacologicaltherapy fails or is not applicable. In patients receiving a surgicalshunt (either selective or total), liver transplantation remainsa future option that is dictated by the patient’s course.

Transjugular Intrahepatic Portal Systemic Shunts. TIPS provides a less invasive method for decompressing varices which is pathophysiologically similar to total and/or partial surgical shunts.¹²⁰ The technical success rate of placing TIPS is 90%, with good initial control of bleeding.^121,122 The major problems with TIPS are the high rates of stenosis and/or thrombosis, and the increased incidence of encephalopathy.

TIPS stenosis occurs in 50% to 70% of patients within the first year,^123-129 requires a systematic follow-up with use of doppler/ultrasound and recatheterization for its identification and management, and is the major cause for TIPS rebleeding rates of 15% to 30%. In addition, the need for reintervention raises cost and is a factor in diminishing quality of life.

Encephalopathy following TIPS is reported at 30% increased incidence.^130-132 In approximately one quarter it is incapacitating, and the others can be managed with diet and medications. In some patients, accelerated liver failure is associated with the encephalopathy.

Controlled trials comparing TIPS with endoscopic therapy have shown significantly better control of bleeding with TIPS, butsignificantly lower encephalopathy with endoscopic therapy.^130,133-135 Meta-analysis of these trials in which both TIPS and sclerotherapywere used as primary therapy to prevent rebleeding, shows no difference in mortality.

The panel felt that: 1) TIPS is accepted therapy for bleeding varices not controlled by pharmacologic and endoscopic therapy in patients who are not surgical candidates; 2) prospective randomized trials are required to compare TIPS with surgical shunts in patients who have failed endoscopic/pharmacologic therapy; 3) TIPS dysfunction is the major problem limiting the use of TIPS; and 4) The mechanism of TIPS dysfunction, its detection, treatment, and prevention are research priorities.

Conclusions. The panel concluded that TIPS is an important new area in managing patients with portal hypertension with variceal bleeding, which requires extensive further clinical and basic study. Liver transplant is a major advance in managing patients with variceal bleeding and end stage liver disease. Surgical shunts offer excellent control of variceal bleeding and should be offered more liberally to good risk patients with rebleeding in spite of endoscopic and pharmacologic therapy.

PORTAL HYPERTENSIVE BLEEDING: GASTRIC VARICES, PORTAL HYPERTENSIVE GASTROPATHY

de Franchis R (Chair), Henderson JM, Laine L, Lebrec D, Miller L, Orloff MJ, Rikkers L, and Pomier-Layrargues G

Gastric Varices. Gastric varices (GV) can be found in the stomachs of patients with portal hypertension, either isolated or in conjunction with esophageal varices. However, there is great variability in the reported figures concerning the prevalence,^136-139 the incidence of bleeding and rebleeding,^139,140 and the fate of gastric varices after endoscopic sclerotherapy of esophageal varices.^{136,138,141-145} Plausible reasons for the discrepancies are (1) difficulty in the diagnosis, (2) differences in the criteria of classification, (3) differences in the stage of portal hypertension, the severity of cirrhosis, and the prevalence of previous sclerotherapy, and (4) differences in the cause of portal hypertension among studies. Several methods have been proposed to treat bleeding gastric varices. These include endoscopic sclerotherapy with conventional sclerosants,^146-149thrombin¹⁵⁰ and butyl-cyanoacrylate,^151,152 TIPS,¹⁵³ and surgery. However, no controlled studies have been published so far.

Diagnosis and Natural History. GV are usually assessed by endoscopy alone, the most widely used classifications by Sarin¹³⁸ and by the NIEC.¹⁵⁴ GV as a whole are fairly common in patients with portal hypertension. However, isolated fundal varices occur in a minority of patients. GV accounts for about 10% of all upper gastrointestinal bleeds in patients with portal hypertension. As a whole, gastric varices tend to bleed less than esophageal varices, but fundal varices tend to bleed and rebleed more frequently and more severely than esophageal varices.¹³⁰

Treatment. There are no data for the optimal protocol to treat GV. Once fundal GV bleed, early surgery or TIPS is advisable.

Directions for Future Research. Future studies should be aimed at assessing factors influencing the development of GV (including endoscopic treatment of esophageal varices) and at determining the natural history of GV (including risk of bleeding and rebleeding). Available therapeutic options (sclerotherapy, glue, TIPS, surgery) should be assessed by means of randomized trials of primary and secondary prophylaxis of bleeding.

Portal Hypertensive Gastropathy. Gastric mucosal changes have always been known to occur in patients with liver cirrhosis, but their relationship with portal hypertension has been recognized only recently.^155,156 Histologically, these lesions are characterized by dilatation of the capillaries and venules of the gastric mucosa,^157,158 although these changes are not specific, whereas mucosalinflammation is infrequent. Collectively, these changes are referredto as Portal Hypertensive Gastropathy (PHG). PHG should be distinguished from gastric antral vascular ectasia, a gastric mucosal lesion that can occur in patients with cirrhosis as well as in subjects without portal hypertension.¹⁵⁹

The prevalence of PHG in published studies varies between 7% and 98%^{155,156,160-165} reflecting differences in patient selection, the absence of a universally accepted classification, and the paucity of data on the specificity^156,160 and reproducibility^163,166,167 of existing classifications. The reproducibility of the NIEC classification¹⁵⁹ shows a fair to good degree of agreement between observers in assessing the elementary lesions of PHG.¹⁴⁸ In thesame study, the prevalence of PHG in cirrhotic patients was 67%,with a specificity of 83%. Evaluation of the relationship betweenPHG and other endoscopic and clinical features in cirrhosis hasgiven conflicting results.^{155-157,160,161,165} The evolution of PHG with time and following endoscopic therapy of esophageal varices is also variable among studies.^{155,161,165,168,169} The reported prevalence of acute bleeding from PHG varies from 4%,^165,170 to 38%¹⁶¹ of all acute bleeds in cirrhotic patients. After a first bleed from PHG, rebleeding appears to be very common, with reported figures of 62% and 75%.^170,171

Very few data exist on the treatment of acutely bleeding PHG. In uncontrolled studies, both retrospective¹⁷² and prospective,¹⁷³ emergency portacaval shunt appeared to be beneficial. Antisecretory drugs are ineffective.¹⁵⁵ Experimental data suggest that vasoactive drugs which decrease mucosal blood flow might be beneficial^174,175 to treat acute bleeds from PHG, but clinical studies are lacking. In the only existing randomized controlled trial, propranolol was more effective than placebo in preventing rebleeding from severe PHG,¹⁷¹ although as many as 50% of patients treated with propranolol in this study rebled during 2 years of follow-up.

The following points were agreed upon: 1) Diagnosis and natural history (PHG is assessed endoscopically; the most widely usedclassifications are those by McCormack¹⁵⁵ and the New ItalianEndoscopic Club¹⁵⁴; PHG is frequently observed in cirrhotics, usually in the presence of gastroesophageal varices; acute bleeding from PHG is relatively rare; and bleeding from PHG is less severe than bleeding from gastroesophageal varices); 2) treatment (pharmacotherapy is the treatment of choice for both acute and chronic bleeding from PHG; if drug therapy fails, decompression (surgery or TIPS) should be performed); and 3) directions for future research (future studies should be aimed at gaining a better understanding of the pathophysiology of PHG; and the New Italian Endoscopic Club classification of PHG¹⁵⁴ should be externally validated and used to determine the natural history, the risk of bleeding, and the impact of PHG bleeding on morbidity and mortality. The value of available therapies (pharmacotherapy, TIPS, decompressive surgery) in the prevention and treatment of bleeding and rebleeding should be assessed in a controlled fashion).

Conclusion

Better understanding of the principles involved in the design of clinical trials and analysis of data will produce high quality RCTs that may help resolve some of the issues raised in this conference. There was a strong sense for the need to use standardized definitions for the critical end points, i.e., bleeding or rebleeding episodes,treatment failures, etc. Issues of quality of life and cost effectiveness are becoming increasingly important concepts to include in trial design. Adherence to these goals may aid in reduction of some of the heterogeneity in RCTs, evaluation of treatment options, and allowance for more widely accepted conclusions for therapeutic outcomes.

Addendum

    Participating Panelists. Andres Blei, Northwestern University School of Medicine, Chicago, IL; Jaime Bosch, University of Barcelona, Barcelona, Spain; Thomas D. Boyer, Emory University School of Medicine, Atlanta, GA; Andrew K. Burroughs, Royal Free Hospital, London, England; Roberto deFranchis, Universita degli Studi di Milano, Milano, Italy; Frances J. Dudley, Monash Medical School, Melbourne, Australia; James Everhart, National Institute of Health, Bethesda, MD; Guadalupe Garcia-Tsao, Yale University School of Medicine, New Haven, CT; Norman D. Grace, Tufts University School of Medicine, Boston, MA; Roberto Groszmann, Yale University School of Medicine, New Haven, CT; J. Michael Henderson, The Cleveland Clinic Foundation, Cleveland, OH; Loren Laine, USC School of Medicine, Los Angeles, CA; Didier Lebrec, Hospital Beaujon, Paris, France; Robert W. Makuch, Yale University School of Medicine, New Haven, CT; Larry Miller, Temple University Medical School, Philadelphia, PA; Marshall Orloff, UC San Diego School of Medicine, San Diego, CA; Luigi Pagliaro, Universita degli Studi di Palermo, Palermo, Italy; Ramon Planas, Hospital Universitari Germans Trias i Pujol, Badelona, Spain; Gilles Pomier-Lagrargues, University of Montreal, Montreal, Canada; Layton Rikkers, University of Wisconsin School of Medicine, Madison, WI; Arun Sanyal, Medical College of VA, Richmond, VA; James Sarfeh, UC Irvine College of Medicine, Orange, CA; Mitchell Schiffman, Medical College of VA, Richmond, VA; Gregory Stiegmann, University of Colorado Health Science Center, Denver, CO; and Judith L. Wagner, Mayo Clinic, Rochester, MN.

Footnotes

Abbreviations: HVPG, hepatic venous pressure gradient; RCTs, randomized controlled trials; TIPS, transjugular intrahepatic portal systemic shunt; MA, meta-analysis; MAID, meta-analysis using individual data; CEA, carcinoembryonic antigen; GV, gastric varices; PHG, portal hypertensive gastropathy.

¹Sponsored by the American Association for the Study of Liver Disease

Received January 5, 1998; accepted April 22, 1998.

Address reprint requests to: Norman D. Grace, M.D., Faulkner Hospital, Division of Gastroenterology, 1153 Centre Street, Boston, MA 02130-3446. Fax: (617) 524-6599.

References

1. Shah V, Garcia-Gardena G, Sessa WC, Groszmann RJ. The hepatic circulation in health and disease: report of a single topic symposium. HEPATOLOGY 1998; 27: 279-288
2. Viallet A, Marleau D, Huet M, Martin F, Farley A, Villeneuve JP, Lavoie P. Hemodynamic evaluation of patients with intrahepatic portal hypertension. Relationship between bleeding varices and the portohepatic gradient. Gastroenterology 1975; 69: 1297-1300
3. Lebrec D, DeFleury P, Rueff B, Nahum H, Benhamou JP. Portal hypertension, size of esophageal varices and risks of gastrointestinal bleeding in alcoholic cirrhosis. Gastroenterology 1980; 79: 1139-1144
4. Garcia-Tsao G, Groszmann RJ, Fisher RL, Conn HO, Atterbury CE, Glickmann M. Portal pressure, presence of gastroesophageal varices and variceal bleeding. HEPATOLOGY 1985; 5: 419-424
5. Graham D, Smith JL. The course of patients after variceal hemorrhage. Gastroenterology 1981; 80: 800-809
6. Burroughs AK, Mezzanotte G, Phillips A, McCormick PA, McIntyre N. Cirrhotics with variceal hemorrhage: the importance of time interval between admission and the start of analysis for survival and rebleeding rates. HEPATOLOGY 1989; 9: 801-807
7. Nevens F, Bustami R, Scheys I, Lesaffre E, Fevery J. Variceal pressure is a factor predicting the risk of a first variceal bleeding: a prospective cohort study in cirrhotic patients. HEPATOLOGY 1998; 27: 15-19
8. The North Italian Endoscopic Club for the Study and Treatment of Esophageal Varices. Prediction of the first variceal hemorrhage in patients with cirrhosis of the liver and esophageal varices: a prospective multicenter study. N Engl J Med 1988; 319: 983-989
9. Pagliaro L, D’Amico G, Pasta L, Politi F, Vizzini G, Traina M, Madonia S, et al. Portal hypertension in cirrhosis: natural history. In: Bosch J, Groszmann RJ, eds. Portal Hypertension: Pathophysiology and Treatment. Oxford: Blackwell Scientific Publications, 1994:72-92.
10. Cales P, Desmorat H, Vinel JP, Caucanas JP, Ravaud A, Gerin P, Brouet P, et al. Incidence of large oesophageal varices in patients with cirrhosis: application to prophylaxis of first bleeding. Gut 1990; 31: 1298-1302
11. D’Amico G, Pagliaro L, Bosch J. The treatment of portal hypertension: a meta-analytic review. HEPATOLOGY 1995; 22: 332-354
12. de Franchis R. Developing consensus in portal hypertension [Editorial]. J Hepatol 1996; 25: 390-394
13. De Franchis R, ed. Portal Hypertension II. Proceedings of the Second Baveno International Consensus Workshop on Definitions, Methodology and Therapeutic Strategies. Oxford: Blackwell Science Ltd., 1996:10-17.
14. de Franchis R, Pascal JP, Ancona E, Burroughs AK, Henderson M, Fleig W, Groszmann R, et al. Definitions, methodology and therapeutic strategies in portal hypertension: a consensus development workshop, Baveno, Lake Maggiore, Italy, April 5 and 6, 1990. J Hepatol 1992; 15: 256-261
15. Feu F, Garcia-Pagan JC, Bosch J, Luca A, Teres J, Escorsell A, Rodes J. Relation between portal pressure response to pharmacotherapy and risk of recurrent variceal hemorrhage in patients with cirrhosis. Lancet 1995; 346: 1056-1059
16. Villanueva C, Balanzo J, Novella MT, Soriano G, Sainz S, Torras X, Cusso X, et al. Nadolol plus isosorbide mononitrate compared with sclerotherapy for the prevention of variceal rebleeding. N Engl J Med 1996; 334: 1624-1629
17. Armonis A, Patch D, Burroughs AK. Hepatic venous pressure measurement: an old test as a new prognostic marker in cirrhosis? HEPATOLOGY 1997; 25: 245-248
18. Byar DP, Schoenfeld DA, Green SB, Amato DA, Davis R, De Gruttola V, Finkelstein DM, et al. Design considerations for AIDS trials. N Engl J Med 1990; 323: 1343-1348
19. Rothman KJ, Michels KB. The continuing unethical use of placebo controls. N Engl J Med 1994; 331: 394-398
20. Collier J. Confusion over use of placebos in clinical trials. BMJ 1995; 311: 821-822
21. Jones B, Jarvis P, Lewis JA, Ebbutt AF. Trials to assess equivalence: the importance of rigorous methods. BMJ 1996; 313: 36-39
22. Makuch R, Johnson M. Issues in planning and interpreting active control eqivalence studies. J Clin Epidemiol 1989; 42: 503-511
23. Sung JJY, Chung SCS, Lai CW, Chan FKL, Leung JWC, Yung MY, Kassianides C, et al. Octreotide infusion or emergency sclerotherapy for variceal hemorrhage. Lancet 1993; 342: 637-641
24. Altman DG. Octreotide infusion versus injection sclerotherapy [Letter]. Lancet 1993; 342: 1486.
25. Martinez JM, Alvarez Fdez-Represa J. Octreotide infusion versus injection sclerotherapy [Letter]. Lancet 1993; 342: 1486.
26. Nicholl J. Is the study worth doing? [Letter]. Lancet 1993; 342: 809.
27. Thomsen BL, Möller S, Sörensen TIA, The Copenhagen Esophageal Varices Sclerotherapy Project. Optimized analysis of recurrent bleeding and death in patients with cirrhosis and oesophageal varices. J Hepatol 1994; 21: 367-375
28. Rothwell PM. Can overall results of clinical trials be applied to all patients? Lancet 1995; 345: 1616-1619
29. Schmid CH, McIntosh M, Cappelleri JC, Lau J, Chalmers TC. Measuring the impact of control rate in meta-analysis of clinical trials. Control Clin Trials 1995; 16: 665.
30. Glasziou PP, Irwig LM. An evidence based approach to individualising treatment. BMJ 1995; 311: 1356-1359
31. Pagliaro L, D’Amico G, Sörensen TIA, Lebrec D, Burroughs AK, Morabito A. Prevention of first bleeding in cirrhosis: a meta-analysis of randomized trials of nonsurgical treatment. Ann Intern Med 1992; 117: 59-70
32. Thompson SG. Why sources of heterogeneity in meta-analysis should be investigated. BMJ 1994; 309: 1351-1355
33. Forrow L, Taylor WC, Arnold RM. Absolutely relative: how research results are summarized can affect treatment decisions. Am J Med 1992; 92: 121-124
34. Naylor CD, Chen E, Strauss B. Measured enthusiasm: does the method of reporting trial results alter perception of therapeutic effectiveness? Ann Intern Med 1992; 117: 916-921
35. Laupacis A, Naylor CD, Sackett DL. How should the results of clinical trials be presented to clinicians? ACP Journal Club 1992 May-June 1992;117:A12-A14.
36. Bobbio M, De Michelis B, Giustetto G. Completeness of reporting trial results: effect on physicians’ willingness to prescribe. Lancet 1994; 343: 1209-1211
37. Laupacis A, Sackett DL, Roberts RS. An assessment of clinically useful measures of the consequences of treatment. N Engl J Med 1988; 318: 1728-1733
38. The Standards of Reporting Trial Groups. A proposal for structured reporting of randomized controlled trials. JAMA 1994; 272: 1926-1931
39. Taddio A, Pain T, Fassos FF, Boon H, Ilersich L, Einarson R. Quality of non-structured and structured abstracts from original research articles in the BMJ, CMAJ and JAMA. Can Med Assoc J 1994; 150: 1611-1615.
40. Begg C, Cho M, Eastwood S, Horton R, Moher D, Olkin I, Pitkin R, et al. Improving the quality of reporting of randomized controlled trials. JAMA 1996; 276: 637-663
41. The Asilomar Working Group on Recommendations for Reporting of Clinical Trials in the Biomedical Literature. Checklist of information for inclusion in reports of clinical trials. Ann Intern Med 1996; 124: 741-743
42. Poynard T, Calès P, Pasta L, Ideo G, Pascal JP, Paglìaro L, Lebrec D, et al. Beta-adrenergic-antagonist drugs in the prevention of gastrointestinal bleeding in patients with cirrhosis and esophageal varices: an analysis of data and prognostic factors in 589 patients from four randomized clinical trials. N Engl J Med 1991; 324: 1532-1538
43. Stewart LA, Parmar MKB. Meta-analysis of the literature or of individual patient data: is there a difference? Lancet 1993; 341: 418-422
44. Friedman HP, Goldberg JD. Meta-analysis: an introduction and point of view. HEPATOLOGY 1996; 23: 917-928
45. Yusuf S, Wittes J, Probstfield J, Tyroler HA. Analysis and interpretation of treatment effects in subgroups of patients in randomized clinical trials. JAMA 1991; 266: 93-98
46. Oxman AD, Guyatt GH. A consumer’s guide to subgroup analysis. Ann Intern Med 1992; 116: 78-84
47. Schwarz RP. Maintaining integrity and credibility in industry-sponsored clinical research. Control Clin Trials 1991; 12: 753-760
48. Shimm DS, Spece Jr. RG. Industry reimbursement for entering patients into clinical trials: legal and ethical issues. Ann Intern Med 1991; 115: 148-151
49. Waldron HA, Cookson RF. Avoiding the pitfalls of sponsored multicentre research in general practice. BMJ 1993; 307: 1331-1334.
50. Chalmers I. Underreporting research is scientific misconduct. JAMA 1990; 263: 1405-1408.
51. Relman AS. Economic incentives in clinical investigation. N Engl J Med 1989; 320: 933-934.
52. DerSimonian R, Laird N. Meta-analysis in clinical trials. Controlled Clinical Trials 1986; 7: 177-186.
53. Draper D, Gaver Jr. DP, Goel PK. Anonymous combining information: statistical issues and opportunities for research: contemporary statistics number 1. Washington DC: National Academy Press , 1992.
54. Fleiss J. The statistical basis of meta-analysis. Stat Methods Med Res 1993; 2: 121-145.
55. Bailey KR. Inter-study differences: how should they influence the interpretation and analysis of results? Stat Med 1987; 6: 351-358.
56. DeMets DL. Methods for combining randomized clinical trials: strengths and limitations. Stat Med 1987; 6: 341-348.
57. Meier P. Statistics and medical experimentation. Biometrics 1975; 31: 511-529.
58. Peto R. Discussion: why do we need systematic overviews of randomized trials? Stat Med 1987; 6: 242.
59. Provenzale D, Lipscomb J. Cost-effectiveness: definitions and use in the gastroenterology literature. Am J Gastroenterol 1996; 91: 1488-1493.
60. Detsky AS, Naglie IG. A clinician’s guide to cost-effectiveness analysis. Ann Intern Med 1990; 113: 147-154.
61. Glass GV. Primary, secondary and meta-analysis of research. Educ Res 1976; 5: 3-90.
62. Stewart LA, Clarke MJ, The Cochrane Working Group on meta-analysis using individual patient data. Practical methodology of meta-analysis (overviews) using updated individual patient data. Stat Med 1995; 14: 2057-2079.
63. Boissel JP, Bossard N. Registry of prospective clinical trials. Eleventh report. Thromb Haemost 1991; 66: 368-383.
64. Felson DT. Bias in meta-analytic research. J Clin Epidemiol 1992; 45: 885-892.
65. Bailar III JC. The practice of meta-analysis. J Clin Epidemiol 1995; 48: 149-157.
66. Burroughs AK, Bosch J. Clinical manifestations and management of bleeding episodes in cirrhotics. In: McIntyre N, Benhamou JP, Bircher J, Rizzeto M, Rodes J, eds. Oxford Textbook of Clinical Hepatology. Oxford: Oxford University Press, 1991:408-425.
67. Grace ND. Prevention of initial variceal hemorrhage. In: Groszmann RJ, Grace ND, eds. Gastroenterology Clinics of North America: Complications of Portal Hypertension: Esophagogastric Varices and Ascites. Philadelphia: Saunders, 1992:149-161.
68. Angelico M, Carli L, Piat C, Gentile S, Capocaccia L. Effects of isosorbide-5-mononitrate compared with propranolol on first bleeding and long-term survival in cirrhosis. Gastroenterology 1997; 113: 1632-1639.
69. Merkel C, Martin R, Enzo E, Donada C, Cavallarin G, Torboli P, Amodio P, et al. Randomized trial of nadolol alone or with isosorbide mononitrate for primary prophylaxis of variceal bleeding in cirrhosis. Lancet 1996; 348: 1677-1681.
70. Besson I, Ingrand P, Person B, Boutroux D, Heresbach D, Bernard P, Hochain P, et al. Sclerotherapy with or without octreotide for acute variceal bleeding. N Engl J Med 1995; 333: 555-560.
71. Levacher S, Letoumelin P, Pateron D, Blaise M, Lapandry C, Pourriat JL. Early administration of terlipressin plus glyceryl trinitrate to control active upper gastrointestinal bleeding in cirrhotic patients. Lancet 1995; 346: 865-868.
72. Bernard B, Lebrec D, Mathurin P, Opolon P, Poynard T. Beta-adrenergic antagonists in the prevention of gastrointestinal rebleeding in patients with cirrhosis: a meta-analysis. HEPATOLOGY 1997; 25: 63-70.
73. García-Pagán JC, Feu F, Bosch J, Rodes J. Propranolol compared with propranolol plus isosorbide-5-mononitrate for portal hypertension in cirrhosis. Ann Intern Med 1991; 114: 869-873.
74. McCormick P, Burroughs A. Prophylaxis for variceal hemorrhage. Gastrointest Endosc Clin N Am 1992; 2: 167-182.
75. Sarin S, Guptan R, Jain A. A randomized controlled trial of endoscopic variceal band ligation for primary prophylaxis of variceal bleeding [Abstract]. Gastroenterol 1994; 106: 975A.
76. Laine L, Cook D. Endoscopic ligation compared with sclerotherapy for treatment of esophageal variceal bleeding: a meta-analysis. Ann Intern Med 1995; 123: 280-287.
77. Stiegmann GV, Goff JS, Michaletz-Onody PA, Korula J, Lieberman D, Saeed ZA, Reveille RM, et al. Endoscopic sclerotherapy as compared with endoscopic ligation for bleeding esophageal varices. N Engl J Med 1992; 326: 1527-1521.
78. Jensen D, Kivacs T, Randall G, You S, Pelayo E, Jensen M. Initial results of a randomized study of emergency banding vs. sclerotherapy for bleeding gastric or esophageal varices [Abstract]. Gastrointest Endosc 1993; 39: 279.
79. Lo G, Lai K, Chang C, Jeng J, Chen S, Lin C, Chiang H. A prospective, randomized trial of injection sclerotherapy vs. banding ligation in the management of bleeding esophageal varices. Gastroenterology 1994; 106: A933.
80. Mundo F, Mitrani C, Rodriguez G, Farca A. Endoscopic variceal treatment, is band ligation taking over sclerotherapy? [Abstract]. Am J Gastroenterol 1993; 88: 1493A.
81. Hou MC, Lin HC, Kuo BIT, Chen CH, Lee FA, Lee SD. Comparison of endoscopic variceal injection sclerotherapy and ligation for the treatment of esophageal variceal hemorrhage. A prospective randomized trial. HEPATOLOGY 1995; 21: 1517-1522.
82. Cooper M, Abedi M, Haber G, Kortan P, Kandel G, Marcon N. Outcomes of rubber band ligation (RBL) of acute variceal hemorrhage: comparing those with an identifiable bleeding site vs. those with varices and no other bleeding sources [Abstract]. Gastrointest Endosc 1996; 43: 332A.
83. Young M, Sanowski R, Rasche R. Comparison and characterization of ulcerations induced by endoscopic ligation of esophageal varices versus endoscopic sclerotherapy. Gastrointest Endosc 1993; 39: 119-122.
84. Grainek I, Jensen D, Jensen M, Cheng S, Gornbein J, Kovacs T, Jutabha R, et al. Assessment of direct costs of medical care in a prospective, randomized, blinded study of banding vs. sclerotherapy for preventing recurrent variceal hemorrhage for patients without active bleeding at endoscopy [Abstract]. Gastrointest Endosc 1996; 43: 313.
85. Liu J, Miller L, Field R, Barbarevech C, Needleman L, Goldberg B. Gastric and esophageal varices: 20 MHz transnasal endoluminal. US Radiol 1993; 187: 363-366.
86. Miller L, Schiano T, Adrian A, Cassidy M, Liu J, Ter H, Bellary S, et al. Comparison of high resolution endoluminal sonography to video endoscopy in the detection and evaluation of esophageal varices. HEPATOLOGY 1996; 24: 552-555.
87. Nagamine N, Ido K, Ueno N, Kimura K, Kawamata T, Kawada H, Hirasawa T, et al. The usefulness of ultrasonic microprobe imaging for endoscopic variceal ligation. Am J Gastroenterol 1996; 91: 523-529.
88. Escorsell A, Bordas JP, Feu F, Garcia-Pagan JC, Gines A, Bosch J, Rodes J. Endoscopic assessment of variceal volume and wall tension in cirrhotic patients: effects of pharmacological therapy. Gastroenterology 1997; 113: 1640-1646.
89. Chouduri G, Srivastava A, Agarwal D, Dhiman R. Endosonographic evaluation of venous changes around the gastroesophageal junction during sclerotherapy and band ligation of varices [Abstract]. Gastroenterol 1996; 110: 1170A.
90. Leung V, Sung J, Ahuja A, Chung S. Large para-esophageal varices on endosonography predict recurrence of esophageal varices and rebleeding [Abstract]. Gastrointest Endosc 1996; 43: 425A.
91. Starzl TE, Donner A, Eliaszin M, Stitt L, Meier P, Fung JJ, McMichael JP, et al. Randomized trialomania? The multicenter liver transplant trials of tacralinius. Lancet 1995; 346: 1346-1350.
92. Turcotte JG. Liver transplantation in the USA. In: Neuberger J, Luce MR, eds. Liver Transplantation. London: BMJ Publishing group, 1994:127-142.
93. Iwatsuki S, Starzl TE, Todo S, Gordon RD, Tzakis AG, Marsh W, Makowka L, et al. Liver transplantation in the treatment of bleeding esophageal varices. Surgery 1988; 104: 697-705.
94. Wood RP, Shaw BW, Rikkers LF. Liver transplantation for variceal hemorrhage. Surg Clin North Am 1990; 70: 449-461.
95. Bismuth H, Adams R, Mathur S, Sherlock D. Options for elective treatment of portal hypertension in cirrhotic patients in the transplantation era. Am J Surg 1990; 160: 105-110.
96. Henderson JM, Gilmore GT, Hooks MA, Galloway JR, Dodson TF, Hood MM, Kutner MH, et al. Selective shunt in the management of variceal bleeding in the era of liver transplantation. Ann Surg 1992; 216: 248-255.
97. Ringe B, Lang H, Tusch G, Pichlmayr R. Role of liver transplantation in management of esophageal variceal hemorrhage. World J Surg 1994; 18: 233-239.
98. Hermann RE, Henderson JM, Vogt DP, Mayes JT, Geisinger MA, Agnor C. Fifty years of surgery for portal hypertension at the Cleveland Clinic Foundation. Ann Surg 1995; 221: 459-468.
99. Esquivel CO, Klintmalm G, Iwatsuki S, Makowka L, Gordon RD, Tzakis A, Starzl TE. Liver transplantation in patients with patent splenorenal shunts. Surgery 1987; 101: 430-432.
100. Brems JJ, Hiatt JR, Klein AS, Millis M, Colonna JO, Quinones-Baldrich WJ, Ramming KP, et al. Effect of a prior portasystemic shunt on subsequent liver transplantation. Ann Surg 1989; 209: 51-56.
101. Broelsch E, Vogelbach P, Emond JC, Thistlewaite JR, Woodle SA, Baker AL, Whitington PF. [Shunt surgery with reference to indications for liver transplantation]. Langenbecks Archiv fur Chirurgie-Supplement II-Verhandlungen der Deutschen Gesellschaft fur Chirurgie 1989;293-299.
102. Mazzaferro V, Todo S, Tzakis AG, Stieber AC, Makowka L, Starzl TE. Liver transplantation in patients with previous portasystemic shunt. Am J Surg 1990; 160: 111-116.
103. Turrion VS, Mora NP, Cofer JB, Solomon H, Morris CA, Gonwa TA, Goldstein RM, et al. Retrospective evaluation of liver transplantation for cirrhosis: a comparative study of 100 patients with or without previous porto-systemic shunts. Transplant Proc 1991; 23: 1570-1571.
104. Aboujaoude M, Grant D, Ghent C, Mimeault R, Wall W. Effect of portasystemic shunts on subsequent transplantation of the liver. Surg Gyn Obstet 1991; 172: 215-219.
105. Boillot O, Houssin D, Santoni P, Ozier Y, Matmar M, Chapuis Y. Liver transplantation in patients with a surgical portasystemic shunts. Gastroenterol Clin Biol 1991; 15: 876-880.
106. Langnas AN, Marujo WC, Stratta RJ, Donovan JP, Sorrell MF, Rikkers LF, Shaw BWJ. Influence of a prior porta-systemic shunt on outcome after liver transplantation. Am J Gastroenterol 1992; 87: 714-718.
107. Ploeg RJ, D’Alessandro AM, Stegall MD, Wojtowycz M, Sproat IA, Knechtle SJ, Pirsch JD, et al. Effect of surgical and spontaneous portasystemic shunts on liver transplantation. Transplant Proc 1993; 25: 1946-1948.
108. Lerut JP, Mazza D, van Leeuw V, Laterre PF, Donataccio M, de Ville de Goyet J, Van Beers B, et al. Adult liver transplantation and abnormalities of splanchnic veins: experience in 53 patients. Transplant Int 1997; 10: 125-132.
109. Orloff J, Orloff MS, Orloff SL, Rambotti M, Girard B. Three decades of experience with emergency portacaval shunt for acutely bleeding esophageal varices in 400 unselected patients with cirrhosis of the liver. J Am Coll Surg 1995; 179: 257-272ID=”.
110. Stipa S, Balducci G, Ziparo V, Stipa F, Lucandri G. Total shunting and elective management of variceal bleeding. World J Surg 1994; 18: 200-204.
111. Sarfeh IJ, Rypins EB. Partial versus total portacaval shunt in alcoholic cirrhosis. Ann Surg 1994; 219: 353-361.
112. Rosemurgy AS, Goode SE, Zweibel BR, Black TJ, Brady PG. A prospective trial of TIPS vs small diameter prosthetic H graft portacaval shunt in the treatment of bleeding varices. Ann Surg 1996; 224: 378-386.
113. Rikkers LF. Is the distal splenorenal shunt better? HEPATOLOGY 1988; 8: 1705-1707.
114. Spina GP, Henderson JM, Rikkers LF, Terés J, Burroughs AK, Conn HO, Pagliaro L, et al. Distal splenorenal shunt versus endoscopic sclerotherapy in the prevention of variceal bleeding. A meta-analysis of 4 randomized clinical trials. J Hepatol 1992; 16: 338-345.
115. Sarfeh IJ, Rypins EB, Mason GR. A systematic appraisal of portacaval H-graft diameters. Ann Surg 1986; 204: 356-362.
116. Warren WD, Zeppa R, Foman JS. Selective transplenic decompression of gastroesophageal varices by distal splenorenal shunt. Ann Surg 1967; 166: 437-455.
117. Rikkers LF, Jin G, Burnett DA, Buchi KN, Cormier RA. Shunt surgery vs. endoscopic sclerotherapy for variceal hemorrhage. Late results of a randomized trial. Am J Surg 1993; 165: 27-33.
118. Idezuki Y, Kakudo N, Sanjo K, Bandai Y. Sugiura procedure for management of variceal bleeding in Japan. World J Surg 1994; 18: 216-221.
119. Dagenasi M, Langer B, Taylor BR, Greig PD. Experience with radical esophagogastric devascularization procedures (Sugiura) for variceal bleeding outside Japan. World J Surg 1994; 18: 222-228.
120. Grace ND. The side-to-side portacaval shunt revisited. N Engl J Med 1994; 330: 208-209.
121. LaBerge JM, Ring EJ, Gordon R, Lake JR, Doherty MM, Somberg KA, Roberts JP, et al. Creation of transjugular intrahepatic portosystemic shunts with the Wallstent endoprosthesis: results in 100 patients. Radiology 1993; 187: 413-420.
122. Rössle M, Haag K, Ochs A, Sellinger M, Nöldge G, Perarnau JM, Berger E, et al. The transjugular intrahepatic portosystemic stent-shunt procedure for variceal bleeding. N Engl J Med 1994; 330: 165-171.
123. La Berge JM, Somberg KA, Lake JR, Gordon RL, Kerlan Jr. RK, Ascher NL, Roberts JP, et al. Two-year outcome following transjugular intrahepatic portosystemic shunt for variceal bleeding: results in 90 patients. Gastroenterology 1995; 108: 1143-1151.
124. Helton WS, Belshaw A, Althans S, Port S, Caldwell D, Johanasen K. Critical appraisal of the angiographic portocaval shunt (TIPS). Am J Surg 1993; 165: 566-571.
125. Echenagusia AJ, Camúñez F, Simó G, Peiró J, Garay MG, Rodriguez Laiz JM. Variceal hemorrhage: efficacy of transjugular intrahepatic portosystemic shunts created with Strecker Stents. Radiology 1994; 192: 235-240.
126. McCormick PA, Dick R, Panagau EB, Chin JK, Greenslade L, McIntyre N, Burroughs AK. Emergency transjugular intrahepatic portosystemic stent shunting as salvage treatment for uncontrolled variceal bleeding. Br J Surg 1994; 81: 1324-1327.
127. Sauer P, Stiehl A, Hermann S, Richter G, Roeren T, Theilmann L. Stent stenosis after transjugular intrahepatic portosystemic shunt [Abstract]. HEPATOLOGY 1994; 20: 108A.
128. Peramau JM, Mrani SM, Sarkis M, Raabe JJ, Arboqast J. TIPS indications are depending on follow up [Abstract]. HEPATOLOGY 1994; 20: 108A.
129. Caldwell DM, Ring EJ, Rees CR, Zemel G, Darcy MD, Haskal ZJ, McKusick MA, et al. Multicenter investigation of the role of transjugular intrahepatic portosystemic shunt in management of portal hypertension. Radiology 1995; 196: 335-340.
130. Sanyal AJ, Freeman A, Shiffman ML, Purdum PP, Luketic VA, Cheatham AK. Portosystemic encephalopathy after transjugular intrahepatic portosystemic shunt: results of a prospective controlled study. HEPATOLOGY 1994; 20: 46-55.
131. Somberg KA, Riegler JL, LaBerge JM, Doherty-Simor MM, Bachetti P, Roberts JP, Lake JR. Hepatic encephalopathy after transjugular intrahepatic portosystemic shunts: incidence and risk factors. Am J Gastroenterol 1995; 90: 549-555.
132. Piotraschke J, Haag K, Berger E. Latent hepatic encephalopathy in outpatients with TIPS. HEPATOLOGY 1995; 22: 476A.
133. Cabrera J, Maynar M, Granados R, Gorriz E, Reyes R, Pulido-Duque JM, Rodriguez sanRoman JL, et al. Transjugular intrahepatic portosystemic shunt versus sclerotherapy in the elective treatment of variceal hemorrhage. Gastroenterology 1996; 110: 832-839.
134. Riggio O, Copocaccia L, Zipaar V. Transjugular intrahepatic portosystemic shunt (TIPS) vs. endoscopic sclerotherapy (Es) in preventing variceal rebleeding [Abstract]. HEPATOLOGY 1994; 20: 107A.
135. Rossle M, Deibert P, Haag K, Ochs A, Olschewski M, Siegerstetter V, Hauenstein KH, et al. Randomized trial of transjugular-intrahepatic-portosystemic shunt versus endoscopy plus propranolol for prevention of variceal rebleeding. Lancet 1997; 349: 1043-1049.
136. Hosking SW, Johnson AG. Gastric varices: a proposed classification and a guide to management. Br J Surg 1988; 75: 195-196.
137. Watanabe K, Kimura K, Matsutani S, Ohto M, Okuda K. Portal hemodynamics in patients with gastric varices. A study in 230 patients with oesophageal or gastric varices using portal vein catheterization. Gastroenterology 1988; 95: 434-440.
138. Sarin SK, Lahoti D, Saxena SP, Murthi NS, Makwane UK. Prevalence and classification and natural history of gastric varices: long term follow-up study in 568 patients with portal hypertension. HEPATOLOGY 1992; 16: 1343-1349.
139. Paquet KJ. Open discussion on technical aspects of injection of sclerotherapy. In: Westaby D, McDougall BRD, Williams R, eds. Variceal Bleeding. London: Pitmann Press, 1982:215-217.
140. Terés J, Cecilia A, Bordas JM, Rimola A, Bru C, Rodés J. Oesophageal tamponade for bleeding varices. Controlled trial between the Sengstaken Blackmore tube and the Linton-Nachlas tube. Gastroenterology 1978; 75: 566-569.
141. Mathur SK, Dalvi AN, Someshwar V, Supe AN, Raniakhantan R. Endoscopic and radiological appraisal of gastric varices. Br J Surg 1990; 77: 432-435.
142. Hedberg CF, Fowler DL, Ryan RLR. Injection sclerotherapy of oesophageal varices using ethanolamine oleate. Am J Surg 1987; 143: 426-431.
143. Takase Y, Ozak A, Orii R. Injection sclerotherapy of oesophageal varices for patients undergoing emergency and elective surgery. Surgery 1982; 92: 474-479.
144. Jorge AD, Adam J, Seittert L, Segal E. Sclerotherapy of oesophageal varicesan Argentinian experience. Endoscopy 1983; 15: 141-143.
145. Zanasi G, Rossi A, Grosso C, Bini M, Gambitta P, Pirone Z, Arcidiacono R. The effect of endoscopic sclerotherapy of esophageal varices on the development of gastric varices. Endoscopy 1966; 28: 234-238.
146. Trudeau W, Prindiville T. Endoscopic injection sclerosis in bleeding gastric varices. Gastrointest Endosc 1985; 32: 264-268.
147. Sarin SK, Sachdev G, Nanda R, Misra SP, Broor SL. Endoscopic sclerotherapy in the treatment of gastric varices. Br J Surg 1988; 75: 747-750.
148. Oho K, Toyonaga A, Iwao T, Sumino M, Shigemori H, Sakaki M, Nakayama M, et al. Sclerotherapy for bleeding gastric varices: ethanolamine oleate versus butyl-cyanoacrylate [Abstract]. HEPATOLOGY 1994; 20: 107A.
149. Gimson AES, Westaby D, Williams R. Endoscopic sclerotherapy in the management of gastric variceal hemorrhage. J Hepatol 1991; 13: 274-278.
150. Williams SGJ, Peters RA, Westaby D. Thrombinan effective treatment for fundal gastric varices? Gut 1994; 35: 1287-1289.
151. Ramond MJ, Valla D, Mosnier JF. Successful endoscopic obliteration of gastric varices with butyl cyanoacrylate. HEPATOLOGY 1989; 10: 488-493.
152. Soehendra N, Nam VC, Grimm H. Endoscopic obliteration of large esophagogastric varices with Bucrylate. Endoscopy 1986; 18: 25-26.
153. Kuradusenge P, Russeau H, Vinel JP, Maquin P, Croizet O, Coustet B, Railhac JJ, et al. Traitement des hemorragies par rupture de varices cardio-tuberositaires par anastomose porto-systemique intrahepatique par voie transjugulaire. Gastroenterol Clin Biol 1993; 17: 431-434.
154. Spina GP, Arcidiacono R, Bosch J, Pagliaro L, Burroughs AK, Santambrogio A, Rossi A. Gastric endoscopic features in portal hypertension: final report of a consensus conference: Milan, Italy, September 19, 1992. J Hepatol 1994; 21: 461-467.
155. McCormack TT, Sims J, Eyre-Brook I. Gastric lesions in portal hypertension: inflammatory gastritis or congestive gastropathy? Gut 1985; 126: 1226-1232.
156. Papazian A, Braillon A, Dupas JL. Portal hypertensive gastric mucosa: an endoscopic study. Gut 1986; 27: 1199-1203.
157. Quintero F, Pique JP, Bombi JA. Gastric mucosal vascular ectasias causing bleeding in cirrhosis. Gastroenterology 1987; 93: 1054-1061.
158. Sarfeh IJ, Tarnawsky A. Gastric mucosal vasculopathy in portal hypertension. Gastroenterology 1987; 93: 1129-1131.
159. Payen JL, Cales P, Voigt JJ, Barbe S, Pilette C, Dubuisson L, Desmorat H, et al. Severe portal hypertensive gastropathy and antral vascular ectasia are distinctive entities in patients with cirrhosis. Gastroenterology 1995; 108: 138-144.
160. Vigneri S, Termini R, Piraino A, Scialabba A, Pisciotta G, Fontana N. The stomach in liver cirrhosis. Gastroenterology 1991; 101: 472-478.
161. D’Amico G, Montalbarro L, Traina M, Pisa R, Menozzi M, Spano C. Natural history of congestive gastropathy. Gastroenterology 1990; 99: 1558-1564.
162. Sarin SK, Misra SP, Singal A. Congestive gastropathy in portal hypertension: variations in prevalence. Am J Gastroenterol 1988; 83: 1235-1239.
163. Calés P, Zabotto B, Meskens C, Caucanas J, Vinel J, Desmorat H. Gastroesophageal endoscopic features in cirrhosis: observer variability, interassociations, and relationships to hepatic dysfunction. Gastroenterology 1990; 98: 156-162.
164. De Weert T, Gostout C, Wiesner R. Congestive gastropathy and other upper endoscopic findings in 81 consecutive patients undergoing orthotopic liver transplantation. Am J Gastroenterol 1990; 85: 573-576.
165. Carpinelli L, Primignani M, Preatoni P, Angeli P, Battaglia G, Beretta L. Portal hypertensive gastropathy: reproductivity of a classification, prevalence of elementary lesions, sensitivity and specificity in the diagnose of cirrhosis of the liver. A NIEC multicenter Study. Ital J Gastroenterol Hepatol 1997; 29: 533-540.
166. Calés P, Burtin P, Person B. Interobserver agreement on endoscopic diagnosis of diffuse gastric mucosal features in patients with cirrhosis. Eur Gastroenterol Hepatol 1993; 5: 263-268.
167. de Franchis R, Carpinelli L, Primignani M. Pathophysiology, diagnosis and treatment of portal hypertensive gastropathy. In: Dobrilla G, Felder M, De Pretis G, eds. Advances in Hepatobiliary and Pancreatic Diseases. Special clinical topics. Dordrecht: Kluwer Academic Publishers, 1995:109-117.
168. Sarin SK, Sreenvias DV, Lahoti D, Saraya A. Factors influencing development of portal hypertensive gastropathy in patients with portal hypertension. Gastroenterology 1992; 102: 994-999.
169. Hon MC, Lin HC, Chen CH, Perng CL, Lee FY, Lee SD. Change of portal hypertensive gastropathy following EVL or sclerotherapy [Abstract]. HEPATOLOGY 1994; 20: 104A.
170. Gostout CJ, Viggiano TR, Balm RK. Acute gastrointestinal bleeding from portal hypertensive gastropathy. Prevalence and clinical features 1993; 88: 2030-2033.
171. Pérez-Ayuso RM, Piqué JM, Bosch J, Panés J, González A, Pérez R, Rigau J, et al. Propranolol in prevention of recurrent bleeding from severe portal hypertensive gastropathy in cirrhosis. Lancet 1991; 337: 1431-1434.
172. Sarfeh J, Juler GL, Stemmer FA, Mason GR. Results of surgical management of hemmorhagic gastritis patients with gastroesophageal varices. Surg Gynecol Abstrt 1982; 155: 67-70.
173. Orloff MJ, Orloff MS, Orloff SL, Haynes KS. Treatment of bleeding from portal hypertensive gastropathy by portacaval shunt. HEPATOLOGY 1995; 21: 1011-1017.
174. Panés J, Piqué JM, Bordas JM, Bosch J, Terés J, Rodés J. Reduction of gastric hyperemia by glypressin and vasopressin administration in cirrhotic patients with portal hypertensive gastropathy. HEPATOLOGY 1994; 19(1): 55-60.
175. Panés J, Piqué JM, Bordas JM, Casadevall M, Terés J, Bosch J, Rodés J. Effect of bolus injection and continuous infusion of somatostatin on gastric perfusion in cirrhotic patients with portal hypertensive gastropathy. HEPATOLOGY 1994; 20: 336-341.

0270-9139/98/2803-0039$3.00/0
Copyright © 1998 by the American Association for the Study of Liver Diseases.