The single topic symposium “Portal hypertension and variceal bleeding – issues yet to be resolved” was held in Atlanta, GA, July 4-6, 2007, co-sponsored by the American Academy of Liver Diseases (AASLD) and the European Association of Liver Diseases. The theme of the meeting was to identify priority issues for research in this area, alternative markers that reflect lesion outcomes, and stratification of patients by risk level. The recommended consensus developed at this meeting was agreed upon by more than 70% of the 23 experts.
I. Natural history of variceal/variceal bleeding
(A) Natural history of cirrhosis. The natural course of cirrhosis can be divided into two stages: compensated and decompensated cirrhosis. Which stage a patient is in depends on the presence of ascites, variceal bleeding, hepatic encephalopathy or jaundice, both of which differ in terms of clinical presentation, outcome, mortality and predictors of death. The average survival of patients with compensated cirrhosis is greater than 12 years (if they are all in the compensated stage during this period) and significantly greater than that of decompensated cirrhosis; the average survival of the latter is about 2 years. Most deaths in compensated cirrhosis are due to liver failure or to causes unrelated to liver disease. In contrast, the majority of deaths in decompensated cirrhosis are related to liver disease. The most common outcome of compensated cirrhosis is conversion from compensated to decompensated, with an incidence of 5-7% per year. the Child-Turcotte-Pugh (CTP) score is the strongest predictor of death from compensated or decompensated cirrhosis.
Recently, a large cohort study of a group of patients with untreated cirrhosis suggested that cirrhosis can be divided into 4 distinct stages of development. Stages 1 and 2 are seen in compensated cirrhosis with or without varices, respectively Stages 3 and 4 are seen in decompensated cirrhosis and are defined by the presence or absence of ascites or varices as stage 3, with ascites, with or without varices and stage 4 with variceal bleeding, with or without ascites, respectively. The 1-year mortality rates for these 4 stages were 1%, 3%, 20% and 57%, respectively.
(B) Significance of hepatic venous pressure gradient in the natural history of portal hypertension. Hepatic venous pressure gradient (HVPG) is a valid predictor for assessing portal pressure and predicting outcome in both compensated and decompensated stages of cirrhosis. HVPG equal to or greater than 10 mmHg is the most important predictor of development of varices in patients with compensated cirrhosis. HVPG is more diagnostically correct for recurrence of hepatitis C after liver transplantation than liver biopsy for predicting the onset of decompensation. In decompensated cirrhosis, HVPG can be used to predict the outcome of variceal bleeding, and is an independent predictor of death in the adjusted end-stage liver disease (MELD) model along with ascites and age, with HVPG 20 mmHg being the most appropriate cut-off point for delineating the prognosis of such patients.
(iii) Recommendations: Compensated and decompensated cirrhosis can be considered as two different clinical components in clinical and research studies. HVPG >10 mmHg is the most reliable predictor of the occurrence of varices and decompensation and can be used in clinical trials to stratify patients with compensated cirrhosis. The stratification of cirrhotic patients needs to be further confirmed by prospective cohort studies.
Second, screening for varices. Screening esophagogastroduodenoscopy should be performed once the diagnosis of cirrhosis is made so that effective prevention of variceal bleeding can be implemented. Capsule gastroscopy as a non-invasive means of evaluating esophagogastric varices requires further study to determine its reproducibility, reliability, correctness, and patient willingness, cost/benefit ratio, but can be a meaningful alternative test for patients who are unable or unwilling to undergo EGD.
III. Prevention of first bleeding from varicose veins
(i) Prevention of varicose vein formation
Non-selective beta-blockers (NSBB) are not effective in preventing the development of varicose veins in patients with portal hypertension and have a high incidence of adverse effects.
(ii) Prevention of first bleeding from varicose veins
The rate of bleeding in small varicose veins with high risk of bleeding is the same as that of large varicose veins, while small varicose veins with low risk of bleeding progress slowly after treatment with nadolol, therefore patients with small varicose veins with high risk of bleeding should receive NSBB, while other patients with small varicose veins can choose to apply NSBB. if patients who have already received NSBB do not need to repeatedly make dye frightened NSBB, they must be reviewed every 2 years. NSBB and transvenous varicose veins can be used as an option.
Both NSBB and transdermal staining are effective in the prevention of first bleeding in patients with medium/large varicose veins (since EVL). Pooled analysis of 12 high-quality randomized controlled studies (RCTs) revealed additional high-quality RCT analyses with no significant difference in the prevention of first bleeding between these 2 treatments, with most serious complications occurring in EVL and 3 deaths reported secondary to bleeding secondary to post-EVL ulceration. When cost/utilization per 1 adjusted quality of life year analysis was introduced, EVL was superior to NSBBB for improving quality of life adjusted year (QALY). EVL significantly reduced initial bleeding but not mortality, and its long-term benefit is uncertain. Therefore NSBB can be used as a first-line treatment to prevent bleeding. the benefits of NSBB combined with EVL in trials have not been concluded, and most participants agreed that NSBB combined with EVL is not recommended.
(iii) Role of HEVG assay in the prevention of variceal/variceal bleeding
In patients without varicose veins and portal hypertension (HVPG ≥6 mmHg), the best predictor of progression to varicose veins is HVPG >10 mmHg at baseline, with HVPG reduction >10% associated with a significant reduction in the incidence of varicose veins. Pharmacological treatment may increase the proportion of patients who achieve a reduction in HVPG and thus effectively prevent the development of varicose veins. In patients with never bleeding varicose veins (mostly medium and large varicose veins), obtaining a reduction in HVPG of >10% to 20% may reduce the incidence of first bleeding, and a reduction in HVPG to <12 mmHg may essentially prevent variceal bleeding. a reduction in HVPG of >15% is associated with a reduced incidence of spontaneous peritonitis.
Recommendation: Clinical trials of existing treatments for primary prevention of variceal bleeding are not necessary unless new treatments are available. Trials for the prevention of first bleeding from varices are limited to patients with HVPG >10 mmHg. Trials of new drugs for the prevention of first variceal bleeding must be compared in a double-blind design with NSBB and include HVPG measurement. In patients with small varicose veins these 2 indices must be combined as endpoints.
IV. Treatment of acute bleeding
(i) Drugs and dye-resistant plague : Short-term antibiotic prophylaxis given to patients with cirrhosis admitted with gastrointestinal bleeding reduces the incidence of spontaneous peritonitis, variceal rebleeding and death. rCTs comparing different drugs (vasopressin, growth inhibitor, terlipressin, octreotide) suggest no significant differences in control of bleeding and early rebleeding, but vasopressin combined with more adverse events. The optimal duration of drug therapy has not been determined, with most participants agreeing on a duration from 2 to 5 days, depending on bleeding control and the presence of predictors of rebleeding (e.g., CTP classification, HVPG).
EVL was superior to endoscopic sclerotherapy in terms of bleeding control, reduction of rebleeding, and adverse events, but there was no significant difference in mortality. Further trials in determining the optimal endoscopic treatment are no longer necessary.
(ii) Measurement of HVPG in acute variceal bleeding
A prospective study of patients with HVPG measured within 48 hours of admission for bleeding showed that HVPG >20 mmHg was associated with increased bleeding and mortality. parameters such as CTP score and blood pressure at admission had the same predictive value.
(iii) Percutaneous intrahepatic portosystemic shunt (TIPS) in acute variceal bleeding
TIPS is an effective option in the face of failed combination drug and endoscopic therapy. A small study suggested that early TIPS placement (within 24 hours of bleeding) was associated with significantly improved survival in patients with HVPG >20 mmHg. Thus, HVPG may provide a stratification of the patient’s risk level and provide useful information for the selection of more urgent treatment for high-risk patients.
(iv) Consensus recommendation: Risk stratification for acute variceal bleeding can help to better estimate the course of drug therapy and determine different treatment strategies (early TIPS for high-risk patients). HVPG greater than 20 mmHg at 48 hours of bleeding is a reliable predictor of poor prognosis. Non-invasive predictors of poor prognosis are under investigation and need further confirmation. Early TIPS treatment in high-risk patients needs further study
V. Prevention of variceal rebleeding
(i) Pharmacological prophylaxis: Patients with cirrhosis who have not yet received primary prophylaxis may receive NSBB or EVL or both. Since NSBB can theoretically prevent rebleeding of varices before their occlusion and prevent the reappearance of varices, the best treatment to prevent rebleeding of esophageal varices is combined NSBB and EVL therapy. Comparative studies of drug combination therapy (NSBB combined with mononitrate) for EVL have shown no difference in rebleeding and mortality.
(ii) HVPG in the prevention of rebleeding: Serial HVPG measurements performed on patients from baseline to 120 days and especially within 1 month suggest that HVPG <12 mmHg allows patients to avoid bleeding, and the risk of rebleeding is very low in patients with HVPG <20% compared to baseline and in non-responders, and about 30%-40% of patients treated with NSBB + mononitrate are HVPG of responders. To date, there are no trials comparing HVPG-guided therapy with conventional therapy for RCTs, and such trials are greatly needed.
(iii) Clinical practice of EVL and surgical shunts in preventing variceal rebleeding: EVL implementation and subsequent dye-guided swishing badgers can be monitored again at 3 months with the 1st dye-guided swishing slow toad implemented at 1 to 3 months and the interval adjusted to 6 to 12 months thereafter depending on the reproduction of variceal veins.
In cirrhotic patients with CTP grade A/B who failed drug/endoscopic therapy, large multicenter studies have shown that TIPS and distal splenorenal shunts are equivalent in terms of rebleeding rates, hepatic encephalopathy, and mortality, and since both procedures have the same outcome, surgery remains an option for low-risk patients only if the surgical risk rate is low (CTP-A, grade B) and an experienced surgeon is present . Based on the fact that there is still a lack of skilled specialists to perform TIPS in many areas and the difficulty of performing liver transplantation, most participants agreed that surgical training in surgical shunts is needed, but further clinical trials in this area are not yet necessary.
(iv) Recommendations: HVPG-guided therapy can be further studied to assess its impact on decision formation and outcomes. New drug therapy reviews must be compared with NSBB in a double-blind design and must include HVPG-containing assays.
VI. HVPG as a surrogate endpoint
HVPG is predictive of decompensation and death, and a sustained decrease in HVPG is a negative predictive marker for the development of varices, variceal hemorrhage, non-variceal complications of developing portal hypertension, and death. HVPG is not only useful as a surrogate for clinical trials in portal hypertension, but is also superior to liver puncture biopsy in assessing the progression of chronic liver disease (viral and metabolic) and in predicting the development of decompensated cirrhosis. superior to liver puncture biopsy.
In studies of RCTs, the threshold for HVPG to successfully predict prevention of variceal occurrence is a 10% decrease from baseline, most appropriately less than 10 mm Hg. In trials of variceal bleeding prevention, the threshold for HVPG to predict successful prevention is a decrease to 12 mm Hg, or a 20% decrease from baseline. In secondary prevention of varicose veins, the 2nd HVPG measurement should be performed as soon as the most appropriate therapeutic dose is obtained and no later than 1 month after the first measurement.
HVPG is an accurate, reproducible and safe technique. The reproducibility of the technique was indirectly confirmed by analyzing repeated HVPG determinations from 20 to 720 minutes after baseline, suggesting that the mean difference between determinations was only 0.4%. In a study of 2,364 HVPG operations performed, the overall complication rate was only 2.3%, with the most common complication being a large hematoma in the neck or groin. No operation-associated deaths were observed.
Instantaneous liver elastography (Fibroscan) is a novel rapid, non-invasive and reusable method to determine liver stiffness (LS). Prospective studies have shown an association of LS with HVPG in patients with chronic hepatitis C and alcoholic cirrhosis. Threshold values for LS reflecting clinically meaningful portal hypertension have not been established, so consecutive large double-blind randomized controlled studies in patients with cirrhosis of various etiologies are needed.
Recommendation: HVPG is the best surrogate marker for trials targeting portal hypertension and must be measured at the time of each trial of drug therapy for portal hypertension. HVPG reflects liver fibrosis formation and progression of chronic liver disease and can be applied in therapeutic trials that include liver fibrosis progression as an endpoint. Although the prognostic value of HVPG in alcoholic/viral hepatitis cirrhosis has been established, prognostic studies of cirrhosis of other etiologies are expected. Transient liver elastography is well-established in detecting clinically significant portal hypertension. Non-invasive techniques to assess hemodynamic response need to be established.
VII. Gastric varices
Gastric varices can occur in approximately 20% of patients with portal hypertension, and fundic varices are a subclass of gastric varices with a high incidence of bleeding and rebleeding. Notably, large fundic varices can bleed at HVPG less than 12 mm Hg. NSBB is recommended as primary prevention of gastric variceal bleeding. Treatments to control bleeding and prevent rebleeding include endoscopy (EVL, gel tamponade, thrombin) and radiological intervention (TIPS, transvenous balloon occlusion with retrograde embolization). Data from uncontrolled trials comparing treatment of fundic variceal bleeding showed that the treatment that achieved the best control of initial bleeding (90-100%) was gel injection, TIPS, or transvenous balloon occlusion retrograde embolization.
Three small, single-center RCTs compared the efficacy of endoscopic variceal gel tamponade (EVO) with that of dyed diathermy (not EVS) or EVL in gastric variceal bleeding. The results of all 3 RCTs favored EVO treatment in terms of control of bleeding, rebleeding, or complication rates. TIPS is recommended for acute bleeding in fundic variceal veins where EVO is not performed or where rebleeding occurs after EVO.
Recommendations: Study of gastric varices with a focus on fundic varices
VIII. Portal hypertension and variceal/ variceal vein bleeding in children
The most common causes of portal hypertension in children are biliary atresia and portal vein embolism, and guidelines for adult cirrhosis are not appropriate for generalized application to this population. Pediatric patients require different treatments depending on the etiology of their portal hypertension formation. EVL may be ideal for acute variceal bleeding and prevention of rebleeding. However, there are no studies comparing these treatments with NSBB. EVL is not suitable for use in children younger than 1 year of age. In children with portal vein thrombosis, Reso-Rex shunts may be the best option for secondary prevention.
RECOMMENDATION: Although studies of RCTs are extremely challenging to implement in children, prospective cohort studies in carefully identified risk populations (compensated or decompensated, with or without variceal bleeding) are still desirable. Non-invasive treatments such as NSBB require further study.
IX. Extrahepatic portal hypertension in non-cirrhotic adults
This type of portal hypertension is mainly due to portal vein thrombosis and occurs mostly in patients with impaired latent coagulation mechanisms. Suitable treatment is limited to methods that have been demonstrated in adults with cirrhosis, in particular NSBB, endoscopic treatment and shunt surgery (splenorenal or mesenteric-venous shunt) and TIPS treatment as second-line therapy. According to the results of retrospective cohort studies, anticoagulant use in patients with impaired coagulation mechanisms, personal history of unexplained thrombosis, history of thrombotic episodes in more than 1 person in the family, history of ischemia or superior mesenteric vein involvement is associated with a reduction in their thrombotic complications and bleeding, and improved patient survival.