Acute variceal bleeding (AVB) is a potentially life-threatening complication of cirrhosis and portal hypertension. Combination therapy with vasoactive drugs and endoscopic variceal ligation is the first-line treatment in the management of AVB after adequate hemodynamic resuscitation. Short-term antibiotic prophylaxis, early resuscitation, early use of lactulose for prevention of hepatic encephalopathy, targeting of conservative goals for blood transfusion, and application of early transjugular intrahepatic portosystemic shunts in patients with AVB have further improved the prognosis of AVB. This article discusses the epidemiology, diagnosis, and nonendoscopic management of AVB.
Key points
- •
Esophageal and gastric varices are common among persons with cirrhosis.
- •
Short-term antibiotic prophylaxis, early resuscitation, targeting of conservative goals for blood transfusion, and management of complications such as infection and renal failure are important.
- •
Combination therapy with vasoactive drugs and endoscopic variceal ligation is the first-line treatment in the management of acute variceal bleeding after adequate hemodynamic resuscitation.
- •
The MELD score is an important predictor of early mortality after variceal bleeding.
Introduction
Epidemiology
At the time of initial diagnosis of cirrhosis, approximately half of the patients have esophageal varices, and with the progression of cirrhosis approximately 90% of patients develop esophageal varices. Varices are present in approximately 40% of patients with compensated cirrhosis and 60% of patients with ascites. Large esophageal varices (>5 mm) are seen in 16% of all patients screened for varices by upper endoscopy. The presence of varices correlates with the severity of liver disease: 20% to 40% of Child-A cirrhosis patients have esophagogastric varices (EGV), compared with up to 85% of Child-C cirrhotics. Patients with primary biliary cirrhosis may develop varices and aute variceal bleeding even in the absence of established cirrhosis.
Esophageal Varices
Patients with cirrhosis who do not have varices at the time of initial upper endoscopy develop varices at a rate of 8% per year. The progression of small (<5 mm) to large varices (>5 mm) occurs at a similar rate of 7% to 8% per year. An elevated hepatic venous pressure gradient (HVPG) of greater than >10 mm Hg is an independent predictor of the development of varices. About one-third of patients who have varices develop AVB. Patients who are found to have small varices at the time of initial endoscopy have a 5% per year risk of bleeding, compared with 15% per year in those with medium-sized or large varices at diagnosis. Moreover, 40% of patients with AVB spontaneously stop bleeding without any intervention, compared with 80% of those with nonvariceal causes of upper gastrointestinal bleeding. However, patients with severe liver disease (Child C) are less likely to stop bleeding spontaneously. With the current standard of therapy, 80% to 90% of patients have cessation of hemorrhage. Despite treatment, 1 in 4 patients will still show either a failure to control the bleeding or an early recurrence of the hemorrhage in the first 6 weeks after the initial bleeding. The risk of rebleeding is highest in the period immediately after the sentinel bleed: 40% of all rebleeding episodes occur within the first 5 days. An elevated HVPG of greater than 20 mm Hg, when measured within 24 hours of variceal hemorrhage, is associated with failure to control bleeding and early rebleeding. A Model for End-Stage Liver Disease (MELD) score of 18 or higher is also an independent predictor of early rebleeding.
Gastric Varices
Gastric varices (GV) are present in 20% of patients with portal hypertension, and are the source of 5% to 10% of all upper gastrointestinal bleeding episodes in patients with cirrhosis. GV carry a 10% to 16% risk of bleeding in 1 year and 25% risk of bleeding in 2 years. Although the prevalence and bleeding risk of GV are lower, the bleeding is usually more severe, requires more transfusions, and is associated with higher mortality. Even after endoscopic injection of tissue glue, GV bleeding is still associated with high rebleeding rates, ranging from 22% to 37%. The risk of recurrent bleeding depends on the location of the varix: isolated varices in the gastric fundus (53%) bear the highest risk for recurrent bleeding, followed by varices along the greater gastric curvature (19%) and lesser gastric curvature (6%). The annual incidence of bleeding is 4% in patients with Child class A with small varices without red wale signs, and 65% in patients with Child class C with large varices with red wale signs. Large fundal varices may occasionally bleed despite HVPG values of less than 12 mm Hg.
Mortality
Any death occurring in the first 6 weeks after the index bleed is considered a bleeding-related death. Three decades ago, AVB was associated with a mortality of 60% at 6 weeks. With recent developments in pharmacologic and endoscopic treatments, this figure has significantly improved in the present era to 20% or less. Immediate mortality from uncontrolled bleeding is in the range of 4% to 8%. Child class C, MELD score 18 and higher, and failure to control bleeding or early rebleeding predict 6-week mortality. Mortality is 0% among patients with Child class A disease and approximately 30% among patients with Child C. Forty percent of deaths are directly related to bleeding and shock, with the remainder being due to renal failure, hepatic encephalopathy (HE), and sepsis. In a recent analysis, MELD greater than 19 predicted 20% or greater mortality, whereas MELD scores of less than 11 predicted less than 5% mortality.
Introduction
Epidemiology
At the time of initial diagnosis of cirrhosis, approximately half of the patients have esophageal varices, and with the progression of cirrhosis approximately 90% of patients develop esophageal varices. Varices are present in approximately 40% of patients with compensated cirrhosis and 60% of patients with ascites. Large esophageal varices (>5 mm) are seen in 16% of all patients screened for varices by upper endoscopy. The presence of varices correlates with the severity of liver disease: 20% to 40% of Child-A cirrhosis patients have esophagogastric varices (EGV), compared with up to 85% of Child-C cirrhotics. Patients with primary biliary cirrhosis may develop varices and aute variceal bleeding even in the absence of established cirrhosis.
Esophageal Varices
Patients with cirrhosis who do not have varices at the time of initial upper endoscopy develop varices at a rate of 8% per year. The progression of small (<5 mm) to large varices (>5 mm) occurs at a similar rate of 7% to 8% per year. An elevated hepatic venous pressure gradient (HVPG) of greater than >10 mm Hg is an independent predictor of the development of varices. About one-third of patients who have varices develop AVB. Patients who are found to have small varices at the time of initial endoscopy have a 5% per year risk of bleeding, compared with 15% per year in those with medium-sized or large varices at diagnosis. Moreover, 40% of patients with AVB spontaneously stop bleeding without any intervention, compared with 80% of those with nonvariceal causes of upper gastrointestinal bleeding. However, patients with severe liver disease (Child C) are less likely to stop bleeding spontaneously. With the current standard of therapy, 80% to 90% of patients have cessation of hemorrhage. Despite treatment, 1 in 4 patients will still show either a failure to control the bleeding or an early recurrence of the hemorrhage in the first 6 weeks after the initial bleeding. The risk of rebleeding is highest in the period immediately after the sentinel bleed: 40% of all rebleeding episodes occur within the first 5 days. An elevated HVPG of greater than 20 mm Hg, when measured within 24 hours of variceal hemorrhage, is associated with failure to control bleeding and early rebleeding. A Model for End-Stage Liver Disease (MELD) score of 18 or higher is also an independent predictor of early rebleeding.
Gastric Varices
Gastric varices (GV) are present in 20% of patients with portal hypertension, and are the source of 5% to 10% of all upper gastrointestinal bleeding episodes in patients with cirrhosis. GV carry a 10% to 16% risk of bleeding in 1 year and 25% risk of bleeding in 2 years. Although the prevalence and bleeding risk of GV are lower, the bleeding is usually more severe, requires more transfusions, and is associated with higher mortality. Even after endoscopic injection of tissue glue, GV bleeding is still associated with high rebleeding rates, ranging from 22% to 37%. The risk of recurrent bleeding depends on the location of the varix: isolated varices in the gastric fundus (53%) bear the highest risk for recurrent bleeding, followed by varices along the greater gastric curvature (19%) and lesser gastric curvature (6%). The annual incidence of bleeding is 4% in patients with Child class A with small varices without red wale signs, and 65% in patients with Child class C with large varices with red wale signs. Large fundal varices may occasionally bleed despite HVPG values of less than 12 mm Hg.
Mortality
Any death occurring in the first 6 weeks after the index bleed is considered a bleeding-related death. Three decades ago, AVB was associated with a mortality of 60% at 6 weeks. With recent developments in pharmacologic and endoscopic treatments, this figure has significantly improved in the present era to 20% or less. Immediate mortality from uncontrolled bleeding is in the range of 4% to 8%. Child class C, MELD score 18 and higher, and failure to control bleeding or early rebleeding predict 6-week mortality. Mortality is 0% among patients with Child class A disease and approximately 30% among patients with Child C. Forty percent of deaths are directly related to bleeding and shock, with the remainder being due to renal failure, hepatic encephalopathy (HE), and sepsis. In a recent analysis, MELD greater than 19 predicted 20% or greater mortality, whereas MELD scores of less than 11 predicted less than 5% mortality.
Predictive models
Multiple studies have identified independent factors that predict the risk of bleeding, rebleeding, and death in AVB ( Table 1 ). However, the aggregate conclusions of these studies are discordant, and the predictive value of the combined results is difficult to assess. Predictors of bleeding include presence of decompensated cirrhosis (Child B or C), size of varices, and presence of high-risk stigmata on endoscopy (red wale signs). The risk of further rebleeding, mortality, and complications is highest within the first 6 weeks. In prognostic studies conducted in the current era using standard-of-care interventions (antibiotics, vasoactive drugs, and endoscopic band ligation), the risk of early rebleeding or treatment failure was 15% to 21%, with the risk of early mortality being 6% to 24%. Predictors of rebleeding or treatment failure include MELD score, Child-Turcotte-Pugh (CTP) score, elevated HVPG (≥20 mm Hg), development of infections, and endoscopic appearance (active bleeding, clot on varix, and shock). Recurrent variceal bleeding is lower among patients who achieve a reduction in HVPG to less than 12 mm Hg or a 20% reduction in baseline HVPG values. However, the role of HVPG-guided therapy remains unclear. Predictors of mortality include the aforementioned factors in addition to hepatocellular carcinoma (HCC), worsening renal function, use of steroids, and advanced age.
| Severity of Variceal Bleed | Treatment Failure | Early Rebleeding | Mortality |
|---|---|---|---|
| HVPG | |||
| Alcoholic liver disease | |||
| Infection | |||
| CTP class/score | CTP class/score | ||
| PRBC transfusion | PRBC transfusion | ||
| Size and morphology of varices | |||
| Ascites | |||
| Portal Vein Thrombosis | |||
| Hematocrit/Hb at presentation | |||
| Platelet Count | Platelet Count | ||
| Degree of Liver Failure | |||
| Active bleeding at endoscopy | |||
| Shock | Shock | ||
| AST | |||
| First bleed | |||
| MELD >18 | |||
| Encephalopathy | |||
| Hepatocellular carcinoma | |||
| Short interval to admission | |||
| Urea | |||
| Hematemesis | |||
| Creatinine | |||
| Albumin | |||
| Age | |||
| Early rebleeding | |||
| Prothrombin Time | |||
| Treatment failure | |||
| Bilirubin | |||
Child-Turcotte-Pugh Score
Traditionally the CTP score has been used to stratify patients at risk for death after AVB, with a higher CTP score associated with an increased mortality. The CTP score is easy to use, but lack of standardized measurement of albumin and prothrombin time across laboratories and the use of subjective parameters limit its clinical use.
Model for End-Stage Liver Disease Score
MELD was shown to be superior to CTP score in predicting mortality after transjugular intrahepatic portosystemic shunt (TIPS) performed to prevent recurrent variceal bleeding. However, in a retrospective study of 212 cirrhotic patients with AVB, CTP and MELD scores similarly predicted overall in-hospital mortality, but MELD was significantly better than CTP score for predicting in-hospital mortality directly related to variceal bleeding.
Patients with a MELD score of 18 or higher were found to be at increased risk for death within the first 6 weeks, and for rebleeding within the first 5 days after the index bleed ( P <.001 and P <.04, respectively). Furthermore, patients with a MELD score of at least 18, requiring 4 or more units of packed red blood cells (PRBC) within the first 24 hours of acute variceal hemorrhage (hazard ratio [HR] 11.3; P <.001) or having active bleeding at endoscopy (HR 9.9, P <.001) were at increased risk for death within 6 weeks. In addition, MELD score was a significant predictor of mortality in patients with cirrhosis who were hospitalized with an AVB: every 1-point increase in the MELD score conferred an 8% and 11% increased risk of death at 5 days and 6 weeks, respectively. A recent study of 128 patients also found MELD to be a powerful predictor of 6-week mortality in patients with early rebleeding after endoscopic band ligation for AVB. The mortality rate was 14.7% in patients with a MELD score of less than 21.5 and 71.7% in patients with MELD score of at least 21.5 at 6 weeks ( P <.001).
Reverter and colleagues recently sought to improve the determination of risk for patients presenting with AVB using objective variables. The primary outcome was mortality within 6 weeks of incident bleed. Several previously proposed prognostic models were compared: these included the MELD score, CTP score, and predictive models proposed by D’Amico and colleagues and Augustin and colleagues. The MELD score was identified as the best model in terms of discrimination and overall performance. A MELD score of at least 19 predicted at least 20% mortality, and a MELD score of less than 11 predicted less than 5% mortality within 6 weeks of the AVB episode.
Hepatic Venous Pressure Gradient
Elevated HVPG is an independent predictor of mortality, with a purported 3% increase in mortality per 1-mm Hg increase in HVPG. A single HVPG measurement greater than 20 mm Hg is associated with worse probability of survival (1-year mortality, 64% vs 20%; P <.01) and early rebleeding from varices. Other factors that encapsulate advanced liver disease, such as serum sodium, are also important. A combination of hyponatremia (serum sodium <130 mmol/L) and HVPG greater than 20 mm Hg is associated with a significantly worse prognosis than either alone. However, it is unclear whether addition of HVPG to MELD of Child-Pugh stage based prediction significantly improves the predictive capabilities of a prognostic model.
Diagnosis
In patients with cirrhosis, 70% of upper gastrointestinal bleeds are due to AVB. Hence all patients with cirrhosis and upper gastrointestinal bleeds should be treated for varices unless evidence for other causes is available. A thorough history and physical examination are important in establishing a diagnosis. Patients usually present with hematemesis, but hematochezia and melena may also be present. On physical examination the presence of signs of chronic liver disease in patients with no known history of cirrhosis should be sought. Severity of blood loss should be estimated by hemodynamic and laboratory parameters. Resting tachycardia in the absence of other causes is the first sign of mild hypovolemic shock, and should not be ignored. In patients who are normotensive, orthostatics should be measured. Orthostatic hypotension suggests a loss of 15% of the blood volume. On the other hand, frank hypotension is associated with a 40% loss of blood volume.
Endoscopy is the gold standard for the diagnosis of AVB. The diagnosis of variceal hemorrhage is made when diagnostic endoscopy shows 1 of the following: active bleeding from a varix, a white nipple overlying a varix, clots overlying a varix, or varices with blood in stomach and no other potential source of bleeding.
Several other modalities have been proposed to predict the presence of varices in subjects who have not overtly bled.
Hepatic Venous Pressure Gradient
HVPG is a reliable method for indirectly evaluating the portal pressure gradient, establishing the effectiveness of treatment, and predicting the occurrence of complications in portal hypertension. Portal hypertension is present when the HVPG is greater than 5 mm Hg, but it is considered clinically significant when the HVPG is greater than 10 mm Hg. Esophageal varices do not bleed at an HVPG less than 12 mm Hg. In patients without varices, an HVPG greater than 10 mm Hg has been shown to be the strongest predictor of the development of varices and clinical decompensation of cirrhosis. In patients with variceal hemorrhage, an HVPG of more than 20 mm Hg (measured within 24 hours after admission) is the best predictor of variceal rebleeding and mortality. By contrast, a reduction in the HVPG to less than 12 mm Hg or a reduction of more than 20% from the baseline value is associated with a decreased risk of variceal hemorrhage and improved patient survival. A meta-analysis that included 10 studies with 595 patients evaluated whether targeted HVPG reduction predicts variceal bleeding in patients with cirrhosis receiving nonselective β-blockers for primary prophylaxis. The relative risk (RR) of bleeding in patients achieving a reduction of HVPG to less than or equal to 12 mm Hg or at least 20% compared with baseline (overall responders) was RR = 0.27 (95% confidence interval [CI] 0.14–0.52), although significant heterogeneity between studies was reported.
Liver and Spleen Stiffness
Measurement of liver and spleen stiffness, as a potential surrogate of elevated HVPG by either transient elastography or acoustic radiation force impulse elastography, has been shown to correlate with the presence of varices. Measurement of liver stiffness correlates reasonably well with the HVPG, particularly at HVPG values lower than 10 mm Hg.
Platelet count and spleen size, when combined with spleen stiffness measurement, have been able to better predict the presence of varices. A platelet count of less than 88,000 has been found to be associated with the presence of esophagogastric varices. In a meta-analysis of 20 studies (N = 3063), the ratio of platelet count to spleen diameter was found to have a high accuracy for diagnosing esophageal varices (EV). A single score obtained by combining measurements of liver stiffness, spleen size, and platelet count has shown good accuracy for diagnosing and ruling out varices and portal hypertension in patients with cirrhosis. In a study of 90 patients with hepatitis B–related liver cirrhosis, a predictive model measuring liver stiffness × spleen diameter/platelet count was found to have a 94.7% negative predictive value and a 93.3% positive predictive value for diagnosing EV, and was able to calculate the risk of future bleeding from EV.
Other Modalities
FibroTest (a panel of 5 biochemical markers of hepatic fibrosis), combined with age and gender, is a good predictor of liver fibrosis and EV. In a prospective study of 130 patients, Fibrotest was found to have good correlation with HVPG, but only a moderate diagnostic value for the detection of severe portal hypertension in patients with cirrhosis.
Endoscopic ultrasonography has also been used to study varices and identify increased risk for bleeding. The role of capsule endoscopy in diagnosing varices has been evolving and also shows promise. Two recent pilot studies demonstrated that capsule endoscopy is a safe and well-tolerated technique with which to diagnose EV.
Multidetector-row computed tomography (MDCT) has shown promising results in the evaluation of EV. A recent study concluded that enhanced MDCT was able to localize the origin of bleeding in 96.4% cases preceding urgent endoscopy, and also led to significantly swifter detection of bleeding etiology at endoscopy.
Treatment
Addressing issues related to stabilizing the airway, breathing, and circulation is paramount. The goals of management include hemodynamic resuscitation, control of bleeding, and prevention and treatment of complications ( Fig. 1 ).
The initial focus of treatment is on delivery of adequate oxygen to the tissues by maintaining oxygen saturation, cardiac output, and hemoglobin concentration.
Patients with AVB are at risk of aspiration of gastric contents, including blood. This risk is especially high in patients with massive hematemesis and reduced consciousness during the endoscopic procedure. Patients with HE or delirium tremens may have difficulties maintaining their airway. Therefore, elective endotracheal intubation before upper endoscopy should be considered in this subset of patients. Intubation before endoscopy not only decreases the risk of aspiration but also makes endoscopic procedures easier to perform, especially when clearance of a large volume of retained blood is expected. Two large-bore intravenous accesses should be established for rapid administration of volume. However, many patients end up with a central venous catheter for aggressive resuscitation, central venous pressure monitoring, and administration of vasoactive medications. Assessment of blood loss should be made by history taking, physical examination, and laboratory tests. While waiting for blood, resuscitation should be initiated with plasma expanders. The choice of fluids for replacing the intravascular volume is a topic of debate. Whatever the type of crystalloid used, the systolic blood pressure should be maintained around 90 to 100 mm Hg with a heart rate of less than 100 beats per minute.
Blood loss should be replaced with PRBC whenever possible to maintain hemoglobin between 7 and 8 mg/dL except in patients with rapid ongoing bleeding or ischemic heart disease, in which case this threshold may be raised. A delicate balance should be maintained when administering blood or fluids, as on one hand hypervolemia has been shown to cause rebound increases in portal pressure and increase the risk of rebleeding and mortality, while on the other hypotension has been shown to increase the risk of infection, renal failure, rebleeding, and death. Recently, the merits of a restrictive strategy of blood transfusion only when the hemoglobin drops below 7 g/dL in comparison with a liberal transfusion strategy (whenever hemoglobin drops below 9 g/dL) were shown, with higher survival seen among persons with Child-A and Child-B cirrhosis. Patients with AVB usually have multiple derangements in clotting parameters secondary to their underlying chronic liver disease. However, these abnormalities do not correlate with the severity of bleeding. It is common practice to administer fresh frozen plasma and platelets to patients with elevated international normalized ratio and thrombocytopenia (platelets <50,000/μL), although the exact benefit of these measures is not established. A recent study suggested that patients with cirrhosis may at times instead be in a hypercoagulable state, owing to an imbalance between procoagulant and anticoagulant factors. The use of fresh frozen plasma and platelets can also lead to expansion of the intravascular volume, leading to an increase in portal pressure. The use of recombinant activated factor VII (the most common deficient clotting factor), which corrects prothrombin time in cirrhotics, is not routinely recommended. Treatment with thrombopoietin or desmopressin, a drug that significantly decreases bleeding time in cirrhosis, has shown no clinical benefit in the setting of AVB.
Hepatic Encephalopathy
Patients having AVB may have HE at the time of presentation. In addition, AVB itself leads to worsening of HE, owing to the breakdown of blood to toxic proteins and portosystemic shunting. Sepsis, azotemia, and electrolyte disturbances that can occur in the context of AVB also tend to worsen HE. Sedation during endoscopy or intubation can also exacerbate HE. Change in mental status caused by HE should be differentiated from delirium resulting from alcohol withdrawal and Wernicke encephalopathy, as the treatment of these disorders is different. Lactulose should be used to treat HE. In a recent study, 70 patients with AVB were randomized to receive either lactulose or no lactulose. Only 5 patients (14%) who received lactulose developed HE, compared with 14 patients (40%) in the nonlactulose group ( P = .03).
Renal Failure
Renal failure develops in 11% of patients with AVB. Hypovolemia, sepsis, liver disease, and nephrotoxic medications contribute to renal failure. The risk of renal failure can be minimized by careful attention to volume status, maintenance of adequate urine output, aggressive treatment of sepsis, and avoidance of nephrotoxic drugs.
Early Antibiotics
Approximately 20% patients of patients with AVB have infection at the time of admission, and an additional 50% acquire infections during their hospital stay.
Infection in AVB is associated with increased mortality and early rebleeding. The most frequent infections are spontaneous bacterial peritonitis (50%), urinary tract infections (25%), and pneumonia (25%), which are often caused by enteric aerobic gram-negative bacteremia. A meta-analysis of 12 trials that compared antibiotic prophylaxis with placebo or no intervention found a significant benefit of prophylactic antibiotic use with regard to all-cause mortality (RR 0.79, 95% CI 0.63–0.98), mortality from bacterial infections (RR 0.43, 95% CI 0.19–0.97), bacterial infections (RR 0.35, 95% CI 0.26–0.47), rebleeding (RR 0.53, 95% CI 0.38–0.74), and days of hospitalization (mean difference −1.9 days, 95% CI −3.8 to 0.02). In the last 2 decades infections caused by gram-positive cocci have markedly increased in the setting of extensive use of invasive procedures, frequent hospital admission, and use of antibiotic prophylaxis.
The use of antibiotics in AVB is therefore recommended to help reduce infections and the risk of rebleeding and mortality. Antibiotics should be started preferably before endoscopy and should be continued for 5 to 7 days. Quinolones have been the most used antibiotic, traditionally because of its ease of administration and low cost, but its use is now limited owing to increasing antibiotic resistance.
A randomized controlled trial of 111 patients with advanced cirrhosis and AVB showed that the probability of developing possible or proven infection is higher with oral norfloxacin than with intravenous ceftriaxone (33% vs 11%, P = .003; 26% vs 11%, P = .03). Therefore, professional societies have recommended that intravenous ceftriaxone should be considered over quinolones in patients with severely decompensated cirrhosis, high prevalence of quinolone resistance, or prior quinolone prophylaxis. Newer antibiotics are being tested for their role in AVB. In a recent study, intravenous cefazolin was shown to be similar to ceftriaxone in preventing infections and reducing rebleeding among Child-A cirrhotic patients after endoscopic interventions for AVB, but inferior in Child-B and Child-C patients. However, more rigorous studies are needed before cefazolin can be recommended. The final selection of antibiotics should be tailored to each individual, with due consideration to the local patterns of antibiotic resistance and patient allergies.
Vasoactive Drugs
The role of vasoactive medications in AVB is well established. This fact was reconfirmed in a recent meta-analysis of 30 randomized controlled trials including 3111 patients with AVB who were treated with vasopressin, somatostatin, or their analogues. The use of vasoactive agents was associated with a significantly lower risk of 7-day mortality (RR 0.74, 95% CI 0.57–0.95; P = .02), improved control of bleeding (RR 1.21, 95% CI 1.13–1.30; P <.001), shorter hospital stay, and lower transfusion requirements. Vasoactive therapy should be started early and continued for 2 to 5 days. Studies comparing the various vasoactive medications have failed to demonstrate a significant difference in efficacy. The choice of vasoactive drug should be made according to local resources and availability.
Timing of Endoscopy
The door to endoscopy time has been a topic of debate. Urgent endoscopy may have the advantages of earlier identification of bleeding, control of bleeding, risk stratification, and prevention of rebleeding, but endoscopy-related complications may compromise these benefits. Few studies concluded that delaying endoscopy in favor of pharmacologic treatment may be as suitable an option, as endoscopic sclerotherapy and pharmacologic therapy were equally effective in controlling bleeding and survival, but the former was associated with more side effects. However, in all these studies sclerotherapy instead of band ligation was performed, which is not the current standard of practice. Moreover, it has now been demonstrated that the combination of endoscopic intervention and pharmacologic treatment is superior to either therapy alone.
Although there is a short delay in endoscopy on the weekends, a study of 36,734 patients presenting with esophageal variceal bleeding (EVB) showed that despite a short delay, there was no difference in mortality. In a contrasting study including 311 cirrhotic patients with EVB, patients were divided into 2 groups of early and delayed endoscopy, taking 15 hours as the cutoff point. Patients in the delayed endoscopy group had a significantly increased risk of in-hospital mortality (odds ratio = 3.67; 95% CI 1.27–10.39; P = .016), independent of the severity of the underlying liver disease, comorbid conditions, hemodynamic status, and endoscopic treatment. In a separate cohort study of 101 patients with AVB, patients presenting with hematemesis and active variceal bleeding observed during endoscopy had the poorest prognosis in terms of early rebleeding and mortality if they received delayed endoscopy. Their 6-week rebleeding and mortality rates were 38.9% and 52.8%, respectively, much higher than the rates of those who received early endoscopy. Contrary to these results, another study showed that the time to endoscopy had no effect on mortality in AVB. However, all patients who were included were hemodynamically stable. Despite conflicting studies, all professional societies have recommended a door to endoscopy time of less than 12 hours in patients with suspected AVB.
Studies have shown that combined treatment with endoscopy and vasoactive drugs is more effective than either treatment alone. In a meta-analysis, combined pharmacologic and endoscopic treatment in comparison with endoscopic treatment alone was significantly more effective for initial control of bleeding (RR 1.12, 95% CI 1.02–1.23) and 5-day hemostasis (RR 1.28, 95% CI, 1.18–1.39). In patients who rebleed, an attempt at repeat endoscopic control should be made. Otherwise, the patient should be considered for TIPS or surgery if liver transplantation is not an option. For patients who survive an episode of AVB, the most common approach for secondary prophylaxis is to combine daily nonselective β-blocker therapy with endoscopic variceal ligation every 3 to 4 weeks until obliteration of varices is achieved. Thereafter, endoscopic surveillance can be performed at varying intervals to determine if and when repeat band ligation is needed.
Rescue Measures
In 10% to 20% of patients, AVB is unresponsive to initial endoscopic and/or pharmacologic treatment. Patients who are clinically stable and have mild bleeding can have a second attempt at endoscopy. If this fails, or if bleeding is severe or the patient is hemodynamically unstable, the patient should be offered rescue treatment.
Transjugular intrahepatic portosystemic shunt
TIPS serves as a side-to-side portosystemic anastomosis, and allows portal decompression without the need for major surgery. Traditionally TIPS has been used as a rescue therapy in patients with EVB who rebleed or do not respond to first-line endoscopic and pharmacologic therapy. Factors associated with poor survival following TIPS include advanced age, emergency TIPS, alanine aminotransferase greater than 100 U/L, bilirubin greater than 3 mg/dL, pre-TIPS encephalopathy that is not related to bleeding, and MELD score. Because it is a procedure with high associated mortality, careful patient selection is needed. One study showed that early mortality is common in patients with a Child-Pugh score greater than 13 undergoing TIPS.
The role of early TIPS in patients with advanced liver disease (Child-Pugh C disease, score 10–13) or HVPG 20 mm Hg or higher has been recently studied in 2 controlled trials. In these selected patients, early TIPS (after the first index bleed) was associated with a significant reduction in failure of treatment and mortality if placed early (within 72 hours), before the patient deteriorates. In the first study, 52 patients with HVPG greater than 20 mm Hg were randomized to receive endoscopic sclerotherapy or uncovered TIPS within 24 hours of admission. Early TIPS placement significantly reduced treatment failure (12% vs 50%; P =.0001), in-hospital mortality (11% vs 38%), and 1-year mortality (35% vs 65%; P <.05). However, the benefits of TIPS may have been overestimated, as the therapy used in the control group was not up to the current standards of endoscopic therapy for AVB. Patients in the control group did not receive combination therapy and the endoscopic treatment used was sclerotherapy, which is not the treatment of choice as per current recommendations. In the other recent study, early TIPS using a polytetrafluoroethylene-covered stent (placed within 72 hours) was compared with optimal medical therapy (endoscopic therapy plus vasoactive medications) in 63 high-risk (Child-Pugh class B with active bleeding or in Child-Pugh class C) patients with cirrhosis and AVB. Patients who were randomly assigned to receive TIPS had a significantly better chance of remaining free of bleeding at 1 year than those who received combined endoscopic and pharmacologic care (97% vs 50%). The rate of survival at 1 year was 86% in the TIPS group, compared with 61% in the medical therapy group. The rate of HE was similar in both groups.
Balloon tamponade
Balloon tamponade is used to temporarily control bleeding in cases of AVB where medical management and endoscopy have failed. Balloon tamponade has been shown to control bleeding in 90% of cases. However, 50% of patients rebleed after removal of the tamponade tube. Complications such as aspiration, esophageal rupture, and necrosis develop in 30% of patients. The tube should be removed within 24 hours to reduce such complications. Balloon tamponade is a life-saving measure to temporarily control bleeding in patients with massive hemorrhage until a more definitive therapy can be undertaken.
Self-expandable metal stents
Self-expandable metal stents (SEMS) control bleeding by compression of the bleeding varices. Similar to balloon tamponade, they are used as a life-saving measure in refractory bleeding. SEMS are inserted over an endoscopically placed guide wire using a stent delivery device without the need for fluoroscopy. Compared with balloon tamponade, these stents can be left in place for up to 2 weeks and can be easily removed by endoscopy. However, they are associated with a high risk of rebleeding after removal. Complications of SEMS include esophageal ulcerations, compression of the bronchi, and stent migration into the stomach. Recent data suggest that a self-expanding covered esophageal metal stent may be an alternative to the Sengstaken-Blakemore balloon in refractory EVB, with the advantage of less severe complications despite longer periods of treatment. More studies are needed to establish the role of SEMS in refractory variceal bleeding.
Surgery
The role of surgery in EVB is limited to patients who fail medical measures and endoscopy. However, today surgical expertise for these procedures is greatly limited. Portal decompressive surgery and esophageal transection were highly effective in achieving hemostasis but were associated with a high mortality of 45% to 75%; such procedures are seldom carried out nowadays. In addition, in patients surviving the bleeding episode, portosystemic shunt surgery significantly increases the incidence of chronic or recurrent HE. In a retrospective study of 82 patients who underwent salvage surgery for variceal bleeding, control of bleeding was achieved in 95% of patients. The in-hospital mortality rate was 15%, and was higher among patients undergoing emergency surgery and in those with cirrhotic portal hypertension. Surgery should be performed in patients with Child-A cirrhosis only. The decision to perform surgery is based on local resources and expertise.
Related posts:
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
