Nonselective beta-adrenergic blockers are the drugs of choice for primary prophylaxis. They lower the portal pressure by causing splanchnic arteriolar vasoconstriction through their β-2 effect and lower cardiac output by blocking β-1 receptors, which reduce portal venous inflow [8]. Propranolol and nadolol are the two most commonly used medications. Their role in preventing the progression of the size of varices and bleeding in small varices is controversial. A multi-center study showed no benefit of using NSBBs compared with placebo in preventing the development of esophageal varices in cirrhotic patients who did not have esophageal varices; on the other hand, there were significant adverse outcomes in that group [9–11]. A recent meta-analysis performed by Qi et al. [12], which included a total of six randomized controlled studies, showed the NSBBs did not show any significant benefit on progression, bleeding events or mortality compared with the placebo group. Yet, some experts advocate using NSBBs in patients with small varices, who are a high risk, as in a Child–Pugh score of B or C, alcoholic liver disease, and in patients who had red wale marks at their initial endoscopic evaluation [13]. Merkel et al. [14] showed in a placebo-controlled trial that using nadolol was effective in reducing the progression and bleeding of small varices, but did not add significant survival benefit.

Nonselective beta-adrenergic blockers play a pivotal role in preventing bleeding in patients with medium and large varices. For esophageal varices to rupture and bleed, HVPG is generally ≥12 mmHg. There is a marked reduction in the risk of bleeding when HVPG was lowered either below 12 mmHg or >20 % from the baseline [15]. It has been shown that either nadolol or propranolol reduces the risk of the first variceal bleed compared with the placebo group (30 % in controls vs 14 % in patients receiving beta-blockers) by lowering the HVPG [8]. Additionally, there was a significant survival benefit in patients who received NSBBs [16]. The starting dose of propranolol is 20 mg twice a day and 40 mg once day for nadolol. The strategy is to reduce the heart rate by at least 20 % from the baseline and to adjust the medication to its maximal tolerated dose. Some 15–20 % of patients cannot tolerate the medication owing to side effects such as lightheadedness and asthma-related symptoms, which lead to complete withdrawal of the medication [17].

Besides nadolol and propranolol, carvedilol is an NSBB with intrinsic alpha-1 adrenergic blocking activity, the use of which for primary prophylaxis against variceal bleeding is becoming more widespread. In a recent study, in patients receiving carvedilol who had not initially responded to propranolol there was a greater reduction in HVPG compared with patients who had responded to propranolol treatment. Not only did patients who had received carvedilol achieved a better hemodynamic response, they also had a lower mortality rate compared with patients treated only with endoscopic variceal banding [18, 19]. More recent data suggest that NSBBs should be avoided in patients with refractory ascites as they result in poor outcomes [20].

To conclude, the use of NSBBs in patients with medium and large varices, in those with high-risk stigmata for bleeding, and in patients with small varices who have not had variceal bleeding results in decreased bleeding episodes and improved survival. Their use is limited to patients who tolerate the medications, requires a hemodynamic response to therapy, and is not advocated in cirrhotic patients with no esophageal varices and those with refractory ascites.

As noted above, all patients with the diagnosis of cirrhosis should undergo screening for esophageal varices [3], but the role of EVB in primary prophylaxis is still controversial. It is clear that EVB is very effective in eradicating esophageal varices and reducing first bleeding events. However, most comparative studies have not shown superiority in reducing bleeding-related mortality compared with NSBBs for primary prophylaxis. A recent Cochrane review by Gluud and Krag [21] included 19 randomized trials comparing EVB with NSBBs for primary prevention. There was a greater reduction in bleeding events in the EVB group (RR = 0.68, 95 % CI: 0.52–0.90); however, EVB was not superior to NSBBs with regard to bleeding-related mortality. Despite this evidence, there is clear agreement that EVB is the option of choice for primary prophylaxis in patients who either cannot tolerate or have a contraindication to NSBBs , those with refractory ascites, and those with high-risk esophageal varices. Of note, trials of endoscopic variceal sclerotherapy for primary prophylaxis do not support its use for primary prophylaxis.

Combination therapy using EVB in addition to NSBBs in primary prevention has been studied with conflicting results [22–24]. A recent retrospective study by Je et al. [25] investigated the long-term efficacy of the combination of EVB and propranolol compared with propranolol alone. A total of 504 patients were enrolled in the study. There was a marked difference in preventing the first variceal bleed in favor of a combination strategy, with no difference in overall mortality between the two groups. On the other hand, Lo et al. conducted a randomized study comparing combination therapy of EVB and nadolol with nadolol alone. There was no difference in bleeding events and bleeding-related mortality in the groups; however, in contrast to the report by Je et al [25]. there were more adverse events in the combined group related to EVB, including esophageal ulcerations [23]. Given the current evidence, there is no clear added benefit of combination EVB and NSBB therapy and it cannot be routinely recommended as the first-line modality for primary prevention.

Patients who survive the first episode of variceal bleeding are at a high risk of recurrent bleeding. This risk is greater than 60 % within the first 2 years, with a higher risk of bleeding within the first 6 weeks after the index bleed and 33 % mortality [15]. In this particular scenario, combination therapy using NSBBs and EVB is the standard treatment and has proved its superiority when compared with either treatment alone. This was shown in a meta-analysis of 9 trials where combination therapy reduced the risk of variceal rebleeding, but not overall mortality, when compared with the use of NSBBs or EVL alone [26]. The added benefit of NSBBs to that of EVB lies in reducing portal pressure, which reduces the risk of bleeding. Additionally, studies demonstrate that NSBBs reduce the risk of spontaneous bacterial peritonitis by 12 %, independent of their hemodynamic effect on HVPG [27].

Once patients recover from the first variceal bleed, treatment should be targeted at eradicating those varices. An endoscopic banding session should be scheduled at 2- to 4-week intervals after the index bleeding episode until the esophageal varices are completely eradicated. This process typically takes 2–4 sessions. Once they are eradicated, the patient should be back in 3 months for a first surveillance, and, if negative, this should be repeated every 6 months. Endoscopic variceal sclerotherapy is also an option that has been shown to decrease rebleeding, but is now infrequently employed (typically in cases in which EVB is not technically feasible) in light of the more favorable complication profile of EVB. The management strategy for secondary prevention is summarized in Table 28.2.

Adverse events of esophageal variceal banding are observed in 14 % of patients. The two most common complications are chest pain and dysphagia, but both are short-lived complications [28]. Post-banding ulcers occur at the banding site in 0.5–3.0 % of cases; these ulcers are shallow in nature and have the tendency to bleed in 5–6.6 % of treated patients. Studies showed that using pantoprazole daily for 9 days following the procedure can significantly reduce the size of post-banding ulcers and their bleeding risk [29]. Life-threatening bleeding associated with post-banding ulceration has been reported and requires urgent attention. Post-banding strictures have also been reported, but are typically easily managed. Endoscopic variceal sclerotherapy is associated with a higher rate of esophageal ulcers and with fever and chest and abdominal pain. It is also associated with risks of perforation, mediastinitis , and portal vein thrombosis.

Patients with liver cirrhosis who present with suspected variceal bleeding should be admitted to the intensive care unit. After rapid stabilization and resuscitation of the patient, including assessing the need for airway protection with a low threshold for intubation and securing adequate intravenous access, urgent endoscopy with therapeutic intent should be undertaken. Blood transfusion should not be undertaken until the hemoglobin falls below 7 mg/dl and caution should be exercised not to over-resuscitate using crystalloids or other blood products so that the portal pressures are not elevated further, hence predisposing to more rebleeding, associated with increased mortality [30]. Once endoscopy is undertaken, EVB is now typically the standard of care if esophageal variceal bleeding is identified. Esophageal variceal sclerotherapy is also effective in this setting.

There is a major established role of using vasoactive pharmacological agents in the acute portal hypertensive bleeding event. Vasopressin and terlipressin have a vasoconstrictive effect on splanchnic circulation and octreotide prevents postprandial elevation and portal pressures. Thus, these agents either lower the portal pressure or prevent a rebound elevation in portal pressure in response to bleeding and subsequently enhance the efficacy of endoscopic therapy. Vasopressin is associated with significant side effects, including potential coronary vasoconstriction, and is thus now rarely used. Terlipressin is very effective, but is not available in the USA, leaving octreotide as the most common agent used in variceal bleeding there. The dose of octreotide is a 50-μg intravenous bolus followed by an infusion of 50 μg/h. Although the optimal duration of infusion has not been well established, given the risk of early rebleeding, which is approximately 50 % in the first 5 days after endoscopic intervention, it is reasonable to continue the octreotide infusion for 5 days, although some studies showed no difference in outcome if octreotide infusion was continued for 2 days or 3 days versus a 5-day strategy [31, 32] In our practice, we commonly continue the octreotide for at least 3 days. It is well-established that the short-term use of prophylactic antibiotics in cirrhotic patients and portal hypertensive bleeding with or without ascites not only decreases the risk of infectious complications, but also improves survival [33]. Antibiotics must be given to patients with liver cirrhosis who suffer from gastrointestinal bleeding irrespective of their ascites status. The use of oral norfloxacin at a dose of 400 mg BID for 7 days or intravenous ceftriaxone in a dose of 1 g/day for 7 days also diminish the incidence of bacterial peritonitis [34].

Gastric varices are present in approximately 17 % of patients with cirrhosis [35]. Although they bleed less frequently than esophageal varices, the episodes of bleeding tend to be more severe, with higher mortality of almost 45 %. The most accepted and well-known classification for gastric varices is the Sarin system (Fig. 28.1). They are classified into: GOV1, esophageal varices that extend below the gastroesophageal junction into the lesser curvature of the stomach; GOV2 extends beyond the gastroesophageal junction into the fundus of the stomach; IGV1 are isolated gastric varices that are limited to the fundus of the stomach and have no contiguity with esophageal varices (Fig. 28.2); and isolated gastric varices in other areas of the stomach are called IGV2. The most prevalent type of gastric varix is GOV1 (75 %), followed by GOV2 (21 %), IGV2 (4 %), and IGV1 (<1 %).