Secondary prophylaxis in patients who have experienced portal hypertensive bleeding


  • Potential conflict of interest: Nothing to report.


Watch the interview with the author

Watch a video presentation of this article

Answer questions and earn CME


nonselective beta-blocker


transjugular intrahepatic portosystemic shunt

Portal hypertension is an inevitable result of the progression of liver fibrosis, which is a feature of liver cirrhosis. As a result of portal hypertension, a collateral circulation forms; this drains blood that cannot go through the liver and is otherwise known as portosystemic collateral circulation. Among portosystemic collaterals, the gastroesophageal one is particularly relevant because it leads to variceal formation.

Gastrointestinal bleeding due to the rupture of esophageal varices used to be the most common cause of death in patients with cirrhosis. Acute mortality resulting from variceal hemorrhage has decreased from 43% (in 1980) to 14% (in 2000) over the last 20 years (Fig. 1).[1] In addition, in subjects who survive the first bleeding episode, the likelihood of experiencing rebleeding within 1 year used to be approximately 50% when no prophylactic treatment was available, and now the rate is approximately 15%.[1]

Figure 1.

Time course of in-hospital mortality, use of NSBB in the prevention of rebleeding, and rebleeding rates from 1980 to 2000 in the same hospital. The data were taken from Carbonell et al.[1]

Over the last 30 years, research has allowed us to define effective therapeutic options for preventing rebleeding. Based on these data, all current guidelines agree that patients who survive the first episode of variceal hemorrhage should receive treatment to prevent rebleeding.[2, 4]

The history of secondary prophylaxis for variceal hemorrhaging is fairly complex and has developed along 2 parallel routes or strategies: (1) medical treatment, which is aimed at reducing portal hypertension as the main determinant of the formation and rupture of varices, and (2) endoscopic treatment, which is aimed at directly removing esophageal varices, which are the ultimate cause of hemorrhage.

In the 1980s, on the basis of pioneering studies by Lebrec[3] and others, the medical treatment of portal hypertension with nonselective beta-blockers (NSBBs) was defined progressively, and the first positive results were obtained within the setting of the prevention of rebleeding. Over the same period, also in relation to technical developments leading to flexible (rather then rigid) endoscopy, endoscopic sclerotherapy for esophageal varices was also investigated.

Clinical studies of both treatments immediately highlighted obvious advantages in comparison with placebos or no treatment. In contrast, studies comparing the 2 treatments, albeit numerous, have been heterogeneous and of poor methodological quality, leading to inconclusive meta-analyses. Overall, sclerotherapy seems to have some advantage over medical treatment, but it is also accompanied by a higher incidence and severity of side effects.

Studies evaluating the combination of NSBB and sclerotherapy have failed to show significant advantages in comparison with either treatment alone. While this issue was being debated, both types of treatments were perfected: medical treatment was improved by the introduction of the association between beta-blockers and nitrates,[5] and sclerotherapy was replaced by band ligation, which is easier to perform, is better standardized, is safer, and leads to fewer side effects.[6] The use of nitrates alone is not indicated. Comparisons of beta-blockers (with or without nitrates) and band ligation have shown comparable results (Fig. 2), whereas the combination of the 2 treatments has been shown to be more effective than each treatment alone in 2 trials.[11, 12] Based on these observations, guidelines from the American Association for the Study of Liver Diseases and the Baveno V consensus workshop conclude that the combination of beta-blockers and band ligation is the best option for secondary prophylaxis (or prophylaxis for rebleeding).[2, 4] It is of the utmost importance that both treatments be performed with meticulous care: the medical treatment by reaching the maximum tolerated beta-blocker dose and the endoscopic treatment by following up the first ligation session with repeat ligation sessions every 1 to 2 weeks until complete eradication. Subsequent endoscopic follow-up needs to be performed at 1 to 3 months and then at 6 to 12 months so that further ligations can be performed if varices that are big enough to be ligated reappear.

Figure 2.

Meta-analysis of beta-blockers plus isosorbide mononitrate versus endoscopic band ligation for the prevention of rebleeding. The outcome was rebleeding.

The observation that the combination of medical and endoscopic treatments is more effective than either treatment alone is reasonable because the 2 types of treatments act on different aspects of the disease: portal hypertension and the varices themselves. In other instances, combinations of treatments are no more effective than a single treatment because their side effects tend to accumulate, and there is no guarantee that patients who do not respond to one treatment are more likely to respond to the other one. However, within the context of variceal bleeding, the reasons for the failure of the treatments are different (Table 1). Medical treatment fails because the reduction in the portal pressure is insufficient, because compliance is poor, or because the treatment is withdrawn on account of side effects. In contrast, endoscopic treatment fails because varices that have not yet been ligated rupture, because there are bleeding ulcers in the area of previous ligations, because there are technical problems, because varices reappear early after eradication, or because there is a variceal rupture before endoscopic follow-up. Because the reasons for failure are different and at least in part are covered by one treatment or the other, it is reasonable to expect that the combination of the two treatments is more effective, as shown in some of the trials.

Table 1. Reasons for Failure in the Prevention of Rebleeding from Esophageal Varices
Medical treatment
Insufficient decrease in portal pressure
Low compliance
Withdrawal because of side effects
Endoscopic treatment
Before variceal eradication
Technical problems
Posttreatment ulcers
Bleeding from varices that are not ligated yet
Withdrawal from treatment
After variceal eradication
Recurrence and rupture before follow-up endoscopy

If the combination of medical and endoscopic treatments fails, the guidelines agree in suggesting a transjugular intrahepatic portosystemic shunt (TIPS)[13] or, where it is still in use, a surgical portocaval shunt in patients classified as Child class A. The indication for TIPS only as a salvage treatment (after the failure of combined medical and endoscopic treatments) is based on the fact that studies comparing TIPS and endoscopic treatment or treatment with beta-blockers plus nitrates as a first-line choice show a lower number of repeat bleeds with TIPS but also a higher incidence of hepatic encephalopathy, similar mortality rates, less common improvements in the Child-Pugh score, and higher costs.[14, 15] In more detail, according to meta-analytic data, endoscopic treatment is associated with more frequent rebleeding (odds ratio = 3.8, 95% confidence interval = 2.8-5.2), less frequent hepatic encephalopathy (odds ratio = 0.43, 95% confidence interval = 0.30-0.60), and similar mortality rates (odds ratio = 0.97, 95% confidence interval = 0.71-1.34) in comparison with TIPS. Thus, it seems reasonable to use TIPS only as a salvage treatment after the failure of the first-line choice.

Exceptions may be represented by patients with contraindications / severe side effects of beta-blockers (Table 2), in whom only endoscopic treatment is indicated, and the rare patients who do not want or cannot undergo band ligation, who can be treated with beta-blockers plus nitrates.

Table 2. Contraindications and Main Side Effects of NSBBs in Cirrhosis
Heart failure
Atrioventricular heart block
Sinus bradycardia and other hypokinetic arrhythmias
Bronchial asthma and chronic obstructive pulmonary disease
Peripheral arterial disease with trophic lesions
Diabetes requiring insulin treatment
Defined allergy to the drug
Side effects
Asthenia, dizziness, arterial hypotension
Tremor, nightmares, depression, impotence
Heart failure
Raynaud's phenomenon

The prevention of rebleeding from gastric varices should be based on endoscopic injection of tissue adhesives up to the eradication of the varices or TIPS. It is not clear which of the two options is preferable. Finally, because of the negative prognostic effects of variceal hemorrhage, all patients who survive a bleeding episode and have no obvious contraindications should be directed to a transplant center for the evaluation and transplantation work-up.

In conclusion, the prevention of rebleeding is a cornerstone of the management of portal hypertension in patients with cirrhosis. In different clinical settings, there is a role for pharmacological, endoscopic, and interventional treatment. It is reasonable that patients will receive the best treatment in centers in which the entire spectrum of management strategies is available with comparable technical skill and experience.