Should children with esophageal varices receive beta-blockers for the primary prevention of variceal hemorrhage ?

Division of Gastroenterology, Hepatology and Nutrition, The Hospital for Sick Children, Toronto, Ontario Correspondence: Dr Simon C Ling, The Hospital for Sick Children, 555 University Avenue, Toronto, Ontario M5G 1X8. Telephone 416-813-8757, fax 416-813-6531, e-mail simon.ling@sickkids.ca Received for publication June 7, 2005. Accepted June 14, 2005 SC Ling. Should children with esophageal varices receive betablockers for the primary prevention of variceal hemorrhage? Can J Gastroenterol 2005;19(11):661-666.

V ariceal hemorrhage is a life-threatening complication of portal hypertension that occurs commonly in children with chronic liver disease or portal vein obstruction (1)(2)(3)(4)(5)(6)(7)(8).In children with biliary atresia, the incidence of variceal hemorrhage ranges from 17% to 29% over a five-to 10-year period (3)(4)(5) and is 50% in children who survive more than 10 years without liver transplantation (6).Among 50 children with esophageal varices, primarily due to cirrhosis, who were prospectively followed and not offered active treatment to prevent variceal bleeding, 42% suffered upper gastrointestinal hemorrhage during a median 4.5 year follow-up period (2).For children with portal vein thrombosis, the available studies suggest that up to 50% suffer a major variceal hemorrhage by 16 years of age (1).The mortality rate from gastrointestinal bleeding ranges from 2.5% to 20% in children with portal hypertension (2,4,7,8).Prevention is, therefore, an important goal.
Although the majority of North American pediatric hepatologists report a willingness to use beta-blockers for primary prophylaxis (9), this approach in children is controversial due to the lack of pediatric data and the wariness among pediatricians of extrapolating results of adult studies to children (10).The present review discusses the mechanism of action of betablockers in the context of the pathophysiology of portal hypertension, compares the evidence for pharmacological and endoscopic prophylaxis in both adults and children, and considers the additional factors that determine whether nonselective beta blockade is an appropriate intervention for children at risk of variceal hemorrhage.

THE PATHOPHYSIOLOGY OF PORTAL HYPERTENSION AND THE MECHANISM OF ACTION OF NONSELECTIVE BETA-BLOCKERS
In portal hypertension caused by cirrhotic liver disease, vascular resistance in the portal system is elevated by distorted hepatic architecture, intrahepatic small vessel thromboses and increased intrahepatic vascular tone arising from the actions of vasoactive substances on myofibroblasts, perisinusoidal activated stellate cells and vascular smooth muscle cells (11)(12)(13)(14)(15)(16).Whereas the intrahepatic vascular bed is constricted, splanchnic arteriolar dilation exacerbates portal hypertension by increasing portal venous inflow (17)(18)(19).
Varices develop when the hepatic venous pressure gradient (HVPG), a measurement obtained by transjugular cannulation of the hepatic veins (20), is elevated above 10 mmHg to 12 mmHg (21).Recent evidence suggests that active angiogenesis may contribute to the development of varices in addition to the effect of increased pressure and flow within potential portosystemic anastamoses (22).Variceal bleeding occurs when increased vein diameter, decreased wall thickness and increased intraluminal pressure elevate variceal wall tension beyond the maximum tolerable threshold (19).Clinical predictors of variceal hemorrhage in adults include the severity of liver disease measured by the Child-Pugh score, the presence of ascites, the size and appearance of the varices at endoscopy, as well as the degree of elevation of HVPG (23).

STUDIES OF BETA-BLOCKERS IN CHILDREN WITH PORTAL HYPERTENSION
Following early reports of potential benefit more than 20 years ago (27,28), nonselective beta-blockers were shown to be superior to placebo or no active treatment in the prevention of first variceal hemorrhage in several randomized controlled trials in cirrhotic adults (29)(30)(31)(32)(33)(34)(35)(36)(37).Meta-analyses of these studies confirm that beta-blockers reduce the incidence of variceal hemorrhage by approximately 50% (38)(39)(40).However, in light of numerous examples of drugs whose efficacy, pharmacokinetics or adverse event profiles differ significantly between adults and children, care must be taken before extrapolating to children the evidence for efficacy of betablockers in cirrhotic adults (41,42).Biliary atresia is the most common cause of advanced liver disease in children and largely accounts for the excess of biliary disease among cirrhotic children, whereas alcoholic liver disease accounted for the majority of the adults included in the randomized studies of beta-blocker prophylaxis.Although there is no difference in response to betablocker prophylaxis between alcoholic and other cirrhotic adults (30)(31)(32)(33)36,43), it is difficult to estimate whether a oneyear-old infant with biliary atresia will respond to beta-blockers similarly to a 60-year-old adult with alcoholic cirrhosis.Pediatric data are, therefore, needed.
The ability of propranolol to reduce portal pressure was initially studied in 13 children with portal hypertension who received 2.1 mg/kg/day to 8.0 mg/kg/day of propranolol that induced an average decrease in heart rate of 25% of baseline (44).Four children with decompensated cirrhosis showed little or no response, and overall, the splenic pulp pressure (a proxy for portal venous pressure) decreased by a mean of only 13%, from 387 mmH 2 O to 337 mmH 2 O.This suggests a limited efficacy of propranolol in these children, although the value of this measurement in predicting variceal hemorrhage is untested.
Clinical experience with propranolol was reported in a retrospective uncontrolled study from Boston, USA, which included a mixed group of 21 children: 19 children had cirrhosis, two had extrahepatic portal vein obstruction, 17 children received primary prophylaxis and four received secondary prophylaxis (45).The propranolol dosage schedule aimed to reduce the heart rate by 25% (achieved in 17 children).Seven of 21 children (33%) had upper gastrointestinal hemorrhage during a median follow-up period of three years.All adverse events (including rash, dizziness, bradycardia, hypotension and depression) were classified as mild and transient, and none required discontinuation of therapy (45).
Additional uncontrolled retrospective studies of propranolol usage in children with portal hypertension have been reported from Turkey.Seven of 45 children (15.6%) and one of 10 children (10%) with cirrhosis, who were not screened for varices, suffered upper gastrointestinal hemorrhage while receiving propranolol for primary prophylaxis during a median five years of follow-up in two studies from different Turkish centres (46,47).Bleeding was less common in children with Child's class A cirrhosis (7.1%) than in those with Child's class B and C combined (29.4%) (46).The incidence of adverse effects was not reported.
Overall, the proportion of children that bled in these studies ranged from 2% to 11% per year of follow-up; no data from untreated control groups are provided (Table 1).In the studies of the natural history of portal hypertension in children that are summarized above, between 2% and 9% of children who

EIS Endoscopic injection sclerotherapy; GI Gastrointestinal; PHT Portal hypertension
did not receive prophylactic treatment suffered gastrointestinal hemorrhage during each year of follow-up.The efficacy of nonselective beta-blockers for the prevention of variceal bleeding in children, therefore, remains unclear, and the adverse event profile associated with this indication has yet to be adequately defined.

POTENTIAL FOR HARM: REPORTED AND THEORETICAL RISKS OF BETA-BLOCKERS IN CHILDREN WITH PORTAL HYPERTENSION
Children who receive nonselective beta-blockers may develop adverse effects, including bradycardia, heart failure, hypotension, conduction disorders, bronchospasm and hypoglycemia following fasting.However, these drugs are generally well tolerated; no clinically significant adverse events were reported in nine randomized controlled trials that included 309 children of whom 188 received propranolol for indications other than the prevention of variceal hemorrhage (48)(49)(50)(51)(52)(53)(54)(55)(56).Although there is no excess mortality among adults who bleed while receiving prophylaxis (29)(30)(31)(32)(33)(34)(35)(36)(37)(38)(39)(40), young children respond differently to shock by relying more heavily on tachycardia due to their relatively fixed stroke volume (57).There is an important theoretical concern that, by limiting tachycardia, nonselective beta-blockers may impair tolerance of hypovolemia and worsen the outcome from variceal hemorrhage.There appear to be no reports that suggest increased morbidity or mortality following hemorrhage in children receiving appropriate therapeutic doses of nonselective beta-blockers, although the large controlled studies required to show such an effect have not been undertaken.

HOW MUCH BETA-BLOCKER?
If beta-blockers are to be used in children with portal hypertension, an effective and safe dosing regimen must be chosen.Heart rate is the most commonly used clinical end point for dosage regulation of beta-blockers in the prevention of variceal hemorrhage in adults.In most studies, the beta-blocker dose is adjusted until the heart rate is reduced to 75% of its baseline value, an arbitrary threshold chosen by Lebrec et al (27) in their original study.However, there is only a weak correlation between the effects of nonselective beta blockade on heart rate and portal hemodynamics, and variceal hemorrhage occurs in some cirrhotic adults taking beta-blockers in spite of an adequate heart rate response (44,58,59).The routine use of HVPG measurements improves the prediction of response to beta-blocker therapy; a decline in HVPG to less than 12 mmHg, or by more than 20% of its baseline value, predicts a significantly lower risk of hemorrhage (0% to 7% incidence) compared with patients who fail to achieve this reduction in HVPG (33% to 41% incidence) (60)(61)(62).The place for routine clinical use of this relatively invasive but reportedly safe test is the subject of much debate and has yet to be tested in prospective trials (63)(64)(65)(66).
While such studies are awaited, common clinical practice among adult hepatologists is to prescribe enough beta-blocker to reduce the heart rate to 75% of its baseline value.The reported experience of propranolol usage in children with portal hypertension follows a similar strategy, with most children requiring between 1 mg/kg/day and 2 mg/kg/day to achieve the required reduction in heart rate; the risk of bleeding is possibly greater in those receiving less than 1 mg/kg/day (45).
Unfortunately, there are no pediatric data to describe the splanchnic hemodynamic response to beta blockade or the optimum dosage schedule that gives maximum efficacy (in reducing hemorrhage episodes) and minimum risk (especially impaired tolerance of shock).

THERAPEUTIC OPTIONS FOR PROPHYLAXIS OF VARICEAL BLEEDING IN CHILDREN
Endoscopic variceal ligation (EVL) provides an alternative approach to the prevention of variceal bleeding that avoids the need for long-term medication, is often suitable for patients with side effects or contraindications to beta-blockers and is at least as effective as nonselective beta-blockers in cirrhotic adults (67)(68)(69)(70)(71)(72)(73)(74)(75)(76).
The evidence supporting endoscopic treatment of varices in children is more limited, although many pediatric gastroenterologists are willing to use this approach for primary prophylaxis (9).The only randomized controlled trial of primary prophylaxis of variceal hemorrhage in children studied the effect of endoscopic injection sclerotherapy (EIS) in 100 children of median age 4.3 years in Brazil (2).Most of the children had biliary atresia or another parenchymal liver disease, while nine had portal vein thrombosis.Statistically significant differences were shown between the EIS group and control group (who received no active treatment) in the all-cause gastrointestinal hemorrhage rate (24% versus 48%, respectively) and the variceal hemorrhage rate (6% versus 42%, respectively) (2).Mortality rates were similar.The study excluded children who had the largest (grade 4) varices, and set no minimum requirement for variceal size, but the hemorrhage rate in the control group was still greater than 40%.The report does not provide details of the method for selection of patients for inclusion in the study or of the timing and method of random assignment, including the process to avoid allocation bias.The interpretation of this study and the application of its results to clinical practice are, therefore, difficult.
EIS proved ineffective for primary prophylaxis in adults and has been largely replaced by EVL.Several uncontrolled case series of EVL in children suggest that it is safe, although there is little experience of its use for primary prophylaxis (77-80) (Table 1).In one series of 31 children with varices who underwent EVL, there were no episodes of variceal hemorrhage and only one incidence of bleeding from portal gastropathy, suggesting that this therapy may offer effective prevention for children with both intrahepatic and prehepatic portal hypertension (78).
As with the pediatric beta-blocker studies, the lack of control groups means that there are inadequate data to base any recommendation concerning the use of an endoscopic approach to primary prophylaxis of variceal hemorrhage in children.For adults, the relative efficacy and practicality of different pharmacological, endoscopic or combined approaches to prevention of hemorrhage is vigorously debated (81)(82)(83).However, the use of nonselective beta-blockers is more cost effective than EVL (84,85) and, therefore, beta-blockers remain the most widely recommended first-line prophylaxis for adult patients with varices that are at risk of bleeding (64,(86)(87)(88)(89).

INDIVIDUAL PATIENT CHARACTERISTICS
It is likely that postpubertal teenagers with portal hypertension respond to beta-blockers and bleeding like adults, whereas the theoretical concerns about adverse outcomes from hemorrhage while receiving beta-blockers apply primarily to infants and younger children.Age may, therefore, be an important factor in the decision to prescribe beta-blockers.
Children and families differ in their perception and management of anxiety and uncertainty.Some accept endoscopic screening for varices without hesitation and are reassured by documentation of the absence of esophageal varices or are encouraged to make clear emergency action plans if large varices are identified, as well as make more informed decisions about such things as vacation travel or relocation to remote areas.Some may prefer pharmacological prophylaxis to a series of EVL procedures with its attendant need in children for general anesthesia.
In making a decision about a treatment plan for which the evidence is unclear, families will differ in their understanding and interpretation of the information provided to them, in their acceptance of risk and their perception of possible benefits.Most clinicians recognize that these differences among families form an important part of treatment decisions and accept that doctors do not always know what is best for each individual child that they treat; their practice has, therefore, moved toward fuller discussion with families and facilitation of a greater role for the child and family in the decision-making process (90).

SUMMARY
Nonselective beta-blocker therapy reduces the incidence of first variceal hemorrhage in cirrhotic adults with large varices and is more cost-effective than EVL.Unfortunately, 25 years after the dawn of pharmacological prevention of variceal hemorrhage in adults, there remains a pressing need for randomized controlled studies in children at risk of fatal variceal hemorrhage.Meanwhile, pediatric gastroenterologists who wish to prescribe nonselective beta-blockers for primary prophylaxis must be mindful of the lack of pediatric evidence and the potential age-related adverse effects.They should carefully consider each patient's circumstances and discuss the situation fully with the patient and parents, whose opinions of what is best should be carefully heard.

Ling
Can J Gastroenterol Vol 19 No 11 November 2005 662

TABLE 1
Studies reporting the use of therapies to prevent the first variceal hemorrhage in children