Advances in the Treatment of Complications of Cirrhosis and Portal Hypertension-

[Guest guest]

Advances in the Treatment of Complications of Cirrhosis and Portal

Hypertension-Variceal Bleeding

Disclosures

K. Herrine, MD

Introduction

Few patients are as ill and difficult to manage as those with

cirrhosis. From subtle abnormalities, such as minimal hepatic

encephalopathy, to the dramatic presentation of the patient with

variceal hemorrhage, the complications of cirrhosis represent a

series of ongoing challenges. This year's meeting of the American

Association for the Study of Liver Diseases (AASLD) was replete with

interesting and important data regarding these topics. The theme of

using prognostic studies to predict outcome continues to be expanded.

Newer agents to prevent first hemorrhage and to reduce bleeding

recurrence were investigated. Insights into changes in the natural

history of portal hypertensive bleeding as gained over the past 3

decades were also presented. New methods to screen for esophageal

varices were subjected to critical review. Additionally,

investigations into the diagnosis and management of hepatic

encephalopathy were reported. This clinical overview summarizes some

of the more important new concepts and data pertaining to these

topics, as presented during these meeting proceedings.

Variceal Bleeding

Hemodynamics

The prevention of portal hypertensive bleeding has been the topic of

many investigations, which is reasonable, given that once bleeding

occurs, rebleeding, morbidity, and mortality are unacceptably

frequent. Previous work by the Portal Hypertensive Collaborative

Group demonstrated that baseline hepatic venous pressure gradient

(HVPG) is predictive of gastroesophageal varices. The aim of their

current study, presented during the Presidential Plenary Session at

this year's AASLD meeting, was to evaluate whether longitudinal

measurement of HVPG was predictive of the development of

gastroesophageal varices in patients with compensated cirrhosis.[1]

This large cohort of 213 patients comprised individuals with

cirrhosis and HVPG > 5 mmHg, but without gastroesophageal varices;

subjects were randomized to receive timolol (a nonselective beta-

blocker) or placebo. Measurement of HVPG and endoscopy were performed

at baseline and yearly, with a median follow-up of 55 months. The

primary endpoint was the development of gastroesophageal varices or

outright variceal hemorrhage. Of the 154 patients who had paired HVPG

measurements at baseline and at 1 year, 45% had a reduction in HVPG

of greater than 10%. This group was significantly less likely to

reach a primary endpoint than the group with a lesser reduction in

pressure. Moreover, subjects who received timolol were significantly

more likely to have a reduction in HVPG greater than 10%. The

investigators concluded that nonselective beta blockade should be

employed with the intention to reduce HVPG in order to decrease the

risk of variceal hemorrhage. This ongoing trial is sure to provide us

with important details regarding the natural history of portal

hypertension and its prevention. Noninvasive surrogates of HVPG are

needed to more effectively apply these pressure reduction guidelines.

Nonselective beta-blockers have been the mainstay of the

pharmacologic approach toward portal decompression in patients with

chronic liver disease. It has been suggested that angiotensin II may

be involved in the pathogenesis of portal hypertension in cirrhosis.

Following up on reports that angiotensin-receptor blockers may

provide another option for lowering portal pressures, Gustavo and

colleagues[2] from Argentina used direct puncture of esophageal

varices (in the setting of treating recent variceal bleeding) to

measure esophageal pressure, variceal pressure, estimated varix

radius, and variceal pressure gradient before and after 14 days of

treatment with losartan* vs placebo. They found no significant change

in variceal pressure of portal blood flow in the treatment group when

compared with placebo. These findings, together with the mixed

results reported in the literature to date, may be explained by

recently described angiotensin II type-1 receptor gene polymorphisms.

To date, the use of angiotensin-receptor blockers can be considered

an efficacious therapy in the management of portal hypertension.

Epidemiology

Data regarding outcome in patients with variceal hemorrhage are truly

grim, with a large percentage of patients dying or rebleeding in most

series. It has been anecdotally observed that various elements in the

medical armamentarium targeting this clinical problem may have had a

positive effect over the past 3 decades. During this year's meeting

proceedings, Stokkeland and colleagues[3] from the Karolinska

Institute in Sweden present data that appear to confirm this trend.

Using the robust Hospital Discharge Register and the Causes of Death

Register, these investigators reviewed data from 1967 to 2002 to

identify all patients with esophageal varices. The results were

encouraging: The 5-year survival in men younger than age 50 years has

increased from 31% in the first decade reviewed to 49% in the last

decade reviewed. Similar improvements were seen in older cohorts of

both men and women. When controlling for other potential causes of

this improvement in survival, the investigators concluded that the

benefit was based on improvement in mortality from esophageal

varices, not a confounding etiology. Although such a study design

cannot identify the reasons for the survival improvement, it seems

reasonable to assume that better treatment for acute variceal

hemorrhage and in prophylaxis has had an effect.

Diagnosis/Screening

If prophylaxis of varices has led to improvement in mortality over

the past 30 years, what is the most effective strategy for achieving

such prophylaxis? The first step in any such program must be to

identify those patients at highest risk for and most likely to

benefit from intervention. Currently, the standard of care is to

perform endoscopic screening in patients with cirrhosis to identify

those with large esophageal varices considered in need of

nonselective beta-blocker therapy. Two reports described potential

alternatives to this invasive approach.

de Franchis from Milan and his international colleagues[4] described

their findings with a wireless capsule endoscopic device used for

screening and surveillance of esophageal varices, during the parallel

session on clinical portal hypertension. Thirty-two patients

underwent both capsule endoscopy and esophagogastroduodenoscopy

(EGD); the studies were performed within 48 hours of each other.

Interobserver agreement for the presence or absence of varices was

similar between EGD and capsule endoscopy (0.701-0.901 and 0.677-

0.761, respectively). Using EGD as the standard, capsule endoscopy

provided a sensitivity for esophageal varices of 100%, a specificity

of 89%, positive predictive value of 96% and negative predictive

value of 100%. Capsule endoscopy was slightly less specific for

detection of portal hypertensive gastropathy, but otherwise showed

similar performance. The number of patients with gastric varices in

this pilot study was too small to reach a conclusion. A larger study

is being designed to confirm these encouraging results.

In a similar vein, Kim and colleagues[5] from Korea described their

experience with multidetector CT (MDCT) compared with EGD for

detection of esophageal varices in 90 patients with cirrhosis. MDCT

can provide high-quality 2-D axial images as well as virtual

endoscopic images. As in the above study with the capsule endoscopic

device, MDCT performed well. Interobserver agreement between EGD and

MDCT esophagography was excellent, with MDCT demonstrating a

sensitivity of 93% and specificity of 97% for the prediction of high-

risk varices. In addition, this imaging modality was able to provide

simultaneous hepatocellular carcinoma screening, and 10 such

(extraesophageal) lesions were discovered during this trial.

Potential downsides to this diagnostic strategy include the need for

intravenous contrast and the insufflation of the esophagus with a 16F

nasal cannula. These pilot data will need to be confirmed before it

is determined how this modality will fit into practice patterns

regarding variceal screening.

Treatment

When prophylaxis fails and variceal hemorrhage occurs, the clinician

must employ strategies for controlling hemorrhage and for prevention

of recurrent bleeding. Endoscopic control of hemorrhage is effective,

and most often is used in combination with pharmacologic agents

designed to lower portal pressure. Salih and colleagues[6] from

Pakistan, during the clinical portal hypertension parallel session,

reported their interim data on terlipressin (vasoconstrictor) vs

octreotide (vasoactive agent) in combination with endoscopic variceal

band ligation. In this large cohort of 209 patients with acute

variceal hemorrhage, 107 received terlipressin* and 102 received

octreotide.* The treatment groups had similar rates of bleeding

control, red blood cell transfusion, and mortality. The length of

hospital stay and total cost were significantly less for the

terlipressin group. It is not clear whether cost and length-of-stay

data can be generalized to other geographic and practice settings.

Although terlipressin is not available in the United States, this

agent appears useful not only for the control of varices but also as

a pharmacologic agent with some promise in the setting of hepatorenal

syndrome. Perhaps findings such as these reported by Salih and

colleagues will help encourage the further availability of this agent.

The portal hypertension group from Barcelona continued to make

important contributions to the field of hepatology with its report

describing predictive indices for failure of variceal treatment.[7]

Such a tool can be helpful in determining the level of care and in

determining transplantation decisions in this very ill patient

population. The investigators described 287 patients with acute

variceal hemorrhage, treated with endoscopic therapy and

somatostatin* infusion. HVPG, MELD (model-for-end-stage liver

disease), and Child-Pugh scores were recorded. Multivariate analysis

found a 6-week survival disadvantage in those patients with HVPG > 16

mmHg, MELD > 8, and ascites. Despite this high-technology approach,

the authors pointed out that MELD and HVPG measures were not better

than Child-Pugh score in predicting therapeutic failure. Bambha and

colleagues[8] from the Mayo Clinic, Rochester, Minnesota, looked at

the utility of MELD in predicting survival in patients with acute

variceal hemorrhage. Using their data from a large lanreotide* study

for acute variceal hemorrhage, the investigators described the

clinical characteristics of those patients who died within 6 weeks.

On arrival at hospital, " eyeball " assessments of Child-Pugh class as

well as the clinical detection of ascites were useful predictors of

poor outcome. Once laboratory data became available, MELD scoring was

more accurate in making the prediction of poor outcome. As the

admissions progressed, transfusion of more than 1 liter of packed red

blood cells was also predictive of poor outcome (ie, predictive of 6-

week mortality in patients with acute variceal hemorrhage). These 2

studies demonstrated that both low-tech and high-tech solutions can

be used to stratify high-risk bleeders in order to aid management.

Park and his colleagues[9] from Korea presented their data regarding

the role of antibiotics in the prevention of variceal rebleeding. In

this cohort of 120 patients, subjects were randomized to receive

prophylactic antibiotics vs " on-demand " (ie, when infection became

evident) antibiotics at the discretion of the clinician. It is not

surprising that the rate of infection was found to be lower in the

prophylactic group. Less intuitive was that subjects given

prophylactic antibiotics were far less likely to have rebleeding from

varices (11/62 vs 23/58; P < .008). No survival benefit was seen.

Although the pathophysiologic mechanism behind this observation is

not clear, these findings add to the current evidence supporting the

use of antibiotic prophylaxis at the time of variceal hemorrhage.

Hepatic Encephalopathy

Hepatic encephalopathy has been recognized as a complication of liver

disease since the earliest days of medicine. Recognition of the

condition does not, at first glance, appear difficult, but

differentiation of mental status changes from other etiologies is

important. Subtle changes of subclinical encephalopathy should also

be sought, because treatment can improve patient performance and

quality of life. During this year's AASLD meeting, data were

presented offering potential enhancements to our standard therapeutic

arsenal. In addition, a closer look at nonabsorbed antibiotics was

undertaken.

Pathophysiology

Hrycewycz and colleagues[10] from Toulouse, France, investigated the

possible contribution of protein malnutrition and zinc deficiency in

the exacerbation of clinical hepatic encephalopathy. Sixty-seven

patients were evaluated by Child-Pugh class, MELD score, the presence

of ascites, the presence of asterixis, and anthropomorphic measures.

Zinc was measured in 2 forms (serum and globular). Of those patients

with cirrhosis, two thirds were alcoholic and the majority were men.

This was an ill cohort, with 43% of patients having Child-Pugh class

C cirrhosis. The investigators found that 88% of cirrhotic patients

had low serum zinc. Low serum zinc was correlated with acute hepatic

encephalopathy, hypoalbuminemia, and nutritional status. Although

zinc repletion was not studied, they appropriately surmised that such

a study should be undertaken. As any clinician who cares for the

patient with hepatic encephalopathy recognizes, attention to detail

can make a significant difference in outcome. It is in this context

that Guevara and colleagues[11] from Barcelona presented their

findings on the potential role of anemia in this syndrome. The aim of

this study therefore was to assess the relationship between anemia

and cognitive function impairment in the setting of cirrhosis. Using

psychometric testing and clinical parameters, these investigators

found that hematocrit was independently correlated with psychometric

tests, specifically the trail-making test, digit symbol values, and

word-list learning. Whether correction of anemia would lead to

clinical improvement remains to be determined.

Diagnosis

The difficulty in diagnosing subclinical or minimal hepatic

encephalopathy is due to the labor-intensive battery of

neuropsychiatric testing that underlies the process. Mardini and

colleagues[12] from Newcastle Upon Tyne, United Kingdom, described

their experience with a computerized assessment of hepatic

encephalopathy compared with the traditional paper and pencil-based

neuropsychological methods. These investigators gathered Child-Pugh

scores and blood ammonia levels from 44 patients with cirrhosis, and

then performed standard " paper-and-pencil " assessment of hepatic

encephalopathy as well as a computerized assessment battery. They

found no association between paper-and-pencil test scores and either

Child-Pugh score or blood ammonia levels, but the computerized

battery did show significant association with Child-Pugh score and

blood ammonia levels. Given the increasing availability of computer

terminals in office settings, this approach may have applicability

not only in research but also in the provision of clinical care.

Treatment

Sleep disturbance is a common clinical manifestation of hepatic

encephalopathy. Spahr and colleagues[13] from Switzerland reported

their findings regarding the safety and efficacy of a histamine H1

blocker in managing sleep alterations in patients with cirrhosis and

hepatic encephalopathy. On the basis of a published report that sleep

disturbance in hepatic encephalopathy is associated with brain

histaminergic hyperactivity, these investigators hypothesized that

the H1 histamine antagonist hydroxyzine* may improve sleep physiology

in this group of patients. Their study group comprised 35 cirrhotic

patients with a mean age of 55 years and a mean Child-Pugh score of

8.6 with either minimal (n = 31) or episodic (n = 4) hepatic

encephalopathy. These individuals had chronic sleep complaints but no

depression or use of benzodiazepines. Hydroxyzine, administered 25 mg

at bedtime, resulted in improved sleeping, as measured by nighttime

activity and sleep efficiency, as compared with placebo. No adverse

effects were reported. This carefully designed study represents the

type of research that will supersede many of our anecdotal approaches

to the management of hepatic encephalopathy.

With the publication of a meta-analysis in the British Medical

Journal in late 2004 that called into question the efficacy of

lactulose (an osmotic cathartic) for the management of hepatic

encephalopathy,[14] a secure therapeutic world was challenged. The

authors of that study concluded that " nonabsorbable disaccharides

should not serve as comparator in randomised trials on hepatic

encephalopathy. " It is in this setting that the recent US release of

rifaximin,* a nonabsorbed oral antibiotic, was reviewed in a meta-

analysis conducted by Bass and colleagues[15] from San Francisco.

Using a MEDLINE search strategy as well as other sources, these

investigators analyzed 11 trials that included the use of rifaximin

in the treatment of hepatic encephalopathy. These data suggested a

beneficial effect of this compound ( " a statistically nonsignificant

increase in risk of no improvement " ). Ammonia levels were

significantly lowered and the drug was found to be well tolerated. We

have entered a new era in the study of the therapeutics of hepatic

encephalopathy, and rifaximin will be a player.

Concluding Remarks

As medicine continues to embrace evidence-based practice and the

incorporation of molecular understanding of pathophysiology and

therapeutics, the care of patients with cirrhosis will become less

historical and more forward-looking. A paradigm shift in the

treatment of hepatic encephalopathy appears to be upon us, and will

be guided by well-designed clinical trials. The Swedish data that

have made progress in the outcome of variceal hemorrhage is cause for

much optimism, although they are also a clarion call for redoubled

efforts. The fact that hepatology can successfully look to its strong

history while also embracing techniques, technologies, and paradigms

confirms that we are in the right field at the right time.