Hepatic Encephalopathy

I. What every physician needs to know.

Hepatic encephalopathy (HE), also known as portal systemic encephalopathy (PSE), is a complex neuropsychiatric disorder in patients with portal systemic shunting, either with or without intrinsic liver dysfunction. When present in acute liver injury, it contributes to the diagnosis of acute liver failure and expedites liver transplantation. In patients with cirrhosis, it signifies decompensation, is associated with decreased survival, worsened quality of life, and increased healthcare costs. Prompt diagnosis, treatment, and prevention is of great importance.

The pathophysiology of HE is incompletely understood, but likely is a complex interaction of increased neurotoxins (e.g., ammonia), changes in blood brain barrier permeability, and alterations in neurotransmitters within the central nervous system. Normally, ammonia absorbed from the gastrointestinal tract enters the liver through the portal vein to be cleared by hepatocytes.

Increased ammonia entering the systemic circulation can be the result of fewer functioning hepatocytes (e.g., cirrhosis), and/or portal systemic shunting, in which some portion of blood from the portal system bypasses the liver entirely. Portal systemic shunting may be due to various causes of portal hypertension (e.g., portal vein thrombosis), but can also be iatrogenic as in patients that have undergone a transjugular intrahepatic portosystemic shunt (TIPS) procedure.

Continue Reading

Symptoms of HE are similar to those seen in other forms of encephalopathy and are characterized by changes in cognition, behavior, emotion, and neuromuscular function. Mental status may range from subtle deficits to coma. While several complex neuropsychiatric tests have been developed to aid diagnosis in research settings, such tools are cumbersome and of little utility in clinical practice. HE is predominantly a clinical diagnosis and relies upon history, patient observation, and physical exam. Although venous ammonia levels are often obtained, there are few circumstances in which this assists with either diagnosis or management.

The first priority in the treatment of HE is the identification and treatment of precipitating factors, such as infection, gastrointestinal bleeding, or electrolyte abnormalities. As increased levels of ammonia have been implicated in the pathogenesis of HE, many current therapies are aimed at decreasing ammonia in systemic circulation. Current mainstay therapies include lactulose and rifaximin, alone or in combination, in addition to an emphasis on proper nutrition and patient education.

The most recent nomenclature for describing HE distinguishes HE associated with acute liver failure (type A), HE associated with portal systemic bypass and no intrinsic liver disease (type B), and HE associated with cirrhosis and portal hypertension/portal systemic shunts (type C). The remainder of this discussion will focus on HE in the setting of cirrhosis and portal hypertension/portal systemic shunts.

A. History Part I: Pattern Recognition:

The neuropsychiatric symptoms of HE are on a spectrum, ranging from subtle deficiencies that are barely recognizable to devastating coma.

Minimal hepatic encephalopathy

Although HE is typically thought of as having overt signs and symptoms, a large portion of patients with cirrhosis show only subtle cognitive deficits, termed minimal HE. These patients are fully alert without obvious neurologic findings but may experience irritability, forgetfulness and confusion detectable only on formal neuropsychiatric testing. Even minor abnormalities can impair daily activities such as working, paying bills, and driving.

Overt hepatic encephalopathy

In overt HE, early symptoms include changes in sleep patterns, slowed response times, impaired computation, shortened attention span, and elevation or depression of mood; these symptoms may be accompanied by the physical signs of tremor, incoordination, asterixis, and impaired handwriting. Later stage symptoms include disorientation, somnolence, and inappropriate behavior; these symptoms may be accompanied by the physical signs of slowed/slurred speech, ataxia, altered reflexes, and nystagmus. The end stage of HE is coma.

B. History Part 2: Prevalence:

Minimal HE is observed in up to 70% of patients with cirrhosis. The prevalence of overt HE in patients with cirrhosis is 35%.

Patients with cirrhosis who also have older age, hypoalbuminemia, hyperbilirubinemia, severe ascites, Child Pugh Class C, diabetes, malnutrition, and H. Pylori infection are at increased risk for developing HE. Transjugular intrahepatic portosystemic shunt (TIPS) placement is also a risk factor, with a 20-30% incidence of new or worsened HE following the procedure. Precipitating factors for the development of HE are described in greater detail below.

C. History Part 3: Competing diagnoses that can mimic hepatic encephalopathy.

Given the similarities between the signs and symptoms of HE, and those of other common causes of altered mental status, a broad differential diagnosis must be entertained in the altered patient with cirrhosis. This should include electrolyte disturbances, uremia, hypoglycemia, hypoxia, hypercapnea, primary central nervous system (CNS) insult (e.g., stroke, intracranial hemorrhage), post-ictal states, toxic ingestions, and delirium tremens associated with acute alcohol withdrawal.

Because there is no pathognomonic physical finding or laboratory test to confirm HE as the etiology of a patient’s altered mental status, a thorough consideration of alternative diagnoses is important, even in patients with cirrhosis and a known history of HE.

D. Physical Examination Findings.

General signs of severe liver disease and portal hypertension should raise the possibility of HE as a cause of altered mental status in any patient. As discussed above, there is no definitive physical exam finding to confirm the diagnosis of HE. The following signs can be seen in HE, but also in neurologic dysfunction caused by a variety of other insults.

Asterixis is classically described as an intermittent, nonrhythmic loss of posture followed by a regaining of posture, and may be observed in any muscle that can be voluntarily commanded to maintain a posture. Thought to be the result of disturbances in oscillatory neural networks, asterixis is seen not only in HE, but also in other metabolic encephalopathies (e.g., renal failure), hypercapnea, and certain drug ingestions (e.g., phenytoin).

When demonstrated in a patient attempting to maintain dorsiflexion of the wrists with arms outstretched (i.e., “stop traffic”), the lapse in sustained posture resembles a bird flapping its wings, leading to the common description of “liver flap”. Patients may also be asked to squeeze the clinician’s fingers, elevate a leg and dorsiflex the foot, close eyes forcibly, or protrude the tongue. The physician may notice an intermittent release by the patient (negative myoclonus) in all of these maneuvers, which also signifies asterixis.

Focal neurologic signs such as hemiplegia or hemiparesis are possible and may be reversed with correction of HE. Other neurologic signs seen in HE include tremor, incoordination, impaired handwriting, dysarthria, ataxia, hypo- or hyper-active reflexes, and Babinski’s sign. Asking a patient to sign their name on the same piece of paper every day is one way to follow subtle changes in handwriting, either for improvement or deterioration.

Timed tests of intellectual function, such as the number connection test and the A deletion test, are primarily research tools and not used routinely in clinical practice; their sensitivity and specificity alone are not adequate for diagnosis, but following serially over time can shed light on treatment response.

E. What diagnostic tests should be performed?

The focus of diagnostic testing in the setting of suspected HE should focus on excluding alternative diagnoses and identifying precipitating factors of HE.

Alternative diagnoses

Alternative etiologies of altered mental status such as electrolyte disturbances, uremia, hypoglycemia, hypoxia, hypercapnea, primary CNS insults (e.g., stroke, intracranial hemorrhage), post-ictal states, and delirium tremens associated with alcohol withdrawal can all present with signs and symptoms similar to HE.

Precipitating factors

Infection, gastrointestinal bleeding, dehydration and electrolyte disturbances, renal failure, medications, alcohol use, recent TIPS, constipation, and sudden substantial increases in dietary protein intake are all factors that can precipitate HE. Hypokalemia increases renal ammonia production, and metabolic alkalosis favors conversion of insoluble ammonium to ammonia. Non-compliance with medications that treat HE is a diagnosis of exclusion.

1. What laboratory studies (if any) should be ordered to help establish the diagnosis? How should the results be interpreted?

Laboratory testing should include chemistries, complete blood count, and liver associated tests, including assessment of synthetic function with coagulation studies. In the correct clinical context, arterial blood gas may be useful. Toxicology testing can help determine if psychoactive drugs (prescribed or otherwise) are contributing factors.

When searching for infectious precipitants, urinalysis and chest x-ray should be considered based on clinical context. A diagnostic paracentesis to rule out spontaneous bacterial peritonitis (SBP) should be performed on admission in any patient with accessible ascites even in the absence of fever, abdominal pain, or leukocytosis.

In patients with established ascites of known etiology, the ascitic fluid should be evaluated with cell count, gram stain, and culture. Ascitic fluid albumin and total protein can be helpful in diagnosing the etiology of new ascites and making determinations about SBP prophylaxis.

2. What imaging studies (if any) should be ordered to help establish the diagnosis? How should the results be interpreted?

If there is clinical suspicion for primary CNS disease (e.g., stroke, intracranial hemorrhage) as an alternative diagnosis, consider brain imaging with computed tomography (CT) or magnetic resonance imaging (MRI). This is especially important in the patient with an atypical presentation (e.g., focal neurologic signs), severe encephalopathy, suspected head trauma (especially in the setting of concurrent coagulopathy) or those patients without an identifiable precipitant for HE. Doppler ultrasonography may be helpful when there is suspicion for hepatic vein thrombosis (Budd-Chiari syndrome), as this may sometimes be associated with development or worsening of hepatic encephalopathy.

While classic triphasic waves on electroencephalogram (EEG) have been described, this study is not routinely used for diagnosis.

F. Over-utilized or “wasted” diagnostic tests associated with this diagnosis.

While venous blood ammonia levels are frequently measured in HE, such testing typically does not provide useful information for either diagnosis or management. Most importantly, reliance on a laboratory test that is neither sensitive nor specific for the diagnosis of HE may prevent the clinician from fully considering the differential diagnosis in a patient with cirrhosis presenting with altered mental status.

The measurement of venous ammonia can have unreliable results, owing to the complexity of obtaining and analyzing the specimen properly. For the test to be accurate, blood should be drawn without the use of a tourniquet or clenched fist, should be placed immediately on ice, and processed within 15 minutes of being drawn to minimize false elevation secondary to cellular metabolism.

Elevated venous ammonia levels are not very specific for HE, and can be seen in a variety of other clinical conditions including recent surgery, hyperalimentation, errors of metabolism, hypothyroidism, and gastrointestinal bleeding without HE. In one study, over 60% of patients with cirrhosis but no clinical signs of overt HE had elevated venous ammonia levels. Several medications also increase venous ammonia levels, including valproic acid, barbiturates, narcotics, and diuretics.

Elevated venous ammonia levels are also not very sensitive for HE. There is wide variation in venous ammonia levels in patients with HE, and patients with a clear clinical diagnosis of HE and response to HE treatment may have normal venous ammonia levels.

There is some evidence that venous ammonia levels can correlate with HE severity, but given the difficulty of obtaining accurate measurements and the wide variation in levels across patients, clinical findings are more typically used to grade severity of HE in the non-research setting.

Arterial ammonia levels have clinical utility in the setting of acute liver failure, where levels greater than 200ug/dl are strongly associated with cerebral herniation and may lead to expedited consideration of liver transplantation, in addition to close monitoring for increased intracranial pressure.

III. Default Management.

A. Immediate management.

In patients with hepatic encephalopathy that is severe enough to warrant admission, immediate management includes searching for alternative diagnoses and identifying and initiating treatment for any precipitating factors. This is often done in tandem with administration of medications to decrease intestinal ammonia concentrations. It may also be appropriate to initially discontinue certain home medications on admission to ensure they are not contributing to the encephalopathy.

Lactulose, a non-absorbable disaccharide, has been shown to improve cognitive function and quality of life in HE, but there is a lack of high quality evidence showing any effect on mortality. Nevertheless, it remains a mainstay of treatment of HE. Lactulose is undigested until it reaches the colon, where it is catabolized by colonic bacteria to form short chain fatty acids. This decreases the pH of the colonic lumen, favoring the conversion of soluble ammonia to insoluble ammonium, functionally trapping ammonia in the colon.

Lactulose also acts as a cathartic, increasing fecal nitrogen excretion. These are two of the proposed mechanisms by which lactulose decreases serum ammonia levels. Dosing strategies vary, but a reasonable starting point is 30-45ml (20-30g) administered orally 3 times per day. This dose should be titrated to produce 2-3 soft bowel movements per day. Initial dosing can be increased for more rapid improvement in episodes of HE, but should be quickly returned to typical daily dosing to avoid overmedication leading to severe diarrhea and possible volume depletion.

If the patient’s mental status prohibits oral intake, a nasogastric tube (NGT) can be placed for administration. If neither oral nor NGT administration are possible, retention enemas can be used, combining 300ml (200g) of lactulose with 700ml of tap water.

The goal of antibiotic use in HE is to decrease the mass of enteric bacteria that produce ammonia. Rifaximin is a poorly absorbed antibiotic with broad spectrum activity against both gram positive and gram negative aerobic and anaerobic bacteria. It is better tolerated that other antibiotics such as neomycin and metronidazole. Meta-analyses of lactulose versus rifaximin suffer from a lack of high quality data, but suggest that rifaximin is at least as effective as lactulose and may be better tolerated.

When used alone for treatment of HE, typical dosing is 400mg administered orally 3 times per day. The relative expense of rifaximin compared to lactulose makes it a less attractive first line therapy. However, for recurrent episodes of HE, the addition of rifaximin (550mg administered orally 2 times per day) to lactulose therapy has been shown to significantly reduce the risk of recurrent episodes of HE, as well as reducing the risk of recurrent hospitalization.

Although patients with HE were once managed with protein restriction, it is now known that protein restriction does not improve HE and can be detrimental in patients with cirrhosis who are already prone to malnutrition as well as rapid protein wasting in times of metabolic stress. A nutrition consult to determine daily requirements to avoid protein catabolism may be helpful.

As noted in an earlier section, hypokalemia will increase renal ammonia production. There is often a concomitant metabolic alkalosis present as well, which promotes conversion of ammonium (NH4
+) to the neutral ammonia (NH3) particle that can cross the blood brain barrier. Thus, correction of hypokalemia is a simple and essential part of treating HE.

More recently, polyethylene glycol 3350-electrolyte solution (PEG) has been evaluated in comparison to lactulose. In one study, PEG use appeared to lead to more rapid resolution of HE. However, this treatment required several liters of PEG (similar to volumes used for pre-colonoscopy bowel preparation), which itself may lead to other problems, particularly in patients at risk for volume depletion.

Other treatments are being explored such as use of branched-chain amino acids (BCAA), probiotics, sodium benzoate, and acarbose. There was also a European cohort study in 2014 that demonstrated improved outcomes in HE after embolization of large spontaneous portosystemic shunts; this did not appear to cause or worsen pre-existing portal hypertensive disease.

B. Physical Examination Tips to Guide Management.

Treatment response is gauged by overall improvement in the patient’s neurologic status and should be assessed on a daily basis. The most important parts of the physical exam to follow daily will be determined by the specific deficits and severity of impairment present on admission. Normalization of speech, improved response times, resolution of asterixis, and improved handwriting skills are examples of exam findings that support improvement.

The number of bowel movements should be accurately documented to assist in proper titration of lactulose. While the therapeutic effects of lactulose are not entirely dependent on catharsis, number of bowel movements per day is a useful surrogate marker of adequate dosing.

C. Laboratory Tests to Monitor Response to, and Adjustments in, Management.

There are no reliable laboratory tests to monitor treatment response. In fact, serial monitoring of ammonia should be discouraged in light of the inaccuracies discussed above. However, diarrhea from overmedication with lactulose can precipitate volume depletion. Therefore, electrolytes and renal function should be monitored at least daily until a stable regimen is achieved. This is especially important in patients on concurrent diuretics or undergoing large volume paracentesis for management of ascites.

D. Long-term management.

Long-term management is focused on encouraging medication compliance with proper titration of lactulose for desired number of daily bowel movements. Patients and families should also be educated regarding avoidance of precipitating factors. As HE is a sign of decompensated cirrhosis, referral for liver transplantation should be initiated.

IV. Management with Co-Morbidities

A. Renal Insufficiency.

Renal function and electrolytes should be carefully monitored as increased bowel movements with lactulose can lead to volume depletion and cause further renal insults.

B. Liver Insufficiency.

No change in standard management.

C. Systolic and Diastolic Heart Failure

No change in standard management.

D. Coronary Artery Disease or Peripheral Vascular Disease

No change in standard management.

E. Diabetes or other Endocrine issues

No change in standard management.

F. Malignancy

No change in standard management.

G. Immunosuppression (HIV, chronic steroids, etc.).

No change in standard management.

H. Primary Lung Disease (COPD, Asthma, ILD)

No change in standard management.

I. Gastrointestinal or Nutrition Issues

No change in standard management.

J. Hematologic or Coagulation Issues

No change in standard management.

K. Dementia or Psychiatric Illness/Treatment

Patients with dementia or psychiatric illness are more likely to have difficulty with self-titration of lactulose to desired number of bowel movements per day, and will often require assistance from family members or caregivers after discharge from the hospital.

V. Transitions of Care

A. Sign-out considerations While Hospitalized.

The patient’s current mental status and any neurologic findings should be documented and conveyed during sign-out so that dramatic changes can be properly assessed. Documenting the patient’s updated medication list on discharge is also important as this can help future providers identify what specific treatments have already been employed.

B. Anticipated Length of Stay.

Length of stay will likely depend on the precipitating factor, but is typically 5-7 days in patients with HE. As may be expected, a higher MELD score is associated with longer duration of stay. Admission duration also appears to be influenced by the patient’s gender, presence of hepatitis C infection, use of antibiotics, and reason for admission.

C. When is the Patient Ready for Discharge.

At discharge, the patient should be back to their baseline mental status, ideally confirmed by someone who knows the patient well. Any precipitating factors identified on admission should be either successfully treated (e.g., hypokalemia) or responding to treatment (e.g., pneumonia) and the patient should be on a steady dose of lactulose and/or rifaximin.

Stable renal function should be documented while on the current regimen of lactulose and/or rifaximin, especially in combination with diuretics.

D. Arranging for Clinic Follow-up

1. When should clinic follow up be arranged and with whom.

Outpatient follow-up should include visits with the patient’s primary care doctor and hepatologist. Ideally, the patient should be seen by a provider within one week of discharge to ensure that self-titration of medication (i.e. lactulose) at home has been successful.

2. What tests should be conducted prior to discharge to enable best clinic first visit.


3. What tests should be ordered as an outpatient prior to, or on the day of, the clinic visit.


F. Prognosis and Patient Counseling.

The most important counseling for patients with HE upon discharge involves instructions for self-titration of lactulose at home to achieve 2-3 bowel movements per day, and avoidance of precipitating factors.

Prognosis for patients with cirrhosis following their first episode of HE is poor; estimated survival is 40% at one year and 20% at three years. It is a sign of decompensated cirrhosis and should prompt evaluation by a hepatologist.

VI. Patient Safety and Quality Measures

B. Appropriate Prophylaxis and Other Measures to Prevent Readmission.

Readmission for patients with decompensated cirrhosis is common. Of the complications of cirrhosis that have been studied in relation to hospital readmission (SBP, variceal hemorrhage, renal failure and HE), HE is implicated as one of the more preventable causes of readmission within a month, citing lack of stable lactulose regimen prior to discharge and inappropriate education regarding lactulose titration at home. This should be in the forefront of the clinician’s mind when approaching discharge decisions and education.

What's the evidence?

Polson, J, Lee, WM. “AASLD Position Paper: The management of Acute Liver Failure”. Hepatology. vol. 41. 2005. pp. 1179-1197.

Volk, ML. “Hospital readmissions among patients with decompensated cirrhosis”. American Journal of Gastroenterology; online publication. vol. 20. 2011. pp. 1-6.

Ferenci, P. “Hepatic Encephalopathy-Definition, nomenclature, diagnosis and quantification”. Hepatology. vol. 35. 2002. pp. 716-721.

Dhiman, RK, Chawla, YK. “Minimal Hepatic Encephalopathy: A Review”. Indian Journal of Gastroenterology. vol. 28. 2009. pp. 5-16.

Cash, WJ. “Current concepts in the assessment and treatment of hepatic encephalopathy”. Q J Med. vol. 103. 2010. pp. 9-16.

Poordad, F. “Review Article. the burden of hepatic encephalopathy”. Alimentary Pharmacology and Therapeutics. vol. 25. 2006. pp. 3-9.

Gokula, RM, Khasnis, A. “Asterixis”. Journal of Postgraduate Medicine. vol. 49. 2003. pp. 272-275.

Dufour, EF. “NACB: Laboratory Guidelines for screening, diagnosis, and monitoring of hepatic injury”. Clinical Chemistry. vol. 46. 2000. pp. 2050-2068.

LaBuzzetta, JN. “Adult non-hepatic hyperammonemia: a case report and differential diagnosis”. American Journal of Medicine. vol. 123. 2010. pp. 885-891.

Ong, JP. “Correlation of ammonia levels and severity of hepatic encephalopathy”. American Journal of Medicine. vol. 114. 2003. pp. 188-193.

Clemmesen. “Cerebral herniation in patients with acute liver failure is correlated with arterial ammonia concentration”. Hepatology. vol. 29. 1999. pp. 648-653.

Lawrence, KR, Klee, JA. “Rifaximin for the treatment of hepatic encephalopathy”. Pharmacotherapy. vol. 28. 2008. pp. 1019-32.

Bass, NM. “Rifaximin treatment in hepatic encephalopathy”. NEJM. vol. 362. 2010. pp. 1071-1081.

Cordoba, J. “Normal protein diet for episodic hepatic encephalopathy: results of a randomized study”. Journal of Hepatology. vol. 41. 2004. pp. 38-43.

Bustamante, J. “Prognostic significance of hepatic encephalopathy in patients with cirrhosis”. Journal of Hepatology. vol. 30. 1999. pp. 890-895.

Plessier, A, Valla, DC. “Budd-Chiari syndrome”. Semin Liver Dis. vol. 28. 2008. pp. 259-69. (Suggests that Budd-Chiari can be a precipitating factor for HE albeit rarely.)

Rahimi, RS. “Lactulose vs polyethylene glycol 3350—electrolyte solution for treatment of overt hepatic encephalopathy: the HELP randomized clinical trial”. JAMA Intern Med. vol. 17. 2014. pp. 1727(Helpful to know that other treatments besides lactulose can be effective.)

Vilstrup, H. “Hepatic encephalopathy in chronic liver disease: 2014 Practice Guideline by the American Association for the Study of Liver Diseases and the European Association for the Study of the Liver”. Hepatology. vol. 60. 2014. pp. 725(Self-explanatory updated practice guideline on the topic.)

Gluud, LL. “Branched-chain amino acids for people with hepatic encephalopathy (Review)”. Cochrane Database Syst Rev. 2015.

McGee, RG. “Probiotics for patients with hepatic encephalopathy”. Cochrane Database Syst Rev. 2011.

Laleman, W. “Embolization of large spontaneous portosystemic shunts for refractory hepatic encephalopathy: a multicenter survey on safety and efficacy”. Hepatology. vol. 57. 2013. pp. 2448(The above three references all put forward evidence for newer treatments of HE.)

Martel-Laferrière, V. “MELD score and antibiotics use are predictors of length of stay in patients hospitalized with hepatic encephalopathy”. BMC Gastroenterol. vol. 14. 2014. pp. 185(Discusses factors associated with longer stay, helpful for identifying patients at higher risk for problems during admission.)

Jump to Section