Portosystemic Encephalopathy 

Updated: Sep 17, 2019
Author: Gagan K Sood, MD; Chief Editor: BS Anand, MD 



Portosystemic encephalopathy (PSE) or hepatic encephalopathy (HE) is a neuropsychiatric syndrome associated with hepatocellular failure or portosystemic venous shunting.

There has been a lack of standardization of terminology used to define hepatic encephalopathy. "Acute" hepatic encephalopathy referred to acute liver failure or acute decompensation in the setting of chronic liver failure. The term "chronic" was used to describe the hepatic encephalopathy seen in chronic liver failure.

In 2002, a working committee task force on hepatic encephalopathy standardized the definition and the classification of hepatic encephalopathy. According to the characteristics of neurologic manifestations, hepatic encephalopathy is classified as episodic (previously, "acute"), persistent (previously, "chronic"), or minimal (previously, "subclinical").

Hepatic encephalopathy is also classified into three types based on the disease state of the liver, as follows:

  • Type A: Hepatic encephalopathy associated with acute liver failure

  • Type B: Hepatic encephalopathy associated with portosystemic bypass with no intrinsic hepatocellular disease

  • Type C: Hepatic encephalopathy associated with cirrhosis and portal hypertension or portosystemic shunts. In cases of chronic liver disease, type C hepatic encephalopathy can be episodic or persistent. The term "subclinical encephalopathy" was replaced with "minimal encephalopathy."

Hepatic encephalopathy is a reversible metabolic encephalopathy with multifactorial pathogenesis. The widely accepted hypothesis is that encephalopathy is due to a failure of hepatic clearance of gut-derived toxins. Although the exact toxins involved remain controversial, ammonia remains the toxin of interest. This has led to many investigative and therapeutic efforts aimed at identifying and eliminating the putative toxin(s) that originate from the gut lumen. A fluctuating level of consciousness is common, and progression to coma may occur rapidly.

A high index of clinical awareness is critical for anticipating and recognizing complications. A precipitating cause is usually discovered after clinical and laboratory evaluations. Although elevated plasma ammonia levels are often seen and therapy based on this observation is generally effective, poor correlation exists between the plasma ammonia levels and the degree of encephalopathy. As noted, multiple mechanisms contribute to the pathogenesis of this disorder. Discrete neuropathologic features are described in PSE but these may represent an epiphenomena (a secondary effect occurring with the disease). Treatment with lactulose is the mainstay of therapy, but novel developmental approaches show promise.


Although the exact pathophysiologic mechanisms of hepatic encephalopathy remain unclear, two areas have received more attention: first, the gut-derived neurotoxins (mainly ammonia) and, second, the changes in astrocyte morphology and physiology.

Hyperammonemia and portosystemic shunting led to the hypothesis in 1877 that enteral production of ammonia is central to the pathogenesis of this disorder. Various other putative toxins, which may also be shunted, may result in portosystemic encephalopathy (PSE), are described.

Portosystemic shunting is a requisite for the development of PSE. Although disturbances in urea cycle metabolism may cause hyperammonemia, similar encephalopathy does not exist in patients with isolated hyperammonemia in the absence of other evidence of hepatic dysfunction. The pathogenesis of portal hypertension is discussed in Portal Hypertension. This complex condition results in the flow of portal blood containing putative toxins produced in the gut to the systemic circulation and, ultimately, the brain via extrahepatic shunts (collateral flow).

A minority of patients with cirrhosis present with recurrent symptoms of hepatic encephalopathy often without any precipitating cause. These patients may have minimal or mild hepatocellular dysfunction but have significant neurologic impairment. In one study, large portosystemic shunts were detected by computed tomography scanning in most patients. Shunting, in part, appears to be a response to increased hepatic vascular resistance in the setting of cirrhosis; however, shunting may also result from other causes, including portal vein thrombosis or compression, congenital hepatic fibrosis, iatrogenic shunt placement, and congenital shunt formation. The latter is an important consideration in younger patients with an otherwise unexplained hepatic encephalopathy (in the absence of cirrhosis or iatrogenic shunts or portosystemic shunts associated with splenic or portal vein thrombosis). These patients may present in middle age and respond to appropriate shunt-reversal surgery.

The intrahepatic shunt (transjugular intrahepatic portosystemic shunt [TIPS]) provides a conduit for portal venous blood to flow directly into the hepatic vein while bypassing the hepatic parenchyma. TIPS is associated with the development of PSE in approximately 25% of cases.

The proposed gut-derived toxins responsible for PSE include ammonia, phenols, thiols, and short-chain fatty acids. Other possible mediators include cytokines and bacterial endotoxins. The enteral production of gamma-aminobutyric acid (GABA) and endogenous benzodiazepines (BZPs) remains somewhat speculative, although alterations in GABA-receptor–mediated neurotransmission may play a role for other reasons. The GABA complex, when provided with an appropriate ligand, leads to the production of an inhibitory signal. Widespread inhibition of cortical function from excessive GABAergic signaling, therefore, has been postulated as a mechanism leading to PSE.

The thiols or mercaptans are small volatile molecules that characteristically are recognized by their pungent odor, which results from the inclusion of a sulfhydryl group. Accordingly, they may lead to the clinical presentation of fetor hepaticus; however, ammonia is the best contender for the most significant gut-derived neurotoxin, and the ammonia hypothesis, therefore, justifies elaboration.

Most successful forms of therapy are based on the concept of ammonia neurotoxicity. Elimination of ammoniagenic luminal bacteria with nonabsorbed antibiotics (eg, neomycin), luminal acidification with nonabsorbed sugars fermented by luminal bacteria (eg, lactulose), avoidance of constipation, and reduction in ammoniagenic substrate intake (eg, protein-restricted diets) support the ammonia hypothesis.

The production of ammonia from the bacterial expression of urease and metabolism of colonic protein accounts for most ammoniagenesis. The bulk of extracolonic ammonia production occurs in the kidneys. Renal failure may promote ammoniagenesis as a consequence of uremia, which increases the available substrate for urease. Ammonia is a neurotoxic compound that principally is eliminated in humans by its hepatic conversion to urea. Periportal hepatocytes in the liver primarily metabolize ammonia. Subsequently, urea is excreted in the urine. Residual ammonia in the hepatic sinusoidal circulation is converted to glutamine by perivenous hepatocytes expressing glutamine synthase.

Besides causing functional changes such as reduction in cerebral perfusion, ammonia may also be responsible for structural changes in the brains of patients with hepatic encephalopathy. In necropsy studies, brains of cirrhotic patients exhibit Alzheimer type II astrocytosis, characterized by swollen astrocytes with enlarged nuclei and chromatin displaced to the perimeter of the cell. Type II astrocytosis is hypothesized to be caused in part by the detoxification of ammonia. Astrocytes, the only cells in the brain that can metabolize ammonia, contain glutamine.

In vivo proton magnetic resonance spectroscopy (MRS) (1H-MRS) shows that astrocyte swelling without increases in intracerebral pressure may occur early in the pathogenesis of PSE.

Ultimately, the development of advanced PSE may be accompanied by cerebral edema, which may contribute to neurologic impairment. Although cerebral edema has its most obvious manifestations in the patient with fulminant hepatic failure (FHF), osmotically active substances do accumulate in the brains of patients without overt cerebral edema. An osmotically sensitive pool of myoinositol is released from astrocytes in response to osmotically induced astrocyte swelling. A depletion of myoinositol is shown by 1H-MRS in patients with chronic PSE, and it appears to correlate with an increase in the signal for glutamine and glutamate.

With the use of MRS, low-grade cerebral edema has been demonstrated in patients with cirrhosis and chronic hepatic encephalopathy.

Despite the demonstration of astrocyte swelling and osmotic phenomena, treatment of hepatic encephalopathy does not include the use of mannitol or hyperventilation unless cerebral edema is suspected, as in FHF. No established role currently exists for routine cerebral magnetic resonance imaging (MRI) or MRS in the evaluation of PSE. The data supporting the ammonia hypothesis in PSE development, therefore, are impressive and follow multiple lines of evidence. Indeed, the past decades have been remarkable for the recognition of ammonia as a key element in the pathogenesis of PSE. However, other small molecules also may contribute, and these theories are not mutually exclusive. Synergistic toxicity of ammonia and other agents likely is important.

Production of so-called false neurotransmitters may contribute significantly to PSE pathogenesis. Putative agents include octopamine and diazepam. Supplementation with branched-chain amino acids (BCAAs) such as isoleucine, leucine, and valine, and avoidance of aromatic amino acids, such as phenylalanine, tryptophan, and tyrosine, may lead to decreased production of false neurotransmitters; however, the clinical benefit of BCAA supplementation has never been demonstrated convincingly (see Diet).

Increased production of endogenous BZPs has been proposed. These agents may represent the best-defined PSE false transmitters; however, their precise role is somewhat unclear. These substances are believed to depress central nervous system (CNS) function by binding to specific high-affinity BZP sites on GABA-receptor complexes. The GABA complex, when provided with an appropriate ligand, leads to the production of an inhibitory signal. Therefore, widespread inhibition of cortical function from excessive GABAergic signaling has been postulated as a mechanism leading to PSE.

Broadly speaking, cytokines are substances produced and released by cells for communication with other cells. Interleukin-1 (IL-1), interleukin-6 (IL-6), and tumor necrosis factor-alpha (TNF-α) are important examples of immunomodulatory cytokines that are increased in the systemic circulation and possibly contribute to the pathogenesis of systemic hemodynamic events in portal hypertension. The rapidly diffusing nitrous oxide (NO) is grouped with these substances for the purposes of this discussion. Although not a cytokine in the strictest sense, NO plays an important, if ill-defined, role in mediating some of the significant communication events resulting from cytokine activation in advanced liver disease.

Endotoxemia, presumably in part from gut mucosal permeability, is demonstrated in cirrhosis with portal hypertension. Associated increased IL-1 and TNF-α concentrations and metabolites of NO in the systemic circulation also are reported. This suggests that shunting of proinflammatory substances from the gut lumen contributes to or perhaps initiates a cascade of events culminating in the hyperdynamic circulation typical of advanced liver disease.

Cerebral ischemia is another mechanism that contributes to PSE, although it may represent a consequence of ammonia toxicity. A loss of cerebral blood flow autoregulation reflexes may accompany the development of FHF; however, cerebral autoregulation, in general, is preserved in patients with cirrhosis if the mean arterial pressure is maintained above 70 mm Hg, even in severe cases of hepatic encephalopathy. In contrast, patients with advanced hepatic encephalopathy have reduced cortical blood flow and increased cerebral vascular resistance.


Precipitating factors that lead to the clinical manifestations of portosystemic encephalopathy may be obvious; however, often a cause is not evident despite concerted efforts to identify one. The importance of trying to determine the precipitating cause cannot be overemphasized. Infection, specifically spontaneous bacterial peritonitis (SBP), is especially important to exclude. Several known causes are categorized below by their proposed mechanisms. In some cases, multiple mechanisms may be responsible.

Increased ammoniagenesis

Findings include the following:

  • Increased substrate (protein) for ammoniagenesis

  • Increased substrate (urea) for ammoniagenesis

  • Increased protein intake

  • Gastrointestinal bleeding

  • Constipation

  • Dehydration

  • Renal failure

  • Increased catabolism of protein

  • Infection

  • Hypokalemia

  • Sepsis

Decreased hepatocellular function

Findings include the following:

  • Dehydration

  • Hypotension

  • Sepsis

  • Hypoxia

  • Anemia

  • Development of hepatocellular carcinoma

  • Worsened intrinsic liver disease

  • Drug toxicity

  • Superimposed viral hepatitis

Increased portocaval shunting

Findings include the following:

  • Portal vein thrombosis

  • Transjugular intrahepatic portosystemic shunt formation

  • Surgical shunt formation

  • Spontaneous shunt formation

Psychoactive drug use

The following agents may cause portosystemic encephalopathy:

  • Benzodiazepines

  • Ethanol

  • Antinauseants

  • Antihistamines

  • Other agents

Other mechanisms

Increased diffusion of ammonia across the blood-brain barrier may occur, potentially resulting in alkalosis, which promotes ammonium ion conversion to less polar and more diffusible ammonia.

Blood transfusion may be a factor. Increased ammoniagenesis from transfusions may not be entirely accurate and is possibly more a theoretical than practical concern. Glutaminase activity and generation of ammonia in stored cellular blood products (especially platelets) may conceivably lead to the infusion of ammonia during transfusion.


United States and international data

The true incidence and prevalence figures for portosystemic encephalopathy are not available. This complication is a frequent and possibly inevitable feature of progressive and chronic liver disease.

Race-, sex-, and age-related demographics

No specific racial data apply.

Presumably, because alcoholic liver disease occurs with greater frequency in men compared to women, a higher proportion of patients appears to be male.

A variety of neurologic conditions common in elderly patients (eg, multiple cerebral infarcts, Alzheimer disease, parkinsonism) may exacerbate the manifestations of portosystemic encephalopathy; however, this condition usually is prominent in patients with advanced liver disease, which may not manifest in very elderly persons because they may not survive that long. No specific age frequency data are available.


In general, the prognosis of patients with advanced liver disease is poor unless they are able to undergo liver transplantation. The contribution that portosystemic encephalopathy makes to the decline of these individuals is limited and, rather than acting as a causal factor, it is essentially a marker of decompensated liver disease. As with other complications of chronic liver disease, survival likely correlates better with their Child-Pugh score than the outcome of the complication itself.


The development of portosystemic encephalopathy indicates decompensated liver disease and, therefore, other features of decompensation, such as impaired coagulation (eg, elevated prothrombin time/international normalized ratio), varices, ascites, and portal hypertension, must be sought. Indeed, these manifestations potentially lead to the recognition of complications that may account for the development of portosystemic encephalopathy, such as spontaneous bacterial peritonitis (SBP) or gastrointestinal bleeding. In this setting, portosystemic encephalopathy will often manifest itself for the first time (acute portosystemic encephalopathy) or suddenly worsen (acute or chronic portosystemic encephalopathy). Identification of the factors responsible for precipitating acute hepatic encephalopathy usually is possible (see Etiology).

Because of the potential for rapid evolution to coma, close clinical monitoring and the anticipation of elective endotracheal intubation for airway protection and ventilatory support is essential. Impairment of consciousness poses the greatest threat to the patient with acute portosystemic encephalopathy and must be evaluated and managed most aggressively. Chronic portosystemic encephalopathy is a common feature of advanced liver disease and, if left untreated, may evolve into a more severe state without any obvious precipitants, although the course of events may be somewhat slower. The level of consciousness in these patients also may fluctuate significantly in an apparently spontaneous manner.


Respiratory and ventilatory insufficiency, and the failure to protect the airway with consequent aspiration are anticipated complications that should be avoided with endotracheal intubation.




Patients with portosystemic encephalopathy may present with vague complaints, such as fatigue, or they may be brought for evaluation by relatives. Indeed, the family members of patients often are the most useful historians. They may describe restlessness (especially at night), somnolence (typically during the day), and instances reflecting episodic confusion.

No neurologic features are entirely specific for this disorder.

Clinical history or physical examination findings suggestive of liver disease, without evidence of another etiology for neurologic dysfunction, form the basis for the diagnosis.

A precipitating factor identified from the clinical history strongly suggests the diagnosis. Therefore, directed questioning in this regard is essential.

Several predisposing conditions also may contribute to the development of chronic portosystemic encephalopathy.

Physical Examination

As with many toxic and metabolic encephalopathies, mental status and, in particular, the level of consciousness may fluctuate dramatically. In addition to impairment of the level of consciousness, patients with portosystemic encephalopathy (PSE) may demonstrate a variety of neurologic signs together or in isolation.

PSE grades are as follows:

  • Subclinical PSE: Early in the course of the condition, patients may appear normal clinically but perform poorly on psychometric testing. This is termed subclinical PSE. It may be particularly important to diagnose subclinical PSE in patients performing tasks that require rapid reaction times because these are prolonged in affected patients. Judgment may also be erratic or questionable. Recognition of subtle impairment, therefore, may lead to a recommendation to avoid driving a motor vehicle or operating machinery.

  • Grade 1 PSE: With deterioration to grade 1 PSE, patients have difficulty with memory, mild confusion, agitation, and irritability. Other complaints may include restlessness or sleepiness during the day and remaining awake at night. Tremor, a rhythmic or regular oscillation, and incoordination may be seen. Handwriting skills are typically impaired. Constructional apraxia may be easily demonstrated at the bedside by requesting the patient to draw a five-pointed star.

  • Grade 2 PSE: Progression to grade 2 PSE involves a slowing of mentation and speech, with the appearance of lethargy. The patient's confusion progresses to difficulty with orientation to time, and loss of inhibition ultimately may result in inappropriate behavior. Neurologic signs include asterixis and an irregular flapping tremor that is observed best with the wrists dorsiflexed and fingers spread. Dysarthria, ataxia, and hypoactive deep-tendon reflexes are characteristic.

  • Grade 3 PSE: This grade portends the significant possibility of coma; therefore, plans to perform endotracheal intubation should be made upon recognition of this state. Patients are drowsy but can be woken up; however, they remain markedly confused. They may exhibit frankly aggressive behavior. Asterixis persists, but the deep-tendon reflexes become hyperactive as they are disinhibited, a process that culminates in the development of the decerebration seen in grade 4 PSE. Babinski signs may be seen.

  • Grade 4 PSE: This also is known as hepatic coma and represents a medical emergency. Patients are unable to protect their airway reliably. Prophylactic endotracheal intubation is mandatory. Patients may remain unresponsive for several days.

  • Stigmata of chronic liver disease and portal hypertension may be evident upon examination. These include the presence of gynecomastia and testicular atrophy in males, palmar erythema, spider nevi, splenomegaly, caput medusae, ascites, peripheral edema, and a shrunken liver.

More specific signs resulting from the underlying liver diseases also may be present, as follows:

  • The Kayser-Fleischer rings of Wilson disease are typically identified only with a slit-lamp examination.

  • Bronzed skin and arthropathy are seen in hereditary hemochromatosis.

  • Cholesterol deposition may be seen as xanthelasma in patients with chronic cholestasis of any cause, but it often is most striking in primary biliary cirrhosis.

  • Alcoholic liver disease is frequently accompanied by advanced malnutrition; however, any long-standing liver disease also may lead to temporal, shoulder girdle, and hip girdle muscle wasting.

  • Cutaneous features of hepatitis C virus infection include a characteristic rash due to leukocytoclastic vasculitis resulting from mixed cryoglobulinemia, porphyria cutanea tarda, and lichen planus.



Diagnostic Considerations

Conditions to consider in patients with suspected portosystemic encephalopathy include the following:

  • Structural central nervous system lesions

  • Subdural hemorrhage

  • Intraparenchymal hemorrhage

  • Cerebral infarct

  • Intracranial infections

  • Encephalitis (eg, herpes simplex virus)

  • Intracerebral abscess

  • Hydrocephalus

  • Toxic or metabolic etiologies

  • Drug intoxication

  • Ethanol

  • Benzodiazepines

  • Hypoglycemia

  • Ketoacidosis

  • Electrolyte disturbances

  • Hypercapnia

  • Hypoxia

  • Postictal encephalopathy

  • Functional psychoses

  • Metabolic encephalopathy

Differential Diagnoses



Laboratory Studies

Laboratory studies include the following:

  • Serum calcium levels

  • Serum ammonia levels, preferably arterial: Elevated blood levels are not diagnostic of hepatic encephalopathy, and normal levels do not rule out hepatic encephalopathy. However, very high levels may suggest an unsuspected urea cycle enzyme deficiency. They are helpful to suggest a diagnosis of hepatic encephalopathy when the cause is obscure.

  • Serum glucose measurements

  • Serum aspartate aminotransferase (AST) levels: A very high AST and/or alanine aminotransferase (ALT) level (eg >1000 U/L) may suggest widespread hepatic necrosis as a consequence of acetaminophen toxicity and, accordingly, may help guide appropriate diagnostic evaluations and therapy (with acetaminophen levels and N-acetylcysteine). However, very high AST and ALT values may be observed in other settings (notably ischemic hepatitis and other causes of submassive or massive hepatic necrosis) and, therefore, are not specific.

  • Serum ALT levels: ALT is more specific for a liver origin because AST also may be released from muscle.

  • Serum bilirubin measurements

  • Complete blood cell (CBC) count

  • Blood and urine screen for drugs

  • Blood alcohol level

  • Serum electrolyte levels: Electrolytes may be disturbed for a variety of reasons in patients with advanced liver disease. Hyponatremia resulting from diuretic use, renal failure, water intoxication, or the syndrome of inappropriate secretion of antidiuretic hormone is particularly important to consider as a contributing cause of encephalopathy.

  • Prothrombin time, and levels of serum albumin and bilirubin: These measurements are true liver function tests that provide an estimate of the severity of liver damage. They also allow the clinician to anticipate certain potential complications (eg, bleeding) and adjust therapy appropriately.

Imaging Studies

Computed tomography (CT) scanning of the head

The rationale for CT scanning is to exclude structural considerations in the differential diagnosis, including intracranial hemorrhage (epidural, subdural, subarachnoid, intraparenchymal), cerebral infarct, intracranial infections (brain abscess with mass effect, meningoencephalitis), and hydrocephalus.

CT scanning of the head may not be necessary in patients with well-documented liver disease and a typical history, especially if no focal or localizing signs are evident. However, if the circumstances leave any doubt, CT scanning is critical in helping exclude structural causes for encephalopathy.

This test is more widely available than magnetic resonance imaging (MRI) and generally can be performed more rapidly. Therefore, it is the imaging modality of first choice in most instances.

Patients may develop portosystemic encephalopathy and subsequently sustain head trauma. This is particularly common among patients with alcoholism, and the event may not be volunteered during history taking or may not be evident upon physical examination. However, in uncomplicated portosystemic encephalopathy, no characteristic clues or findings are present.

Magnetic resonance imaging (MRI)

The availability and speed with which CT scans can be performed make them preferable in most instances for helping exclude mass lesions and, especially, intracranial hemorrhage. However, MRI findings may be of particular value in cases in which a diagnosis is not clear, based on other clinical and laboratory data.

The presence of hyperintense-appearing regions on T1-weighted MRI studies of the brains of patients with cirrhosis is described as a characteristic feature. The increased MRI signal intensity may be the result of manganese (Mn) deposition in these structures.

The globus pallidus, putamen, and caudate nucleus of the basal ganglia and the frontal and occipital cortex of patients with cirrhosis who died with hepatic encephalopathy contain increased Mn concentrations when compared to matched control specimens. Pallidal hyperintensity on T1-weighted MRI does not appear to be present in well-compensated patients with cirrhosis who do not have hepatic encephalopathy. This finding appears to correlate with blood ammonia levels but not the severity of hepatic encephalopathy itself.

The amount of Mn deposition that can be identified at autopsy of patients with hepatic coma appears to be independent of the patient's age, etiology of cirrhosis, or the presence of chronic hepatic encephalopathy. In an experimental model using both cirrhotic and portacaval-shunted rats, Mn levels in the basal ganglia were significantly elevated above control values. These levels also were significantly higher in portacaval-shunted rats when compared to those with experimental cirrhosis; therefore, although the precise etiology responsible for Mn deposition is unclear, it is enhanced by portal hypertension.

Therefore, signs of extrapyramidal toxicity in hepatic encephalopathy conceivably may result from Mn deposition. The neurologic and radiologic changes may resolve gradually with time following liver transplantation. Deposition of Mn also may potentiate the effects of benzodiazepines (BZPs), natural or otherwise, by increasing the number of available peripheral-type BZP binding sites (possibly by promoting receptor expression).

These intriguing issues unveiled by the advent of MRI are complemented by the metabolic data derived from the application of magnetic resonance spectroscopy (MRS). This technique has demonstrated findings of altered glutamine metabolism. Its role in clinical evaluation and management of hepatic encephalopathy is unclear. In one series, MRS findings (ie, decreased myoinositol, increased glutamine) correlated poorly with neurologic status. These markers were suggested to be more representative of the underlying chronic hepatic dysfunction.

Positron emission tomography (PET) scanning

PET scanning has demonstrated reduced metabolic activity for glucose utilization in the parietal cortex of patients with cirrhosis with mild hepatic encephalopathy. At present, this technique is best reserved for research applications because no clear clinical guidelines are available for its use, and moreover its availability is limited.

Other Tests

Psychometric testing

Psychometric evaluations are of value for establishing the diagnosis and perhaps for monitoring the response to therapy in subclinical portosystemic encephalopathy. The number-connection test and the trail-making test are pragmatic approaches, and they are used widely at the bedside.

More formal testing may not be feasible with many patients, in part due to the length of time taken to administer the tests and also because of patient uncooperativeness. However, the results of psychometric testing in subclinical portosystemic encephalopathy may be of prognostic value independent of the Child-Pugh score of disease severity.

Patients with alcohol-induced liver disease exhibit poorer test scores than those with liver disease from other causes, presumably due to cerebral toxicity intrinsic to chronic alcohol use and independent of the hepatic insufficiency.

A grading scheme that incorporates the level of consciousness, personality and intellect, neurologic signs, and electroencephalographic (EEG) abnormalities has been proposed for hepatic encephalopathy. The clinical portion of this grading approach has the advantage of being easily administered at the bedside, and it is helpful as a guide to progress.


EEG studies of patients in portosystemic encephalopathy grades 1-3 may demonstrate high voltage and low-frequency triphasic waves of 1-3 Hz. These also may be seen in uremia but are characteristic of hepatic encephalopathy. With progression to coma, the EEG typically shows delta-wave activity, representing a generalized slowing of the cortex, a nonspecific pattern seen in toxic and metabolic encephalopathies.

The EEG is most helpful in excluding the presence of other causes for encephalopathy, such as status epilepticus and akinetic seizures, or the demonstration of postictal slowing with or without focal spike and wave activity that suggests prior seizures.

EEG monitoring frequently is useful in assisting with the diagnosis of hepatic encephalopathy, especially subclinical hepatic encephalopathy. Computer-assisted or spectral EEG analysis may demonstrate characteristic abnormalities, but the incremental benefit over conventional EEG is unclear.

Evoked potentials

Further electrophysiologic assessment may be performed with evoked-potential studies, but whether this approach is of value remains unclear, except when significant doubt exists with respect to the underlying diagnosis of portosystemic encephalopathy as the cause for neuropsychiatric dysfunction. This is rarely the case in practice.

These studies include visual-evoked potentials, somatosensory-evoked potentials, or brainstem auditory–evoked potentials, and they represent the externally recorded voltage from synchronous firing of neurons in a network response to specific stimuli.

The chief value of this approach may be to document abnormal cortical function in subclinical portosystemic encephalopathy and to establish the diagnosis. Brainstem auditory–evoked responses appear particularly sensitive as a marker of perturbed cortical function. Evoked-potential studies in acute hepatic encephalopathy conceivably may be useful for monitoring clinical response to treatment; however, the use of these electrophysiologic diagnostic modalities is not widespread.

Histologic Findings

In neuropathologic studies of hepatic encephalopathy, Alzheimer type II astrocytosis is typical and likely represents the end result of these mechanisms. The astrocytes demonstrate swollen nuclei, margination of the chromatin, and a prominent nucleolus.



Approach Considerations

Consultation with a neurologist is of value, especially if doubts exist regarding the etiology of the encephalopathy.

Alcohol rehabilitation is a requirement for acceptance for listing for orthotopic liver transplantation. Regardless of whether the patient is a candidate for orthotopic liver transplantation or not, the development of portosystemic encephalopathy is a potential revelation that permits progress toward achieving and maintaining abstinence. This is a critical step for increasing the probability of survival and is often best approached initially by taking advantage of the inpatient status.

Medical Care

Nonabsorbable disaccharides

Despite the availability of multiple approaches, treatment with nonabsorbed disaccharides remains the mainstay of therapy for hepatic encephalopathy. These agents act by at least three mechanisms, as follows:

  • The luminal bacteria metabolize lactulose and lactitol (beta galactoside-sorbitol). The consequent acidification of the gut lumen leads to ammonia being protonated to the ammonium ion (NH4+), which is relatively membrane impermeable; therefore, less ammonia is absorbed from the colon.

  • The second benefit to gut luminal acidification is a reduction in the number of bacteria present, which reduces the presence of bacterial urease and consequent ammoniagenesis.

  • The osmotic effect of nonabsorbed disaccharides enhances gastrointestinal transit.

A typical starting dose for the treatment of chronic hepatic encephalopathy is 20 g of oral (PO) lactulose twice daily with the goal of producing two to three soft bowel movements daily. Dose increases or reductions may be necessary based on the patient's response.

In the acute setting, hepatic encephalopathy may be treated with 20 g lactulose every few hours until a satisfactory result is achieved, but care must be taken to avoid diarrhea that leads to electrolyte depletion and dehydration. Lactulose enemas may be of benefit when a paralytic ileus precludes PO or nasoenteral tube administration. Note: Because the bacterial metabolism of lactulose results in the production of hydrogen, this agent should not be used as a lavage preparation for colonoscopy, as electrocautery under these circumstances may produce explosive results.

Clinical evaluations regarding the efficacy of nonabsorbed disaccharides are limited, especially when their widespread use and the potential for adverse reactions are considered. The toxicity of lactulose and lactitol includes gastrointestinal bloating due to bacterial gas production, dehydration, and electrolyte disturbances. The latter may result in paralytic ileus and, therefore, must be carefully distinguished from the more common bloating. Indeed, these complications of lactulose therapy may paradoxically worsen hepatic encephalopathy. Gastrointestinal upset infrequently leads to a requirement for dose reduction or, ultimately, discontinuance. It may be less problematic with lactitol.


Treatment with nonabsorbable antibiotics also is advocated as a treatment of hepatic encephalopathy, and it may be of particular value in a patient intolerant of lactulose. This may be relevant for a patient with intestinal ileus, in whom the administration of lactulose may generate large volumes of hydrogen gas upon bacterial fermentation. Similarly, patients with multiple electrolyte disturbances or dehydration may benefit from this antibiotic approach rather than from the use of lactulose.

Neomycin at 1 g PO three or four times daily may be used, but it is not recommended for prolonged use. As much as 3% of a PO dose of neomycin may be absorbed systemically, and nephrotoxicity may result. The efficacy of neomycin has been questioned when no clinical benefit could be demonstrated when it was compared with placebo in the treatment of acute hepatic encephalopathy.

Oral metronidazole is used with success, but long-term therapy may result in toxicity, including peripheral neuropathy.

Rifaximin is a relatively newer nonabsorbable antibiotic for the treatment of hepatic encephalopathy. Its efficacy in hepatic encephalopathy has been studied in more than 20 studies, including 14 randomized controlled studies. In seven studies, rifaximin was compared to lactulose or lactilol. Results of Cochrane meta-analysis from these studies suggested rifaximin to be significantly more effective than nonabsorbable disaccharides (lactulose or lactilol) in the treatment of hepatic encephalopathy.

In March 2010, rifaximin was approved by the Food and Drug Administration (FDA) to reduce recurrence of hepatic encephalopathy. The approval was based on a phase 3 clinical trial conducted by Bass et al that evaluated rifaximin’s ability to reduce the risk of recurrent hepatic encephalopathy.[1] In this trial, 299 patients received either rifaximin 550 mg or placebo twice daily. Each group also received lactulose. The primary endpoint—the risk of a recurrent episode of hepatic encephalopathy—was reduced by 58% in the rifaximin group compared to the placebo group (P <  0.0001).[1] The key secondary endpoint, risk of experiencing hepatic encephalopathy-related hospitalization, was reduced by 50% in patients who received rifaximin compared to those who received placebo (P = 0.0129).

An open-label maintenance extension of the phase 3 trial of rifaximin in hepatic encephalopathy reported benefits of long-term treatment with rifaximin in maintenance of remission from overt encephalopathy.[2] This study reported results of long-term efficacy and survival of 152 rollover patients from the registration trial and 128 new patients treated with rifaximin. The rollover patients were rifaximin (n = 70) or placebo (n = 82) patients who completed or withdrew from the registration trial with a Conn Score of 2 of less.

Sixty of the 70 rifaximin-treated patients from the registration trial who enrolled in the long-term follow-up remained in remission at study completion or withdrawal, and 43 (72%) of these patients did not experience breakthrough overt encephalopathy.[2] The risk of experiencing breakthrough encephalopathy was decreased by 79% compared to their prior 6-month placebo treatment. Changes in the Model for End-Stage Liver Disease (MELD) score were minimal in both the registration trial and the open-label extension, regardless of the treatment (see the MELD Score calculator). The authors concluded that longer therapy with rifaximin is associated with continued protection from breakthrough hepatic encephalopathy, with no adverse effect on expected mortality.[2]

Altering gut flora

The presence of urease-expressing bacterial organisms in the gut microflora forms the basis for efforts to repopulate the gut with nonureolytic organisms, such as Lactobacillus acidophilus and Enterococcus faecium. In theory, this should result in a reduction in colonic ammoniagenesis, but few well-designed studies exist to support the routine clinical application of this approach. Results of small trials with Lactobacillus species have been mixed. The use of orally administered Enterococcus species resulted in sustained protection from hepatic encephalopathy in one study and appeared to be safe. Further evaluation of this approach is justified and needed.

The presence of Helicobacter pylori in the gastric mucosa represents another potential source of ammonia because this organism produces urease. Helicobacter ammoniagenesis may be most significant when accompanied by achlorhydria, in part due to increased absorption of nonprotonated ammonia across the gastric mucosa and, possibly, from increased numbers of bacteria. The role of H pylori in the pathogenesis of hepatic encephalopathy remains contentious; some investigators have identified it as an independent risk factor for the development of hepatic encephalopathy, whereas others have not.

One possible explanation for improvement in hepatic encephalopathy following eradication therapy for H pylori is that the antibiotics decreased the gut's colonic population of urease-expressing organisms and those of the gastric mucosa. It appears reasonable to treat patients for H pylori when dictated by routine clinical circumstances (eg, in the treatment of peptic ulcer disease) but not as prophylaxis for hepatic encephalopathy.


Probiotics are not as useful in overt hepatic encephalopathy but have been used with some success in minimal HE. The species that are most efficacious are lactobacilli and bifidobacteria. Probiotics may also reduce bacterial translocation and subsequent endotoxemia and ameliorate the hyperdynamic circulation.

Although probiotics appear to reduce plasma ammonia concentration when compared with placebo or no intervention, a Cochrane meta-analysis concluded that probiotics are efficacious in altering clinically relevant outcomes,[3] but further randomized clinical trials are needed. Another study showed lactulose and probiotics are effective for secondary prophylaxis of hepatic encephalopathy in patients with cirrhosis.[4]

Ammonia scavengers and activated charcoal

Intravenous sodium benzoate and sodium phenylacetate or the phenylacetate prodrug oral sodium phenylbutyrate can combine with glycine or glutamine to form water-soluble compounds excreted through the kidneys. These agents are not yet approved by the FDA for use in hepatic encephalopathy; they depend on normal renal function for ammonia excretion, and the large therapeutic doses confer a significant sodium load, which can increase fluid retention.

Newer ammonia scavengers and orally ingested activated charcoal are being studied for the treatment of hepatic encephalopathy.

Glycerol phenylbutyrate (HPN-100) is a compound that is a prodrug of sodium phenylbutyrate with much lower therapeutic doses needed. It has been used for urea cycle disorders and continues to undergo clinical trials for chronic hepatic encephalopathy.[5, 6, 7]

With regard to orally ingested activated charcoal, AST-120 is a spherical carbon adsorbent of small molecules (ammonia, lipopolysaccharides, and cytokines) that has been used to improve symptoms of hepatic encephalopathy.[6] A pilot study showed equal efficacy as lactulose and fewer adverse events.[8]

Increasing ammonia metabolism

Another treatment approach is to increase the metabolism of ammonia with the administration of substrates that permit its incorporation.

Ornithine is a substrate for urea, and aspartate is a substrate for glutamine. Both enteral and intravenous administration of ornithine aspartate (a mixture of the two amino acids) in some controlled trials have been shown to lower serum ammonia levels and improve mild hepatic encephalopathy by increasing the conversion of ammonia to urea. Trials using ornithine alpha-glutarate did not demonstrate a benefit. In part, this may be because it only supplies one substrate for incorporation of ammonia. Relatively large doses of amino acids (18 g/day PO) appear to be necessary for any clinical benefit.

The mechanism of action of L-ornithine L-aspartate may extend beyond the urea cycle. Administration of ornithine aspartate to portal hypertensive rats results in high concentrations of glutamate in the plasma and cerebrospinal fluid (CSF) and an associated reduction in plasma ammonia. The elevated glutamate concentrations facilitate synthesis of glutamine by glutamine synthase, which is expressed at high levels in the liver, brain, and skeletal muscle. This mechanism may permit further significant reductions in ammonia levels within both the central nervous system and the systemic circulation. Indeed, increased glutamine synthase expression is induced in the skeletal muscle by portocaval shunting.

An increase in plasma concentrations of BCAAs also is an anticipated metabolic consequence of increased glutamate availability. It remains of uncertain significance and does not necessarily contribute to the improvement in hepatic encephalopathy documented in this experimental model.

Sodium benzoate has also been shown to be efficacious in reducing serum ammonia. It is conjugated to glycine to form hippuric acid, which is excreted in the urine. Similarly, phenylacetate is conjugated with glutamine to form phenacetylglutamine. Both of these organic acids have been used successfully to treat hepatic encephalopathy in some clinical trials.

Zinc supplementation

The urea cycle allows the conversion of ammonia to urea. Because two of the enzymes in this metabolic pathway require zinc as a cofactor and because reduced plasma zinc levels from increased urinary zinc losses are documented in hepatic encephalopathy, oral zinc supplementation is proposed as a treatment of this condition.

The measurement of serum zinc levels may not accurately reflect whole-body zinc status, but it would appear reasonable to supplement patients found to have low serum zinc levels with zinc gluconate.


Treatment efforts with flumazenil, a competitive antagonist of BZPs, are based on the GABA hypothesis; however, results of the small clinical trials performed to date are variable.

In two well-designed studies, flumazenil was found to be of value in a limited number of patients but clear factors that might permit their identification were not proposed; therefore, because of the difficulty in establishing a more generalized improvement in the patient's condition and the relatively short duration of action of the drug, it is not of convincing benefit.

Dopamine agonists

Parkinsonian or extrapyramidal symptoms may manifest with hepatic encephalopathy. Treatment with levodopa or bromocriptine is shown to result in improvement in clinical and electroencephalographic findings in anecdotal reports and small studies.

Although the use of bromocriptine is advocated for cases of refractory hepatic encephalopathy, well-designed prospective controlled trials have not been conducted.


Simón-Talero et al found evidence that a subgroup of patients with advanced cirrhosis and episodic hepatic encephalopathy may benefit from treatment with albumin. In a randomized, prospective, double-blind, controlled trial, 56 cirrhotic patients with an acute episode of hepatic encephalopathy received albumin (n = 26) or saline (n = 30), in addition to commonly administered therapy consisting of laxatives and 1200 mg of rifaximin per day.[9]

The investigators determined that albumin did not aid in reducing the percentage of patients with encephalopathy during the hospitalization period, with no significant difference found between the albumin and saline groups with regard to the percentage of patients at day 4 whose encephalopathy had resolved. At day 90, however, the survival rate in the albumin and saline groups did differ significantly (69.2% vs 40.0%, respectively). The investigators suggested that the development of encephalopathy possibly signals which patients with advanced cirrhosis may benefit from albumin therapy.[9]

Surgical Care

The definitive approach to management of portosystemic encephalopathy is orthotopic liver transplantation (OLT). Portosystemic encephalopathy as a complication of end-stage liver disease may warrant discussion of OLT. Indeed, even central nervous system structural changes evident on magnetic resonance imaging may be reversed slowly following OLT; however, a detailed discussion of OLT is beyond the scope of this article (see Liver Transplantation).

A retrospective cohort study by Laleman et al indicated that embolization can be used safely and effectively to treat hepatic encephalopathy in patients with large spontaneous portosystemic shunts (SPSSs).[10] The study involved 37 patients with refractory hepatic encephalopathy and a single large SPSS who underwent embolization. Within 100 days following the procedure, 22 patients (59.4%) were free of the encephalopathy, with 18 of them remaining free of it over a mean follow-up period of 697 days. One major, albeit nonlethal, procedure-associated complication occurred in the study.[10]


Nutritional therapy includes the following:

  • Therapy with branched-chain amino acids (BCAA) has been evaluated extensively as a treatment for hepatic encephalopathy. They are of suggested benefit in lowering the production of false neurotransmitters. Vegetable protein–based diets that exclude meat are relatively high in BCAAs and low in aromatic amino acid content. Clinical trials performed to date with vegetable protein–based diets or BCAAs have not demonstrated any consistent or convincing benefit with respect to the clinical manifestations of hepatic encephalopathy and do not support their widespread use for the treatment or prevention of hepatic encephalopathy.

  • Vegetable protein, however, may have other benefits. In patients with subclinical hepatic encephalopathy, one study demonstrated computer-analyzed electroencephalographic (EEG) findings improved with vegetable protein–based diets, despite the lack of any change in ammonia levels. Urinary 3-methyl-histidine excretion was increased with the use of an animal-protein diet, and the vegetable diet was associated with a decrease in urinary nitrogen excretion; thus, the nitrogen balance tended to be more positive with the vegetable-protein diet.

  • Patients with advanced liver disease often present with poor nutritional status and may develop a negative nitrogen balance exacerbated by acute or recurrent efforts to restrict dietary protein. Although the practice of strictly limiting dietary protein in these patients was advocated until relatively recently, adverse nutritional consequences now appear to outweigh any potential benefit for the prevention of hepatic encephalopathy. In general, patients with chronic liver disease should be advised to eat approximately 1 g of dietary protein per kilogram body weight per day. In this context, BCAAs may be a useful dietary supplement to help preserve skeletal muscle mass via protein sparing, with little risk of precipitating hepatic encephalopathy.

  • With a well-tolerated exogenous source of amino acids, the catabolic metabolism that otherwise results in negative nitrogen balance therefore may be offset. This is of particular importance in patients awaiting orthotopic liver transplantation, in whom a protracted nutritional decline may occur during the long wait for a suitable donor liver.


Consider restricting patients with encephalopathy from driving motor vehicles or operating potentially dangerous machinery. This may be difficult to justify to patients and their families in the absence of formal psychometric testing. Reaction times and judgment are typically impaired. Similarly, a patient may need to perform different job duties in order to avoid physical or, possibly, financial harm. Clearly, these issues may settle themselves because mental status changes resolve completely with therapy. These difficult issues need review on an individual basis and require periodic reassessment.

Patients need adequate rest, and their goal should be to sleep at least 8 hours at night. Strenuous activity should be avoided, but regular mild exercise is distinctly advantageous for maintaining bone mass and cardiovascular conditioning in anticipation of the long wait for a donor organ required for an orthotopic liver transplantation.



Medication Summary

The mainstay of therapy is lactulose, a nonabsorbed disaccharide. The actions of this agent are multiple and culminate in reduced delivery of ammonia to the brain. Lactitol, another disaccharide, may be better tolerated, but it is used relatively infrequently. Oral antibiotics are not used as often, but they have a role in patients who are intolerant of lactulose. Flumazenil, bromocriptine, branched-chain amino acids (BCAAs), and L-ornithine have been investigated but cannot be considered as first-line therapy.

Nonabsorbed disaccharides

Class Summary

Promote acidification, sterilization, and ammonium ion trapping in colon lumen. Increased stool frequency occurs, and all mechanisms likely reduce ammonia delivery to the systemic circulation and the brain. Inexpensive and generally well tolerated.

Lactulose (Cephulac, Cholac, Constulose)

DOC; inhibits diffusion of NH3 into blood by producing an acidic pH that causes conversion of NH3 to NH4.

Nonabsorbed antibiotics

Class Summary

Used for gut sterilization.

Neomycin PO (Neo-Tabs)

Second-choice drug. Small percentage of neomycin may be absorbed with each dose. Chronic administration may lead to nephrotoxicity or ototoxicity.

Rifaximin (Xifaxan)

Oral antibiotic that reduces ammonia-producing enteric bacteria in patients with hepatic encephalopathy. In vitro, elicits broad-spectrum activity against gram-positive and gram-negative aerobic and anaerobic enteric bacteria. Minimal systemic absorption (< 0.4%); concentrated in GI tract; low risk for inducing bacterial resistance. Rifampin structural analog. Binds to beta-subunit of bacterial DNA-dependent RNA polymerase, thereby inhibiting RNA synthesis.