Laboratory Evaluations

The diagnosis of HE is based on the signs and symptoms of cerebral dysfunction in a setting of hepatic failure. Usually, standard laboratory test results, including serum bilirubin and hepatic enzymes, are abnormal. Products of normal hepatic function, including serum albumin and clotting factors, often are low, leading to elevation of the International Normalized Ratio (INR). Measurements of the arterial ammonia level may be helpful in diagnosing HE. When obtaining blood samples for an ammonia determination, care must be taken to be certain that the sample is of arterial origin (venous ammonia levels may be artificially high, especially after the outpouring of ammonia by muscle made ischemic by applying a tourniquet). The sample should be placed on ice and carried by hand to the laboratory for immediate analysis. Delays can result in ammonia production in the specimen, producing a spuriously elevated result.

Several consensus conferences sponsored by the International Society for Hepatic Encephalopathy and Nitrogen Metabolism have made recommendations concerning the use of electrophysiological and neuropsychological tests to evaluate patients with HE. The favored electrophysiological tests are those that are responsive to cortical function and include event-related electrical potentials (ERPs) such as P300 tests, the EEG, and visual and somatosensory ERPs. Neuropsychological tests are useful for diagnosing minimal HE. Domains to be evaluated, in descending order of desirability, include processing speed, working memory, anterograde verbal memory, visuospatial ability, anterograde visual memory, language, reaction time, and motor functions.

The EEG may be the most useful of the commonly used laboratory diagnostic tests. Bursts of moderate- to high-amplitude (100-300 µV), low-frequency (1.5-2.5 Hz) waves are the most characteristic abnormality. There are three stages in the EEG evolution: a theta stage with diffuse 4- to 7-Hz waves; a triphasic phase with surface-positive maximum deflections; and a delta stage characterized by random arrhythmic slowing with little bilateral synchrony. Computerized analysis of the EEG, designed to identify abnormalities in the spectra, may become a valuable means to identify patients with minimal encephalopathy. Abnormal ERPs may be found in patients with minimal encephalopathy. A combination of visual-evoked potentials, auditory P300s, and selected neuropsychological tests (such as Trailmaking Tests A and B) may be useful in detecting minimal encephalopathy in cirrhotic subjects. Although there has been less experience with auditory P300 potential recordings, in which the subject is asked to discriminate between a rare and common tone, differences in latencies and waveforms also have been associated with encephalopathy. It is uncertain whether this less complex approach to detecting minimal encephalopathy will prove to be more reliable and cost-effective than a focused neuropsychological test battery.

Neuropsychological tests are an underused and valuable means of diagnosing encephalopathy and monitoring the response to therapy (see Chapter 34). Sixty percent or more of all patients with cirrhosis with no overt evidence of encephalopathy exhibit significant abnormalities when given a battery of neuropsychological tests. Tests of attention, concentration, and visuospatial perception are the most likely to be abnormal. A test battery consisting of Trailmaking Tests A and B, serial dotting, line tracing, and the digit-symbol subtest of the Wechsler Adult Intelligence Scale, Revised, has been recommended for evaluating patients who may have hepatic encephalopathy. This battery is sensitive and relatively specific for the disorder, compared with other metabolic encephalopathies. Patients with alcoholic cirrhosis typically have more difficulty with memory deficits than patients with nonalcoholic cirrhosis. Even though these patients appear to be normal, the degree of impairment, particularly in the visuospatial sphere, may be severe enough to interfere with the safe operation of an automobile or other dangerous equipment. A study comparing patients with minimal encephalopathy with nonencephalopathic patients with cirrhosis and a third group with gastrointestinal disease, found that those with minimal encephalopathy performed the worst during an on-the-road driving test. Specific problems centered on handling, adaptation to road conditions, and accident avoidance. Language functions are usually normal. These data, combined with other studies showing that the quality of life is affected by these abnormalities, suggest that neuropsychological tests should be used more extensively for routine evaluation of all patients with cirrhosis, particularly those without overt evidence of HE.

Although the diagnosis of HE is typically made on the basis of clinical criteria, neuroimaging techniques are commonly employed to exclude structural lesions. Magnetic resonance imaging (MRI) and spectroscopic studies have revealed new insights into the pathophysiology of HE (Lockwood et al., 1997). On T1-weighted images, it is common to find abnormally high signals arising in the pallidum. These are seen as whiter-than-normal areas in this portion of the brain, as shown in Fig. 56.1. In addition to these more obvious abnormalities, a systematic analysis of MR images shows that the T1 signal abnormality is widespread and found in the limbic and extrapyramidal systems, and generally throughout the white matter. A generalized shortening of the T2 signal also occurs. This abnormality is less evident on visual inspection of the images because of the generally short duration of T2 signals. These abnormalities have been linked to an increase in the cerebral manganese content. The abnormalities become more prominent with time and regress after successful liver transplantation. The unexpected finding of high T1 signals in the pallidum should suggest the possibility of liver disease.

Fig. 56.1 T1-weighted magnetic resonance images from a patient with cirrhosis of the liver. Note high signal in basal ganglia, cerebral peduncles, and substantia nigra.

Proton MR spectroscopic techniques also have been applied to the study of patients with cirrhosis and are available in many centers. In the absence of absolute measures that are referable to concentrations, the signal of specific compounds is usually referenced to creatine and expressed as a compound-to-creatine ratio. There is general agreement among studies that an increase in the intensity of the signal occurs at approximately 2.5 ppm; this is attributed to glutamine plus glutamate. With high field–strength magnets, this peak can be resolved into its components, and the increase is attributed to glutamine, as expected on the basis of animal investigations. Correlations between the glutamine concentration, generally considered to be a reflection of exposure of the brain to ammonia and the severity of the encephalopathy, have led some to propose that MR spectroscopy may be useful in the diagnosis of HE.

Myoinositol and choline signals decrease, whereas N-acetylaspartate resonances (a neuronal marker) are consistently normal. Neuroimaging studies are not generally required in patients with HE. Imaging is useful in the diagnosis of coexisting structural lesions of the brain, such as subdural hematomas or other evidence of cerebral trauma, or complications of alcohol abuse or thiamine deficiency, or both, such as midline cerebellar atrophy, third ventricle dilatation, mamillary body atrophy, or high signal–strength lesions in the periventricular area on T2 FLAIR images.

Pathophysiology

The pathophysiological basis for the development of HE is still not completely known. However, treatment strategies for the disorder are all founded on theoretical pathophysiological mechanisms. A number of hypotheses have been advanced to explain the development of the disorder. Suspected factors include hyperammonemia, altered amino acids and neurotransmitters—especially those related to the γ-aminobutyric acid (GABA)-benzodiazepine complex—mercaptans, and short-chain fatty acids. Although none of the current hypotheses is completely capable of explaining the development of HE, it is likely that ammonia plays a central role. Because of the complexity of the metabolic derangements that attend liver disease, other factors may contribute synergistically to the development of this complex disorder.

Cerebral Blood Flow and Glucose Metabolism

Whole-brain measurements of cerebral blood flow (CBF) and metabolism are normal in patients with grade 0 to 1 HE. Reductions occur in more severely affected patients. Sophisticated statistical techniques designed to analyze images have made it possible to identify specific brain regions in which glucose metabolism is abnormal in patients with low-grade encephalopathy and abnormal neuropsychological test scores (Lockwood et al., 2002). These positron emission tomography (PET) data show clearly that minimal forms of HE are caused by the selective impairment of specific neural systems rather than global cerebral dysfunction. Reductions occur in the cingulate gyrus, an important element in the attentional system of the brain, and in frontal and parietal association cortices. These PET data are in accord with cortical localizations based on the results of neuropsychological tests. Fig. 56.2 shows the results of correlation analyses between scores on selected neuropsychological tests and sites of reduced cerebral glucose metabolism.

Fig. 56.2 Correlations between performance, as measured by age-corrected z scores, and metabolism for various subtests in the composite battery. Only those subjects able to complete the test are included in the analyses. The statistical parametric mapping Z image projections are as in the other figures. They all show significant correlations with bilateral parietal associative cortex, with increasing correlations with frontal regions.

(Used with permission from Lockwood, A.H., Weissenborn, K., Bokemeyer, M., et al., 2002. Correlations between cerebral glucose metabolism and neuropsychological test performance in nonalcoholic cirrhotics. Metab Brain Dis 17, 29-40.)

Role of Ammonia

Hepatic encephalopathy is linked to hyperammonemia. Patients with encephalopathy have elevated blood ammonia levels that correlate to a degree with the severity of the encephalopathy. Metabolic products formed from ammonia—most notably glutamine and its transamination product, α-ketoglutaramic acid—also are present in excess cerebrospinal fluid (CSF) in patients with liver disease. Treatment strategies that lower blood ammonia levels are the cornerstone of therapy.

Tracer studies performed with ¹³N-ammonia have helped clarify the role of this toxin in the pathophysiology of HE. Ammonia and other toxins are formed in the GI tract and carried to the liver by the hepatic portal vein, where detoxification reactions take place. Portal systemic shunts cause ammonia to bypass the liver and enter the system circulation, where it is transported to the various organs as determined by their blood flow. The liver is the most important organ for the detoxification of ammonia. However, in patients with portacaval shunting of blood, because of the formation of varices, TIPS, or other surgically created shunts, skeletal muscle becomes more important as the fraction of blood bypassing the liver increases. Under the most extreme conditions, muscle becomes the most important organ for ammonia detoxification. It is partly for this reason that nutritional therapy for patients should be designed to prevent development of a catabolic state and muscle wasting.

Ammonia is always extracted by the brain as arterial blood passes through the cerebral capillaries. When ammonia enters the brain, metabolic trapping reactions convert free ammonia into metabolites (Fig. 56.3). The adenosine triphosphate (ATP)-catalyzed glutamine synthetase reaction is the most important of these reactions. The blood-brain barrier is approximately 200 times more permeable to uncharged ammonia gas (NH³) than it is to the ammonium ion (NH⁴⁺); however, because the ionic form is much more abundant than the gas at physiological pH values, substantial amounts of both species appear to cross the blood-brain barrier. Because of this permeability difference and because ammonia is a weak base, relatively small changes in the pH of blood relative to the brain have a significant effect on brain ammonia extraction. As blood becomes more alkalotic, more ammonia is present as the gas and cerebral ammonia extraction increases; however, the role this has in the production of HE is not known. The permeability surface-area (PS) product of the blood-brain barrier may be affected by prolonged liver disease. However, the experimental data about this change are in conflict: one study reported an increase in the PS product, and another reported no change. An increased PS product could explain in part the toxin hypersensitivity that develops with time.

Fig. 56.3 Human ammonia metabolism. The brain becomes more sensitive to ammonia as time progresses. Although the reasons for this are largely unknown, studies of cerebral ammonia metabolism have shown that the blood-brain barrier is more permeable to ammonia in patients with liver disease than in controls. Thus at a given arterial ammonia concentration, more ammonia enters the brain of a patient with cirrhosis than a patient without cirrhosis. This observation may explain the presence of encephalopathy in patients with near-normal arterial ammonia levels. In addition, ammonia may cause anorexia by stimulating hypothalamic centers, leading to reductions in muscle mass and an impaired ability of muscle to detoxify ammonia. GI, Gastrointestinal.

(Adapted from Lockwood, A.H., McDonald, J.M., Reiman, R.E., et al., 1979. The dynamics of ammonia metabolism in man: effects of liver disease and hyperammonemia. J Clin Invest 63, 449-460.)

Other Pathophysiological Mechanisms

Neuropathology

The Alzheimer type II astrocyte is the neuropathological hallmark of hepatic coma. An account of the original descriptions of this change was provided in translation by Adams and Foley in 1953. In this report, they presented their own findings of this astrocyte change in the cerebral cortex and the lenticular, lateral thalamic, dentate, and red nuclei, offering the tentative proposal that the severity of these changes might be correlated with the length of coma. The cause of the astrocyte change was established by studies that reproduced the clinical and pathological characteristics of HE in primates by continuous infusions of ammonia. In studies of rats with portacaval shunts, astrocyte changes become evident after the fifth week. Before coma develops, astrocytic protoplasm increases and endoplasmic reticulum and mitochondria proliferate, suggesting that these are metabolically activated cells. After the production of coma, the more typical signs of the Alzheimer type II change became evident as mitochondrial and nuclear degeneration appeared. Norenberg (2007) suggested that HE is an astrocytic disease, although oligodendroglial cells are affected as well. More recent evidence from his laboratory has shown that ammonia affects a wide variety of astrocytic functions and aquaporin-4.

The neuropathological-neurochemical link between astrocytes and the production of hyperammonemic coma is strengthened by immunohistochemical studies that localized glutamine synthetase to astrocytes and their end-feet. Similar findings for glutamate dehydrogenase have been described. Long-standing or recurrent HE may lead to the degenerative changes in the brain characteristic of non-Wilson hepatocerebral degeneration. Brains of these patients have polymicrocavitary degenerative changes in layers five and six of the cortex underlying white matter, basal ganglia, and cerebellum. Intranuclear inclusions that test positive by periodic acid-Schiff also are seen, as are abnormalities in tracts of the spinal cord.

Treatment

Ideally, the management of cirrhosis should involve a cooperative effort between hepatologists, surgeons, neurologists, and psychologists, with additional input from nurses and dieticians. Practice guidelines published by the American College of Gastroenterology identify four goals: (1) provide supportive care, (2) identify and treat precipitating factors, (3) reduce the nitrogenous load from the gut, and (4) assess need for long-term therapy (Blei and Cordoba, 2001; Ferenci et al., 2002).

Initial diagnostic and therapeutic efforts should be directed at the identification and mitigation of precipitating factors and reducing the nitrogenous load arising from the GI tract. This is accomplished by a brief withdrawal of protein from the diet and the administration of cleansing enemas, followed by the use of lactulose. Antibiotics may be used as an alternative to lactulose. After the acute phase of HE, patients should receive the maximum amount of protein that is tolerated. Prolonged periods of protein restriction should be avoided. Protein is required for the regeneration of hepatocytes and prevention of a catabolic state and muscle wasting.

In patients who have cirrhosis without overt encephalopathy, diagnostic efforts should be directed toward identifying patients with minimal encephalopathy and monitoring the effects of treatment. The inappropriate terms, subclinical or latent HE have been too commonly applied to patients with minimal encephalopathy. Patients with minimal encephalopathy have a diminished quality of life and benefit from therapy, typically lactulose. Although rigorous criteria have not been developed to establish this diagnosis, deficits on neuropsychological test scores are usually used as the criterion. Some have advocated the use of computerized EEG analysis for this purpose, with a focus on abnormal slowing seen on an analysis of the spectrum. Follow-up testing is needed to monitor treatment.