Neurologic Consequences of Liver Disease

Ammonia

Sources of ammonia production derive mainly from the gastrointestinal tract through bacterial metabolism of urea and proteins in the large intestine, through deamination of glutamine in the small intestine via glutaminase, and from nitrogenous food products. The kidneys make a small amount of ammonia, and ammoniagenesis is increased during diuretic treatment and in hypokalemic states. In addition, in states of sarcopenia, release of glutamine from muscle cells further contributes to ammoniagenesis.

Under normal physiologic conditions, ammonia that is created enters the portal circulation and is converted to urea in the liver, and is subsequently cleared by the kidneys. However, as a result of cirrhosis and portosystemic shunting, the ammonia concentration rises in the blood and ammonia crosses the BBB. In the brain, astrocytes are the only cells that metabolize ammonia by the enzyme glutamine synthetase, converting glutamate and ammonia to glutamine. Glutamine is osmotically active, and thus an increase in glutamine concentration (due to an increase in ammonia concentration) leads to astrocyte swelling and edema. Furthermore, glutamine enters the mitochondria and is cleaved by glutaminase to ammonia and glutamate, which subsequently increases the intracellular ammonia concentration. This increase in intracellular ammonia concentration causes a “feed forward loop,” also known as the Trojan horse hypothesis , whereby intracellular ammonia leads to production of reactive oxygen and nitrogen species, causing further edema.

Over time, continued exposure of the brain to ammonia leads to physiologic disturbances. Neurally active compounds such as myoinositol and taurine are released from cells, of which lower levels of myoinositol have been associated with decompensated HE. Moreover, increased ammonia levels cause abnormal cerebral blood flow. Single photon emission CT studies have revealed altered cerebral perfusion with hyperammonemia results in redistribution of blood flow from the cortex to subcortical areas, which was associated with impaired performance on neuropsychiatric testing.

Inflammation

Sepsis is a frequently encountered precipitating factor for HE, and there is evidence that there are worse outcomes and severity of HE in patients with marked inflammation, in those patients with acute liver failure, and in those with cirrhosis. In addition to infection, inflammation can also result from gastrointestinal bleeding, obesity, and alterations in the intestinal microbiota. Proinflammatory cytokines, such as interleukin-1 (IL-1), IL-6, and tumor necrosis factor (TNF), work in conjunction with ammonia to worsen cerebral edema in HE. Both TNF and IL-6 have been shown to induce changes in the morphology of astrocytes with increased permeability of the BBB. Lipopolysaccharide has also been shown to enhance ammonia-induced changes to the BBB.

Microbiota

Alterations in the microbiota and its contribution to the development of HE have been well elucidated. Such alterations may be responsible for the formation or release of products such as ammonia, endotoxins, indoles, and oxindoles, which may lead to cognitive impairment. There is a dysbiosis in cirrhosis (dysbiosis rate) whereby there is reduced levels of members of autochthonous taxa ( Lachnospiraceae , Ruminococcaceae , and Clostridiales XIV) and increased levels of members of Enterobacteriaceae and Streptococcaceae with disease progression. A lower dysbiosis rate has been associated with HE and higher levels of endotoxemia. Endotoxemia contributes to liver damage, which activates the immune response and the inflammatory cascade, which as described already, can further exacerbate brain edema via its synergism with ammonia.

Neuromodulators

γ -Aminobutyric acid (GABA) is an inhibitory neurotransmitter that has been implicated in the pathogenesis of HE. There is increased GABAergic tone in the CNS via increased sensitivity of the astrocyte benzodiazepine receptor that activates the GABA system. It has been shown that an excess of benzodiazepine-like compounds that contribute to HE is derived from the intestinal flora, vegetables in the diet, and medications. Ammonia may also bind to the GABA receptor complex on the astrocyte, which may trigger synthesis of neurosteroids, which are GABA agonists. Other neurotransmitters, such as serotonin, acetylcholine, glutamate, and monoamines, have also been suggested to contribute in the pathogenesis of HE.

Manganese Deposition

The role of manganese in HE remains unclear. Manganese is cleared by the liver and excreted in bile in healthy individuals. As expected, excretion is impaired in cirrhosis secondary to portosystemic shunting. This results in increased systemic levels of manganese and cerebral manganese deposition. In vitro studies suggest that manganese toxicity may cause Alzheimer type II astrocytosis, brain tissue that can be found in cirrhotic patients, and altered astrocyte proteins. Furthermore, manganese may also alter glutamatergic neurotransmission, and increase expression of benzodiazepine-like receptors.

Overt Hepatic Encephalopathy

The prevalence of OHE in the United States ranges from 30% to 45% of patients with cirrhosis, with an annual incidence of 20%. This incidence is associated with a high rate of hospitalization, which has continued to rise. Associated with this high rate of hospitalization is the rising charge per hospitalization. The estimated cost of HE-related disease in the United States between 1993 and 2003 was $932 million, which is a conservative estimate. There is a substantial effect of OHE on the patients' health-related quality of life (HRQOL), which affects patients' mental and physical functioning.

OHE is associated with poor survival rates in patients with cirrhosis Bustamante et al. showed that in those patients with the first episode of HE, the survival rate probability was 42% at 1 year and 23% at 3 years despite controlling for other factors associated with death in cirrhosis. Stewart et al. found that advanced HE that required hospitalization remained a statistically significant predictor of mortality (hazard ratio, 2.6; 95% confidence interval, 1.7 to 3.8; p < 0.01), despite taking into account the Model for End-Stage Liver Disease score. Therefore, not only is HE an independent marker of mortality, its burden continues to rise in the United States.

Covert Hepatic Encephalopathy

The estimated prevalence of CHE ranges from 30% to 84% of the patients with cirrhosis tested worldwide. These estimates differ because of the differing populations and tests used to classify patients as having CHE. CHE has a wide-ranging effect on patients' daily lives and their functioning as productive members of society. Compared with healthy controls and patients without CHE, CHE patients exhibit severe impairments in the psychosocial aspects of social interaction, alertness, and emotional behavior, and in the physical domains of ambulation, mobility, body care, and movement. CHE also adversely affects sleep, work, home management, recreation, pastimes, and eating behavior. This leads to an overall reduction in all aspects of the HRQOL in CHE and engenders several complaints that are voiced in the hepatology clinic. Lastly, CHE is also a predictor of death and hospitalizations independently of the Model for End-Stage Liver Disease score.

Earning capacity in patients with CHE is reduced compared with that of cirrhotic patients who do not have CHE. This is especially true for those who have blue collar jobs because the profile of impairment in MHE affects these professions.

Driving skills are considered a balancing act among strategic, operational, and tactical strategies that use most skills that are adversely affected in CHE. Evaluation of driving skills is sensitive for patients because it could have medicolegal implications. It can be performed with a driving simulator, by on-road driving tests, and by actual driving outcome results. Studies performed by on-road driving tests in Germany and Japan have shown that patients with CHE are unfit to drive in 52% to 100% of cases when a driving instructor, who was masked as to their underlying psychometric function, assessed their driving. Wein et al. showed that the instructor had to intervene to prevent an accident 10 times more frequently in CHE patients than in cirrhotic patients without CHE. Driving simulator assessment has consistently shown in several studies that CHE patients have a higher risk of collisions. Their navigation skills on the simulator are also impaired compared with controls and cirrhotic patients without CHE (i.e., they got lost more frequently while driving). A study of divided attention (i.e., when the patient has to perform two tasks at once, reminiscent of using a cell phone while driving) showed poor ability of CHE patients to multitask while driving. Simulator studies also highlighted the difficulty with fatigue in patients with CHE, who had significantly more collisions during the second half of a 25-minute simulator task compared with the first half.

Actual driving accidents and violations by patients with cirrhosis were investigated with use of both anonymous and identified questionnaires. Patients with CHE responding to anonymous questionnaires had a significantly higher rate of traffic accidents and violations during the previous year and previous 5 years compared with those without CHE and those who were actively drinking alcohol. A study citing data from the patients' own statements and Department of Transportation records showed that patients with cirrhosis who had an abnormal performance on the inhibitory control test (ICT) had a significantly higher risk of traffic accidents during the previous year compared with those who did not have cirhossis. On prospective follow-up, patients in who CHE was diagnosed by the ICT and those with prior driving offenses were significantly likelier to have traffic accidents. There was no significant difference between the patient's admission of accidents and the official driving record. Another aspect of the driving assessment is the lack of insight patients with CHE have regarding their driving skills; patients with CHE consider themselves to be significantly better drivers than adult observers who were familiar with their driving skills. In contrast, patients without CHE and controls had personal driving skill assessments similar to those of their adult observers. Driving skills for those with CHE are therefore affected right from the simulator to the actual driving outcomes, and studies are needed to assess the effects of treatment on the skills of these individuals.

Diagnosis of Overt Hepatic Encephalopathy

Pure clinical diagnostic strategies for HE are flawed because of their inherent subjectivity. The West Haven criteria, which are modifications of the original Parsons-Smith criteria, are the most recognized system for clinical classification of HE ( Table 14-3 ). They divide patients into those with grade 0 HE, which is normal, through grade 4 HE, which is a coma ( Fig. 14-3 ).

Physical Examination in Patients With Overt Hepatic Encephalopathy

There is considerable variation in the physical examination findings of patients with cirrhosis, and the depth of the examination can often reveal signs of extrapyramidal disorders or changes in ocular movements that are consistent with liver disease–associated cognitive impairment. OHE tends to be a global process; therefore a previously unknown focal motor deficit is not considered typical. Hyperreflexia, positive Babinski sign, and asterixis are neurologic signs often found in OHE. Asterixis is defined as a flapping tremor associated with the disturbance in the oscillatory networks in the brain that is seen in the tongue, and in the upper and lower extremities. Tremulousness due to alcohol abuse or withdrawal can be confused with asterixis, and the nonspecificity of asterixis, which can also be found in carbon dioxide intoxication and uremia, should always be kept in mind.

Brain Imaging for the Diagnosis of Hepatic Encephalopathy

Patients with OHE rarely require imaging of the head for diagnosis unless there are doubts about their clinical condition. Imaging is advised for the diagnosis of HE only if there are atypical features present (i.e., seizures, appearance of new focal changes) to exclude other diagnoses. It can also unearth consequences of falls in patients that can masquerade as HE. Although a head CT scan can show loss of cerebral volume in HE, this is nonspecific and is often found in patients with cirrhosis without signs of HE as well. In patients who are imaged though, the predominant finding on an MRI scan is hyperintensity of the basal ganglia on T1-weighted imaging. This finding is probably related to manganese deposition and is often reversible after liver transplantation. Other modalities such as magnetic resonance spectroscopy, PET, and functional MRI are predominantly used for research purposes. Magnetic resonance spectroscopy shows an increase in glutamine and glutamate levels, and a compensatory decrease in brain myo -inositol levels. There are characteristic changes in blood flow and cerebral activation seen on PET and functional MRI that mirror the ammonia transport and assimilation in the brain. However, because of the expense, time, and nonspecificity of these imaging findings, it is unlikely that they will be used in clinical practice for the diagnosis of HE.

Abdominal Imaging

In patients with persistent HE, however, there is a role for a CT scan of the abdomen to evaluate the presence of collaterals that could sustain this condition despite adequate therapy. Large spontaneous portosystemic shunts were detected in 71% of HE patients compared with only 14% of those without HE. This is important because interventional radiologic embolization techniques can obliterate these shunts and improve the control of HE.

Miscellaneous Imaging

Chest imaging is a must for the evaluation of a possible chest infection or another possible trigger of OHE (see Management of Acute Overt Hepatic Encephalopathy section).

Ammonia Levels for the Diagnosis of Overt Hepatic Encephalopathy

There is considerable controversy regarding the need to obtain ammonia levels in the routine clinical management of HE. Although it is well known that ammonia metabolism abnormalities play a key role in the pathogenesis of HE, it is inaccurate to assume a direct correlation between ammonia levels and consciousness levels for individual patients. Although this is true in groups of several hundred patients, the same cannot be said for an individual patient. Clinical experience shows that there are patients who are severely obtunded with HE with normal ammonia levels and vice versa; however, the clinical decision to initiate therapy should always and only be based on the clinical and mental status rather than the ammonia level alone.

Complicating these assessments are the methods for drawing the ammonia sample, which should ideally be performed without a tourniquet, should be tested within 30 minutes, and should be kept on ice until testing. Also the false elevation of the ammonia level after seizure activity adds another twist to the interpretation of ammonia levels. The specific situations where ammonia levels may help are in the rare patients with urea-cycle disorders, in the initial diagnosis of a patient with a coma who does not have a history of cirrhosis, and in predicting the development of HE.