Is minimal hepatic encephalopathy completely reversible following liver transplantation?

Hepatic encephalopathy (HE) is a metabolic neuropsychiatric syndrome that has the potential for full reversibility with liver transplantation (LT). However, in recent years there have been an increasing number of reports that challenge this traditional view. Cirrhotic patients with minimal hepatic encephalopathy (MHE) have no recognizable clinical symptoms of HE but have mild cognitive and psychomotor deficits. These abnormalities impact significantly on the patients' quality of life. Current research in the area of MHE has focused upon the pathophysiological basis of this syndrome to better define the condition, which would allow for the development of better diagnostic modalities and new therapies. From the pathophysiological perspective, the pathogenesis of MHE is thought to be similar to more advanced grades of HE. Ammonia has been suggested to be a factor in the pathogenesis of this syndrome. We have recently shown that induction of hyperammonemia in patients with cirrhosis results in an increase in brain glutamine (a consequence of brain ammonia metabolism) and a reduction in the brain magnetization transfer ratio, which is indicative of brain edema.1 Cordoba et al. showed evidence of low-grade cerebral edema in cirrhotic patients with MHE, which was completely resolved following LT.2 Supporting post-LT neurological benefits, many other studies have demonstrated substantial improvement with neuropsychological tests in MHE2-8 and quality of life.9, 10 after LT. In addition, others have demonstrated neurophysiological improvement in MHE after LT with brain metabolic alterations detected by means of proton magnetic resonance spectroscopy,3, 6, 11 positron emission tomography,12 and electroencephalography (EEG)13 HE, hepatic encephalopathy; MHE, minimal hepatic encephalopathy; LT, liver transplantation; EEG, electroencephalography; WE, Wernicke's encephalopathy. In this issue of Liver Transplantation, Mechtcheriakov et al. report on the results of a study evaluating changes in neuropsychological tests in 14 liver transplant recipients pre- and post- (average, 21 months) LT. They used a neuropsychological test battery Visuo-Motor and Visuo-Constructive Performance (VMCP): average percentile score of four selected neuropsychological tests) in which overall visuomotor and visuoconstructive performance prior to and after LT was evaluated. A trend toward improvement in average VMCP score was demonstrated in post-LT versus pre-LT recipients. However, the tests remained abnormal, compared with a group of healthy volunteers. Furthermore, this incomplete improvement of VCMP resulted from an “inhomogeneous response of the patients to LT.” The paper demonstrates the distribution of the improvement rates within the group of patients and shows that there are two opposite tendencies in the post-LT course of treatment, one toward improvement and one toward deterioration. In the improvement group of seven patients, an increase in VCMP score by 2.52 was concluded as sufficient to be considered as completely reversible, while in the other group the VCMP score deteriorated post-LT. Even though a small number of patients were studied, this finding is important and needs to be evaluated and further discussed. Proposed hypothesis. Hepatic Encephalopathy manifests clinical features that are consistent with “deliriumlike” or “dementialike” features. Both of these feature may produce alteration in psychometric tests that would be consistent with a diagnosis of minimal hepatic encephalopathy. The former is likely to be metabolic in origin, whereas the latter is likely to be due to a structural brain lesion, which may or may be specific to liver disease. Hence, reversibility of the deranged neuropsychological tests with liver transplantation would be determined by primary pathology underlying the abnormal psychometric tests rather than a diagnosis of minimal hepatic encephalopathy. Mechtcheriakov et al.'s study confirms previous reports that neuropsychological tests may not resolve completely with LT. Tarter et al.14 first prospectively evaluated neuropsychological function in 62 patients prior to and after LT. They showed deficits on tasks measuring memory and visuospatial capacity prior to LT (compared with matched controls), most of which were ameliorated but not completely restored following LT. Interestingly, Hockerstedt et al.15 reported that the majority of neurological impairment disappeared after LT in eight patients; however, four subjects showed new, mild changes in neurophysiological and neuropsychological tests post-LT, suggesting an operative factor causing MHE, and evidence that MHE is a multidimensional syndrome. A possible “dimension” of MHE has been suggested to be parkinsonism related to cirrhosis, which has been recently described to be largely underestimated.16 Lazeyras et al.17 studied 14 unselected patients with cirrhosis and MHE (n = 8) and no HE (n = 6) pre- and post-LT. Four months following LT in patients with preoperative MHE, they found a persistence of mild parkinsonian signs. Although they observed a normalization of the spectroscopic changes, pallidal hyperintensity remained abnormal in five out of eight cases. In patients without MHE pre-LT, these abnormalities were not detected post-LT, indicating that the Parkinsonian features were unlikely to be related to the operation or the immunosuppressive drugs. More recently, O'Carroll et al.9 performed a large prospective study over a 3-year period in which neuropsychological data were prospectively collected from 164 patients who were assessed for LT. One-year post-LT, follow-up data from transplant recipients showed significant improvement in most psychological domains. However, the reaction time remained significantly worse than controls who were also followed prospectively over the same period. In addition, the quality of life measure did not reach normal control values. The data at 1-year follow-up remained consistent after 3 years and supports the notion that neuropsychological function independent of post-LT time may not reach normal values. Furthermore, in a recent study, Lewis and Howdle demonstrated that patients who survived 10 years post-LT had significant cognitive dysfunction and poor health-related quality of life.18 This is supporting evidence for a “dementia like” parameter of MHE that is irreversible following LT (Fig. 1). It is important to note that in Mechtcheriakov et al.'s study, none of the clinical factors such as duration of liver cirrhosis and its severity (as determined by Child-Pugh classification), presence and severity of esophageal varices, history of gastrointestinal bleeding, previous episodes of HE, post-LT immunosuppressive treatment, time between pre- and post-LT neuropsychological testing, relative level of functioning pre-LT, education, and age did not influence whether the patients improved or deteriorated after LT. These important findings open up new areas of research and require further investigation into the pathophysiology of MHE. This is supported by the study of O'Carroll et al.,9 in which the Child-Pugh rating at pre-LT assessment was not predictive of psychological outcome after 1 year. However, more severe liver disease at pre-LT assessment was associated with greater slowing of reaction times and an increased bioelectric brain dysfunction.13 In addition, regardless of the severity of underlying the liver disease, HE independently has detrimental effects on quality of life.19 This irreversibility of MHE (cognitive deficits) may reflect the neurological sequelae of underlying brain pathology. This is supported in studies in which abnormalities demonstrating brain edema and gross atrophy were documented by CT scan in cirrhotic patients with advanced liver disease who were not overtly encephalopathic.14 Therefore, the neurological deficits measured after LT reflect permanent central nervous system disturbance. However, in the study by Mechtcheriakov et al., all patients were screened for significant cortical atrophy or other structural brain changes as revealed by conventional neuroimaging. The implications of post-LT neurological deficits are important (particularly when the patient is being considered for LT), and attempts should be made to diagnose this subgroup pre-LT to maximize post-LT recovery of functional daily activities and quality of life. It is unfortunate in Mechtcheriakov et al.'s study that the demographic or clinical characteristics of the patients within the “improved” and “deteriorated” groups were not shown, even though the etiology of cirrhosis did not influence pre-LT functional level and post-LT improvement. However, of the 14 patients, five were alcoholic. Alcoholic versus nonalcoholic patients is an important comparison that needs to be carefully followed. Kril and Butterworth and 20 performed detailed histological evaluation of brain tissue from patients (30 alcoholic, six nonalcoholic) with end-stage chronic liver disease who died in hepatic coma. Alzheimer type II astrocytosis, a pathological characteristic of chronic liver disease, was apparent in all cases. None of the patients were diagnosed with acquired non-Wilsonian hepatocerebral degeneration, and nine patients (all alcoholic) were diagnosed with Wernicke's encephalopathy (WE) or alcoholic cerebellar degeneration conditions associated with neuronal loss. Results showed a high incidence of diencephalic (25%) and cerebellar (56%; 17 alcoholic, three nonalcoholic) lesions in patients with end-stage liver failure. Both pathologies were present in 17% of cases. Since no neuropathological characteristics of either acute or chronic WE were observed in the brains of six nonalcoholic patients despite equivalent severity of decompensated liver failure and of Alzheimer's type II astrocytosis, it may be suggested that thiamine deficiency resulting from chronic alcoholism rather than liver disease is the major contributing factor in WE. Also, in the Kril and Butterworth study, a similar incidence of cerebellar degeneration was observed in alcoholic and nonalcoholic cirrhotic patients, suggesting that chronic liver disease per se may contribute to cerebellar degeneration. These findings show that the brains of a high proportion of cirrhotic patients with end-stage liver disease manifest concomitant unsuspected diencephalic and cerebellar pathology. Results of Mechtcheriakov et al.'s study suggest a possible neuropathological explanation for the observed “deteriorating” group in patients with advanced liver disease and MHE following LT. Another important issue in relation to persistent MHE post-LT is whether the incomplete improvement in neuropsychological and neurophysiological function relates to factors related to the immunosuppressive regime. It is, unfortunately, impossible to test this hypothesis on LT recipients, since it would be unethical to withhold a needed treatment. However, Tarter et al.14 compared LT recipients with Crohn's disease subjects to control for the nonspecific effects of medical illness and immunosuppressive therapy. They found that not all of the deficits observed were relative to the Crohn's control group and that these impairments were not attributable to encephalopathy alone but also may reflect the effects of complications of having a chronic illness. Furthermore, no spectral EEG alteration was observed in two controls to whom cyclosporine was administered following long-term treatment of severe psoriasis.13 In addition, in the study by O'Carroll et al.,9 the type of immunosuppression (cyclosporine vs. tacrolimus) did not have differential effects on quality of life, fatigue, or affective status, although those given cyclosporine showed greater improvements at 1-year follow-up on simple and choice reaction times. The other factor that needs to be considered as an explanation for the “persistence” and/or deterioration of the neuropsychological function after LT is the possible effect of the operative procedure itself. The best data are derived from studies evaluating cognitive function in patients undergoing cardiopulmonary bypass. Depending upon the sensitivity of the test used, the incidence of cognitive deficits after coronary artery bypass grafting varies between 15 and 40%, 1-year after the operation. Van Dijk et al.21 studied cognitive function in patients undergoing coronary artery grafting with or without bypass and showed no significant differences in the incidence of cognitive dysfunction. Although the population of patients was different, the operative procedure may have produced alterations in cerebral perfusion and oxygenation, microemboli, or other undefined changes, which may have critically affected postoperative neuropsychological function. In keeping with this hypothesis, Phillips et al.22 studied changes in cerebral blood flow in cirrhotic patients undergoing LT. They showed that in patients with HE, the cerebral metabolic rate for oxygen was low during all stages of LT. These data suggest that patients with HE undergoing LT may be more susceptible to hypoxemic brain injury during LT. It has been demonstrated that psychometric test abnormalities are not specific for MHE, and therefore these test results may be persistently abnormal after LT for reasons unrelated to liver function. Psychometric tests are less reversible than EEG tests, and this suggests that these tests may access different components of HE.23 This discrepancy between EEG and psychometric testing supports the concept of a “multidimensional definition of HE,” whereby MHE may be an entity different from either episodic or persistent encephalopathy, with a distinct pathogenesis.24 It is very difficult to predict whether neurological symptoms will disappear after LT. As Ciancio et al. elegantly describe, “…additional studies should consider the different pathogenetic background and prognostic meaning of different tools used in the diagnosis of HE.”13 In conclusion, these recent studies suggesting that MHE may not completely recover with LT seriously question the validity of the hypothesis that HE is purely a metabolic syndrome. Nonreversibility of MHE following LT may result from preexisting other pathologies (WE, non-Wilsonian hepatocerebral degeneration, alcoholic cerebellar degeneration), or it is possible that MHE indeed represents a collection of syndromes in which some are reversible and some are not. The neuropathological studies identifying either collectively or individually the features of atrophy, diencephalic and cerebellar pathology, and basal ganglia hyperintensity in patients with HE, together with emerging clinical observations suggesting an element of “irreversibility” of MHE with LT, merits a reexamination of the existing concepts of HE. The question of whether MHE is fully reversible remains unanswered and therefore sets the stage for a new area of research in which studies of detailed neuropsychological and radiological assessments prior to and after LT, monitoring for hypoxemic brain injury during LT and assessing the neuropsychological effects of immunosuppressive drugs, should be thoroughly investigated.