Nearly 100 years ago, Van Woerkem (1) reported a patient with cirrhosis who at the age of 50 years had developed progressive rigidity, tremor, somnolence and emotional instability. A post-mortem neuropathological examination of this patient showed bilateral striatal nerve cell degeneration accompanied with a widespread proliferation of astrocytes. This patient is considered to represent the first published case of the so-called acquired hepatolenticular degeneration (AHD) – a rare complication of cirrhosis or portosystemic shunting in the absence of intrinsic hepatocellular disease, and the counterpart of the ‘hereditary hepatolenticular degeneration’ or ‘Wilson's disease’. In 1965, Victor et al. published a landmark paper on this disorder reviewing a couple of case reports as well as the characteristic clinical findings of 27 of their own patients supplemented by neuropathological findings in 17 of these patients (2). The most prominent clinical features in their case series were dementia, dysarthria, ataxia, tremor and choreoathetosis. A few patients, in addition, showed rigidity as did some of those cases published before. Victor et al. pointed out that ‘All of the patients had gross portal-systemic collaterals and twelve had had surgically-created shunts’, and ‘In two patients the shunt alone seemed to have been responsible for the neurological syndrome, the liver being normal at a time when the neurological symptomatology was already well developed’ (2). No substantial new aspect has been added to the understanding of the syndrome after their excellent description for decades. But then, in the 90s of the last century, a possible role of a manganese deposition in the basal ganglia in the development of hepatic encephalopathy (HE) achieved increasing interest (3) triggered by the observation of characteristic bilateral pallidal signal alterations on brain magnetic resonance images from patients with cirrhosis (4), and several papers focused upon the parkinsonian symptoms in patients with cirrhosis (5–7). Movement alterations were acknowledged as one of the most characteristic features of HE and bradykinesia, tremor and rigidity were described to be detectable with precise neurological examination even in about one-third of patients with cirrhosis who were considered neurologically unaffected on the first view (8).
While these observations clearly improved our understanding of HE in general, they did not help to really understand what is going on in patients who develop acquired hepatocerebral degeneration (AHD). We do not know why some patients with excessive portosystemic shunts (with and without severe liver disease) develop AHD while others do not. We also do not know whether AHD represents an extreme manifestation of HE as it is usually assumed or if it comprises an entity of its own. Neither an excessive manganese deposition in the basal ganglia nor the combined toxic effects of manganese and ammonia suffice as an explanation for the development of AHD, because the vast majority of patients who are exposed to both toxins do not react with the development of progressive extrapyramidal and/or cerebellar symptoms as in AHD, which do usually not respond to the classical plasma ammonia lowering therapy of HE. New strategies have to be applied to be able to succeed in the clarification of the pathophysiology of AHD.
Today, our knowledge is still based on some case series and multiple single case reports. Thereby, even the prevalence of AHD is unknown. Two recent retrospective analyses of about 800 and 1000 patients, respectively, estimated a prevalence of about 1% (9, 10). A prospective study in a small group of patients who were examined while they were assessed as potential candidates for liver transplantation in contrast, detected extrapyramidal symptoms of an extent, that was suspicious for parkinsonism in about 20% of the patients (11). Kang et al. in their study published in the current issue of this journal chose a totally different approach (12). They identified all patients from a representative national health insurance database who had been given the diagnosis of cirrhosis throughout the year 2001, and then built up a comparison group of patients from the same data base that was matched in terms of sex, age and the year of their index ambulatory care visit. Patients aged <40 years were excluded from the analysis as were patients who received antipsychotic or dopaminergic drugs at the time point of their index ambulatory care visit. Both cirrhosis and control group were then checked for the appearance of the diagnosis of idiopathic Parkinson's disease or secondary parkinsonism over a time period of 7 years after the index ambulatory care visit. Patients with the diagnosis of Parkinson's plus syndrome were not included in the analysis, nor were patients who had a diagnosis of HE or took lactulose, or patients who had been diagnosed with idiopathic Parkinson's disease or parkinsonism but had not been treated with dopaminergic drugs. During the follow-up period, 3.5% of the patients with cirrhosis and 1.4% of the patients in the comparison group received a diagnosis of Parkinson's disease or parkinsonism. The data indicate a prevalence of AHD above the 1 or 2% level that has been discussed so far (9, 10, 13). And as the authors point out, the ‘strength of (their) study lies in its longitudinal database and large population size capable of representing the risk of parkinsonism among patients with cirrhosis in Taiwan’. This epidemiological approach is commendable, and provides the opportunity to clarify at least the incidence and prevalence of AHD in patients with cirrhosis. But, several questions remain unanswered.
What are we looking for, if we talk about parkinsonism or Parkinson's disease associated with cirrhosis? Kang et al. state that the aim of their study is ‘to investigate the risk for parkinsonism during a 7-year follow-up period after a diagnosis of cirrhosis’ (12). They considered patients with the diagnosis ‘idiopathic Parkinson's disease’ and ‘secondary parkinsonism’. So, what were they looking for?
Were they interested to find out if idiopathic Parkinson's disease is more frequent in patients with cirrhosis, or whether the presence of liver disease alters the natural course of Parkinson's disease – an aspect that could be discussed considering their data showing a similar prevalence of Parkinson's disease and parkinsonism in liver patients and controls above the age of 65 years, but a significantly increased prevalence in the patients younger than 65 years. The latter might of course also be an artefact owing to the age distribution of the patients with non-alcoholic steatohepatitis-associated cirrhosis. Probably, Kang and colleagues were looking for the prevalence of AHD. Then, however, they should have also included patients younger than 40 years, because AHD also develops in adolescents and young adults (2, 13, 14). And they should also have included patients with Parkinson plus syndrome, because the combination of dysarthria, ataxia, tremor, rigidity, hypo- and dyskinesia and cognitive dysfunction, which is characteristic for AHD, is very likely to be classified as Parkinson plus syndrome if people do not consider the symptoms being related to the accompanying severe liver disease. In addition, they should have included patients with parkinsonism but no dopaminergic treatment, because a lack of treatment does not argue against a diagnosis.
The different case series and single cases so far reported show that AHD may present not only as parkinsonism but also as dystonia or choreoathetosis (2, 13). Recently, even a pure cerebellar syndrome has been qualified as a manifestation of AHD (15). Thus, the true prevalence of AHD is likely to be underestimated by the data provided by Kang et al. (12) who considered only patients with parkinsonian symptoms.
There is one other aspect that needs to be discussed: is it sensible to exclude patients with the diagnosis of HE or patients who are treated with lactulose if we are looking for AHD? We all know from the literature and also from personal experience that patients develop AHD usually after they have also experienced episodes of HE, and that progressive extrapyramidal symptoms because of AHD may be episodically accompanied by HE featuring as temporarily clouding of consciousness – for example – and responding to plasma ammonia-lowering therapy (2, 13). Thus, neither a history of HE episodes nor treatment with lactulose would disprove the diagnosis of AHD, especially as the latter must be considered uneffective with regard to the extrapyramidal symptoms if parkinsonism is still diagnosed.
Further questions such as the pathophysiology behind the clinical symptoms, the susceptibility of some patients to develop AHD or the most efficient course of treatment are far from being answered today. Recently, Romero-Gómez et al. (16) suggested a genetically determined difference in the conversion rate of glutamine to ammonia in the peripheral tissues and brain as a potential factor accounting for the variability in cirrhotic patients' predisposition to develop HE. Susceptibility genes are also discussed with regard to the development of manganism (17) and thus genetic factors might well also play a role for a patient's susceptibility to develop AHD. The treatment of AHD remains challenging. The symptoms usually do not respond to the classical HE therapy. Inconsistent results have been reported for the effect of bromocryptine, levodopa or dopa agonists and for the effect of liver transplantation with positive reports predominating (9, 13, 18). Negative results, however, are less probable to be published. Thus, reliable data can be expected only from the prospective assessment, registration and follow-up observation of well-defined patient cohorts using standardized means – an approach that is worthy to be realized by a co-operative task force of HE experts in the future.