In the current study, the etiology of liver cirrhosis was not related to HE or health-related QoL. Diabetes mellitus was more common among patients with hepatocellular cirrhosis compared to those with cholestatic cirrhosis, in accordance to published data [7
], and neither cirrhosis severity nor the prevalence of HE differed between the two groups. Furthermore, the prevalence of HE did not differ among the groups of cirrhosis due to alcohol, chronic hepatitis C, or cholestatic liver disease. To our knowledge, this is the first study to investigate the impact of etiology of cirrhosis on minimal and overt HE, as assessed according to currently accepted criteria [16
], with simultaneous assessment of diabetes mellitus and measurement of plasma ammonia levels.
These findings are in line with previous reports on the relation of cognitive impairment with the etiology of liver disease [4
]. Despite initial reports that cognitive impairment detected by means of neuropsychological testing varied according to the type -hepatocellular vs. cholestatic- of cirrhosis [3
], a subsequent study demonstrated that cerebral computed tomography scan abnormalities were similarly common in both hepatocellular and cholestatic liver disease [4
]. A recent investigation showed that cognitive dysfunction was frequent in PBC patients and unrelated to the severity of liver disease [12
]. However, few of the patients included in this study had frank cirrhosis and patients with PBC were not compared with controls with another chronic liver disease [12
]. Therefore, these findings may not be extrapolated to cirrhotic populations due to PBC or PSC.
In previous imaging studies cerebral abnormalities have been detected with computed tomography in non-alcoholic cirrhotic patients [29
] as well as with magnetic resonance imaging in cirrhotic patients irrespective of liver disease etiology [30
]. Although cirrhotic alcoholics have been reported to exhibit more gross pathology than non-alcoholic cirrhosis on ratings of cerebral atrophy assessed by computed tomography in another imaging study, the authors noted that the two groups were more similar to each other than they were different on planimetric measurements [31
]. Cognitive impairment has been observed in non-cirrhotic patients with alcoholism [8
] and with chronic hepatitis C [10
]. However, in some hepatitis C cohorts cognitive decline has not been shown to differ from that of patients with chronic liver disease of other etiologies [10
]. Liver disease severity is thought to be the main determinant of cognitive dysfunction in alcoholic cirrhosis [13
] although some controversy exists [33
]. In a recent study in patients with cirrhosis without overt HE (according to the West Haven criteria), minimal HE was proposed to be more common in cirrhotic patients with hepatitis C [34
]. However, this difference was only observed in univariate analysis of data and was not tested in multivariate analysis [34
]. Apart from this report data comparing cognitive function of patients with hepatitis C cirrhosis with that of patients with cirrhosis due to other etiologies are largely lacking. Our findings support previously published data that alcoholic etiology is not a major determinant of cognitive dysfunction in cirrhotic patients [13
] and further indicate that neither hepatitis C nor cholestatic liver disease have a major effect on hepatic encephalopathy in cirrhosis.
Diabetes mellitus has been shown to be associated to HE in hepatitis C cirrhosis [5
]. Insulin resistance, which is common in cirrhotic patients, is also related to plasma ammonia levels [6
]. We did observe an independent relation of diabetes mellitus with the time needed to perform NCT-A, in accordance with a previous report [6
], but diabetes did not affect the prevalence of HE in our patient cohort. The same was true when the potential effect of diabetes on HE was analyzed separately in the subgroups of hepatocellular or cholestatic cirrhosis. Although our study was not designed to assess the impact of diabetes on HE our results indicate that factors other than diabetes mellitus might be more important in determining HE in cirrhosis of various etiologies.
Most previous studies have not detected any difference in health-related QoL indices among patients with liver cirrhosis of different etiologies [17
]. Younossi et al compared health-related QoL in patients with hepatocellular and cholestatic cirrhosis and found that physical dimensions of QoL were less impaired in patients with cholestatic disease than in those with hepatocellular disease [18
]. We found that health-related QoL was related to hepatic encephalopathy in line with previously published data [17
] but we did not observe any significant differences in any QoL dimension between hepatocellular and cholestatic cirrhosis. These apparently different findings may be explained by the different patient groups recruited in the two studies. In the study of Younossi et al the group of hepatocellular cirrhosis consisted mainly of patients with alcoholic and cryptogenic cirrhosis as well as patients with autoimmune hepatitis [18
]. On the other hand, patients with cryptogenic cirrhosis were excluded from the current investigation and 39/127 patients with hepatocellular disease had viral cirrhosis. Furthermore, in the previous study the proportion of patients with Child-Pugh C cirrhosis was 3.3% in the cholestatic and 26.6% in the hepatocellular group [18
] whereas there were not any significant differences in cirrhosis severity between the two groups in the current study. Our findings are in accordance with previously published data indicating HE and cirrhosis severity are important determinants of QoL in cirrhosis whereas the etiology of liver disease does not seem to play a major role [17
Certain methodological limitations of the currents study should be taken into consideration. First, HE was evaluated by means of clinical assessment and psychometric tests but no quantitative neurophysiologic tools (such as electroencephalography) were used. Although neurophysiologic tools are often used in research studies there is no clear consensus as to the validity of these tests when used alone or in combination [35
]. In the current study, the guidelines of the 11th
World Congress of Gastroenterology were followed in the evaluation of HE [16
]; clinical assessment and simple bedside psychometric tests were chosen as they are easily applied in everyday practice. Second, although in the current study no major differences were observed in demographic variables or in severity of liver disease among groups of patients with different etiologies, ideally patients in the different groups should be matched for age, gender, and severity of liver disease. Third, the current study is a cross-sectional one. Thus, a cause-effect relationship between the measured variables would be hard to establish. Last, although our investigation is the largest, to date, exploring the relation of the type of liver cirrhosis (cholestatic vs. hepatocellular) with the presence of hepatic encephalopathy and although all patients with cholestatic cirrhosis under our care were asked to participate, a type-II error cannot be excluded. Further multicenter studies might be necessary to fully delineate the potential role of the type and etiology of cirrhosis in hepatic encephalopathy.