Subclinical impairment of brain function in chronic
hepatitis C infection
The authors of this article have been reporting on this research over the past few years at various scientific conferences, and I've been reporting on their data and posting these reports all along the way on the internet and on my website (www.natap.org). These authors are among several who've conducted studies on this question: does HCV lead to neurologic dysfunction. Many of the papers on this research are also archived on my website and can be found by using our search engine on the website. The authors of this paper use objective testing to make a good case that HCV does effect neurologic functioning and can cause cognitive impairment in some HCV-infected individuals. The authors are, however, clear that we don't know whether this is due to HCV entering the brain or CSF or because HCV causes immune dysfunction which in turn affects the brain. This study makes a good case that these findings of HCV effecting cognitive and neurologic functioning in this study is not due to cofactors such as cirrhosis, IVDU, or alcohol. In self-reported SF-36 questionaires HCV-infected patients on average reported reduced physical functioning, role physical, bodily pain, general health, vitality, social functioning, role emotional, and mental health. Patients on average also reported more fatigue. In my experience not all HCV-infected individuals experience these symptoms but I think some do. |
Background/Aims: Central nervous system abnormalities such as fatigue and depression occur more frequently in chronic hepatitis C virus (HCV) infection than in many other causes of chronic liver disease. The finding that fatigue is
unrelated to activity of hepatitis or mode of infection could indicate an independent effect of HCV on brain function. This study tested the hypothesis of a subclinical cognitive dysfunction in HCV-infected patients.
Methods: One-hundred untreated HCV-RNA positive biopsy-proven patients were investigated by P300 event-related potentials, a sensitive electrophysiologic test of cognitive processing. Health-related quality of life and fatigue were assessed using the SF-36 questionnaire and the Fatigue Impact Scale, respectively.
Results: Cognitive brain function was subclinically impaired in the cohort of HCV-infected patients as indicated by significantly prolonged P300 latencies
(P =0.01 for comparison to matched healthy subjects) and reduced P300 ampli-tudes (P <0.001, respectively). Seventeen of the 100 HCV-infected patients had P300 latencies outside the age-adjusted normal range. Abnormal P300 characteristics were not related to the degree of histologic or biochemical activity of hepatitis, severity of fatigue or mental health impairment.
Conclusions: This study demonstrates that patients with HCV infection showed a slight but significant neurocognitive impairment, possibly indicating a further extrahepatic manifestation of chronic hepatitis C.
Journal of Hepatology 37 (2002) 349-354
Ludwig Kramer, Edith Bauer, Georg Funk, Harald Hofer, Wolfgang Jessner, Petra Steindl-Munda, Friedrich Wrba, Christian Madl, Alfred Gangl, Peter Ferenci. Department of Medicine IV, University Hospital Vienna, Wa¨ hringer Gu¨rtel 18-20, A-1090 Vienna, Austria Department of Pathology, University of Vienna Medical School, Wa¨hringer Gu¨ rtel 18-20, A-1090 Vienna, Austria
Symptoms of central nervous system dysfunction such as fatigue and depression are common in patients with chronic hepatitis C and occur more frequently than in chronic hepatitis B contributing to a significant reduction of health-related quality of life (HRQL). Both fatigue and the impairment of HRQL are unrelated to biochemical and histologic activity, or autoimmune features of hepatitis C. HRQL improved in patients who cleared the virus on therapy but remained unchanged in non-responders. Based on these observations it was proposed that the hepatitis C virus (HCV) per se could cause brain dysfunction.
This hypothesis is supported by findings like replication of HCV in the astroglia of patients with AIDS, detection of HCV-RNA in brain tissue, identification of brain-specific quasispecies HCV variants, and by demonstration of metabolic changes resembling HIV-1 encephalopathy in the basal ganglia of patients with chronic hepatitis C. Thus, it is possible that fatigue and depression in chronic hepatitis C could be due to a neuropathogenic effect of the HCV. We tested this hypothesis using P300 event-related potentials, a sensitive neurophysiologic measure of cognitive function. P300 latency correlates with information processing speed, reaction time and age; its amplitude reflects attention devoted to a stimulus. The P300 approach avoids inherent bias of test batteries or impaired self-rating due to awareness of a potentially progressive disease. Subclinical brain dysfunction was detected more sensitively by the P300 method than by electroencephalogram (EEG) and psychometric tests in a
variety of medical conditions including hypoglycaemia and chronic renal failure. Marked P300 abnormalities occur early in HIV-1 encephalopathy and precede overt cognitive dysfunction.
2.1.1. Study population
We studied 100 consecutive patients with untreated chronic hepatitis C
(44±11 years, range, 21-74, 68 males) referred for evaluation of antiviral
therapy trials. Thus, patients with cirrhosis stage Child B/C, alcoholic liver
disease and injection drug use within 6 months prior to study were not
eligible. Current or past alcohol dependency was an exclusion criterion
for any trial. Furthermore, we excluded patients with cryoglobulinaemia,
organ transplantation, extrahepatic infections, renal failure, insulin-depen-dent diabetes and a history of alcohol consumption within the last 4 weeks.
Also patients with neurologic or psychiatric illness, attention and learning
deficits or use of centrally active substances were excluded. When patients
reported a known mode of transmission, disease duration was determined
based on time since transfusion, icteric hepatitis or diagnosis of acute non-A-
non-B hepatitis (before 1990). For individuals reporting prior injection
drug use, we made an estimate of disease duration based on 1st year of
injection drug use. The control population consisted of 100 sex- and age-matched healthy individuals (44±13 years). The study protocol was
approved by the human subject board of the Vienna Medical Faculty and
patients gave written informed consent.
2.2.5. Psychometric assessment
The Mini-mental state examination, a bedside test of neurocognitive
function that can be performed within 10 min was used for global assess-ment
of cognitive and psychomotor function.
2.2.6. Assessment of health-related quality of life
The SF-36 (QualityMetric, Inc., Lincoln, RI) is a self-report, generic
HRQL measure including eight multi-item scales (36 items) that evaluate
the extent to which an individual's health limits his or her physical,
emotional, and social functioning. The SF-36 measures eight domains
of HRQL: physical functioning, role physical, bodily pain, general health,
vitality, social functioning, role emotional, and mental health and is scored
from 0 to 100, with higher scores indicating better HRQL. Summary scores
for physical and mental aspects of HRQL can be calculated. Reliability and
validity of the SF-36 are well-established.
2.2.7. Assessment of fatigue
The fatigue impact scale (FIS) was developed to assess functional limita-tions
due to chronic fatigue. In this test, 40 statements within three subscales (physical, cognitive and psychosocial) are presented in a random
order to be scored from 0 (no problem') to 4 (extreme problem'), with
scores between 0 and 160. The FIS has been validated in HCV infection and
was found to be internally consistent and reproducible.
3.1.1. Patients characteristics
Histology showed chronic hepatitis in 75 patients (75%) and cirrhosis in 25 patients (25%). All cirrhotic patients were in Child A clinical stage; no patient had ascites. None of the patients had overt cognitive dysfunction as indicated by normal scores in the mini mental state examination (median, 30; range 26-30). Data on questionnaires could not be analysed in ten patients: five declined to complete the questionnaire and in the other five, parts of the questionnaire were left unanswered. HCV genotypes were available in 90 patients: 72 patients had genotype 1; three had genotype 2; 21 had genotype 3a; and four patients had genotype 4. Twelve patients were coinfected by two different genotypes or subtypes.
3.1.2. Event-related potentials
Compared to matched controls, patients with chronic HCV infection had delayed P300 peak latencies (P=0.01) and reduced amplitudes (P<0.001).
Seventeen of the 100 HCV-infected patients had P300 latencies
outside the age-adjusted normal range. (edit note: that means that test results find these patients to have slower information processing speed, reaction time. There was average age difference of 51 vs 43 years, is this enough age difference to be significant?). The difference between patients and healthy control subjects was similar in central (Cz, International 10/20 classification) and frontal brain areas (Fz). In addition, the typical correlation of P300 latency and age was significantly (P=0.008) less pronounced in patients (r=0.35; P=0.001) than in matched controls, ( r=0.65, P<0.001) indicating
interference by factors other than age. There was no correlation between cognitive dysfunction and histologic or biochemical features of hepatitis C activity.
Patients with chronic HCV infection scored significantly below population-derived normative values for most SF-36 domains. There were only weak negative correlations of P300 latency and the SF-36 scores for physical
function (R= -0.27, P=0.01), general health (R= -0.31, P=0.003) and summary scores for physical impairment (R= -0.28, P=0.02). No other correlation
between HRQL and neurophysiologic function was detected.
The FIS in patients (median 48, range 1-124) was higher than in matched controls (median 23, range 0-61, P<0.001). Again there was no correlation to cognitive brain function in patients and controls. However, a clear negative correlation between FIS and SF-36 physical and mental health summary scales was observed (r= -0.49 and -0.74, P<0.001, respectively).
3.2. Subgroup analysis
3.2.1. Injection drug use - related HCV infection
Patients with a history of injection drug use (n=23) were younger (35±8 versus 48±11 years, P <0.001). Duration of disease (3.6 years) was similar to that of other patients (4.6 years, P=0.77). More patients with a history of injection drug use were anti-HBc positive (54 versus 28%, P=0.05) and had HCV subtype 3a (42 versus 13%, respectively; P=0.04). The incidence of cirrhosis was similar (21 versus 25%, P=0.96). Patients with a history of injection drug use had slightly earlier P300 latencies (345 versus 360 ms, P=0.06) and higher amplitudes (7.7 versus 7.0 mV, P=0.04); while HRQL and fatigue scores were not different (data not shown).
Cirrhotic patients were older (52±10 versus 41±11 years, P<0.001) and had higher mean aspartate amino-transferase (AST) levels (53 versus 32 U/l, P=0.004). Platelet count, white blood count and serum cholinesterase
but not albumin concentrations were significantly lower in cirrhotic patients (data not shown). There was, however, no difference in cognitive brain function (mean P300 latency, 364 in cirrhotic versus 355 ms in non-cirrhotic patients, P=0.26; P300 amplitude: 12.5 versus 12.4 mV, P=0.96, respectively). Moreover, also P300 results of the 37 patients with stages 3-4 fibrosis were similar compared to patients with stages 0-2 fibrosis (356 versus
355 ms, P=0.95; 14.1 versus 12.3 mV, P=0.29, respectively). We also found no appreciable impact of cirrhosis on fatigue severity (FIS, 44±30 in cirrhotic versus 50±34 in non-cirrhotic patients, P=0.50) and most domains of the
SF-36 (data not shown); with only general health' being more severely impaired in cirrhotic patients (49±21 versus 60±22, P=0:04).
Out of the group of HCV-positive patients, 29% denied any current or past alcohol consumption. Thus 71% of patients had ever consumed alcohol. Within this group, severity of previous alcohol consumption was reported to be minimal in 38%, occasionally in 56%, and moderate or substantial in 3% each, respectively; with no statistically significant difference in P300 parameters between patients with or without a history of alcohol consumption and those
with minimal or occasional compared to moderate or substantial alcohol consumption, respectively. Median time between last drink and inclusion in HCV treatment studies was 12 months (range 1-144). Patients with prolonged P300 latencies were not different compared to other patients in mean values of y glutamyl transpeptidase (51 versus 54 U/l, P=0.84), AST (37 versus 37 U/l,
P=0.94) and mean corpuscular volume (86 versus 84 fl, P=0.054).
3.2.4. Subclinical cognitive impairment
Patients with abnormal P300 latency did not differ from other patients with respect to severity of fatigue (mean FIS score, 49 versus 48, P=0.84), presence of cirrhosis, HCV genotypes, viral load and activity of hepatitis. Patients with abnormal P300 latency were older (51 versus 43 years, P=0.01 and tended to have a more prolonged duration of HCV infection (7.2 versus 3.8 years, P=0.22) Among the measures of HRQL, physical function, bodily pain and general health were more severely impaired in the patients with cognitive dysfunction.
To compare cognitive dysfunction in chronic hepatitis C with that of other chronic diseases, mean P300 latencies and amplitudes were calculated from age-matched patients withinsulin-dependent diabetes or dialysis-dependent renal failure. P300 latencies of HCV-infected patients were similar to those of matched diabetic (n=61; 352±21 ms; p=0.44) and uraemic patients (n=15; 36 434 ms, P=0.43, respectively). Compared to diabetic patients, P300 amplitudes were significantly reduced in patients with chronic hepatitis C infection (P ,0:001) indicating impaired attention similar to uraemic patients (15.2±5.6
mV, P=0.30, respectively).
This study demonstrates that patients with HCV infection showed a slight but significant neurocognitive impairment, possibly indicating a further extrahepatic manifestation of chronic hepatitis C. Overall, one in six patients with HCV infection exhibited grossly abnormal P300 latencies. Among other reasons, we employed the P300 method for its relative independence from socio-cultural background, language limitations and level of education (in which it clearly differs from psychometric tests). Subclinical cognitive
dysfunction was unrelated to the degree of fibrosis, inflammatory activity, virus genotype, viral load, or a history of injection drug use but correlated weakly with physical and general health impairment. Moreover, it cannot
be excluded that increased age and HCV infection contribute synergistically to subclinical cognitive dysfunction.
The hypothesis that HCV could affect brain function was first raised by Foster et al., who observed that HCV-infectedpatients were more severely fatigued and scored worse than matched controls or otherwise similar patients with chronic hepatitis B on HRQL indices. The same group has recently reported a high prevalence of neurocognitive abnormalities in HCV infection. Impaired HRQL scores improved with successful antiviral treatment but remained
unchanged in non-responders. Thus it seems possible that neurocognitive dysfunction indicates a cerebral effect of HCV. Whether cerebral dysfunction reflects a direct action of the HCV or develops secondary to non-specific conse-quences of a chronic inflammatory disease awaits further study. Replication in mononuclear cells and neutrophils could theoretically enable the HCV to enter the brain. HCV-RNA was detected in microglia and astrocytes of patients infected with HIV. Unfortunately, no data is available on the neuropathology of HCV infection and it is currently unclear whether brain-specific quasispecies HCV variants, central effects of cytokines or other factors contribute to cerebral dysfunction. Magnetic resonance spectroscopy (MRS) yielded higher basal ganglia and white matter choline/creatine ratios in patients with mild chronic HCV infection compared to chronic hepatitis B infection and
healthy controls. Similar abnormalities have been extensively documented in HIV encephalopathy. Nonetheless, no clear correlation exists between MRS abnormalities and cognitive dysfunction. Confirming previous studies, HCV-infected patients had lower HRQL scores and were more severely fatigued than
controls, although the absence of a validation study for the German FIS is a limitation. The pathophysiologic mechanisms leading to fatigue as the most frequent symptom of chronic HCV infection remain unknown: Analysis of cogni-tive tasks in HCV infection yielded a preferential impairment of concentration and memory processes, suggesting neurocognitive disturbance. Profound fatigue in hepatitis C could involve a disturbance of central serotoninergic
neurotransmission since was relieved by the 5-HT3 antagonist ondansetrone. The weak association of cognitive dysfunction and physical health impairment observed in our study could also indicate a link of cerebral dysfunction and
impaired physical function. Accordingly, fatigue severity was inversely related to hand grip strength in patients with primary biliary cirrhosis. In fatigue, neurophysiological disturbances are considered to affect brainstem, hippocam-pal and subthalamic brain areas, which are also involved in the process of P300 generation. Thus, a postviral fatigue syndrome was characterised by impairment of attention, memory and stimulus evaluation on P300 examination. In the current study, cerebral dysfunction was largely unrelated
to fatigue, suggesting that other yet unidentified factors could affect brain function in HCV infection.
In this context, a comparison to chronic hepatitis B infection would be interesting. Unfortunately, we were unable to recruit a suitable control group due to the low incidence of hepatitis B infection in Austria. It is unlikely that neurophysiologic dysfunction in patients with chronic HCV infection represented minimal hepatic encephalopathy. First, patients with advanced
cirrhosis at potential risk were excluded. Second, we found no correlation of neurocognitive function with biochemical indices of liver function. There was also no difference between cirrhotic and non-cirrhotic patients with respect to cognitive function, most HRQL domains and fatigue severity. Given rigorous exclusion criteria and the lack of a difference in GGT, AST or mean corpuscular volume, alcohol is unlikely as a cause of cognitive dysfunction. Similar to the study by Forton et al., our cohort of HCV-infected patients showed no appreciable impact of previous intravenous drug abuse on cognitive function. According to their younger age, patients with a history of injection drug
use had even better P300 findings than those without such a history. We were also unable to demonstrate a meaningful difference in HRQL parameters between groups. Thus, our data do not support the hypothesis that previous drug abuse caused cognitive deficits in HCV infection. Several studies have focused on the assessment of HRQL characteristics in HCV infection. In the current study, patients with subclinical brain dysfunction scored worse than
other patients in physical function, general health and bodily pain, although hepatic function was not different. It is very difficult to interpret HRQL impairment as direct evidence of cerebral involvement by HCV: While the
SF-36 can provide a basis for comparison with other patient groups or normative general population data, limitations of a self-reported score in patients with a chronic and potentially progressive disease have to be kept in mind. For obvious reasons, patients are not blinded for their disease status and the reported improvement of SF-36 scores following successful antiviral treatment could be partially explained by subjective awareness of health improvement. Accordingly, former injection drug users aware of HCV seropositivity scored lower for some of the SF-36 domains than otherwise
comparable injection drug users unaware of their HCV status. Conversely, HCV infection is frequently associated with depressive symptoms, which could partially account for impaired HRQL and fatigue scores but do not explain P300
In summary, chronic hepatitis C is associated with a subclinical neurophysiologic impairment that does not correlate with severity of liver disease or a history of injection drug use. These results could therefore indicate a direct or indirect action of the HCV on the brain. However, further
data including neurophysiologic follow-up studies of successfully treated patients with HCV infection are required to prove the causality of such a connection.