PATIENTS

Patients with liver disease of unknown cause were identified among 1075 consecutive patients who underwent liver biopsy in our liver unit between 1991 and 1997. They were identified as cases with a persistent increase in alanine aminotransferase (ALT) (>1.5 times the upper limit of normal on at least two different occasions over a minimum period of six months) and who met the following criteria: (1) absence of HBsAg and anti-HCV antibodies in serum; (2) ferritin, alpha₁antitrypsin, and ceruloplasmin levels within the normal range; (3) antinuclear, antimitochondrial, and antismooth muscle antibodies at a titre of ⩽1/80; (4) ethanol intake less than 80 g/day; (5) absence of treatment with potentially hepatotoxic drugs; (6) negative human immunodeficiency virus (HIV) serology; (7) absence of decompensated diabetes, thyroid dysfunction, morbid obesity (defined as body mass index ⩾35 kg/m²) or other systemic diseases which might affect the liver, including severe hyperlipidaemia (cholesterol or triglyceride serum levels greater than 350 mg/dl); and (7) absence of focal intrahepatic lesions or biliary abnormalities at ultrasonography.

Epidemiological data (previous history of acute hepatitis, transfusion, intravenous drug addiction, acupuncture, tattooing, etc), clinical data (patient age, sex, body mass index (body weight in kg/height in metres)), and biochemical parameters, including glucose, triglycerides, cholesterol, and proteinogram were recorded. Fifty eight consecutive sex and age matched healthy blood donors from the hospital blood bank were used as controls.

LIVER HISTOPATHOLOGY

Liver biopsies were reviewed by an independent pathologist who was blinded to the serological and biochemical patient data. Specimens were fixed in 10% neutral buffered formalin and stained with haematoxylin and eosin, trichrome, orcein, and periodic acid-Schiff with diastase digestion for reticulin and iron. Different histological parameters were evaluated including fibrosis, portal inflammation, piecemeal necrosis, lobular inflammation, lobular necrosis, steatosis, cholestasis, bile duct damage, sinusoidal dilation, and iron deposition. These histological features were graded as absent, sparse-mild, or moderate-severe. Fibrosis was graded on a scale of absent, sparse-mild, or severe bridging-cirrhosis. According to these data, liver biopsies were classified into three groups: (1) chronic hepatitis and cirrhosis according to international criteria13; chronic hepatitis in turn was subclassified into mild (lymphocytic inflammation in more than one third of portal tracts, leaving the parenchymal limiting plate preserved, with minimal intra-acinous changes), moderate (dense inflammatory portal infiltration with piecemeal necrosis involving less than 50% of the circumference of most portal tracts but still restricted to periportal areas), or severe (piecemeal necrosis in the periportal region involving more than 50% of the circumference of most portal tracts, and along fibrous septa); (2) non-alcoholic steatohepatitis (NASH): moderate-gross macrovesicular fatty changes with hepatocyte degeneration and ballooning (with or without Mallory hyaline bodies) or fibrosis in zone 314 ,15; and (3) non-specific/minimal changes: a variety of mild abnormalities including intrahepatocytic cholestasis, steatosis, sinusoidal dilation, and mild lobular inflammation or necrosis.

Immunoperoxidase staining for HBV surface and core proteins was performed in liver samples of all PCR-B positive cases.

SEROLOGY

HBsAg, anti-HBc (IgG), and anti-HBs were tested by standard ELISA kits (Cobas Core, Roche, Madrid). Anti-HCV antibodies were detected using the ELISA 3 test (Ortho-Clinical Diagnostics GmbH, Germany). Anti-HIV antibodies were tested by ELISA (Cobas Core, Roche, Madrid). Antiendomysium antibodies were tested using the indirect fluorescent antibody procedure (MARDX Diagnostics, Inc., California, USA).

PCR ASSAY FOR HBV DNA

DNA was extracted from 200 μl of serum using the High Pure Viral Nucleic Acid Kit (Boehringer Mannheim, Mannheim, Germany). Every extraction and PCR set included 12 samples: seven serum samples to be tested, three water samples as negative controls to detect cross contamination, and two positive controls to establish sensitivity, corresponding to two dilutions of serum from a HBsAg positive patient representing 300 and 30 HBV particles, respectively. Nested PCR was performed using the following as primers: HBV-S1 (nt 222-241), HBV-S2 (nt 757-738), HBV-S3 (324-343), and HBV-S4 (707-688).16Briefly, half of the DNA volume was amplified for 36 cycles (95°C for 20 seconds, 60°C for 15 seconds, and 72°C for 25 seconds) followed by a final extension at 72°C for 10 minutes in a total reaction mixture of 30 μl containing 10 mM Tris HCl, pH 8.3, 1.5 mM MgCl₂, 50 mM KCl, 400 mM dNTP, 1.5 U ofTaq polymerase (Boehringer Mannheim), and 0.15 mM of each outer primer HBV-S1 and HBV-S2. The first PCR product (1 μl) was mixed with 9 μl of the reaction mixture (as described above) containing each of the inner primers HBV-S3 and HBV-S4. A second round of PCR was performed with the same protocol and the 10 μl of reaction mixture were loaded on a 2% agarose gel for electrophoresis. All precautions for avoiding PCR contamination were followed stringently. Each positive result was confirmed by a second independent determination. DNA from liver was extracted using the DNAzol Reagent (Life Technologies, Gibco BRL, Gaithersburg, USA). DNA (1 μg) was used for amplification with the PCR protocol described above. The sensitivity of the assay was 10³ genome eq/ml.

CLONING AND SEQUENCE ANALYSIS

Positive samples were loaded after amplification on a 2% agarose gel. Bands were excised from the gel and DNA was purified using the Geneclean II kit (Bio 101, Vista, California, USA). Aliquots were cloned into the pCR2.1-TOPO plasmid using the TOPO-TA cloning kit (Invitrogen, Carlsbad, California, USA). At least 10 white colonies were tested for the presence of the insert by PCR. Minipreps were performed using the Qiaprep Spin Miniprep kit (Qiagen, Hilden, Germany). A minimum of five clones were sequenced with M13 primers using the dRhodamine Terminator Cycle Sequencing Ready Reaction Kit in an ABI PRISM 310 Genetic Analyzer (PE Applied Biosystems, Foster City, California, USA). Bidirectional sequencing was performed to confirm mutations. Sequences from each patient were compared using the Clustal program and each representative sequence was compared for homology in the GenBank.

RT-PCR ASSAY FOR DETECTION OF HCV RNA

RNA was extracted from sera and liver using the Ultraspec Reagent (Biotecx Laboratories, Inc., Houston, Texas, USA). RT-PCR was performed using primers from 5′UTR as described previously.17

PCR ASSAY FOR DETECTION OF HGV RNA

RNA was extracted from sera using the Ultraspec Reagent (Biotecx Laboratories, Inc., Houston, Texas, USA). HGV-RNA was detected by RT-PCR using a commercial kit (Hepatitis G Virus-Primer and Capture Probe Set; Boehringer Mannheim).

STATISTICAL ANALYSIS

Fisher's exact test was used to compare proportions and the Student's t test to compare continuous variables. The Mann-Whitney U test was used to compare non-parametric variables in independent samples. All statistical tests were two tailed.

PREVALENCE OF CRYPTOGENIC LIVER DISEASE AMONG PATIENTS WHO UNDERWENT LIVER BIOPSY BECAUSE OF PERSISTENT ALT ELEVATION

We investigated 1075 patients referred for liver biopsy because of persistent (more than six months) elevation of serum transaminases. In 966 cases (89.8%), clinical, biochemical, and serological testing allowed identification of the disease entity or aetiology of the process while in 109 patients (10.1%) the cause of the hepatic lesion was not identified before the liver biopsy was performed, despite an appropriate diagnostic work up (see material and methods). Of the 109 cases of liver disease of unknown cause, serum samples from eight patients were not available. Thus 101 patients (71 male, 30 female, mean age 45.8 (13.5) years) were included in the study.

HISTOPATHOLOGICAL FINDINGS IN CRYPTOGENIC LIVER DISEASE

Liver biopsies from the 101 cases with hypertransaminasaemia of unknown aetiology included in the study were reviewed blindly by the same pathologist. Examination of the liver tissue showed non-specific changes in 33 subjects (32.7%) and NASH in 16 patients (15.8%). All patients with NASH presented moderate to severe steatosis in liver tissue while in the group recorded as non-specific changes, moderate to severe steatosis (simple fatty liver) was observed in four cases and mild steatosis in 17 (other criteria for NASH are given in material and methods). Chronic hepatitis was found in 39 patients (38.6 %), being mild in 11 and moderate/severe in 28. Cirrhosis was present in 13 subjects (12.9%). These diagnoses were coincident with the initial report except that four cases previously recorded as simple fatty liver were included in the NASH group and four cases diagnosed as minimal changes were considered as mild chronic hepatitis after histological review.

PREVALENCE OF HBV-DNA, HCV-RNA, AND HGV-RNA IN CRYPTOGENIC LIVER DISEASE

The prevalence of HBV-DNA and/or HCV-RNA was significantly higher in patients with hepatic disease of unknown cause than in healthy blood donors (26.7% v 3.4%; p<0.001). HBV-DNA and HCV-RNA were detected in the serum of 19 (18.8%) and nine (8.9%) patients with cryptogenic disease, respectively. One healthy blood donor was HBV-DNA positive (1.7%) and another HCV-RNA positive (1.7%). Only one patient and none of the controls showed positivity for both viral genomes (table 1). In the group of 966 cases (representing 89.9% of the total number of 1075 patients studied) in which the aetiology of the liver lesion was established by conventional diagnostic tests, 96 (9.9%) had HBV infection, 548 (56.7%) had HCV infection, and 10 (1.03%) were infected by both HCV and HBV. Therefore, occult HBV infection was more prevalent than occult HCV infection despite the fact that seropositive hepatitis C is more prevalent than HbsAg positive liver disease in our geographic area.

We tested 74 patients for the presence of HGV-RNA: only one patient (1.3%) was found to be positive. This patient was negative for HBV-DNA and HCV-RNA. HGV-RNA was not present in any of the healthy blood donors tested.

In five HBV-DNA positive and two HCV-RNA positive subjects, investigation of these viral sequences in serum was repeated in samples obtained on three or more occasions. In three of the former group and in one of the latter, a positive PCR test was obtained in two or more samples. These data indicate that in our patients with positive HBV-DNA or HCV-RNA, viral load was low and viraemia was detectable only intermittently.

Four patients with cryptogenic cirrhosis underwent orthotopic liver transplantation (two who tested positive for serum HBV-DNA and two who tested negative). In these four cases we studied the presence of HBV-DNA and HCV-RNA in the explanted liver and in post-transplantation serum (1–28 weeks after liver transplantation). HBV-DNA was detected in the explanted liver only in the two cases who were serum HBV-DNA positive before transplantation. One of these patients continued to have detectable HBV-DNA in serum after transplantation. HCV-RNA was negative in both post-transplantation serum and explanted liver in all four cases.

MARKERS OF HEPATITIS VIRUS IN CRYPTOGENIC LIVER DISEASE

The presence of anti-HBV antibodies was tested in 98 patients. As shown in table 2 30.6% patients showed positivity for either anti-HBc or anti-HBs, or both. Fifty per cent of cases with positive serum HBV-DNA and 26.2% with negative HBV-DNA had positive HBV antibodies. These differences were not statistically significant. Interestingly, all HCV-RNA positive cases showed negative anti-HBV antibodies (table2). These results (and the fact that we found positivity for both HBV-DNA and HCV-RNA in only one case) indicate that HBV positive and HCV positive cases represent two serologically differentiated groups. HBs and HBc proteins were investigated in liver biopsies by immunohistochemistry. All cases were found to be negative. Finally, it is worth noting that for all viral markers (HCV-RNA, HBV-DNA, and anti-HBV antibodies), only 52 of 98 cases lacked serological signs of contact with hepatitis viruses.

CLINICAL AND BIOCHEMICAL FINDINGS IN CRYPTOGENIC LIVER DISEASE ACCORDING TO THE PRESENCE OR ABSENCE OF HBV-DNA OR HCV-RNA IN SERUM

We compared clinical and biochemical data from patients with liver disease of unknown cause and who were positive for HBV-DNA or HCV-RNA with those in whom viraemia was not detected. As shown in table3, patients from the first group were significantly older (50.6 (14.3)v 43.9 (12.8) years; p=0.03) and showed clinical signs of decompensated cirrhosis more frequently (25.9%v 5.4%; p=0.007) than patients in whom both HBV-DNA and HCV-RNA were negative. The number of patients with dyslipaemia (elevated serum cholesterol and/or triglyceride levels) was significantly higher in patients without detectable viraemia than in positive cases (p=0.04), probably reflecting a higher proportion of NASH among these patients. Also, ALT levels were significantly higher in patients who tested negative for HBV-DNA or HCV-RNA than in cases with detectable viraemia, but no significant differences in other clinical and biochemical data (including glycaemia, obesity, gamma globulin levels, presence of low titres of autoantibodies, parenteral exposure, or past history of acute hepatitis) were observed. With respect to cases diagnosed as NASH, none showed fasting glycaemia in excess of 130 mg/dl, in only two patients did serum cholesterol exceed 300 mg/dl, and in only three cases were triglyceride levels greater than 300 mg/dl with 332 mg/dl being the highest value (patients with decompensated diabetes, morbid obesity, and severe dyslipaemia were excluded from the category of hypertransaminasaemia of unknown aetiology). Antiendomysium antibodies were tested in 64 patients from the first group and in 10 from the second (cases with negative and positive viraemia, respectively); the test was negative in all cases, indicating a low percentage of occult coeliac disease in our series of cryptogenic liver disease. There were no significant differences in clinical or biochemical parameters between patients with positivity for HBV-DNA and those with positivity for HCV-RNA.

PREVALENCE OF HBV-DNA OR HCV-RNA IN THE DIFFERENT HISTOPATHOLOGICAL FORMS OF LIVER DAMAGE IN PATIENTS WITH CRYPTOGENIC LIVER DISEASE

With regard to the prevalence of HBV-DNA or HVC-RNA in the different histopathological forms of liver damage in patients with cryptogenic liver disease, viraemia was found in 14.3% of patients with NASH or non-specific changes, in 30.7% of those with chronic hepatitis, and in 61.5% of subjects with cirrhosis (fig 1). In patients with non-specific changes or NASH, the prevalence of viraemia was not significantly different from that found in controls (p=0.08). In contrast, in patients with chronic hepatitis or cirrhosis, the percentage of cases with positive HBV-DNA or HCV-RNA was significantly higher than in healthy controls (38.5% v 3.4%; p<0.001) or patients with non-specific changes or NASH (38.5% v 14.3%; p=0.007). Cirrhosis was significantly more prevalent in the group of patients with detectable HBV-DNA or HCV-RNA than in cases with negative viraemia (29.6%v 6.8%; p=0.005) (table 3). No significant differences in liver histopathology were found between subjects with anti-HBV antibodies and those without anti-HBV antibodies in serum.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Percentage of patients with cryptogenic liver disease and healthy controls who had detectable hepatitis B virus-DNA and/or hepatitis C virus-RNA by polymerase chain reaction in serum. NASH, non-alcoholic steatohepatitis.

SEQUENCE OF HBV IN PATIENTS WITH POSITIVE HBV-DNA IN SERUM

The HBV fragment amplified by PCR comprised the immunodominant “a” determinant of HBsAg as well as 383 nucleotides from the polymerase gene. This fragment was cloned and sequenced in 17 of the 19 HBV-DNA positive patients. In all cases a minimum of five clones per sample were analysed.

All HBV sequences corresponded to genotype A serotype adw2 (50% of samples) or to genotype D serotype ayw (50% of samples). No differences were found in clinical, biochemical, or pathological findings in relation to the HBV genotype present in serum.

Sequences in seven patients showed total homology with wild-type sequences published previously. In the remaining 10 patients at least one nucleotide change was found, two of which were associated with deletion of one and six nucleotides, respectively, in some of the clones. In eight of these patients two or three different HBV sequences were found. The observed mutations caused at least one amino acid change in both polymerase and HBsAg genes in nine patients and one amino acid change in the polymerase gene in the remaining patient. In three cases the changes affected the “a” determinant.