The Significance of Hepatitis G Virus in Serum of Patients With Sporadic Fulminant and Subfulminant Hepatitis of Unknown Etiology

FULMINANT HEPATITIS (FH) is an uncommon hepatic disorder characterized by the development of hepatic encephalopathy within 8 to 12 weeks from the onset of symptoms of acute liver disease.1 Acute infection with the hepatitis viruses A or B can be identified as the etiologic agent in 5% to 20% of patients with FH.2,3 It is now generally accepted that infection with the hepatitis C virus is an exceedingly rare cause of FH.4-14 Likewise, FH due to acute infection with the hepatitis D or E viruses accounts for only a small number of cases in the United States.7,10,11,13,14 Thus, excluding FH related to acetaminophen or other drug overdose, nearly 40% to 50% of FH cases have no discernible etiology and could possibly be due to a yet-unidentified viral agent. The recent identification and molecular characterization of the hepatitis G virus (HGV) has led to the investigation of a potential role for this agent in the etiology of a variety of chronic and acute cryptogenic hepatic disorders.15-21 Although it is now apparent that HGV is ubiquitous and not etiologically related to chronic liver diseases, there are conflicting data about a putative role of HGV in patients with cryptogenic FH.22-26 The aim of this work was to investigate the presence of circulating HGV in a relatively large number of patients with a sporadic and cryptogenic form of FH from 3 separate geographic locations in the United States.

Thirty-one patients who met the diagnostic criteria for fulminant hepatitis and 5 patients with late onset (subfulminant) hepatitis were studied.1,27,28 The diagnosis of FH of unknown cause (cryptogenic; non–A-E) was established by the absence of the following markers in the patients’ serum: HAV-IgM, HBsAg, HBc-IgM, HCV antibodies (RIBA II), HCV RNA (reverse transcriptase-polymerase chain reaction [RT-PCR] with 3 sets of primers),7HBV DNA (PCR),7 HEV RNA (RT-PCR),7 anti-HEV antibody, antinuclear antibody (ANA), and anti-smooth muscle antibody (SMA). Sera from 19 patients was tested and found negative for serologic markers of remote or past infection with HBV and HAV, by determination of hepatitis B core and surface antibodies and hepatitis A antibodies. Other causes of FH, such as acetaminophen overdose, other toxic ingestion, Budd Chiari syndrome, Wilson’s disease, and Epstein-Barr virus (EBV) and cytomegalovirus (CMV) infections were excluded by history and appropriate biochemical or serologic assays. Serum samples from the 36 patients were stored at −70°C and aliquots were thawed for duplicate testing of HGV RNA. HGV was detected in serum using an RT-PCR assay with primers from the NS5 region of the HGV genome. Briefly, RNA was extracted from 125 μL of serum using RNA isolator (Genosys, Woodlands, TX). After ethanol precipitation, the pellet was dissolved in 25 μL of diethyl pyrocarbonate (DEPC)-treated water containing 10 U/μL of RNAsin (Promega, Madison, WI). RT-PCR was performed using an RNA PCR Core Kit and PCR carry-over prevention kit (Perkin Elmer, Branchburg, NJ) with primers (0.5 mmol/L each) from the NS5 region of the HGV genome: sense, 5′-CTCTTTGTGGTAGCCGAGAGAT-3′ (nucleotides 6904-6928); and antisense, 5′-CGAATGAGTCAGAGGACGGGGTAT-3′ (nucleotides 7004-7036). Forty-five cycles of PCR (94°C for 1 minute, 55°C for 1 minute, and 72°C for 1 minute for each cycle) were performed, followed by an extension reaction at 72°C for 7 minutes. PCR products were analyzed by dot blot hybridization with a ³²P-labeled oligonucleotide probe: 5′-TCGGTTACTGAGAGCAGCTCAGATGAG-3′ (nucleotides 6978-7054). The sensitivity of this assay is 10 HGV RNA copies per reaction, corresponding to approximately 200 copies/mL of serum. Positive results were confirmed by an independent assay. To assess the validity of the HGV RNA assay, the serum samples were tested for HGV RNA without knowledge of the patient diagnoses. Subsequently, 30 serum samples were recoded and reanalyzed for HGV RNA using different primer sets from the NS5 region as provided by the manufacturer (Boehringer Mannheim Gmbh, Mannheim, Germany).29 

In 5 patients, serum samples were tested for HGV RNA before and after they underwent urgent liver transplantation for FH. Serum samples from 5 other patients were tested for HGV RNA at different time points during their disease. HGV RNA was also determined in the serum of 10 patients with FH of known etiology (8 acetaminophen overdose and 2 fulminant Wilson’s disease) and in 10 individuals without liver disease, who served as negative controls. Twenty-two FH patients were from a Mid-Atlantic location, 9 patients from the Southeast United States, and 5 patients from the Northeast United States. Statistical analysis for differences in proportions was performed using the χ² test.

The demographic, clinical, and laboratory characteristics of the 36 patients with sporadic non–A-E FH are presented in Table 1. Twenty-two patients (61%) ultimately died, and 14 (38.8%) patients underwent urgent liver transplantation. HGV RNA was positive by duplicate assays in 14 of 36 patients (38.8%). HGV RNA was found in 1 of 5 (20%) patients from the Northeast, in 1 of 9 (11%) from the Southeast, and in 12 of 22 (54.5%) patients from the Mid-Atlantic region (Table 2). None of the patients from the Northeast and Southeast had received therapeutic blood transfusions before sample collection for HGV RNA, whereas 17 of 22 patients from the Mid-Atlantic region received blood products before testing (Table2). The distribution of patients according to HGV RNA status and blood transfusion is shown in Table 3. The presence of HGV RNA in serum was associated with the use of therapeutic blood transfusion (P = .0203). Of the 12 patients from the Mid-Atlantic region who tested positive for HGV, 10 had received transfusion of blood products (Table 3). Other than blood product transfusion in the 10 patients from the Mid-Atlantic region, none of the patients had risk factors associated with transmission of hepatitis viruses before the onset of FH. Analysis of patients with non–A-E fulminant hepatitis with respect to HGV positivity and transfusions of blood products demonstrated a significant association of the 2 parameters (P < .02; Table 3). Serum from 5 patients was obtained on different dates and tested for HGV RNA; in 2 patients (patients no. 18 and 23), HGV RNA was persistently positive at different times (samples obtained 2 and 4 days apart), whereas in 2 other patients (no. 17 and 34), HGV RNA was consistently absent in all samples (samples also obtained 2 and 4 days apart). In a fifth patient (no. 33), the initial serum sample was negative for HGV RNA but became positive 12 days later after plasmapheresis and perioperative blood transfusions. Five patients with non–A-E FH were tested before and after liver transplantation (LTx). HGV RNA was absent from serum in 2 of these patients (no. 29 and 31) before and after LTx, whereas 2 other patients (no. 26 and 30) demonstrated circulating HGV RNA before and after LTx. The fifth patient (no. 18) had HGV RNA in serum before LTx, but lost it 4 weeks after the transplantation procedure. HGV RNA was also found in 2 patients with FH of known cause and in 1 of 10 individuals without liver disease. Retesting for HGV RNA using a different set of primers in 30 serum samples confirmed the results in all cases but 1 patient (no. 24). Because the second set of primers might span regions with sequence heterogeneity explaining the PCR negativity, this patient was considered HGV RNA positive for the prevalence analysis.

The etiology of fulminant hepatitis remains unclear in many of these patients, because they generally lack evidence of acute infection with known hepatotropic viruses or other known causes of acute hepatic failure.7-11,14,30 The identification of HGV led to the hypothesis that this agent could be involved in the pathogenesis of at least some cases of cryptogenic FH. Several studies with relatively small numbers of patients have addressed this issue31-42; a summary of published reports is shown in Table 4. It is clear that observations are widely diverse among the various preliminary reports. Yoshiba et al33 first reported the presence of HGV sequences in serum of 3 of 6 patients with cryptogenic FH. However, the 3 HGV-positive patients had at least 1 risk factor for acquiring hepatitis viruses before the onset of FH. In a subsequent brief report, these investigators expanded their series to 16 patients with non–A-E FH and detected HGV RNA in 6 (37.5%) of these patients.38 In this report, only 1 of the patients had received therapeutic blood product transfusion after the onset of FH but before obtaining the serum sample for HGV RNA. This led the investigators to conclude that the presence of HGV RNA was not the result of therapeutic transfusions in most of their cases.38 However, the uncertain reliability of historical transfusion data obtained from referring hospitals may have resulted in underestimation of the true frequency of therapeutic transfusion in these patients. There are 4 full publications to date on the prevalence of serum HGV RNA in patients with cryptogenic FH, describing German, Australian, Japanese, and Spanish patients, respectively.22-25 Heringlake et al25 found that HGV RNA was present in the serum of 50% of 10 German patients with non–A-E FH. Although serum samples in this study were obtained on admission and therefore presumably before transfusion of blood products, no specific information was provided regarding risk factors in these patients or the history of transfusion administered at the referring hospitals. Thus, pre-existing HGV infection or passive acquisition via blood product transfusions could explain the high frequency of HGV RNA positivity in this study as well. In contrast, Moaven et al24 found no trace of HGV RNA in the serum of 14 Australian patients with fulminant FH at the time of admission; most of them became positive after liver transplantation, suggesting acquisition of HGV via perioperative blood transfusion, but no information on preoperative risk factors, including transfusion history, was provided in this report. In the study by Tanaka et al,23 only 2 of 11 patients with cryptogenic FH tested positive for serum HGV RNA, and both patients had previously received blood or plasma transfusion. Saiz et al22 recently reported the presence of HGV RNA in serum of 2 of 19 Spanish patients with idiopathic FH, but no information was provided as to previous transfusion history in this cohort of patients.

HGV RNA has been detected in up to 2% of blood donors; therefore, there is a high probability of transmission from the large amounts of fresh frozen plasma used to treat the severe coagulopathy of FH. Our study included 2 cohorts of patients with sporadic FH that differed precisely in this aspect. Our observations indicate that positivity for HGV RNA in FH was strongly associated with the use of therapeutic blood product transfusion. HGV RNA was rarely found in patients without this risk factor. The presence of HGV RNA in the serum of patients with non–A-E FH does not necessarily imply a causative role for HGV. The critical question is whether HGV infection preceded the fulminant hepatitis or occurred as a result of whole blood, plasma, or platelet transfusion in these critically ill patients. Given the unpredictability of the onset of FH, no serum samples were generally available before the onset of FH to distinguish between the 2 possibilities. It is conceivable that some of our HGV-positive patients are chronic HGV carriers who then developed fulminant hepatitis from another cause. Our observation of a consistent pattern of serum HGV RNA in patients tested on different dates supports the concept of active viremia in these patients and a lack of causation of the FH in those persistently negative for HGV RNA.

Our study is the first to examine the prevalence of HGV in a large American cohort of patients with cryptogenic FH and provided an opportunity to directly assess the role of blood product transfusion in the significance of HGV positivity. The frequency of HGV RNA in serum of transfused patients was much higher than those who did not receive transfusion and is thus in agreement with the results from the reports by Moaven et al,24 Tanaka et al,23 and Saiz et al.22 One limitation of our study is the lack of determination of HGV RNA and HBV DNA in liver tissue, because previous reports suggested that a few patients with non–A-E FH may actually have HBV DNA in hepatic tissue.9,30 Only 1 report is available on the presence of HGV RNA in liver tissue obtained at the time of transplantation for seronegative FH.32 HGV RNA was not detected in liver tissue in any of these patients.

Our observations, taken in the context of the reports summarized in Table 4, strongly suggest that HGV RNA is not causally related to non–A-E FH and that the finding of HGV RNA in patients with non–A-E FH is most likely the result of blood product transfusions administered after the onset of fulminant hepatitis.