Hepatitis C and G Virus Infection and Liver Dysfunction After Allogeneic Bone Marrow Transplantation: Results From a Prospective Study

LIVER DYSFUNCTION is an important cause of morbidity and mortality after allogeneic bone marrow transplantation.1 Among the causes of liver damage are acute and chronic graft-versus-host disease (GVHD), toxic liver cell damage due to chemotherapy, including venoocclusive liver disease (VOD), and viral infections2-4 (hepatitis C virus [HCV]).

Two hepatitis-associated viruses termed GB virus C (GBV-C) or hepatitis G virus (HGV)6 have recently been isolated. Both viruses share a 95% nucleotide homology demonstrating that GBV-C and HGV are different isolates of the same virus. This new agent is parenterally transmitted,7-9 its prevalence is from 1% to 2% in voluntary donors from the United States, and it is more frequent than HCV in the general population.10 Although there has been rapid increase in the molecular and epidemiologic knowledge of HGV, its clinical relevance remains largely unresolved. Patients who undergo bone marrow transplantation (BMT) are exposed to a large number of blood transfusions and suffer from a profound immunosuppression. Accordingly, the hypothesis that this new agent may infect this patient population and play a role in post-BMT liver abnormalities seems plausible. Thus, the aim of this study was to analyze the prevalence of both HCV and HGV infections, as well as their influence on liver function abnormalities after allogeneic BMT.

We prospectively studied 59 patients who had undergone an allogeneic BMT from January 1994 to December 1995. The clinical and technical characteristics are shown in Table 1. The median follow-up period for surviving patients has been 502 days (range, 184 to 822 days). The aim of the study was to determine the possible role of hepatitis C and G viruses in both acute and chronic liver dysfunction after allogeneic BMT. This study was approved by the Ethics Committees of our institutions.

The methods used in allogeneic BMT have previously been reported.11 All of the patients were hospitalized in single rooms ventilated through high-efficiency particular air (HEPA) filtration systems. Trimethoprim/sulfamethoxazole was administered to all patients from day −7 to day −1 and given again twice per week during engraftment for 6 months after transplant. All patients received intravenous parenteral nutrition, which started on day +1 and continued until oral tolerance was achieved. Broad spectrum antibiotics and amphotericin-B were used as previously described.11,12 In recipients seropositive for cytomegalovirus (CMV) or in those whose donors were seropositive for CMV, high-dose acyclovir (500 mg/m²/intravenously [IV]/8 hours) was administered from day −5 to day +30. None of the patients received growth factors or intravenous gammaglobulin.

Busulfan (16 mg/kg)-cyclophosphamide (120 mg/kg) (BuCy2) and Cy (120 mg/kg) with fractioned total body irradiation (TBI) (1,200 cGy divided in six doses administered on days −3, −2, and −1) (CyTBI) were the conditioning regimens more commonly employed. Irradiation was delivered by a ⁶⁰Co irradiation source at a dose rate of 4 to 7 cGy/min. Patients with severe aplastic anaemia were treated with either Cy and total lymphoid irradiation or anti-thymocyte globulin (ATG). VP-16 (30 mg/kg) was added to BuCy2 in two patients and the patient with multiple myeloma received TBI (900 Gy), cyclophosphamide (120 mg/kg), and thiotepa (500 mg/m²).

All patients received non-T–cell-depleted bone marrows, which were infused on day 0. A median of 3.4 × 10⁸ nucleated bone marrow cells per kilogram of receiving body weight was infused (range, 1.8 to 5.2). Cyclosporine (CsA) and a short course of methotrexate according to the Seattle protocol for GVHD prophylaxis13 was given in all cases. After day +100, the CsA dose was decreased by 5% per week and discontinued 6 months after transplantation in patients with no evidence of GVHD. Acute GVHD was classified according to Deeg's grading14 and chronic GVHD was graded as either limited or extensive according to classical criteria.¹².

Early liver dysfunction was defined as either an increase in alanine aminotransferase (ALT) levels above the normal value or a bilirubin level greater than 2 mg/dL. Venoocclusive disease (VOD) was diagnosed according to clinical findings when the three major criteria (weight gain, jaundice, and hepatomegaly) were present.15 Patients without acute GVHD, VOD criteria, or other identified causes of liver dysfunction were classified as having acute liver toxicity related to the preparatory regimen and the transplantation procedure. For patients who survived longer than 180 days after BMT, the presence of liver dysfunction was defined as an increase in ALT levels above the normal value that persisted for more than 3 months.

Hepatitis-B virus surface antigen (HBsAg), as well as antibodies against HBsAg (anti-HBs) and hepatitis B virus core antigen (Anti-HBc), were tested by commercially available enzyme immunoassays (Abbott, North Chicago, IL).

Antibodies against HCV (anti-HCV) were detected by first- and second-generation enzyme-linked immunosorbent assay (ELISA) tests (Ortho, Raritan, NJ). Serum HCV-RNA was determined by reverse transcription and nested polymerase chain reaction (PCR) (RT-PCR), using primers from the highly conserved 5′noncoding region of the HCV genome.16 HCV-RNA quantitation was performed using the Amplicor HCV Monitor test kit (Roche Diagnostic System, Brandiburg, NJ). HCV-RNA genotyping was performed as described by Davidson et al.17 

For the detection of HGV RNA, total RNA from serum samples was isolated as described.18 The total RNA thus obtained was amplified by RT-PCR, using specific primers of the putative helicase/NS3 gene designed according to the HGV sequence reported by Simons et al,5 which amplifies a 272-bp fragment of the HGV genome, as well as primers from the 5′ noncoding region, according to the sequence reported by Leary et al,19 which amplifies a 320-bp fragment. The antisense oligonucleotides AMS3 (NS3 region) and S1 (5′noncoding region) were used to prime cDNA synthesis of the genomic HGV strand. After an RNA denaturation step (5 minutes at 95°C), AMV-RT was added, and reverse transcription was performed at 42°C for 30 minutes. Finally, avian myeloblastosis virus reverse-transcriptase (AMV-RT) was inactivated by heating at 95°C for 15 minutes.

For NS3 amplification, the first 35-cycle PCR rounds were performed using the AMS1 primer at 94°C for 1.5 minutes, 55°C for 1.5 minutes, and 72°C for 1 minute, with a final extension step at 72°C for 7 minutes. The second PCR round was performed under the same conditions, with 10% of the first PCR product, using the AMS2 and AMS4 primers. For the 5′noncoding region amplification, the S1 primer was used in the first PCR. The conditions for the first 35 PCR cycles were 95°C, 1 minute, 50°C, 1 minute, and 72°C, 1 minute. For the second PCR, 10% of the first PCR product was amplified using the same conditions as for the first PCR round and the primers S2 y A2 (sequence and position of the primers used is shown in Table 2).

Tests for HCV (serology and HCV-RNA detection) and HGV were performed on the day of admission for BMT, as well as on days +30, +180, and +360 in all enrolled patients.

A liver biopsy was scheduled for all patients with persistent liver dysfunction or without a clear clinical diagnosis. A total of 24 liver biopsies was obtained from 23 patients. All histologic samples were reviewed by two independent pathologists.

Values are expressed as either medians and intervals or percentages. For qualitative variables, both the χ² test and the Fischer exact test were used, while the Student's t-test was used for quantitative variables. All of the analyses were performed through SPSS v 6.0 (Chicago, IL) statistical packages on a Macintosh computer.

Among the 59 patients enrolled in this study, HGV-RNA alone, HCV-RNA alone, and both viral genomes were detected in 12 (20%), 5 (8.5%), and 6 patients (10%), respectively, before BMT. Nine of the 11 patients with HCV-RNA had anti-HCV antibodies. The HCV genotype was 1b in all cases. HBsAg was not found in any patient, although 17 had antibodies against HBV (anti-HBs or anti-HBc).

All patients (n = 23) infected with hepatitis viruses before BMT had undergone pre-BMT blood transfusions, while this occurred in only 15 of the 36 (41.6%) patients with no evidence of viral infection (P < .05). Furthermore, infected patients were exposed to a significantly higher number of donors than noninfected patients (95.7 ± 80.6 v 27.5 ± 39.4, P < .0001).

Five of the 11 patients with HCV-RNA had received transfusions before the introduction of anti-HCV screening, 2 of 11 had been transfused with blood products screened by first-generation ELISA for anti-HCV, and 4 of 11 had received blood tested by second-generation tests.

Abnormally high ALT levels before BMT were more common in patients infected with both viruses (5 of 6, 83%), followed by those infected with HCV alone (2 of 5, 40%), and HGV alone (3 of 12, 25%). In contrast, only 4 of 36 (11%) of the patients not infected with any of the hepatotropic viruses had abnormal baseline ALT levels. However, no statistical differences were observed in ALT levels among the four groups of patients (Table 3).

Fourteen of the 18 patients with HGV-RNA had received blood transfusions screened by second-generation tests for HCV determination.

Of the 36 patients negative for both HCV-RNA and HGV RNA before BMT, seven became positive for HGV RNA and one for HCV-RNA after BMT (the latter patient never developed antibodies or liver dysfunction).

During the 100 days after BMT, ALT levels remained normal in only nine patients receiving chemotherapy or other drugs, irrespective of the presence of either new or past hepatitis virus infection (data not shown). Nine patients developed VOD after BMT, and two of them died. Table 3 shows the pre-BMT clinical features of early liver dysfunction in all of our patients, divided according to virus status. Maximum ALT levels were significantly higher in those patients infected with both types of viruses than in the other groups (P = .01). VOD was also significantly more common in the former group of patients (P = .05). However, four of these six patients were transplanted while having an active hematologic disease; the two remaining cases (chronic myeloid leukemia [CML] and acute lymphoblastic leukemia [ALL]) had undergone a second BMT. At that time, four cases had elevated ALT levels, a very well-known risk factor for VOD. Five of the 59 patients (8.5%) died of hepatic disease in this early period. Two patients infected with HCV and HGV died of severe VOD (both of them had been transplanted while in refractory relapse with high ALT levels of 240 IU and 92 IU and in one case, a second BMT); three other patients died of therapy-resistant grade IV hepatic GHVD.

During this period, liver samples were obtained from 9 patients (by transjugular biopsy in 5 cases, percutaneous liver biopsy in 3, and autopsy in 1). The histologic diagnosis was acute GVHD and VOD in 6 and 3 cases, respectively.

Characteristics of the patients infected with HGV alone.Thirteen of the 19 patients infected with HGV were evaluable for liver function after day 100 (Table 4). Six patients were not evaluable for various reasons: fatal grade IV-acute GVHD (3 patients), multiorgan failure and sepsis (1 patient), and early relapse (2 patients). With a median follow-up of 536 days after BMT (range, 347 to 942 days), 8 of these 13 patients had an increase in ALT levels after BMT. In 4 cases, the diagnosis was chronic GVHD of the liver (proven by liver biopsy in all cases). The other 4 patients were diagnosed by liver biopsy of iron overload (two patients), steatohepatitis (one patient) and undetermined (one patient, not biopsied). Serum HGV-RNA remained detectable in all patients.

Characteristics of the patients infected with HCV alone.Of the six patients infected with HCV alone, two died due to early relapse. Three of the four remaining patients had consistently abnormal ALT values and detectable HCV-RNA. Two patients exhibited a chronic active hepatitis and a moderate cytolysis with iron overload, as shown by liver biopsy. In the other patient, ALT levels reached normal values after venesections. When comparing baseline with last serum samples, the concentration of HCV RNA increased in patients infected with HCV alone (baseline 0.24 × 10⁶ ± 0.4 × 10⁶ genome copies/mL, final 0.37 × 10⁶ ± 0.34 × 10⁶ genome copies/mL) or in those infected with HGV and HCV simultaneously (baseline 0.27 × 10⁶ ± 0.34 × 10⁶ genome copies/mL, final 0.29 × 10⁶ ± 0.39 × 10⁶ genome copies/mL); the differences, however, were not stastistically significant. Genotype 1b remained detectable in all HCV-infected patients throughout the study.

Characteristics of the patients infected with both HCV and HGV.Three of the six patients infected with both types of viruses before BMT died of BMT-related complications (VOD, two patients; acute GVHD, one patient). Another patient relapsed from AML (day +215) and developed a histologically proven acute hepatitis on day +172 with an increase in HCV-RNA concentration (0.049 × 10⁵v 0.065 × 10⁶). The other two patients developed liver function abnormalities after day 100. In one of these patients, who relapsed from CML after the second allogeneic BMT, HGV disappeared from the serum after treatment with interferon (IFN)-α 2a for CML relapse.20 The other surviving patient had a histologically proven chronic GVHD, but the presence of cytolysis and portal infiltration was also noted.

Characteristics of the 28 patients with neither HCV nor HGV virus infection.Eight of 28 patients were not evaluable for liver dysfunction after day 100 due to acute GVHD (n = 4), early relapse (n = 2), fatal hemorrhage (n = 1), and loss to follow up (one patient). Ten of the 20 evaluable cases had developed some type of liver function abnormality after day 100. The diagnosis was chronic GVHD in 6 cases (proven by liver biopsy in 5 cases), iron overload in 3 cases (proven by liver biopsy in 2 cases), and undetermined in 1 case.

ALT levels at the end of follow-up.Table 4 displays the percentage of patients with normal ALT levels in the last clinical evaluation according to their hepatitis virus status.

HGV and HCV infection and engraftment.There were no statistical differences in terms of neutrophil (>500 and >1,000/μL) or platelet (>20,000/μL >100,000/μL) recovery and HCV and HGV status (data not shown).

In the present study, we have found a high incidence (42%) of HGV infection in a group of patients who had undergone BMT. This percentage is higher than that reported in other risk populations, such as hemophiliacs (18%),6 multitransfused patients (18%),6 or patients on maintenance hemodialysis (3.1%)²¹; however, it is similar to that communicated by Neilson et al22 for long-term survivors of hematologic malignancy (47%). All of the patients who were positive for HGV-RNA before BMT had been transfused before transplant. Seven patients acquired the virus after BMT, which represented 19% of the previously noninfected patients. These data support the hypothesis that HGV is transmitted by blood. On the other hand, 25 of the 29 patients had received blood donations tested by second generation ELISA test for anti-HCV antibodies. This finding shows that anti-HCV is not useful as a surrogate marker of HGV infection.

A high prevalence of HCV infection among this group of patients (20%) was observed. This prevalence is related to the period when the patients were transfused, as 7 patients received multiple transfusions before adequate screening tests for HCV infection were available. However, 4 infected patients had received blood screened by second generation ELISA assays before BMT, and they probably became infected by blood anti-HCV negative, but with HCV-RNA. The fact that 3 of our patients had serum HCV-RNA, but no anti-HCV, suggests that testing for anti-HCV alone may not be useful for the diagnosis of HCV infection in BMT patients, as previously reported.23 In this study, 13 evaluable patients were infected with HGV alone so that the possible effects of this infection on BMT could be studied. Among these 13 patients, 5 had consistently normal ALT levels, which argues against a possible pathogenic role of HGV in chronic liver disease. Although fulminant hepatitis has been associated with HGV,24 we did not find any case of acute or fulminant hepatitis in our patients. Because posttransplant immunosuppression increased viral titers of HCV, a similar effect would be expected with respect to HGV; if so, the development of liver disease due to this agent would be more likely, although that was not the case in our patients. All of these results reinforce the argument that HGV might not be a hepatitis virus at all, as has been proposed recently by other investigators.10,25,26 

Hepatitis G virus has been linked to aplastic anemia.27,28 We examined the effect of HGV infection on engraftment parameters, although no differences were found in terms of neutrophil or platelet recovery among patients with or without HGV infection. Differences were not even noted in the early or late outcome of BMT between patients with or without HGV infection. All of these results demonstrate that although HGV infection is frequent among BMT patients, no clear hematologic or hepatic pathogenicity due to this agent was evident. Nevertheless, long-term follow-up of HGV-infected patients should be performed to confirm the benign prognosis of this infection.

In contrast to HGV infection, patients with HCV had abnormal ALT levels and a histologically proven chronic hepatitis. Therefore, the prognosis of HCV-infected patients would be different from that of HGV-infected patients. The latter patients may thus require antiviral therapy after BMT.

With respect to the HGV and HCV-coinfected patients, the behavior was similar to those who had only HCV infection. This finding demonstrates that HCV has a predominant pathogenic role in patients infected with both viruses.

Finally, the frequency of abnormal ALT levels at the last visit in patients with neither HGV nor HCV infection, was statistically lower (20%) than that found in patients with hepatitis C or hepatitis C and G virus infection (75% to 100%), although it was similar to that in patients with HGV infection alone. These results show again that HGV has no role in causing liver damage in BMT patients.