Infection with hepatitis G virus among recipients of plasma products

THE LANCET

Early report

Infection with hepatitis G virus among recipients of plasma products L M Jarvis, F Davidson, J P Hanley, P L Yap, C A Ludlam, P Simmonds

Summary Background Hepatitis G virus (HGV or GBV-C) is a newly discovered human flavivirus distantly related to hepatitis C virus (HCV). Little information is available on its natural history or routes of transmission, although it can be transmitted parenterally. We investigated the prevalence of persistent infection of HGV and HCV in patients exposed to non-virus-inactivated pooled blood products associated with transmission of HCV. Methods RNA was extracted from the plasma of 112 patients with haemophilia and 57 with hypogammaglobulinaemia, as well as from 64 different batches of archived coagulation-factor concentrates and immunoglobulins. RNA was reverse transcribed and amplified with primers from the 5' non-coding region of HCV and HGV by a nested polymerase chain reaction (PCR). Viral RNA was quantified by titration of complementary DNA before amplification. Findings Among non-remunerated UK blood donors HGV infection (detected by PCR) was more common than HCV infection (four [3·2%] of 125 compared with 137 [0·076%] of 180 658 in southeast Scotland). Testing of batches of factor VIII and factor IX concentrates prepared without viral inactivation procedures showed high frequencies of contamination with HGV (16 of 17 factor VIII batches positive; six of six factor IX batches positive), with no difference between remunerated and non-renumerated donors. However, among 95 haemophiliacs who had received non-virus-inactivated concentrates, 13 (14%) were positive for HGV compared with 79 (83%) who were positive for HCV. Two of 37 recipients of long-term immunoglobulin replacement therapy were positive for HGV. Virus inactivation of blood products substantially reduced or eliminated contamination by HGV RNA sequences. Interpretation Despite the extremely high level of HGV contamination of non-virus-inactivated blood products, their use was not associated with high rates of persistent infection in recipients. The infectivity of HGV in blood products may be lower than that of HCV, or the virus may be less able to establish persistent infection in humans. Whatever the case, the high prevalence of active HGV infection in the general population remains difficult to explain.

Lancet 1996; 348: 1352–55 Departments of Medical Microbiology (L M Jarvis PhD, P Simmonds MRCPath); and Haematology (J P Hanley MRCP , C A Ludlam FRCP ) University of Edinburgh; and Edinburgh and Southeast Scotland Blood Transfusion Service (F Davidson PhD, P L Yap FRCPath) Edinburgh, UK Correspondence to: Dr L M Jarvis, Department of Medical Microbiology, University of Edinburgh EH8 9AG, UK

1352

Introduction The newly discovered hepatitis G virus (HGV or GBVC)1,2 has an RNA genome, and has a similar genome organisation to hepatitis C virus (HCV) and other flaviviruses.2,3 HGV can cause acute and persistent infection in humans but the clinical significance of such infection remains uncertain. HGV RNA sequences have been detected by polymerase chain reaction (PCR) of the plasma from individuals with acute, fulminant hepatitis.4 The causative role of HGV in this disease has been questioned.5 Active infection with HGV can be detected in a surprisingly high proportion of the general population. For example, among non-renumerated blood donors 1·7% were PCR positive;2 in whom infection was presumably symptomless. The detection of HGV in donated blood has prompted the concern that HGV infection may be transmitted by blood transfusion and to recipients of non-virusinactivated pooled blood products such as factor VIII and factor IX concentrates and immunoglobulins. For this study, we developed a nested PCR to amplify the conserved 5' non-coding region of HGV, a potentially more sensitive assay than those based on NS3 or NS5 primers.2,6 This assay was used to determine the frequency of HGV and HCV infection in blood donors, and to compare the frequency of contamination with these viruses in a range of blood products prepared from donated blood and from commercial donor plasma. To investigate the transmissibility of HGV to recipients, we measured the frequency of infection of HGV in haemophiliacs and long-term users of immunoglobulin replacement treatment.

Methods Factor VIII concentrates derived from plasma of remunerated (commercial) and non-remunerated blood donors were of intermediate purity. Batches of factor VIII prepared before 1985 were not virally inactivated, whereas those prepared after 1985 were inactivated by heat or solvent detergent treatment. Intravenous immunoglobulin prepared from plasma from nonrenumerated donors was derived by cold ethanol fractionation and given intramuscularly. Commercial immunoglobulins were prepared by cold ethanol fractionation, followed by either pH4/pepsin or ion-exchange chromatography and administered intravenously. Factor VIII and factor IX concentrates and immunoglobulins were reconstituted as recommended with sterile water (5, 10, or 20 mL). 5 mL of reconstituted factor concentrate or 2·5 mL of reconstituted intravenous immunoglobulin were made up to 7 mL with RPMI medium and ultracentrifuged, and viral RNA extracted.7 The study population comprised four groups. All 95 patients (89 male and six female) in the first group were haemophiliacs who were anti-HCV positive by Abbott second-generation enzyme immunoassay (A-EIA; Abbott, Weisbaden, Dalkenheim, Germany) and by second-generation recombinant immunoblot assay (RIBA-2; Chiron Corp, Emeryville, CA, USA). The severity of the coagulation disorder was measured by factor VIII

Vol 348 • November 16, 1996

THE LANCET

Donor and product type

Viral inactivation

HGV

HCV

Number positive/ number tested

Geometric mean concentration (copies/mL)

Number positive/ number tested

Geometric mean concentration (copies/mL) ·· ·· 1100 ··

Factor VIII NR NR R R

No Yes No Yes

9/10 1/11 7/7 0/6

162 1·6 21 000 ··

0/10 0/13 7/7 0/6

Factor IX NR NR R

No Yes Yes

6/6 2/4* 0/2

370 16 ··

2/6 0/4 0/2

4 ·· ··

*Two positive batches of factor IX were prothrombin complex concentrates. NR=non-renumerated, volunteer donor. R=renumerated donor.

Table 1: Detection of HGV and HCV RNA in factor VIII and factor IX concentrates concentrations. Of the 95 patients, 40 had severe haemophilia (factor concentrations 聿2%), 25 moderate (concentrations 2–10%), 18 mild (concentrations 肁10%), and 12 had von Willebrand’s disease. All had received blood products including non-virus-inactivated factor concentrates, and 71 of the 95 patients had persistently elevated serum alanine aminotransferase concentrations. 20 patients were anti-HIV positive and two were HBsAg-positive. The second group comprised 17 male patients with severe haemophilia A, who were anti-HCV negative and who had received only virus-inactivated factor concentrates. The third group comprised 57 immunoglobulin recipients with primary immunodeficiency (X-linked agammaglobulinaemia or common variable immunodeficiency) and all had received longterm replacement therapy. The fourth group consisted of 125 consecutive volunteer blood donors, who gave blood in Edinburgh in November, 1995, and who were negative for other virus markers. For testing of immunoglobulin recipients and blood donors identifying features were removed. Initial experiments were designed to evaluate different primers and conditions for PCR of HGV. RNA was extracted from plasma and reverse transcribed with the outer, antisense primer used for PCR.7 Primers originally described to amplify the NS3 region of HGV (NS3·2–s1, NS3·2-a2, GBV-C-s1, and GBV-Cal)6 were combined with internal primers derived from conserved sequences among published1,2 sequences and our own sequences obtained from haemophiliacs and patients with fulminant hepatitis. In the first-round PCR primers GBV-C-s1 and GBVC-a1 at positions 4257 and 4470 in the published sequence PNF21612 were used with the following thermal cycles: 94°C 18 s, 50°C 21 s, and 72°C 90 s for 35 amplification cycles. In the second round a sense primer at position 4269 (5'ATCCCCTTTTATGGGCATGG-3') and an antisense primer at position 4435 (5'-YTCRTTGATGATGGAACTGTC-3') were used with the same temperature and time cycle for the first round of PCR but for only 30 cycles. Buffer and primer concentrations were as described for PCR of the 5' non-coding region of HCV. Conserved nested primers were designed from published sequences from the 5' non-coding region of PNF2161, R10291, and GBV-C.1,2 The primers were located at positions 108 (5'-AGGTGGTGGATGGGTGAT-3'; sense, outer), 134 (5'-TGGTAGGTCGTAAATCCCGGT-3'; sense, inner), 476 (5'-GGRGCTGGGTGGCCYCATGCWT-3'; antisense, inner) and 531 (5'-TGCCACCCGCCCTCACCCGAA-3'; antisense, outer). Amplification was over 25 cycles for both first and second rounds of PCR, with the following times and temperatures: 94°C 18 s, 55°C 21 s, and 72°C 90 s. RNA was extracted from 0·1 mL plasma from each of the 95 patients with haemophilia who had been exposed to noninactivated concentrates of factor VIII and IX.7 Nine of the RNA samples were positive with the NS3 primers, whereas 13 were positive with the 5' non-coding-region primers (including the nine NS3-positive samples). Similarly, of a range of factor VIII and factor IX concentrates 13 were positive with NS3 primers

Vol 348 • November 16, 1996

and 20 with those from the 5' non-coding region (including all the NS3 positives). The 5' non-coding-region primers were selected for the current study because they were more sensitive than NS3 primers. All four of the HGV 5' non-coding-region positive, NS3 negative samples, and eight of the nine PCR positive samples in both regions were further studied by sequence analysis. All patients were infected with variants of HGV with sequences similar but not identical to each other or to the published sequences PNF2161, R10291, and GBV-C. HCV PCR and genotyping by restriction-fragment length polymorphism analysis were done as previously described.7,8 Quantification of HCV and HGV was carried out by titration of complementary DNA (cDNA) before amplification. All necessary precautions were taken to prevent contamination and several positive and negative controls were included with each assay. Estimation of the number of copies of RNA per mL of plasma was calculated on the assumption of a 5% efficiency of reverse transcription, as previously established for HCV.9

Results RNA was extracted from plasma from 125 nonrenumerated blood donors and amplified with primers from the 5'non-coding-region. Circulating HGV RNA sequences were detected in four (3·2%) of these donors at concentrations ranging from 4⫻104 to more than 107 copies of RNA/mL. The rate of active infection with HCV in the same population was 0·076%, based on summary data from blood-donor screening in Scotland in the first months of anti-HCV screening from September, 1991, to February, 199210 (609 of 180 658 donations were repeat reactive on screening for anti-HCV, of which 137 were PCR positive). HGV was frequently detected in pooled blood products prepared from non-renumerated donors (table 1). For example, nine of ten batches of factor VIII concentrate prepared before 1985 were positive for HGV RNA at concentrations ranging from 80 to 8000 copies of RNA/mL. However, HCV was not found in these batches, which confirms the substantially lower frequency of HCV infection than HGV infection among blood donors, even before screening for HCV started. Virus

Donor type

HGV

Geometric mean concentration (copies/mL)

Immunoglobulin preparations NR R

2/12 6/6

3·4 312

Immunoglobulin recipients NR R

2/37 2/20

·· ··

Table 2: Detection of HGV in immunoglobulin and in immunoglobulin recipients

1353

THE LANCET

Group

Treated haemophiliacs Haemophilia A (n=66) Haemophilia B (n=17) von Willebrand’s disease (n=12) Total (n=95) Previously untreated haemophiliacs (n=17)†

Number positive HGV positive*

HCV positive*

10 (15%) 2 (12%) 1 (8%) 13 (14%)

54 (82%) 15 (88%) 10 (83%) 79 (83%)

1 (6%)

0 (0%)

*PCR positive in 5'non-coding region.† Recipients of only virus-inactivated-factor VIII concentrates.

Table 3: Frequency of HGV and HCV infection in haemophiliacs

inactivation measures (heating for 72 h at 80°C) substantially reduced the HGV concentration of factor VIII concentrate. Only one of 11 batches of factor VIII prepared after 1990 was positive for HGV, with a concentration 100 times lower than the mean of the preinactivation samples. Factor IX manufactured before 1985 was contaminated with both HCV (two of six batches) and HGV (six of six batches). HGV RNA was detected in two batches of factor IX (prothrombin complex concentrate) that had been heat treated (72 h 80°C), but not in high-purity factor IX concentrate inactivated by heating at 80°C for 72 h as well as solvent/detergent treatment. Factor VIII (six batches from three manufacturers) and factor IX (two batches from one manufacturer) prepared from remunerated donors showed concentrations of HGV RNA as high as 8⫻105 copies of RNA/mL. Virus inactivation steps (eg, solvent detergent treatment or pasteurisation) substantially reduced the amounts of HGV and HCV RNA detectable in products; none of the eight virus inactivated batches tested was positive for either HCV or HGV. Although heat treatment greatly reduced the concentrations of detectable HGV RNA sequences in clotting-factor concentrates, only solvent/ detergent inactivation was associated with their complete elimination. Immunoglobulins were frequently contaminated with HGV, though at concentrations generally lower than those of clotting factor concentrates, whether prepared from plasma of non-remunerated or remunerated donors (table 2). The use of non-virus inactivated concentrates was associated with high frequencies of HCV infection among recipients. For example, the haemophiliacs in this study who had received factor VIII prepared from volunteer donor plasma (n=52) or from commercial sources (n=14) showed high rates of HCV viraemia (54 of 66 patients; table 3). High rates of HCV viraemia were also found among recipients of factor IX concentrates (haemophilia B; 15 of 17 patients) or intermediate-purity factor VIII concentrates (von Willebrand’s disease; ten of 12 patients). The other haemophiliacs (16 of 95) were antiHCV positive and PCR negative, indicating past resolved infection with HCV. None of the haemophiliacs treated with only virally inactivated clotting concentrates showed evidence of current or past HCV infection by PCR or serological testing. In contrast, persistent HGV infection in haemophiliacs was relatively uncommon, with the highest frequency (15%) in haemophilia A patients (table 3). Of the 13 haemophiliacs infected with HGV, all but one showed active coinfection with HCV. The single individual who was HCV PCR negative had evidence of chronic hepatitis on a liver biopsy sample. There was no association 1354

between HGV viraemia and HCV genotypes. Of the 79 haemophiliacs in whom HCV genotypes was identified, the following rates of HGV infection were observed: type 1a five of 35, type 1b one of 15, type 2 none of 5, type 3 six of 23, and type 5a none of one. There was no association between HGV infection and severity of haemophilia. Of the 40 patients with severe haemophilia A or B, five were HGV PCR positive compared with four of 25 with moderate, and three of 18 with mild disease. Similarly, there was no association between concurrent HIV infection and HGV viraemia (two of 20 HIV-positive haemophiliacs were HGV PCR positive, compared with 11 of 75 in HIV-negative haemophiliacs). HGV infection was detected in one of 17 recipients of exclusively inactivated concentrates, this frequency was not very different from that of the blood donors. The mother of this HGV-positive individual was also found to be HGV PCR positive. The immunoglobulin recipients had all been on longterm replacement therapy and had received large numbers of infusions of immunoglobulins. All of the immunoglobulin recipients had primary immunodeficiency syndromes or secondary immunodeficiency (three cases). Nevertheless, the frequencies of HGV infection, and therefore ability to transmit by transfusion and/or establish long-term chronic infection, was low (four of 57; table 2).

Discussion In this study we found a major difference between HCV and HGV in their ability to establish chronic infection in humans. HGV infection appears commonly in the general population (3·2% in Edinburgh, 1·7% in nonremunerated blood donors in the USA2), but there was little evidence that this rate of infection has led to persistent infection among recipients of blood products manufactured from the plasma of such individuals. The frequency of HGV infection among haemophiliacs, many of whom were immunosuppressed as a result of concurrent HIV infection, was low, consistent with findings in recipients of commercial factor VIII.2 Similarly, HGV viraemia was not frequently detected in recipients of long-term immunoglobulin replacement therapy (most with primary immunodeficiencies). These findings are in contrast with those for HCV, for which infection among haemophiliacs receiving non-inactivated concentrates is almost universal,11–13 despite a generally much lower frequency of HCV infection in the population of blood donors from whom such blood products were prepared.14 There are two possible explanations for this difference in frequencies of HCV and HGV infection in bloodproduct users. One possibility is that the two viruses are equally infectious but that exposed individuals are more able to clear infection with HGV. This hypothesis is supported by retrospective investigations of cases of HGV transmission by blood transfusion. For example, in a study of patients exposed to HGV by blood transfusion, acute HGV viraemia was detected in three of 13 patients, but only one remained viraemic after 18 months’ followup.2 Another possible biological explanation for the greater rate of clearance of HGV infection lies in the structural differences between the envelope glycoproteins of HCV and HGV. By analogy with HIV, another virus able to establish persistent infection in humans, HCV

Vol 348 • November 16, 1996

THE LANCET

may be able to persist because of the shielding effect of carbohydrate moieties on the surface of the virus which prevent antibody binding to the underlying protein. Persistence of HCV may also occur by a process of immune escape, in which rapid sequence change in the hypervariable region of the gene for the envelope protein E2 may prevent neutralisation by antibodies.15 Neither of these mechanisms of persistence may be possible for HGV; the putative envelope proteins of HGV (E1, E2) are probably less extensively glycosylated than those of HCV, with only four potential N-linked glycosylation sites compared with 15 for HCV.16 Furthermore, there is no evidence for the existence of a hypervariable region to allow immune evasion, with E1 and E2 showing similar degrees of variability to other parts of the HGV genome. In these respects, infections with HGV in humans may be more akin to infections with the related pestiviruses and flaviviruses, which are generally capable of establishing only transient infections in the immunocompetent host. Alternatively, the low rate of HGV infection in bloodproduct recipients could be explained by low infectivity of HGV compared with HCV. HGV RNA sequences detected in blood products could represent non-infectious virus particles, either because they were inactivated by steps in the manufacturing process or because neutralising antibody was present. The lack of glycosylation and the relatively conserved E1 and E2 sequences of HGV would clearly contribute to broad and effective neutralisation of HGV on mixing of plasma donations. Another, less likely, possibility is that the core protein of HGV is truncated or defective in some way preventing the production of fully infectious virus particles. Whatever the underlying reasons for the low rate of HGV infection in individuals who have had multiple transfusions, there remains the paradox of the high rate of HGV infection in the general population, several times that of HCV and presumably not related to past parenteral exposures, which is the main route of transmission of HCV in the UK and other western countries. However, cross-sectional studies of blood donors such as ours do not provide information of the incidence of new infection. For example, we do not know whether each of the HGV-infected donors represents a case of new, acute infection, or whether HGV may cause persistent infection in a proportion of individuals, as is the case for hepatitis B virus (HBV). Extending this analogy, mother-to-baby transmission of HGV (which may be relatively common17,18) could be associated with a higher rate of persistent infection. This may be the explanation for the HGV PCR positivity of the one haemophiliac who had received only virus-inactivated factor concentrate. For HBV, perinatal transmission is associated with persistent infection in around 90% of cases, compared with 2–10% in adults, and these individuals constitute the main reservoir of infection in communities where HBV infection is endemic.

Vol 348 • November 16, 1996

Longitudinal studies of individuals identified as HGV PCR positive are clearly essential for the clarification of the natural history of infection, and would help resolve the unusual discrepancy between the transmissibility of HCV and HGV we found. We thank Chris Healey and Helen Chapel (Department of Gastroenterology, John Radcliffe Hospital, Oxford), Billie Reynolds and Norah Davidson (Edinburgh Haemophilia Centre), and Carol Lycett (Department of Medical Microbiology, Edinburgh University).

References 1 Leary TP, Muerhoff AS, Simons JN, et al. Sequence and genomic organization of GBV-C: a novel member of the flaviviridae associated with human non-A-E hepatitis. J Med Virol 1996; 48: 60–67. 2 Linnen J, Wages J, ZhangKeck ZY, et al. Molecular cloning and disease association of hepatitis G virus: a transfusion-transmissible agent. Science 1996; 271: 505–08. 3 Muerhoff AS, Leary TP, Simons JN, et al. Genomic organization of GB viruses A and B: two new members of the flaviviridae associated with GB agent hepatitis. J Virol 1995; 69: 5621–30. 4 Yoshiba M, Okamoto H, Mishiro S. Detection of the GBV-C hepatitis virus genome in serum from patients with fulminant hepatitis of unknown aetiology. Lancet 1995; 346: 1131–32. 5 Kao JH, Chen PJ, Chen DS. GBV-C in the aetiology of fulminant hepatitis. Lancet 1996; 347: 120. 6 Simons JN, Leary TP, Dawson GJ, et al. Isolation of novel virus-like sequences associated with human hepatitis. Nature Med 1995; 1: 564–69. 7 Jarvis LM, Watson HG, McOmish F, Peutherer JF, Ludlam CA, Simmonds P. Frequent reinfection and reactivation of hepatitis C virus genotypes in multitransfused hemophiliacs. J Infect Dis 1994; 170: 1018–22. 8 Davidson F, Simmonds P, Ferguson JC, et al. Survey of major genotypes and subtypes of hepatitis C virus using restriction fragment length polymorphism of sequences amplified from the 5' non-coding region. J Gen Virol 1995; 76: 1197–204. 9 Simmonds P, Zhang LQ, Watson HG, et al. Hepatitis C quantification and sequencing in blood products, haemophiliacs, and drug users. Lancet 1990; 336: 1469–72. 10 Crawford RJ, Gillon J, Yap PL, et al. Prevalence and epidemiological characteristics of hepatitis C in Scottish blood donors. Transfus Med 1994; 4: 121–24. 11 Fletcher ML, Trowell JM, Craske J, Pavier K, Rizza CR. Non-A, nonB hepatitis after transfusion of factor VIII in infrequently treated patients. BMJ 1983; 287: 1754–57. 12 Tedder RS, Briggs M, Ring C, et al. Hepatitis C antibody profile and viraemia prevalence in adults with severe haemophilia. Br J Haematol 1991; 79: 512–15. 13 Watson HG, Ludlam CA, Rebus S, Zang LQ, Peutherer JF, Simmonds P. Use of several second generation assays to determine the true prevalence of hepatitis C infection in haemophiliacs treated with non-virus inactivated factor VIII and IX concentrates. Br J Haematol 1992; 80: 514–18. 14 Follett EAC, Dow BC, McOmish F, et al. HCV confirmatory testing of blood donors. Lancet 1992; 339: 928. 15 Weiner AJ, Geysen HM, Christopherson C, et al. Evidence for immune selection of hepatitis C virus (HCV) putative envelope glycoprotein variants: potential role in chronic HCV infections. Proc Natl Acad Sci USA 1992; 89: 3468–72. 16 Choo QL, Richman KH, Han JH, et al. Genetic organization and diversity of the hepatitis C virus. Proc Natl Acad Sci USA 1991; 88: 2451–55. 17 Feucht HH, Zollner B, Polywka S, Laufs R. Vertical transmission of hepatitis G. Lancet 1996; 347: 615–16. 18 Moaven LD, Tonnakoon PS, Bowden DS, Locarnini SA. Mother-tobaby transmission of hepatitis G virus. Med J Aust 1996; 165: 84–85.

1355