zyxwvutsrq zyxwvu zyxwvutsrqponm zyxwvuts zy Iournalof Viral Hepatitis. 1996. 3, 197-202 Hepatitis E virus infection in north-east Italy: serological study in the open population and groups at risk G.Gessoni' and F. Manoni' 'TransfusionaI Centre. Community Hospital, Chioggia (VeniceJ.Ztalg and 'Clinical Pathologg SPrvice. Community Hospital, ChiogBia (Venice). Italy Received 8]anuarg 1996; acceptedforpublication 7 March 1 9 9 6 SUMMARY. Developed western countries are considered zyxwvu zyxwv to be relatively free from endemic foci of hepatitis E virus (HEV) infections. The aim of this study was to assess the seroepidemiologyof HEV in north-east Italy. Of the 2361 individuals studied 1889 were representative of the general population and 472 were from groups at high risk for viral infections: 2 79 drug users and 193 patients on chronic haemodialysis. All sera were tested for hepatitis C virus antibody (HCVAb), human immunodeficiency virus antibody (HIVAb)and for hepatitis B virus (HBV) serology. Two solid-phase enzyme-linked immunosorbent assays (ELISA) were used to study the seroepidemiologyof HEV IgG, the first (using recombinant antigens) for screening, the second (using synthetic peptides) for confirmation of initially reactive samples. The prevalence of circulating hepati- INTRODUCTION North-east Italy is characterized by a close association of water and land owing to the presence of two large lagoons: the Venetian lagoon and the Ca Leri lagoon. In this area the estuaries of Adige and Brenta and the large Po delta are found. Chioggia is a town of about 60 000 inhabitants, with a harbour and large fishingfleet. In the Venetian lagoon there are many mollusc farms. This is an area characterized by high mortality Abbreviations:U S A . enzyme-linked immunosorbent assay: HAVAb, antibody to hepatitis A virus; HBsAg. hepatitis B surface antigen:HBV, hepatitis B virus; HCVAb. antibody to hepatitis C virus: HEV, hepatitis E virus: HEVAb, antibody to hepatitis E virus: HIVAb. antibody to human immunodeficiency virus; IVDU. intravenous drug users: ORF.open reading frame;. Correspondence:Dr Gianluca Gessoni. Transfusional Centre Community Hospital, Via Madonna Marina 5 0 0 . 3 0 0 1 9 Chioggia (Venice),Italy. 8 1 9 9 6 Blackwell Science Ltd tis E virus antibody (HEVAb)was 2.6%in the open population, 5.4% among drug users and 9.3% among patients on chronic haemodialysis. In the open population a positive relationship between age and prevalence of HEVAb was observed. A relationship between presence of HEVAb and serological evidence of previous HBV or HCV infections was also observed in this study. It was concluded that HEV infections are present in north-east Italy and are more frequent among subjects at risk for blood-borne viral infections. The positive correlation, observed in the open population, between age and prevalence of HEVAb suggests the presence of a cohort effect. Keywords: HEVAb. seroepidemiology, open population, Italy, parenteral transmission. from chronic liver disease and hepatocellular carcinoma. Seroepidemiological studies show a high prevalence of hepatototropic viral infections: in the open population the prevalence of hepatitis B surface antigen (HBsAg) chronic carriers is 3.9% and 42% have serological evidence of previous hepatitis B virus (HJ3V) infection [l]. More than 90% of the local adults have antibodies to hepatitis A virus (HAVAb) (unpublished data) and the prevalence of hepatitis C virus antibody (HCVAb) is higher than in other areas of north Italy [2-31. Hepatitis E virus (HEV)is the major aetiological agent of enteric non-A non-B hepatitis. It is associated with large epidemic outbreaks, particularly in developing countries, but sporadic infectious are described in western countries [4]. The virus is transmitted by the faecal-oral route through contamination of water or food [S] and secondary attack rates among exposed household members seem to be low in endemic areas [6]. Type E hepatitis is usually a mild self-limiting z zyxwvutsrqpo zy zyxwvutsr zyxwvu zyxwvuts zyxwvutsrqpo 198 G. Gessoni and F. Manoni disease [7]. but it is associated with fulminant liver disease in endemic areas [8] with a high rate of mortality (20%)among pregnant women [9]. HEV has a genome of single-stranded RNA of approximately 7.5 kb, which encodes three open reading frames (ORFs): at the 5' terminal there is a 2 7 nucleotide non-coding sequence and at the 3' terminal there is a 68 nucleotide non-coding sequence, which probably contains a poly-A tract [ 101. The 5' region of the genome, of approximately 5.0 kb, contains the ORF2 and ORF3 regions that are partially overlapping and encode structural proteins. The ORF1, at the 3'end of the genome is approximately 2,5 kb long, and encodes the viral replication proteins [ 101. Morphological studies and genome organization of HEV are consistent with those of the Caliciviriduefamily [ 111. Two immunodominant epitopes encoded by ORF2 and ORF 3 were identified by immunoscreening of HEV cDNA clones and solid-phase enzyme-linked immunosorbent assays (ELISA)have been developed to detect circulating HEV IgG (or IgM) by using synthetic or recombinant proteins from the Burmese or Mexican HEV strains [ 121. Western European countries are considered to be free of natural foci of HEV infection at present, although this disease was common in the past [ 1 31. There are several reports of sporadic cases of HEV infection in northern Europe in individuals from areas of high endemicity and under these circumstances very few secondary cases have been observed among contacts. In the Mediterranean European countries (Italy, Spain and Greece), sporadic HEV infections have been reported, related to the consumption of shellfish cultivated in polluted waters [14-16]. In 1994, Zanetti reported the results of a study on the prevalence of hepatitis E virus antibody (HEVAb) in north-west Italy (Torino): a low prevalence of circulating HEVAb in blood donors and drug addicts was observed [ 171. The aim of this retrospective seroepidemiological study, from subjects living in north-east Italy, was to assess the prevalence of circulating HEVAb in the open population and groups at high risk for viral infections. MATERIALS A N D METHODS Subjiii-ts To assess the seroprevalence of HEVAb in the open population 1889 sera were tested. They were collected from blood donors, women post childbirth and outpatients attending our Clinical Pathology Service for routine examination. We also tested two groups of subjects at high risk of blood-borne viral infections-2 79 intravenous drug users (IVDU), 221 males and 58 females with ages ranging from 18 to 3 5 years (mean 24.7), and 193 patients on long-term haemodialysis117 males and 76 females with ages ranging from 19 to 79 years (mean 59.5). All subjects were living in north-east Italy. Venous blood was sampled, serum separated within 4 h, and stored at -30°C. Each subject was tested for antibody to HIV (HIVAb) 1 and 2. and HCVAb, with third generation ELISA methods (Abbott Laboratories, Chicago, IL), and for HBV markers using commercial ELISA assays supplied by Hoffman la Roche (Basel CH). Sera reactive by ELISA for HCVAb and/or HIVAb were confirmed by immunoblotting (RIBA I11 HCV and RIBA HIV 1-2, Ortho Diagnostic System, Raritan, NJ), Detection ofHEVAb To detect circulating HEV IgG two commercial E1,ISA tests were used: the first (HEV HA. Abbott Laboratories) (ELISA1) for screening, and the second (HEV antibodies ELISA Kit. Nuclear Laser Medicine, MI) (ELISA2) for confirmation of sera initially reactive by ELISA 1. In the ELISAl test, human serum (or plasma) was diluted 1:400O and incubated with polystyrene beads precoated with recombinant HEV proteins representing immunodominant epitopes from ORF2 and ORF 3 of the Burmese strain. If HEVAb were present in the tested sample it bound to the coated beads and, after removing unbound material, bound human imniunoglobulins were detected by incubating the bead-antigen-antibody complex with horse-radish peroxidasrlabelled goat antibodies directed against human I&;. Unbound enzyme was then removed and the beads were washed. 0-Phenylenediamine solution, containing hydrogen peroxide, was then added to the beads: after incubation the enzyme substrate reaction was terminated by the addition of a 1N solution of sulphuric acid. The absorbance at 492 nm was read by a photometer. Samples with absorbance values less than the cut-off value were considered negative for HEVAb. samples with absorbance greater than the cut-off plus lo'%were considered initially reactive and retested in duplicate with the ELISA2 test. In the ELISA2 test human serum (or plasma) was diluted 1:20 and incubated in a polystyrene microplate 0 1996 Blackwell Science Ltd. jourrinl oJViral Hrpntilis, 3. 197-202 z zy zyxw zyx zyx zyxw Circulating HEVAb in north-east ltaly well coated with synthetic HEV proteins, representing immunodominant epitopes from ORF2 and ORF3 of the Chinese strain. If HEVAb were present in the tested sample it bound to the coated wells and after removing the unbound materials, bound human immunoglobulins were detected by incubating the wells-antigen-antibody complex with horse-radish peroxidaselabelled goat antibodies directed against human IgG, Unbound enzyme was then removed and the beads were washed. Tetramethylbenzidene solution containing hydrogen peroxide was added to the beads and the reaction terminated by addition of a 1N solution of sulphuric acid. The absorbance at 450 nm was read by a photometer. Samples with absorbances values lower than the cut-off value were considered negative for HEVAb, samples with absorbance values higher than the cut-off were are considered reactive. Only samples initially reactive by ELISA1 and confirmed by U S A 2 were considered positive for circulating HEV IgG. Statistical significance was assessed by the x2 test, and Pearson's test with Yate's correction, according to a computerized EPI Info program release 5.0 (CDCAtlanta, GA). RESULTS Table 1 summarizes the demographic data of the patients studied. Of the 2361 individuals tested for circulating HEVAb, 8 7 were initially reactive by ELISAl, and 4 of these samples were negative by ELISA2, therefore only 8 3 sera were reactive by both ELISA 1 and 2 and were considered positive for HEVAb. The overall prevalence of HEVAb in this study was 3.5%, with important differences between the open population and groups at high risk for blood-borne viral infections, such as drug users and patients on chronic haemodialysis. Serologicalresults obtained for HEW, HCV and HIV 199 infections are shown in Table 1.In the general population 38.9% had serological evidence of previous HBV infections, among IVDU this prevalence was 90.6%) and among haemodialysis patients it was 50.3%. In the general population the prevalence of HCVAb was 4.2% among IVDU 58.4% and among dialysis patients 27.9%. In the general population the prevalence of HIVAb was 0.05%, compared to 70.6 in IVDU and 0.5% in haemodialysis patients. In the open population the overall prevalence ofHEVAb was 2.6%(50/1889). In subjects less than 1 5 years old the prevalence was 0.3% (1/301),between 1 6 and 30years 1.7%(8/482), between 31 and 4 5 2.4'k (9/379),between 4 6 and 60 3.2% (13/403) and among subjects more than 60 years old the prevalence of HEVAb was 5.8% (19/324) (Table 2). In the general population subjects less than 4 5 years old had a lower prevalence of HEVAb than older subjects (1.1%vs 3.8%, PeO.01). The prevalence of HEVAb in the general population was 2.5% (25/973) among females and 2.7% (25/916) among males (P>0.05). Among IVDU the prevalence of HEVAb was 5.4% (151279) and among dialysis patients was 9.3% (18/193). This difference was statistically significant ( P < 0.05). The prevalence of HEVAb was significantly lower in the open population compared to the IVDUs (P e 0.05) and dialysis patients (P c 0.01). A significant association was observed between the presence of HEVAb and serological evidence of previous HBV (P e 0.05) and HCV (P < 0.01) infection. DISCUSSION There is little data regarding the prevalence of HEV infection and/or the prevalence of circulating HEV antibodies in developed countries. Grabow reported a prevalence of 1.8% among canoeists and 2.6% among zyxwvutsrq zyxwvu Table 1 Description of the study population n Male Mean age HEV positive HBV positive HCV positive HIV positive General population IVDU Haemodialysis patients 1889 48.5% 36.9 2.6% 38.9% 4.2% 2 79 74.6% 24.7 5.4% 90.6% 58.4% 70.6% 193 60.6% 59.5 9.3% 50.3% 27.9%1 0.5% U.05'% IVDU. Intravenous drug users. 8 1996 Blackwell Science Ltd. ]ournalof Viral Hepatitis. 3, 197-202 200 Age zyxwvutsr zyxwvutsrqpon zyxw zy zyxwvutsrqpo G. Gessoni and F. Manoni Less than 1 5 16-30 3145 46-60 Over 6 0 Total HEV positive (n) I1 HEV positive (%I) 301 1 0.3 482 3 79 403 324 8 9 13 19 50 1.7 2.4 3.2 5.8 2.6 1889 medical students in South Africa [ 181, Vandenvelde observed a seroprevalence of 2.3%in Belgian soldiers after return from Somalia [19], and Lavanchy in Switzerland observed a prevalence of 2.1% among pregnant women, 3.2% among blood donors and 6.4% among drug addicts [20]. In France, Halfon reported that 10.8'%, of haemodialysis patients were HEVAb positive [21] and in the Netherlands, Zaaijer [22,23] reported a prevalence of HEVAb from 0.4% to 1.8% among blood donors. In north-west Italy the prevalence of HEVAb is 0.95% among blood donors, 1.94% in drug addicts and 0.74%among healthy people [17]. In our study the overall prevalence of circulating HEVAb was 3.5% with great deviations in the four groups studied, ranging from 2.6% among the open population to 9.3%in haemodialysis patients. Among volunteer blood donors in Wisconsin, a region where HEV is not known to be endemic 5/200 (2.5%)were found to be HEV positive [24] and this is comparable to the prevalence reported in this study. Pregnant women are considered representative of the open unselected adult population for seroepidemiologicalstudies of viral infections: in this group in Switzerland the prevalence of HEVAb is 2.1% a result similar to that reported in the same area among blood donors (3.2%).In our study the HEVAb prevalence in blood donors was the same as that observed in the open population. This observation suggests that the usually adopted criteria for selection of repeat blood donors (screening for HBV, HCV and HIV, alanine aminotransferase (ALT) activity, history of previous jaundice or liver disease) are not sufficient to identify subjects with serological evidence of previous HEV infection. The frequency of circulating HEV IgC in subjects who had never visited tropical countries proves that HEV is present in north-east Italy. Perhaps the large number of restaurants serving frozen imported exotic food may be associated with a n increased risk of HEV infection. The need to assess the seroepidemiology of HEV among blood donors is Table 2 HEVAb in the open population zy zy emphasized by the reports of protracted viraeinia during acute sporadic HEV infection 1251 and by the observation of post-transfusional HEV infection [ 2 h ] .It is noteworthy that among the open population the positivity rate increased with age, varying from 0 . 3 % i n subjects aged less than 1 5 years old to 5.8%in subjects aged over 65 (x' for linear trend. P < 0.05). Among IVDU the prevalence of HEVAb was 5.4% higher than data reported by Zanetti (1.94%)117) but quite similar to results obtained by Lavancy (6.4'%)[20]. The high prevalence of HEVAb among IVDU might be the result of poor hygiene (faecal-oral route) and possibly also to a parenteral route of infection, as recently suggested 1261. Among IVDU in the IJSA, hepatitis outbreaks have been reported to be caused by other eiiterically transmitted infections, such as hepatitis A 1271.HEV. in common with other agents, could be transmitted in this population through the parenteral route (injection with contaminated needlestick). enteral route (ingestion of contaminated drugs) and even direct person-toperson contact facilitated by sexual practices. Previously, Montella [ 2 8 ] reported an association between the presence of circulating HEVAb and HIV infection in homosexual men but not among IVDU. Our study confirms the lack of a n association between HIV and HEV infection. This observation may be explained as follows: among IVDU HEV infection may be mainly transmitted through contaminated equipment, this route of infection being more eficient for HIV than HEV transmission. Among homosexual men HEV infection may be mainly transmitted through the high frequency of high risk sexual practices. The HEVAb prevalence among haemodialysis patients in this study was 9.3%. similar to the results reported by Halfon (10.3'X1) [ 2 1 1. This is significantly higher than that in the general population. The results obtained among IVDU and haemodialysis patients, groups at high risk for bloodborne viral infections, suggest that faecal-oral transmission may not be the only route of transmission of z 0 1996 Blackwell Science Ltd. loirrriril of Virril Hepititis. 3 , 19 7-202 Circulating HEVAb in north-east Italy HEV and that patients at high risk of HBV and HCV infections could also be infected with HEV. The results obtained in this study, mostly in the open population, could reflect a cohort effect: in a remote past HEV might be present in Chioggia, imported by immigrants, travellers or sailors, and became endemic owing to faecal contamination of the drinking water. In Chioggia there are many mollusc farms, and contamination of water by sewage and consumption of crude shellfish cultivated in these polluted waters may to be related to HEV infections. In Chioggia, HAV infection, the most important enterically acquired viral disease in western countries, is relatively common: a large proportion of adults, of 20-30 years, have circulating HAV antibodies, and 2461247 (99.6%)of the elderly patients in the city nursing home had HAV IgG (personal observation). Thomas found that HEVAb positivity rates increase with age and suggested a correlation with HCV infection [29],these observations were confirmed by Pisanti in Naples [30]. In the present study we confirmed the presence of a relationship between circulating HEVAb and the presence of HCVAb, and this relationship was statistically significant; a positive correlation ( P c 0.05) was also observed between the prevalence of HEVAb and serological evidence of previous HBV infection. The association between HEV and HCV, or HCV. suggests the presence of common routes of infection and indicates parenteral transmission of HEV [ 2 h ,3 I]. Because individuals with serological evidence of previous HBV infection and/or circulating HCVAb were older than negative subjects, and circulating HEVAb was found mainly in older patients, it seems that HEV circulated in this area many years ago. In an epidemiological setting characterized by an active circulation of hepatotropic viruses, even blood-borne or parenterally acquired, intrafarniliar transmission even though inefficient, could have an important role in the spread of both HCV and HEV in the past, during the years in which both infections peaked. Hence the epiderniologicaI features of HEV infection may be quite different in developed western countries than in developing countries in Asia or Africa. In endemic areas the importance of secondary attacks is low and the diffusion of the infection is sustained mainly from polluted drinking water. In non-endemic areas the importance of secondary attacks owing to patient-patient contact may play an important role in the spread of infection. Although at present HEV infection is not a major public health problem in western countries and a screening zy 201 for blood donors is not reformed the increasing number of travellers to HEV endemic areas and immigration from developing countries requires careful study. REFERENCES 1 Gessoni G. Manoni F. Antico F et al. Prevalence of serum markers’ Hepatitis B Virus (HBV) in Chioggia. In: Schiraldi 0.Pastore G and Dentico P. eds. Progress nnd Prospects in Viral Hepntitis. San Severo (FG): Gerni Editore, 199 1: pp 9 1-9 3. Chiaramonte M, Stroffolini T, Caporaso N et (11. Hepatitis C virus infection in Italy: a multicentric study. ztd ] Oatroeritrrol1991: 23: 555-558. Gessoni G, Manoni F. Prevalence of anti-hepatitis C virus antibodies among teenagers in the Venetian area: a seroepidemiological study. i h r ] Med 199 3; 2: 79-82. Khuroo M. Stdy of an epidemic of non-A non-B hepatitis. Possibility of another hepatitis virus distinct from posttransfusion non-A non-B type. A m ] Med 1980: 6 8 : 818-824. Parcell R, Ticehurst J. Enterically transmitted non-A non-B hepatitis: Epidemiology and clinical characteristics. In: Zuckerman A, ed. Viral Hepatitis rind Liver Disense. NewYork:AlanRiss. 1988:pp 131-137. Aggarwal R. Naik S. Hepatitis E: intrafamilial transmission versus waterborne spread. ] Hepntol 1994: 2 1 : 718-723. Balayan M. HEV infection: historical perspective. global epidemiology and clinical features. In: Hollinger 17. Lemon S and Purcell R, eds. Virril Hepititis ntid Liver Disease. Baltimore: Williams & Wilkins. 1991: pp 49 8-501. Nanda S. Yalcinkaya K, Panigrahi A, Acharya S, Jameel S, Panda S. Etiological role of hepatitis E virus in sporadic fulminant hepatitis, ] Med Viral 1994: 42: 13 3-1 37. Balayan M. Andjaparidze A. Saviskaya S . Ketiladze E. Braginsky D. Savinov A. Evidence for a virus in non-A non-B hepatitis transmitted via the fecal oral route. Zntervirologg 1983; 20: 23-31. Bradley D, Purdy M, Reyes G. Hepatitis E virus genome. Molecular features, expression of immunoreactive proteins and sequence divergence. ] Hepatol 199 1; 1 3 : 152-1 54. Bradley D. Hepatitis E: epidemiology, aetiology and molecular biology. Rev Mrd Virol 1992; 2: 19-28. Dawson G, Chau K. Cabal C. Yarbough P, Reyes G, Mushahwar I. Solid phase enzyme linked immunosorbent assays for hepatitis E virus IgG and IgM antibodies utilizing recombinant antigens and synthetic peptides. ] V i r d M r t h 1992; 38: 175-1 8 6 . Balayan M. Hepatitis E virus infection in Europe: regional situation regarding regarding laboratory diagnosis and epidemiology. Clin Diug Viral 1993: 1: 1-19. zyxwvu zyxwvutsr zyxwvutsrq zyxwvuts zyxwvu zyxwvutsrqp 2 3 4 5 zyxwvu h 7 8 9 10 11 12 zyxwvutsrqp 8 1996 Blackwell Science Ltd. Iourrml ofviral Hepatitis. 3. 197-202 13 202 zyxwvutsrqpon zyxwvutsrqp G. Gessoni and F. Manoni zyxwvuts zy zyxwvutsrq zyxwvut zyxwv 14 Tassopoulos N, Krawczynski K, Hatzakis A. Karsoulidou A. Uelladetsima I, Koutelou M. Case report: role of hepatitis E virus in the etiology of community acquired non-A non-B hepatitisinGreece.]MedVirol1994; 42: 124-128. 1 5 Coursaget P, Depril N, Buisson Y. Molinik C, Roue R. Hepatitis type E in a French population: detection of antiHEV by a synthetic peptide based enzyme linked immunosorbent assay. Res Virol1994; 145: 51-57. 1h Caredda F, Antinori S, Pastecchia C. Coppin P. Moroni M. Sporadic non-A non-B hepatitis. h n r e t 1988; I: 58-59. 17 Zanetti A. Dawson G and the study group of hepatitis E. Hepatitis type E in Italy: a seroepiderniological survey. ] M~dVirol1994: 42: 318-320. 18 Crabow W, Favorov M, Khudyakova N, Taylor M, Fields H. Hepatitis E seroprevalence in selected individuals in South Africa. Med Virol1994; 44: 384-388. 19 Vandenvelde C, Heyvaert F. Maes F, Kadi L, Opstellen R. Hepatitis E virus infection in Belgian soldiers. Lancet 1994; 344: 747. 2 0 Lavanchy D, Morel B, Frei P. Seroprevalence of hepatitis E virus in Switzerland. Lancet 1994; 344: 747-748. 2 1 Halfon P, Ouzan D. Chanas M, Khiri H, Feryn J, Mangin L. High prevalence of hepatitis E virus antibodies in haemodialysis patients. Lancet 1994; 344: 746. 22 Zaaijer H, Yin M. Lelie P. Seroprevalence of hepatitis E in the Netherlands. Lancet 1992; 340: 681. 2 3 Zaaijer H, Kok M, Lelie P, Timmerman R, Chan K, Van der Pod H. Hepatitis E in the Netherlands, imported and endemic. h r i c d 1992; 341: 826. 2 4 Paul D, Knigge M, Ritter A, Gutierrez R. Pilot-Matias T. Chau K. Determination of hepatitis E virus seroprevalence by using recombinant fusion protein and synthetic peptides. InJect Dis 1994; 169: 801-80h. 25 Nanda S, Ansari I. Acharya S. Jameel S. Panda S . Protracted viremia during acute sporadic hepatitis E virus infection. Gastromterologg 199 5; 108: 225-2 3 0 . 2 6 Wang C, Fleming B, Moeckli R. 'l'ransmission of hepatitis E virus by transfusion? Lancet 199 3; 341: 825-826. 2 7 CDC. Hepatitis A among drug abusers. Morb Mort Wrekl Rep 1988; 37: 297-305. 28 Montella F, Rezza G. Di Sora F, Pezzotti P, Recchia 0 . Association between hepatitis E virus and HIV infection in homosexual men. k m e t 1994; 344: 14 3 3. 2 9 Thomas D. Mahley R, Badur S. Palaoglu K, Quinn T. Epidemiology of hepatitis E virus infection in Turkey. h n c e t 1993: 341: 1561-1562. 30 Pisanti F, Coppola A. Galli C. Association between hepatitis C and hepatitis E virus in southern Italy. I m c r t 1994; 344: 746-747. 31 Chauhan A. Jameel S, Dilawari C, Chawla Y, Kaur IJ. Ganguly N. Hepatitis E virus transmission to a volunteer. Lancet 199 3: 341: 149-1 50. 0 1996 Blackwell Science Ltd. Iournnl oJVirnl Hepatitis, 3, 19 7-202