Journal of Medical Virology 81:15311538 (2009)

Association of Concurrent Hepatitis B Surface Antigen and Antibody to Hepatitis B Surface Antigen With Hepatocellular Carcinoma in Chronic Hepatitis B Virus Infection
Ji Sun Jang,1,2 Hyoung Su Kim,1 Ha Jung Kim,3 Woon Geon Shin,1 Kyung Ho Kim,1 Jin Heon Lee,1 Hak Yang Kim,1 Dong Joon Kim,1 Myung Seok Lee,1 Choong Kee Park,1 Byung-Hoon Jeong,4 Yong-Sun Kim,4 and Myoung Kuk Jang1*
1 2

Department of Internal Medicine, Hallym University Medical Center, Seoul, South Korea Department of Internal Medicine, Seoul Veterans Hospital, Seoul, South Korea 3 Laboratory for Gastroenterology and Hepatology, Hallym University Medical Center, Seoul, South Korea 4 Ilsong Institute of Life Science, Hallym University, Anyang, South Korea

Antibody to hepatitis B surface antigen (HBsAg) (anti-HBs) can exist in patients with chronic hepatitis B virus (HBV) infection. To date, little is known about the association of concurrent HBsAg and anti-HBs (concurrent HBsAg/ anti-HBs) with hepatocellular carcinoma (HCC). The aim of this study was to investigate the clinical relevance of concurrent HBsAg/anti-HBs with preS deletion mutations and HCC in chronic HBV infection. A total of 755 patients with chronic HBV infection were included consecutively at a tertiary center. Logistic regression analysis was used to identify risk factors for HCC, and serum HBV DNA was amplied, followed by direct sequencing to detect preS deletions. The prevalence of concurrent HBsAg/anti-HBs was 6.4% (48/755) and all HBVs tested were genotype C. HCC occurred more frequently in the concurrent HBsAg/anti-HBs group than in the HBsAg only group [22.9% (11/48) vs. 7.9% (56/707), P 0.002]. In multivariate analyses, age >40 years [odds ratio (OR), 14.712; 95% condence interval (CI), 4.36549.579; P < 0.001], male gender (OR 2.431; 95% CI, 1.2264.820; P 0.011), decompensated cirrhosis (OR, 3.642; 95% CI, 1.7887.421; P < 0.001) and concurrent HBsAg/anti-HBs (OR, 4.336; 95% CI, 1.9569.613; P < 0.001) were associated independently with HCC. In molecular analysis, preS deletion mutations were more frequent in the concurrent HBsAg/anti-HBs and HCC groups than in the HBsAg without HCC group (42.3% and 32.5% vs. 11.3%; P 0.002 and 0.012, respectively). In conclusion, concurrent HBsAg/anti-HBs is associated with preS deletion mutations and may be one of the risk factors for HCC in chronic HBV infection
2009 WILEY-LISS, INC.

with genotype C. J. Med. Virol. 81:1531 1538, 2009. 2009 Wiley-Liss, Inc. KEY WORDS: concurrent hepatitis B surface antigen and antibody; genotype C; hepatitis B virus; hepatocellular carcinoma

INTRODUCTION Chronic hepatitis B is a major global public health problem. Approximately 400 million people worldwide have chronic hepatitis B virus (HBV) infection and are at high risk of developing complications such as liver cirrhosis or hepatocellular carcinoma (HCC) [Beasley, 1988; Poland and Jacobson, 2004]. Based on epidemiological evidence, particular ethnicity (Asians in the Far East, Africans), old age (males >40 years, females >50 years), underlying liver cirrhosis, and family history of HCC are well known to be risk factors for HCC in HBV carriers. Vigorous surveillance on a
Abbreviations: anti-HBs, antibody to hepatitis B surface antigen; HBsAg, hepatitis B surface antigen; HBV, hepatitis B virus; HCC, hepatocellular carcinoma. Parts of the results were presented at 43rd Annual Meeting of the European Association for the Study of the Liver, April 2327, 2008, Milan, Italy. *Correspondence to: Myoung Kuk Jang, Department of Internal Medicine, Kangdong Sacred Heart Hospital of Hallym University Medical Center, 445, Gildong, Kangdong-gu, Seoul 134-701, South Korea. E-mail: mkjang@hallym.or.kr Accepted 19 May 2009 DOI 10.1002/jmv.21577 Published online in Wiley InterScience (www.interscience.wiley.com)

1532

Jang et al.

regular basis should therefore be recommended to detect early HCC in patients with chronic HBV infection [Bruix and Sherman, 2005]. In addition to the host factors, various viral factors of HBV are associated with HCC development, for example, higher serum HBV DNA levels, genotype C, basal core promoter mutations, and preS (S1, S2) deletion mutations [Chen et al., 2006, 2008; Fang et al., 2008; Lin and Kao, 2008]. Generally, the development of antibody to hepatitis B surface antigen (HBsAg) (anti-HBs) leads to the clearance of infecting HBV. However, the concurrent presence of HBsAg and anti-HBs (concurrent HBsAg/ anti-HBs) has been reported occasionally in chronic hepatitis B patients. In the past, this phenomenon was regarded simply as superinfection with a second HBV strain [Heijtink et al., 1982; Shiels et al., 1987]. Recent studies, however, have demonstrated the probability of mutations in the viral surface gene and/or defective host immunity against HBV [Margeridon et al., 2005; Lada et al., 2006]. HBV strains harboring preS deletion mutations can not only induce dysplasia of hepatocytes and HCC in transgenic mice but are also detected often in the sera and/or the liver tissues of patients with HCC [Fan et al., 2000, 2001; Wang et al., 2006; Zhang et al., 2007]. Clinically, concurrent HBsAg/anti-HBs may be associated with mutations in the preS gene, and has been detected more frequently in advanced liver diseases than mild cases [Weinberger et al., 1999; Colson et al., 2007]. Taken together, the previous data suggest that concurrent HBsAg/anti-HBs may be related to advanced liver disease via preS deletion mutations. However, a clinical impact of concurrent HBsAg/anti-HBs on HCC in chronic hepatitis B remains to be claried. The aim of this study was to investigate the clinical relevance of concurrent HBsAg/anti-HBs for the development of HCC as well as the occurrence of preS deletion mutations in chronic hepatitis B infection with genotype C. METHODS Patients and Study Design A total of 755 patients with chronic HBV infection, HBsAg positive for at least 6 months, were included consecutively in this study at Hallym University Medical Center, Seoul, Korea, from March 2003 to April 2008. Exclusion criteria were as follows: patients who had other viral markers (IgM anti-HAV, anti-HCV) were serious alcoholics (>20 g alcohol/day), showed evidence of drug-induced hepatic injury within the past 3 months, and had any malignancy apart from HCC. All patients underwent clinical, biochemical, and virological evaluations at every visit. Clinical parameters were compared between the HBsAg only group and the concurrent HBsAg/anti-HBs group, or the non-HCC group and the HCC group. Univariate and multivariate analyses were performed to identify independent risk factors for HCC. Stored sera from the concurrent HBsAg/anti-HBs group, the HCC group, and a subset
J. Med. Virol. DOI 10.1002/jmv

of the control group (HBsAg only without HCC) were examined to investigate the occurrence of preS (S1, S2) deletion mutations. The study was approved by the Investigation and Ethics Committee for Human Research at the Hallym University Medical Center, Seoul, Korea. Denitions The denition of chronic hepatitis B was based on the American Association for the Study of Liver Diseases Practice Guidelines [Lok and McMahon, 2007], and anti-HBs positivity was dened as serum anti-HBs higher than 10 mIU/ml in at least two independent measurements taken 6 months apart. Decompensated cirrhosis was diagnosed by histology or ultrasonographic/ CT imaging features, supplemented with a relevant portal hypertension (esophageal and/or gastric varices, ascites, splenomegaly with a platelet count of <100,000/mm3) or hepatic encephalopathy clinically. Compensated cirrhosis was therefore dened as being cirrhosis without any evidence of decompensation [de Jongh et al., 1992]. Diagnosis of HCC was based on histology or on the presence of hypervascular liver masses with serum alpha-fetoprotein (AFP) levels higher than 400 ng/ml [Tangkijvanich et al., 2000]. Perinatal transmission was dened according to the following criteria: (1) having tested positive to HBsAg at any point in the patients lifetime, (2) patients mother tested positive for HBsAg prior to delivery, (3) patients father tested negative for HBsAg, and (4) no history of blood transfusion, surgical operations, parenteral drug abuse, needle stick injury, or any other risk factors for HBV infection such as undergoing hemodialysis. Serological Testing Serum HBsAg, anti-HBs, HBeAg, and anti-HBe were tested using commercial enzyme-linked immunoassay kits (Abbott Laboratories, North Chicago, IL). Serum HBV DNA levels were tested by a hybrid capture II HBV DNA test (Digene, Gaithersburg, MD; detection limits of 1.4 105 to 1.7 109 copies/ml) or VERSANT HBV DNA 3.0 Assay (bDNA) (Bayer HealthCare LLC, Tarrytown, NY; detection limits of 2.0 103 to 1.0 108 copies/ml). HBV DNA Extraction and HBV Genotyping by PCR Serum HBV DNA was extracted using commercial kits (QIAamp1 DNA Blood Mini Kit; Qiagen, Hilden, Germany), according to the manufacturers protocol. All serum samples were stored at 708C until use. PCR with type-specic primers, which were reported originally by Naito et al. [2001], was used to conrm HBV genotype. This assay was designed based on the conserved nature of nucleotide sequences in regions of the preS1 through S genes and the differences in the sizes of the genotype-specic bands. As the HBV genotype is genotype C in >99% of chronic hepatitis B patients in Korea [Song et al., 2005], the PCR could be performed with only two primers: common universal

HBsAg/Anti-HBs and Hepatocellular Carcinoma

1533

primer (types A to E-specic, sense) and type C-specic primer (antisense) (Table I). A negative control was included for every 16 samples tested. Analysis of preS Deletion Mutations in HBV DNA PreS (preS1, preS2) genes of the HBV genome were amplied by nested PCR, and the puried PCR products were sequenced directly as described previously [Gao et al., 2007]. The primers used for the rst and second round PCR are listed in Table II. The predicted length of the target sequence was 699 bp, which covered the whole preS region. In order to nd deletion mutations in the preS gene, serum was obtained from 81 patients who had concurrent HBsAg/anti-HBs or HCC. After excluding the PCR-negative cases, HBV DNA was sequenced directly using ABI prismTM BigDyeTM Terminator Cycle Sequencing Ready Reaction Kit (Applied Biosystems, Inc., Foster, CA) on an ABI automated uorescent sequencer (ABI 3730XL; Applied Biosystems, Inc.). HBV DNA sequences were analyzable fully in 66 out of 81 patients. For comparative analysis, sera from 53 patients who were positive for HBsAg only and did not develop HCC were used as controls. The age, gender distribution, and disease status of these patients were similar to the study groups when selected (data not shown). If a serum sample showed two or more clones when PCR products were run on a gel, all HBV DNA sequences were analyzed. Nucleotide sequences were compared with the published sequences of HBV genotype C subtype adr registered in GenBank (NCBI) AF286594 (GI: 22651877) using Chromas viewer and DNAMAN (version 4.1) software. Statistical Analysis The MannWhitney U-test for continuous variables and the chi-squared test for categorical variables were used in the analyses as appropriate. Logistic regression analysis (multivariate analysis) was performed with candidate variables in univariate analysis (P-value less than 0.1). All P-values were two-sided and a P-value of less than 0.05 was considered signicant statistically.

Statistical analysis was performed using the dBSTAT program (dBSTAT version 4.5, Seoul, Korea). RESULTS The median age of the total study population was 45 years (range 191) and 65.2% of the cohort was male (492/755). All HBVs tested (n 119) were genotype C (Fig. 1). Eighteen percent (133/755) had evidence of decompensated cirrhosis and 8.9% (67/755) had HCC. Clinical Signicance of Concurrent HBsAg/Anti-HBs The prevalence of concurrent HBsAg/anti-HBs was 6.4% (48/755). There were no signicant differences in baseline characteristics such as age, gender, transmission mode, family history of HCC, serum ALT levels, serum HBV DNA levels, and the existence of decompensated cirrhosis between the concurrent HBsAg/ anti-HBs group and the HBsAg only group. HBeAg positivity was higher in the concurrent HBsAg/anti-HBs group than in the HBsAg only group (67.4% vs. 50.2%; P 0.024). Intriguingly, HCC was more frequent in the concurrent HBsAg/anti-HBs group than in the HBsAg only group (22.9% vs. 7.9%; P 0.002) (Table III). Risk Analysis of Concurrent HBsAg/Anti-HBs on HCC In univariate analysis, age >40 years, male gender, family history of HCC, decompensated cirrhosis, and concurrent HBsAg/anti-HBs were associated signicantly with HCC and perinatal transmission had marginal signicance (Table IV). These variables were considered candidates for multivariate analysis. In multivariate logistic regression analyses, age >40 years [odds ratio (OR), 14.712; 95% condence interval (CI), 4.36549.579; P < 0.001], male gender (OR, 2.431; 95% CI, 1.2264.820; P 0.011), decompensated cirrhosis (OR, 3.642; 95% CI, 1.7887.421; P < 0.001), and concurrent HBsAg/anti-HBs (OR, 4.336; 95% CI, 1.9569.613; P < 0.001) were independent risk factors for HCC (Table V).

TABLE I. Primer Sequences Used for HBV Genotyping Specicity Polarity

0

Sequencesa 5 -GGCTCMAGTTCMGGAACAGT-3 50 -GGTCCTAGGAATCCTGATGTTG-30

0

Position nt 6786 nt 165186

Types A to E-specic Sense Type C-specic Antisense

a

An M represents a nucleotide that could be either an A or a C.

TABLE II. Primer Sequences for Nested PCR Use 1st round PCR 2nd round PCR Polarity Sense Antisense Sense Antisense Sequences 50 -ACATACTCTGTGGAAGGCTG-30 50 -TTGAGAGAAGTCCACCACGA-30 50 -GGAAGGCTGGCATTCTATAT-30 50 -C TGTGGTATTGTGAGGATTC-30 Position nt 27502769 nt 273254 nt 27612780 nt 244225
J. Med. Virol. DOI 10.1002/jmv

1534

Jang et al.

Fig. 1. HBV genotyping. All PCR products with type-specic primers showed an estimated size of 122 bp, indicating genotype C. M, molecular size standard.

The risk of HCC increased gradually with the number of risk factors exhibited by the patient. When the male group with concurrent HBsAg/anti-HBs was compared to the female group with HBsAg only, the OR for HCC was 6.512 [95% CI, 2.54216.682, P < 0.001; 28.1% (9/32) vs. 5.7% (14/247), respectively]. As for the decompensated cirrhosis group with concurrent HBsAg/anti-HBs, the OR was 12.867 compared to the HBsAg only group with compensated liver disease [95% CI, 2.76259.948, P 0.006; 42.9% (3/7) vs. 5.5% (32/581), respectively]. Notably, the OR for HCC in the

concurrent HBsAg/anti-HBs group with age >40 years peaked at 146.667 [95% CI, 17.2071250.113, P < 0.001; 34.8% (8/23) vs. 0.004% (1/276), respectively] compared to the HBsAg only group with age <40 years. Association of Concurrent HBsAg/Anti-HBs or HCC With preS Deletion Mutations The existence of preS deletion mutations was assessed in the 119 patients whose sera were nested PCR positive for HBV DNA and for whom DNA sequences were

TABLE III. Baseline Characteristics of the Patients (n 755) Variables, n (%) Age (years) Male gender, n (%) Perinatal transmission, n (%) FHx of HCC, n (%) Serum ALT (IU/L)a HBeAg(), n (%)b Log (HBV-DNA)a,c,d Decompensated cirrhosis, n (%) HCC, n (%)
a

HBsAg only, 707 (93.6) 45 (191) 460 (65.1) 196 (27.7) 55 (7.8%) 43 (21,651) 329/656 (50.2) 6.4 (<3.38.0) 126 (17.8) 56 (7.9)

Concurrent HBsAg/ anti-HBs, 48 (6.4) 49 (1872) 32 (66.7) 16 (33.3) 6 (12.5%) 59 (6817) 31/46 (67.4) 6.2 (<3.38.0) 7 (14.6) 11 (22.9)

P-value 0.07 0.82 0.40 0.27 0.97 0.02 0.80 0.57 0.002

ALT, alanine aminotransferase; FHx, family history; HBV, hepatitis B virus; HCC, hepatocellular carcinoma. a Median (range). b HBeAg positivity was not determined in 53 patients. c Serum levels, log10(copies/ml). d Cut-off levels <2,000 (3.3 log) copies/ml.

TABLE IV. Univariate Analysis of Risk Factors for HCC Variables, n (%) Age (>40 years), n (%) Male gender, n (%) Perinatal transmission, n (%) FHx of HCC, n (%) Serum ALT (IU/L)a HBeAg(), n (%)b HBV-DNAa,c,d Decompensated cirrhosis, n (%) Concurrent HBsAg/anti-HBs, n (%) HCC, 67 (9) 64 (96) 51 (76) 12 (18) 14 (20.9%) 46 (6495) 28/64 (44) 5.9 (<3.38.0) 27 (40) 11 (16) Non-HCC, 688 (91) 399 (58) 441 (64) 200 (29) 47 (6.8%) 43 (21651) 332/638 (52) 6.5 (<3.38.0) 106 (15) 37 (5) P-value <0.001 0.049 0.052 <0.001 0.18 0.21 0.39 <0.001 0.002

ALT, alanine aminotransferase; FHx, family history; HBV, hepatitis B virus; HCC, hepatocellular carcinoma. a Median (range). b HBeAg positivity was not determined in 53 patients. c Serum levels, log10(copies/ml). d Cut-off levels <2,000 (3.3 log) copies/ml.

J. Med. Virol. DOI 10.1002/jmv

HBsAg/Anti-HBs and Hepatocellular Carcinoma

TABLE V. Multivariate Analysis of Risk Factors for HCC Odds ratio (95% CI) Age (>40 years) Male gender Perinatal transmission FHx of HCC Decompensated cirrhosis Concurrent HBsAg/anti-HBs 14.712 (4.36549.579) 2.431 (1.2264.820) 0.938 (0.4641.895) 0.574 (0.2311.425) 3.642 (1.7887.421) 4.336 (1.9569.613) P-value <0.001 0.011 0.938 0.231 <0.001 <0.001

1535

evaluable. Twenty-ve percent of the patients (30/119) had preS deletion mutations ranging from 6 to 207 bp long. With respect to the locations of these mutations, deletions in the preS2 region were more common (22/119, 18.5%) than preS1 deletions (13/119, 10.9%). These percentages included start codon mutations for preS1 (4/119, 3.4%) and preS2 (3/119, 2.5%). Four percent of the patients (5/119) had deletions in both preS1 and preS2 regions (Fig. 2). These patients were reassigned into four groups: Group I, HBsAg only without HCC (control group, n 53); Group II, HBsAg only with HCC (n 40); Group

III, concurrent HBsAg/anti-HBs without HCC (n 18); Group IV, concurrent HBsAg/anti-HBs with HCC (n 8). The prevalence of preS deletion mutations was higher signicantly in the concurrent HBsAg/anti-HBs and HCC groups (groups IIIV) than in the HBsAg only without HCC group (group I) (all P < 0.05) (Fig. 3). Specically, preS1 deletion mutations showed more signicant relationships with concurrent HBsAg/ anti-HBs (Groups III and IV) than HBsAg only groups (Groups I and II) [34.6% (9/26) vs. 4.3% (4/93); P < 0.001] (Fig. 4). Although preS2 deletion mutations occurred more frequently in HCC groups (Groups II and IV) than

Fig. 2. Detection and location of preS deletion mutations in each group. The preS1 and preS2 DNA sequences are represented at the top of this gure: spans in base pairs are shown in the boxes and nucleotide numbers are shown above the boxes. Deletion mutations are indicated by solid lines, while preS1 and preS2 start codon mutations are shown by the symbols * and **. Group I, HBsAg only without HCC; Group II, HBsAg only with HCC; Group III, concurrent HBsAg/anti-HBs without HCC; Group IV, concurrent HBsAg/anti-HBs with HCC.

J. Med. Virol. DOI 10.1002/jmv

1536

Jang et al.

Fig. 3. Prevalence of preS deletion mutants was higher signicantly in groups IIIV than in group I (32.5% (13/40), 38.9% (7/18), and 50.0% (4/8) vs. 11.3% (6/53), respectively, *P 0.012, **P 0.015, ***P 0.020).

Fig. 5. Prevalence of preS2 deletion mutants was higher in HCC groups than in non-HCC groups (P 0.014).

non-HCC groups (Groups I and III) [29.2% (14/48) vs. 11.3% (8/71); P 0.014], there was no signicant association with concurrent HBsAg/anti-HBs (Fig. 5). DISCUSSION To date, some researchers have shown an association between concurrent HBsAg/anti-HBs and preS deletion mutations, while others revealed the relationship between the preS deletion mutations and HCC development [Fan et al., 2000, 2001; Wang et al., 2006; Zhang et al., 2007; Chen et al., 2008; Fang et al., 2008]. However, little is known about the clinical relevance of concurrent HBsAg/anti-HBs with HCC development and preS deletion mutations in humans. This crosssectional study showed the clinical relationship between concurrent HBsAg/anti-HBs and HCC, and evaluated preS deletion mutations as signicant contributors to HBV-associated hepatocarcinogenesis. This nding suggests the role of concurrent HBsAg/anti-HBs as a risk factor for the development of HCC. This study population presented homogenously genotype C, which was comparable to a previous report in which most patients had genotype C regardless of various clinical outcomes in Korea [Song et al., 2005]. Therefore, this was an opportunity to analyze the association of concurrent HBsAg/anti-HBs and HCC without confounding effects of HBV genotype.

Fig. 4. Concurrent HBsAg/anti-HBs groups had preS1 deletion mutations more frequently than HBsAg only groups (P < 0.001).

The prevalence of concurrent HBsAg/anti-HBs in patients with HBV infection had been reported as less than 5% to over 30%. In the present study, 6.4% (48/755) of patients with chronic HBV infection had anti-HBs concomitantly, which was lower than other previous studies. However, these studies included acute hepatitis B in addition to chronic hepatitis B [Tsang et al., 1986; Shiels et al., 1987] or failed to mention the analytical threshold of the anti-HBs titer [Heijtink et al., 1982; Tsang et al., 1986; Shiels et al., 1987; Hayashi et al., 1990]. Therefore, several previous studies might have overestimated the prevalence of concurrent HBsAg/ anti-HBs. To date, anti-HBs titer over 10 mIU/ml is regarded as a protective level, since a lower anti-HBs titer might be directed against a subdeterminant with little reactivity against a determinant of HBsAg [Swenson et al., 1983] or might be non-specic antibodies to other antigens occurring naturally that cross-react with a subdeterminant of HBsAg [Wang et al., 1991]. Recent studies using the criteria of anti-HBs titer >10 mIU/ml for positive anti-HBs have shown that the prevalence of concurrent HBsAg/anti-HBs varies from less than 3% to approximately 10% in chronic hepatitis B patients [Wang et al., 1991; Lada et al., 2006; Colson et al., 2007]. In the present study, anti-HBs positivity was dened as serum anti-HBs higher than 10 mIU/ml in at least two independent measurements, and the prevalence of concurrent HBsAg/anti-HBs was similar to that of recent studies. This study revealed a clear statistical association of concurrent HBsAg/anti-HBs with HCC, even though only a small number of patients with concurrent HBsAg/ anti-HBs (n 48) were enrolled. Concurrent HBsAg/ anti-HBs has been considered a challenge for serological diagnosis and problem for determining clinical signicance or prognosis during the natural course of chronic HBV infection. While some studies have described that concurrent HBsAg/anti-HBs does not reect a distinct clinical entity [Tsang et al., 1986; Dienstag, 1987; Hayashi et al., 1990], others have suggested that concurrent HBsAg/anti-HBs is related to more advanced liver diseases such as chronic hepatitis B with advanced brosis, cirrhosis, and a worse prognosis

J. Med. Virol. DOI 10.1002/jmv

HBsAg/Anti-HBs and Hepatocellular Carcinoma

1537

[Trepo et al., 1976; Heijtink et al., 1982; Shiels et al., 1987]. It was thought to be caused by an activated immunological response followed by active inammation for a prolonged period. Advanced liver brosis in patients with concurrent HBsAg/anti-HBs may be related to a long history of HBV infection since the median age of these patients was found to be higher than controls [Colson et al., 2007]. However, previous reports did not consider concurrent HBsAg/anti-HBs as a risk factor for HCC development. In this study, there were no signicant differences in demographic features (age, gender, transmission mode, family history of HCC), laboratory ndings (serum ALT levels, serum HBV DNA levels), or disease status (existence of decompensated cirrhosis) between the concurrent HBsAg/anti-HBs group and the HBsAg only group (control). Although HBeAg positivity was higher in the concurrent HBsAg/ anti-HBs group than in the HBsAg only group, it was not related to HCC in either univariate or multivariate analyses. Meanwhile, the concurrent HBsAg/anti-HBs group exhibited signicantly higher prevalence of HCC than the HBsAg only group. In multivariate analyses, concurrent HBsAg/anti-HBs remained an independent risk factor for HCC. As mentioned above, age >40 years, male gender, and decompensated cirrhosis were also independent risk factors for HCC. Notably, the risk of HCC increased when concurrent HBsAg/anti-HBs was present in combination with these other factors. For instance, the OR for HCC peaked at 146.667 (P < 0.001) in the concurrent HBsAg/anti-HBs group with age >40 years. This nding suggests that concurrent HBsAg/anti-HBs may be a risk factor for the development of HCC. The present study also demonstrated the molecular association of preS deletion mutations with HCC and concurrent HBsAg/anti-HBs. Previous studies revealed that concurrent HBsAg/anti-HBs was associated with deletion mutations in the preS region of the HBV genome [Wang et al., 1999; Lu et al., 2006], which are associated with intracellular retention of the viral envelope proteins and the histologic appearance of ground-glass hepatocytes [Gerber et al., 1974; Chisari et al., 1987; Xu and Yen, 1996]. That is, the HBsAg with the preS mutants accumulates in the endoplasmic reticulum (ER) and evokes signicant ER stress [Chisari et al., 1987; Hsieh et al., 2004] and oxidative stress, which results in DNA damage (genomic instability) of hepatocytes [Hagen et al., 1994]. These responses cause severe, prolonged hepatocellular injuries that can initiate a programmed response characterized by inammation, regenerative hyperplasia, transcriptional deregulation, and aneuploidy, which can lead to development of HCC in turn [Chisari et al., 1989; Wang et al., 2006; Zhang et al., 2007]. Therefore, the preS region of the HBV genome was analyzed in the sera of the patients with concurrent HBsAg/anti-HBs or HCC and compared with the control group (HBsAg only group without HCC). In accordance with expectations, preS deletion mutations were detected more frequently in the concurrent HBsAg/anti-HBs group than in the control

group (42.3% vs. 11.3%; P 0.002). Deletion mutations in the preS region were also detected in HCC patients with a considerable prevalence (32.5%). That is, the concurrent HBsAg/anti-HBs group not only demonstrated a signicantly higher prevalence of HCC, but also more frequent preS deletion mutations. In particular, a signicant higher prevalence of preS2 deletion mutations was found in HCC patients. The prevalence of deletions in the preS1 and preS2 regions in the present study was comparable to that of a recent study in patients with genotype C HBV infection [Choi et al., 2007]. Although deletions in the preS2 region have been proposed to contribute to HCC development previously [Blackberg and Kidd-Ljunggren, 2003; Choi et al., 2007], a subsequent study showed a higher prevalence of preS1 deletions in Korean patients with HCC and genotype C HBV infection [Mun et al., 2008]. Both preS1 and preS2 deletions are related to ER stress and induction of oxidative DNA damage, as described above. Therefore, deletions of the preS1 region also may be important in the development of HCC, at least in chronic hepatitis B patients with concurrent HBsAg/anti-HBs. In this study, ER stress and oxidative DNA damage in hepatocytes were not assessed. Therefore, further studies will be needed to evaluate the ER stress in hepatocytes caused by preS deletion mutations in chronic hepatitis B with concurrent HBsAg/anti-HBs. Although basal core promoter mutations were detected frequently in chronic hepatitis B patients with genotype C in Korea [Song et al., 2006] and also have been associated with HCC [Chen et al., 2008], this sequence could not be analyzed in the present study due to a limited quantity of serum. In conclusion, concurrent HBsAg/anti-HBs may increase independently the risk of HCC development in chronic genotype C HBV infection through the frequent occurrence of preS deletion mutations, especially in the preS1 region. In the future, prospective larger scale studies will be needed to conrm the role of concurrent HBsAg/anti-HBs as a risk factor for HCC and to clarify the function of preS deletions or other mutations in the development of HCC in chronic hepatitis B patients with concurrent HBsAg/anti-HBs.

REFERENCES
Beasley RP. 1988. Hepatitis B virus. The major etiology of hepatocellular carcinoma. Cancer 61:19421956. ckberg J, Kidd-Ljunggren K. 2003. Mutations within the hepatitis B Bla virus genome among chronic hepatitis B patients with hepatocellular carcinoma. J Med Virol 71:1823. Bruix J, Sherman M. 2005. Practice Guidelines Committee, American Association for the Study of Liver Diseases. Management of hepatocellular carcinoma. Hepatology 42:12081236. Chen CJ, Yang HI, Su J, Jen CL, You SL, Lu SN, Huang GT, Iloeje UH, REVEAL-HBV Study Group. 2006. Risk of hepatocellular carcinoma across a biological gradient of serum hepatitis B virus DNA level. J Am Med Assoc 295:6573. Chen CH, Changchien CS, Lee CM, Hung CH, Hu TH, Wang JH, Wang JC, Lu SN. 2008. Combined mutations in pre-S/surface and core promoter/precore regions of hepatitis B virus increase the risk of hepatocellular carcinoma: A case-control study. J Infect Dis 198:16341642.

J. Med. Virol. DOI 10.1002/jmv

1538
Chisari FV, Filippi P, Buras J, McLachlan A, Popper H, Pinkert CA, Palmiter RD, Brinster RL. 1987. Structural and pathological effects of synthesis of hepatitis B virus large envelope polypeptide in transgenic mice. Proc Natl Acad Sci USA 84:69096913. Chisari FV, Klopchin K, Moriyama T, Pasquinelli C, Dunsford HA, Sell S, Pinkert CA, Brinster RL, Palmiter RD. 1989. Molecular pathogenesis of hepatocellular carcinoma in hepatitis B virus transgenic mice. Cell 59:11451156. Choi MS, Kim DY, Lee DH, Lee JH, Koh KC, Paik SW, Rhee JC, Yoo BC. 2007. Clinical signicance of pre-S mutations in patients with genotype C hepatitis B virus infection. J Viral Hepatol 14:161168. Colson P, Borentain P, Motte A, Henry M, Moal V, Botta-Fridlund D, Tamalet C, Gerolami R. 2007. Clinical and biological signicance of the co-existence of HBsAg and anti-HBs antibodies in hepatitis B chronic carriers. Virology 367:3040. de Jongh FE, Janssen HL, de Man RA, Hop WC, Schalm SW, van Blankenstein M. 1992. Survival and prognostic indicators in hepatitis B surface antigen positive cirrhosis of liver. Gastroenterology 103:16301635. Dienstag JL. 1987. Concurrent hepatitis B surface antigen and antibody and the clonal selection theory of antibody diversity. Gastroenterology 93:899902. Fan YF, Lu CC, Chang YC, Chang TT, Lin PW, Lei HY, Su IJ. 2000. Identication of a pre-S2 mutant in hepatocytes expressing a novel marginal pattern of surface antigen in advanced diseases of chronic hepatitis B virus infection. J Gastroenterol Hepatol 15:519528. Fan YF, Lu CC, Chen WC, Yao WJ, Wang HC, Chang TT, Lei HY, Shiau AL, Su IJ. 2001. Prevalence and signicance of hepatitis B virus (HBV) pre-S mutants in serum and liver at different replicative stages of chronic HBV infection. Hepatology 33:277286. Fang ZL, Sabin CA, Dong BQ, Wei SC, Chen QY, Fang KX, Yang JY, Huang J, Wang XY, Harrison TJ. 2008. Hepatitis B virus pre-S deletion mutations are a risk factor for hepatocellular carcinoma: A matched nested casecontrol study. J Gen Virol 89:28822890. Gao ZY, Li T, Wang J, Du JM, Li YJ, Li J, Lu FM, Zhuang H. 2007. Mutations in preS genes of genotype C hepatitis B virus in patients with chronic hepatitis B and hepatocellular carcinoma. J Gastroenterol 42:761768. Gerber MA, Hadziyannis S, Vissoulis C, Schaffner F, Paronetto F, Popper H. 1974. Electron microscopy and immunoelectron microscopy of cytoplasmic hepatitis B antigen in hepatocytes. Am J Pathol 75:489502. Hagen TM, Huang S, Curnutte J, Fowler P, Martinez V, Wehr CM, Ames BN, Chisari FV. 1994. Extensive oxidative DNA damage in hepatocytes of transgenic mice with chronic active hepatitis destined to develop hepatocellular carcinoma. Proc Natl Acad Sci USA 91:1280812812. Hayashi J, Noguchi A, Nakashima K, Morofuji M, Kashiwagi S. 1990. Frequency of concurrence of hepatitis B surface antigen and antibody in a large number of carriers in Okinawa, Japan. Gastroenterol Jpn 25:593597. Heijtink RA, van Hattum J, Schalm SW, Masurel N. 1982. Co-occurrence of HBsAg and anti-HBs: Two consecutive infections or a sign of advanced chronic liver disease? J Med Virol 10:83 90. Hsieh YH, Su IJ, Wang HC, Chang WW, Lei HY, Lai MD, Chang WT, Huang W. 2004. Pre-S mutant surface antigens in chronic hepatitis B virus infection induce oxidative stress and DNA damage. Carcinogenesis 25:20232032. Lada O, Benhamou Y, Poynard T, Thibault V. 2006. Coexistence of hepatitis B surface antigen (HBsAg) and anti-HBs antibodies in chronic hepatitis B virus carriers: Inuence of a determinant variants. J Virol 80:29682975. Lin CL, Kao JH. 2008. Hepatitis B viral factors and clinical outcomes of chronic hepatitis B. J Biomed Sci 15:137145.

Jang et al.
Lok AS, McMahon BJ. 2007. Chronic hepatitis B. Hepatology 45:507 539. Lu HY, Zeng Z, Xu XY, Zhang NL, Yu M, Gong WB. 2006. Mutations in surface and polymerase gene of chronic hepatitis B patients with coexisting HBsAg and anti-HBs. World J Gastroenterol 12:4219 4223. Margeridon S, Lachaux A, Trepo C, Zoulim F, Kay A. 2005. A quasimonoclonal anti-HBs response can lead to immune escape of wild-type hepatitis B virus. J Gen Virol 86:16871693. Mun HS, Lee SA, Jee Y, Kim H, Park JH, Song BC, Yoon JH, Kim YJ, Lee HS, Hyun JW, Hwang ES, Kook YH, Kim BJ. 2008. The prevalence of hepatitis B virus preS deletions occurring naturally in Korean patients infected chronically with genotype C. J Med Virol 80:11891194. Naito H, Hayashi S, Abe K. 2001. Rapid and specic genotyping system for hepatitis B virus corresponding to six major genotypes by PCR using type-specic primers. J Clin Microbiol 39:362364. Poland GA, Jacobson RM. 2004. Clinical practice: Prevention of hepatitis B with the hepatitis B vaccine. N Engl J Med 351:2832 2838. Shiels MT, Taswell HF, Czaja AJ, Nelson C, Swenke P. 1987. Frequency and signicance of concurrent hepatitis B surface antigen and antibody in acute and chronic hepatitis B. Gastroenterology 93:675680. Song BC, Cui XJ, Kim H. 2005. Hepatitis B virus genotypes in Korea: An endemic area of hepatitis B virus infection. Intervirology 48:133137. Song BC, Cui XJ, Kim HU, Cho YK. 2006. Sequential accumulation of the basal core promoter and the precore mutations in the progression of hepatitis B virus-related chronic liver disease. Intervirology 49:266273. Swenson PD, Escobar MR, Carithers RL, Jr, Sobieski TJ III. 1983. Failure of preexisting antibody against hepatitis B surface antigen to prevent subsequent hepatitis B infection. J Clin Microbiol 18:305309. Tangkijvanich P, Anukulkarnkusol N, Suwangool P, Lertmaharit S, Hanvivatvong O, Kullavanijaya P, Poovorawan Y. 2000. Clinical characteristics and prognosis of hepatocellular carcinoma: Analysis based on serum alpha-fetoprotein levels. J Clin Gastroenterol 31:302308. Trepo CG, Robert D, Motin J, Trepo D, Sepetjian M, Prince AM. 1976. Hepatitis B antigen (HBSAg) and/or antibodies (anti-HBS and antiHBC) in fulminant hepatitis: Pathogenic and prognostic signicance. Gut 17:1013. Tsang TK, Blei AT, OReilly DJ, Decker R. 1986. Clinical signicance of concurrent hepatitis B surface antigen and antibody positivity. Dig Dis Sci 31:620624. Wang YJ, Lee SD, Wu JC, Lo KJ. 1991. Acute hepatitis B viral infection in a patient with anti-HBs. J Med Virol 35:133135. Wang YM, Ng WC, Lo SK. 1999. Detection of pre-S/S gene mutants in chronic hepatitis B carriers with concurrent hepatitis B surface antibody and hepatitis B surface antigen. J Gastroenterol 34:600 606. Wang HC, Huang W, Lai MD, Su IJ. 2006. Hepatitis B virus pre-S mutants, endoplasmic reticulum stress and hepatocarcinogenesis. Cancer Sci 97:683688. Weinberger KM, Zoulek G, Bauer T, Bohm S, Jilg W. 1999. A novel deletion mutant of hepatitis B virus surface antigen. J Med Virol 58:105110. Xu Z, Yen TS. 1996. Intracellular retention of surface protein by a hepatitis B virus mutant that releases virion particles. J Virol 70:133140. Zhang KY, Imazeki F, Fukai K, Arai M, Kanda T, Mikata R, Yokosuka O. 2007. Analysis of the complete hepatitis B virus genome in patients with genotype C chronic hepatitis and hepatocellular carcinoma. Cancer Sci 98:19211929.

J. Med. Virol. DOI 10.1002/jmv