Journal of Medical Virology 86:1828–1837 (2014)

Decreasing Hepatitis B Viral Load is AssociatedWith a Risk of Significant Liver Fibrosis inHepatitis B e Antigen Positive Chronic Hepatitis B

Qinxiu Xie,1 Xiangyang Hu,2 Yafei Zhang,1 Xiaoping Jiang,1 Xu Li,1* and Jiabin Li1

1Department of Infectious Disease, the First Affiliated Hospital of Anhui Medical University, Hefei, China2Department of Pathology, the First Affiliated Hospital of Anhui Medical University, Hefei, China

Alanine aminotransferase (ALT), hepatitis Bvirus (HBV) DNA levels and age are usedcommonly to assess liver histology in chronichepatitis B. Increasing levels of HBV DNA areassociated with the increasing prevalence ofsignificant fibrosis in HBeAg-negative patients.It is unclear whether these data can be appliedto HBeAg-positive patients. In present study,liver biopsies were performed and clinicalparameters were measured in 234 treatment-naive chronic HBeAg-positive patients. Theproportion of significant fibrosis in patientswith ALT 1–2�ULN was similar to in patientswith ALT more than 2�ULN (48.4% vs. 51.8%).Patients over 30 years of age (>30 years) had ahigher prevalence of significant fibrosis thanpatients 30 years of age and younger (61.0%vs. 33.6%). Negative correlation between HBVDNA levels and significant fibrosis was ob-served in patients >30 years. The optimal levelof serum HBV DNA to evaluate low risk ofsignificant fibrosis was �6.7 log10 IU/ml. Pa-tients with serum HBV DNA levels �8.5 log10IU/ml all had no significant fibrosis, however,patients with HBV DNA levels <4.7 log10 IU/mlall had significant fibrosis. Logistic regressionsshowed that age, aspartate aminotransferase,platelet count, and HBV DNA levels wereindependent predictors of significant fibrosis.In summary, older age, elevated ALT, andlower HBV DNA levels are associated withsignificant fibrosis. Decreasing levels of HBVDNA are associated with increasing prevalenceof significant fibrosis in patients >30 years.The threshold of HBV DNA levels for treatmentof HBeAg-positive patients needs to be com-bined with age. J. Med. Virol. 86:1828–1837, 2014. # 2014 Wiley Periodicals, Inc.

KEY WORDS: chronic hepatitis B; liver histol-ogy; HBV DNA levels; ALT; age

INTRODUCTION

The clinical conditions associated with the hepatitisB virus (HBV) infection are variable, ranging frominactive carrier status, active chronic hepatitis B(CHB) to the development of clinical complications,including cirrhosis and hepatocellular carcinoma(HCC). Chronic hepatitis B is prevalent in China,where most of the patients acquired HBV in theperinatal period or in their early stage of life [Laiand Yuen, 2007]. Histological injury and fibrosis havea good correlation with long-term risk, becauserepeated hepatitis flares in patients infected withhepatitis B were found to have increased necroin-flammation and fibrosis on liver biopsies, resulting inincreasing fibro genesis and eventual disease progres-sion [Mani and Kleiner, 2009]. Current antiviraltreatments may slow the disease progression incirrhotic patients and reduce hepatic events anddeaths in patients infected with hepatitis B with livercirrhosis [Du et al., 2013; Wong et al., 2013]. Severalguidelines for patients infected with hepatitis Brecommend the use of both serum alanine amino-transferase (ALT) and HBV DNA in selecting pa-tients for therapy, with a persistent ALT level ofmore than 2�upper limit of normal (ULN) [Liawet al., 2008; Chinese guideline of prevention andtreatment for chronic hepatitis B, 2011; EASL clinicalpractice guidelines for HBV infection, 2012]. Somepatients with elevated ALT levels less than two timesthe ULN (�2�ULN) levels in association with HBV

Grant sponsor: National Natural Science Foundation of China;Grant number: 81072342.

�Correspondence to: Xu Li, Department of Infectious Disease,the First Affiliated Hospital of Anhui Medical University, No 218Jixi Road, Hefei, Anhui 230022, China.E-mail: 2416194232@qq.com

Accepted 30 May 2014

DOI 10.1002/jmv.24000Published online in Wiley Online Library(wileyonlinelibrary.com).

�C 2014 WILEY PERIODICALS, INC.

DNA levels >20,000 IU/ml (or 1� 105 copies/ml) forHBeAg positive patients and >2,000 IU/ml (or 1� 104

copies/ml) for HBeAg negative patients, requiredconsideration of a liver biopsy to determine whethera significant histological disease was present, espe-cially if the patient were over 40 years age.ALT is an important biochemical marker that is

used in the assessment of hepatic injury, but it islimited by its poor correlation with disease severity.Some studies have shown minimal histological changesin HBeAg-positive patients with normal ALT, whileother studies have shown significant necroinflamma-tion, fibrosis, and even cirrhosis on liver biopsies[Kumar et al., 2008; Gui et al., 2010; Alam et al.,2011]. The serum HBV DNA levels reflect directly thedegree of HBV replication, and are correlated stronglywith long-term mortality. The risk for cirrhosis in-creases significantly with increasing HBV DNA levels,regardless of HBeAg status and ALT [Chen et al.,2006; Iloeje et al., 2007]. Increasing levels of HBVDNA are associated with an increasing prevalence ofsignificant fibrosis in HBeAg-negative patients [Croaghet al., 2010]. It is unclear whether these data can beapplied to HBeAg-positive patients. Previous studies inHBeAg-positive patients did not find HBV DNA to be apredictor of fibrosis [Shao et al., 2007; Park et al.,2008]. However, a previous study [Wang et al., 2008]in 28 HBeAg-positive immune-tolerant patients andreported that a lower serum HBV DNA levels, alongwith being over 30 years of age, was correlatedindependently with stage 2 fibrosis or more on liverbiopsy. However, that study is limited by the smallnumber of samples and only involved immune-tolerantstage patients. Few studies on threshold value of HBVDNA level for evaluating low risk of significant fibrosisin HBeAg-positive patients. The aims of the presentstudy were as follows: (1) to evaluate the associationbetween clinical parameters and significant histologicabnormalities in HBeAg-positive patients; (2) to identi-fy high risk factors of significant fibrosis and inflam-mation; (3) to determine a threshold value for the HBVDNA levels which could differentiate patients whohave progressive liver disease from those who are onlycarrying the virus. This differentiation is important foridentifying the group of patients infected with HBVwho are more likely to progress to advanced stages ofthe disease.

MATERIALS AND METHODS

Patients

The present study included 234 treatment-naiveHBeAg-positive patients infected with hepatitis Bwho were recruited to assess the status of liverhistology from January 2010 to October 2013 in theDepartment of Infectious Disease at the First Affiliat-ed Hospital of Anhui Medical University, China.Demographic, clinical and laboratory data were re-corded, including age, sex, alcohol intake, ALT level,aspartate aminotransferase (AST) level, gamma-glu-

tamyl transpeptidase (GGT), prothrombin time inter-national normalized ratio (PT-INR), albumin, Platel-ets counts (PLT), HBeAg status, HBV DNA levels atthe time of liver biopsy, and family history of HBVinfection. All patients were HBsAg-positive andHBeAg-positive for at least 12 months before enroll-ing in the study. Patients with concomitant liverdiseases, including chronic hepatitis C or D infection,Wilson’s disease, autoimmune hepatitis, primary bili-ary cirrhosis, significant alcohol intake (30 g per dayfor male, 20 g per day for female), overt cirrhosis orliver cancer, decompensated liver disease and priorantiviral treatment were excluded.

Liver Biopsy

All patients received ultrasonography-guided per-cutaneous biopsy of the right lobe of the liver. Aquick-cut and 16-gauge sheathed cutting needle(BARD, purchased from American) was used for thisprocedure. The biopsy lengths were 1.5–2.5 cm withnine or more available portal areas. The biopsy speci-mens were fixed with 10% formalin, paraffin-embed-ded, and stained with hematoxylin and eosin formorphological evaluation and Masson’s trichromestain for the assessment of fibrosis (Fig. 1). Histologicstaging of fibrosis and grading of inflammation wereperformed using the Scheuer’s classification [Scheuer,1995] and the guideline on prevention and treatmentof chronic hepatitis B in China, the grade of hepaticinflammation was G0–4, and the stage of fibrosis wasS0–4. G0–1 was considered as insignificant inflamma-tion, G2–4 as significant hepatic inflammation, S0–1as insignificant fibrosis, and S2–4 as significanthepatic fibrosis, G� 2 or S� 2 was considered start-ing points for treatment.

Laboratory Assays

Serum samples used for measurements were takenat the day of biopsy and stored at �80˚C. Allparameters were measured using standard methodol-ogies. Serum HBsAg, HBeAg, and antibody to thehepatitis B e antigen (anti-HBe) were measuredusing commercially available immunoassays (AbbottLaboratories, Chicago, IL). A serum real-time PCRtechnique was used for the quantitation of the serumHBV DNA levels, and the dectection range was1,000–300,000,000 IU/ml, samples with HBV DNAlevels higher than 3.0� 108 IU/ml were diluted at1:100 for retesting. All laboratory analysises wereperformed at the Department of Clinical Laboratoryof First Affiliated Hospital of Anhui MedicalUniversity.

Definition of Normal ALT

The ULN of serum ALT was defined as 30U/L formen and 19U/L for women. All patients identified inthe present study had ALT measured on at leastthree occasions, which were at intervals of more than

J. Med. Virol. DOI 10.1002/jmv

Liver Fibrosis and HBV DNA, ALT, Age 1829

3 months apart over a period of 12 or more monthsbefore liver biopsy. The following ranges were used:�1�ULN: persistently normal ALT; 1–2�ULN: in-termittent or continuously minimally elevated ALT;and >2�ULN patients with an elevated ALT >2�ULN once on the liver biopsy day or within 2 weeksprior to the liver biopsy, but having persistentlynormal or minimally elevated ALT (�2�ULN) levelsduring the previous 12 months.

Ethics Statement

The present study was approved by the Institution-al Review Board, the First Affiliated Hospital ofAnhui Medical University, China. All patients provid-ed written consent prior to the liver biopsy and studyentry with all clinical investigations conducted ac-cording to the principles expressed in the Declarationof Helsinki.

Statistical Analysis

All continuous variables are expressed as themedian (range). Statistical analyses were performed

using SPSS version 16.0 (SPSS, Inc., Chicago, IL)software package. The Mann–Whitney U-test wasused to compare continuous variables with a skeweddistribution and t-test was used to compare continu-ous variables with a normal distribution; the Chi-squared test and Fisher’s exact test was used forcategorical variables. A Receiver operating curve(ROC) analysis was performed to find a cut-off valuewhich could differentiate immune tolerant phasefrom immune clearance phase in HBeAg-positivepatients infected with hepatitis B. Multivariate logis-tic regression was used to identify factors that wereassociated independently with significant fibrosis orinflammation. A two-sided P value of<0.05 wasconsidered statistically significant.

RESULTS

Patients Baseline Characteristics

As stated above, 234 subjects met the inclusioncriteria and were included in the present study. Themean age was 35.8� 10.7 years (range, 14–63 years),and 74.4% (174/234) of the patients were male. Of

Fig. 1. Histological inflammation grading and fibrosis stagingwere evaluated by hematoxylin-eosin and Masson’s trichromestain method for liver biopsy specimen. A was read as havingstage 2–3 fibrosis using Masson’s trichrome stain method.Original magnification is 100�. B was read as having stage 2–3

fibrosis and grade 1 inflammation using HE stain method.Original magnification is 100�. C was read as having grade 2inflmmation using HE stained method. Original magnification is100�. D was read as having stage 1 fibrosis using Masson’strichrome stain method. Original magnification is 100�.

J. Med. Virol. DOI 10.1002/jmv

1830 Xie et al.

these, 29 (12.4%) had normal persistent ALT levels(<1�ULN), 93 (39.8%) had intermittent or continu-ously minimally elevated (1–2�ULN) ALT levels and112 (47.9%) had ALT levels >2�ULN. Patients withALT >2�ULN had significantly higher ALT(P< 0.001), AST (P< 0.001), GGT (P< 0.001), andhad higher proportion of significant inflammation(P¼ 0.006) than did those with ALT �2�ULN.However, there were no significant differences ingender, age, albumin, HBV DNA levels, PT-INR,platelet count, or significant fibrosis (all P> 0.05).The baseline characteristics of the study patients aredepicted in Table I.

Overall Liver Histology

Significant inflammation and fibrosis were found in16.2% (38/234) and 47.4% (111/234) of the total studypatients, respectively. Of these, 39 (16.7%) hadestablished cirrhosis. Significant difference were not-ed in age, HBV DNA, PT-INR, PLT, albumin, ALT,AST, GGT and proportion of significant inflammation

between �S2 group and <S2 group (all P< 0.05)(data were shown in Table II).

Correlation Between ALT and SignificantFibrosis

A strong positive correlation was noted betweenALT and inflammation and significant fibrosis(Z¼ 3.506, P< 0.001 and Z¼ 2.407, P¼ 0.016). Ac-cording to the ALT levels, patients were divided intothree groups: group 1: ALT �1�ULN, group 2: ALT1–2�ULN, and group 3: ALT> 2�ULN. In theabove three groups, the distribution of significantfibrosis was 27.6% (8/29), 48.4% (45/93), and 51.8%(58/112), respectively, and significant inflammationwas 6.9% (2/29), 10.8% (10/93), and 23.2% (26/112),respectively. The patients in group 2 had a higherproportion of significant fibrosis than did the patientsin group 1 (P¼ 0.048), but exhibited no significantdifference with group 3 (P¼ 0.628). However, thepatients in group 2 had a lower proportion of signifi-cant inflammation than did the patients in group 3

TABLE I. Baseline Characteristics of All 234 Patients and Histologic Findings (n¼ 234)

Clinical characteristics Total ALT�2�ULN ALT>2�ULN P-value�

n, % 234 (100) 122 (52.14) 112 (47.86)Age (years) (median, range) 31 (14–63) 32 (14–63) 29 (15–62) 0.144Male (n, %) 174 (74.4) 97 (79.5) 77 (68.8) 0.060Albumin (g/l) (median, range) 44.4 (28.3–55.8) 45.1 (30.3–55.8) 43.9 (28.3–52.3) 0.078GGT (U/L) (median, range) 27 (7–611) 22 (7–207) 34.5 (7–611) <0.001�ALT (U/L) (median, range) 55 (11–1743) 38 (11–60) 80 (39–1743) <0.001�AST (U/L) (median, range) 36 (14–660) 29 (14–50) 52 (23–660) <0.001�PT-INR (median, range) 1.06 (0.9–1.5) 1.04 (0.9–1.5) 1.07 (0.9–1.5) 0.099PLT (�109/l) (median, range) 160 (26–318) 154 (57–318) 162 (26–277) 0.719HBV DNA (log IU/ml) (median, range) 7.34 (3.0–8.8) 7.29 (3.0–8.8) 7.42 (3.2–8.7) 0.320Patients with significant fibrosis (n, %) 111 (47.4) 53 (43.4) 58 (51.8) 0.202Patients with significant inflammation (n, %) 38 (16.2) 12 (9.8) 26 (23.2) 0.006�

ALT, alanine aminotransferase; AST, aspartate aminotransferase; GGT, gamma-glutamyl transpeptidase; PT-INR, prothrombin timeinternational normalized ratio; PLT, platelet count; HBV DNA, hepatitis B virus deoxyribonucleic acid.Comparison was between patients with ALT �2�ULN and >2�ULN.�GGT, ALT, AST, patients with significant inflammation.

TABLE II. Comparison of Clinical Characteristics Between Patients With Significant Fibrosis and Patients WithInsignificant Fibrosis

Clinical characteristics �S2 <S2 P-value�

n, % 111 (47.4) 123 (52.6)Age (years) (median, range) 36 (15–63) 27 (14–53) <0.001�Male (n, %) 85 (48.9) 26 (43.3) 0.461Albumin (g/l) (median, range) 42.3 (28.8–55.8) 45.3 (28.3–54.8) <0.001�GGT (U/L) (median, range) 41 (9–661) 21 (7–368) 0.002�ALT (U/L) (median, range) 58.7 (17–1743) 49 (11–602) 0.044�AST (U/L) (median, range) 43 (14–660) 32 (14–207) 0.001�PT-INR (median, range) 1.1 (0.9–1.5) 1.0 (0.9–1.5) 0.001�PLT (�109/l) (median, range) 136.2 (26–265) 174.6 (54–318) <0.001�HBV DNA (log IU/ml) (median, range) 7.1 (3.0–8.58) 7.6 (4.1–8.8) <0.001�Patients with significant inflammation (n, %) 28 (25.2) 10 (8.1) <0.001�

ALT, alanine aminotransferase; AST, aspartate aminotransferase; GGT, gamma-glutamyl transpeptidase; PT-INR, prothrombin timeinternational normalized ratio; PLT, platelet count; HBV DNA, hepatitis B virus deoxyribonucleic acid.Comparison was between patients with �S2 and <S2.�S2, significant fibrosis; < S2, insignificant fibrosis.�All clinical characteristics except “male” are significant.

J. Med. Virol. DOI 10.1002/jmv

Liver Fibrosis and HBV DNA, ALT, Age 1831

(P¼ 0.020), but had no significant difference withgroup 1 (P¼ 0.729).

Correlation Between Age and SignificantFibrosis

The current study shows an increasing prevalenceof significant fibrosis with age in the total studypatients (t¼ 4.897, P< 0.001). The prevalence ofsignificant fibrosis increased from 33.6% in patients30 years and younger to 85.7% in patients older than50. Compared with patients older than 30 (>30years), patients 30 years or younger (�30 years) hada lower proportion of significant fibrosis (33.6% vs.61.0%, P< 0.001). On the contrary, patients olderthan 50 years had higher proportion of significantfibrosis and inflammation than did those in thesegroup of patients aged 50 years and younger (85.7%vs. 45.0%, P¼ 0.004 and 35.7% vs. 15.0%, P¼ 0.042,respectively). The prevalence of significant fibrosisalso increased with age in the subgroup of patientswith ALT �2�ULN (t¼ 4.027, P< 0.001; Fig. 2).

Correlation Between Serum HBV DNA Levelsand Significant Fibrosis

This study shows that the serum HBV DNA levelswere associated with significant fibrosis in both thewhole study patients (t¼ 4.559, P< 0.001) and sub-

group patients with ALT� 2�ULN (t¼ 4.997,P< 0.001). A decreasing prevalence of significantfibrosis with HBV DNA levels was noted. Accordingto the levels of HBV DNA (log10 IU/ml), the patientswere divided into five groups ranges: �5, 5–6, 6–7, 7–8, and �8. The proportion of significant fibrosis ineach group for the whole study patients was 65.2%,74.2%, 43.3%, 45.3%, and 30.9%, respectively, and76.9%, 75.0%, 47.1%, 35.9%, and 18.2%, respectivelyin the subgroup of patients with ALT� 2�ULN. Sixpatients with HBV DNA levels<4 log10 IU/ml all hadsignificant fibrosis.According to the guidelines for patients infected

with hepatitis B, the patients were divided into HBVDNA (IU/ml) �1� 105 group and >1� 105 group.Then, in each group, the patients were furtherdivided into different subgroups according to thedifferent strata ALT levels and age (Table III). Theprevalence of significant fibrosis in the whole studypatients was 65.2% (15/23) for the HBV DNA (IU/ml)�1� 105 group and 45.5% (96/211) for the HBV DNA(IU/ml) >1� 105 group (P¼ 0.072). However, basedon the different strata of ALT, further analysis wasperformed between the HBV DNA (IU/ml) �1� 105

and >1� 105 groups. The prevalence of significancedecreased with HBV DNA levels, from 75.0% to20.0% in the subgroup with ALT �1�ULN(P¼ 0.052) and from 87.5% to 44.7% in the subgroupwith ALT 1–2�ULN (P¼ 0.021). Among the sub-group patients with ALT levels �2�ULN, patientswith HBV DNA levels �1� 105 were more likely tohave significant fibrosis compared with patients withHBV DNA (IU/ml) levels >1� 105 (P¼ 0.003). Theprevalence of significant fibrosis increased with ALTlevels in the HBV DNA (IU/ml) >1� 105 group(P¼ 0.026), but it was not noted in the HBV DNA(IU/ml) �1� 105 group (P¼ 0.096).Patients were then analyzed according to different

strata of age (�30 and >30 years). A strong negativecorrelation was noted between HBV DNA levels andsignificant fibrosis in patients over 30 years of age(t¼ 4.342, P< 0.001), but no correlation was found inpatients 30 years of age and younger (t¼ 0.365,P¼ 0.716). Among patients over 30 years of age, theproportion of significant fibrosis decreased from77.8% in the HBV DNA (IU/ml) �1� 105 group to58.0% in the HBV DNA >1� 105 group (P¼ 0.126).However, among patients 30 years of age and youn-ger, the proportion of significant fibrosis increasedfrom 20.0% in the HBV DNA (IU/ml) �1� 105 groupto 34.2% in the HBV DNA >1� 105 group (P¼ 0.662).

Value of Serum HBV DNA Levels for Low Risk ofSignificant Fibrosis

The ROC curves and AUC values of serum HBVDNA levels in evaluating low risk of significantfibrosis were performed. Serum HBV DNA levelsproduced a better AUC for low risk of significantfibrosis in patients with ALT� 2�ULN and age over

Fig. 2. The proportion of significant fibrosis and inflammationin patients with different ages (years) among the total patientsand the subgroups with an ALT�2�ULN. The proportion ofsignificant fibrosis in patients aged �30, 31–40, 41–50, >50years among the total patients was 32.8%, 55.4%, 59.0%, and85.7%, respectively, and among the same subgroups with anALT �2�ULN, the proportion was 26.9%, 48.7%, 54.2%, and100%, respectively. The proportion of significant inflammationin patients aged �30, 31–40, 41–50, >50 years among the totalpatients was 12.9%, 21.5%, 10.3%, and 35.7%, respectively, andamong the subgroups with an ALT�2�ULN, it was 7.7%,15.4%, 4.2%, and 14.3%, respectively. The number of patients inthese groups was 116, 65, 39, and 14, respectively, among thetotal patients, and it was 52, 39, 24, and 7, respectively, amongthe subgroups with an ALT�2�ULN. �S2, significant fibrosis;�G2, significant inflammation.

J. Med. Virol. DOI 10.1002/jmv

1832 Xie et al.

30 years (AUC 0.820, P< 0.001, 95% CI: 0.720–0.919)compared to the whole study patients with ALT� 2�ULN (AUC 0.770, P< 0.001, 95% CI: 0.686–0.854).Serum HBV DNA levels did not have any evaluatingvalue for low risk of significant inflammation (AUC0.580 and AUC 0.520 for whole patients and patientswith ALT� 2�ULN, respectively). The sensitivity,specificity and predictive values of different HBVDNA levels in patients with ALT� 2�ULN and ageover 30 years are shown in Table IV. Based on theYouden Index, the optimal level of serum HBV DNAto evaluate low risk of significant fibrosis was �6.7log10 IU/ml (Youden Index 0.557, sensitivity 83.9%,specificity 71.8%). Patients with serum HBV DNAlevels �8.5 log10 IU/ml all had no significant fibrosis,however, patients with HBV DNA levels �4.7 log10IU/ml all had significant fibrosis.

Univariate and Multivariate Analysis forSignificant Fibrosis

Univariate analysis showed that in HBeAg-positivepatients, age, albumin, AST, PT-INR, platelet count,GGT and HBV DNA levels were associated withsignificant fibrosis, and that ALT, AST, GGT, albu-min, PT-INR, and platelet count (PLT) were associat-

ed with significant inflammation. Gender was notassociated with either significant fibrosis or inflam-mation. Multivariate analysis showed that in thewhole study patients, older age, lower HBV DNAlevels, lower PLT and higher AST were associatedindependently with significant fibrosis, while higherAST and PLT were associated independently withsignificant inflammation. In patients with ALT �2�ULN, an older age (P¼ 0.043), higher AST (P< 0.011)and lower HBV DNA (P¼ 0.001) were associatedindependently with significant fibrosis, while nofactor was associated independently with significantinflammation. The univariate and multivariate analy-sis of clinical parameters associated with significantinflammation or fibrosis are shown in Table V.

DISCUSSION

This study demonstrates that progressive liverdisease in HBeAg-positive patients is driven by threemain factors: which is age, a surrogate of diseaseduration; immune activation, as measured by ALT;and viral replication, as measured by HBV DNAlevels. ALT levels are used commonly to assess liverdisease, and elevated levels have been shown to beassociated with active liver disease on histology while

TABLE III. The Distributions of Significant Fibrosis and Inflammation in Different Level of HBV DNA According toDifferent Strata of ALT and Age (n¼234)

Characteristics

HBV DNA (IU/ml)

P-value�1�105 (n, %) >1� 105 (n, %)

ALT (U/L) n �S2 �G2 n �S2 �G2�1�ULN 4 3 (75.0)a 0 25 5 (20.0) 2 (8.0) 0.052,a 0.021b

1–2�ULN 8 7 (87.5)b 0 85 38 (44.7)c 10 (11.8) 0.026,c 0.026d

>2�ULN 10 5 (50.0) 3 (30) 102 53 (52.0) 23 (22.5)d

Age (years)�30 5 1 (20.0) 2 (40.0) 111 38 (34.2) 13 (11.7) 0.033,e 0.001f

>30 18 14 (77.8)e 1 (5.6) 100 58 (58.0)f 22 (22.0)g 0.045g

ALT, alanine aminotransferase; ULN, upper limit of normal; HBV DNA, hepatitis B virus deoxyribonucleic acid; n, number of patients;�S2, significant fibrosis; �G2, significant inflammation.aComparison was between patients with serum HBV DNA (log10 IU/ml) �1� 105 and >1� 105 among patients with ALT �1�ULN.bComparison was between patients with serum HBV DNA (log10 IU/ml) �1� 105 and >1� 105 among patients with ALT 1–2�ULN.cComparison was between patients with ALT 1–2�ULN and �1�ULN among patients with serum HBV DNA (log10 IU/ml) >1� 105.dComparison was between patients with ALT 1–2�ULN and >2�ULN among patients with serum HBV DNA (log10 IU/ml) >1� 105.eComparison was between patients with age >30 years and �30 years among each group.fComparison was between patients with age >30 years and �30 years among each group.gComparison was between patients with age >30 years and �30 years among each group.

TABLE IV. Sensitivity, Specificity and Predictive Values of Different Serum HBV DNA Levels for Predicting InsignificantFibrosis Among Patients With ALT�2�ULN and Age >30 Years (n¼70)

HBV DNA (log10 IU/ml) n Sensitivity Specificity PPV NPV LRþ LR��6.00 45 83.9% 64.1% 65.00% 86.21% 2.34 0.251�6.70 37 83.9% 71.8% 70.27% 84.85% 2.98 0.224�7.00 35 80.6% 71.8% 69.44% 82.35% 2.86 0.270�7.50 25 58.1% 82.1% 72.00% 71.11% 3.25 0.510�8.00 17 45.2% 92.3% 82.35% 67.92% 5.87 0.594�8.47 4 9.7% 100% 100% 59.09% — 0.903

Insignificant fibrosis: Stage of fibrosis <S2.ALT, alanine aminotransferase; ULN, upper limit of normal; HBV DNA, hepatitis B virus deoxyribonucleic acid; PPV, positive predictivevalue; NPV, negative predictive value; LRþ, positive likelihood ratio; LR�, negative likelihood ratio.

J. Med. Virol. DOI 10.1002/jmv

Liver Fibrosis and HBV DNA, ALT, Age 1833

normal ALT is shown to be associated with inactivehistology [Sarin and Kumar, 2008]. Nevertheless,studies have shown that ALT levels may not alwaysshow significant correlations with liver fibrosis orinflammation [Prati et al., 2002; Lai et al., 2007].Several guidelines for patients infected with hepatitisB require meeting a certain ALT threshold (>2�ULN) for treatment decisions. Treatment is optionalfor patients with ALT less than 2�ULN, and liverbiopsy is recommended. Using such an ALT thresh-old is open to debate, especially when previousstudies have shown that patients infected withhepatitis B with a normal ALT at the upper rangehave an increased risk of long-term cirrhotic compli-cations and HCC [Yuen et al., 2005a; Kumada et al.,2010]. In present study, an association was foundbetween ALT and significant fibrosis and inflamma-tion. There was a high proportion of significantfibrosis (27.6%) in patients with a normal ALT inpresent study, even using the ULN criteria revisedby Prati (30U/L for men and 19U/L for women)[Prati et al., 2002]. This finding is consistent withthose of a previous study [Kumar et al., 2008]. Inother words, some patients with a serum ALT atpresent normal ranges may, in fact, be in theimmune clearance phase, thus leading to significantfibrosis on liver histology. The proportion of signifi-cant fibrosis in patients with ALT 1–2�ULN similarto ALT >2�ULN, was higher than that in patientswith normal ALT. Therefore, using the ALT thresh-old of more than 2�ULN in selecting patients fortreatment has the potential of not treating patients

with significant fibrosis or cirrhosis. The presentstudy results show that a single serum ALT alonecan give a false impression. About one-third of thepatients may be missed when evaluating advancedliver disease, even when using the Prati criteria fornormal ALT. Therefore, these findings suggestedthat the present normal ranges for ALT should belowered or that antiviral therapy should be initiatedwhen the ALT exceeds normal (>1�ULN), but doesnot exceed 2�ULN. In fact, previous studies havesuggested that liver biopsy was recommended inpatients with high-normal ALT (20–40U/L) [Funget al., 2008; Gui et al., 2010; Lee et al., 2010; Baiet al., 2013]. Because significant fibrosis derivedfrom long-term liver inflammation, ALT may bemore associated with liver inflammation, but notfibrosis, and the current study confirmed this find-ing. Hence, further longitudinal studies should beperformed to determine the exact relationship be-tween ALT and fibrosis.The role of age in evaluating liver damage was

investigated. Increasing age was associated indepen-dently with significant fibrosis in HBeAg-positivepatients, which is consistent with previous studies[Fung et al., 2008; Kumar et al., 2008; Croagh et al.,2010; Lee et al., 2010]. Compared with patients30 years of age or younger, patients over 30 years ofage had more significant fibrosis. This finding mayexplained that most of patients 30 years of age oryounger are in an immune-tolerance phase. This issimilar to a previous study [Seto et al., 2012], inwhich patients over 30 years of age had a higher

TABLE V. Analysis of Predictors for Significant Fibrosis and Inflammation in HBeAg-Positive Patients by LogisticRegression (n¼ 234)

Variable B Wald OR 95% CI P-value

Univariate analysis of predictors for significant fibrosisHBV DNA �0.491 17.985 0.612 0.468–0.768 <0.001Age 0.064 20.151 1.066 1.037–1.097 <0.001PT-INR 4.174 11.068 64.998 5.557–760.193 0.001PLT �0.014 22.05 0.986 0.981–0.992 <0.001ALB �0.123 16.201 0.885 0.833–0.939 <0.001AST 0.011 8.821 1.011 1.004–1.019 0.003GGT 0.013 13.24 1.013 1.006–1.020 <0.001

Univariate analysis of predictors for significant inflammationPT-INR 5.038 13.626 154.175 10.624–2237.42 <0.001PLT �0.012 10.477 0.988 0.988–0.995 0.001ALB �0.085 5.955 0.919 0.858–0.983 0.015ALT 0.005 11.361 1.005 1.002–1.007 0.001AST 0.012 16.095 1.012 1.006–1.018 <0.001GGT 0.007 9.154 1.007 1.003–1.012 0.002

Multivariate analysis of predictors for significant fibrosisAge 0.048 8.875 1.049 1.017–1.083 0.003HBV DNA �0.267 4.063 0.766 0.551–0.993 0.044PLT �0.009 8.301 0.991 0.985–0.997 0.004AST 0.011 8.539 1.011 1.004–1.018 0.003

Multivariate analysis of predictors for significant inflammationPT-INR 3.368 4.287 29.23 1.197–703.808 0.038AST 0.012 14.74 1.012 1.006–1.018 <0.001

HBV DNA, hepatitis B virus deoxyribonucleic acid; PT-INR, prothrombin time international normalized ratio; PLT, platelet; ALB, albumin;AST, aspartate aminotransferase; GGT, gamma-glutamyl transpeptidase; ALT, alanine aminotransferase; OR, odd ratio; 95% CI, 95%confidence interval.

J. Med. Virol. DOI 10.1002/jmv

1834 Xie et al.

degree of histological changes compared with those30 years of age or younger. On the contrary, a higherproportion of fibrosis was observed in the group over50 years of age than in the group who were 50 yearsof age or younger. This finding is in line withprevious studies showing that older patients weremore likely to have significant fibrosis [Lai et al.,2007; Gui et al., 2010]. This may be explained by thefact that Chinese patient who acquired HBV duringthe perinatal period or in their early life, experiencedfluctuating levels of viral replication with recurrentdisease flares and remission, which are believed tocontribute to the progression of liver disease overtime. Several longitudinal studies have demonstratedthat the incidence of cirrhosis and HCC increaseswith age, especially in patients older than 40 [Realdiet al., 1994; Chu et al., 2004; Yu et al., 1997]. Thus,according to this study, liver biopsy should be consid-ered in patients older than 30 and ALT abnormal orwith patients older than 50, regardless of ALT.HBV DNA measurements are used to evaluate

disease activity, assess the efficiency of antiviraltherapy and predict treatment outcomes [Sarin andKumar, 2008]. Several reports have shown a positiverelationship between HBV DNA levels and liverfibrosis or inflammation among HBeAg-negative pa-tients [El-Zayadi et al., 2009; Wang et al., 2013].Previous studies [Shao et al., 2007; Park et al., 2008;Bai et al., 2013] showed that no relationship betweenHBV DNA levels and liver fibrosis among HBeAg-positive patients. Other studies have shown thateven among subjects with low HBV DNA levels, HBVcomplications and liver damage may arise, indepen-dent from other disease factors, especially amongpersons who acquire the virus early in their lives[Yuen et al., 2005b]. Present study showed thatsignificant fibrosis was correlated negatively withHBV DNA levels, which is consistent with a previousstudy of HBeAg-positive patients [Croagh et al.,2010]. Present study showed that patients with HBVDNA levels<4 log10 IU/ml all had significant fibrosis,whereas the prevalence of significant fibrosis de-creased to 18.2% in patients with HBV DNA level�8 log10 IU/ml. This finding is in line with those of aprevious study. In that study, 60.0% of patients withHBV DNA< 5 log10 IU/ml had significant histology inHBeAg-positive patients [Park et al., 2008]. Similar-ly, the results of Kumar et al [Kumar et al., 2008]showed that approximately 21.0% of patients withnormal ALT and HBV DNA levels <1� 105 IU/ml hadsignificant inflammation and that approximately50.0% had fibrosis. Negative correlation betweenHBV DNA levels and significant fibrosis was found inpatients over 30 years of age, but no correlation wasfound in patients 30 years of age or younger. In otherwords, patients with lower HBV DNA level may havesevere fibrosis with age increasing. The median ageof these patients with HBV DNA levels <4.7 log10 IU/ml in the present study conformed this. This can beexplained by an extensive and immune-mediated

repeatedly response leading to low viremic levels andliver damage. A more comprehensive understandingof the immune-mediated response in patients infectedwith hepatitis B, T cell mediated antiviral andcytolytic function or non-cytolytic function participatethe hepatic inflammation and hepatitis progressionand viral load suppression. Over time, increasedimmune activation results in ALT “flares” and subse-quent reductions in serum HBV DNA levels [Lok,2002]. These findings suggest that the maximalsuppression of HBV DNA as early as possible mayreduce the risk of disease progression. In addition,this finding suggests that the current threshold ofHBV DNA >5 log10 IU/ml for treatment regardless ofage in HBeAg-positive patients is inappropriate.Lower HBV DNA level was the risk of significantfibrosis in patients over 30 years of age. No correla-tion was found between ALT levels and liver fibrosisin patients with HBV DNA levels <1� 105 IU/ml inthis study. In other words, patients with HBV DNAlevels <1� 105 IU/ml may have significant fibrosiseven with normal ALT levels. Thus, we can initiateliver fibrosis assessment or antiviral treatment inpatients with HBV DNA levels <1� 105 IU/ml regard-less their ALT level. The present study also foundthe optimal serum HBV DNA cut-off value to evalu-ate low risk of significant fibrosis was �6.7 log10 IU/ml, but its predictive value (PPV) and negativepredictive value (NPV) were not ideal. However,patients with serum HBV DNA levels �8.5 log10 IU/ml all had no significant fibrosis. In other words,patients with HBV DNA levels �8.5 log10 IU/ml andALT� 2�ULN can be observed closely without need-ing a liver biopsy, regardless of their age. Thisfinding is inconsistent with a previous study [Bai etal., 2013], in that study, liver fibrosis correlatedpositively with HBV DNA levels among patients 35years of age or more in HBeAg-positive patients. Weassume that the contradictory conclusion of thatstudy may be caused by patients with different stageof immune. Most of patients with HBV DNA levels�8 log10 IU/ml were in immune-tolerant stage in thisstudy. Thus, this cut-off value of HBV DNA for lowrisk of significant fibrosis needs additional studies tovalidate its clinical applicability.The univariate analysis showed that age, albumin,

AST, PT-INR, platelet count, GGT and HBV DNAlevels were associated with significant fibrosis, butthe multivariate analysis showed that only age, AST,PLT and serum HBV DNA levels were independentfactors for the evaluating of significant fibrosis. Itmay be explained that the risk of liver histology inchronic hepatitis B is the complex result of manyinteracting factors. Thus, we propose that it is bestconsidered in the context of the traditional phase ofdisease model rather than a simple function of anysingle factor alone.There are several limitations in this study, includ-

ing its cross-sectional evaluation, non-Asian patientsinfected with chronic hepatitis B, the lack of a HBV

J. Med. Virol. DOI 10.1002/jmv

Liver Fibrosis and HBV DNA, ALT, Age 1835

genotype, core-promoter mutation, the course of HBVinfection and serum complement, which is to beassociated with significant liver histology abnormali-ties [Chu et al., 2004; Yuen et al., 2005a,b; Bugdaciet al., 2011]. In addition, the HBsAg levels were notevaluated in present study, although prior studieshave shown that high HBsAg levels can predictinsignificant fibrosis in HBeAg-positive patients[Shao et al., 2007]. Because serum clinical param-eters tend to fluctuate with time, the measurement ofeach clinical parameter at several time points mightbe more representative. Hence, a focused approachusing HBeAg status, age, ALT, AST, PT, alpha fetalprotein (AFP), viral load and ethnic origin withlongitudinal assessment of liver histology should bestudied in patients with ALT less than 2�ULN toclassify appropriately patients and select patientswho require an assessment of liver fibrosis to informtreatment decisions. Additionally, it must be ac-knowledged that the biopsy group presented here isbiased toward patients with a higher likelihood ofsignificant fibrosis because clinicians used markers ofhigher ALT and signs of chronic liver disease onexamination to guide the decision to conduct a biopsy.Thus, these risk of factors for liver fibrosis need tofurther studies to validate its clinical applicability.In conclusion, serum ALT, serum HBV DNA levels

and age were associated with significant fibrosis inHBeAg-positive patients. Decreasing levels of HBVDNA are associated with increasing prevalence ofsignificant fibrosis in patients over 30 years of age.The threshold of HBV DNA levels for treatment needto be combined with age. We can initiate liver fibrosisassessment or antiviral treatment in patients withHBV DNA levels <1� 105 IU/ml regardless their ALTlevel. In contrast, patients with an HBV DNA levels>8.5 log10 IU/ml and an ALT levels �2�ULN maynot require immediate liver biopsy and treatment butonly close follow up, regardless of age.

ACKNOWLEDGMENTS

This project was supported by a grant from theNational Natural Science Foundation of China, No.81072342. We thank for Yibo Shao, Mei Wang,Zhenhua Zhang and Yingguang Fan for contributingto the writing of this manuscript.

REFERENCES

Alam S, Ahmad N, Mustafa G, Shrestha A, Alam AK, Khan M.2011. Evaluation of normal or minimally elevated alaninetransaminase, age and DNA level in predicting liver histologicalchanges in chronic hepatitis B. Liver Int 31:824–830.

Bai H, Liu H, Chen X, Xu C, Dou X. 2013. Influence of age andHBeAg status on the correlation between HBV DNA and hepaticinflammation and fibrosis in chronic hepatitis B patients. DigDis Sci 58:1355–1362.

Bugdaci MS, Alkim C, Karaca C, Kesici B, Bayraktar B, Sokmen M.2011. Could complement C4 be an alternative to biopsy forchronic hepatitis B histopathologic findings? J Clin Gastro-enterol 45:449–455.

Chen CJ, Yang HI, Su J, Jen CL, You SL, Lu SN, Huang GT, IloejeUH. 2006. Risk of hepatocellular carcinoma across a biological

gradient of serum hepatitis B virus DNA level. JAMA 295:65–73.

Chinese Society of Hepatology, Chinese Society of Infectious Dis-ease, Chinese Medical Association. 2011. The guideline ofprevention and treatment for chronic hepatitis B (2010 version).Chin J Hepatol 19:13–24.

Chu CM, Hung SJ, Lin J, Tai DI, Liaw YF. 2004. Natural history ofhepatitis B e antigen to antibody seroconversion in patients withnormal serum aminotransferase levels. Am J Med 116:829–834.

Croagh CM, Bell SJ, Slavin J, Kong YX, Chen RY, Locarnini S,Desmond PV. 2010. Increasing hepatitis B viral load is associat-ed with risk of significant liver fibrosis in HBeAg-negative butnot HBeAg-positive chronic hepatitis B. Liver Int 30:1115–1122.

Du X, Wang J, Shao L, Hu X, Yang C, Shen L, Weng X, Zhang W.2013. Histological improvement of long-term antiviral therapy inchronic hepatitis B patients with persistently normal alanineaminotransferase levels. J Viral Hepat 20:328–335.

European Association for the Study of the Liver. 2012. EASLclinical practice guidelines: Management of chronic hepatitis Bvirus infection. J Hepatol 57:167–185.

El-Zayadi AR, Badran HM, Saied A, Shawky S, Attia Mel-D, ZalataK. 2009. Evaluation of liver biopsy in Egyptian HBeAg-nagativechronic hepatitis B patient at initial presentation: Implicationsfor therapy. Am J Gastroenterol 104:906–911.

Fung J, Lai CL, But D, Wong D, Cheung TK, Yuen MF. 2008.Prevlence of fibrosis and cirrhosis in chronic hepatitis B:Implications for treatment and management. Am J Gastro-enterol 103:1421–1426.

Gui HL, Wang H, Yang YH, Wu YW, Zhou HJ, Guo SM, Lin LY,Wang L, Cai W, Chen R, Guo Q, Zhou XQ, Bao SS, Xie Q. 2010.Significant histopathology in Chinese chronic hepatitis B pa-tients with persistently high-normal alanine aminotransferase. JViral Hepat 1:44–50.

Iloeje UH, Yang HI, Jen CL, Su J, Wang LY, You SL, Chen CJ.2007. Risk and predictors of mortality associated with chronichepatitis B infection. Clin Gastroenterol Hepatol 5:921–931.

Kumada T, Toyoda H, Kiriyama S, Sone Y, Tanikawa M, HisanagaY, Kanamori A, Atsumi H, Takagi M, Arakawa T, Fujimori M.2010. Incidence of hepatocellular carcinoma in patients withchronic hepatitis B virus infection who have normal alanineaminotransferase values. J Med Virol 82:539–545.

Kumar M, Sarin SK, Hissar S, Pande C, Saksuja P, Sharma BC,Shauhan R, Bose S. 2008. Virologic and histologic features ofchronic hepatitis B virus-infected asymptomatic patients withpersistently normal ALT. Gastroenterology 134:1376–1384.

Lai CL, Yuen MF. 2007. The natural history and treatment ofchronic hepatitis B: A critical evaluation of standard treatmentcriteria and end points. Ann Intern Med 147:58–61.

Lai M, Hyatt BJ, Nasser I, Curry M, Afdhal NH. 2007. The clinicalsignificance of persistently normal ALT in chronic hepatitis Binfection. J Hepatol 47:760–767.

Lee JK, Shim JH, Lee HC, Lee SH, Kim KM, Lim YS, Chung YH,Lee YS, Suh DJ. 2010. Estimation of the healthy upper limitsfor serum alanine aminotransferase in Asian populations withnormal liver histology. Hepatology 51:1577–1583.

Liaw YF, Leung N, Kao JH, Piratvisuth T, Gane E, Han KH, GuanR, Lau GK, Locamini S. 2008. Asian-Pacific consensus statementon the management of chronic hepatitis B: A 2008 update.Hepatol Int 2:263–283.

Lok AS. 2002. Chronic hepatitis B. N Engl J Med 346:1682–1683.

Mani H, Kleiner DE. 2009. Liver biopsy findings in chronichepatitis B. Hepatology 49:S61–S71.

Park JY, Park YN, Kim DY, Paik YH, Lee KS, Moon BS, Han KH,Chon CY, Ahn SH. 2008. High prevalence of significant histologyin asymptomatic chronic hepatitis B patients with genotype Cand high serum HBV DNA levels. J Viral Hepat 15:615–621.

Prati D, Taioli E, Zanella A, Della Torre E, Butelli S, Del VecchioE, Vianello L, Zanuso F, Mozzi F, Milani S, Conte D, ColomboM, Sirchia G. 2002. Updated definitions of healthy ranges forserum alanine aminotransferase levels. Ann Intern Med 137:1–10.

Realdi G, Fattovich G, Hadziyannis S, Schalm SW, Almasio P,Sanchez-Tapias J, Christensen E, Giustina G, Noventa F. 1994.Survival and prognostic factors in 366 patients with compensat-ed cirrhosis type B: A multicenter study. The Investigators ofthe European Concerted Action on Viral Hepatitis (EUROHEP).J Hepatol 21:656–666.

J. Med. Virol. DOI 10.1002/jmv

1836 Xie et al.

Sarin SK, Kumar M. 2008. Should chronic HBV infected patientswith normal ALT treated: Debate. Hepatol Int 2:179–184.

Scheuer PJ. 1995. The nomenclature of chronic hepatitis: Time fora change. J Hepatol 22:112–114.

Seto WK, Lai CL, Ip PP, Fung J, Wong DK, Yuen JC, Hung IF,Yuen MF. 2012. A large population histology study showing thelack of association between ALT elevation and significantfibrosis in chronic hepatitis B. PLOS ONE 7:e32622.

Shao J, Wei L, Wang H, Sun Y, Zhang LF, Li J, Dong JQ. 2007.Relationship between hepatitis B virus DNA levels and liverhistology in patients with chronic hepatitis B. World J Gastro-enterol 13:2104–2107.

Wang CC, Lim LY, Deubner H, Tapia K, Lau AW, Manansala J,Krows M, Shuhart MC, Kowdley KV. 2008. Factors predictiveof significant hepatic fibrosis in adults with chronic hepatitisB and normal serum ALT. J Clin Gastroenterol 42:820–826.

Wang Y, Xu MY, Zheng RD, Xian JC, Xu HT, Shi JP, Li SB, Qu Y,Dong YW, Lu LG. 2013. Prediction of significant fibrosis and

cirrhosis in hepatitis B e-antigen negative patients with chronichepatitis B using routine parameters. Hepatol Res 43:441–451.

Wong GL, Chan HL, Mak CW, Lee SK, Ip ZW, Lam AT, Iu HW,Leung JM, Lai JW, Lo AO, Chan HY, Wong VW. 2013.Entecavir treatment reduces hepatic events and deaths inchronic hepatitis B patients with liver cirrhosis. Hepatology 6.doi: 10.1002/hep.26301 [Epub ahead of print].

Yu MW, Hsu FC, Sheen IS, Chu CM, Lin DY, Chen CJ, Liaw YF.1997. Prospective study of hepatocellular carcinoma and livercirrhosis in asymptomatic chronic hepatitis virus carriers. Am JEpidemiol 145:1039–1047.

Yuen MF, Yuan HJ, Wong DK, Yuen JC, Wong WM, Chan AO,Wong BC, Lai KC, Lai CL. 2005a. Prognostic determinants forchronic hepatitis B in Asians: Therapeutic implications. Gut54:1610–1614.

Yuen MF, Tanaka Y, Ng IQ, Mizokami M, Yuen JC, Wong DK,Yuan HJ, Sum SM, Chan AO, Lai CL. 2005b. Hepatic necroin-flammation and fibrosis in patients with genotypes Ba and C,core-promoter and precore mutations. J Viral Hepal 12:513–518.

J. Med. Virol. DOI 10.1002/jmv

Liver Fibrosis and HBV DNA, ALT, Age 1837