Non-linear association between liver fibrosis scores and viral load in patients with chronic hepatitis B

Article information

Clin Mol Hepatol. 2024;30(4):793-806

Publication date (electronic) : 2024 July 19

doi : https://doi.org/10.3350/cmh.2024.0252

Gi-Ae Kim ¹^,^*, Seung Won Choi ²^,^*, Seungbong Han ³, Young-Suk Lim^,⁴

¹Department of Internal Medicine, Kyung Hee University School of Medicine, Seoul, Korea

²University of Ulsan College of Medicine, Seoul, Korea

³Department of Biostatistics, Korea University, Seoul, Korea

⁴Department of Gastroenterology, Liver Center, Asan Medical Center, University of Ulsan College of Medicine, Seoul, Korea

Corresponding author : Young-Suk Lim Department of Gastroenterology, Asan Medical Centre, University of Ulsan College of Medicine, 88 Olympic-ro 43-gil, Songpa-gu, Seoul 05505, Korea Tel: +82-2-3010-5933, Fax: +82-2-485-5782, E-mail: limys@amc.seoul.kr

*GA Kim and SW Choi contributed equally to this work and deserve co-first authorship.

Editor: Chang Wook Kim, The Catholic University of Korea, Korea

Received 2024 April 12; Revised 2024 July 13; Accepted 2024 July 18.

Abstract

Background/Aims

Serum hepatitis B virus (HBV) DNA levels and non-invasive liver fibrosis scores are significantly associated with hepatocellular carcinoma (HCC) risk in chronic hepatitis B (CHB) patients. Nonetheless, the relationship between HBV DNA levels and liver fibrosis scores is unclear.

Methods

A historical cohort comprising 6,949 non-cirrhotic Korean CHB patients without significant alanine aminotransferase elevation was investigated. The association of HBV DNA levels with the aspartate aminotransferase to platelet ratio index (APRI) and fibrosis (FIB)-4 score at baseline was analyzed using general linear models.

Results

In HBeAg-negative patients (n=4,868), HBV DNA levels correlated linearly with both APRI and FIB-4 scores. In contrast, in HBeAg-positive patients (n=2,081), HBV DNA levels correlated inversely with both APRI and FIB-4 scores. Across the entire cohort, a significant non-linear parabolic relationship was identified between HBV DNA levels and fibrosis scores, independent of age and other covariates. Notably, moderate viral loads (6–7 log₁₀ IU/mL) corresponded to the highest APRI and FIB-4 scores (P<0.001). Over a median 10-year follow-up, 435 patients (6.3%) developed HCC. Higher APRI scores ≥0.5 and FIB-4 scores ≥1.45 were significantly associated with elevated HCC risk (P<0.001 for both). HBV DNA level remained a significant predictive factor for HCC development, even after adjusting for APRI or FIB-4 scores.

Conclusions

HBV viral load is significantly correlated with APRI and FIB-4 scores, and is also associated with HCC risk independent of those scores in CHB patients. These findings suggest that HBV DNA level is associated with hepatocarcinogenesis through both direct and indirect pathways.

Keywords: Hepatitis B virus; Hepatocellular carcinoma; Liver cancer; Liver fibrosis

Graphical Abstract

INTRODUCTION

Chronic hepatitis B virus (HBV) infection (CHB) is the most common chronic viral infection worldwide, with an estimated global prevalence of 3.9%, corresponding to 296 million infected individuals [1]. CHB is the most common cause of primary liver cancer, which is the third leading cause of cancer-related mortality worldwide [2]. Hepatocellular carcinoma (HCC) accounts for more than 90% of all primary liver cancers [1-3].

Serum HBV DNA levels are significantly associated with the risk of HCC in CHB patients. A prospective cohort (REVEAL) study showed that baseline serum HBV DNA level up to 10⁶ copies/mL (approximately 5 log₁₀ IU/mL) is linearly associated with the risk of HCC [4]. Emerging studies indicate that the pattern of association between a comprehensive range of baseline HBV DNA levels (up to 9 log₁₀ IU/mL) and HCC risk is non-linear parabolic [5-7]. Moderate HBV DNA level around 6 log₁₀ IU/mL was associated with the highest HCC risk in both untreated and treated CHB patients. However, the mechanisms underlying these observations were not well elucidated.

Non-invasive liver fibrosis scores are also significantly associated with HCC risk in CHB patients [8,9]. The aspartate aminotransferase (AST) to platelet ratio index (APRI) and fibrosis-4 (FIB-4) scores are the most commonly used liver fibrosis scores in clinical practice, since they are widely available, affordable, and reliable [10]. Nonetheless, there is virtually no data on the correlation between these non-invasive liver fibrosis scores and serum HBV DNA levels.

Therefore, in this large-scale cohort study, our objective was to explore the correlation between HBV DNA levels and non-invasive liver fibrosis markers in non-cirrhotic CHB patients who are not candidates for antiviral treatment due to the absence of significant alanine aminotransferase (ALT) elevation. Additionally, we aimed to analyze the association between baseline HBV DNA levels and HCC risk during follow-up, after adjustment for liver fibrosis scores. This study is expected to provide insights into the intricate interplay between serum HBV DNA levels, liver fibrosis, and HCC risk. The findings may improve our ability to better understand and stratify HCC risk in CHB patients, thereby refining antiviral treatment strategies.

MATERIALS AND METHODS

Study population

The source population was a historical cohort of 10,897 treatment-naïve adult patients with non-cirrhotic CHB registered between January 2000 and December 2013 at Asan Medical Center, a 2,700-bed tertiary referral hospital located in Seoul, Korea (Fig. 1). The cohort comprised both HBeAg-positive and HBeAg-negative patients who met the following criteria: (1) age over 20 years, (2) no evidence of cirrhosis, (3) serum ALT levels <2× the upper limit of normal (ULN), (4) no history of malignancy or organ transplantation, and (5) follow-up for >1 year. All patients tested positive for the serum hepatitis B surface antigen (HBsAg) for >6 months.

Figure 1.

Patient disposition. ULN, upper limit of normal; ALT, alanine aminotransferase; CHB, chronic hepatitis B; HBV, hepatitis B virus; HCV, hepatitis C virus; HDV, hepatitis D virus; HIV, human immunodeficiency virus.

Cirrhosis was defined as the presence of any of the following features: coarse liver echotexture or nodular liver surface on ultrasonography, clinical features of portal hypertension (splenomegaly, ascites, or varices), thrombocytopenia (<100,000/mm³), or histological evidence of cirrhosis. The ULN for ALT was established as 40 U/L for both men and women. The exclusion criteria were as follows: (1) seropositivity for hepatitis C virus (HCV), hepatitis D virus, and human immunodeficiency virus; (2) history of anti-HBV therapy or immunosuppressive treatment; and (3) ALT elevation >2x ULN within the first year from baseline.

Finally, the study population comprised a total of 6,949 patients (2,081 HBeAg-positive and 4,868 HBeAg-negative) who were included in the analyses (Fig. 1).

This study was approved by the institutional review board of Asan Medical Center.

Clinical and laboratory variables

The baseline clinical information and outcomes were systematically extracted from the electronic medical records of Asan Medical Center. Clinical assessments, liver function tests, assays for HBeAg, anti-HBe, and serum HBV DNA levels were performed at baseline and every 3–6 months. Serological markers including HBsAg, anti-HBs, HBeAg, anti-HBe, and anti-HCV were measured using enzyme immunoassays (Abbott Laboratories, Chicago, IL, USA). Serum HBV DNA was measured using real-time polymerase chain reaction (linear dynamic detection range, 10 IU/mL– 1×109 IU/mL; Abbott Laboratories). Given that >98% of Korean CHB patients have HBV genotype C, we assumed that our study population had HBV genotype C.

The baseline or index date was determined as the first day of gathering all necessary data regarding the patients’ demographics, serologic markers, chemical analyses, serum HBV DNA levels, and platelet counts. The APRI and FIB-4 scores were calculated at baseline using the following formulae: APRI=[AST (U/L)/ULN]/[platelet count (109/L)]×100; FIB-4=[age (years)×AST (U/L)]/[platelet count (10⁹/L)×ALT (U/L)]1/2 [11,12]. A cutoff value of 0.5 of the APRI score was used to categorize patients into the high and low APRI groups. This cutoff is reported to have a sensitivity and specificity of 73.0% and 54.5%, respectively for advanced hepatic fibrosis [13]. A cutoff value of 1.45 of the FIB-4 score was used to categorize patients into the high and low FIB-4 score groups. This cutoff is reported to have a sensitivity and specificity of 62.7% and 55.7%, respectively, for advanced hepatic fibrosis [13].

Assessment

The primary assessment was the association of serum HBV DNA levels with the platelet count, APRI, and FIB-4 scores at baseline. The secondary assessment was the adjusted association of the platelet count, APRI, FIB-4 scores, and HBV DNA levels with HCC risk. The serum HBV DNA levels were categorized into ≤3 log₁₀ IU/mL, 3–4 log₁₀ IU/mL, 4–5 log₁₀ IU/mL, 5–6 log₁₀ IU/mL, 6–7 log₁₀ IU/mL, 7–8 log₁₀ IU/mL, and >8 log₁₀ IU/mL.

Follow-up started at baseline, which was the first date of serum HBV DNA level, platelet count, ALT, and AST measurement. Patients underwent regular HCC surveillance with liver ultrasonography and serum alpha-fetoprotein measurements at baseline and every 6 months. They were followed up until the date of HCC diagnosis, death, transplantation, or the last follow-up date (31 December 2021). The diagnosis of HCC was based on histological examination and/or the presence of typical findings (nodule >1 cm with arterial hypervascularity and portal/delayed-phase washout) on dynamic computed tomography and/or magnetic resonance imaging. To verify the diagnosis of HCC, the data of the study population stored in the Korean National Health Insurance Service database was cross-referenced.

The data of patients who started antiviral therapy were censored at 6 months after treatment initiation, as antiviral treatment reduces HCC risk [14,15]. To avoid potential measurement bias from missing HCC events due to censoring exactly at the date of antiviral treatment initiation, we censored the patients 6 months after they started antiviral therapy. Antiviral therapy was indicated when the following reimbursement criteria were met: ALT levels ≥80 U/L with serum HBV DNA levels ≥20,000 IU/mL for HBeAg-positive patients, ALT levels ≥80 U/L with serum HBV DNA levels ≥2,000 IU/mL for HBeAg-negative patients, and ALT levels ≥40 U/L and HBV DNA levels ≥2,000 IU/mL in the presence of cirrhosis.

Statistical analysis

All patients who met the eligibility criteria at baseline were included in the analyses. Categorical variables were compared using the chi-square test and continuous variables were compared using the t-test.

One-way analysis of variance with Tukey’s post hoc test was used to investigate the association of serum HBV DNA levels with the platelet count, APRI, and FIB-4 scores. A general linear model (GLM) with gamma distribution (loglink) was constructed to investigate the association between HBV DNA levels and liver fibrosis scores after adjusting for age, sex, ALT levels, and HBeAg status. This analysis ascertained the differences between groups and the arithmetic mean ratio (AMR) of the non-invasive liver fibrosis scores between groups.

The cumulative incidence rates of HCC were estimated using the Kaplan–Meier method and compared using the log-rank test. Univariate and multivariable Cox proportional hazards models were used to assess the predictive factors for HCC risk. The variables adjusted in the multivariable analysis were age, sex, HBeAg, ALT levels, serum HBV DNA levels, and platelet count. Multivariable analysis using the APRI and FIB-4 scores replacing the platelet count was also performed. The performance of the non-invasive fibrosis scores for predicting HCC risk was assessed using the time-dependent area under the receiver operating characteristic curve (AUROC) at 3, 5, and 10 years.

All reported P-values are two-sided, and p<0.05 was considered statistically significant. SAS (version 9.1; SAS, Cary, NC, USA) and R (version 3.0, http://cran.r-project.org/) software were used to conduct the statistical analyses.

RESULTS

Patient characteristics

The baseline characteristics of the patients are summarized in Table 1 and Table 2. The mean age of the patients was 45 years, 55.9% were men, and 29.9% were HBeAg-positive. The median HBV DNA level was 3.1 log₁₀ IU/mL (interquartile range [IQR] 1.0–6.6 log₁₀ IU/mL), and 29.2% (n=2,029) of the patients had the HBV DNA level >6 log₁₀ IU/mL. The median ALT and AST levels were 25 U/L (IQR 18–36 U/L) and 28 U/L (IQR 22–36 U/L), respectively. The median platelet count was 196×1,000/mm³ (IQR 161– 231×1,000/mm³) and the median APRI and FIB-4 scores were 0.4 (IQR 0.3–0.5) and 1.3 (IQR 0.9–1.9), respectively.

Table 1.

Baseline characteristics of patients stratified by baseline serum hepatitis B virus DNA levels

Table 2.

Baseline characteristics of the patients by HBeAg status

The distribution of patients by the seven HBV DNA level categories was as follows: ≤3 log₁₀ IU/mL (n=3,406), 3–4 log₁₀ IU/mL (n=586), 4–5 log₁₀ IU/mL (n=431), 5–6 log₁₀ IU/mL (n=497), 6–7 log₁₀ IU/mL (n=564), 7–8 log₁₀ IU/mL (n=590) and >8 log₁₀ IU/mL (n=875). Younger patients were more likely to have higher HBV DNA levels (Table 1 and Supplementary Fig. 1). Older patients were more likely to have lower platelet counts and higher APRI and FIB-4 scores (Supplementary Fig. 2). HBeAg-negative patients generally presented with lower HBV DNA levels compared to HBeAg-positive patients, although some overlap was noted between the two groups (Supplementary Fig. 3).

Association between liver fibrosis scores and HBV viral load

In HBeAg-negative patients, the platelet count was inversely associated with serum HBV DNA levels: the platelet count was the highest in patients with HBV DNA levels ≤3 log₁₀ IU/mL and the lowest in patients with HBV DNA levels 6–7 log₁₀ IU/mL (Fig. 2A). In contrast, in HBeAg-positive patients, the platelet count had a linear association with the HBV DNA levels: the platelet count was the lowest in patients with HBV DNA levels of 5–6 log₁₀ IU/mL and increased with elevation in the HBV DNA levels (Fig. 2B). In the entire study population, the platelet count showed a non-linear, U-shaped association with the HBV DNA levels (Fig. 2C). The platelet count was the lowest (mean 175×1,000/mm³) in the HBV DNA 5–7 log₁₀ IU/mL group and the highest in the >8 log₁₀ IU/mL group (mean 216×1,000/mm³).

Figure 2.

Association between the platelet count and serum HBV DNA levels. (A) HBeAg-negative patients. (B) HBeAg-positive patients. (C) Entire patient population. The average values of platelet count by serum HBV DNA level are shown as error bars with 95% CIs. Groups with less than 100 patients are not included in the figure. HBeAg, hepatitis B e antigen; HBV, hepatitis B virus.

The APRI scores of HBeAg-negative patients showed a linear association with the HBV DNA levels: the APRI scores were the lowest in the HBV DNA ≤3 log₁₀ IU/mL group and the highest in the HBV DNA 6–7 log₁₀ IU/mL group (Fig. 3A). In contrast, the APRI scores of HBeAgpositive patients showed an inverse association with the HBV DNA levels: the APRI score was the highest with HBV DNA levels of 6–7 log₁₀ IU/mL and decreased with elevation in the HBV DNA level (Fig. 3B). In the entire study population, the APRI scores showed a non-linear parabolic association with the HBV DNA levels (Fig. 3C). The APRI scores were the highest in the HBV DNA 6–7 log₁₀ IU/mL group and the lowest in the HBV DNA ≤3 log₁₀ IU/mL and >8 log₁₀ IU/mL groups.

Figure 3.

Association between the APRI score and serum HBV DNA levels. (A) HBeAg-negative patients. (B) HBeAg-positive patients. (C) Entire patient population. The average values of the APRI score by the serum HBV DNA level are shown as error bars with 95% CIs. Groups with less than 100 patients are not included in the figure. HBeAg, hepatitis B e antigen; HBV, hepatitis B virus; APRI, aspartate aminotransferase to platelet ratio index.

The FIB-4 scores of HBeAg-negative patients showed a linear association with HBV DNA levels (Fig. 4A). In contrast, the FIB-4 scores of HBeAg-positive patients showed an inverse association with HBV DNA levels: the FIB-4 scores were the highest in patients with HBV DNA of 6–7 log₁₀ IU/mL and decreased with the rise in the HBV DNA levels (Fig. 4B). In the entire study population, the FIB-4 scores showed a non-linear parabolic association with the HBV DNA levels (Fig. 4C). The FIB-4 scores were the highest in patients with HBV DNA levels of 6–7 log₁₀ IU/mL and the lowest in patients with HBV DNA levels of ≤3 log₁₀ IU/mL and >8 log₁₀ IU/mL.

Figure 4.

Association between the FIB-4 score and serum HBV DNA levels. (A) HBeAg-negative patients. (B) HBeAg-positive patients. (C) Entire patient population. The average values of the FIB-4 score by the serum HBV DNA level are shown as error bars with 95% CIs. Groups with less than 100 patients are not included in the figure. HBeAg, hepatitis B e antigen; HBV, hepatitis B virus; FIB-4, Fibrosis-4.

Adjusted association between liver fibrosis scores and HBV viral load

Adjusted analysis using GLM with gamma distribution revealed non-linear associations of serum HBV DNA levels with the platelet count, APRI scores, and FIB-4 scores (Fig. 5). The platelet count was the lowest with HBV DNA levels of 6–7 log₁₀ IU/mL (arithmetic mean ratio [AMR] 0.87; 95% CI 0.84–0.90; p<0.001), showing a U-shaped curve (Fig. 5A, Supplementary Table 1). The APRI score was the highest with HBV DNA levels of 6–7 log₁₀ IU/mL (AMR 1.42; 95% CI 1.34–1.51; p<0.001), showing an inverse U-shaped or parabolic curve (Fig. 5B). The FIB-4 score was also the highest with HBV DNA levels of 6–7 log₁₀ IU/mL (AMR 1.42; 95% CI 1.33–1.50; p<0.001), showing an inverse U-shaped or parabolic curve (Fig. 5C).

Figure 5.

Adjusted association between liver fibrosis scores and HBV DNA levels by general linear models with gamma distribution. (A) HBV DNA levels and platelet counts. (B) HBV DNA levels and APRI score. (C) HBV DNA levels and FIB-4 score. HBeAg, hepatitis B e antigen; HBV, hepatitis B virus; FIB-4, Fibrosis-4; APRI, aspartate aminotransferase to platelet ratio index; AMR, arithmetic mean ratio.

Predictive factors for HCC

During the median follow-up period of 10.0 years, 435 patients (6.3%) developed HCC. The annual incidence rate of HCC was 0.63 per 100 person-years and the estimated cumulative incidence rate was 1.6%, 3.2%, and 6.4% at 3, 5, and 10 years, respectively.

Patients were divided into three groups based on the serum HBV DNA levels: low viral load of ≤4 log₁₀ IU/mL, moderate viral load of 4–8 log₁₀ IU/mL, and high viral load of >8 log₁₀ IU/mL. Kaplan–Meier analysis revealed that the risk of HCC was significantly higher in the moderate viral load group (p<0.001; Supplementary Fig. 4). HCC risk was significantly lower in the high and low viral load groups (p<0.001). Kaplan–Meier analysis revealed that platelet count <150×1,000/mm³, APRI score ≥0.5, and FIB-4 score ≥1.45 were associated with a significantly higher risk of HCC (p<0.001 for all; Supplementary Fig. 5).

Multivariate analysis revealed that male sex, older age, and moderate HBV DNA levels were independent predictive factors of HCC (Table 3). Using the platelet count ≥200×1,000/mm³ as reference, platelet counts <150×1,000/mm³ (hazard ratio [HR] 5.08; 95% CI 3.87–6.68; p<0.001] and 150–199×1,000/mm³ (HR 2.12; 95% CI 1.59–2.81; p<0.001) were found to be independently associated with a higher risk of HCC. APRI scores ≥0.5 were independently associated with a higher risk of HCC compared to APRI scores <0.5 (HR 4.30; 95% CI 3.41–5.42; p<0.001). FIB-4 scores ≥1.45 were independently associated with a higher risk of HCC compared to FIB-4 scores <1.45 (HR 3.70; 95% CI 2.83–4.84; p<0.001).

Table 3.

Univariable and multivariable analyses for the predictive factors for hepatocellular carcinoma

An analysis that separately adjusted for platelet count, APRI score, and FIB-4 score as covariates found that serum HBV DNA level was significantly and independently associated with HCC risk in a non-linear parabolic pattern (Table 3). HBV DNA level of 6–7 log₁₀ IU/mL consistently demonstrated the highest adjusted risk of HCC (p<0.001 for all).

Non-invasive fibrosis scores predicting HCC

The AUROCs of the platelet count for predicting HCC development at 3, 5, and 10 years were 0.76 (95% CI 0.71– 0.80), 0.78 (95% CI 0.75–0.81), and 0.78 (95% CI 0.76– 0.80), respectively (Supplementary Fig. 6). The AUROCs for APRI scores for predicting HCC development at 3, 5, and 10 years were 0.78 (95% CI 0.74–0.83), 0.82 (95% CI 0.79–0.84), and 0.83 (95% CI 0.81–0.85), respectively. The AUROCs of FIB-4 scores for predicting HCC development at 3, 5, and 10 years were 0.79 (95% CI 0.75–0.83), 0.82 (95% CI 0.80–0.85), and 0.82 (95% CI 0.80–0.84), respectively.

DISCUSSION

This study, which enrolled 6,949 treatment-naïve patients with non-cirrhotic CHB without significant ALT elevation, investigated the relationship between serum HBV DNA levels and non-invasive liver fibrosis scores of the platelet count, APRI, and FIB-4 scores. Our research uncovered a distinct and significant non-linear correlation between serum HBV DNA levels and liver fibrosis scores. Remarkably, patients with moderate HBV DNA levels (6–7 log₁₀ IU/mL) exhibited the lowest platelet counts and the highest APRI and FIB-4 scores.

Our findings are in line with those of recent histological studies which demonstrated that, in CHB patients with normal ALT levels, moderate HBV viral load (5–7 log₁₀ IU/mL) was independently associated with significant liver inflammation and fibrosis [16-18]. In a cohort of 414 Chinese CHB patients with detectable HBV DNA and normal ALT who underwent liver biopsy, those with a moderate viral load (5–7 log₁₀ IU/mL) exhibited a notably higher incidence of significant liver inflammation (≥grade 2), fibrosis (≥stage 2), and overall significant histological disease (≥grade 2 inflammation or ≥stage 2 fibrosis), compared to patients with a high or low HBV viral load [18]. When the HBV viral loads were in the 5–7 log₁₀ IU/mL range, the likelihood of significant histological disease was approximately four times higher than that in cases with HBV DNA levels <3 log₁₀ IU/mL. In contrast, the histological impact of an HBV viral load ≥7 log₁₀ IU/mL did not significantly differ from that of levels <3 log₁₀ IU/mL. Notably, this study also identified platelet count as an independent predictor of significant histological disease. These findings reinforce the association of the HBV viral load with significant liver histopathological changes and non-invasive liver fibrosis scores, highlighting a non-linear, parabolic relationship. It is worth noting that these studies used a single measurement of serum HBV DNA levels at the time of the liver biopsy. The utility of the baseline serum HBV DNA levels was affirmed in longitudinal studies. It served as a significant predictor of HCC risk during longterm follow-up of up to 10 years in both untreated and treated non-cirrhotic CHB patients [5,6,19,20]. The REVEAL cohort study showed that among 1,140 participants with HBV DNA levels ≥104 copies/mL at baseline, the majority (78.0%, n=889) had persistently high HBV viral loads during up to 10 years of follow-up (38,330 person-years) [21].

Our recent historical cohort studies consistently demonstrated that moderate HBV DNA levels, approximately 6 log₁₀ IU/mL, are associated with the highest risk of HCC in CHB patients, irrespective of antiviral therapy [5-7]. Several studies validate the findings of this study by showing the non-linear association between HBV DNA levels and HCC risk. PAGED-B risk scores demonstrated that incorporating baseline HBV DNA levels improved the performance in predicting HCC development during the first 5 years of entecavir or tenofovir treatment in HBeAg-positive, non-cirrhotic CHB patients [22]. However, the mechanisms underlying these observations were not well elucidated.

Mounting evidence supports two primary HBV-specific mechanisms in the pathogenesis of HCC [23]. The first is direct hepatocarcinogenesis through the integration of the hepatitis B viral genome into the host chromosome, resulting in the loss of tumor suppressor gene functions and/or activation of tumor-promoting genes. The second mechanism encompasses indirect hepatocarcinogenesis via the inflammation–fibrosis pathway, leading to high rates of hepatocyte proliferation. During the initial phase of infection, HBeAg-positive CHB patients typically show very high HBV DNA levels (≥8 log₁₀ IU/mL), as fully infected hepatocytes can generate 109 to 1010 viruses per mL of serum [24]. However, immune-mediated destruction of HBV-infected hepatocytes induces hepatic inflammation and fibrosis, thereby reducing the HBV DNA titer (<8 log₁₀ IU/mL) in serum. Consequently, a moderate HBV viral load may indicate the presence of significant hepatic inflammation and fibrosis, and a subsequent increase in the hepatocyte proliferation rate, which is a major risk factor for HCC development [25,26]. Our current study showed that HBV viral load is significantly correlated with APRI and FIB-4 scores, and is also associated with HCC risk independent of those liver fibrosis scores in CHB patients. These findings suggest that HBV DNA level is associated with hepatocarcinogenesis through both direct (HBV DNA integration, etc.) and indirect (inflammation and fibrosis) pathways.

Notably, the patients in this study did not receive antiviral treatment as per current guidelines, due to the absence of significant ALT elevation [27-29]. Nevertheless, patients with moderate HBV viral loads exhibited the lowest platelet counts along with the highest APRI and FIB-4 scores, and a significantly high incidence of HCC. This aligns with growing evidence indicating that progression to significant liver fibrosis and hepatocarcinogenesis can occur even without ALT level elevation in patients with CHB [23,30,31]. Our recent research also indicated that, while long-term potent antiviral therapy may reduce the risk of HCC associated with moderate HBV viral loads, it does not completely reverse the risk [6,7]. These findings collectively point to a potential shortcoming of the current approach of delaying antiviral treatment until ALT elevation, with respect to HCC prevention, especially in CHB patients with moderate HBV viral loads.

Despite the existence of effective antiviral treatments that significantly reduce HCC risk [28,29,32], only 2.2% CHB patients received antiviral therapy worldwide in 2019 [33]. One of the main reasons for this extremely low rate of treatment coverage for CHB is the strictness and complexity of the current practice guidelines. Currently, most clinical practice guidelines recommend antiviral treatment after the identification of significant histological liver disease via liver biopsy or elevation in serum ALT levels, despite the presence of high HBV viremia [28,29,32]. Although liver biopsy is the gold standard method for assessing histological liver disease, its use in real-world clinical settings is severely limited owing to its invasive and repetitive nature. Therefore, ALT elevation is used as a criterion to initiate antiviral treatment in most CHB patients [11,34].

This study has several notable limitations. First, it is inherently susceptible to bias and confounding owing to the observational design. To address this issue, we implemented stringent inclusion criteria consistently across the cohort and conducted thorough follow-ups, ensuring near-complete data. This methodology is particularly pertinent, given the large sample size and the relatively low HCC incidence in non-cirrhotic CHB patients. Second, our study exclusively recruited Korean patients, necessitating further research in diverse populations with various ethnicities, genotypes, and modes of HBV transmission. It should be noted that most Koreans are infected with genotype C HBV acquired through the vertical mode of transmission [35,36]. Third, our study does not include data on liver stiffness measurements, as this test was not routinely conducted in clinical practice at the time of cohort assembly (between 2000 and 2013). Liver stiffness measurements could potentially show a stronger association with serum HBV DNA levels and serve as a more accurate predictor for HCC compared to APRI and FIB-4 scores. Lastly, there was a lack of direct liver histology data, reflecting the practical difficulties in performing liver biopsies in a large cohort of patients with CHB.

In summary, our study uncovers a significant and non-linear, parabolic relationship between HBV viral load and non-invasive liver fibrosis scores. A critical finding is that moderate HBV DNA levels, approximately 6 log₁₀ IU/mL, correspond to the lowest platelet counts and the highest APRI and FIB-4 scores. Importantly, this level of HBV viral load was also associated with an elevated risk of HCC, independent of liver fibrosis scores. These findings offer valuable insights into the complex interplay between serum HBV DNA levels, liver fibrosis, and HCC risk in CHB patients. This understanding is essential for enhancing our approach to antiviral therapy in CHB, allowing for more refined and effective treatment strategies.

Notes

Authors’ contribution

Substantial contributions to study conception and design: GAK, SWC, and YSL; substantial contributions to analysis and interpretation of the data: GAK, SWC, SH, and YSL; drafting the article or revising it critically for important intellectual content: GAK, SWC, and YSL; final approval of the version of the article to be published: GAK, SWC, SH, and YSL.

Conflicts of Interest

Y-SL is an advisory board member of Gilead Sciences and receives investigator-initiated research funding from Gilead Sciences. All the other authors have no conflict of interest to declare.

Acknowledgements

There was no industry involvement in the design, conduct, or analysis of the study. This study was supported by grants from the Patient-Centered Clinical Research Coordinating Centre (PACEN; grant number HC20C0062) of the National Evidence-based Healthcare Collaborating Agency and the National R&D Program for Cancer Control through the National Cancer Centre (grant number: HA21C0110), funded by the Ministry of Health & Welfare, Republic of Korea. The funding sources had no role in the design of this study, its execution, analyses, interpretation of the data, or decision to submit the results.

SUPPLEMENTAL MATERIAL

Supplementary material is available at Clinical and Molecular Hepatology website (http://www.e-cmh.org).

Supplementary Table 1.

Adjusted association between HBV DNA levels and liver fibrosis scores by generalized linear models with gamma distribution

cmh-2024-0252-Supplementary-Table-1.pdf

Supplementary Figure 1.

Association between age and serum HBV DNA levels. HBV, hepatitis B virus.

cmh-2024-0252-Supplementary-Fig-1.pdf

Supplementary Figure 2.

Association between age and platelet counts, APRI, and FIB-4 score. APRI, aspartate aminotransferase to platelet ratio index; Fib-4, Fibrosis-4.

cmh-2024-0252-Supplementary-Fig-2.pdf

Supplementary Figure 3.

Frequency and HBV DNA titers by HBeAg status. HBV, hepatitis B virus; HBeAg, hepatitis B e antigen.

cmh-2024-0252-Supplementary-Fig-3.pdf

Supplementary Figure 4.

Risk of HCC by HBV viral load groups. HBV, hepatitis B virus; HCC, hepatocellular carcinoma.

cmh-2024-0252-Supplementary-Fig-4.pdf

Supplementary Figure 5.

Risk of HCC by liver fibrosis scores. HCC, hepatocellular carcinoma.

cmh-2024-0252-Supplementary-Fig-5.pdf

Supplementary Figure 6.

ROC curves of the liver fibrosis scores for predicting HCC development. HCC, hepatocellular carcinoma; ROC, receiver operating curve.

cmh-2024-0252-Supplementary-Fig-6.pdf

Abbreviations

ALT

alanine aminotransferase

AMR

arithmetic mean ratio

anti-HBe

antibody to hepatitis B e antigen

APRI

aspartate aminotransferase to platelet ratio index

AST

aspartate aminotransferase

CHB

chronic hepatitis B

confidence interval

FIB-4

fibrosis-4

HBeAg

hepatitis B e antigen

HBV

hepatitis B virus

HCV

hepatitis C virus

HCC

hepatocellular carcinoma

hazard ratio

IQR

interquartile range

standard deviation

ULN

upper limit of normal

References

1. Thomas DL. Global elimination of chronic hepatitis. N Engl J Med 2019;380:2041–2050.

2. Foreman KJ, Marquez N, Dolgert A, Fukutaki K, Fullman N, McGaughey M, et al. Forecasting life expectancy, years of life lost, and all-cause and cause-specific mortality for 250 causes of death: reference and alternative scenarios for 2016-40 for 195 countries and territories. Lancet 2018;392:2052–2090.

3. European Association for the Study of the Liver. EASL clinical practice guidelines: management of hepatocellular carcinoma. J Hepatol 2018;69:182–236.

4. Chen CJ, Yang HI, Su J, Jen CL, You SL, Lu SN, et al. Risk of hepatocellular carcinoma across a biological gradient of serum hepatitis B virus DNA level. JAMA 2006;295:65–73.

5. Kim GA, Han S, Choi GH, Choi J, Lim YS. Moderate levels of serum hepatitis B virus DNA are associated with the highest risk of hepatocellular carcinoma in chronic hepatitis B patients. Aliment Pharmacol Ther 2020;51:1169–1179.

6. Choi WM, Kim GA, Choi J, Choi GH, Lee YB, Sinn DH, et al. Non-linear association of baseline viral load with on-treatment hepatocellular carcinoma risk in chronic hepatitis B. Gut 2024;73:649–658.

7. Choi WM, Kim GA, Choi J, Han S, Lim YS. Increasing on-treatment hepatocellular carcinoma risk with decreasing baseline viral load in HBeAg-positive chronic hepatitis B. J Clin Invest 2022;132e154833.

8. Suh B, Park S, Shin DW, Yun JM, Yang HK, Yu SJ, et al. High liver fibrosis index FIB-4 is highly predictive of hepatocellular carcinoma in chronic hepatitis B carriers. Hepatology 2015;61:1261–1268.

9. Kim MN, Lee JH, Chon YE, Ha Y, Hwang SG. Fibrosis-4, aspartate transaminase-to-platelet ratio index, and gammaglutamyl transpeptidase-to-platelet ratio for risk assessment of hepatocellular carcinoma in chronic hepatitis B patients: comparison with liver biopsy. Eur J Gastroenterol Hepatol 2020;32:433–439.

10. European Association for the Study of the Liver. EASL clinical practice guidelines on non-invasive tests for evaluation of liver disease severity and prognosis - 2021 update. J Hepatol 2021;75:659–689.

11. Wai CT, Greenson JK, Fontana RJ, Kalbfleisch JD, Marrero JA, Conjeevaram HS, et al. A simple noninvasive index can predict both significant fibrosis and cirrhosis in patients with chronic hepatitis C. Hepatology 2003;38:518–526.

12. Vallet-Pichard A, Mallet V, Nalpas B, Verkarre V, Nalpas A, Dhalluin-Venier V, et al. FIB-4: an inexpensive and accurate marker of fibrosis in HCV infection. comparison with liver biopsy and fibrotest. Hepatology 2007;46:32–36.

13. Xiao G, Yang J, Yan L. Comparison of diagnostic accuracy of aspartate aminotransferase to platelet ratio index and fibrosis-4 index for detecting liver fibrosis in adult patients with chronic hepatitis B virus infection: a systemic review and metaanalysis. Hepatology 2015;61:292–302.

14. Hosaka T, Suzuki F, Kobayashi M, Seko Y, Kawamura Y, Sezaki H, et al. Long-term entecavir treatment reduces hepatocellular carcinoma incidence in patients with hepatitis B virus infection. Hepatology 2013;58:98–107.

15. Wu CY, Lin JT, Ho HJ, Su CW, Lee TY, Wang SY, et al. Association of nucleos(t)ide analogue therapy with reduced risk of hepatocellular carcinoma in patients with chronic hepatitis B: a nationwide cohort study. Gastroenterology 2014;147:143–151.e5.

16. Liu J, Wang J, Yan X, Xue R, Zhan J, Jiang S, et al. Presence of liver inflammation in asian patients with chronic hepatitis B with normal ALT and detectable HBV DNA in absence of liver fibrosis. Hepatol Commun 2022;6:855–866.

17. Wang J, Yan X, Zhu L, Liu J, Qiu Y, Li Y, et al. Significant histological disease of patients with chronic hepatitis B virus infection in the grey zone. Aliment Pharmacol Ther 2023;57:464–474.

18. Huang R, Liu J, Wang J, Qiu Y, Zhu L, Li Y, et al. Histological features of chronic hepatitis B patients with normal alanine aminotransferase according to different criteria. Hepatol Commun 2024;8e0357.

19. Choi WM, Yip TC, Kim WR, Yee LJ, Brooks-Rooney C, Curteis T, et al. Chronic hepatitis B baseline viral load and on-treatment liver cancer risk: a multinational cohort study of HBeAgpositive patients. Hepatology 2024;80:428–439.

20. Liu S, Li Y, Sun J. Concerns about the inverse relationship between baseline HBV DNA and on-treatment hepatocellular carcinoma risk. J Clin Invest 2022;132e161134.

21. Chen CF, Lee WC, Yang HI, Chang HC, Jen CL, Iloeje UH, et al. Changes in serum levels of HBV DNA and alanine aminotransferase determine risk for hepatocellular carcinoma. Gastroenterology 2011;141:1240–1248. 1248.e1-e2.

22. Chun HS, Papatheodoridis GV, Lee M, Lee HA, Kim YH, Kim SH, et al. PAGE-B incorporating moderate HBV DNA levels predicts risk of HCC among patients entering into HBeAgpositive chronic hepatitis B. J Hepatol 2024;80:20–30.

23. Chemin I, Zoulim F. Hepatitis B virus induced hepatocellular carcinoma. Cancer Lett 2009;286:52–59.

24. Zoulim F, Mason WS. Reasons to consider earlier treatment of chronic HBV infections. Gut 2012;61:333–336.

25. Nkongolo S, Mahamed D, Kuipery A, Sanchez Vasquez JD, Kim SC, Mehrotra A, et al. Longitudinal liver sampling in patients with chronic hepatitis B starting antiviral therapy reveals hepatotoxic CD8+ T cells. J Clin Invest 2023;133e158903.

26. Luxenburger H, Neumann-Haefelin C. Liver-resident CD8+ T cells in viral hepatitis: not always good guys. J Clin Invest 2023;133e165033.

27. Terrault NA, Bzowej NH, Chang KM, Hwang JP, Jonas MM, Murad MH, et al. AASLD guidelines for treatment of chronic hepatitis B. Hepatology 2016;63:261–283.

28. Sarin SK, Kumar M, Lau GK, Abbas Z, Chan HL, Chen CJ, et al. Asian-Pacific clinical practice guidelines on the management of hepatitis B: a 2015 update. Hepatol Int 2016;10:1–98.

29. European Association for the Study of the Liver. EASL 2017 clinical practice guidelines on the management of hepatitis B virus infection. J Hepatol 2017;67:370–398.

30. Kim GA, Lim YS, Han S, Choi J, Shim JH, Kim KM, et al. High risk of hepatocellular carcinoma and death in patients with immune-tolerant-phase chronic hepatitis B. Gut 2018;67:945–952.

31. Choi GH, Kim GA, Choi J, Han S, Lim YS. High risk of clinical events in untreated HBeAg-negative chronic hepatitis B patients with high viral load and no significant ALT elevation. Aliment Pharmacol Ther 2019;50:215–226.

32. Terrault NA, Lok ASF, McMahon BJ, Chang KM, Hwang JP, Jonas MM, et al. Update on prevention, diagnosis, and treatment of chronic hepatitis B: AASLD 2018 hepatitis B guidance. Hepatology 2018;67:1560–1599.

33. Cui F, Blach S, Manzengo Mingiedi C, Gonzalez MA, Sabry Alaama A, Mozalevskis A, et al. Global reporting of progress towards elimination of hepatitis B and hepatitis C. Lancet Gastroenterol Hepatol 2023;8:332–342.

34. Gleeson J, Barry J, O’Reilly S. Use of liver imaging and biopsy in clinical practice. N Engl J Med 2017;377:2296.

35. Kim H, Jee YM, Song BC, Shin JW, Yang SH, Mun HS, et al. Molecular epidemiology of hepatitis B virus (HBV) genotypes and serotypes in patients with chronic HBV infection in Korea. Intervirology 2007;50:52–57.

36. Lin CL, Kao JH. The clinical implications of hepatitis B virus genotype: recent advances. J Gastroenterol Hepatol 2011;26 Suppl 1:123–130.

Article information Continued

This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/3.0/) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.

Notes

Study Highlights

• Emerging evidence indicates that serum HBV DNA levels have a non-linear association with the risk of HCC in CHB patients. Non-invasive liver fibrosis scores, such as APRI and FIB-4 scores, are associated with HCC risk. Our study indicates that HBV viral load is significantly associated with non-invasive liver fibrosis scores showing a non-linear and parabolic pattern. Moderate viral load of 6–7 log₁₀ IU/mL exhibited the highest APRI and FIB-4 scores and was also associated with HCC risk.

Characteristics	Entire patients	HBV DNA levels, log10 IU/mL							P-value
Characteristics	Entire patients	≤3	3–4	4–5	5–6	6–7	7–8	>8	P-value
Number of patients	6,949	3,406	586	431	497	564	590	875
Male sex	3,885 (55.9%)	1,846 (54.2%)	330 (56.3%)	257 (59.6%)	288 (57.9%)	356 (63.1%)	318 (53.9%)	490 (56.0%)	0.003
Age, mean±SD, y	45±12	47±11	48±11	45±10	46±10	45±11	39±12	39±11	<0.001
<30	742 (10.7%)	231 (6.8%)	31 (5.3%)	30 (7.0%)	39 (7.8%)	54 (9.6%)	159 (26.9%)	198 (22.6%)
30–39	1,430 (20.6%)	596 (17.5%)	103 (17.6%)	78 (18.1%)	93 (18.7%)	121 (21.5%)	161 (27.3%)	278 (31.8%)
40–49	2,279 (32.8%)	1,156 (33.9%)	184 (31.4%)	182 (42.2%)	180 (36.2%)	182 (32.3%)	158 (26.8%)	237 (27.1%)
50–59	1,765 (25.4%)	967 (28.4%)	198 (33.8%)	102 (23.7%)	139 (28.0%)	147 (26.1%)	82 (13.9%)	130 (14.9%)
≥60	733 (10.5%)	456 (13.4%)	70 (11.9%)	39 (9.0%)	46 (9.3%)	60 (10.6%)	30 (5.1%)	32 (3.7%)
HBeAg
Negative	4,868 (70.1%)	3,406 (100%)	586 (100%)	371 (86.1%)	305 (61.4%)	169 (30.0%)	28 (4.7%)	3 (0.3%)	<0.001
Positive	2,081 (29.9%)	0	0	60 (13.9%)	192 (38.6%)	395 (70.0%)	562 (95.3%)	872 (99.7%)
Platelets, ×1,000/mm³	196 (161–231)	199 (166–233)	194 (165–230)	186 (153–219)	174 (139–205)	168 (137–205)	198 (161–233)	212 (183–246)	<0.001
ALT, U/L	25 (18–36)	20 (15–27)	23 (17–32)	30 (21–42)	37 (28–48)	41 (29–55)	38 (26–54)	33 (24–50)	<0.001
AST, U/L	28 (22–36)	25 (21–30)	25 (21–31)	30 (24–38)	37 (30–47)	40 (33–51)	36 (27–47)	30 (24–39)	<0.001
Albumin, g/dL	4.1 (3.9–4.3)	4.2 (4.0–4.4)	4.2 (4.0–4.4)	4.1 (3.9–4.4)	4.1 (3.9–4.3)	4.0 (3.8–4.2)	4.0 (3.8–4.2)	4.0 (3.8–4.3)	<0.001
Bilirubin, mg/dL	0.9 (0.7–1.1)	0.9 (0.7–1.1)	0.9 (0.7–1.1)	0.9 (0.8–1.2)	1.0 (0.8–1.2)	1.0 (0.8–1.2)	0.9 (0.7–1.1)	0.9 (0.7–1.1)	<0.001
APRI score	0.4 (0.3–0.5)	0.3 (0.2–0.4)	0.3 (0.3–0.4)	0.4 (0.3–0.6)	0.5 (0.4–0.8)	0.6 (0.4–0.8)	0.5 (0.3–0.7)	0.3 (0.3–0.5)	<0.001
<0.5	5,075 (73.0%)	2,922 (85.8%)	487 (83.1%)	285 (66.1%)	202 (40.6%)	191 (33.9%)	336 (56.9%)	652 (74.5%)
≥0.5	1,874 (27.0%)	484 (14.2%)	99 (16.9%)	146 (33.9%)	295 (59.4%)	373 (66.1%)	254 (43.1%)	223 (25.5%)
FIB-4 score	1.3 (0.9–1.9)	1.3 (1.0–1.8)	1.3 (0.9–1.8)	1.3 (1.0–1.9)	1.7 (1.1–2.5)	1.8 (1.1–2.7)	1.1 (0.7–1.9)	1.0 (0.6–1.4)	<0.001
<1.45	4,029 (58.0%)	1,987 (58.3%)	342 (58.4%)	241 (55.9%)	192 (38.6%)	211 (37.4%)	379 (64.2%)	677 (77.4%)
≥1.45	2,920 (42.0%)	1,419 (41.7%)	244 (41.6%)	190 (44.1%)	305 (61.4%)	353 (62.6%)	211 (35.8%)	198 (22.6%)
REACH-B score	7 (5–10)	5 (3–7)	7 (6–9)	10 (8–12)	10 (9–12)	11 (9–12)	10 (8–12)	10 (9–11)	<0.001
CU-HCC score	3.0 (1.5–4.5)	1.5 (0.0–3.0)	2.5 (1.0–4.0)	2.5 (1.0–4.0)	5.5 (4.0–7.0)	5.5 (4.0–7.0)	4.0 (4.0–5.5)	4.0 (4.0–5.5)	<0.001
GAG-HCC score	65.0 (54.4–75.8)	57.0 (47.0–67.0)	68.2 (59.8–76.3)	70.4 (61.6–77.1)	72.6 (65.3–80.3)	76.7 (67.1–84.5)	70.0 (60.9–80.6)	74.8 (66.2–83.2)	<0.001
Duration of follow-up, y	10.0 (6.0–13.6)	11.6 (8.1–14.9)	8.4 (5.2–11.1)	9.0 (5.0–11.9)	6.7 (3.9–11.5)	7.3 (4.0–11.6)	9.6 (4.8–13.8)	7.9 (4.9–12.0)	<0.001
HCC development	435 (6.3%)	147 (4.3%)	15 (2.6%)	31 (7.2%)	71 (14.3%)	113 (20.0%)	38 (6.4%)	20 (2.3%)	<0.001

Characteristics	Whole study population	HBeAg-negative patients	HBeAg-positive patients	P-value
Number of patients	6949	4,868 (70.1%)	2,081 (29.9%)
Male sex	3,885 (55.9%)	2,708 (55.6%)	1,177 (56.6%)	0.47
Age, mean±SD, years	45±12	47±11	40±11	<0.001
<30	742 (10.7%)	313 (6.4%)	429 (20.6%)
30–39	1,430 (20.6%)	845 (17.4%)	585 (28.1%)
40–49	2,279 (32.8%)	1,667 (34.2%)	612 (29.4%)
50–59	1,765 (25.4%)	1,418 (29.1%)	347 (16.7%)
≥60	733 (10.5%)	625 (12.8%)	108 (5.2%)
HBV DNA, median (IQR), log₁₀ IU/mL	3.1 (1.0–6.6)	1.0 (1.0–3.4)	7.8 (6.8–8.4)	<0.001
HBV DNA ≤3	3,406 (49.0%)	3,406 (70.0%)	0
3< HBV DNA ≤4	586 (8.4%)	586 (12.0%)	0
4< HBV DNA ≤5	431 (6.2%)	371 (7.6%)	60 (2.9%)
5< HBV DNA ≤6	497 (7.2%)	305 (6.3%)	192 (9.2%)
6< HBV DNA ≤7	564 (8.1%)	169 (3.5%)	395 (19.0%)
7< HBV DNA ≤8	590 (8.5%)	28 (0.6%)	562 (27.0%)
HBV DNA >8	875 (12.6%)	3 (0.1%)	872 (41.9%)
Platelets, median (IQR), ×1,000/mm³	196 (161–231)	194 (161–229)	198 (161–235)	0.06
ALT, median (IQR), U/L	25 (18–36)	22 (16–31)	36 (25–52)	<0.001
AST, median (IQR), U/L	28 (22–36)	26 (22–32)	34 (26–46)	<0.001
Albumin, median (IQR), g/dL	4.1 (3.9–4.3)	4.2 (3.9–4.4)	4.0 (3.8–4.2)	<0.001
Bilirubin, median (IQR), mg/dL	0.9 (0.7–1.1)	0.9 (0.7–1.1)	0.9 (0.7–1.1)	0.02
APRI score	0.4 (0.3–0.5)	0.3 (0.3–0.5)	0.4 (0.3–0.7)	<0.001
<0.5	5,075 (73.0%)	3,847 (79.0%)	1,228 (59.0%)
≥0.5	1,874 (27.0%)	1,021 (21.0%)	853 (41.0%)
FIB-4 score	1.3 (0.9–1.9)	1.4 (1.0–1.9)	1.1 (0.7–1.8)	0.002
<1.45	4,029 (58.0%)	2,713 (55.7%)	1,316 (63.2%)
≥1.45	2,920 (42.0%)	2,155 (44.3%)	765 (36.8%)
REACH-B score	7 (5–10)	6 (4–8)	10.0 (9–12)	<0.001
CU-HCC score	3.0 (1.5–4.5)	2.5 (0.0–4.0)	5.5 (4.0–6.3)	<0.001
GAG-HCC score	65.0 (54.4–75.8)	61.6 (51.0–72.1)	73.1 (63.5–81.4)	<0.001
Duration of follow-up, median (IQR), years	10.0 (6.0–13.6)	10.5 (6.9–13.9)	8.4 (4.7–12.5)	<0.001
HCC development	435 (6.3%)	261 (5.4%)	174 (8.4%)	<0.001

Variables	Univariable analysis		Multivariable analysis
			Model 1^a		Model 2^b		Model 3^c
	HR (95% CI)	P-value	HR (95% CI)	P-value	HR (95% CI)	P-value	HR (95% CI)	P-value
Sex
Women	1 [Reference]	–	1 [Reference]	–	1 [Reference]	–	1 [Reference]	–
Men	2.48 (2.00–3.08)	<0.001	2.50 (2.01–3.11)	<0.001	2.62 (2.11–3.26)	<0.001	2.67 (2.15–3.31)	<0.001
Age, years	1.05 (1.04–1.06)	<0.001	1.06 (1.05–1.07)	<0.001	1.06 (1.05–1.07)	<0.001	1.04 (1.03–1.05)	<0.001
ALT levels
ALT, <1× ULN	1 [Reference]	–	1 [Reference]	–	1 [Reference]	–	1 [Reference]	–
ALT, 1–2× ULN	2.73 (2.24–3.34)	<0.001	1.16 (0.92–1.47)	0.22	0.82 (0.65–1.04)	0.10	1.17 (0.92–1.48)	0.20
HBeAg
HBeAg-negative	1 [Reference]	–	1 [Reference]	–	1 [Reference]	–	1 [Reference]	–
HBeAg-positive	1.85 (1.53–2.25)	<0.001	1.37 (1.03–1.82)	0.03	1.34 (1.01–1.78)	0.045	1.35 (1.01–1.80)	0.04
HBV DNA, log10 IU/mL
≤3	1 [Reference]	–	1 [Reference]	–	1 [Reference]	–	1 [Reference]	–
3–4	0.83 (0.49–1.42)	0.50	0.83 (0.49–1.42)	0.50	0.87 (0.51–1.49)	0.62	0.81 (0.48–1.38)	0.44
4–5	2.19 (1.48–3.22)	<0.001	1.91 (1.28–2.85)	0.002	1.69 (1.13–2.53)	0.01	2.12 (1.42–3.17)	<0.001
5–6	4.91 (3.69–6.53)	<0.001	3.67 (2.63–5.12)	<0.001	2.89 (2.06–4.05)	<0.001	3.84 (2.76–5.34)	<0.001
6–7	6.58 (5.14–8.41)	<0.001	4.05 (2.86–5.75)	<0.001	3.30 (2.32–4.69)	<0.001	4.45 (3.15–6.31)	<0.001
7–8	1.80 (1.26–2.57)	0.001	1.75 (1.10–2.78)	0.02	1.46 (0.92–2.33)	0.11	1.90 (1.19–3.03)	0.007
>8	0.71 (0.44–1.13)	0.15	0.87 (0.50–1.51)	0.61	0.71 (0.41–1.24)	0.23	0.82 (0.47–1.44)	0.49
Platelet count, ×1,000/mm³
≥200	1 [Reference]	–	1 [Reference]	–	–		–
150–199	2.80 (2.12–3.71)	<0.001	2.12 (1.59–2.81)	<0.001	–		–
<150	9.48 (7.30–12.33)	<0.001	5.08 (3.87–6.68)	<0.001	–		–
APRI score
<0.5	1 [Reference]	–	–		1 [Reference]	–	–
≥0.5	7.36 (6.02–9.00)	<0.001	–		4.30 (3.41–5.42)	<0.001	–
FIB-4 score
<1.45	1 [Reference]	–	–		–		1 [Reference]	–
≥1.45	6.54 (5.17–8.28)	<0.001	–		–		3.70 (2.83–4.84)	<0.001