Initially, a cross-section study carried out from January 2017 to March 2018, involving 206 medical inpatients with CHB at Beijing You an Hospital, Capital Medical University, were enrolled as Group 1 (coexistence of HBsAg and anti-HBs) according to the inclusion and exclusion criterion (Table 1). As controls, 206 medical inpatients were randomly selected as Group 2 (HBsAg alone). Medical records of these 412 patients were reviewed, including clinical evaluations, imaging features and laboratory results. Demographic, clinical and serological characteristics of Group 1 and Group 2 were compared, and the risk of developing HCC was assessed through logistic regression analysis.

Furthermore, a cohort study with chronic HBV infection patients compared the development of HCC in Group 3 (130 patients with coexistence of HBsAg and anti-HBs) and Group 4 (130 patients with HBsAg alone). The cohorts were retrospectively collected between September 2010 and November 2013, according to the eligibility criteria listed in Table 2. The flow chart is presented in Fig. 1. The follow-up duration was calculated from the initial visit to the last visit or the diagnosis of HBV-associated HCC. The medical records all patients were reviewed, including clinical evaluations, imaging features and laboratory results. All patients were followed by ultrasonography (US) or liver dynamic three-dimensional computed tomography (CT) and serum alpha-fetoprotein every 6–12 months. During follow-up, that ranged from 9 to 92 months, HCC developed in 20 (15.4%) of Group 3 and in 10 (7.7%) patients of Group 4. Since early studies had confirmed an increasing risk of HCC when HBV DNA level was over 2000 IU/mL (high HBV load),13,14 the cohort was further divided into two subgroups (high HBV load and low HBV load). Cumulative incidence of HCC was then estimated using Kaplan–Meier analysis.

Fig. 1.

Flow chart of the study sample of the cohort study. LT: Liver transplantation.

(0.17MB).

Case definitions of chronic CHB and HBV-associated cirrhosis were in accordance with the criteria issued by the Association of Infectious Diseases of China in 2015.15 HBV-associated HCC was diagnosed according to standardization of diagnosis and treatment for primary liver cancer.16 The liver failure was diagnosed according to the guideline from the Association of Infectious Diseases of China in 201217 and categorized into four groups: acute liver failure (ALF), subacute liver failure (SALF), acute-on-chronic liver failure (ACLF), and chronic liver failure (CLF).

The HBV serological markers HBsAg, anti-HBs, HBeAg, anti-HBe, and anti-HBc were detected under Roche Elecsys e601 (Roche, German). Anti-HBs with results over 10 mIU/mL were considered positive. Serum anti-HBs levels ≥100 mIU/mL were regarded as ‘High’ anti-HBs (high levels of anti-HBs), for individuals with anti-HBs ≥100 mIU/mL were recognized to have sufficient immunity against HBV infection. It was necessary to maintain anti-HBs ≥100 mIU/mL so as to prevent reinfection of HBV in liver transplant recipients by using HBV immunoglobulin (HBIG).18 Therefore, samples with positive anti-HBs and serum levels <100 mIU/mL were regarded as ‘Low’ anti-HBs (low levels of anti-HBs). It was reported that HBsAg ≥1000 IU/mL was a key factor associated with HCC development in patients with HBV DNA level <2000 IU/mL.13 Thus, serum levels ≥1000 IU/mL were regarded as ‘High’ HBsAg (high levels of HBsAg). Similarly, samples with positive HBsAg with serum levels <1000 IU/mL were regarded as ‘Low’ HBsAg (low levels of HBsAg).

Serum alanine aminotransferase (ALT) levels were determined under ADVIA 2400 (Siemens, Germany). Serum alpha-fetoprotein (AFP) levels were also determined by Roche Elecsys e601 (Roche, German), and serum levels induced by vitamin K absence or antagonist II (PIVKA-II) were detected by using the commercial kit (Fujirebio, Japan) of chemiluminescent microparticle immunoassay. Serum HBV DNA levels were quantified by COBAS Taqman HBV DNA (Roche, German) with a limit detection threshold of 20 IU/mL.

The Mann–Whitney U-test was used to compare age, ALT, AFP, PIVKA-II, and HBV DNA among groups, while the chi-squared test was used to compare categorical variables such as the proportions of age ranges, sex, the grade of Child–Pugh. Logistic regression analysis was performed using candidate variables based on univariate analysis. Kaplan–Meier analysis was performed to estimate the cumulative incidence of HCC in patients including Group 3 and Group 4, and the two cumulative incidences were compared with the log rank test. Data from medical records were stored using Microsoft Office Excel 2010 (Microsoft Corp.). The software of SPSS v. 24.0 (IBM Corp.) was used for data analysis. All p-values were two-sided and p-values <0.05 were considered statistically significant.

The overall (n = 412) median age study subjects was 52 years (range, 19–91), 76.5% (315/412) male; 44.4% (183/412) were diagnosed with HCC, 10.4% (43/412) had evidence of liver failure, 18.4% (76/412) were evaluated as Child–Pugh C.

Demographic, clinical and serological characteristics of the study groups are summarized in Table 3. In brief, there were no significant differences between Group 1 and 2 regarding sex (p = 0.131), grade of Child–Pugh (p = 0.488), presence of liver failure (p = 0.903), serum ALT level (p = 0.231), HBeAg positive rate (p = 0.767) and serum levels of HBV DNA (p = 0.727). However, the median age of patients in Group 1 was significantly higher than in Group 2 (54 years vs. 49 years; p < 0.001). It should be pointed out that there was a higher proportion of patients with HCC and also a higher proportion of patients with liver failure in Group 1 than in Group 2 (51.0% vs. 37.8%, p = 0.011; 13.6% vs. 7.3%, p = 0.036, respectively). The proportion of patients with HBsAg ≥1000 IU/mL in Group 1 was significantly higher than Group 2 (38.3% vs. 54.9%, p = 0.001). Moreover, the median serum AFP level of the patients was 32.7 (P25–P75, 6.2–296.2) ng/mL in Group 1 and 16.4 (P25–P75, 4.9–156.5) U/L in Group 2 (p = 0.024). Similarly, the median serum PIVKA-II level of patients in Group 1 was higher than in Group 2 (p < 0.001). Interestingly, the coexistence of HBeAg and anti-HBe was more frequently found in patients with coexistence of HBsAg and anti-HBs than patients in Group 2 (39.8% vs. 22.1%, p = 0.008).

Four serological patterns of different serum levels combinations of HBsAg and anti-HBs are exhibited in Fig. 2. Out of the 206 patients with coexistence of HBsAg and anti-HBs, the proportion of ‘Low’ HBsAg / ‘Low’ anti-HBs was the highest (40.29%), while the number of patients with ‘High’ HBsAg / ‘High’ anti-HBs was the lowest (9.71%). The rate of HCC were higher in Group 1 than in Group 2. However, there was no significant difference in the proportions of HCC patients within these different serological patterns (P = 0.827).

Fig. 2.

The proportion of different serological patterns of HBsAg and anti-HBs in the cross-sectional study. H/H: ‘High’ HBsAg / ‘High’ anti-HBs. H/L: ‘High’ HBsAg / ‘Low’ anti-HBs. L/H: ‘Low’ HBsAg / ‘High’ anti-HBs. L/L: ‘Low’ HBsAg / ‘Low’ anti-HBs. +/−: HBsAg positive and anti-HBs negative. CHB: chronic hepatitis B. HCC: hepatocellular carcinoma. The proportions of HCC in patients with all four serological patterns of coexistence of HBsAg and anti-HBs are higher than patients with HBsAg positive and anti-HBs negative. However, there are no obvious differences in the proportions of HCC in patients with these different serological patterns (p = 0.827).

(0.06MB).

In univariate analysis, age ≥50 years, male sex, Child–Pugh C, liver failure, initial serum ALT level, coexistence of HBsAg and anti-HBs, HBeAg positive, and serum HBV DNA level were significantly associated with HCC (Table 4).

In multivariate logistic regression analysis, age ≥50 years [odds ratio (OR), 4.486; 95% confidence interval (CI), 2.592–7.762; p < 0.001], male sex (OR, 3.683; 95% CI, 2.025–6.699; p < 0.001), and coexistence of HBsAg and anti-HBs (OR, 2.063; 95% CI, 1.259–3.382; p = 0.004) were considered as independent risk factors for HCC (Table 5). The different serological patterns for coexistence of HBsAg and anti-HBs, except for ‘High’ HBsAg and ‘Low’ anti-HBs, were also considered as independent risk factors for HCC. In particular, compared to patients with HBsAg positive and anti-HBs negative, patients with the ‘High’ HBsAg /‘High’ anti-HBs (OR, 4.295; 95% CI, 1.104–16.699; p = 0.035) and ‘Low’ HBsAg / ‘High’ anti-HBs (OR, 3.207; 95% CI, 1.299–7.919; p = 0.012) exhibited significantly higher risk for HCC (Table 5).

The baseline characteristics of the study sample are presented in Table 6. Lower median serum HBV DNA levels and shorter median follow-up duration in Group 3 than Group 4 (5.3 log10 IU/mL vs. 6.0 log10 IU/mL, p = 0.041; 54.5 months vs. 68.5 months, p = 0.041, respectively) were found. There were no significant differences between these two groups, in terms of age (p = 0.567), male sex (p = 0.274), proportion of Child–Pugh C (p = 0.169), serum ALT (p = 0.186), proportion of cirrhosis (p = 0.380), and rate of patients with HBeAg positive.

The incidence of HCC during the follow-up period was 15.4% (20/110) in patients with coexistence of HBsAg and anti-HBs and 7.7% (10/120) in those with HBsAg alone. Out of the 260 patients, 223 (85.8%) received antiviral therapy, and there was no significant difference in the proportion of patients who received antiviral therapy between the two groups (p = 0.214). The cumulative incidence of HCC was not related to coexistence of HBsAg and anti-HBs in a sub-cohort of 59 patients with HBV-DNA level <2000 IU/mL (p = 0.902) (Fig. 3A). However, in a subcohort of 201 patients with HBV-DNA level ≥2000 IU/mL, the cumulative incidence of HCC was associated with this serological pattern (p = 0.018) (Fig. 3B). Moreover, the cumulative incidence of HCC was also associated with different serological patterns of HBsAg and anti-HBs in patients with high HBV load (p < 0.001) (Fig. 3C). Particularly, ‘Low’ HBsAg / ‘High’ anti-HBs might be associated with higher risk of HCC development. To validate this finding, patients with HBV DNA level ≥2000 IU/mL were categorized by the different serological pattern: ‘Low’ HBsAg/‘Low’ anti-HBs (Fig. 3D), ‘High’ HBsAg/‘Low’ anti-HBs (Fig. 3E) and ‘Low’ HBsAg/‘High’ anti-HBs (Fig. 3F). Only the serological pattern of ‘Low’ HBsAg/‘High’ anti-HBs was associated with increased risk of HCC in patients with high HBV load (log rank p < 0.001). Unfortunately, only four patients with ‘High’ HBsAg and ‘High’ anti-HBs were included in our cohort study, and none of them developed HCC at the end of follow-up.

Fig. 3.

The cumulative incidence of HCC in patients with CHB. (A) The cumulative incidence of HCC is not related to coexistence of HBsAg and anti-HBs in a subcohort of 59 patients with HBV-DNA level <2000 IU/mL (p = 0.902). (B) However, in a subcohort of 201 patients with HBV-DNA level ≥2000 IU/mL, the cumulative incidence of HCC is associated with coexistence of HBsAg and anti-HBs (P = 0.018). (C) Comparison of cumulative incidence of HCC in different serological patterns of HBsAg/anti-HBs. These patients with HBV DNA level ≥2000 IU/mL are categorized by the different serological pattern: ‘Low’ HBsAg / ‘Low’ anti-HBs (D), ‘High’ HBsAg / ‘Low’ anti-HBs (E), and ‘Low’ HBsAg / ‘High’ anti-HBs (F).

(0.54MB).