Baseline characteristics

A total of 719 HBeAg-positive patients with normal ALT were included in this study. Male patients represented 59.8% and the median age was 33.0 (30.0, 40.0) years. The median levels of ALT, HBsAg, HBeAg and HBV DNA were 23.9 U/L, 4.5 log10 IU/mL, 3.1 log10 S/CO, and 7.5 log10 IU/mL, respectively (Table 1). 586 (81.5%) and 133 (18.5%) patients with HBV DNA > 106 IU/mL and HBV DNA ≤ 106 IU/mL were classified into IT phase and indeterminate phase, respectively.

Table 1 Clinical characteristics between patients with HBeAg-positive indeterminate phase and immune-tolerant phase

In comparison, patients in the indeterminate phase had older age (37.0 years vs. 33.0 years, P < 0.001), higher proportion of male sex (68.4% vs. 57.8%, P = 0.025) compared to those in the IT phase. Patients in the IT phase had higher median levels of serum HBsAg (4.6 log10 IU/mL vs. 3.4 log10 IU/mL, P < 0.001), HBeAg (3.1 log10 S/CO vs. 0.8 log10 S/CO, P < 0.001), PLT (203.0 × 109/L vs. 176.0 × 109/L, P < 0.001) and lower median levels of liver function indicators, including AST (21.9 U/L vs. 24.3 U/L, P < 0.001), gama-glutamyl transpeptidase (GGT, 17.0 U/L vs. 25.3 U/L, P < 0.001) and total bilirubin (TBil,12.2 µmol/L vs. 13.7 µmol/L, P = 0.001). The median level of ALT and albumin (ALB) were comparable between IT patients and indeterminate patients.

Comparison of liver fibrosis between patients with HBeAg-positive indeterminate phase and immune-tolerant phase

The median values of APRI and FIB-4 were 0.28 and 0.75 in the overall patients, respectively. A total of 156 patients had LSM data with a median value of 5.75 kPa. Liver biopsies were performed in 86 patients and 44.2% of patients had significant fibrosis (≥ S2). In comparison, the APRI (0.33 vs. 0.27, P < 0.001), FIB-4 (1.07 vs. 0.72, P < 0.001), and LSM values (7.80 kPa vs. 5.65 kPa, P = 0.011) were significantly higher in patients with indeterminate phase than those in the IT phase (Table 1). Patients in the indeterminate phase had significant higher proportions of significant fibrosis (27.1% vs. 11.3%, P < 0.001) and cirrhosis (14.3% vs. 3.2%, P < 0.001) compared to those in the IT phase (Fig. 1).

Fig. 1

The proportion of significant liver fibrosis and cirrhosis in patients with HBeAg-positive indeterminate phase and immune-tolerant phase

Logistic regression analysis of factors associated with liver fibrosis

Due to the multicollinearity with APRI and FIB-4, several parameters including age, PLT, ALT, and AST were not included in the logistic regression analysis (Table 2). In the univariate analysis, male sex, higher TBil and GGT, lower ALB and HBV DNA levels, and indeterminate phase was associated with higher risk of significant fibrosis. Indeterminate phase (odds ratio [OR] 2.138, 95% confidence interval [CI] 1.253–3.649, P = 0.005) remained an independent risk factor of significant liver fibrosis in the multivariate analysis. Besides, we also explored the association of indeterminate phase with cirrhosis. Multivariate analysis showed that indeterminate phase (OR 4.163, 95%CI 1.921–9.025, P < 0.001) was also associated with the increased risk of the presence of cirrhosis (Table 3).

Table 2 Logistic regression analysis of clinical factors associated with significant liver fibrosisTable 3 Logistic regression analysis of clinical factors associated with liver cirrhosisAssociation between indeterminate phase and liver fibrosis in different age subgroups

A further subgroup analysis was performed in patients with different age (Table 4). In patients younger than 30 years, 14.0% were in the indeterminate phase and 86.0% were in the IT phase. The APRI, FIB-4, and LSM values, along with the proportion of significant liver fibrosis and cirrhosis, were not significantly different between these two groups. Although more indeterminate patients (16.7%) had significant liver fibrosis than IT patients (8.1%), the differences were not statistically significant (P = 0.215). However, in patients aged ≥ 30 years, patients in the indeterminate phase had higher values of APRI (0.36 vs. 0.27, P < 0.001), FIB-4 (1.21 vs. 0.80, P < 0.001), and LSM (8.20 kPa vs. 5.65 kPa, P = 0.006) than those in the IT phase. The proportion of significant liver fibrosis (29.4% vs. 12.3%, P < 0.001) and cirrhosis (14.7% vs. 3.9%, P < 0.001) were also significantly higher in patients in the indeterminate phase compared to those in the IT phase (Fig. 1). Patients aged ≥ 30 years presented a higher proportion of significant liver fibrosis (15.7% vs. 9.3%, P = 0.035) compared to patients aged < 30 years.

Table 4 Clinical characteristics between patients with HBeAg-positive indeterminate phase and immune-tolerant phase stratified by age

To determine the factors associated with significant fibrosis, the logistic regression analysis was conducted in patients younger than 30 years and those 30 years or older, respectively (Table 5). In patients with age < 30 years, indeterminate phase was not associated with significant fibrosis. However, indeterminate phase (OR 2.360, 95% CI 1.316–4.231, P = 0.004) was an independent risk factor of significant fibrosis in patients with age ≥ 30 years.

Table 5 Logistic regression analysis of clinical factors associated with significant liver fibrosis stratified by age

We performed additional correlation analysis to explore the relationship between HBV DNA levels and APRI, FIB-4, LSM, as well as fibrosis stages (Table S1). In general, HBV DNA levels were negatively correlated with the values of APRI (r =-0.269, P < 0.001), FIB-4 (r =-0.306, P < 0.001) and LSM (r =-0.243, P = 0.002), and fibrosis stages (r =-0.299, P < 0.001). We observed similar relationship in patients aged ≥ 30 years, while there was no significant correlation between HBV DNA levels and liver fibrosis in patients aged < 30 years.

Sensitivity analysis in patients who underwent liver biopsy

A total of 86 patients underwent liver biopsy, including 21 patients (24.4%) in the indeterminate phase and 65 patients (75.6%) in the IT phase (Table S2). The age, sex, and ALT levels were comparable, while the serum HBsAg level (4.5 log10 IU/mL vs. 2.4 log10 IU/mL, P < 0.001) and PLT counts (180.0 × 109/L vs. 147.0 × 109/L, P = 0.002) were higher in the IT phase compared to patients with indeterminate. Patients in the indeterminate phase had higher proportion of significant liver fibrosis (66.7% vs. 36.9%, P = 0.017)   than patients in the IT phase. There was no significant difference of liver inflammation grades and significant liver injury between the two groups (P = 0.076, P = 0.080). We further investigated the association of HBV DNA level with liver fibrosis (Table S3) and found that lower HBV DNA level remained associated with significant liver fibrosis (OR 0.632, 95%CI 0.443–0.901, P = 0.011). Multivariate logistics regression analysis showed that HBV DNA level was not associated with significant liver inflammation (Table S4) but associated with significant liver injury (OR 0.657, 95%CI 0.443–0.987, P = 0.038) (Table S5). However, indeterminate phase was not a risk factor of significant liver fibrosis (P = 0.152), significant liver inflammation (P = 0.459) and significant liver injury (P = 0.328).

Sensitivity analysis by a more stringent normal value of ALT

Using a more stringent normal value of ALT of 30 U/L for male and 19 U/L for female, 394 of patients had normal ALT. 81 (20.6%) and 313 (79.4%) patients were classified into indeterminate and IT groups, respectively. In comparison, patients in indeterminate phase had higher proportion of significant liver fibrosis than those in IT phase (Table S6), which was consistent with the results based on ULN of ALT of 35 for men and 25 for women. Logistics regression analysis also demonstrated that indeterminate phase (OR 2.061, 95% CI 1.055–4.028, P = 0.034) was an independent risk factor of significant fibrosis (Table S7).