Patient characteristics and treatment outcome
A total of 110 patients with CHB were recruited for this study from February 2016 to December 2018 in Shanghai Changhai Hospital. Finally, 79 patients with CHB (54 HBeAg-positive) were enrolled in this study. All of them were treated for more than 96 weeks with first-line NAs (ETV or TDF).
At baseline, HBeAg-positive patients with CHB had significantly higher viral replication level compared with HBeAg-negative patients with CHB. However, during the 96 weeks of NA treatment, HBeAg-positive and HBeAg-negative patients achieved similar serum HBV DNA below the MDL (< MDL), HBsAg loss rate, and ALT normalization rate (Table 1). Of these 79 patients with CHB, only one patient achieved HBsAg loss, whereas four patients (two HBeAg-positive and two HBeAg-negative patients with CHB) achieved HBsAg < 100 IU/mL during the 96 weeks of antiviral treatment. Fourteen (25.93%) HBeAg-positive patients with CHB achieved SC response. HBeAg-positive patients with SC had significantly higher ALT normalization rate and serum HBV DNA < MDL rate than NSC patients (Table 1).
Table 1
Clinical Characteristics of patients with CHB
Characteristics | Total (n = 79) | HBeAg pos. (n = 54) | HBeAg Neg. (n = 25) | P value | HBeAg Pos. | P value |
SC (n = 14) | NSC (n = 40) |
Sex (M: F) | 55:24 | 40:14 | 15:10 | 0,206 | 10:4 | 30:10 | 1.000 |
Age(year) | 36.0(30.0–45.0) | 32.0(29.0–40.0) | 45.0(37.5–55.5) | ༜0.001 | 31.5(27.00–40.00) | 32.00(30.00-41.5) | 0.553 |
Baseline characteristics |
HBsAg (log 10, IU/ml) | 3.47(3.08–4.14) | 3.69(3.22–4.39) | 3.11(2.75–3.55) | 0.004 | 3.80(3.20–4.19) | 3.63(3.20–4.46) | 0.845 |
HBeAg(log 10)S/CO | 2.81(1.69–3.11) | 2.81(1.69–3.11) | / | / | 2.75(1.80–2.96) | 2.81(1.57–3.16) | 0.612 |
HBV DNA(log10)IU/ml | 6.11(4.41–7.70) | 7.01(5.51–7.90) | 4.34(3.51–5.25) | ༜0.001 | 7.05(5.32–7.96) | 7.01(5.64–7.93) | 0.942 |
HBV RNA(log 10)copies/ml | 6.64(5.58–8.08) | 7.54(5.96–8.30) | 5.78(5.23–6.42) | 0.001 | 7.20(5.59–7.97) | 7.70(6.09–8.68) | 0.302 |
ALT(U/L) | 105.5(51.5-227.75) | 105.50(53.75-230.25) | 101.0(33.25–225.0) | 0.561 | 255.00(205.5-444.75) | 73.5(44.5-153.25) | 0.000 |
AST(U/L) | 66.5(36.0-112.0) | 68.00(38.75–112.00) | 63.0(27.0-121.0) | 0༎663 | 119.5(73.75–195.0) | 44.5(31.75-87.00) | 0.000 |
Treatment outcome |
HBV DNA below MDL | 63(78.48) | 41(75.93) | 22(88.00) | 0.214 | 14/14(100) | 27/40(67.5) | 0.037 |
HBsAg loss rate,n(%) | 1(1.27) | 1(1.85) | 0(0) | 1.00 | 1/14(7.14) | 0/40(0) | 0.259 |
HBsAg < 100IU/mL rate,n(%) | 4(5.06) | 2(3.70) | 2(8.0) | 0.796 | 2/14(14.29) | 0/40(0) | 0.064 |
ALT normalization rate,n(%) | 60(75.95) | 39(72.22) | 21(84.00) | 0.255 | 14/14(100) | 25/40(62.5) | 0.000 |
Continuity data were shown as median (interquartile range). SC: seroconversion; NSC: no seroconversion; pos. positive; Neg. negative. MDL: minimum detection limit; ALT, alanine aminotransferase; AST, aspartate aminotransferase; HBeAg, hepatitis B e antigen; HBsAg, hepatitis B surface antigen; HBV, hepatitis B virus |
Correlation between serum HBV RNA and other HBV biomarkers in patients with CHB on NA treatment by baseline HBeAg status
In HBeAg-positive patients with CHB, before NA treatment, strong correlation was found between serum HBV RNA and HBV DNA (Spearman’s r = 0.73; Fig. 1A), and this strong correlation gradually decreased with treatment (Spearman’s r = 0.53 and 0.41; Figs. 1B and 1C). Furthermore, sustained moderate correlation was noted between serum HBV RNA and HBeAg and HBsAg titer during treatment (Spearman’s r = 0.59–0.40; Figs. 1A–1C). However, in HBeAg-negative patients with CHB on NA treatment, a significant correlation only existed between serum HBV RNA and HBV DNA level before treatment (Spearman’s r = 0.53; Fig. 1D) and the relationship decreased and disappeared gradually with treatment (Spearman’s r = 0.42 and 0.25; Figs. 1E and 1F).
Kinetics of serum HBV RNA in patients with CHB on NA treatment by HBeAg status
During NA treatment, HBeAg-positive and HBeAg-negative patients with CHB had similar dynamic change patterns of serum HBV biomarkers, including serum HBV RNA, HBV DNA, and HBsAg titers (Figs. 2A–2C). Serum HBV DNA level decreased more rapidly than serum HBV RNA level in total CHB patients and HBeAg-positive patients during the early stage of NA treatment (Figs. 2D and 2E). However, serum HBVRNA and DNA level had similar decrease degrees in HBeAg-negative patients (Fig. 2F).
Dynamic change patterns of serum HBV RNA in HBeAg-positive patients with CHB on NA treatment by serum HBeAg conversion response
Before initiation of NA treatment, there were no significance difference in the serum HBV RNA and HBV DNA levels between HBeAg-positive patients with or without the HBeAg SC response during the 96 weeks of treatment (Figs. 3A and 3B). However, from week 12 of treatment, absolute values of serum HBV RNA, HBsAg, and HBeAg in patients with the SC response were significantly lower than those in patients without the SC response (Figs. 3B–3D). Consistent with this result, serum HBsAg and HBeAg titers in patients with the HBeAg SC response decreased rapidly from baseline during the first 12 weeks of treatment than that in patients without the HBeAg SC response (Figs. 2G and 2H). However, no significant difference was found between serum HBV RNA and DNA decrease degree in patients with different HBeAg SC responses (Figs. 3E and 3F).
Value of serum HBV RNA in monitoring the HBeAg SC response of NA therapy
Considering the strong correlation between serum HBV RNA and HBeAg level in HBeAg positive CHB patients before initiation and during NA treatment (Figs. 1A-C), and the significant difference in serum HBV RNA, HBsAg, and HBeAg titers in patients who achieved different HBeAg SC responses on NA treatment (Figs. 3), we illustrated the potential utility of serum HBV RNA in monitoring the HBeAg response of NA therapy. The ROC curves and area under curve (AUROC) scores of absolute and change from baseline levels for all viral indicators are shown in Figs. 4A–4E. Based on the AUROC scores of the absolute values of serum HBV RNA, HBeAg, and HBsAg, the SC responders could be well identified as early as 12 weeks of treatment, with the AUROC scores of 0.703, 0.780, and 0.761, respectively (Fig. 4B). At 24 weeks of treatment, only the absolute values of HBeAg and HBsAg demonstrated good discriminatory ability, with AUROC scores of 0.789 and 0.717, respectively (Fig. 4C). The absolute value of serum HBV DNA consistently showed poor discriminatory ability, with AUROC score ≤ 0.631 (Figs. 4A–4C).
Serum HBV RNA level decrease degree from baseline at 12 and 24 weeks of NA treatment showed poor discriminatory ability, with low AUROC scores (Figs. 4D and 4E), but those for serum HBsAg and HBeAg titers showed similar discriminatory ability to their absolute value at the same time point, especially the HBeAg titer (Figs. 4D and 4E).
Cutoff value of HBV biomarkers to predict the HBeAg SC response of NA therapy
Considering the higher discriminatory ability of HBV biomarkers’ absolute value and decrease degree from baseline at the early stage of treatment for the HBeAg response of NA treatment and the characteristics of HBV biomarkers’ dynamics during NA treatment, the cutoff values of these HBV biomarkers at 12 and 24 weeks of treatment were identified. Moreover, their performance was compared.
At the early stage of treatment, serum HBV RNA level consistently showed good predictive ability for the HBeAg response during 96 weeks of NA treatment. The absolute values of serum HBV RNA of 4.94 (log cps/mL) at 12 weeks and 4.31 (log cps/mL) at 24 weeks of treatment both had good negative predictive value (NPV) for the HBeAg SC response (90.48% vs. 94.74%), and both possessed high sensitivity (83.33% vs. 90.0%, respectively). Classical viral indicators, HBsAg cutoff value of 3.46 log10 IU/mL at 12 weeks of treatment, and HBeAg value of 2.87 log10 S/CO at 12 weeks of treatment also had good NPV% (> 80%) for the HBeAg SC response (Table 2). The 24-week HBeAg cutoff value of 0.306 log10 S/CO showed a two-way predictive capability for the HBeAg response (PPV% and NPV% both > 80%; Table 2). In addition to the absolute cutoff value of HBeAg, the titer decreases from baseline with a cutoff value of 1.55 log IU/mL at 24 weeks of treatment had a good NPV% (86.47%) and high specificity (86.49%) (Table S1).
Table 2
Absolute cutoff value of HBV biomarkers at 12 and 24 weeks of treatment for predict HBeAg seroconversion.
characteristics | Cutoff values | Sensitivity, % | Specificity, % | PPV, % | NPV, % |
12 weeks of treatment |
HBV RNA, log10 cps/mL | 4.94 | 83.3 | 55.88 | 40 | 90.48 |
HBeAg, log S/CO | 2.87 | 78.57 | 32.5 | 28.95 | 81.25 |
HBsAg, log10(IU/ml) | 3.46 | 84.62 | 59.46 | 42.31 | 91.67 |
24weeks of treatment |
HBV RNA, log10 cps/mL | 4.31 | 90.0 | 62.07 | 45 | 94.74 |
HBeAg, log10 S/CO | 0.306 | 57.15 | 95 | 80 | 86.36 |
HBsAg, log10(IU/ml) | 2.89 | 57.14 | 85 | 57.14 | 57.14 |
HBV: hepatitis B virus, HBsAg: Hepatitis B surface antigen, HBeAg: hepatitis B envelop antigen; PPV positive predictive value, NPV negative predictive value; cps/mL: copies/mL; cutoff value of HBV biomarkers was obtained from ROC analysis; |
Serum HBV RNA combined with a sharp decline in HBeAg can better predict the HBeAg response
To find better predictors for the HBeAg response, we explored the predictive value of serum HBV RNA in combination with other classical HBV biomarkers and found that serum HBV RNA combined with HBeAg had good predictive value and high AUROC scores (0.854, 95% CI: 0.754–0.954; Table S2). The absolute cutoff value of serum HBV RNA of 4.31 combined with HBeAg decrease degree of 1.55 at 24 weeks of treatment had a good two-way predictive capability for the HBeAg response (PPV%: 83.33% and NPV%: 81.25%), and the specificity was 96.30% (Table S2).