Tenofovir Alafenamide in the Treatment of Chronic Hepatitis B Virus Infection in the High‐Replicative Low‐Inflammatory Phase: A 48‐Week Randomized Controlled Trial Authors: Qiumin Luo, Ruixuan Xu, Yeqiong Zhang, Wenxiong Xu, Xiangyong Li, Jing Lai, Jianguo Li, Xingrong Zheng, Hong Deng, Lubiao Chen, Xiang Zhu, Chan Xie, Liang Peng Published: 2026 ABSTRACT ABSTRACTAntiviral treatment for chronic HBV infection in the “high‐replicative low‐inflammatory” phase has not been widely recommended. This study aimed to evaluate the safety and efficacy of antiviral therapy in this patient population and analyze peripheral immunological characteristics in relation to treatment response. In this randomized controlled trial, HBeAg‐positive patients with normal ALT and elevated HBV DNA were randomly allocated 1:1 to receive TAF 25 mg/day or observation. The primary endpoint was the decline from baseline in HBsAg at Week 48. Secondary endpoints included HBV DNA response rates and the magnitude of HBV DNA reduction. The treatment group was further divided into virological response (VR) and low‐level viremia (LLV) subgroups based on HBV DNA response at Week 48. PBMCs from VR (n = 3) and LLV (n = 3) patients underwent scRNA‐seq for comparative immunological analysis. A total of 59 patients were allocated to the treatment (n = 30) and control (n = 29) groups. No serious adverse events occurred, except for one control patient who developed an ALT flare and required antiviral initiation. At Week 48, the treatment group demonstrated significantly greater reductions in HBsAg (0.19 vs. 0 log10 IU/mL, p = 0.005) and HBV DNA (6.32 vs. 0.19 log10 IU/mL, p < 0.001) compared to controls. In the treatment group, 20% achieved HBV DNA < 20 IU/mL and 3.3% achieved HBeAg seroconversion, while none occurred in the untreated patients. PBMCs scRNA‐seq revealed dominant distributions of NK_FCER1G, NK_KLRF1, NK_XCL1, and NKT_FCGR3A subclusters in the VR group, with upregulated expression of genes such as SLC27A4 and PTMA. Taken together, our findings revealed that TAF demonstrates favorable safety and antiviral efficacy in patients with high‐replicative, low‐inflammatory HBV infection during the 48‐week follow‐up period, although complete virological suppression rates remain suboptimal. NK‐mediated mechanisms may contribute to antiviral success. Trial Registration: The trial was registered on ClinicalTrials.gov (NCT 04231565). Abbreviations alanine aminotransferase cell–cell interactions chronic hepatitis B chitinase‐3‐like protein 1 differentially expressed genes full analysis set hepatitis B virus hepatocellular carcinoma high‐replicative low‐inflammatory low‐level viremia liver stiffness measurement ligand‐receptor modified intention‐to‐treat mononuclear phagocytes nucleos(t)ide analog natural killer peripheral blood mononuclear cells principal component analysis plasmacytoid dendritic cells single‐cell RNA sequencing statistical package for the social sciences tenofovir alafenamide fumarate tumor necrosis factor receptor superfamily uniform manifold approximation and projection virological response Background Chronic hepatitis B virus (HBV) infection poses a significant global health burden. HBeAg‐positive chronic HBV infection constitutes the first stage in its natural history, traditionally termed the “immune‐tolerant phase”. Based on the Polaris Observatory Collaborators' report and China's epidemiological data on chronic hepatitis B (CHB), a study estimated 59.4 million individuals worldwide remain in this phase, including 15.84 million cases in China [1]. Emerging evidence demonstrates the presence of HBV DNA integration, clonal hepatocyte expansion, and HBV‐specific T cell responses in immune‐tolerant phase patients, challenging the notion that so‐called immune‐tolerant patients are immunologically inert and suggesting that the term “high‐replicative low‐inflammatory (HRLI)” phase may more accurately reflect the clinical characteristics of this stage, which is defined by positive HBeAg, highly elevated HBV DNA, normal alanine aminotransferase (ALT) levels, and absence of significant necroinflammation or fibrosis [2, 3]. Current guidelines maintain conservative recommendations regarding antiviral therapy for HRLI patients compared to those with chronic active hepatitis [4, 5], stemming from the traditional view that the histological progression risk is minimal. Additional considerations include the prolonged treatment duration requirements, along with concerns about long‐term drug safety, risk of resistance, and patient adherence challenges [6, 7, 8, 9]. However, recent studies have demonstrated that even in HBeAg‐positive patients with normal ALT levels, a certain proportion still exhibits significant hepatic inflammation (≥ G2) or significant fibrosis (≥ S2) [10, 11]. Although the risk of hepatocellular carcinoma (HCC) is lower than in chronic active hepatitis or cirrhosis, it remains significantly higher than in non‐HBV carriers [12]. A Korean retrospective cohort study of non‐cirrhotic patients with high viral load demonstrated significantly higher 10‐year HCC incidence in untreated patients with normal ALT versus treated chronic hepatitis B patients (12.7% vs. 6.1%) [13]. Accordingly, the Chinese Chronic Hepatitis B Prevention and Treatment Guidelines have expanded antiviral indications [14]. However, for HRLI patients without liver fibrosis or a family history of HCC, particularly young individuals, the “Treat All” approach has not been implemented. Considerations for the long‐term efficacy and safety of current first‐line oral antiviral agents (e.g., entecavir, Tenofovir Disoproxil Fumarate [TDF], Tenofovir Alafenamide Fumarate [TAF]). Several studies have suggested that long‐term entecavir therapy is associated with an increased risk of drug resistance (especially in lamivudine‐exposed patients), while long‐term TDF treatment is associated with an increased risk of renal and bone impairment [15, 16]. TAF was approved for the treatment of HBV‐infected adults and adolescents over 12 years old by the Food and Drug Administration (FDA) in November 2016. Compared with TDF, TAF exhibits improved renal and bone safety without a loss of antiviral efficacy [17]. A retrospective study showed that TAF treatment in chronic HBV‐infected patients with normal ALT was well tolerated, with no severe adverse events occurring [18]. In this prospective study, we aimed to assess the safety and efficacy of TAF in HRLI patients with chronic HBV infection, providing further evidence for whether to further expand treatment indications. Methods Patients and Study Design This prospective, open‐label randomized controlled trial enrolled HBeAg‐positive chronic HBV‐infected individuals with normal ALT levels from the Infectious Diseases Outpatient Clinic of the Third Affiliated Hospital of Sun Yat‐Sen University. Inclusion criteria were: (1) HBsAg positivity > 6 months; (2) HBeAg positivity with HBV DNA > 20 IU/mL; (3) ALT ≤ 1 × upper limit of normal (ULN, 40 U/L); (4) age 18–65 years; (5) no prior interferon or nucleos(t)ide analog (NA) therapy within 1 year; (6) liver stiffness measurement (LSM) < 5.8 kPa. Exclusion criteria included: (1) use of immunosuppressive agents; (2) pregnancy; (3) coinfection with hepatitis A, C, D or E viruses; (4) liver fibrosis, cirrhosis or HCC on imaging; (5) coexisting liver diseases of other etiologies; (6) HIV or other immunodeficiency disorders; (7) comorbidities including diabetes, renal insufficiency, autoimmune diseases; and (8) family history of cirrhosis or HCC. Eligible participants were randomly assigned to treatment or control groups in a 1:1 ratio, using computer‐generated randomization sequences derived from the Statistical Package for the Social Sciences (SPSS). Following approval by the ethics committee and obtaining informed consent, patients sequentially drew sealed envelopes to determine their group assignment. The treatment group received TAF 25 mg/day with meals, while controls underwent observation only, with TAF initiated if they presented with ALT > 1 × ULN along with detectable HBV DNA during follow‐up. All participants underwent comprehensive evaluations at weeks 0, 24, and 48, including virological (HBV DNA, HBV pgRNA), serological (HBsAg quantification, HBeAg), biochemical, and fibrotic markers (Chitinase‐3‐like protein, LSM). Adherence monitoring was also conducted. The study protocol has previously been published [19], with the study flowchart shown in Figure S1. Peripheral Blood Single‐Cell RNA Sequencing TAF treatment group patients were categorized based on their HBV DNA response at Week 48 into the virologic response (VR, defined as HBV DNA < 20 IU/mL) and the low‐level viremia (LLV, defined as HBV DNA 20 ~ 2000 IU/mL) subgroups. Three patients were randomly selected from each subgroup for peripheral blood collection (2 mL) at Week 48 of treatment. Peripheral blood mononuclear cells (PBMCs) were isolated using Ficoll‐Paque Plus according to the manufacturer's instructions. Cell viability and concentration were recorded using a fluorescence counter. Single‐cell suspensions at a concentration of 1 × 10⁵ cells/mL were loaded into microfluidic chips. The scRNA‐seq libraries were constructed using the GEXSCOPE single‐cell sequencing kit, followed by cDNA synthesis via mRNA reverse transcription and amplification. cDNA underwent fragmentation, adapter ligation, and was sequenced using 150 bp paired‐end reads on the Illumina Novaseq. 6000 platform. Table S2 shows the data quality control report for the six peripheral blood samples used for single‐cell RNA sequencing in our study. These data are not derived from a public database and do not involve citations. Supplementary Methods include the method for peripheral blood single‐cell RNA sequencing and the relevant analysis software. Laboratory Assessments Serum HBV DNA levels were quantified using the qPCR internal standard method (Abbott), with a detection range of 10 IU/mL to 1 × 109 IU/mL. Serum HBsAg levels were measured using a quantitative assay kit from Roche, with a detection range of 0.05 IU/mL to 52 000 IU/mL. HBeAg and anti‐HBe status were determined using enzyme‐linked immunosorbent assay (ELISA) kits. HBV pgRNA was detected using real‐time RT‐PCR with fluorescent probes (Supbio), with a detection range of 15 copies/mL to 1 × 10⁸ copies/mL. The Fibro‐CHI chemiluminescent assay (Proprium Diagnostics) was employed to measure chitinase‐3‐like protein 1 (CHI3L1) levels, with values ≥ 79 ng/mL indicating significant hepatic fibrosis. LSM values were obtained using transient elastography (FibroScan; Echosens). Outcomes The primary endpoint was the reduction in HBsAg levels from baseline to Week 48. Secondary endpoints included: HBsAg loss (defined as HBsAg < 0.05 IU/mL) rate, HBeAg seroconversion (defined as HBeAg < 1 COI) rate, reduction in HBeAg levels from baseline, virologic response (defined as HBV DNA < 20 IU/mL) rate, reduction in HBV DNA levels from baseline, and decline in HBV pgRNA levels from baseline. Safety evaluations included monitoring for biochemical abnormalities and documenting adverse events. Statistical Analysis Given the limited number of previous studies on TAF treatment for HBV‐infected patients with normal ALT levels, sample size estimation was based on data from antiviral therapy outcomes in CHB patients. Previous research revealed that 1‐year treatment with TAF resulted in an average HBsAg decline of 0.65 log₁₀ IU/mL from baseline in HBeAg‐positive CHB patients [20]; while in untreated patients, HBsAg declines slowly through the natural course [21]. The annual rate of spontaneous HBsAg clearance is approximately 0.5%–1% [22]. This indirectly indicates that changes in HBsAg levels within a 1‐year time window may be minimal or show no significant decrease in untreated patients with HBV chronic infection. Therefore, we set the 1‐year decline in HBsAg among untreated patients to 0. Using PASS software, with 80% statistical power, a two‐sided α level of 0.05, and a 10% dropout rate, the required sample size was 42 patients per study group. All analyses adhered to the modified intention‐to‐treat (mITT) principle, where all randomized patients who received at least one dose of TAF or had post‐baseline data were included in the full analysis set (FAS) for efficacy and safety assessments. If missing data were < 20%, these data were imputed using the average of the two adjacent measurements from the same patient for the respective variable. Continuous variables were compared using Student's t‐test (for normally distributed data) or Wilcoxon rank‐sum test (non‐normally distributed), while categorical variables were analyzed with Pearson's χ 2 test or Fisher's exact test. Statistical analyses were performed using SPSS 23.0, and graphs were generated using GraphPad Prism 6. For single‐cell analysis, cell annotation was performed based on transcriptomic profiles and marker gene expressions. The Seurat package (v3.1.2; https://github.com/satijalab/seurat) was used for cell type identification and clustering analysis of RNA‐seq data. To identify differentially expressed genes (DEGs), we used the Seurat FindMarkers function based on the Wilcoxon rank sum test with default parameters, and selected the genes expressed in more than 10% of the cells in both of the compared groups of cells and with an average log2(fold change) value greater than 0.25 as DEGs. Adjusted p value was calculated by Bonferroni Correction, and the value 0.05 was used as the criterion to evaluate the statistical significance. This study adhered to the ethical guidelines of the 1975 Declaration of Helsinki. Approval was granted by the Medical Ethics Committee of the Third Affiliated Hospital, Sun Yat‐sen University (approval number [2019]02‐599‐01). The trial was registered on ClinicalTrials.gov (NCT 04231565). Results Patient Enrollment From June 4, 2020, to September 19, 2023, a total of 59 eligible patients were randomly assigned to the treatment group (n = 30) and control group (n = 29), and included in the FAS set. By September 2023, four protocol deviations were documented: one participant in the treatment group voluntarily discontinued treatment at Week 16 but resumed antiviral treatment at Week 24 following virologic rebound, while three participants in the control group required TAF initiation during follow‐up. Nine patients were lost to follow‐up, yielding a cohort of 46 participants who adhered to the protocol and completed the 48‐week follow‐up, as detailed in Figure 1. Flowchart of study subjects. @, According to the modified intention‐to‐treat (mITT) principle, all patients enrolled were classified into the full analysis set (FAS) to evaluate efficacy and safety. *A 32‐year‐old female patient was administered glucocorticoids 4 weeks post‐enrollment following a diagnosis of scleroderma, with concurrent initiation of TAF antiviral therapy. ^A 22‐year‐old patient exhibited an ALT flare (199 U/L) at 24 weeks post‐enrollment, indicative of hepatitis activity, and thus initiated TAF therapy. After 24 weeks of treatment, HBV DNA became undetectable and ALT levels normalized. ※A 38‐year‐old patient exhibited coarsened hepatic echotexture on ultrasonography at 24 weeks post‐enrollment, consistent with chronic hepatitis sonographic findings, and thus initiated TAF therapy despite normal ALT levels. Baseline Characteristics The TAF group exhibited marginally higher baseline ALT levels than the control group (26.6 vs. 22.8 U/L, p = 0.033), though both were within the normal range. No significant differences were observed in baseline demographic and clinical characteristics between the two groups (Table 1). Table 1Baseline characteristics of all patients in the FAS.Total (n = 59)TAF (n = 30)Control (n = 29) p a Age (years)28.9 ± 5.529.1 ± 5.128.7 ± 6.00.776Male, n (%)19 (32.2%)9 (30%)10 (34.5%)0.713WBC (109/L)6.0 ± 1.16.2 ± 1.15.8 ± 1.10.147HGB (g/L)138.6 ± 15.3137.6 ± 15.9139.6 ± 14.90.623PLT (109/L)252 (225, 289)264 (240, 305.3)246 (207, 282)0.050ALT (U/L)24.7 ± 6.926.6 ± 6.922.8 ± 6.50.033AST (U/L)22 (19, 27)22 (19.8, 27.3)22 (19, 24.5)0.416ALB (g/L)46.6 ± 3.146.8 ± 2.746.5 ± 3.60.712TBIL (μmol/L)10.7 (6.7, 15.7)11.2 (6.3, 16.9)10.1 (7.0, 15.2)0.802DBIL (μmol/L)2.3 (1.5, 3.4)2.4 (1.5, 4.0)2.1 (1.5, 3.2)0.524GGT (U/L)16.3 (13, 20)17.5 (12.8, 23.6)16 (13, 19)0.251TBA (μmol/L)3.8 (2.9, 5.3)3.7 (2.8, 4.2)4 (3.4, 6.0)0.107BUN (mmol/L)4.4 ± 1.14.4 ± 1.34.4 ± 0.80.902Cr (μmol/L)55 (47, 66)56 (47, 64.5)55 (46, 71.3)0.844eGFR (mL/min/1.73m2)123.6 ± 9.9123.9 ± 8.9123.4 ± 10.90.843P (mmol/L)1.2 ± 0.21.2 ± 0.11.1 ± 0.20.288CHOL (mmol/L)5.2 ± 0.95.0 ± 0.85.3 ± 1.10.129TG (mmol/L)1.0 ± 0.40.9 ± 0.41.0 ± 0.50.213LDL‐C (mmol/L)3.0 ± 0.72.9 ± 0.73.1 ± 0.80.553HDL‐C (mmol/L)1.4 ± 0.31.4 ± 0.21.5 ± 0.40.229AFP (ng/mL)2.2 ± 1.02.1 ± 1.02.2 ± 1.10.666HBV DNA (log10, IU/mL)8.21 (7.77, 8.42)8.22 (7.61, 8.43)8.18 (8.03, 8.43)0.524HBV pgRNA (log10, copies/mL)6.32 (5.97, 6.65)6.24 (5.85, 6.63)6.38 (6.09, 6.82)0.554HBsAg (log10, IU/mL)4.69 (4.42, 4.72)4.68 (4.37, 4.72)4.70 (4.48, 4.72)0.397HBeAg (log10, COI)3.19 (3.13, 3.23)3.20 (3.14, 3.23)3.18 (3.11, 3.23)0.633CHI3L1 b(ng/mL)33.5 ± 15.034.9 ± 17.031.9 ± 12.80.545LSM (KPa)5.0 ± 0.64.9 ± 0.55.1 ± 0.60.257NAFLD, n (%)4 (6.8%)3 (10%)1 (3.4%)0.306 Note: Data are expressed as mean ± SD, median (IQR), or n (%). a Indicates the significance of the difference between the TAF group and the control group. b 20 and 19 patients underwent CHI3L1 testing in the TAF group and the control group, respectively.John Wiley & Sons, Ltd. Baseline characteristics of all patients in the FAS. Note: Data are expressed as mean ± SD, median (IQR), or n (%). Indicates the significance of the difference between the TAF group and the control group. 20 and 19 patients underwent CHI3L1 testing in the TAF group and the control group, respectively. Safety of Antiviral Therapy No severe adverse events were documented during the study. During follow‐up, patients in the TAF (n = 1) and control (n = 1) groups exhibited a transient increase in CK (> 1000 U/L) after intense exercise, which resolved spontaneously with rest. At Week 48, the incidence of ALT elevation (ALT > 35 U/L) was comparable between the TAF group and the control group (23.3% vs. 27.6%, respectively; p = 0.708). Only one patient in the control group exhibited an ALT flare (increased to 199 U/L), prompting immediate TAF initiation. All other ALT elevations remained below 2×ULN. Besides, there were no significant differences between the groups in renal function parameters, blood lipid profiles, or other biochemical indices. The biochemical profiles at Week 48 are shown in Table 2. Table 2Biochemical parameters of both groups at Week 48.TAF (n = 26)Control (n = 23) p 95% CI for difference Z ALT (U/L)23.2 ± 927 ± 10.50.183(−9.4, 1.8)AST (U/L)20.5 (18, 24.3)22 (19, 26)0.376−0.885ALB (g/L)45.5 ± 2.446.2 ± 2.70.360(−2.2, 0.8)TBIL (μmol/L)11 ± 3.711 ± 4.30.998(−2.3, 2.3)GGT (U/L)15 (12.8, 23.5)18 (13, 21)0.841−0.201Cr (μmol/L)58.9 ± 1357 ± 19.50.694(−7.6, 11.4)eGFR120 ± 9.3120.9 ± 11.10.748(−6.9, 5.0)P (mmol/L)1.1 ± 0.11.2 ± 0.10.068(−0.1, 0)CHOL (mmol/L)4.9 ± 0.75 ± 0.90.844(−0.5, 0.4)Δ CHOL (mmol/L)−0.01 ± 0.71−0.19 ± 0.740.416(−0.26, 0.61)TG (mmol/L)1 ± 0.61.1 ± 0.50.503(−0.5, 0.2)Δ TG (mmol/L)0.03 ± 0.420.07 ± 0.350.709(−0.27, 0.19)LDL‐C (mmol/L)2.9 ± 0.72.9 ± 0.90.842(−0.4, 0.5)Δ LDL‐C (mmol/L)0.01 ± 0.58−0.13 ± 0.680.435(−0.23, 0.53)HDL‐C (mmol/L)1.4 ± 0.31.5 ± 0.80.291(−0.5, 0.2)Δ HDL‐C (mmol/L)0.01 ± 0.190.13 ± 0.780.486(−0.47, 0.23) Note: Δchanges in CHOL, TG, LDL‐C, and HDL‐C from baseline to Week 48.John Wiley & Sons, Ltd. Biochemical parameters of both groups at Week 48. Note: Δchanges in CHOL, TG, LDL‐C, and HDL‐C from baseline to Week 48. At Week 48, patients (n = 7, 26.9%) in the TAF and control (n = 2, 8.7%) groups showed evidence of fatty liver on follow‐up ultrasound examinations (p = 0.091). No imaging evidence of liver cirrhosis was observed in any patient at this time point. Besides, no significant differences were found in fibrosis markers between the two groups, including CHI3L1 levels (38.1 vs. 37.5 ng/mL, p = 0.951) and LSM (4.7 vs. 4.6 kPa, p = 0.757). Treatment Efficacy As shown in Table 3 and Figure 2, serological analyses indicated that the TAF group showed a significantly greater decrease in HBsAg from baseline compared to the control group starting from Week 12. At Week 48, the median difference in HBsAg from baseline between the two groups was −0.19 and 0 log10 IU/mL (p = 0.005), respectively. However, no patient experienced HBsAg loss. The reduction in HBeAg from baseline was also significantly greater in the TAF group. One patient in the TAF group (3.3%) experienced HBeAg loss and seroconversion, while none occurred in the control group. Assessment of virological outcomes revealed that HBV DNA levels in the TAF group decreased significantly from baseline as early as Week 12. By Week 48, the median reductions from baseline were 6.32 log10 IU/mL in the TAF group versus 0.19 log10 IU/mL in the control group (p < 0.001). The median HBV DNA levels at Week 48 were 2.01 log10 IU/mL in the TAF group and 8.15 log10 IU/mL in the control group (p < 0.001). As shown in Table S1, the results from the per‐protocol set (PPS) analysis were consistent with those from the FAS analysis. The TAF group achieved an HBV DNA response rate of 20% (6/30), with the remaining patients having HBV DNA levels between 20 ~ 2000 IU/mL. In the control group, three patients required initiation of antiviral therapy during the observation period (Figure 1). Notably, one case involved treatment initiation following an ALT elevation event; subsequent TAF administration resulted in virologic suppression (HBV DNA < 20 IU/mL) by Week 24 of therapy. The longitudinal course of the patient who developed an ALT flare during the observation period is presented in Figure S2. No significant differences were observed in HBV pgRNA levels between the two groups. Table 3The dynamic changes in serological and virological parameters.TAFControl Z p 12 weeks(n = 30)(n = 29)HBsAg (log10, IU/mL)4.59 (4.24, 4.71)4.67 (4.52, 4.72)−1.5150.130Δ HBsAg (log10, IU/mL)−0.03 (−0.11, 0)0 (−0.02, 0.03)−2.4000.016HBeAg (log10, COI)3.19 (3.15, 3.22)3.20 (3.16, 3.23)−0.5990.549Δ HBeAg (log10, COI)−0.01 (−0.06, 0.02)0 (−0.04, 0.05)−1.6540.098HBV DNA (log10, IU/mL)3.83 (2.86, 4.57)8.28 (7.99, 8.42)−6.153< 0.001Δ HBV DNA (log10, IU/mL)−4.15 (−4.91, −3.48)0.02 (−0.18, 0.16)−6.313< 0.00124 weeks(n = 30)(n = 28)HBsAg (log10, IU/mL)4.59 (4.14, 4.68)4.67 (4.50, 4.72)−2.2430.025Δ HBsAg (log10, IU/mL)−0.08 (−0.13, 0)0 (−0.02, 0.02)−3.0440.002HBeAg (log10, COI)3.18 (3.12, 3.24)3.21 (3.14, 3.25)−0.7080.479Δ HBeAg (log10, COI)−0.01 (−0.08, 0.02)0.01 (−0.02, 0.07)−1.5560.120HBV DNA (log10, IU/mL)2.77 (1.67, 3.11)8.22 (7.94, 8.35)−5.711< 0.001Δ HBV DNA (log10, IU/mL)−5.33 (−5.86, −4.91)−0.04 (−0.32, 0.13)−5.944< 0.001HBV pgRNA (log10, copies/mL)6.59 (6.29, 7.38)6.73 (6.13, 7.07)−0.2560.798Δ HBV pgRNA (log10, copies/mL)0.45 (0.07, 0.71)−0.05 (−0.18, 0.65)−1.9020.05748 weeks(n = 26)(n = 23)HBsAg (log10, IU/mL)4.46 (3.91, 4.61)4.66 (4.51, 4.72)−3.2390.001Δ HBsAg (log10, IU/mL)−0.19 (−0.25, −0.06)0 (−0.13, 0)−2.8010.005HBeAg (log10, COI)3.14 (3.09, 3.21)3.19 (3.14, 3.24)−1.1920.233Δ HBeAg (log10, COI)−0.07 (−0.15, 0.01)0.01 (−0.06, 0.06)−2.2140.027HBV DNA (log10, IU/mL)2.01 (1.29, 2.13)8.15 (7.88, 8.46)−5.412< 0.001Δ HBV DNA (log10, IU/mL)−6.32 (−6.52, −5.69)−0.19 (−0.63, 0.12)−5.289< 0.001HBV pgRNA (log10, copies/mL)6.73 (6.11, 7.16)6.88 (6.23, 7.63)−1.1000.271Δ HBV pgRNA (log10, copies/mL)0.46 (0.03, 0.77)0.55 (−0.11, 1.12)−0.3630.717 Note: Δ, changes from baseline.John Wiley & Sons, Ltd. The dynamic changes in serological and virological parameters. Note: Δ, changes from baseline. Virological and serological dynamics in the two groups. (A–D) Dynamics of HBV DNA, HBV pgRNA, HBsAg, and HBeAg values; (E–H) Changes from baseline in HBV DNA, HBV pgRNA, HBsAg, and HBeAg in both groups. The figures show the median and IQR. # p > 0.05; *p < 0.05; **p < 0.01; ***p < 0.001. Peripheral Immunological Characteristics of Patients With Different Virological Responses Peripheral Blood Immunocellular Profiling of the VR and LLV Patients PBMCs were extracted from VR (n = 3) and LLV (n = 3) patients at Week 48 and subjected to single‐cell transcriptome sequencing (Figure 3A) to explore peripheral immunological differences between patients with differing virological responses. The serological, virological, and biochemical parameters of patients who underwent scRNA‐seq are presented in Figure S3. Single‐cell transcriptome profiles of PBMC in chronic HBV‐infected patients following TAF treatment. (A) Experimental design and virological dynamics in relevant patients following TAF treatment; (B) UMAP of the 54994 single cells from 6 patients, coloured by 7 major clusters; (C) violin plots of expressions of marker genes in each cell type in each cell cluster; (D) the proportion of major cell clusters from each individual. The numbers around the pie chart represent the average cell proportion of these cell clusters; (E) UMAP plots of 17 subclusters identified from T and NK cell cluster; (F) UMAP plots of T and NK cell subclusters coloured by VR group and LLV group; (G) the fraction of 17 subclusters from T and NK cells in each group; (H) comparison of the distribution of 17 subclusters from T and NK cells between the VR group and LLV group estimated by Ro/e score. Following rigorous quality control and screening (Table S2), we established a comprehensive single‐cell atlas comprising 54 994 cells from six samples, identifying seven major cell clusters, including Plasma cells, B cells, T and natural killer (NK) cells, neutrophils, mononuclear phagocytes (MPs), plasmacytoid dendritic cells (pDCs), and platelets (Figure 3B). Cell type identification was validated through characteristic marker gene expression patterns, including IGHG1 (plasma cells), MS4A1 (B cells), and GNLY (NK cells) (Figure 3C). T and NK cells were the most abundant across all samples, with significantly higher proportions observed in the VR group (Figure 3D). High‐resolution subclustering of T/NK cells identified 17 distinct cell subclusters (Figure 3E). To explore immunological differences between patients with varying antiviral responses, we compared the distribution of these subclusters between the VR and LLV groups. The results showed that NK_FCER1G, NK_KLRF1, NK_XCL1, and NKT_FCGR3A cell subclusters were significantly enriched in the VR group (Figure 3F–H). Differences in NK and NKT Cell Subclusters Between the VR and LLV Groups Further analysis of differentially expressed genes (DEGs) in the NK_FCER1G, NK_KLRF1, NK_XCL1, and NKT_FCGR3A subsets revealed that the VR group exhibited significantly upregulated expression of 91, 138, 125, and 115 genes, respectively, compared to the LLV group. Figure 4A illustrates the inter‐group differences in DEGs across these four cell subclusters. Notably, the SLC27A4 gene exhibited highly significant upregulation in the VR group across all subclusters. Besides, a Venn diagram analysis identified 15 shared highly expressed DEGs (such as PTMA and RPS3A) among the four cell subsets in the VR group (Figure 4B). To assess immune cell exhaustion, Ucell was applied to score exhaustion‐related gene signatures in the target cell subsets. As shown in Figure 4C–D, low overall exhaustion scores were obtained across all samples, with exhaustion scores in the VR group being significantly lower than those in the LLV group. We further analyzed CCIs based on L‐R binding strength. Both VR and LLV groups showed active communication among MPs, NK_FCER1G, NKT_FCGR3A, and Teff_BCAS3 subclusters with other cell types (Figure 4E). Moreover, the NK_FCER1G subcluster in the VR group exhibited significantly more interactions with NKT_FCGR3A and Teff_BCAS3 cells compared to the LLV group (Figure 4F). Some L‐R pairs exhibited highly significant upregulation only in the VR group, including BAG6‐NCR3 and CD58‐CD2 (between NKT_FCGR3A as ligand and NK_FCER1G as receptor), as well as LTA‐TNFRSF1A, LTA‐TNFRSF14, TNF‐TNFRSF1A, and CD58‐CD2 (between Teff_BCAS3 as ligand and NK_FCER1G as receptor), as shown in Figure 4G. Differences in NK/NKT cell subclusters between the VR and LLV groups. (A) Volcano plots showing the top 10 DEGs of NK_FCER1G, NK_KLRF1, NK_XCL1 and NKT_FCGR3A subclusters between VR and LLV patients; (B) the expression of 15 upregulated DEGs shared by these four cell subclusters in the VR patients between the two groups; (C) UMAP plots of exhaustion in the overall samples (scoring of exhaustion gene sets in T and NK cluster by Ucell); (D) violin plots showing intergroup differences in exhaustion of NK_FCER1G, NK_KLRF1, NK_XCL1 and NKT_FCGR3A subclusters, with statistical significance determined by Wilcoxon rank‐sum test between VR and LLV groups. **p < 0.01; ***p < 0.001; (E) Heatmap displaying the CCIs. The scale bar shows the count of ligand‐receptor (L‐R) pairs between two subclusters; (F) heatmap displaying the difference in the count of L‐R pair interactions between the two groups. Red indicates more L‐R pair interactions in the VR group, while blue represents more in the LLV group; (G) bubble plot showing L‐R pairs that significantly increased in interactions only in the VR group, with NK_FCER1G as the receptor cell. Discussion This study focused on chronic HBV‐infected patients in the HRLI phase, a population for whom antiviral therapy is not yet strongly recommended by the American Association for the Study of Liver Diseases and European Association for the Study of the Liver guidelines. Our results showed that patients treated with TAF had no severe adverse events within 48 weeks. No significant differences were observed between the two groups in biochemical parameters during follow‐up, indicating good tolerability of TAF in this specific population. The ALT elevation rates were comparable between the two groups. The sole case of ALT flare occurred in the control group, consistent with the natural history of CHB. Elevations in ALT levels in other patients were mild (< 2×ULN). In contrast, a study on TAF treatment for HBeAg‐positive CHB patients reported an ALT flare incidence of 11% during therapy [23]. This suggests that initiating antiviral therapy early in patients with high viral loads, before significant liver inflammation or fibrosis occurs, may reduce the risk of hepatitis flare‐ups. However, as the follow‐up period was only 48 weeks and the study population was relatively young (mean age, 28.9 years), the safety profile with long‐term use and in older patients remains to be investigated. Regarding efficacy, only 20% of patients in this study achieved HBV DNA < 20 IU/mL after 48 weeks of TAF therapy, which was lower than the response rates reported in previous antiviral treatments for CHB. In a phase 3 clinical study of TAF prior to its approval (Study 110), 64% of CHB patients with mildly elevated ALT achieved HBV DNA < 29 IU/mL after 48 weeks of TAF [23]. The primary reason for the observed differences in outcomes is likely attributable to disparities in the study populations. In the present study, participants presented with high HBV DNA (median HBV DNA > 8 log10 IU/mL) levels and normal ALT, characteristics more closely aligned with an immune‐tolerant phase or a state of low immune activity. A retrospective study consistently indicated that, compared to high baseline ALT ( > 100 U/L), low‐to‐moderate ALT elevation (< 100 U/L) combined with high HBV DNA load (> 6 log10 IU/mL) was an independent risk factor for LLV [24]. A higher baseline ALT may indicate greater immune activity, facilitating viral clearance. Although a substantial proportion of our treatment group achieved LLV at Week 48, our data still demonstrated positive and clinically meaningful outcomes: a substantial HBV DNA reduction from baseline (> 6 log10 IU/mL) and a significant HBsAg decline (0.19 log10 IU/mL), consistent with prior studies. A cohort study showed that HBeAg‐positive CHB patients experienced an HBsAg decline of 0.238 log10 IU/mL after 1 year of NA therapy [25], suggesting that HRLI patients can also achieve significant HBsAg reduction with NA treatment, similar to the response observed in traditionally recommended antiviral therapy for CHB. Persistent high‐level HBsAg stimulation is known to impair and exhaust HBV‐specific immune responses [26]. NA therapy, by reducing HBV DNA and HBsAg levels, may facilitate the restoration of antigen‐specific immunity [27], thereby establishing conditions for pursuing HBsAg clearance. Building upon the above findings, it can be inferred that patients in the HRLI phase may benefit from early antiviral intervention. However, no significant differences were observed in HBV pgRNA levels between the TAF group and the control group during the 48‐week follow‐up period. This potentially suggests that TAF primarily focuses on reverse transcriptase inhibition, while the direct effect on cccDNA‐derived transcription requires a longer treatment duration to manifest as a significant change in pgRNA levels. Recent research has shown that an analysis of liver scRNA‐seq data reveals a distinct increase in NK cells in the liver biopsies of patients with functionally cured CHB compared to those with uncured CHB, and these cells display a primed state [28]. To explore immunological differences between patients with varying virological responses, scRNA‐seq of PBMC samples from VR and LLV subgroups was performed in our study. Similarly, our results demonstrated predominance of NK and NKT cells in the VR group, suggesting that innate immunity may play a critical positive role in TAF treatment outcomes among HRLI individuals. Analysis of DEGs across these four dominating cell subclusters revealed that the SLC27A4 gene exhibited highly significant upregulation in the VR group across all the cell subtypes. SLC27A4, which encodes fatty acid transporter 4, a member of the solute carrier family, is now understood to be involved in fatty acid synthesis, β‐oxidation, and phospholipid synthesis [29]. Overexpression of SLC27A4 has been reported in multiple tumor cells. For instance, hepatic carcinoma cells overexpressing SLC27A4 exhibit enhanced uptake of exogenous monounsaturated fatty acids, thereby suppressing lipid peroxidation and ferroptosis, accelerating HCC progression [30]. Therefore, we hypothesize that overexpression of SLC27A4 in NK cells might similarly promote NK cell proliferation and activation. However, it warrants further functional validation experiments to test this hypothesis in future studies. Moreover, we observed increased PTMA expression in the VR group. PTMA encodes thymosin α, a peptide with immunomodulatory functions. A study by Geretz et al. showed that PTMA gene expression in CD4+T cells from HIV‐1‐infected individuals negatively correlated with viral RNA levels. In vitro experiments further confirmed that PTMA inhibits viral replication by limiting HIV‐1 transcription [31]. Further investigation is warranted to determine whether NK cells might similarly harness PTMA to inhibit HBV replication. Analysis of intergroup differences in CCIs revealed that NK_FCER1G receptor cells in the VR group exhibited close interactions with the ligand cells NKT_FCGR3A and Teff_BCAS3. Besides, several L‐R pairs exhibited increased expression exclusively in the VR group, including CD58‐CD2, BAG6‐NCR3, LTA‐TNFRSF1A, LTA‐TNFRSF14, and TNF‐TNFRSF1A. The binding of ligand CD58 and receptor CD2 is a crucial component of the immunological synapse, promoting NK cell activation and triggering its cytolytic effect [32]. Upon binding to the natural cytotoxicity receptor NKp30 (encoded by NCR3) on NK cells, the ligand BAG6 transmits potent activating signals that further stimulate NK cell cytotoxicity [32, 33]. Members of the tumor necrosis factor receptor superfamily (TNFRSF) mediate signaling through ligand–receptor interactions: binding of receptor TNFRSF1A to its ligands LTα or TNF, and receptor TNFRSF14 binding to ligand LTα, could trigger NK cell activation to produce pro‐inflammatory cytokines IFN‐γ and TNF‐α, thereby amplifying NK cell cytotoxicity [34]. Thus, we hypothesize that during TAF treatment of such HRLI patients, these enhanced CCIs may represent the critical signaling events required to break immune tolerance or initiate immune clearance, thereby activating NK cells and inducing pro‐inflammatory cytokine production to drive antiviral effects. However, further validation of this hypothesis is required. Previous study has showed that the frequencies of IFN‐γ produced by NKT and NK cells and TNF‐α produced by NK cells of patients in the “immune active” phase were significantly higher than those of patients in the “immune tolerance” phase [35]. Therefore, based on our study, we hypothesize that for patients in the “immune active” phase, whose NK cells already produce higher levels of pro‐inflammatory cytokines, NK cells may also play a crucial role in the treatment process. This will likewise constitute our future research agenda. The main limitation of this study is the relatively short follow‐up period of 48 weeks, which is insufficient to evaluate the impact of TAF therapy on long‐term outcomes such as the development of liver cirrhosis or HCC in patients. Additionally, the sample size enrolled in our study is limited. Furthermore, the scRNA‐seq comparison is cross‐sectional, rather than assessing longitudinal immune changes before and after TAF therapy; the observed immunological differences between patients with differing virological responses are based solely on transcriptomic analysis, and peripheral immune profiling may not fully reflect intrahepatic immune mechanisms. Thus, further experimental validation and exploration of the underlying mechanisms are required to confirm these findings. Conclusion This study demonstrates that TAF is well‐tolerated in HRLI patients with chronic HBV infection after 48 weeks of treatment, and significant reductions in HBsAg, HBeAg, and HBV DNA levels were observed, although the LLV rate remained relatively high. Therefore, early initiation of NA treatment may benefit this patient population. Furthermore, mechanistic insights suggest that NK cells may play a critical positive role in the virological response of these patients. Author Contributions Conceptualization by L.P., C.X., and W.X.X.; data curation by Q.M.L., R.X.X., and Y.Q.Z.; funding acquisition by L.P. and C.X.; investigation by W.X.X., X.Y.L., J.L., J.G.L., X.R.Z., H.D., L.B.C., and X.Z.; methodology by Q.M.L. and R.X.X.; project administration by L.P. and C.X.; supervision by L.P.; writing – original draft by Q.M.L.; writing – review and editing by Y.Q.Z. and L.P. Ethics Statement This study adhered to the ethical guidelines of the 1975 Declaration of Helsinki. Approval was granted by the Medical Ethics Committee of the Third Affiliated Hospital, Sun Yat‐sen University (approval number [2019]02‐599‐01). Consent Informed consent was obtained from all patients for being included in the study. Conflicts of Interest The authors declare no conflicts of interest. Supporting information Figure S1: Study protocol. Figure S2: The longitudinal course of the patient who developed an ALT flare during the observation period. A 22‐year‐old patient exhibited an ALT flare (199 U/L) at 24 weeks post‐enrollment, indicative of hepatitis activity, and thus initiated TAF therapy. After 24 weeks of treatment, HBV DNA became undetectable and ALT levels normalized. Figure S3: The serological, virological, and biochemical parameters of patients who underwent scRNA‐seq. Table S1: The dynamic changes in serological and virological parameters (per‐protocol set, PPS). Table S2: Data quality control report for peripheral blood single‐cell RNA sequencing. Supporting File