Abstract
Background and Aim
We assessed the effect of hepatitis B surface antigen (HBsAg) seroclearance (HBsAg‐loss) on liver fibrosis regression in patients with chronic hepatitis B (CHB) infection.
Method
CHB patients with recent documented HBsAg‐loss were age‐ and gender‐matched with treatment‐naïve HBeAg‐negative CHB infection. Paired assessment with transient elastography and enhanced liver fibrosis (ELF) measurements were performed and repeated at 3 years. Fibrosis regression was arbitrarily defined as decrease in ≥ 1 fibrosis stage by ELF, or combining with reduction > 30% in liver stiffness.
Results
A total of 142 HBsAg‐loss and 142 CHB subjects were recruited (median age 58.1 years, 51.4% male). A total of 1.8% (1.4% HBsAg‐loss vs 2.1% CHB) achieved combined endpoint of fibrosis regression at 3 years. When ELF‐only definition of fibrosis regression was used, 14.5% HBsAg‐loss and 16.9% CHB subjects achieved this endpoint, which was significantly associated with baseline ELF (hazard ratio (HR) 1.827, 95% confidence interval (CI) 1.085–3.075) and time since HBsAg‐loss (HR 2.688, 95% CI 1.257–5.748). While increasing time since HBsAg‐loss increased the proportion of ELF‐defined fibrosis regression, increasing age was also associated with significant fibrosis. Age of achieving HBsAg‐loss (ageSC) was independently associated with high baseline ELF values. Up to 52.3% and 63.8% subjects with ageSC > 50 had advanced fibrosis/cirrhosis at baseline and 3 years, respectively, compared with 5.9% and 20.6% in subjects with ageSC < 50.
Conclusion
Fibrosis regression occurred in a minority of subjects achieving HBsAg‐loss, which was not significantly different compared with subjects with persistent overt CHB. Subjects after achieving HBsAg‐loss, especially among those with ageSC > 50, should receive ongoing surveillance for liver‐related complications.
Keywords: Age, Chronic viral hepatitis, Fibrosis regression, Functional cure, HBV
Introduction
Liver cirrhosis is one of the leading causes of digestive disease‐related death, accounting for 1.32 million deaths worldwide and has increased by 47% in the past two decades, 1 accounting for 2.4% of all deaths globally. 2 Forty percent of cirrhosis‐related deaths in East Asia are related to chronic hepatitis B (CHB) infection. 3 Most CHB patients will be infected lifelong, and only a minority of patients (annually 0.5%–2.0%) will achieve functional cure, which is defined as seroclearance of hepatitis B surface antigen (HBsAg) from the blood according to international guidelines. 4 , 5 , 6 HBsAg seroclearance, also known as HBsAg‐loss, are associated with obvious clinical benefits. The risk of hepatocellular carcinoma (HCC) will be substantially lowered especially if HBsAg‐loss occurs before the age of 50 years, 7 , 8 and risk of liver decompensation, transplantation, and death is also reduced. 9 On the other hand, whether HBsAg‐loss produces similar beneficial effect on cirrhosis is uncertain. There are minimal data on the change in liver fibrosis burden after HBsAg‐loss. The scarcity of fibrosis data in patients with HBsAg‐loss is inevitable in the past when liver biopsy was the only validated method to assess liver fibrosis, as it would be unethical to subject these very stable patients to the potential risks of liver biopsy. One small study described the liver histology in 11 patients with HBsAg‐loss, which showed improved necro‐inflammation, hepatocyte necrosis, but no change in hepatic fibrosis. 10 It would be important to ascertain the beneficial effect of HBsAg‐loss, if any, on liver fibrosis regression.
Non‐invasive means to quantify liver fibrosis are now widely accepted as part of standard of care for patients with CHB. Examples of these modalities include serum‐based markers (e.g. enhanced liver fibrosis [ELF] test) as well as imaging‐based (e.g. transient elastography) methods. ELF test is a serum‐based measurement of three different markers for advanced liver fibrosis and cirrhosis, including tissue inhibitor of metalloproteinase 1 (TIMP‐1), amino‐terminal pro‐peptide of type III pro‐collagen (PIIINP) and hyaluronic acid (HA). ELF has been validated in histology‐based studies in CHB patients. 11 Transient elastography has good performance characteristics for liver fibrosis. It is safe, fast, reproducible, and is therefore widely adopted as a surrogate marker for liver fibrosis as recommended by international guidelines for CHB patients and other liver diseases. 12 To improve the performance of non‐invasive tests for liver fibrosis, a combination of tests for fibrosis assessment is increasingly being advocated. 13 Combining a serum‐based test and an imaging‐based test would increase the diagnostic accuracy and avoid more liver biopsies than using either test alone. 14
We aimed to compare the proportion of patients with fibrosis regression among subjects who achieved HBsAg‐loss and CHB patients without HBsAg‐loss. We hypothesize that HBsAg‐loss, as ‘functional cure’ endpoint and denotes disease quiescence, is associated with on‐going favorable effects on liver fibrosis and will increase the chance of fibrosis regression, defined by improvement in transient elastography‐determined liver stiffness and/or ELF.
Methods
Subjects
All patients were prospectively recruited from the Liver Clinic, Department of Medicine, Queen Mary Hospital, Hong Kong, which is a tertiary referral centre for the management of chronic liver diseases. CHB is defined as HBsAg positivity for > 6 months. Asian CHB patients, aged 18–75 years, with spontaneous HBsAg‐loss were recruited, where HBsAg‐loss was defined by sustained HBsAg‐negativity > 6 months in a prior known CHB patient. Matched controls were CHB patients with no HBsAg‐loss who were not on antiviral treatment, and they were recruited in a 1:1 ratio after matching for age (± 5 years) and gender. We specifically excluded antiviral‐treated patients as it is known to influence fibrosis regression significantly. 15 Other exclusion criteria were as follows: abnormal alanine aminotransferase (ALT; defined as > 1 time upper limit of normal, taken as 40 U/L 4 ) documented on two separate occasions of at least 6 months apart before the baseline assessment (elevated ALT will confound liver stiffness measurement), 16 concomitant liver disease (chronic hepatitis C and D infection, histologically confirmed non‐alcoholic steatohepatitis, Wilson's disease, autoimmune hepatitis, primary biliary cholangitis, and primary sclerosing cholangitis), significant alcohol intake (> 30 g/day for men or > 20 g/day for women), prior history of HCC or liver transplantation, or other significant medical illnesses. All recruited subjects received baseline liver assessment by transient elastography and serum ELF measurement, and the assessment were repeated at 3 years. The study protocol was approved by the Institutional Review Board of the University of Hong Kong/Hospital Authority Hong Kong West Cluster (HKU/HA HKW IRB reference number: UW 18‐700). The study was conducted in compliance with the principles of Good Clinical Practice and Declaration of Helsinki. All recruited subjects provided written informed consent prior to any study‐related procedures.
Clinical and laboratory assessment
Subjects age at recruitment, gender, body mass index (BMI), and presence of type 2 diabetes were documented. For subjects with HBsAg‐loss, the age at HBsAg‐loss (ageSC), and the time since HBsAg‐loss was recorded. Laboratory markers, including aspartate aminotransferase (AST), ALT, and platelet, were recorded. Qualitative HBsAg was assessed every 6 months, and hepatitis B virus (HBV) DNA levels were determined using Cobas Taqman assay (Roche Diagnostics, Branchburg, NJ, USA) with a lower limit of detection of 20 IU/mL.
Transient elastography
Liver stiffness measurement was performed by Fibroscan (Echosens®, Paris, France) using M probe for patients whose BMI was < 30 kg/m2 and XL probe for patients whose BMI was ≥ 30 kg/m2, respectively. Liver stiffness was expressed as the median value of ≥ 10 successful acquisitions in units of kilopascals (kPa). Measurements were considered reliable with a success rate of ≥ 60%, combined with an interquartile range of < 30%. Continued attenuation parameter (CAP), a transient elastography‐based modality for estimation of liver steatosis, was documented and values expressed in decibel per meter (dB/m). Operators performing the measurements were from our own centre who have received proper training from Echosens® and had prior experience of at least 500 transient elastography procedures. The degree of liver fibrosis was classified in accordance with the European Association for Study of Liver, Asociación Latinoamericana para el Estudio del Hígado clinical guidelines, 12 where minimal fibrosis (F1) was defined as liver stiffness (LS) < 6 kPa, gray zone as LS 6–9 kPa, advanced fibrosis (F3) as LS 9–12 kPa, and cirrhosis (F4) as LS > 12 kPa. As the optimal cut‐off values of LS for F3/F4 fibrosis in CHB with concurrent hepatic steatosis are not clear, the same cut‐off value for CHB was used in this study with or without concurrent hepatic steatosis. Steatosis was categorized as mild (CAP 248–267 dB/m), moderate (CAP 268–279 dB/m), and severe (≥ 280 dB/m). 17
Serum enhanced liver fibrosis test
Serum ELF score was calculated using the recommended formula: 2.278 + 0.851 × ln (HA) + 0.751 × ln (PIIINP) + 0.394 × ln (TIMP‐1). ELF scores of ≥ 8.5, ≥ 9.4, and ≥ 10.1 have been shown to accurately identify histological F2, F3, and F4, respectively, in Asian CHB patients. 18
Definition of fibrosis regression
Baseline and 3‐year assessment with transient elastography and ELF was performed. The primary outcome of this study was the proportion of patients with fibrosis regression, defined by a combination of LS and ELF, at 3 years in patients who achieved HBsAg‐loss compared with those without HBsAg‐loss. The time interval was decided based on the observation that significant fibrosis regression occurred after achieving disease quiescence by nucleoside analogue or HBsAg‐loss after a minimum duration of 3 years. 15 , 19 Dynamic change of LS and ELF over time has been well described, 20 , 21 , 22 although the optimal range for a clinically meaningful change has not been well defined. For LS, as there were no universal definitions for ‘regression of fibrosis’, it was arbitrarily defined as ≥ 30% reduction of LS from baseline. 23 There were also no data on the longitudinal change of ELF in patients with chronic liver disease nor a definition of ‘significant change’. 24 Therefore, it was arbitrarily defined as down‐staging of ELF (i.e., > 10.1 ➔ < 9.4, or > 9.4 ➔ < 8.5, or > 8.5 ➔ < 8.5). 18
Sample size
We hypothesize that CHB patients with HBsAg‐loss will have a higher likelihood and faster rate of fibrosis regression. If we estimate the proportion of fibrosis regression in the HBsAg‐loss group to be 88%, 15 and that in the non‐HBsAg‐loss group to be 24%, 23 together with a 85% power and an alpha risk of 5% (two‐sided), we will need to recruit 260 study subjects (130 subjects per group). Allowing a dropout rate of 10%, we plan to recruit a minimum of 286 subjects (143 subjects per group).
Data processing and analysis
Analysis was performed according to the intention‐to‐treat (ITT) principle, with all patients who complete 3 years of follow‐up included in the endpoint analysis. In addition, analysis according to the per‐protocol (PP) principle was performed after excluding CHB subjects who were started on antiviral therapy or developed HBsAg‐loss (for control group) during the 3 years of follow‐up period. Continuous variables were expressed as median (range or interquartile range as specified). Mann–Whitney U test and Kruskal–Wallis test were used for comparison of median between two groups and multiple groups, respectively. Categorical variables, expressed as proportions, were compared using χ 2 test and Fisher's Exact test when appropriate. Multivariate analysis was performed using binary logistic regression. In view of significant differences in age between the two groups and the impact of age on burden of liver fibrosis, as a sensitivity analysis, we performed 1:1 propensity‐score matching using a caliper width of 0.05. Standardized mean difference (SMD) of age between the two groups after propensity‐score matching were calculated, which was interpreted as balanced when the SMD was below the threshold of 0.1. 25 All statistical analysis was performed using Statistical package for Social Sciences (SPSS) version 27.0 (SPSS Inc, Chicago, IL, USA). A two‐tailed P value of < 0.05 was considered statistically significant.
Results
Patient characteristics
A total of 284 (142 HBsAg‐loss vs 142 CHB) gender‐matched (51.4% male) patients were recruited and completed study evaluation. In spite of our study intention to match the age (limit to 5 years discrepancy), the baseline age was still significantly higher in HBsAg‐loss patients than CHB patients (60 vs 55.7 years old, P < 0.001). Overweight (defined by BMI > 23 kg/m2), type 2 diabetes and hepatic steatosis was present in 61.6%, 18%, and 47.9%, respectively, without significant differences between HBsAg‐loss and CHB subjects (Table 1). For HBsAg‐loss group, the median ageSC was 57.4 (50.5–63.2) years old, with a median duration of 5.5 (4.9–6.2) years from the time of HBsAg‐loss to the time of baseline assessment. For CHB group, the median HBV DNA level was 267 (69–887) IU/mL. ELF‐defined advanced fibrosis or cirrhosis (F3/F4) was present in significantly more patients with HBsAg‐loss than CHB patients (41.1% vs 14.1%, P < 0.001). However, LS‐defined F3/F4 was present in less than 5% of subjects in both groups without significant differences (Table 1). Upon 3 years of follow‐up, no patients developed HCC in both groups.
Table 1.
Baseline characteristics of recruited subjects
| HBsAg‐loss (n = 142) | CHB (n = 142) | P value | |
|---|---|---|---|
| Age (years old) | 60 (53.3–66.9) | 55.7 (49.5–61.4) | < 0.001 |
| Gender (male) | 51.4% | 51.4% | 1.000 |
| Overweight | 56.3% | 66.9% | 0.087 |
| BMI (kg/m2) | 23.8 | 24.3 | 0.185 |
| CAP (dB/m) | 234 (207–294) | 251 (208–297) | 0.208 |
| Type 2 diabetes | 17.6% | 18.3% | 1.000 |
| ALT (U/L) | 21 (16–29) | 24 (19–31) | 0.012 |
| AST (U/L) | 24 (21–28) | 24 (21–31) | 0.364 |
| Platelet (x 109/L) | 214 (183–254) | 216 (183–250) | 0.837 |
| HBV DNA (IU/mL) | ‐ | 267 (69–887) | — |
| AgeSC (years old) | 57.4 (50.5–63.2) | — | — |
| Time after HBsAg‐loss (years) | 5.5 (4.9–6.2) | — | — |
| Steatosis (CAP ≥ 248 dB/m) | 43.7% | 52.1% | 0.191 |
| LS (kPa) | 5.1 (4.1–6.2) | 5.2 (4.3–6.8) | 0.420 |
|
4.9% | 3.5% | 0.770 |
| ELF | 9.3 (8.8–9.9) | 8.4 (7.9–9.3) | <0.001 |
|
41.1% | 14.1% | <0.001 |
Abbreviations: AgeSC, age of achieving hepatitis B surface antigen seroclearance; ALT, alanine aminotransferase; AST, aspartate aminotransferase; BMI, body mass index; CAP, controlled attenuation parameter; CHB, chronic hepatitis B; ELF, enhanced liver fibrosis; F3, advanced fibrosis; F4, cirrhosis; HBV, hepatitis B virus; LS, liver stiffness; HBsAg‐loss, hepatitis B surface antigen seroclearance.
Median (interquartile range) shown.
Fibrosis burden at baseline and at 3 years
The distribution of liver fibrosis stage according to ELF definition and LS definition was shown in Figure 1. Compared with LS, ELF classified more patients with F3/F4 at both baseline and 3 years. For ELF defined liver fibrosis, the proportion of subjects with F1 decreased from 37.8% to 28.5%, in parallel with an increase in the proportion of subjects with F3 (16.3% to 24.9%) and F4 (11.3% to 12.1%). For LS defined liver fibrosis, similar changes were observed where the proportion of F1 decreased (66.2% to 60.1%) and that of F3/F4 increased (4.2% to 11.3%).
Figure 1.
Distribution of liver fibrosis stage at baseline and 3 years. (a) Left panel: ELF‐defined fibrosis stages. (b) Right panel: LS‐defined fibrosis stages. ELF, enhanced liver fibrosis; F0/F1, no or minimal fibrosis; F2, significant fibrosis; F3, advanced fibrosis; F4, cirrhosis; LS, liver stiffness. (a) 
, F0/F1; 
, F2.Gray zone; 
, F3; 
, F4. (b) 
, F0/F1; 
, F2/Gray zone; 
, F3; 
, F4.
Dynamic change in liver fibrosis
There was significant increase in the median serum ELF in both CHB and HBsAg‐loss subjects at 3 years, but not LS values (Figs S1 and 2a). The change in absolute value and percentage of ELF and LS were not significantly different between HBsAg‐loss and CHB subjects (Fig. 2b). For the primary outcome (ITT), a total of 1.8% subjects (1.4% HBsAg‐loss vs 2.1% CHB; P = 1.000) had fibrosis regression at 3 years when both ELF and LS definitions were met. When less stringent criteria was used (i.e., LS decline by ≥ 30%, ELF down‐staging, or any parameter showing regression), more subjects met the criteria of fibrosis regression, without significant differences between CHB and HBsAg‐loss (Fig. 2c). When the PP principle was applied, a total of 23 CHB patients (21 started on antiviral treatment and 2 developed HBsAg‐loss during follow‐up) were excluded, which showed no significant difference in achieving fibrosis regression regardless the modality of assessment used for definition (Fig. S2).
Figure 2.
(a) ELF and LS values at baseline and 3 years in CHB and HBsAg‐loss subjects. (b) ELF and LS change (absolute value and percentage) at 3 years in CHB and HBsAg‐loss subjects. (c) Proportion of subjects achieving fibrosis regression at 3 years using various definitions, stratified by HBsAg‐loss or CHB. CHB, chronic hepatitis; ELF, enhanced liver fibrosis; HBsAg, hepatitis B surface antigen; LS, liver stiffness; n.s., not significant; S‐loss, HBsAg seroclearance. (c) 
, CHB (n = 142); 
, HBsAg seroclearance.
Factors associated with fibrosis regression by ELF down‐staging
As the proportion of patients achieving the primary outcome (combined ELF and LS‐defined fibrosis regression) was too low, the following analysis evaluated the factors associated with ELF‐defined fibrosis regression, which was observed in 14.5% HBsAg‐loss subjects and 16.9% CHB subjects (Fig. 2c). Patients with ELF down‐staging had higher baseline CAP, higher proportion of T2D, longer time since HBsAg‐loss, and higher baseline ELF. Multivariate binary logistic regression showed that time since HBsAg‐loss (hazard ratio [HR] 2.688, 95% confidence interval [CI] 1.257–5.748, P = 0.011) and baseline ELF (HR 1.827, 95% CI 1.085–3.075, P = 0.023) were independent associated with ELF‐defined fibrosis regression at 3 years (Table 2). Fibrosis regression was observed in 7.3%, 11.5%, and 24.4% patients with HBsAg‐loss for a duration of < 5, 5–6, or > 6 years of HBsAg‐loss, respectively (P = 0.059) (Fig. S3).
Table 2.
Factors associated with fibrosis regression (defined by ELF down‐staging) at 3 years in the entire cohort
| No (n = 236) | Yes (n = 44) | P value | Multivariate HR | 95% CI | P value | |
|---|---|---|---|---|---|---|
| Age | 58.1 | 58.2 | 0.332 | |||
| Gender (male) | 49.6% | 59.1% | 0.256 | |||
| BMI | 24 | 24.1 | 0.336 | |||
| CAP | 243 | 258 | 0.026 | 1.006 | 0.996–1.015 | 0.233 |
| Type 2 diabetes | 15.3% | 31.8% | 0.016 | 1.901 | 0.613–5.888 | 0.266 |
| ALT | 22 | 23 | 0.314 | |||
| AST | 24 | 24 | 0.873 | |||
| Platelet (x 109/L) | 214 | 227 | 0.091 | |||
| HBsAg‐loss (vs CHB) | 50% | 45.5% | 0.625 | |||
| Time since HBsAg‐loss † | 5.40 | 6.20 | 0.017 | 2.688 | 1.257–5.748 | 0.011 |
| ELF | 8.7 | 9.5 | <0.001 | 1.827 | 1.085–3.075 | 0.023 |
| LS | 5.2 | 5.0 | 0.713 |
Abbreviations: ALT, alanine aminotransferase; AST, aspartate aminotransferase; BMI, body mass index; CAP, controlled attenuation parameter; CHB, chronic hepatitis B; CI, confidence interval; ELF, enhanced liver fibrosis; HBsAg‐loss, hepatitis B surface antigen seroclearance; HR, hazard ratio; LS, liver stiffness.
Median (interquartile range) shown.
Time since HBsAg‐loss was a continuous variable for HBsAg‐loss group. CHB group with persistent HBsAg seropositivity do not have a numerical value for this parameter and excluded from multivariate analysis.
Predictors of high baseline ELF
Baseline ELF was found to be associated with fibrosis regression, considering that ELF per se was used in the definition of fibrosis regression. To decipher the factors predictive of high baseline ELF, which was defined by baseline value > 8.5, multivariate regression (model 1) showed that older age (HR 1.084, 95% CI 1.048–1.121), baseline AST (HR 1.049, 95% CI 1.011–1.089), and presence of HBsAg‐loss (HR 4.445, 95% CI 2.463–8.020) were significant factors. In a separate model (model 2) that analyzed the subgroup with HBsAg‐loss, ageSC was found to be independently associated with high baseline ELF (HR 1.061, 95% CI 1.016–1.109) (Table S1).
Interaction between age and HBsAg‐loss
In view of the observation that subject age, ageSC, and HBsAg‐loss was associated with baseline liver fibrosis burden, detailed analysis was performed to evaluate the interaction between these parameters. When baseline age was divided into quartiles, the proportion of patients with HBsAg‐loss increased with advancement in age (Fig. 3a). Paradoxically, the proportion of patients with ≥ F2 by ELF or ≥ gray zone by LS also increased with higher age at recruitment (Fig. 3b), despite a higher percentage of patients with HBsAg‐loss in older age groups. In the subgroup of HBsAg‐loss patients, ageSC < 50 was associated with significantly lower prevalence of ≥ F2 defined by ELF or gray zone defined by LS (Fig. S4). In addition, ageSC < 50 was associated with a significantly lower prevalence of ELF‐defined F3/F4 at baseline (5.9% vs 52.3%, P < 0.001) and 3 years (20.6% vs 63.8%, P < 0.001) compared with patients with ageSC > 50 (Fig. 4). Similar observations were made for LS‐defined F3/F4 although statistical significance was not reached.
Figure 3.
(a) Proportion of subjects with HBsAg‐loss in different age groups (stratified by the quartile of age). (b) Proportion of subjects with ≥ F2 defined by ELF or LS at baseline and 3 years in different age groups (stratified by the quartile of age). ELF, enhanced liver fibrosis; ≥ F2, at least significant fibrosis; HBsAg, hepatitis B surface antigen; LS, liver stiffness. (b) 
, Age < 51.6; 
, Age 51.6–58.1; 
, Age 58.1–64.1; 
, Age > 64.1. n = 71 in each group.
Figure 4.
Proportion of subjects with F3/F4 defined by ELF or LS at baseline and 3 years, stratified by ageSC (age of achieving HBsAg‐loss). ELF, enhanced liver fibrosis; F/F4, advanced fibrosis or cirrhosis; HBsAg, hepatitis B surface antigen; LS, liver stiffness. 
, Age at HBsAg seroclearance < 50 (n = 34); 
, Age at HBsAg seroclearance ≥ 50 (n = 108).
Sensitivity analysis
After 1:1 propensity score matching for age, 109 subjects with HBsAg‐loss and 109 subjects with CHB were included, with median age of 58.6 and 58.1 years, respectively (SMD 0.01) (Table S2). Time since HBsAg‐loss (HR 1.674, 95% CI 1.135–2.471) and ELF (HR 1.285, 95% CI 1.115–1.482) were independently associated with fibrosis regression (Table S3). The overall findings were compatible with the main analysis.
Discussion
HBsAg‐loss is advocated as a clinical endpoint of functional cure in CHB, due to the observed clinical benefits of lower risk of HCC and decompensated liver disease. In this prospective case–control study, only 1.4% HBsAg‐loss subjects achieved fibrosis regression—defined by both ELF and LS—at 3 years, which was not statistically different to patients with persistent CHB (2.1%). When less stringent criteria of fibrosis regression was used, although a higher proportion of subjects achieved the endpoint, HBsAg‐loss subjects fared similarly compared with CHB subjects. Our hypothesis that HBsAg‐loss is associated with fibrosis regression is rejected, which could be partially contributed by the older age in the HBsAg‐loss group. However, findings were consistent after propensity score matching for age. Fibrosis regression does not occur in the majority of HBsAg‐loss subjects in this cohort within a short period of time despite HBsAg‐loss, these patients are still at risk of advanced fibrosis and cirrhosis‐related complications.
The unexpectedly low proportion of patients with fibrosis regression could be attributed to the protocol‐defined endpoint of fibrosis regression, which included a combination of LS‐ and ELF‐defined endpoints. The LS values at baseline belonged to F0/F1 or gray zone for the majority of subjects (95.8%; Fig. 1). This did not allow for significant room for improvement at 3 years. The low prevalence of at least significant fibrosis at baseline stems from the low disease burden in the enrolled subjects (low HBV DNA, normal ALT, treatment‐naïve, non‐cirrhotic) and is reflected by the lack of HCC development after 3 years of follow‐up. Therefore, when less stringent criteria of ELF‐defined fibrosis regression were adopted, a much higher percentage of subjects achieved this endpoint (14.5% HBsAg‐loss and 16.9% CHB subjects). It should be noted that ELF‐defined advanced fibrosis or cirrhosis was present in significantly more patients with HBsAg‐loss than CHB patients (41.1% vs 14.1%, P < 0.001) who were significantly older (60.0 vs 55.7 years old, P < 0.001), when LS‐defined F3/F4 was only present in 4.9% and 3.5%, respectively (P = 0.770; Table 1). There was moderate correlation between serum ELF and age (r = 0.526, P < 0.001), while the correlation between LS and age was weak (r = 0.215, P < 0.001). One point to note is the lack of well‐defined LS cut‐off values in patients with concurrent hepatitis B infection and hepatic steatosis, which is seen in 48.8% of this cohort. In addition, age per se is an independent variable for high baseline ELF (HR 1.084; Table S1). This highlighted that age‐related perpetuation in liver fibrosis is more readily detected by ELF than LS. Overall, ELF is probably more sensitive than LS to identify advanced liver fibrosis in patients with apparently low‐risk of liver complications (normal ALT, treatment‐naïve, half of the group already achieved functional cure).
Time since HBsAg‐loss was also associated with > 2.5‐fold likelihood to develop ELF‐defined fibrosis regression, which is coherent with a previous report from our group. 19 With establishment of viral quiescence, ongoing liver insults from the hepatotropic virus are halted, allowing gradual resorption of fibrous deposition in the extracellular matrix by tissue metalloproteinases. 26 This is in analogy with the effects of antiviral treatment on reversal of histological fibrosis or cirrhosis after a few years of therapy. 15 , 27 However, we also noted that the deleterious impact of older age might have partially offset the potential benefits of increasing time since HBsAg‐loss, as shown in Figure 3b where an incremental proportion of subjects would have at least significant fibrosis with advancement in age. In this regard, the ageSC seems to mitigate the risk of having a high baseline liver fibrosis burden, where ageSC < 50 was associated with significantly lower risk of having baseline F3/F4 compared with ageSC > 50 (5.9% vs 52.3%; Fig. 4). The relatively advanced ageSC in the HBsAg‐loss group (57.4 in the main analysis, 55.3 years old in the PSM cohort) could have diminished or offset any potential benefits on fibrosis regression brought about by viral quiescence. Our findings not only reinforced the general understanding that the favorable effects of HBsAg‐loss are most manifest when ageSC < 50 7 but also highlights that patients can still have F3/F4 despite achieving HBsAg‐loss, especially if ageSC was > 50 years old (> 50%). Ongoing surveillance for liver‐related complications should be conducted for these subjects. In addition, drug trials currently in development for CHB should prioritize patients younger than 50 years old to maximize the potential benefits from drug‐induced HBsAg seroclearance. 28
Metabolic factors including overweight and type 2 diabetes, although were not significantly associated with fibrosis progression, were present in a high proportion of HBsAg‐loss subjects compared with CHB patients (56.3% and 17.6%, respectively) and could have further contributed to the high prevalence of ELF‐defined F3/F4 at baseline (41.1%; Table 1). Similarly, hepatic steatosis was present in around half of all subjects, and a higher CAP was found among subjects with fibrosis regression (258 vs 243 dB/m, P = 0.026) although it was no longer significant after adjusting for other factors in the multivariate analysis model (Table 2). While it was reported that concomitant hepatic steatosis was associated with fibrosis progression but also HBsAg‐loss, 29 the current study cannot address the complicated interaction between hepatic steatosis and CHB due to the inclusion of subjects who already achieved HBsAg‐loss.
The main strengths of the current study included the prospective nature, the use of a combination of modalities for non‐invasive assessment of liver fibrosis, and the inclusion of a homogenous population (Chinese CHB patients or HBsAg‐loss patients, presumed low‐risk due to normal ALT and without the need of antiviral therapy). The main limitations would be the lack of liver biopsy for histological assessment of liver fibrosis. Also, despite the intention to match the age between both groups, the age gap between patients with HBsAg‐loss and CHB was still statistically significant. Also, the high proportion of patients with minimal fibrosis or gray zone fibrosis at baseline limits the ability to observe significant change in fibrosis burden. Moreover, baseline was not at the time of HBsAg seroclearance but sometime after, which might have diminished early changes in fibrosis burden, if any. Future research should observe the fibrosis burden right after the occurrence of HBsAg seroclearance. Finally, follow‐up duration could be lengthened to detect more events.
In conclusion, fibrosis regression occurred in a minority of subjects achieving HBsAg‐loss who achieved HBsAg‐loss after 50 years old, which was not significantly different compared with subjects with persistent overt CHB. Although increasing duration of HBsAg‐loss was associated with higher likelihood of fibrosis progression, these effects are compounded by the advancement in age. Age of achieving HBsAg‐loss appears to carry the most prognostic implications in risk of advanced fibrosis/cirrhosis. Subjects after achieving HBsAg‐loss, especially among those with ageSC > 50, should receive ongoing surveillance for liver‐related complications.
Presentation
The preliminary data of this study was presented at The 30th Conference of the Asian Pacific Association for the Study of the Liver (APASL) in a parallel oral session (Mak LY, Wong DKH, Hui WRH et al. Regression of liver fibrosis after HBsAg loss: a prospective matched case–control evaluation using transient elastography and serum Enhanced Liver Fibrosis (ELF) test. APASL Annual Meeting, virtual 2021; Young Investigator Award).
Supporting information
Table S1. Factors associated with a high baseline ELF (≥8.5).
Table S2. Baseline characteristics of recruited subjects after 1:1 propensity score matching for age.
Table S3. Factors associated with fibrosis regression (defined by ELF down‐staging) at 3 years after 1:1 propensity score matching for age.
Figure S1. Paired values of serum ELF and liver stiffness at baseline and 3 years in patients with HBsAg‐loss and CHB.
Figure S2. Proportion of subjects achieving fibrosis regression at 3 years using various definitions, stratified by HBsAg‐loss or CHB (per‐protocol principle).
Figure S3. Proportion of patients achieving ELF‐defined fibrosis regression according to time since HBsAg‐loss.
Figure S4. Proportion of subjects with ≥F2 defined by ELF or ≥gray zone defined by LS at baseline and 3 years, stratified by ageSC (age of achieving HBsAg seroclearance).
Acknowledgments
We would like to thank Miss Delanda Wong and Mr John Yuen for the help in subject recruitment.
Mak, L.‐Y. , Hui, R. W.‐H. , Chung, M. S. H. , Wong, D. K.‐H. , Fung, J. , Seto, W.‐K. , and Yuen, M.‐F. (2024) Regression of liver fibrosis after HBsAg loss: A prospective matched case–control evaluation using transient elastography and serum enhanced liver fibrosis test. Journal of Gastroenterology and Hepatology, 39: 2826–2834. 10.1111/jgh.16728.
Declaration of conflict of interest: L. Y. Mak is an advisory board member of Gilead Sciences. J. Fung is an advisory board member of Gilead Sci.ences. M. F. Yuen is an advisor/consultant for and/or received grant/research support from AbbVie, Aligos Therapeutics, AiCuris, Antios Therapeutics, Arbutus Biopharma, Arrowhead Pharmaceuticals, Assembly Biosciences, Bristol‐Myers Squibb, Clear B Therapeutics, Dicerna Pharmaceuticals, Finch Therapeutics, Fujirebio Incorporation, GlaxoSmithKline, Gilead Sciences, Immunocore, Janssen, Roche, Silverback Therapeutics, Sysmex Corporation, Tune Therapeutics, Vir Biotechnology and Visirna Therapeutics. W. K. Seto received speaker's fees from AstraZeneca and Mylan, is an advisory board member of Abbott, is an advisory board member and received speaker's fees from AbbVie, and is an advisory board member, received speaker's fees and researching funding from Gilead Sciences. The other authors have nothing to disclose.
Author contributions: The authors declare they have participated in the preparation of the manuscript and have seen and approved the final version. L. Y. Mak was involved in study concept and design, acquisition of data, analysis, and interpretation of data and drafting of manuscript. R. W. H. H., M. S. H. C., and D. K. H. W. were involved in acquisition of data and analysis and interpretation of data. J. F. and W. K. S. were involved in critical revision of manuscript. M. F. Y was involved in study concept and design, analysis, and interpretation of data, critical revision of manuscript, and overall study supervision.
Financial support: This study was supported by the Health and Medical Research Fund of the Hong Kong Special Administrative Region (reference number: 16171251).
Guarantor of the article: Professor Man‐Fung Yuen.
Contributor Information
Wai‐Kay Seto, Email: wkseto@hku.hk.
Man‐Fung Yuen, Email: mfyuen@hkucc.hku.hk.
Data availability statement
The data that support the findings of this study are available from the corresponding author upon reasonable request.
References
- 1. Ye F, Zhai M, Long J et al. The burden of liver cirrhosis in mortality: Results from the global burden of disease study. Front. Public Health 2022; 10: 909455. 10.3389/fpubh.2022.909455. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Collaborators GBDC . The global, regional, and national burden of cirrhosis by cause in 195 countries and territories, 1990‐2017: a systematic analysis for the Global Burden of Disease Study 2017. Lancet Gastroenterol. Hepatol. 2020; 5: 245–266. 10.1016/S2468-1253(19)30349-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Mokdad AA, Lopez AD, Shahraz S et al. Liver cirrhosis mortality in 187 countries between 1980 and 2010: a systematic analysis. BMC Med. 2014; 12: 145. 10.1186/s12916-014-0145-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. European Association for the Study of the Liver. Electronic address eee, European Association for the Study of the L . EASL 2017 clinical practice guidelines on the management of hepatitis B virus infection. J. Hepatol.; 2017: 370–398. 10.1016/j.jhep.2017.03.021. [DOI] [PubMed] [Google Scholar]
- 5. Terrault NA, Bzowej NH, Chang KM, Hwang JP, Jonas MM, Murad MH, American Association for the Study of Liver Diseases . AASLD guidelines for treatment of chronic hepatitis B. Hepatology 2016; 63: 261–283. 10.1002/hep.28156. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Zu J, Zhuang G, Liang P, Cui F, Wang F, Zheng H, Liang X. Estimating age‐related incidence of HBsAg seroclearance in chronic hepatitis B virus infections of China by using a dynamic compartmental model. Sci. Rep. 2017; 7: 2912. 10.1038/s41598-017-03080-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Yuen MF, Wong DK, Fung J et al. HBsAg Seroclearance in chronic hepatitis B in Asian patients: replicative level and risk of hepatocellular carcinoma. Gastroenterology 2008; 135: 1192–1199. 10.1053/j.gastro.2008.07.008. [DOI] [PubMed] [Google Scholar]
- 8. Simonetti J, Bulkow L, McMahon BJ et al. Clearance of hepatitis B surface antigen and risk of hepatocellular carcinoma in a cohort chronically infected with hepatitis B virus. Hepatology 2010; 51: 1531–1537. 10.1002/hep.23464. [DOI] [PubMed] [Google Scholar]
- 9. Anderson RT, Choi HSJ, Lenz O et al. Association between seroclearance of hepatitis B surface antigen and long‐term clinical outcomes of patients with chronic hepatitis B virus infection: systematic review and meta‐analysis. Clin. Gastroenterol. Hepatol. 2021; 19: 463–472. 10.1016/j.cgh.2020.05.041. [DOI] [PubMed] [Google Scholar]
- 10. Fong TL, Di Bisceglie AM, Gerber MA, Waggoner JG, Hoofnagle JH. Persistence of hepatitis B virus DNA in the liver after loss of HBsAg in chronic hepatitis B. Hepatology 1993; 18: 1313–1318. [PubMed] [Google Scholar]
- 11. Parkes J, Roderick P, Harris S et al. Enhanced liver fibrosis test can predict clinical outcomes in patients with chronic liver disease. Gut 2010; 59: 1245–1251. 10.1136/gut.2009.203166. [DOI] [PubMed] [Google Scholar]
- 12. European Association for Study of Liver . Asociacion Latinoamericana para el Estudio del Higado Clinical Practice Guidelines: non‐invasive tests for evaluation of liver disease severity and prognosis. J. Hepatol. 2015; 63: 237–264. 10.1016/j.jhep.2015.04.006. [DOI] [PubMed] [Google Scholar]
- 13. Castera L, Vergniol J, Foucher J et al. Prospective comparison of transient elastography, Fibrotest, APRI, and liver biopsy for the assessment of fibrosis in chronic hepatitis C. Gastroenterology 2005; 128: 343–350. [DOI] [PubMed] [Google Scholar]
- 14. Papastergiou V, Tsochatzis E, Burroughs AK. Non‐invasive assessment of liver fibrosis. Ann. Gastroenterol. 2012; 25: 218–231. [PMC free article] [PubMed] [Google Scholar]
- 15. Chang TT, Liaw YF, Wu SS et al. Long‐term entecavir therapy results in the reversal of fibrosis/cirrhosis and continued histological improvement in patients with chronic hepatitis B. Hepatology 2010; 52: 886–893. 10.1002/hep.23785. [DOI] [PubMed] [Google Scholar]
- 16. Fung J, Lai CL, Wong DK, Seto WK, Hung I, Yuen MF. Significant changes in liver stiffness measurements in patients with chronic hepatitis B: 3‐year follow‐up study. J. Viral Hepat. 2011; 18: e200–e205. 10.1111/j.1365-2893.2010.01428.x. [DOI] [PubMed] [Google Scholar]
- 17. Karlas T, Petroff D, Sasso M et al. Individual patient data meta‐analysis of controlled attenuation parameter (CAP) technology for assessing steatosis. J. Hepatol. 2017; 66: 1022–1030. 10.1016/j.jhep.2016.12.022. [DOI] [PubMed] [Google Scholar]
- 18. Kim BK, Kim HS, Yoo EJ et al. Risk assessment of clinical outcomes in Asian patients with chronic hepatitis B using enhanced liver fibrosis test. Hepatology 2014; 60: 1911–1919. 10.1002/hep.27389. [DOI] [PubMed] [Google Scholar]
- 19. Mak LY, Seto WK, Hui RW, Fung J, Wong DKH, Lai CL, Yuen MF. Fibrosis evolution in chronic hepatitis B e antigen‐negative patients across a 10‐year interval. J. Viral Hepat. 2019; 26: 818–827. 10.1111/jvh.13095. [DOI] [PubMed] [Google Scholar]
- 20. Gardner AR, Ma Y, Bacchetti P et al. Longitudinal assessment of the enhanced liver fibrosis score in the era of contemporary HIV and hepatitis C virus treatment. J Infect Dis 2023; 227: 1274–1281. 10.1093/infdis/jiac315. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Liang LY, Wong VW, Tse YK, Yip TCF, Lui GCY, Chan HLY, Wong GLH. Improvement in enhanced liver fibrosis score and liver stiffness measurement reflects lower risk of hepatocellular carcinoma. Aliment. Pharmacol. Ther. 2019; 49: 1509–1517. 10.1111/apt.15269. [DOI] [PubMed] [Google Scholar]
- 22. de Franchis R, Bosch J, Garcia‐Tsao G, Reiberger T, Ripoll C, Baveno VIIF. Baveno VII—renewing consensus in portal hypertension. J. Hepatol. 2022; 76: 959–974. 10.1016/j.jhep.2021.12.022. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Wong GL, Wong VW, Choi PC et al. On‐treatment monitoring of liver fibrosis with transient elastography in chronic hepatitis B patients. Antivir. Ther. 2011; 16: 165–172. 10.3851/IMP1726. [DOI] [PubMed] [Google Scholar]
- 24. Martinez SM, Fernandez‐Varo G, Gonzalez P et al. Assessment of liver fibrosis before and after antiviral therapy by different serum marker panels in patients with chronic hepatitis C. Aliment. Pharmacol. Ther. 2011; 33: 138–148. 10.1111/j.1365-2036.2010.04500.x. [DOI] [PubMed] [Google Scholar]
- 25. Austin PC. Some methods of propensity‐score matching had superior performance to others: results of an empirical investigation and Monte Carlo simulations. Biom. J. 2009; 51: 171–184. 10.1002/bimj.200810488. [DOI] [PubMed] [Google Scholar]
- 26. Naim A, Pan Q, Baig MS. Matrix metalloproteinases (MMPs) in liver diseases. J. Clin. Exp. Hepatol. 2017; 7: 367–372. 10.1016/j.jceh.2017.09.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Marcellin P, Gane E, Buti M et al. Regression of cirrhosis during treatment with tenofovir disoproxil fumarate for chronic hepatitis B: a 5‐year open‐label follow‐up study. Lancet 2013; 381: 468–475. 10.1016/S0140-6736(12)61425-1. [DOI] [PubMed] [Google Scholar]
- 28. Ghany MG, Buti M, Lampertico P, Lee HM, A‐EH‐HTEC Faculty . Guidance on treatment endpoints and study design for clinical trials aiming to achieve cure in chronic hepatitis B and D: report from the 2022 AASLD‐EASL HBV‐HDV Treatment Endpoints Conference. Hepatology 2023; 78: 1654–1673. 10.1097/HEP.0000000000000431. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Mak LY, Hui RW, Fung J et al. Diverse effects of hepatic steatosis on fibrosis progression and functional cure in virologically quiescent chronic hepatitis B. J. Hepatol. 2020; 73: 800–806. 10.1016/j.jhep.2020.05.040. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Table S1. Factors associated with a high baseline ELF (≥8.5).
Table S2. Baseline characteristics of recruited subjects after 1:1 propensity score matching for age.
Table S3. Factors associated with fibrosis regression (defined by ELF down‐staging) at 3 years after 1:1 propensity score matching for age.
Figure S1. Paired values of serum ELF and liver stiffness at baseline and 3 years in patients with HBsAg‐loss and CHB.
Figure S2. Proportion of subjects achieving fibrosis regression at 3 years using various definitions, stratified by HBsAg‐loss or CHB (per‐protocol principle).
Figure S3. Proportion of patients achieving ELF‐defined fibrosis regression according to time since HBsAg‐loss.
Figure S4. Proportion of subjects with ≥F2 defined by ELF or ≥gray zone defined by LS at baseline and 3 years, stratified by ageSC (age of achieving HBsAg seroclearance).
Data Availability Statement
The data that support the findings of this study are available from the corresponding author upon reasonable request.