Introduction
Chronic hepatitis B virus (HBV) infection is a major cause of chronic liver diseases, ranging from viral hepatitis to end-stage liver diseases, including cirrhosis and liver cancer.1 Although universal hepatitis B vaccination among newborns and infants has significantly reduced new infections, HBV infections remain a major global health burden, with an estimated 296 million people living with chronic infection.2
Hepatitis B e antigen (HBeAg) is a secreted non-structural viral protein that is not involved in HBV replication. However, the presence of HBeAg is necessary for the establishment of chronic HBV infection either through vertical or horizontal transmission due to its inhibitory effect on innate and adaptive host immunity against HBV.3,4 On the other hand, HBeAg expression decreased/abolished virus strain (1,762/1,764/1,896 mutations) could avoid the host-specific adaptive immune response to HBeAg/HBcAg and become the dominant virus strain, resulting in the persistence of chronic HBV infection,5 which makes HBeAg usually the first HBV viral protein being lost after the activation of host immunity against HBV.
Patients with chronic hepatitis B are usually divided into four phases by HBeAg status and hepatic inflammation (based on serum ALT and/or liver biopsy): HBeAg-positive infection, HBeAg-positive chronic hepatitis B (CHB), HBeAg-negative chronic infection, and HBeAg-negative CHB.6 Previous studies have suggested that for patients with HBeAg-positive, active liver inflammation is more likely to cause HBV DNA decrease and HBeAg clearance, while for patients who are HBeAg-negative, active liver inflammation is associated with viral reactivation and persistent liver damage. Therefore, most guidelines emphasize in their recommendation that HBeAg-negative patients with chronic hepatitis B should be given timely antiviral treatment to inhibit viral replication and reduce liver damage.6
It is well known that the disease progression of chronic HBV infection is driven by the interactions between the HBV virus and the host immune response against HBV infection, which results in liver immunopathological damage manifested by liver inflammation and fibrosis/cirrhosis. However, the correlation between viral replication and the host immune response has not been fully demonstrated. Accumulating studies have indicated that anti-HBc could be considered a surrogate biomarker for liver inflammation in patients with chronic HBV infection. Our recent study reported that the level of anti-HBc was negatively correlated with intrahepatic covalently closed circular DNA (cccDNA) and serum HBV DNA in HBeAg-positive patients (r = −0.387, P<0.05; r=−0.220, P<0.05), but positively in those HBeAg-negative patients (r =0.419, P<0.05; r=0.570, P<0.05).7 However, it is still unclear whether this difference is only related to the presence of HBeAg, and whether this transition occurs before or after HBeAg clearance.
In this study, we evaluated the relationship between the severity of liver inflammatory damage and viral replication in CHB patients in different HBeAg status and HBV DNA levels through the analysis of two independent cross-sectional cohorts, one validation cohort, and meta-analyses. We observed an opposite relationship between viral replication and liver inflammation for patients with different HBeAg status. What’s more, an opposite relationship existed between HBV replication and liver inflammation in HBeAg-positive patients with HBV DNA ≤2×106 IU/mL and >2×106 IU/mL.
Methods
Clinical cohorts
This study included two independent cohorts and one validation cohort. Cohort A is a retrospective cohort from the Fifth Medical Center of the PLA General Hospital of Beijing, China, with the majority of enrolled patients having been previously reported.8,9 Briefly, the patients were enrolled from 2010 to 2020, and those with available laboratory evidence of chronic HBV infection (serum HBsAg and/or HBV DNA positive for at least 6 months) were included (approved by the Ethics Committee of the Fifth Medical Center of the PLA General Hospital of Beijing, No. KY-2022-1-4-1). None of the enrolled patients had received antiviral treatment in the past 6 months, and the liver necroinflammatory grade of each patient was evaluated by liver biopsy,10,11 as previously described. Cohort B is the retrospective cohort from Ruijin Hospital, as previously described.12 Briefly, the database was retrospectively derived from CHB patients who had undergone liver biopsy. The criteria for enrollment were:1) with chronic HBV infection for more than 6 months; 2) not receiving anti-HBV treatment. 3) with clinical information such as serum ALT, HBV DNA, HBeAg, and anti-HBe available. The exclusion criteria were patients with HIV infection, other viral hepatitis infections (hepatitis A, C, D, and E), clinical or imaging evidence of decompensated liver cirrhosis, hepatocellular carcinoma, non-alcoholic fatty liver disease, autoimmune liver disease, drug-induced liver disease, metabolic or genetic disease, and alcohol consumption. A validation cohort enrolled 2,624 treatment-naïve patients with HBeAg-positive CHB from Nanjing Drum Tower Hospital between 2015 and 2022, including 457 who underwent liver biopsy; some of these patients have been described previously.13
The following clinical indices were collected for analysis: ALT, aspartate aminotransferase (AST), gamma-glutamyl transpeptidase (GGT), liver biopsy-determined liver necroinflammatory grade (G), HBV DNA, HBeAg, and anti-HBe. In addition, quantitative serum HBsAg measurement and hepatic immunohistochemistry (IH) and scoring of HBcAg were available for some patients in cohort A to confirm the relationship between hepatic virus replication and liver injury.
Statistical analysis
The data obtained were analysed with IBM SPSS Statistic 22.0 software (International Business Machines Corporation, New York, USA) and GraphPad Prism version 5.0 (GraphPad Software Inc., La Jolla, California). Clinical characteristics of the patients are described as the mean ± standard error (SEM) or median (-IQR, +IQR), depending on the data distribution. Differences between sample groups were tested using ANOVA following rank transformation or the t-test, as determined by the distribution of data. Spearman’s rank correlation coefficient test, linear regression analysis, and logistic regression were used to describe the association between variables. A two-sided P-value <0.05 was considered statistically significant.
The systematic review and meta-analysis aimed to explore the relationship between HBV DNA and ALT in CHB patients with different HBeAg statuses.
Search strategies and selection criteria
Databases including Embase and PubMed were searched for relevant literature using search terms (Supplementary Fig. 1). Literature searches were performed on April 25, 2022. A manual search was also performed as supplemental. Only English language studies were reviewed. Briefly, a study was regarded as eligible for inclusion in the meta-analysis if it provided the R-value between HBV DNA and ALT in CHB patients with a specific HBeAg status. Study screening flows are described in Supplementary Figure 1. 16 studies (HBeAg positive: n=7;14–20 HBeAg negative: n=1414,16,17,19,21–29) were finally included for analysis.
Data extraction and Risk of Bias assessment
The following information was extracted from each study: name of the first author, year of publication, number of patients, baseline HBV DNA, ALT level, and R-value between HBV DNA and ALT. The basic information of studies included in Part 1 is summarized in Supplementary Tables 1 and 2. Two authors (Wang. L and Zhao. K) independently extracted the information from the studies included using the same standardized form. Discordances were resolved by consensus and/or by discussion with a third and senior author. Study quality and risk of bias were assessed using the modified ROBINS-E Tool. This scale consists of 7 domains. The risk of bias was evaluated as follows: low risk of bias, some concerns, high risk of bias, and very high risk of bias (Supplementary Tables 3 and 4).
Statistical methods for meta-analysis
Stata 16.0 (StataCorp, Texas, USA) and RevMan 5.4.1 (The Cochrane Collaboration, Oxford, UK) were used to perform meta-analyses. The R-value was transformed to Fisher’s Z to calculate the overall R-value. Forest plots were drawn, and pooled ORs and 95%CIs were calculated. Heterogeneity was evaluated by the Cochrane Q test and I2 test. Briefly, I2>50% was considered as heterogeneity exists. Random-effects meta-analysis was used to pool the results from the included studies.
Results
Clinical characteristics of patients
Two independent cohorts of treatment-naïve patients with chronic HBV infection were enrolled in this cross-sectional clinical study. Cohort A was composed of 993 individuals from the Fifth Medical Center of Chinese PLA General Hospital; Cohort B included 357 individuals from Ruijin Hospital. The clinical characteristics of the two cohorts are summarized in Supplementary Table 5. In cohort A, quantitative serum HBsAg measurements were available in 760 patients, and hepatic HBcAg expression was semi-quantitatively scored by immunohistochemistry in 567 patients. The clinical characteristics of the validation cohort including 2,624 HBeAg-positive individuals are summarized in Supplementary Table 6.
Opposite relationships between HBV replication and the severity of hepatic necroinflammation were observed in HBeAg-positive and -negative individuals
As shown in Table 1, the level of serum HBV DNA correlated negatively with the grade of liver necroinflammation (rho=−0.176, P<0.01) and GGT (rho=−0.157, P<0.01) in patients with HBeAg-positive status, both in cohort A and in cohort B (rho=−0.260, P<0.01; and rho=−0.249, P<0.01, respectively), though the coefficient correlations were weak. In contrast, serum HBV DNA levels exhibited a significantly positive correlation with the severity of liver inflammatory grade in both cohort A (rho=0.434, P<0.01) and in cohort B (rho=0.426, P<0.01), in those HBeAg-negative patients, which was supported by a weak to moderately strong correlation of serum HBV DNA levels with biochemical hepatic inflammatory indicators including ALT, AST and GGT. Moreover, the above opposite correlations between patients’ serum HBV DNA level and ALT were partially confirmed by meta-analysis (Supplementary Fig. 2, Supplementary Tables 1 and 2). HBV DNA did exhibit a moderate positive correlation with ALT (n=13, r=0.481, 95% CI: 0.286, 0.638, I2=97.0%), in HBeAg-negative patients (Supplementary Figs. 2A and 3), though forest plots of the meta-analysis showed no correlation between HBV DNA and ALT (n=7, r=−0.004, 95% CI: −0.083, 0.075, I2=40.8%) in HBeAg-positive patients (Supplementary Fig. 2B).
Table 1Spearman relationship between HBV DNA and clinical index for liver inflammation
| Spearman Rho (P-Value) | HBV DNA
|
|---|
| Cohort A (n=993) | Cohort B (n=357) |
|---|
| HBeAg positive (n=649) | HBeAg negative (n=344) | HBeAg positive (n=231) | HBeAg negative (n=126) |
|---|
| ALT (IU/mL) | −0.036 (P=0.36) | 0.450 (P<0.01) | 0.144 (P=0.029) | 0.363 (P<0.01) |
| AST (IU/mL) | 0.068 (P=0.85) | 0.459 (P<0.01) | 0.059 (P=0.370) | 0.448 (P<0.01) |
| GGT (IU/mL) | −0.157 (P<0.01) | 0.208 (P<0.01) | −0.249 (P<0.01) | 0.220 (P<0.01) |
| Liver Necroinflammatory grade | −0.176 (P<0.01) | 0.434 (P<0.01) | −0.260 (P<0.01) | 0.426 (P<0.01) |
In addition to HBV DNA, hepatic HBcAg may also reflect the transcriptional activity of hepatic HBV cccDNA. To confirm the above findings, immunohistochemistry (IH) staining and scoring of intrahepatic HBcAg were performed for 569 patients in cohort A, including 329 HBeAg-positive and 240 HBeAg-negative patients. As expected, the HBcAg expression level in the liver tissue of HBeAg-positive patients was significantly higher than that in those HBeAg-negative patients [0.527 (0.071, 1.061) vs. 0.008 (0.001, 0.039), P<0.01]. Among the 79.9% (263/329) HBeAg-positive patients accompanied by positive hepatic HBcAg staining. Spearman analysis revealed a marginal negative correlation between the in-situ expression intensity of hepatic HBcAg (semi-quantitated by IH scores) and the liver necroinflammatory grade (Scheuer grade) (r=−0.208, P=0.026). Such a tendency was also demonstrated by the decrease in the hepatic HBcAg-positive rate, along with increased severity of liver necro-inflammatory activities (G0–1: 83.85%, G2: 75.98%, and G3–4: 68.89%) (Table 2). Concordantly, the expression intensity of HBcAg in liver tissues of G3–4 patients was significantly lower than that of G2 (P=0.014) and G0–1 (P=0.001) in the HBeAg-positive patients. In contrast, in HBeAg-negative patients, the hepatic HBcAg-positive staining rate was found to correlate positively with the increased severity of liver necroinflammation, which was supported by the increased rate of hepatic HBcAg-positive staining from 3.86% to 21.27% and then to 46.15%, respectively in parallel with the severity of liver necroinflammatory grade from G0–1 to G2 and G3–4 (Table 2), though only 9.58% HBeAg-negative patients accompanied by positive hepatic HBcAg.
Table 2Hepatic HBcAg expression in HBeAg-positive or HBeAg-negative patients with different liver necroinflammatory grades
| Liver necroinflammation grade/Hepatic HBcAg positive percentage and HBcAg IH scores | G0–1 | G2 | G3–4 |
|---|
| HBeAg-positive | 83.85% (n=135/161) | 76.98% (n=97/126) | 68.89% (n=31/45) |
| 11.87 (8.49, 13.41) | 10.96 (5.01, 13.54) | 8.54 (0.00, 11.59) |
| HBeAg-negative | 3.86% (n=7/181) | 21.27% (n=10/47) | 46.15% (n=6/13) |
| 0.30 (0.18, 0.43) | 1.98 (1.38, 2.58) | 3.06 (2.01, 4.11) |
Negative relationship between HBV replication and severity of liver inflammation only present in HBeAg-positive patients with moderate to high serum HBV DNA levels
The results above to some extent reminded us that there could be efficient control of HBV DNA replication by the host immune system against HBV infection in the early phase of immune activation (IA) in the HBeAg-positive stage. First, a threshold of HBV DNA at 2×106 IU/mL was explored and determined by Figure 1 and the serial Spearman test. Then, those HBeAg-positive patients were sub-grouped into HBV DNA high (HBeAg+/DNA high) and low (HBeAg+/DNA low) groups, accordingly. The clinical characteristics of HBeAg-positive patients with different HBV DNA levels for cohort A, cohort B, and Nanjing cohort were summarized in Supplementary Table 6. HBeAg-positive patients with lower HBV DNA levels in cohorts A and B were accompanied by higher ALT compared with patients in the validation cohort. As shown in Table 3, the relationship between viral replication and liver inflammation exhibited the most obvious difference in HBV DNA >2×106 IU/mL HBeAg+/DNA high group. Such HBeAg+/DNA high patients exhibited a marginal negative relationship of HBV DNA level with ALT (rho=−0.160, P<0.001), AST (rho=−0.188, P<0.001), and the severity of liver necroinflammation (rho=−0.211, P<0.001), respectively in cohort A. However, in HBeAg-positive patients with lower HBV DNA levels, this negative correlation weakened or disappeared, and even changed to a positive correlation, similar to HBeAg-negative patients (Table 1). Consistently, data from cohort B and the Nanjing cohort exhibited similar tendencies (Table 3). Next, when stratified by liver fibrosis stage or anti-HBe, similar tendencies still existed, though not obvious in cohort B patients with significant fibrosis stage (S≥2) (Supplementary Tables 7 and 8). Besides, we also used public data of CHB patients to perform a similar analysis.30 Clinical data of 124 CHB patients were available in this article.30 Among them, 94 patients had complete information for HBV DNA level and HBeAg status. As shown in Supplementary Tables 9 and 10, similar tendencies were also observed in these data.
Table 3Spearman relationship between serum HBV DNA and ALT, AST, GGT and liver necroinflammatory grade for HBeAg-positive CHB with different HBV DNA levels
| Spearman Rho (P value) between | HBV DNA and | DNA >2×106 IU/mL subgroup | DNA ≤2×106 IU/mL subgroup |
|---|
| Cohort A | | (n=555) | (n=94) |
| ALT | −0.160 (<0.001) | 0.278 (0.007) |
| AST | −0.188 (<0.001) | 0.328 (0.001) |
| GGT | −0.141 (0.001) | 0.083 (0.429) |
| Liver necroinflammatory grade | −0.211 (<0.001) | 0.005 (0.958) |
| Cohort B | | (n=152) | (n=79) |
| ALT | −0.042 (0.612) | 0.260 (0.021) |
| AST | −0.169 (0.038) | 0.279 (0.013) |
| GGT | −0.334 (<0.001) | 0.074 (0.539) |
| Liver necroinflammatory grade | −0.305 (<0.001) | −0.296 (0.008) |
| Independent validation Nanjing Cohort | (n=2,041) | (n=583) |
| ALT | −0.181 (<0.001) | 0.201 (<0.001) |
| AST | −0.236 (<0.001) | 0.215 (<0.001) |
| GGT | −0.194 (<0.001) | 0.104 (0.012) |
| Liver necroinflammatory grade | −0.158 (0.004) * | 0.011 (0.903)# |
Furthermore, with the use of linear regression adjusting for age, sex and anti-HBe, the results (Table 4) showed that for HBeAg positive patients with HBV DNA >2×106 IU/mL subgroup, patients with active liver inflammation accompanied by a lower HBV DNA level (−0.177 log HBV DNA for ALT elevated, −0.204 log HBV DNA for AST elevated, −0.229 log HBV DNA for GGT elevated and −0.335 log HBV DNA for liver necroinflammation ≥G2) in cohort A. Whereas, for HBeAg positive patients with HBV DNA ≤2×106 IU/mL, subgroup with active liver inflammation accompanied by a higher HBV DNA level (0.504 log HBV DNA for ALT elevated, 0.657 log HBV DNA for AST elevated, 0.290 log HBV DNA for GGT elevated and 0.252 log HBV DNA for liver necroinflammation ≥G2) in cohort A. Binary logistic regression analysis revealed that active liver inflammation was a protective factor for higher HBV DNA in HBeAg positive patients with >2×106 IU/mL subgroup, but a risk factor for higher HBV DNA in HBeAg positive patients with ≤2×106 IU/mL subgroup (Table 5) in the cohort A. Data from cohort B exhibited similar tendencies (Supplementary Tables 11 and 12).
Table 4Association of HBV DNA change with liver inflammation indicators in HBeAg positive patients with different HBV DNA levels (cohort A)
| Variables | Change in log HBV DNA (IU/mL)
|
|---|
HBV DNA >2×106 IU/mL subgroup
| HBV DNA ≤2×106 IU/mL subgroup
|
|---|
| β* | SE | T value | P value | β* | SE | T value | P value |
|---|
| ALT>2×ULN | −0.177 | 0.068 | −2.613 | 0.009 | 0.504 | 0.246 | 2.051 | 0.043* |
| AST>1.5×ULN | −0.204 | 0.069 | −2.973 | 0.003 | 0.657 | 0.242 | 2.712 | 0.008 |
| GGT>1×ULN | −0.229 | 0.086 | −2.663 | 0.008 | 0.290 | 0.265 | 1.094 | 0.277 |
| Liver necroinflammation grade ≥G2 | −0.335 | 0.066 | −6.103 | <0.001 | 0.252 | 0.244 | 1.035 | 0.304 |
Table 5Binary logistic regression analysis between HBV DNA and liver inflammation indicator in HBeAg-positive patients with different HBV DNA levels (cohort A)
| Variables | Liver inflammation indicators | OR (95% CI) * | P value | β |
|---|
| HBV DNA >2×106 IU/mL subgroup | | | |
| HBV DNA increased (>1×108 IU/mL) | ALT (>2×ULN) | 0.572 (0.403–0.813) | 0.002 | −0.558 |
| AST (>1.5×ULN) | 0.564 (0.396–0.806) | 0.002 | −0.572 |
| GGT (>ULN) | 0.678 (0.435–1.057) | 0.086 | −0.388 |
| Liver necroinflammatory grade (≥G2) | 0.459 (0.323–0.651) | <0.001 | −0.779 |
| HBV DNA ≤2×106 IU/mL subgroup | | | |
| HBV DNA increased (>3×105 IU/mL) | ALT (>2×ULN) | 2.991 (1.133–7.895) | 0.027 | 1.096 |
| AST (>1.5×ULN) | 4.708 (1.680–13.189) | 0.003 | 1.549 |
| GGT (>ULN) | 2.258 (0.840–6.068) | 0.106 | 0.814 |
| Liver necroinflammatory grade (≥G2) | 1.494 (0.622–3.591) | 0.369 | 0.402 |
Moreover, analysis was also performed among a subgroup of patients with available hepatic HBcAg IH score or quantitative serum HBsAg data. Weak to moderate negative correlations of hepatic HBcAg with ALT, AST, and liver necroinflammation grade were observed in those HBeAg+/DNA high patients, but not in HBeAg+/DNA low patients with HBV DNA ≤2×106 IU/mL (Table 6). Moderate negative correlations of serum HBsAg with ALT, AST, and liver necroinflammation grade were observed in HBeAg+/DNA high patients but became weakened in HBeAg+/DNA low patients in both cohort A and the Nanjing cohort (Supplementary Table 13).
Table 6Spearman relationship between serum HBsAg/Hepatic HBcAg and clinical index for liver injury for HBeAg positive patients with different HBV DNA levels in cohort A
| Spearman Rho (P value) between Different variables | HBV DNA >2×106 IU/mL (n=286)
| HBV DNA ≤2×106 IU/mL (n=43)
|
|---|
| Serum HBsAg | Hepatic HBcAg | Serum HBsAg | Hepatic HBcAg |
|---|
| ALT | −0.325 (0.000) | −0.290 (0.000) | −0.248 (0.109) | 0.006 (0.972) |
| AST | −0.402 (0.000) | −0.279 (0.000) | −0.231 (0.135) | 0.018 (0.908) |
| GGT | −0.278 (0.000) | −0.243 (0.000) | −0.075 (0.635) | −0.083 (0.596) |
| Liver necroinflammatory grade | −0.403 (0.000) | −0.236 (0.000) | −0.011 (0.942) | 0.117 (0.455) |
The results above reveal a statistically significant negative relationship between the level of HBV DNA and the severity of liver inflammation only in HBeAg-positive patients with higher HBV DNA levels (with HBV DNA >2×106 IU/mL in this study). However, in HBeAg positive patients with lower HBV DNA levels, this negative correlation weakened or disappeared, and even changed to a positive correlation, similar to in those HBeAg negative patients.
Discussion
In this study, we explore the relationship between serum HBV DNA and hepatic necroinflammatory damage, especially for patients with different HBeAg status and HBV DNA levels. In general, in both cohort A and cohort B, weak but statistically significant negative correlations between serum HBV DNA level and the severity of hepatic necroinflammation were observed in HBeAg-positive patients. In contrast, such a negative correlation converted to a moderate positive with the loss of HBeAg. In line with these, similar correlation patterns between HBV DNA and biochemical indicators of liver injury were also observed. These phenomena were partially confirmed by meta-analysis. Unexpectedly, we observed the disappearance or reversal of the negative correlation between viral replication activity and most indicators of liver inflammation damage even in HBeAg-positive patients with HBV DNA ≤2×106 IU/mL, which were confirmed by the validation cohort composed of 2,624 HBeAg-positive patients, including 457 patients with liver biopsy available.
The key finding of our study demonstrates that a marginal negative relationship between HBV DNA and liver damage mainly exists in HBeAg+/DNA-high (herein in this study, >2×106 IU/mL) patients. Noticeably, the relationship between HBV DNA and liver damage had already become positive in HBeAg+/DNA-low (≤2×106 IU/mL) patients, similar to that observed in HBeAg-negative patients. A schematic figure was drawn to illustrate these “opposite relationships” observed in this study (Fig. 2). It demonstrated that patients in the yellow zone with moderate HBV DNA levels, especially for those with HBV DNA surrounding the threshold 2×106 IU/mL, suffer from more active inflammation and probably possess a higher risk of disease progression, with/without HBeAg loss. In line with this, recent studies reported an increased risk of HCC development for CHB patients with moderate (104–108 IU/mL) viral replication levels compared to those with HBV DNA >108 IU/mL or <104 IU/mL, particularly for those with HBV DNA levels between 106–107 IU/mL.31,32 Thus, we think that a more aggressive treatment strategy should be recommended for patients with moderate HBV DNA levels regardless of HBeAg status (especially for patients with HBV DNA around 2×106 IU/mL).
Notably, in our study, most HBeAg-positive patients with HBV DNA >2×106 IU/mL were not in the “real” immune tolerance status, only 11.17% of patients in cohort A and 17.76% of patients in cohort B showed no indication of liver damage, characterized by normal ALT, AST, GGT, and no to mild liver necroinflammation (G=0/1). Thus, a negative correlation between HBV DNA level and the severity of hepatic inflammatory damage was observed in these patients. Unfortunately, antiviral treatment is rarely initiated in HBeAg-positive patients with HBV DNA >2×106 IU/mL, who are usually assumed to be in the immune-tolerance phase as their HBV DNA >1×106 IU/mL.6,33 Based on our observations, we sought to determine whether therapeutic efficacy may be improved if antiviral treatment is launched in HBeAg-positive patients with HBV DNA around 2×106 IU/mL, supporting the suggestion to “treat-all” individuals with chronic HBV infection
The characteristics of “the negative relationship of active virus replication and hepatic inflammation” and “the positive relationship of virus replication and hepatic inflammation” observed in our studies seemed to constitute two patterns of ALT flares, “good flares” and “bad flares”, as previously described.34,35 The former is usually related to a decrease in HBV DNA and probably leads to the loss of HBeAg and even HBsAg and is therefore considered “good” host immune-induced inflammation. The latter is always relevant to persistent liver injury without efficient control of HBV replication and has been considered a virus-induced injury.36 Based on our current study, a “good flare” could be expected to occur in HBeAg-positive patients with higher HBV DNA levels (>2×106 IU/mL). On the opposite, a “bad flare” would occur most likely in HBeAg-negative patients. A potential explanation for this phenomenon could be that in HBeAg-positive patients, especially those with high HBV DNA levels, almost all hepatocytes are accompanied by HBV infection and active transcription.37,38 Thus, even nonspecific immune-induced hepatocyte injury would result in the number of a reduction in HBV-infected hepatocytes, which ultimately reduces the HBV DNA level. While in HBeAg-negative patients, activation of the host’s immune response induced by HBV replication might cause more serious immunopathological damage, than the clearance of cccDNA-positive hepatocytes, because usually only a few hepatocytes were cccDNA transcriptionally active as indicated by the extremely lower rate of HBcAg positive hepatocytes in previous and current studies.38
Although this study provides insight into the relationship between viral replication and the severity of liver inflammatory damage in CHB patients with different virological characteristics, there are still some limitations. First, this was a retrospective study. To address this, we expanded the sample size through meta-analysis to minimize the impact of study bias. Second, due to the shortcomings of retrospective and cross-sectional studies, detection and collection of the immune indicators of CHB patients, especially HBV-specific immune data were not available. Thus, we use liver damage as the indirect indicator to reflect the host’s immune response in this study. Though most likely the severity of liver damage is mediated by the host’s immune response against HBV, hepatic damage is the outcome but not the direct indicator of the host’s immune response against HBV infection. We expect that a prospective, large-scale, follow-up cohort study with multiple immune index detections (especially HBV-specific immunity) will be conducted in the future to investigate the causes of this phenomenon.
In summary, our study reports distinguishable relationship patterns between serum HBV DNA and liver damage, in patients with different HBeAg statuses and/or HBV DNA levels. The relationship between viral replication and the severity of hepatic necroinflammatory damage changes before the loss of HBeAg in patients with chronic hepatitis B virus infection.
Supporting information
Supplementary Fig. 1
Flow diagram of Study selection.
Search terms were described as follow: (1) (relation) OR (relationship) OR (correlation) OR (relevance) OR (relevant) OR (correlate) OR (related); (2) (Alanine Transaminase) or (Transaminase, Alanine) or (Glutamic-Alanine Transaminase) or (Glutamic Alanine Transaminase) or (Transaminase, Glutamic-Alanine) or (Alanine-2-Oxoglutarate Aminotransferase) or (Alanine 2 Oxoglutarate Aminotransferase) or (Aminotransferase, Alanine-2-Oxoglutarate) or (Alanine Aminotransferase) or (Aminotransferase, Alanine) or (Glutamic-Pyruvic Transaminase) or (Glutamic Pyruvic Transaminase) or (Transaminase, Glutamic-Pyruvic) or (SGPT) or (ALT); (3)HBV DNA; (4)HBeAg. SGPT, serum glutamic pyruvic transaminase; ALT, alanine aminotransferase; HBeAg, hepatitis B e antigen.
(DOCX)
Supplementary Fig. 2
Forrest plots for value the relationship between HBV DNA and ALT in different studies of patients with (A) HBeAg-negative status; (B) HBeAg-positive status.
Forrest plots summarize Fisher’s Z transformed from R value to calculated overall Fisher’s Z value and R-value. HBeAg, hepatitis B e antigen; ALT, alanine aminotransferase.
(DOCX)
Supplementary Fig. 3
Forrest plots for the relationship between HBV DNA and ALT in different studies of patients with HBeAg negative.
After excluded four studies (Schwarz 2019; Hou 2013; Ijaz 2011; Sangfelt 2004), I2 decreased to 77.2%, overall R value was 0.411 (95%CI: 0.321–0.494). HBeAg, hepatitis B e antigen; ALT, alanine aminotransferase.
(DOCX)
Supplementary Table 1
The raw data extracted from each study for CHB patients with negative HBeAg.
(DOCX)
Supplementary Table 2
The raw data extracted from each study for CHB patients with positive HBeAg.
(DOCX)
Supplementary Table 3
Study quality and risk of bias from each study for CHB patients with negative HBeAg.
(DOCX)
Supplementary Table 4
Study quality and risk of bias from each study for CHB patients with positive HBeAg.
(DOCX)
Supplementary Table 5
Clinical characteristics of cross-sectional cohorts A and B.
(DOCX)
Supplementary Table 6
Clinical characteristics of HBeAg positive patients with different HBV DNA levels for cohort A, cohort B, and Nanjing cohort.
(DOCX)
Supplementary Table 7
Spearman relationship between serum HBV DNA and ALT, AST, GGT, and liver necroinflammatory grade for HBeAg-positive CHB with different HBV DNA levels (under different fibrosis stages).
(DOCX)
Supplementary Table 8
Spearman relationship between serum HBV DNA and ALT, AST, GGT, and liver necroinflammatory grade for HBeAg-positive patients (anti-HBe negative) with different HBV DNA levels and HBeAg-negative patients (anti-HBe positive).
(DOCX)
Supplementary Table 9
Spearman relationship between serum HBV DNA and liver inflammation in patients with different HBeAg statuses.
(DOCX)
Supplementary Table 10
Spearman relationship between serum HBV DNA and liver inflammation in HBeAg positive patients with different HBV DNA level.
(DOCX)
Supplementary Table 11
Multiple linear regression analysis between HBV DNA change and active liver inflammation in HBeAg-positive patients with different HBV DNA levels (cohort B).
(DOCX)
Supplementary Table 12
Binary logistic regression analysis between HBV DNA and liver inflammation in HBeAg-positive patients with different HBV DNA levels (cohort B).
(DOCX)
Supplementary Table 13
Spearman relationship between serum HBsAg and clinical index for liver injury for HBeAg-positive patients with different HBV DNA levels in cohort A and independent validation Nanjing cohort.
(DOCX)
Abbreviations
- ALT:
alanine transaminase
- AST:
aspartate aminotransferase
- cccDNA:
covalently closed circular DNA
- GGT:
gamma-glutamyl transpeptidase
- HBcAg:
hepatitis B core antigen
- HBeAg:
hepatitis B e antigen
- HBV:
hepatitis B virus
- IH:
immunohistochemistry
Declarations
Ethical statement
This study was approved by the Ethics Committee of the Fifth Medical Center of the PLA General Hospital of Beijing (No. KY-2022-1-4-1). All procedures were conducted according to the principles of the Declaration of Helsinki. Written informed consent was waived.
Data sharing statement
Data supporting the findings of this study are available from the corresponding author upon reasonable request.
Funding
This work was supported by the National Natural Science Foundation of China (No. 82072280) and Beijing Natural Science Foundation (No. 7212063).
Conflict of interest
FL has been an Editorial Board Member of Journal of Clinical and Translational Hepatology since 2013, JL has been an Editorial Board Member of Journal of Clinical and Translational Hepatology since 2024, XZ has been an Editorial Board Member of Journal of Clinical and Translational Hepatology since 2013. The other authors have no conflict of interests related to this publication
Authors’ contributions
FL and LW designed and revised the paper; LW analysed data, wrote the paper, and made figures; KZ analysed data; JW and XZ collected data and analysed the data; LJ revised the paper and analysed data. JL and JZ contributed to the study design and manuscript revision. All authors read and approved the final manuscript