Age as a Predictor of Significant Fibrosis Features in HBeAg-Negative Chronic Hepatitis B Virus Infection with Persistently Normal Alanine Aminotransferase

Background Although alanine aminotransferase (ALT) levels reflect the degree of liver damage, not all patients with chronic hepatitis B virus (HBV) infection exhibit persistently elevated ALT levels. In the present study, we aimed to comprehensively evaluate the characteristics of histological abnormalities in a large population of Chinese patients with chronic HBV and persistently normal ALT levels. Methods In total, 2303 consecutive patients who underwent liver biopsy were screened. Of these patients, 273 were categorized as having persistently normal ALT levels (PNALT), whereas 618 were categorized as having persistently or intermittently elevated ALT levels (PIALT). All these patients had at least three ALT values recorded in the year prior to the baseline liver biopsy. Results Significant necroinflammation was observed in 9.7% (11/113) patients with PNALT, 23.3% (42/180) patients with PIALT (ALT 1–2× upper limit of normal [ULN]), and 27.8% (42/151) patients with PIALT (ALT > 2× ULN), whereas significant fibrosis was observed in 8.8% (10/113) patients with PNALT, 27.8% (42/151) patients with PIALT (ALT 1–2× ULN), and 21.2% (32/151) patients with PIALT (ALT > 2× ULN). Multiple logistic regression analysis indicated that age parameters were associated with significant histological abnormalities in patients with PNALT. The area under the curve showed that age was associated with significant fibrosis characteristics in patients with hepatitis B extracellular antigen (HBeAg)-negative PNALT. Conclusion Significant histological abnormalities are not often observed in Chinese patients with PNALT. Interestingly, age appears to be a predictor of significant fibrosis in patients with HBeAg-negative PNALT.


Introduction
Approximately one-third of the world's population possesses serological evidence of past or present hepatitis B virus (HBV) infection, and 350-400 million people are known to be chronic HBV surface antigen (HBsAg) carriers. The disease spectrum and natural history of chronic HBV infection are diverse and variable; they range from having an inactive carrier status to having progressive chronic hepatitis B (CHB), which may then progress to cirrhosis and hepatocellular carcinoma (HCC) [1,2]. Chronic HBV infection is a dynamic process, and its natural history was schematically divided into five phases by the European Associated for the Study of the Liver (EASL) Clinical Practice Guidelines (2012) [1]: (1) the "immune tolerant" phase is characterized by hepatitis B extracellular antigen (HBeAg) positivity, high levels of HBV replication (reflected by high levels of serum HBV DNA), normal or low aminotransferase levels, mild or no liver necroinflammation, and no or slow fibrosis progression; (2) the "immune reactive HBeAg-positive phase" is characterized by HBeAg positivity, relatively lower levels of viral replication compared to the immune tolerant phase (as reflected by lower serum HBV DNA levels), increased or fluctuating aminotransferase levels, moderate or severe liver necroinflammation, and more rapid progression of fibrosis compared to the previous phase; (3) the "inactive HBV carrier state," which may follow seroconversion from HBeAg to anti-HBe, is characterized by very low or undetectable serum HBV DNA levels and normal serum aminotransferase levels; (4) "HBeAg-negative CHB" may follow seroconversion from HBeAg to anti-HBe during the immune reactive phase or may develop after years or decades in an inactive carrier state; and (5) the "HBsAg-negative CHB" or "HBsAg-negative" phase.
Although serum levels of alanine transaminase (ALT), an enzyme released by hepatocytes during liver injury, usually reflect the degree of liver damage [3], not all patients with chronic HBV infection have persistently elevated ALT levels. Patients in the immune tolerant phase and inactive HBV carrier state have persistently normal ALT (PNALT) levels, whereas a proportion of patients with HBeAg-negative CHB may have intermittently normal ALT levels. There have also been reports of histological injury in patients with normal ALT levels [4][5][6][7][8][9]. Furthermore, some large cohort studies have shown that patients with CHB and normal serum ALT levels were at risk of developing cirrhosis and HCC, [10,11] irrespective of the HBeAg status.
In the present study, we aimed to comprehensively evaluate the characteristics of histological abnormalities in a large population of Chinese patients with CHB and PNALT or persistently or intermittently elevated ALT (PIALT) levels. Moreover, we aimed to analyze the factors associated with significant histological changes in patients with PNALT and PIALT.

Patients
This was a retrospective study of patients with CHB between January 2005 and June 2010 at the Liver Clinic and Department of Hepatosis, The Third Hospital of Zhenjiang Affiliated Jiangsu University. Patients were examined every 3-6 months, or more often if clinically indicated. At each visit, the results of liver biochemistry, ultrasonography, and HBV serology, including HBsAg, HBeAg, anti-HBe, HBV DNA levels, and genotype were assessed. The inclusion criteria were as follows [12]: (1) HBsAg-positive for at least the previous 6 months; (2) HBV DNA level of >1,000 copies/mL; and (3) categorized as having PNALT levels (at least three ALT values taken in the year prior to baseline liver biopsy, of which all values were >40 IU/L and had remained so until the start of treatment or the last follow-up if not treated). Patients were categorized as having PIALT levels if they had at least three ALT values taken, at least one of which was >40 IU/L in the year prior to the baseline biopsy, or any time prior to the start of treatment or last follow-up if not treated (intermittently elevated) [4,5,7,9,13,14]. The exclusion criteria were as follows: (1) hepatitis A, C, or D, or human immunodeficiency virus co-infection; (2) evidence of liver disease of another etiology; (3) use of hepatotoxic drugs or regular consumption of alcohol; (4) received previous antiviral (HBV) therapy or any liver functional protection therapy to alleviate hepatic inflammation; and (5) less than three normal ALT values taken prior to the biopsy. When a patient fulfilled the inclusion criteria and had no contraindications to liver biopsy including the presence of ascites, an international normalized ratio of >1.5, activated partial thromboplastin time of >1.5 times, or platelet count of <100,000/mL, the doctor would recommend a liver biopsy. Ultimately, 2303 consecutive patients who underwent liver biopsy were screened. Informed written consent for the liver biopsy was obtained from every patient.
The study protocol was conducted as per the guidelines of the 1975 Declaration of Helsinki and approved by the ethics committee of The Third Hospital of Zhenjiang Affiliated Jiangsu University. Written informed consent was obtained from all subjects.

Biochemical and Serological Tests
Biochemical tests and complete blood cell counts were performed using routine automated analyzers. The normal upper limit of the ALT level was 40 IU/L. Insulin resistance (IR) was determined using homeostasis model of assessment (HOMA) equations [15], HOMA-IR = [fasting plasma insulin (mIU/l) × fasting plasma glucose (FPG) (mmol/l)]/22.5. HBsAg, HBeAg, and anti-HBe levels were assayed with commercially available enzyme-linked immunosorbent assay (ELISA) kits. The HBV DNA level was measured by real-time PCR, with a lower detection limit of 1000 copies/mL (DaAn Gene Co, shanghai,China).

Genotype Determination by Multiplex PCR
Genotyping was performed by multiplex PCR using specific primers for each HBV genotype (A-F) [16].

Statistical Analysis
Results are presented as median (range) or mean ± SD, as appropriate. The patients' demographic and clinical features were analyzed using the Statistical Package for the Social Sciences (SPSS) version 21.0 (SPSS Inc., Chicago, IL, USA). Statistical analyses were performed using chi-squared and Fisher's exact tests for categorical variables, while Student's t-test or one-way analysis of variance was used for group comparisons of parametric quantitative data. Multinomial (binary) logistic regression and an ordinal logit model were applied to evaluate factors predicting significant fibrosis based on histology. The predicted probabilities of the parameters were used as a surrogate marker to construct receiver operating characteristic (ROC) curves. ROC and regression analyses were performed using MedCalc (Version 10.4.7.0; MedCalc, Mariakerke, Belgium). All P values were two-sided.   Table 1. Patients with PIALT tended to be older and male and have higher levels of HBV DNA than patients with PNALT. The platelet counts of the patients with PIALT were lower than those with PNALT. Significant necroinflammation was noted in 9.7% (11/113)  We found that age, aspartate aminotransferase (AST) level, HBsAg level, HBV DNA level, and histological abnormalities differed significantly between the PNALT and PIALT groups. When PNALT/PIALT was considered as a binary dependent variable, we used multiple logistic (binary) regression analysis to assess the factors associated with PNALT and PIALT in HBeAgpositive patients ( Table 1). The results suggested that age of 30-40 years, HBsAg level, HBV DNA10 7 copies/Land AST level were associated with PNALT and PIALT in HBeAg-positive patients. We also found significant histological abnormalities associated with PNALT and PIALT.
We found that age, AST level, HBsAg level, and histological abnormalities differed significantly between the PNALT and PIALT groups. When PNALT/PIALT was considered as a binary dependent variable, we used multiple logistic (binary) regression analysis to assess factors associated with PNALT and PIALT in HBeAg-negative patients ( Table 2). Using the "enter" method, the results suggested that an age of 40-60 years, AST level, and albumin level were associated with significant histological abnormalities in patients with PNALT and PIALT.

Analysis of Factors Associated with Significant Necroinflammation in HBeAg-Positive and -Negative Patients with PNALT
When the presence or absence of significant necroinflammation was considered as a binary dependent variable in HBeAg-positive and -negative patients with PNALT, univariate analysis indicated that ALT and age were associated with significant histological abnormalities in HBeAg-positive and -negative patients with PNALT, respectively (Tables 3 and 4). However, multiple logistic regression analysis indicated that no parameter was associated with significant histological abnormalities in patients with PNALT, regardless of the HBeAg-positive or -negative status (Tables 3 and 4).

Analysis of Factors Associated with Significant Fibrosis in HBeAg-Positive and -Negative Patients with PNALT
When the presence or absence of significant fibrosis was considered as a binary dependent variable in HBeAg-positive and -negative patients with PNALT, univariate and multiple logistic regression analyses indicated that no parameter was associated with significant fibrosis in HBeAg-positive patients with PNALT (Table 5). However, in HBeAg-negative patients with PNALT, univariate analysis indicated that age and PLT level were associated with significant fibrosis, whereas multiple logistic regression analysis indicated that age 40-60 years was associated with significant fibrosis (Table 6).     Fig 2A). The AUC of the ALT level associated with the diagnosis of significant necroinflammation in HBeAg-negative patients was 0.618 (95% CI, 0.554-0.681; sensitivity = 82.3%, specificity = 39.9%, P<0.001; Fig 2B). The AUC of the ALT level associated with the diagnosis of significant fibrosis in HBeAg-positive patients was 0.564 (95% CI, 0.523-0.645; sensitivity = 61.9%, specificity = 55.1%, P = 0.0526; Fig 2C). The AUC of the ALT level associated with the diagnosis of significant fibrosis in HBeAg-negative patients was 0.59 (95% CI, 0.527-0.653; sensitivity = 79.3%, specificity = 39.3%, P = 0.0049; Fig 2D).

AUC of Age Associated with Significant Histological Fibrosis in HBeAg-Positive and -Negative Patients with PNALT
Age was used to predict the probability of being diagnosed with significant fibrosis abnormalities in HBeAg-positive and -negative patients with PNALT. The AUC of the age associated with the diagnosis of significant fibrosis abnormalities in HBeAg-positive PNALT patients was 0.612 (95% CI, 0.516-0.702; sensitivity = 54.5%, specificity = 64.6%, P = 0.192; Fig 3A). The AUC of the age associated with the diagnosis of significant fibrosis in HBeAg-negative patients with PNALT was 0.672 (95% CI, 0.593-0.74; sensitivity = 66.7%, specificity = 75.4%, P = 0.032; Fig 3B).

Discussion
Despite its invasive nature, liver biopsy remains the gold standard for assessing histological liver abnormalities [20]. The 2012 Asian-Pacific Association for the Study of the Liver (APASL) guidelines recommended that liver biopsies should be considered in patients with viremic CHB aged >40 years, especially in those with high normal or minimally increased ALT levels [1,21]. Previous studies focusing on the histological characteristics of the liver in such cases have been limited by a small sample size, a focus on patients with elevated ALT levels, or a follow-up time of <6-12 months [22][23][24][25]. Therefore, in the present study, we considered HBV DNA levels of >1000 copies/mL and limited ALT level assessments to at least every 3 months for at least 1 year [26]. We aimed to comprehensively evaluate the histological characteristics of a large population of Chinese patients with CHB and PNALT or elevated ALT levels. Some reports indicated that HBeAg-positive patients with PNALT usually have high levels of HBV DNA with no or minimal liver histological changes [27,28]. We examined the histological features in HBeAg-positive patients with PNALT and PIALT. Although these patients had lower active significant necroinflammatory and fibrosis scores than HBeAg-positive patients with elevated ALT levels, our results demonstrate that significant necroinflammatory and fibrosis scores were found in HBeAg-positive patients with PNALT (9.7% and 8%). We also  [7,29]. These studies reported more significant histological abnormalities than that observed in the present study. We believe that the reason for this is that these patients had elevated ALT levels at follow-up, whereas the histological abnormality results were similar to those of our patients with PIALT. A systematic review [5] illustrated 215 patients with serum HBV DNA levels of <20,000 IU/mL from four studies with good or acceptable definitions of PNALT [7,[30][31][32]. Necroinflammatory activity was more than just minimal in five (7%) and fibrosis was more than just mild in seven (10%) of 73 patients with serum HBV DNA levels of 2000-20,000 IU/mL as well as in two (1.4%) and one (0.7%) of 142 patients with serum HBV DNA levels of <2000 IU/mL, respectively. However, the 2012 EASL guidelines recommended that liver biopsies should be considered in immune tolerant patients aged >30 years in addition to patients with fluctuating ALT levels [26]. Advice on which patients should be recommended for liver biopsy, especially those with normal ALT levels, vary among these guidelines. Multiple logistic regression analysis of age revealed that an age of 30-40 years was associated with PNALT and PIALT in patients with CHB if the AST level was excluded. We stratified patients based on an age of 30 years according to the EASL guidelines. We found significant necroinflammation in 4.3% (4/93) and significant fibrosis in 4.4% (4/90) patients with PNALT. The results suggest that patients aged >30 years in the PIALT group had a higher frequency of necroinflammation (25.89%, 160/618) than those aged <30 years in the PNALT group (9.89%, 27/273; P<0.001). Similarly, patients aged >30 years in the PIALT group had a higher frequency of necroinflammation (23.75%, 153/618) than those aged <30 years in the PNALT group (9.89%, 33/273, P<0.001).
One study by Wong et al. from Hong Kong indicated that the risk of liver fibrosis increases after the age of 35 years in HBeAg-positive patients [33]. Nevertheless, patients with advanced fibrosis (F3-F4) were identified based on transient elastography measurement instead of histological evidence. Interestingly, we found that the AUC of the age associated with the diagnosis of significant fibrosis in HBeAg-negative patients with PNALT had a cutoff value of 42 years. Zoulim et al. recently suggested starting therapy in all patients with normal ALT levels who show relatively low levels of viremia, not just in patients aged >40 years, but also among patients in their 20s. This represents a step further than that reported in the major guidelines. The author demonstrated that many unnecessary deaths might be prevented by earlier antiviral intervention; additionally, considering its long duration, initiation and promotion may both be significant during the immune tolerant phase and may increase the risk of HCC later in life, even in the absence Age as a Predictor in Chronic Hepatitis B Patients of cirrhosis [34]. However, in a prospective randomized, double-blind, placebo-controlled study, 380 patients with CHB and PNALT were screened while receiving entecavir therapy for 1 year and achieved virological improvements but no histological benefits [35]. Therefore, we supported the proposal that patients with CHB and PNALT usually have high levels of HBV DNA with no or minimal liver histological changes. In addition, trials have shown that these patients tend to not have a good response to the currently available antiviral therapy [36,37].
ALT, as a biomarker of liver injury, also predicts sensitivity and specificity in liver disease and is associated with significant histological characteristics to some extent in HBeAg-positive and -negative patients. The AUC of the ALT level associated with the diagnosis of significant necroinflammation in HBeAg-positive patients was 0.584 (P = 0.007). The AUC of the ALT level associated with diagnosis of significant necroinflammation in HBeAg-negative patients was 0.618 (P = 0.003). The AUC of the ALT level associated with the diagnosis of significant fibrosis in HBeAg-negative patients was 0.59 (P = 0.0049). Obviously, from these results, it can be concluded that the ALT level was also good at predicting the probability of a diagnosis of significant histological characteristics in PIALT patients compared to PNALT patients.
In conclusion, significant histological abnormalities are not often observed in Chinese patients with PNALT levels when good or acceptable definitions of PNALT are used. However, ALT level is a predictor of significant histological characteristics between PNALT and elevated ALT levels. Interestingly, age appears to be a predictor of a diagnosis of significant fibrosis in HBeAg-negative patients with PNALT.