Outcome in Caucasian patients with hepatitis B e antigen negative chronic infection: A long‐term observational cohort study

Abstract Sensitive polymerase chain reaction assays to measure hepatitis B virus (HBV) DNA became only available the last decade. Hence, the long‐term outcome of Caucasian patients in Western Europe with hepatitis B e antigen (HBeAg)‐negative chronic infection, especially with a baseline HBV DNA level ⩾2000 IU/mL, is still unclear. Out of a cohort of 1936 chronic HBV patients, 413 Caucasian individuals were identified with HBeAg‐negative chronic infection, defined as persistently normal alanine aminotransferase (ALT) levels and HBV DNA levels <20 000 IU/mL. During a mean follow‐up of 12 years, 366 (88.6%) maintained an HBeAg‐negative chronic infection status, whereas 25 (6.1%) developed chronic active hepatitis (CAH). In total, Nine of these 25 CAH cases were related to immunosuppression. In total, 22 (5.3%) individuals had ALT > 2 × upper limit of normal due to non‐HBV‐related causes. The cumulative probability of spontaneously developing CAH after 10 years was almost exclusively seen in patients with baseline HBV DNA level ⩾2000 IU/mL (11.7% vs 1.2%; P < .001). Advanced liver disease developed significantly more in patients with baseline HBV DNA level ⩾2000 IU/mL (5.2% vs 1.5%; P = .018) and occurred especially in patients with obesity (16.7% vs 4.2%; P = .049). The incidence of hepatocellular carcinoma was 0.0%. Caucasian patients with HBeAg‐negative chronic infection and baseline HBV DNA level <2000 IU/mL have an excellent long‐term prognosis in the absence of immunosuppressive therapy. However, patients with baseline HBV DNA level ⩾2000 IU/mL are at risk to develop advanced liver disease.


| INTRODUCTION
Infection with hepatitis B virus (HBV) remains a global health challenge with approximately 257 million people living with chronic HBV infection, of whom 887 000 die annually from HBV-related complications such as cirrhosis and hepatocellular carcinoma (HCC). 1 The natural history of chronic HBV infection can be divided in four phases, taking into account the presence of hepatitis B e antigen (HBeAg), serum HBV DNA levels, and serum alanine aminotransferase (ALT) values. [2][3][4] One of these phases is the HBeAg-negative chronic infection state. 4 This stage was previously called "inactive carrier" phase, but this terminology was abandoned since previous studies indicated that some of these patients developed advanced liver disease. 4 According to current management guidelines on hepatitis B, therapy is not indicated in patients with HBeAg-negative chronic infection but follow-up for the risk of HBV reactivation, advanced liver disease, and HCC is recommended. [2][3][4] Most evidence on the natural history of HBeAg-negative chronic infection is based on studies in Asian patients, wherein the high sensitivity of polymerase chain reaction (PCR)-based HBV DNA assays has contributed to our understanding that even low HBV DNA levels may still be associated with the risk of liver disease progression. [5][6][7] However, only a limited number of studies have assessed the risk of disease progression and its predictors in Caucasian patients. Previous European studies on the natural history of patients in the chronic infection phase were limited by the short duration of follow-up (less than 10-year followup), small groups of patients (<200 subjects), poorly defined criteria for HBeAg-negative chronic infection, and the use of the low sensitivity branched HBV DNA assays for quantification. [8][9][10][11][12][13][14][15][16][17][18] Recent sensitive PCR assays to measure HBV DNA have demonstrated that, in general, patients with HBeAg-negative chronic infection have HBV DNA levels <2000 IU/mL. However, some patients in this phase have HBV DNA levels between 2000 and 20 000 IU/mL. 3,4 The risk of liver disease progression is unclear in this subgroup of Caucasian patients with baseline HBV DNA levels ⩾2000 IU/mL. This is the first long-term follow-up study of more than 10 years in HBeAg-negative Caucasian patients making use of the high sensitive PCR-based assays. The main objectives were to investigate the disease outcome of those with HBV DNA levels ⩾2000 IU/mL and to find out whether a cut-off value of serum HBV DNA level of 2000 IU/ mL can predict those patients who would benefit from a strict followup program and those who do not require a stringent monitoring.

| PATIENTS AND METHODS
The study involved three large educational hospitals, two in Belgium and one in the Netherlands. The data were to be collected according to a protocol that followed the strengthening the reporting of observational studies in epidemiology statement with clear definitions of the study population, follow-up, and outcome data.

| Study population
The study identified all consecutive chronic HBV patients of the participating centers between 1 January 1987 and 31 July 2018 who fulfilled the following inclusion criteria: (a) Caucasian race, (b) persistence of HBsAg for at least 1 year, (c) the presence of antibodies to HBeAg (anti-HBe) without HBeAg, (d) low HBV DNA (<20 000 IU/mL), (e) persistently normal ALT levels, defined as ⩾3 ALT determinations (⩽40 IU/L) at least 2 months apart over a period of at least 12 months, 5,6,19 and (f) follow-up at the enrolling center for a minimum period of 12 months. Patients previously treated with HBV antiviral agents were excluded as well as HBsAg-positive patients with prophylactic administration of antiviral therapy. Other

| Definition of clinical events
The primary outcome was the development of chronic active hepatitis (CAH), defined by increased ALT levels to more than twice the upper limit of normal (ULN) on two occasions at least 2 weeks apart with HBV DNA levels ⩾2000 IU/mL whether or not with HBeAg reversion. [2][3][4] Increased ALT levels > 2 x ULN that could not be classified as CAH were designated as non-HBV-related cause. 22 According to the report of the Baveno VI Consensus Workshop, advanced liver disease was suspected on transient elastography values >10 kPa. 23 The development of cirrhosis was defined as a clinical syndrome consisting of either histological confirmation of 3374 | cirrhosis or ultrasonographic findings of cirrhosis. 14 Cirrhosis was further classified according to the Child-Pugh score. 24 HCC diagnosis was based on noninvasive criteria (positive lesion detected by at least two different imaging techniques) or pathology. 14 In addition, this study evaluated the incidence of mortality and liver-related mortality.

| Immunosuppressive treatment
This study also assessed whether patients who developed a CAH were exposed to immunosuppressive or cancer chemotherapy. The immunosuppressive therapies were classified into those with low (eg, azathioprine), moderate (cyclosporine), and high (eg, rituximab, highdose corticosteroids, infliximab) risk of reactivation as previously outlined by Loomba and Liang. 25 CAH resulting from immunosuppressive therapy was considered up to 6 months after cessation of immunosuppression, and in the case of B-cell depleting drugs (eg, rituximab) as late as 12 months posttreatment. 26

| Statistical analysis
Statistical analysis was performed with the SPSS software version 25 (IBM Corp, Armonk, NY). Continuous variables are expressed as mean + standard deviation or medians + interquartile range as appropriate. For the comparison of categorical variables, either the χ 2 test or the Fisher's exact test was used. The Student t test or Mann-Whitney U nonparametric test was used to analyse continuous variables between two independent groups. The level of statistical significance was set at P < .05 in two-tailed tests.
Estimates on the rate of CAH and advanced liver disease were calculated using the Kaplan-Meier method, and the difference was determined using the logrank test. Univariate analyses (logrank tests) to identify variables associated with CAH or advanced liver disease included age at diagnosis (<40 vs ⩾40 years), sex (male vs female), obesity (yes vs no), baseline ALT level (low-normal < 0.5 × ULN vs high-normal 0.5 -1 × ULN), baseline quantitative HBsAg (qHBsAg) level (<1000 vs ⩾1000 IU/mL) and baseline HBV DNA level (<2000 vs ⩾2000 IU/mL). Multivariate Cox regression analysis was not performed due to expected data sparseness, that is, CAH and advanced liver disease were not expected to occur in all converges of the abovementioned variables.     In studies with a longer follow-up, the prevalence of cirrhosis varied widely from 0.4% to 17.5%. [12][13][14][15] Correlation between baseline HBV DNA level and worse prognosis was not evaluated in these studies. [12][13][14][15] Moreover, in the prior studies, patients' serum HBV DNA level was quantified by an insensitive hybridization method.

| Clinical outcome
Hence, some patients may have been misclassified as HBeAgnegative chronic infection without use of PCR-based assays.
That the level of viremia can be important has been illustrated in the REVEAL study in Asian patients. There was an increased risk of cirrhosis, HCC, and liver-related mortality in HBeAg-negative patients with baseline normal ALT level and HBV DNA level ⩾2000 IU/mL. 7 There are only a few studies that have addressed the issue of CAH development in Caucasian HBeAg-negative patients with persistently normal ALT values and HBV DNA levels below 20 000 IU/mL on sensitive PCR-based assays. 8,9,11,17 With a limited follow-up between 1 and 5 years, the overall prevalence of progression to CAH in these studies was 0.0% in patients with HBV DNA level <2000 IU/mL and varied widely from 0.0% to 50.0% in those with HBV DNA level ⩾2000 IU/mL.
Our finding of spontaneous progression to CAH over 10 years was 1.2% and 11.7% in those patients with baseline HBV DNA level <2000 IU/mL and HBV DNA level ⩾2000 IU/mL, respectively. This is lower than previously reported and could be ascribed to differences in study population, follow-up duration, and the use of different criteria for CAH. 8,9,11,17 Moreover, the number of patients with baseline HBV DNA level ⩾2000 IU/mL in the previous studies was small (n = 4-46). 8,9,11,17 In the current cohort, 81 patients had a baseline HBV DNA level ⩾2000 IU/mL. In line with our findings, Papatheodoridis et al 9 confirmed the relation between baseline HBV DNA level ⩾2000 IU/mL and the risk of HBV reactivation in patients in Greece.
KOC ET AL.

| 3377
Recent data have shown that hepatitis B surface antigen quantification might be helpful in the decision on the frequency of followup in such patients. [2][3][4]27 However, this test is not routinely available in daily practice. In line with prior studies, our patients-in which we could do this test-had a median qHBsAg level <1000 IU/mL and qHBsAg level could identify patients at higher risk of progression in the subgroup with baseline HBV DNA level ⩾2000 IU/mL. 11,27 As a consequence of the growing obesity epidemic, more and more chronic HBV patients with coexisting NASH are expected. 28 NASH is an independent risk factor for cirrhosis and HCC, and in our study obesity was also a risk factor for advanced liver disease among those patients with baseline HBV DNA level ⩾2000 IU/mL. 28 We could also confirm that CAH was triggered by immunosuppressive therapy or cancer chemotherapy. In that respect, we found that immunosuppression was associated with the development of CAH in more than one-third of the patients, underlying the importance of screening for hepatitis B before starting immunosuppressive therapies. [2][3][4] One limitation of the current study is that variables such as genotype, precore/core mutations, and platelets level were not included in the current study to assess their association with disease F I G U R E 3 Cumulative probabilities of progression to advanced liver disease by baseline hepatitis B virus DNA level (n = 404). After excluding nine individuals with HBV reactivation due to immunosuppressive therapy, progression to advanced liver disease was higher in patients with baseline HBV DNA level >2000 IU/mL than in patients with baseline HBV DNA level <2000 IU/mL (P = .018, logrank test). Cumulative probabilities of advanced liver disease were 0 of 53 (0.0%) vs 0 of 231 (0.0%), 2 of 31 (6.5%) vs 0 of 163 (0.0%), and 3 of 20 (15.0%) vs 1 of 106 (0.9%) at 5, 10, and 15 years follow-up, respectively. HBV, hepatitis B virus. † Patients were censored on the date of last outpatient clinic visit