Clinical Microbiology and Infection 23 (2017) 860e867 Contents lists available at ScienceDirect Clinical Microbiology and Infection journal homepage: www.clinicalmicrobiologyandinfection.com Original article Serum hepatitis B core-related antigen is more accurate than hepatitis B surface antigen to identify inactive carriers, regardless of hepatitis B virus genotype M. Riveiro-Barciela 1, 2, *, M. Bes 2, 3, F. Rodríguez-Frías 2, 4, D. Tabernero 2, 4, A. Ruiz 2, 4, lez 1, M. Homs 2, 4, L. Nieto 4, S. Sauleda 2, 3, R. Esteban 1, 2, R. Casillas 5, J. Vidal-Gonza 1, 2 M. Buti noma de Barcelona, Barcelona, Spain Liver Unit, Department of Internal Medicine, Hospital Universitari Vall d'Hebron and Universitat Auto ticas y Digestivas (CIBERehd), Instituto de Salud Carlos III, Madrid, Spain n Biom Centro de Investigacio edica en Red de Enfermedades Hepa 3) de la Salut, Barcelona, Spain Transfusion Safety Laboratory, Banc de Sang i Teixits, Servei Catala 4) noma de Liver Pathology Unit, Departments of Biochemistry and Microbiology (Virology Unit) Hospital Universitari Vall d'Hebron and Universitat Auto Barcelona, Barcelona, Spain 5) Vall d'Hebron Institut de Recerca, Barcelona, Spain 1) 2) a r t i c l e i n f o a b s t r a c t Article history: Received 12 September 2016 Received in revised form 24 February 2017 Accepted 5 March 2017 Available online 11 March 2017 Objectives: To investigate whether hepatitis B surface antigen (HBsAg) and hepatitis B core-related antigen (HBcrAg) levels are useful to identify inactive carriers among HBeAg-negative patients infected by different hepatitis B virus (HBV) genotypes. Methods: In all, 202 consecutive HBeAg-negative patients with chronic hepatitis B, 135 inactive carriers and 67 with HBV activity, were prospectively followed for 1 year. Results: In HBeAg-negative patients, HBsAg levels differed across the different genotypes (p <0.001). The highest levels were observed in genotypes F or H (4.2 ± 0.6 logIU/mL), followed by genotype E (3.4 ± 1.1 logIU/mL), genotype A (3.4 ± 0.8 logIU/mL), and the lowest in genotype D (2.7 ± 1.1 logIU/mL). Variations in HBsAg levels were similar in inactive carriers and patients with HBV activity. HBsAg <3 logIU/mL showed good performance for identifying genotype D inactive carriers: 76% of genotype D inactive carriers met this cut-off versus 31% for genotypes A, E, F or H. However, in patients with genotype A, HBsAg levels 3.7 logIU/mL better classified inactive carriers. The combination of a single measurement of HBcrAg 3 logU/mL plus HBV DNA 2000 IU/mL yielded a positive predictive value and diagnostic accuracy >85% in all HBV genotypes, except genotype H or F, with values of 62.5% and 72.7%, respectively, for the two parameters. Conclusions: HBsAg levels varied across genotypes in HBeAg-negative patients. HBsAg levels <3 logIU/mL were only useful for identifying genotype D inactive carriers. A single HBcrAg measurement 3 logU/mL plus HBV DNA 2000 IU/mL was highly accurate for identifying inactive carriers, regardless of their HBV genotype. M. Riveiro-Barciela, Clin Microbiol Infect 2017;23:860 © 2017 European Society of Clinical Microbiology and Infectious Diseases. Published by Elsevier Ltd. All rights reserved. Editor: G. Antonelli Keywords: HBcrAg HBeAg negative chronic hepatitis B HBsAg HBV inactive carrier Hepatitis B Hepatitis B virus genotype Introduction Hepatitis B e antigen (HBeAg) -negative chronic hepatitis B (CHB) infection is the most common form of hepatitis B virus (HBV) infection in many countries, including southern Europe [1,2]. One of * Corresponding author. M. Riveiro-Barciela, Hospital Universitari Vall d'Hebron, Passeig Vall Hebron, 119-129, 08035, Barcelona, Spain. E-mail address: [email protected] (M. Riveiro-arciela). the most challenging issues in CHB is to distinguish between the active HBeAg-negative phase of the infection and the inactive carrier state. The prognosis of the disease and the indication for therapy differ in these two clinical situations; hence, their differentiation is crucial. Therapy is not recommended in HBeAgnegative inactive carriers, and monitoring can be reduced because of their favourable long-term prognosis. In contrast, HBeAg-negative patients with HBV activity have a higher risk of http://dx.doi.org/10.1016/j.cmi.2017.03.003 1198-743X/© 2017 European Society of Clinical Microbiology and Infectious Diseases. Published by Elsevier Ltd. All rights reserved. M. Riveiro-Barciela et al. / Clinical Microbiology and Infection 23 (2017) 860e867 developing liver fibrosis and hepatocellular carcinoma, so regular monitoring and, in many cases, antiviral therapy are needed. To accurately identify inactive carrier status, the current guidelines recommend at least three alanine aminotransferase (ALT) and HBV DNA determinations during 1 year. Since Blumberg's discovery of hepatitis B surface antigen (HBsAg) in 1965, this parameter has been accepted as a laboratory marker to establish HBV infection [3]. HBsAg production is controlled, at least in part, by the amount of intrahepatic covalently closed circular DNA and the specific host immune response against the HBV envelope proteins. This is particularly evident in HBeAgpositive patients who show a high correlation between covalently closed circular DNA and HBsAg levels [4]. The amount of circulating HBsAg varies during the different phases of chronic infection: levels are higher in immunotolerant patients and they decrease after HBeAg seroconversion, and levels are lower in inactive carriers than in HBeAg-negative CHB patients [5e8]. For this reason, HBsAg determination has been cited as a potentially useful tool to identify inactive carriers [9]. Brunetto et al. [10] proposed a combination of HBsAg level <3 logIU/mL plus HBV DNA 2000 IU/mL to identify inactive carrier status at a single time-point in HBeAg-negative patients. The reported diagnostic accuracy was 93%, but the cohort was limited to genotype D-infected patients. Currently, ten different HBV genotypes have been described, each with a characteristic geographical distribution [11,12]: genotype A predominates in northern Europe and America, genotypes B and C in Asia, genotype D in southern Europe, genotype E in Africa, and genotypes H and F in South America [2,12]. Recently, hepatitis B core-related antigen (HBcrAg) has emerged as a new serological marker for CHB. As HBcrAg correlates well with intrahepatic covalently closed circular DNA levels in both naïve patients and those treated with nucleoside analogues [13,14], it may also be useful for proper classification of HBeAg-negative patients. The aim of this study was to investigate the performance of HBsAg and HBcrAg determinations to categorize HBeAg-negative patients across the various HBV genotypes. Patients and methods Patient selection and definitions Consecutive treatment-naive HBeAg-negative patients chronically infected with HBV (HBsAg-positive for more than 6 months) were prospectively selected in the outpatient clinics of a tertiary hospital. According to the guidelines of the European Association for the Study of the Liver [9], patients were classified as follows: inactive carriers, persistently normal ALT levels and HBV DNA 2000 IU/mL in three determinations over 1 year; active carriers, HBV DNA >20 000 IU/mL, or at least one fluctuation of HBV DNA >2000 IU/mL plus abnormal ALT levels (above the upper limit of normality). Patients were regarded as ‘intermediate’ on HBV DNA fluctuation between 2000 and 20 000 IU/mL regardless of ALT levels within the year of follow up [15]. For the purposes of this study, patients classified as either intermediate or active carriers were analysed together as patients with HBV activity, as all had HBV DNA >2000 IU/mL and therefore, a three-fold higher probability of disease progression [9,16]. Patients were excluded if they had undergone liver transplantation, were co-infected with hepatitis C virus, hepatitis D virus, or human immunodeficiency virus (HIV), had high alcohol intake, evidence of liver cirrhosis based on ultrasound findings (hepatic parenchyma nodules, spleen >12 cm, portal vein >16 mm), or analytical data (platelet count persistently below 140 10E9/ mL). This study was conducted in accordance with the Declaration of Helsinki guidelines and the principles of Good Practice, and 861 was approved by the Ethic Review Board of the Vall d'Hebron Hospital. Data acquisition Data on demographics (sex, age and race) were prospectively collected, and the clinical history was collected from the patients' medical records at the time of enrolment. Data on chemistry (ALT, aspartate aminotransferase (AST), g-glutamyl transferase, bilirubin, and albumin), haematology (platelet count, prothrombin time), and HBV serology and virology (serum HBsAg and HBcrAg, HBV DNA, HBV genotype) were also collected. Laboratory measurements HBsAg was quantified using the COBAS 80000 HBsAg II assay (Roche Diagnostics, Mannheim, Germany): lower limit of detection, 0.05 IU/mL. HBcrAg was quantified by an electrochemiluminescent assay: Lumipulse® G HBcrAg assay (Fujirebio, Fujirebio Europe, Gent, Belgium): lower limit of detection, 2 logU/ mL. This technique simultaneously determines denatured HBeAg, HBcAg and a 22-kDa precore protein (p22cr), as all three share an identical sequence of 149 amino acids. Hepatitis C virus, hepatitis D virus and HIV were detected by commercially available immunoassays. Serum HBV DNA was quantified by PCR with a COBAS 6800 HBV test (Roche Diagnostics): lower limit of quantification 20 IU/ mL, lower limit of detection 10 IU/mL. For HBV genotyping, HBV DNA was first enriched by ultracentrifugation of 9.6 mL of serum. Subsequently, Sanger sequencing was carried out after amplification of two different viral regions: PreC/Core (nucleotides 1774e2389, 615 bp) and PreS/Surface (nucleotides 2828e176, 561 bp), as previously reported [17]. Phylogenetic analysis was performed with HBV reference sequences using neighbour-joining analysis with the MEGA program, version 6. Genotypes H and F were combined due to their phylogenetic proximity and geographic distribution in the same areas [18,19]. Statistical analysis Normally distributed quantitative variables were compared with the Student's t test and expressed as the mean ± standard deviation. Variables with a non-normal distribution were analysed with the ManneWhitney U test and expressed as the median and interquartile range. Categorical variables were compared between groups using the chi-squared or Fisher exact test, as appropriate. Correlations were tested with the Spearman correlation test. The diagnostic performance of HBsAg and HBcrAg levels was evaluated by receiver operating characteristic curve analysis. The cut-off value to differentiate inactive carriers from patients with HBV activity was selected considering the highest Youden index. All statistical analyses were performed using IBM SPSS, 20 (SPSS Inc., Chicago, IL, USA). Results Baseline characteristics Of the 213 consecutive treatment-naive HBeAg-negative patients evaluated, 202 were included: 135 (66.8%) were classified as inactive carriers, 12 (6%) as active carriers, and 55 (27%) as intermediates (see Supplementary material, Fig. S1). Therefore, 67 individuals met the ‘HBV activity’ criteria. Eleven patients (5.2%) with HBV DNA <10 IU/mL were excluded because HBV genotype could not be determined at this level. 862 M. Riveiro-Barciela et al. / Clinical Microbiology and Infection 23 (2017) 860e867 Table 1 Baseline characteristics according to the stage of HBV infection Age, years, median (IQR) Male, n (%) Genotype, n (%) A B or C D E F or H Mixed A/E ALT, IU/mL, median (IQR) Platelets, 10E9/L, median (IQR) Albumin, mg/dL, median (IQR) HBsAg, logIU/mL, median (IQR) HBcrAg, logU/mL, median (IQR)a HBV DNA, logIU/mL, median (IQR) Total (n ¼ 202) Inactive carriers (n ¼ 135) HBV activity (n ¼ 67) p value 46 (36e56) 112 (55%) 48 (39e57) 81 (60%) 43 (32e54) 31(47%) 0.004 0.07 75 (37%) 5 (2%) 58 (29%) 36 (18%) 24 (12%) 4 (2%) 18 (13e28) 225 (188e251) 4.4 (4.2e4.6) 3.5 (2.6e4.1) 2.0 (2e2.5) 2.7 (2.1e3.3) 52 (38%) 1 (1%) 38 (28%) 26 (19%) 16 (12%) 2 (2%) 16 (11e24) 222 (185e248) 4 (4.2e4.6) 3.3 (2.4e4.0) 2.0 (2.0e2.0) 2.5 (1.9e2.9) 23 (34%) 4(6%) 20 (30%) 10 (15%) 8 (12%) 2 (3%) 23 (16e33) 227 (197e255) 4.4 (4.1e4.6) 3.7 (3.2e4.1) 2.0 (2.0e3.3) 3.5 (3.0e3.8) 0.98 <0.001 0.25 0.81 0.003 <0.001 <0.001 Abbreviations: ALT, alanine aminotransferase; HBcrAg, hepatitis B core-related antigen; HBsAg, hepatitis B surface antigen; HBV, hepatitis B virus; IQR, interquartile range. a Available in 134 patients. Serum HBcrAg levels according to the phase of chronic hepatitis B infection and HBV genotype Baseline characteristics are summarized in Table 1 and in the Supplementary material (Table S1). The predominant genotypes were A (37%), D (29%), E (18%), and F or H (12%). Patients with HBV activity were younger, and had higher ALT, HBV DNA, HBsAg and HBcrAg levels than inactive carriers (Table 1). Distribution of HBV genotypes was similar between inactive carriers and patients with HBV activity (p 0.98). Overall, mean HBV DNA levels were similar between the main genotypes (p 0.91). Univariate and multivariate analyses of baseline characteristics between inactive carriers and patients with HBV activity are summarized in the Supplementary material (Table S2). HBcrAg levels varied among the HBV genotypes in patients with HBV activity (p 0.02) (Table 2) and were similar in inactive carriers (p 0.052) (see Supplementary material, Table S3). In the comparison of HBcrAg levels between inactive carriers and patients with HBV activity according to genotype, differences were only observed in genotype A-infected patients (Fig. 1). Correlations between HBcrAg and HBV DNA were only observed in genotype A (r 0.43, p <0.001), although there was a trend to a correlation in genotype E patients (r 0.49, p 0.057) (Fig. 2b). All HBV genotype H and F patients had HBcrAg levels <2 logU/mL, the lower limit of detection, which precluded the study of correlations. There was a moderate correlation between HBsAg and HBcrAg levels in patients with genotypes A (r 0.55, p <0.001) and D (r 0.41, p 0.01), and no correlations in genotype E, F or H infection (Fig. 2b). Serum HBsAg levels according to the phase of chronic hepatitis B infection and HBV genotype Overall, HBsAg levels were lower in inactive carriers than in patients with HBV activity (p 0.003). HBsAg levels differed significantly (p <0.001) across the HBV genotypes, being highest for genotype F or H, followed by E, A, and the lowest, D (Table 2). When HBsAg levels were compared between these two groups according to HBV genotype, significant differences were only observed for genotypes A and D (p 0.002 and p <0.001, respectively) (Fig. 1). A positive correlation between HBsAg and HBV DNA was found in patients with HBV genotype A (r 0.46, p <0.001) and D (r 0.56, p <0.001) infection, but there were no correlations in genotypes E, or F or H (Fig. 2a). HBsAg and HBcrAg levels to identify HBV inactive carriers Overall, the area under the receiver operating characteristic curve (AUC) for identification of inactive carrier status was 0.67 (95% CI 0.56e0.77; p 0.003) for HBcrAg and 0.63 (95% CI 0.55e0.71; p 0.001) for HBsAg. The cut-off associated with the highest Youden index was 3.21 logIU/mL for HBsAg (sensitivity 49.6%; specificity 77.6%) and 3 logU/mL for HBcrAg (sensitivity 97.84%; specificity Table 2 Comparison of HBsAg and HBcrAg levels according to HBV genotype and disease activity in HBV infection Total Inactive carriers HBV activity n HBsAg, logIU/mL, median (IQR) HBcrAg, logU/mL, median (IQR) n (%) HBsAg, logIU/mL, median (IQR)a HBcrAg, logU/mL, median (IQR)b n (%) HBsAg, logIU/mL, median (IQR)c HBcrAg, logU/mL, median (IQR)d Total GT A GT D GT E GT F or H p value 202 3.5 (2.6e4.1) 2.0 (2.0e2.5) 135 (67%) 3.3 (2.4e4.0) 2.0 (2.0e2.0) 67 (33%) 3.7 (3.2e4.1) 2.0 (2.0e3.3) 75 3.5 (3.1e4.0) 2.0 (2.0e2.7) 52 (39%) 2.7 (3.3e3.8) 2.0 (2.0e2.2) 23 (34%) 3.8 (3.5e4.2) 2.9 (2.0e3.4) 58 2.7 (2.1e3.5) 2.0 (2.0e2.0) 38 (28%) 2.4 (1.8e2.9) 2.0 (2.0e2.0) 20 (30%) 3.4 (2.9e3.7) 2.0 (2.0e2.6) 36 3.7 (3.0e4.2) 2.0 (2.0e2.5) 26 (19%) 3.7 (2.9e4.2) 2.0 (2.0e2.3) 10 (15%) 3.8 (3.2e4.2) 2.0 (2.3e5.8) 24 4.3 (4.1e4.4) 2.0 (2.0e2.0) 16 (12%) 4.3 (3.9e4.5) 2.0 (2.0e2.0) 8 (12%) 4.3 (4.2e4.4) 2.0 (2.0e2.0) d <0.001 0.052 d <0.001 0.63 d 0.003 0.02 Abbreviations: GT, genotype; HBcrAg, hepatitis B core-related antigen; HBsAg, hepatitis B surface antigen; HBV, hepatitis B virus; IQR, interquartile range. a HBsAg in inactive carriers: p <0.001 for genotype A versus D; p 0.57 for genotype A versus E; p <0.001 for genotype A versus F or H; p <0.001 for genotype D versus E; p <0.001 for genotype D versus F or H; p 0.013 for genotype E versus F or H. b HBcrAg in inactive carriers: p 0.55 for genotype A versus D; p 0.75 for genotype A versus E; p 0.24 for genotype A versus F or H; p 0.52 for genotype D versus E; p 0.48 for genotype D versus F or H; p 0.14 for genotype E versus F or H. c HBsAg in patients with HBV activity: p 0.02 for genotype A versus D; p 0.87 for genotype A versus E; p 0.06 for genotype A versus F or H; p 0.09 for genotype D versus E; p <0.001 for genotype D versus F or H; p 0.07 for genotype E versus F or H. d HBcrAg in HBV activity: p 0.03 for genotype A versus D; p 0.94 for genotype A versus E; p 0.01 for genotype A versus F or H; p 0.34 for genotype D versus E; p 0.21 for genotype D versus F or H; p 0.13 for genotype E versus F or H. M. Riveiro-Barciela et al. / Clinical Microbiology and Infection 23 (2017) 860e867 863 Fig. 1. Comparison of mean hepatitis B surface antigen (HBsAg) and hepatitis B core-related antigen (HBcrAg) levels between inactive carriers and patients with hepatitis B virus (HBV) activity showed statistical differences in patients infected by genotype A (HBsAg and HBcrAg) and D (HBsAg), but not in those with genotypes E, F or H infection. 27.3%). In genotype A, the HBcrAg AUC was higher than that of HBsAg (AUC 0.80, 95% CI 0.65e0.95; p <0.001, and AUC 0.73, 95% CI 0.60e0.85; p 0.002, respectively). In genotype D, only HBsAg determination had sufficient discriminatory power to identify inactive carriers (AUC 0.78, 95% CI 0.66e0.90; p <0.001). Higher HBsAg levels in genotype A than genotype D-infected patients at baseline resulted in different cut-offs to identify inactive carriers. HBsAg <3 logIU/mL showed a diagnostic accuracy of 74.1% for genotype D, but only 49.3% for genotype A (Table 3). The cut-off showing the best performance in genotype A was HBsAg 3.7 logIU/mL (positive predictive value 82.2%; negative predictive value 50%; diagnostic accuracy 69.3%). However, the previous cutoffs were not useful for proper classification of inactive carriers infected by other HBV genotypes (genotype E, AUC 0.57; p 0.48 and genotype F or H, AUC 0.49; p 0.95). The diagnostic accuracy of HBcrAg determination was higher than that of HBsAg for identifying inactive carriers. A single HBV DNA determination 2000 IU/mL together with HBcrAg 3 logU/ mL yielded a positive predictive value and diagnostic accuracy >85% for all genotypes except H or F, which showed values of 63% and 73%, for these two indexes, respectively (Table 3). Inclusion of 11 patients with undetectable HBV DNA in the cohort did not change the diagnostic accuracy or predictive values of the abovementioned cut-offs (see Supplementary material, Table S4). HBsAg determination with or without HBV DNA values did not achieve higher diagnostic accuracy than the combined cut-offs of HBcrAg 3 logU/mL plus HBV DNA 2000 IU/mL. Variability of HBsAg levels In total, 103 (51%) and 81 (40%) patients had consecutive quantitative HBsAg determinations separated by 6 and 12 months, respectively. The comparisons showed no differences in HBsAg levels over this time period (month 6: mean difference 0.036, 95% CI e0.035 to 0.11; p 0.32; month 12: mean difference 0.069, 95% CI e0.024 to 0.16; p 0.15). In the comparison of consecutive HBsAg levels in inactive carriers, 48 patients had at least two determinations separated by 12 months, and a significant decay in HBsAg levels was observed in this group (mean difference 0.16, 95% CI 0.078e0.23; p <0.001). However, in the 33 patients with HBV activity and an HBsAg follow up of at least 1 year, no changes were observed (p 0.67). Discussion In this prospective study of well-classified HBeAg-negative patients infected by HBV genotypes A and D to F, we found considerable variability in HBsAg levels across the genotypes, with statistically higher levels in patients with genotype F or H, followed by E, A and D. There was a high correlation between HBsAg and HBV DNA levels in genotype A and D patients, which made HBsAg quantification a good biomarker for proper identification of inactive carriers infected by either of these genotypes. Our results confirmed that the previously described HBsAg cut-off of <3 logIU/ mL [10] is particularly useful to identify genotype D inactive carriers, with a positive predictive value of 76% and diagnostic accuracy of 74%. These results stem from the low HBsAg levels in patients with genotype D, the genotype associated with the lowest levels (mean 2.7 logIU/mL). In the remaining HBV genotypes, the percentage of inactive carriers who met this criterion was 31%, a finding that indicates the need for HBV genotyping to use this HBsAg cut-off value. For example, in genotype A patients, an HBsAg value of 3.7 logIU/mL yielded the best performance for identifying inactive carriers: positive predictive value 82% and diagnostic accuracy 69%. HBcrAg is a recently described serum marker formed by three products from the preCore/Core gene that have an identical sequence of 149 amino acids. As HBcrAg level decreases across the different phases of chronic hepatitis B, it has been suggested as a potential tool for identifying patients at risk of developing hepatocellular carcinoma or HBV reactivation after discontinuing nucleoside analogues [20,21]. Data are scarce on the possible role of HBcrAg levels to help classify the chronic hepatitis B infection stage. In our cohort, HBcrAg levels were associated with excellent predictive values for identifying HBV inactive carriers, in particular when a single measurement of this biomarker was combined with HBV DNA 2000 IU/mL. HBcrAg levels achieved higher diagnostic accuracy for classifying inactive carriers across all the HBV 864 M. Riveiro-Barciela et al. / Clinical Microbiology and Infection 23 (2017) 860e867 Fig. 2. Correlations between hepatitis B virus (HBV) DNA and both hepatitis B surface antigen (HBsAg) (a) and hepatitis B core-related antigen (HBcrAg) (b) reached statistical significance. Independent analysis by each genotype showed that correlations were only present in patients infected by genotype A (HBsAg and HBcrAg) and D (only HBsAg). Moderate correlations between HBsAg and HBcrAg levels (c) were found in genotype A and D patients. genotypes than HBsAg quantification: a single HBV DNA determination 2000 IU/mL plus HBcrAg levels of 3 logU/mL showed an overall positive predictive value and diagnostic accuracy of >85%. There is extensive information concerning HBsAg levels in HBV genotypes B and C [5,7,22e24]. However, data are scarce for genotypes E to H, the most common genotypes in developing countries, where HBV is more prevalent [18,19,25,26]. As HBV DNA monitoring is not currently available worldwide [27], it would be useful to have a serological marker that can identify inactive carriers in a single determination with less sophisticated technology and easier manipulation and preservation of serum samples [9]. HBsAg quantification could be a good candidate, but the significant M. Riveiro-Barciela et al. / Clinical Microbiology and Infection 23 (2017) 860e867 865 Fig. 2. (continued). variation in HBsAg levels across the HBV genotypes indicates that genotyping would be needed for proper classification of inactive carriers. High HBsAg levels were also reported in a cohort of 128 HBeAg-negative HIVeHBV co-infected patients with HBV genotype E (mean 3.82 logIU/mL) [28] and in a European cohort of HBV- infected patients, where higher HBsAg levels were reported in genotype A than in genotype D (p 0.007) [6]. However, this difference was not seen in a small number of inactive carriers (8 A and 18 D). Similar results have been described in cell cultures. Sugiyama et al. showed that extracellular expression of viral DNA and surface Table 3 Diagnostic accuracy and predictive values of HBsAg and HBcrAg cut-offs for identifying inactive carriers status according to HBV genotype Population Total GT A GT D GT E GT F or H Total n ¼ 202 55/69 (80%) 53/133 (40%) 108/202 (53%) GT A n ¼ 75 16/18 (89%) 21/57 (37%) 37/75 (49%) GT D n ¼ 58 29/35 (83%) 14/23 (61%) 43/58 (74%) GT E n ¼ 36 8/10 (80%) 8/26 (31%) 16/36 (44%) GT F or H n ¼ 24 1/1 (100%) 8/23 (35%) 9/24 (38%) Total n ¼ 202 PPV 67/82 (82%) NPV 52/120 (43%) Diagnostic accuracy 119/202 (59%) HBsAg 3.21 logIU/mL + HBV DNA 2000 IU/mL Population Total n ¼ 202 PPV 67/75 (89%) NPV 59/127 (46%) Diagnostic accuracy 126/202 (62%) HBcrAg 3 logU/mL Population Total n ¼ 134 PPV 88/120 (73%) NPV 12/14 (86%) Diagnostic accuracy 100/134 (75%) HBcrAg 3 logU/mL + HBV DNA 2000 IU/mL Population Total n ¼ 134 PPV 88/103 (86%) NPV 29/31 (91%) Diagnostic accuracy 117/134 (87%) GT A n ¼ 75 24/26 (92%) 21/49 (43%) 45/75 (60%) GT D n ¼ 58 31/38 (82%) 13/20 (65%) 44/58 (76%) GT E n ¼ 36 10/12 (83%) 8/24 (33%) 18/36 (50%) GT F or H n ¼ 24 1/1 (100%) 8/23 (35%) 9/24 (38%) GT A n ¼ 75 24/26 (92%) 21/49 (43%) 45/75 (60%) GT D n ¼ 58 31/33 (94%) 18/25 (67%) 49/58 (84%) GT E n ¼ 36 10/11 (91%) 9/25 (36%) 19/36 (53%) GT F or H n ¼ 24 1/1 (100%) 8/23 (35%) 9/24 (38%) GT A n ¼ 65 47/57 (82%) 7/8 (88%) 54/65 (83%) GT D n ¼ 38 24/35 (69%) 2/3 (67%) 26/38 (68%) GT E n ¼ 16 11/14 (79%) 2/2 (100%) 13/16 (81%) GT F or H n ¼ 11 5/11 (46%) 0/0 (0%) 5/11 (46%) GT A n ¼ 65 47/52 (90%) 12/13 (92%) 59/65 (91%) GT D n ¼ 38 24/28 (86%) 9/10 (90%) 33/38 (87%) GT E n ¼ 16 11/12 (92%) 4/4 (100%) 15/16 (94%) GT F or H n ¼ 11 5/8 (63%) 3/3 (100%) 8/11 (73%) HBsAg <3 logIU/mL Population PPV NPV Diagnostic accuracy HBsAg 3.21 logIU/mL Population Abbreviations: GT, genotype; HBcrAg, hepatitis B core-related antigen; HBsAg, hepatitis B surface antigen; HBV, hepatitis B virus; NPV, negative predictive value; PPV, positive predictive value. 866 M. Riveiro-Barciela et al. / Clinical Microbiology and Infection 23 (2017) 860e867 antigens differs between genotypes, with higher HBsAg levels in genotype A than genotype D [29]. Moreover, HBsAg levels have a tendency to decay over a 1-year period in inactive carriers, whereas they remain stable in patients with HBV activity. In a retrospective multicentre study including 292 HBeAg-negative patients, the mean annual HBsAg decrease was more prominent in inactive carriers than in those with HBV activity [15]. A correlation between HBcrAg levels and HBV DNA has been reported in HBeAg-negative genotype A and D patients [30], but there are no available data for genotypes E, F or H. In the present study, the correlations between HBV DNA and both HBsAg and HBcrAg are lower than previously reported [30,31]. This may be because only HBeAg-negative patients were enrolled on our study, and most were inactive carriers with HBV DNA persistently 2000 IU/mL. Our study has several limitations, one of them being the heterogeneous distribution of HBV genotypes. This occurred because patients were unselected and prospectively included, and that implied a higher percentage of the more prevalent HBV genotypes in our area (A and D). Another limitation is that HBcrAg could not be determined in a small number of patients. Therefore, the use of HBcrAg for proper classification of inactive carriers regardless of HBV genotype should be validated in a larger cohort. One of the strengths of our study is inclusion of patients with less extensively investigated HBV genotypes, such as E, F and H, in whom data in this regard are scarce. The results of this study could have implications for future research. The variability of HBsAg levels among HBeAg-negative patients with different genotypes should be taken into account when this serum marker is used for clinical purposes. In conclusion, in a large cohort of well-categorized HBeAgnegative patients with chronic HBV infection, HBsAg levels varied among inactive carriers according to genotype, with higher levels in genotype F or H. In contrast, HBcrAg levels were similar regardless of the genotype, and when combined with HBV DNA determination, HBcrAg enabled identification of inactive carrier status without HBV genotyping. Funding This study was funded in part by the Instituto de Salud Carlos III (grant number PI12/01893) and by Gilead (grant number GLD13/ 00137). Transparency declaration MB and RE have served as advisors to Gilead, Bristol-Myers Squibb and Novartis. MB and MRB have received research grants from Gilead. The other authors have no personal interests to declare. Acknowledgements HBcrAg reagents were provided free of charge by Fujirebio Europe. The funder had no role in the study design, data collection, analysis or preparation of the manuscript. English writing support was provided by Celine Cavallo. Contribution to authorship MB and FRF act as guarantors of this article. MB, RE and MRB designed the study. MRB, MB and SS selected the patients. MH and DT performed quantification of HBsAg and HBV DNA. RC, MB and LN carried out HBV DNA amplification and genotyping. MRB, MB and JV collected the clinical and laboratory data. MH, DT, LN and FRF analysed and interpreted the data. MRB, MB and FRF drafted the manuscript. MB, RE and FRF reviewed the manuscript. All authors approved the final version of the article. Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.cmi.2017.03.003. References [1] Liu CJ, Kao JH. Global perspective on the natural history of chronic hepatitis B: Role of hepatitis B virus genotypes A to J. Semin Liver Dis 2013;33:97e102. [2] McMahon BJ. Chronic hepatitis B virus infection. Med Clin North Am 2014;98: 39e54. [3] Blumberg BS, Sutnick AI, London WT. Hepatitis and leukemia: their relation to Australia antigen. Bull N Y Acad Med 1968;44:1566e86. [4] Thompson AJ, Nguyen T, Iser D, Ayres A, Jackson K, Littlejohn M, et al. Serum hepatitis B surface antigen and hepatitis B e antigen titers: disease phase influences correlation with viral load and intrahepatic hepatitis B virus markers. Hepatology (Balt, MD) 2010;51:1933e44. [5] Nguyen T, Thompson AJ, Bowden S, Croagh C, Bell S, Desmond PV, et al. Hepatitis B surface antigen levels during the natural history of chronic hepatitis B: a perspective on Asia. J Hepatol 2010;52:508e13. [6] Jaroszewicz J, Calle Serrano B, Wursthorn K, Deterding K, Schlue J, Raupach R, et al. Hepatitis B surface antigen (HBsAg) levels in the natural history of hepatitis B virus (HBV) infection: a European perspective. J Hepatol 2010;52: 514e22. [7] Chan HL, Wong VW, Wong GL, Tse CH, Chan HY, Sung JJ. A longitudinal study on the natural history of serum hepatitis B surface antigen changes in chronic hepatitis B. Hepatology (Balt. MD) 2010;52:1232e41. [8] Kim YJ, Cho HC, Choi MS, Lee JH, Koh KC, Yoo BC, et al. The change of the quantitative HBsAg level during the natural course of chronic hepatitis B. Liver Int 2011;31:817e23. [9] Easl clinical practice guidelines: Management of chronic hepatitis B virus infection. J Hepatol 2012;57:167e85. [10] Brunetto MR, Oliveri F, Colombatto P, Moriconi F, Ciccorossi P, Coco B, et al. Hepatitis B surface antigen serum levels help to distinguish active from inactive hepatitis B virus genotype D carriers. Gastroenterology 2010;139: 483e90. [11] Kao JH. Hepatitis B viral genotypes: clinical relevance and molecular characteristics. J Gastroenterol Hepatol 2002;17:643e50. [12] Kramvis A. Genotypes and genetic variability of hepatitis B virus. Intervirology 2014;57:141e50. [13] Wong DK, Tanaka Y, Lai CL, Mizokami M, Fung J, Yuen MF. Hepatitis B virus core-related antigens as markers for monitoring chronic hepatitis B infection. J Clin Microbiol 2007;45:3942e7. [14] Suzuki F, Miyakoshi H, Kobayashi M, Kumada H. Correlation between serum hepatitis B virus core-related antigen and intrahepatic covalently closed circular DNA in chronic hepatitis B patients. J Med Virol 2009;81:27e33. [15] Brouwer WP, Lik-Yuen Chan H, Brunetto MR, Martinot-Peignoux M, Arends P, Cornberg M, et al. Repeated measurements of hepatitis B surface antigen identify carriers of inactive HBV during long-term follow-up. Clin Gastroenterol Hepatol 2016;14:1481e9. e5. [16] Iloeje UH, Yang HI, Su J, Jen CL, You SL, Chen CJ. Predicting cirrhosis risk based on the level of circulating hepatitis b viral load. Gastroenterology 2006;130: 678e86. [17] Bes M, Vargas V, Piron M, Casamitjana N, Esteban JI, Vilanova N, et al. T cell responses and viral variability in blood donation candidates with occult hepatitis b infection. J Hepatol 2012;56:765e74. [18] Norder H, Courouce AM, Magnius LO. Complete genomes, phylogenetic relatedness, and structural proteins of six strains of the hepatitis B virus, four of which represent two new genotypes. Virology 1994;198:489e503. [19] Arauz-Ruiz P, Norder H, Robertson BH, Magnius LO, Genotype H. A new Amerindian genotype of hepatitis B virus revealed in Central America. J Gen Virol 2002;83:2059e73. [20] Matsumoto A, Tanaka E, Minami M, Okanoue T, Yatsuhashi H, Nagaoka S, et al. Low serum level of hepatitis B core-related antigen indicates unlikely reactivation of hepatitis after cessation of lamivudine therapy. Hepatol Res 2007;37:661e6. [21] Tada T, Kumada T, Toyoda H, Kiriyama S, Tanikawa M, Hisanaga Y, et al. Hbcrag predicts hepatocellular carcinoma development: an analysis using time-dependent receiver operating characteristics. J Hepatol 2016;65:48e56. [22] Seto WK, Wong DK, Fung J, Hung IF, Fong DY, Yuen JC, et al. A large caseecontrol study on the predictability of hepatitis B surface antigen levels three years before hepatitis B surface antigen seroclearance. Hepatology (Balt. MD) 2012;56:812e9. [23] Tseng TC, Liu CJ, Chen CL, Yang WT, Yang HC, Su TH, et al. Higher lifetime chance of spontaneous surface antigen loss in hepatitis B carriers with genotype C infection. Aliment Pharmacol Ther 2015;41:949e60. M. Riveiro-Barciela et al. / Clinical Microbiology and Infection 23 (2017) 860e867 [24] Tseng TC, Liu CJ, Su TH, Wang CC, Chen CL, Chen PJ, et al. Serum hepatitis B surface antigen levels predict surface antigen loss in hepatitis B e antigen seroconverters. Gastroenterology 2011;141. 517e25, 525.e511e12. [25] Kidd-Ljunggren K, Miyakawa Y, Kidd AH. Genetic variability in hepatitis B viruses. J Gen Virol 2002;83:1267e80. [26] Okamoto H, Tsuda F, Sakugawa H, Sastrosoewignjo RI, Imai M, Miyakawa Y, et al. Typing hepatitis B virus by homology in nucleotide sequence: Comparison of surface antigen subtypes. J Gen Virol 1988;69:2575e83. [27] WHO. WHO guidelines approved by the guidelines review committee. Guidelines for the prevention, care and treatment of persons with chronic hepatitis B infection. Geneva: WHO; 2015. [28] Boyd A, Maylin S, Moh R, Mahjoub N, Gabillard D, Eholie SP, et al. Hepatitis B surface antigen quantification as a predictor of seroclearance during 867 treatment in HIV-hepatitis B virus co-infected patients from sub-Saharan Africa. J Gastroenterol Hepatol 2016;31:634e44. [29] Sugiyama M, Tanaka Y, Kato T, Orito E, Ito K, Acharya SK, et al. Influence of hepatitis B virus genotypes on the intra- and extracellular expression of viral DNA and antigens. Hepatology (Balt. MD) 2006;44:915e24. [30] Maasoumy B, Wiegand SB, Jaroszewicz J, Bremer B, Lehmann P, Deterding K, et al. Hepatitis B core-related antigen (HBcrAg) levels in the natural history of hepatitis B virus infection in a large European cohort predominantly infected with genotypes A and D. Clin Microbiol Infect 2015;21:606. e601e610. [31] Seto WK, Wong DK, Fung J, Huang FY, Liu KS, Lai CL, et al. Linearized hepatitis B surface antigen and hepatitis B core-related antigen in the natural history of chronic hepatitis B. Clin Microbiol Infect 2014;20:1173e80.