Introduction

After natural hepatitis B virus (HBV) infection in people living with HIV (PLWH), an indeterminate serostatus, such as “isolated anti-HBc (IAHBc)”, representing negative hepatitis B surface antibody (anti-HBs) and positive hepatitis B core antibody (anti-HBc) coupled with negative hepatitis B surface antigen (HBsAg), is more often observed in such individuals than in non-PLWH [1–3]. While IAHBc is clinically interpreted as either 1) resolved infection, 2) occult infection, 3) false positive, or 4) a mutant hepatitis B surface antigen (HBsAg) strain [4,5], the prevailing consensus considers IAHBc mainly reflects resolved infection, based on stable serologic values associated with absence of viral rebound and re-infection [1,2,6]. Hence, the major guidelines assume that the majority of IAHBc cases reflect robust HBV immunity where “anti-HBs levels had waned” years or decades after stable seroconversion [4,7,8]. However, some studies reported strong association between IAHBc and occult hepatitis B infection (OBI) [9–12] and hepatocellular carcinoma [13]. To our knowledge, there are no long-term IAHBc follow-up studies with repeat serology that have demonstrated a decline/disappearance of anti-HBs over years or decades.

“Anti-HBs alone” with positive anti-HBs and negative anti-HBc is interpreted as “immune from receipt of prior vaccination (if documented complete vaccine series), and “negative anti-HBs/anti-HBc” with negative anti-HBs and negative anti-HBc is interpreted as “never infected (if no documentation of vaccine series completion)” [4]. However, several studies have reported that both are not rare after HBV infection in unvaccinated patients [1–3], but to our knowledge, the clinical interpretation of these two patterns has never been discussed extensively.

There is a growing need for reassessment of HBV immunity in PLWH. Evidence suggests that the rate of new HBV infection among men-who-have-sex-with-men (MSM) is high [14], and that the efficacy of HBV vaccines in PLWH is low [15–17]. Even in those PLWH who had been vaccinated during infancy, HBV infection can occur during adulthood [18]. Furthermore, HBV reactivation has been observed in individuals with past infection under various settings, such as administration of anti-hepatitis C therapy and immunosuppressive therapy [19,20]. Today, PLWH are protected against HBV infection or reactivation because the majority currently use antiretroviral therapy that includes the combination of tenofovir (TFV) [tenofovir disoproxil fumarate (TDF) or tenofovir alafenamide (TAF)] plus [lamivudine (3TC) or emitricitabine (FTC)], both of which have potent anti-HBV activity. However, the recent use of two-drug antiretroviral therapies, such as those that exclude TFV (dolutegravir/3TC) [21] or completely omit anti-HBV agents (dolutegravir/rilpivirine and cabotegravir/ rilpivirine) [22,23], raises concern about the potential for new HBV infection [23,24] or reactivation [25]. Although current guidelines recommend the use of TFV plus FTC/3TC in chronic or occult HBV infection, there are no recommendations for PLWH with indeterminate serostatus [7].

The major aims of the present study were first to clarify whether IAHBc is the result of waning and disappearance of anti-HBs, and second, to provide clinical interpretation to “anti-HBs alone” and “lost anti-HBs/anti-HBc” as well as IAHBc, based on long-term follow-up of serostatus following HBV infection.

Methods

Study participants and study protocol

Fig 1 shows the study flow and subject classification. First, in this single-center and retrospective cohort study, we included only PLWH with past natural HBV infection and negative history for HBV vaccination based on comprehensive review of the medical record of each individual. Past natural HBV infection was defined as any record of a positive test for HBsAg, HBeAg, anti-HBe, anti-HBc or HBV-DNA. Second, based on the most recent serology status during the period between January 2021 and September 2023, the selected 878 individuals were classified into either 1) seroconversion (n = 640), 2) chronic infection (n = 60), or 3) indeterminate (n = 178) groups. The seroconversion group represented PLWH who were HBsAg-negative, anti-HBs-positive, and anti-HBc-positive. The chronic infection group represented PLWH who were HBsAg-positive, anti-HBs-negative, and anti-HBc-positive. All other PLWH were categorized as indeterminate serostatus, and they included those with inconsistent test results of several tests or a single test results inconsistent with seroconversion and chronic infection.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Flow diagram of subject enrolment and classification.

Serostatus screening conducted in 878 people living with HIV (PLWH) and past hepatitis B virus (HBV) infection allowed their classification into either seroconversion, chronic infection, or indeterminate pattern. Serial changes in serological tests occurring during the long observation period were reviewed and analyzed in PLWH with indeterminate serostatus.

https://doi.org/10.1371/journal.pone.0336924.g001

The study was approved by the Ethics Review Committee of Tokyo Medical University (T2023-0184) and conducted according to the principles of the Declaration of Helsinki. All data were accessed in fully anonymized style on January 29, 2024, and any information that identifies individuals was not accessible. According to the Japanese ethical guidelines for medical and biological research involving human subjects, the requirement for consent was waived due to the retrospective nature of the study.

Definitions of sub-serostatuses

The sub-serostatus of the group of PLWH with an indeterminate serostatus was determined by reviewing the titers of HBsAg, HBsAb, HBeAg, HBeAb, HBcAb, HBV-DNA and HBcrAg measured between October 2003 and September 2023. Generally, serostatus, such as IAHBc, is defined by each single set of anti-HBV antibodies; however, this definition could not be applied in our study to determine the long-term serostatus since the serostatus changed during the period, with unstable titers of anti-HBV antibody in PLWH with indeterminate serostatus. For this reason, we applied new definitions for the long-term sub-serostatus, which were based on chronological changes in serology (Table 1). Thus, PLWH with one or more tests negative for anti-HBs and positive for anti-HBc were defined as “IAHBc”, irrespective of the results of other serologic tests. Similarly, “anti-HBs alone” was defined as any history of serostatus with positive anti-HBs and negative anti-HBc, whereas “lost anti-HBs/anti-HBc” represented any history of disappearance of both anti-HBs and anti-HBc.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Definition of sub-serostatus classification.

https://doi.org/10.1371/journal.pone.0336924.t001

Clinical interpretations of serological changes

Based on the assessment of serological changes during the study period, we assigned clinical interpretation to PLWH with indeterminate serostatus (Table 2). First, the majority of PLWH without primary infection data (positive HBsAg or HBV-DNA) were classified as “unknown” since other possible interpretations could be assigned, while definite clinical interpretations were assigned to those PLWH with primary infection data. Second, we classified the HBV immunity robustness into “weak”, “robust”, and “resolved infection”, based on the chronological stability of anti-HBs or the appearance of HBsAg/HBV-DNA (Table 2). Weak immunity was defined as unstable anti-HBs after primary infection or positive HBsAg/HBV-DNA, whereas robust immunity was defined as persistently positive anti-HBs after initial seroconversion. Third, we further assigned detailed interpretations to address highly variable serologic patterns. For example, when PLWH had no record of positive anti-HBs after primary infection, they were labeled as “non-seroconversion”. When PLWH showed repeated appearance/disappearance of anti-HBs, they were labeled as “weak/unstable seroconversion”.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Simplified classification of HBV serologic profiles and inferred immune status.

https://doi.org/10.1371/journal.pone.0336924.t002

Furthermore, for the “unknown” group, the adopted “possible” clinical interpretations were based on definitions. PLWH with primary infection data of the “Definite resolution” group represented those who showed a gradual fall in anti-HBs level after seroconversion, while those without primary infection data of the “possible resolution” group showed gradual fall of anti-HBs.

Results

Clinical characteristics of study participants

Screening of 878 cases with past HBV infection without history of vaccination showed seroconversion (positive anti-HBs and anti-HBc) in 640 (73%) patients, chronic infection (positive HBsAg, negative anti-HBs and positive anti-HBc) in 60 (7%), and indeterminate patterns in 178 (20%). Comparison of the three serostatus groups demonstrated that PLWH of the indeterminate group were older, while those of the seroconversion group had lower CD4 cell count. Anti-HBV agents were more often used by the chronic infection and indeterminate groups than the seroconversion group (Table 3).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Characteristics of participants at screening.

https://doi.org/10.1371/journal.pone.0336924.t003

Clinical characteristics of indeterminate group by sub-serostatus

Analysis of the serological tests showed a median [interquartile range (IQR)] observation period of 13.0 (9.3–15.5) years, during which various serologic tests were performed (average: 9.0, range: 7–13) times per person, with inter-serologic test interval of 15.5 (range: 9.5–20.5) years. Of the 178 PLWH of the indeterminate group, 118 (66%) were classified as IAHBc, 35 (20%) as anti-HBs alone, and 25 (14%) were lost anti-HBs/anti-HBc. The type of sub-serostatus was independent of age, sex, observation period, test numbers, and test intervals (Table 4).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 4. Characteristics of patients with indeterminate serostatus against hepatitis B virus.

https://doi.org/10.1371/journal.pone.0336924.t004

Clinical interpretation of indeterminate serostatus

Weak HBV immunity was observed in 47/118 (39.8%) of IAHBc and in 11/25 (44.0%) of lost anti-HBs/anti-HBc, whereas it was observed only in 3/35 (8.6%) of anti-HBs alone. Specifically, chronic infection, viral rebound, non-seroconversion and weak/unstable seroconversion were observed in IAHBc and lost anti-HBs/anti-HBc, but not among anti-HBs alone (Table 5). In contrast, all PLWH with sustained seroconversion were positive for anti-HBs alone, but not IAHBc or lost anti-HBs/anti-HBc. Surprisingly, despite the current consensus interpretation, no PLWH showed a resolved infection pattern in any of the serostatus groups (Table 5).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 5. Clinical interpretations of indeterminate serostatus based on long-term serological changes.

https://doi.org/10.1371/journal.pone.0336924.t005

Logistic multivariate analysis confirmed the association between sub-serostatus and HBV immunity. Even in adjusted models by the history of TFV-based treatment, age, CD4 cell counts, and HIV suppression at screening point, IAHBc was a significant predictor of weak HBV immunity [odds ratio (OR) 2.170, 95% confidence interval (95%CI) (1.058–4.452)], whereas anti-HBs alone inferred robust HBV immunity [OR: 0.143, 95%CI: 0.042–0.494] (Table 5).

Chronological changes in serostatus

Tentative seroconversion was observed in 114/178 (64%) of the entire indeterminate group, 73/118 (61.9%) of IAHBc, 33/35 (94.3%) of anti-HBs alone, and 8/25 (32.0%) of lost anti-HBs/anti-HBc. Only 39/178 (26%) showed stable serostatus (no change in serology) throughout the observation period, while the count of serostatus changes was > 4 in 14 (8%) PLWH (Table 6).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 6. Count of serological changes by indeterminate sub-serostatus and clinical interpretations.

https://doi.org/10.1371/journal.pone.0336924.t006

The count of serostatus changes did not correlate with clinical interpretation. For example, in subjects who showed no seroconversion pattern, anti-HBs remained negative without any change during the observation period, indicating weak immunity. On the other hand, some cases of late seroconversion showed frequent changes in serology until stable seroconversion, indicating robust immunity (Table 6).

Case presentations of indeterminate serostatus group

Fig 2 shows serial changes in various serological tests (CD4, HIV-RNA, HBsAg, HBsAb, HBV-DNA, HBcrAg, and anti-HBV Ab titer), during the indicated long observation periods in representative cases of various serostatuses who received various therapeutic regimens and considered to have different clinical interpretations. The case shown in Fig 2A, who was labeled as Lost anti-HBs/anti-HBc serostatus, never showed positive anti-HBs after primary infection and exhibited waning anti-HBc. The case was clinically interpreted as “non-seroconversion”. On the other hand, the case depicted in Fig 2F who was labeled as IAHBc showed repeated viral rebound due to interruption of anti-HBV agents.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Serologic changes in 8 representative cases of sub-serostatus and clinical interpretation groups.

(A) Lost anti-HBs/anti-HBc case interpreted as “non-seroconversion”. Anti-HBs was never detected since primary infection. Sub-serostatus was initially IAHBc during primary infection, and then changed to “Lost anti-HBs/anti-HBc” following disappearance of anti-HBc. (B) An IAHBc case interpreted as “unstable seroconversion”. Note the frequent disappearance of Anti-HBs titer after the primary infection. (C) An IAHBc case interpreted as “early loss of anti-HBs”. Despite the high titer of anti-HBs during primary infection, the level decreased rapidly within 3 years. Although anti-HBs was negative during tenofovir-sparing antiretroviral therapy, its level increased again after tenofovir-based regimen. (D) An IAHBc case interpreted as “late seroconversion”. Note the appearance and rise of anti-HBs to >100 IU/mL about 3 years after the primary infection. (E) An IAHBc case interpreted as a complex of “late seroconversion, early loss of anti-HBs, and viral rebound”. Note the appearance of Anti-HBs at 5 years after primary infection, and its disappearance after 2 years with occasional viral rebound throughout the observation period. (F) An IAHBc case interpreted as “viral rebound” by treatment interruption. Note that treatment interruption resulted in a rebound in both HIV and HBV, suggesting that IAHBc inferred poor immunity against HBV. (G) A case positive for anti-HBs alone interpreted as “possible resolution”. Due to lack of primary infection data, it was not clear how many years had passed since the primary infection. Note the persistently negative Anti-HBc after the first 4 years of appearance. The alternative diagnosis of late seroconversion could not be ruled out. (H) A case positive for anti-HBs alone interpreted as “post-exposure protective reaction”. Note the abrupt increase in Anti-HBs from 24.9 to 1000 IU/mL in the third year of observation, followed by its decrease to 37.5 IU/mL by the fifth year. Although the parallel rise in aminotransferase and anti-HBc suggested exposure to HBV, HBV-DNA was negative throughout the observation period. TDF: tenofovir disoproxil fumarate, FTC: emitricitabine, ATV: atazanavir, TDF: tenofovir disoproxil fumarate, DTG: dolutegravir, ABC: abacavir, 3TC: lamivudine, ETV: entecavir, TAF: tenofovir alafenamide, RAL: raltegravir.

https://doi.org/10.1371/journal.pone.0336924.g002

Discussion

The key findings of the present study are: 1) IAHBc indicates weak immunity, while “anti-HBs alone” strongly infers robust immunity, 2) none of the IAHBc group showed a “resolved pattern”, with fall and disappearance of anti-HBs years or decades after the initial seroconversion, and 3) unstable serostatus in the indeterminate group, highlighting the importance of repeated serologic tests to establish the correct diagnosis. These findings emphasize the fragility of HBV immunity in PLWH.

Although numerous studies have attempted clinical interpretation of IAHBc, several major guidelines and studies interpret IAHBc as resolved infection. Evidence supporting the concept of resolution includes: 1) a stable serostatus in long-term follow-up [2], 2) comparable HBV-specific T cell and memory B cell responses between the IAHBc and seroconversion groups [26], and 3) the relatively favorable vaccine response to HBV [17,27,28]. However, one study from Taiwan with extensive serological testing [3] reported the finding of unstable antibodies in IAHBc, which was also seen in our study. Interestingly, there are no studies that have examined T cell or memory B cell responses in PLWH, who are known to have significantly impaired T cell function [26]. In the present study, some PLWH with late seroconversion showed stable seroconversion, suggesting that IAHBc is a mixture of robust and weak/unstable immunity. Based on such heterogeneous HBV immunity in IAHBc, some groups offered “diagnostic HBV-vaccine booster” to identify PLWH with robust immunity in IAHBc [5]. However, it is noteworthy that: 1) none of the IAHBc cases showed a resolved pattern with gradual fall in anti-HBs after seroconversion; and 2) post-exposure protection was observed in only 1/118 case (1%) of IAHBc, whereas it was observed in 5/35 cases (14%) of anti-HBs alone. These findings suggest that IAHBc in immunodeficient individuals is rather more indicative of latent or chronic infection than resolved infection. Indeed, such concern is reflected in the practice guidelines for prevention of HBV reactivation in candidates for immunosuppressive therapy [29]. Furthermore, despite the lack of significant association with weak HBV immunity in logistic analysis, the clinical interpretation of lost anti-HBs/anti-HBc was strictly identical to that of IAHBc, suggesting that lost anti-HBs/anti-HBc also indicates weak HBV immunity.

The prevalence and significance of HBV serostatus can be affected by HBV epidemiology and genotypes [5]. For example, in Japan, HBV is highly endemic in MSM and PLWH [14] most likely due to delayed vaccine policy. Extrapolation of our findings may be limited to the economically developed countries, where HBV prevalence is low due to early induction of universal vaccination programs. In contrast, HBV remains endemic in most African, several Asian, and Eastern Mediterranean countries, largely due to the delayed implementation of universal vaccination programs [30,31]. Furthermore, although genotype C is traditionally predominant among elderly Japanese individuals [32], genotype A is more common in younger Japanese populations and PLWH [33]. In our study, among 85 PLWH with documented HBV genotypes—most of whom had chronic infections—47 (55%) were genotype A, 13 (15%) were genotype C, and 20 (24%) were unclassified (data not shown). Genotype A is distributed globally, particularly in Europe and South-East Africa [34,35]. Accordingly, the applicability of our findings may extend to select Asian, Eastern Mediterranean, and South-East African countries.

Various serological anomalies in HBV infection result from a complex interplay between viral persistence mechanisms [36] and host immune dysregulation, such as exhaustion and functional impairment of HBV-specific T cells and B cells [37,38]. In addition, one proposed explanation for IAHBc involves the consumption of anti-HBs through immune complex formation with their cognate antigens [39]. Since HBcAg is the most recognized immunogenic viral component, anti-HBc may persist long after acute infection [4,40]; however, its concentration also reflects latent HBV-DNA levels and is associated with hepatic disease severity [9,12]. Recent ex vivo findings by Wang et al. [29] offer insight to HBV-specific immune dynamics. Their data demonstrated suppression of intrahepatic covalently closed circular DNA (cccDNA) and HBV-DNA even in IAHBc cases, alongside preserved HBV-specific T-cell responses in individuals who had undergone seroconversion.

In non-PLWH, HBsAg can seldom be suppressed by nucleos(t)ide analogues (NAs) [41] because of host immune dysregulation. Meanwhile, in HIV/HBV coinfection, HBsAg can disappear by NAs [42] possibly due to the immune reconstitution by HIV treatments. In IAHBc, insufficient immune reconstitution hinders the production of sufficient anti-HBs to exceed elevated HBsAg, indicative of a substantial cccDNA reservoir. Conversely, in cases with anti-HBs alone, relatively preserved T/B-cell function may suppress HBsAg, but sometimes fail to mount a detectable response because of minimal HBcAg stimulation, which are likely suppressed by anti-HBV therapy (S1 Fig).

Recently, there has been a shift from conventional triple-agent antiretroviral therapy to dual therapy. However, switching to an anti-HBV agent-sparing therapy raises the concern of new HBV infection or, more importantly, reactivation [25,43]. Our study also found viral rebound in an IAHBc case due to interruption of anti-HBV agents (Fig 2F). Hence, in IAHBc and cases with lost anti-HBs/anti-HBc, careful interpretation is needed in switching ART to anti-HBV agents-sparing therapy. On the other hand, the risk of viral rebound after switching to tenofovir-sparing regimen in IAHBc and lost anti-HBs/anti-HBc is not clear. Although HBV infection can be prevented with 3TC monotherapy, resistant viruses could emerge in infected cases, whilst no new HBV infections or resistant HBV viruses have been reported in patients receiving TDF [44]. Tenofovir-containing regimens have better clinical outcomes than 3TC-containing regimens in liver-related mortality rate as well as virologic suppression [45,46]. Moreover, the combination of TAF and FTC, which increases intracellular concentration, resulted in a higher rate of HBsAg disappearance than TDF/FTC, but low seroconversion rates were noted in patients of both treatment groups (9% vs. 7%) [42]. These findings indicate that HBsAg suppression in PLWH is strongly dependent on anti-HBV therapy. For individuals with undetectable or low anti-HBs titers (e.g., < 100 IU/mL), we recommend the following approach: 1) Repeated serological testing: Concurrent assessment of HBsAg, anti-HBs, anti-HBc, and HBV-DNA should be performed at least two to three times to ensure diagnostic accuracy and capture potential fluctuations in serological markers. 2) Continuation of HBV-active ART: Maintenance of ART regimens containing agents with anti-HBV activity is advised, both to suppress HBV replication and to mitigate the risk of HBV reactivation.

Apart from the above strengths of our study, the study design has several limitations that need to be discussed. First, although IAHBc in immunocompetent individuals is expected to have different entities from those in immunodeficient people [17,26,47,48], the present study did not compare these groups of subjects. Second, we did not conduct long-term serologic reviews in cases with HBsAg(-)/anti-HBs(+)/anti-HBc(+) or HBsAg(+)/anti-HBs(-)/anti-HBc(+). Based on the highly changeable and different anomalies in HBV serostatus, it is possible that a percentage of such cases might have been missed despite transient indeterminate patterns. Third, one can argue that the 13-year observation period was not long enough to identify the resolved pattern. Our findings suggest that some of possible non-seroconversion can include resolved pattern, with waning of anti-HBs for longer than 20 years. Finally, although assay platforms for HBV markers changed over the 13-year duration of this retrospective study, we consider the impact on serostatus trends to be minimal. First, the lower limits of detection remained largely consistent across assay generations, and second, results obtained using Lumipulse (2003–2012) aligned closely with those from Architect, which was employed in a recent study covering 2013–2023 [45].

In conclusion, PLWH often show indeterminate HBV serostatus after HBV infection, with high instability chronologically, raising risks of misdiagnosis. IAHBc and lost anti-HBs/anti-HBc in PLWH indicate weak and unstable immunity rather than resolved infection, whereas anti-HBs alone infers robust immunity.

Key points: In people living with HIV, indeterminate serostatus of hepatitis B virus (HBV) can change widely, raising risk of misdiagnosis. Isolated anti-HBc and lost anti-HBs/anti-HBc do not indicate resolved but weak/unstable immunity against HBV, whereas anti-HBs alone infers robust immunity.

Supporting information

Acknowledgments

The authors thank the enrolled participants and all staff of the Outpatient Clinics of the Department of Laboratory Medicine and of the Clinical Laboratory Center, Tokyo Medical University Hospital.